USFDC Home  USF Electronic Theses and Dissertations   RSS 
Material Information
Subjects
Notes
Record Information

Full Text 
xml version 1.0 encoding UTF8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchemainstance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd leader nam 22 Ka 4500 controlfield tag 007 crbnuuuuuu 008 s2010 flu s 000 0 eng d datafield ind1 8 ind2 024 subfield code a E14SFE0004561 035 (OCoLC) 040 FHM c FHM 049 FHMM 090 XX9999 (Online) 1 100 Sullivan, JoAnn. 0 245 Characterization of an urban heat island (uhi) in the tampa region of florida h [electronic resource] / by JoAnn Sullivan. 260 [Tampa, Fla] : b University of South Florida, 2010. 500 Title from PDF of title page. Document formatted into pages; contains X pages. 502 Dissertation (PHD)University of South Florida, 2010. 504 Includes bibliographical references. 516 Text (Electronic thesis) in PDF format. 538 Mode of access: World Wide Web. System requirements: World Wide Web browser and PDF reader. 3 520 ABSTRACT: Numerous research studies have been conducted on the modification of weather and local climate by the urban environment. In studying the urban environment effects, researchers have investigated the urban heat island (UHI), anthropogenic cloud condensation nuclei, anthropogenic heat emissions and other factors. Many of these studies used data sampling networks, while other studies relied solely on computer modeling. This research has taken an approach between the sampling network studies (which were often limited in spatial density) and the pure computer model studies (which lacked the benefits of observational data) to investigate the Tampa Bay Region UHI. The research utilized inexpensive commercially available temperature logging sensors within a 525 km2 study area. One hundred temperature logging sensors, deployed within the study area, generated in excess of 250,000 time and temperature data points for analysis. The large number of temperature sensors enabled the generation of detailed spatiotemporal maps of the Tampa Bay Region UHI. Analysis of the data revealed a significant relationship between the percentage of impervious surface in the study area and the intensity of the local UHI delta temperatures. In addition, the analysis identified the existence of micro UHIs within residential areas. These micro UHIs affected readings within the residential areas. In conjunction with the investigation of the relationship between the percentage of impervious surface and the generation of a UHI, wind speed's role as a moderating factor was also investigated. It was found that increases in wind speed are correlated with a lessoning of the observed UHI. Wind speeds above approximately 2 ms1 exhibit a significant negative relationship to the development of a UHI. The results of this study add to the field of UHI research in subtropical environments. 590 Advisor: Jeniffer Collins, Ph.D. 653 Climate Urban Impervious UHI LULC 690 Dissertations, Academic z USF x Geography & Env Sci & Policy Masters. 773 t USF Electronic Theses and Dissertations. 4 856 u http://digital.lib.usf.edu/?e14.4561 PAGE 1 Characterization of an Urban Heat Island (UHI) in the Tampa Region of Florida by JoA nn Sullivan A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Geography College of Art s and Sciences University of South Florida Major Professor: Jennifer Collins Ph D Robert Brinkmann Ph D Ruiliang Pu Ph D Steven Reader Ph D Amy Stuart Ph D Date of Approval: May 7, 2010 Keywords: climate, urban, impe rvious UHI, LULC Copyright 2010 JoAnn Sullivan PAGE 2 ii A CKNOWLEDGEMENTS Thank you to all of you that have given me assistance and guidance over the past fe w years. Dr. Jennifer Collins my advisor has been instrumental in helping me through both the good times and the ro ugh times. She provided me inspiration and guidance in navigating the course towards my degree. She has shown great patience and understanding. I would also like to thank the members of my committee, Dr. Robert Brinkmann, Dr. Steven Reader, Dr. Ruiliang Pu and Dr. Amy Stuart for their feedback and their advi c e during my research. A special thanks goes out to the personnel at the Tampa Electric company for the usage of their light pole s during my research. Last but not least thanks to my fellow graduate st udent s in the D epartment of Geography at the Uni versity of South Florida for their camaraderie PAGE 3 i TABLE OF CONTENTS LIST OF TABLES iv LIST OF FIGURES vi ABSTRACT xii 1. 0 INTRODUCTION 1 1.1 Early studies detailing urban environment effects on weather and 4 c limate 1.2 Underlying physical processes 9 1.3 Density of sensors in UHI studies 14 1.4 Increases in computing power and resources 15 1.5 Satellite imagi ng of the environment 17 1.6 Land Use Land Cov er (LULC) in the study of UHI s 19 1.7 Wind speeds and the UHI 22 1.8 UHI societal i mpacts 28 1. 9 Research q uestions 29 1. 10 Hypotheses 30 2. 0 METHODS 32 2 .1 Research a rea 36 2.2 Temperature d ata l ogging s ensor p lacement 40 2.3 Temperature d ata l ogging s ensors 48 2.4 Temperature d ata l ogger program s etup 54 2.5 Sensor data r eduction 56 PAGE 4 ii 2.6 Sensor p roxim ate percent impervious s urface 57 2.7 Wind speed data c ollection 60 2.8 Analysis m ethods 61 3. 0 DETAILS OF THE URBAN AND RURAL TEMPERATURES AND UHI 62 3.1 Introduction 62 3.2 Data collection 62 3.3 91 3.4 Commercial, i ndustrial and r esidential UHI comparisons 118 3.5 Summary 124 4. 0 RELATIONSHIP OF IMPERVIOUS SURFACE TO THE UHI 125 4.1 Summary 141 5. 0 RELATIONSHIP OF WIND SPEED TO UHI 143 5.1 Sea breeze wind s a ffect on UHI 149 5.2 Summary 153 6 0 CONCLUSIONS 154 6.1 Perceived weaknesses of this study 15 7 6.2 Areas of potential future work 159 7 0 REFERENCES 161 APPENDICES Appendix A Thermochron DG1921H specification s 169 Appendix B Delta t emperature 0 F valu es at sensor locations 171 Appendix C Correlation values of impervi ous percent to time period delta 190 temperature PAGE 5 iii Appendix D 2007 Period OLS regression results with constant value 195 Appendix E 2007 Period OLS regression results without a co nstant value 244 Appendix F 2008 Period OLS regression results zero crossing 293 Appendix G Aggregated 2007 and 2008 Period OLS regression results 342 without constant ABOUT THE AUTHOR End Page PAGE 6 iv L IST OF TABLES Table 1.1 Percentage of growth in population of the counties in the 3 Tampa Bay Region from 1950 to 2000 Table 1.2 Thermal properties 13 Table 1.3 T ypical surface roughness value z 0 for urbanized terrain 22 Table 1.4 Heat island intensity and the corresponding wind speeds in 23 the G reater Paris area Table 3.1 Sample raw data download 63 Table 3.2 E xample of c alculated 30 minu te sample mean temperatures o F 64 Table 3.3 Urban and rural daily tem perature descriptive statistics 66 Table 3.4 ArcMap attribute table including o F) and impervious 91 surface values ( percent ). Table 4.1 Correlation results 126 Table 4.2 r egression model with constant descriptive statistics 128 Table 4.3 r egression model with constant summary 128 Table 4.4 r egression model with constant ANOVA 128 Table 4.5 r egression model with constant coefficients 128 Table 4.6 r egression model with constant residual statistics 129 Table 4.7 2007 r egression model without constant summary 1 3 1 Table 4.8 2007 r egression model without constant ANOVA values 131 Table 4.9 2007 r egression model without constant coefficients 131 Table 4.10 2007 r egression model without constant residual stat istics 131 PAGE 7 v Table 4.11 2008 r egression model without constant summary 133 Table 4.12 2008 r egression model without constant coefficients 133 Table 4.13 2008 r egression model without constant coefficients 13 3 Table 4.14 2007 and 2008 c ombined regression model without constant 134 summary Table 4.15 2007 and 2008 c ombined regression model without constant 134 ANOVA values. Table 4.16 2007 and 2008 c ombined regression model without constant 134 coefficients Table 5.1 Q uadratic m odel s ummary 145 Table 5.2 Quadratic c oefficients 145 Table 5.3 Natural l og m odel s ummary 146 Ta ble 5. 4 Natural l og ANOVA values 1 46 Table 5.5 Natural l og c oefficient s 146 Table 5.6 2008 Q uadratic model summary 148 Table 5.7 2008 Anova values 148 Table 5.8 2008 Coefficients summary 148 PAGE 8 vi LIST OF FIGURES Figure 1.1 Flo rida population growth between the years 1900 and 2000 3 Figure 1.2 METROMEX instrumentation 7 Figure 1.3 Cross section view of an UHI 10 Figure 1.4 Graphic representa t ion of surface energy balance 11 Figure 1.5 Florida s ea breeze flow 26 Figure 2.1 Research design flow 33 Figure 2.2 Aerial image of the Tampa Bay R egion 36 Figure 2.3 Tampa study area boundaries 38 Figure 2.4 Study area in relation to the Greater Tampa Bay Regio n 39 Figure 2.5 Rural se nsor number 99 mounting pole 40 Figure 2.6 Area near rural sensor number 99 41 Figure 2. 7 2009 image of the rural sens or nu mber 1 surrounding area 42 Figure 2. 8 2009 image of the area surroun ding rural se nso r number 100 43 Figure 2.9 Hawths a nalysis toolset for ArcGIS 44 Figure 2.10 Hawths a nalysis tool random point g eneration 44 Figure 2.11 Temperature data logging sensor p lacement 4 7 Figure 2.12 C ommercially available data log ging weather station 48 Figure 2.13 Ther mo chron sensor relative size 52 Figure 2.14 Typical sensor installation on a TECO utility pole 53 Figure 2.15 Scanning Devices programming sample page 55 PAGE 9 vii Figure 2.16 Percentage of impervious surface at temperature logging sensor 59 locations Figure 3.1 24 hour study area mean urban and rural near surface air temperatures 65 Figure 3.2 Near surface air temperatures 00:00 67 Figure 3.3 Near surface air temperatures 00:30 67 Figure 3.4 Near surface air temperatures 01:00 68 Figure 3.5 Near surface air temperatures 01:30 68 Figure 3.6 Near surface air temperatures 02:00 69 Figure 3.7 Near surface air temp eratures 02:30 69 Figure 3.8 Near surface air temperatures 03:00 70 Figure 3.9 Near surface air temperatures 03:30 70 Figure 3.10 Near surface air temperatures 04:00 71 Figure 3.11 Near surface air temperatures 04:30 71 Figure 3.12 Near surfa ce air temperatures 05:00 72 Figure 3.13 Near surface air temperatures 05:30 72 Figure 3.14 Near surface air temperatures 06:00 73 Figure 3.15 Near surface air temperatures 06:30 73 Figure 3.16 Near surface air temperatures 07:00 74 Figure 3.17 N ear surface air temperatures 07:30 74 Figure 3.18 Near surface air temperatures 08:00 7 5 Figure 3.19 Near surface air temperatures 08:30 75 Figure 3.20 Near surface air temperatures 09:00 76 Figure 3.21 Near surface air temperatures 09:30 76 Figur e 3.22 Near surface air temperatures 10:00 77 PAGE 10 viii Figure 3.23 Near surface air temperatures 10:30 77 Figure 3.24 Near surface air temperatures 11:00 78 Figure 3.25 Near surface air temperatures 11:30 78 Figure 3.26 Near surface air temperatures 12:00 79 Figure 3.27 Near surface air temperatures 12:30 79 Figure 3.2 8 Near surface air temperatures 13:00 80 Figure 3.29 Near surface air temperatures 13:30 80 Figure 3.30 Near surface air temperatures 14:00 81 Figure 3.31 Near surface air temperatur es 14:30 81 Figure 3.32 Near surface air temperatures 15:00 82 Figure 3.33 Near surface air temperatures 15:30 82 Figure 3.34 Near surface air temperatures 16:00 83 Figure 3.35 Near surface air temperatures 16:30 83 Figure 3.36 Near surface air t emperatures 17:00 84 Figure 3.37 Near surface air temperatures 17:30 84 Figure 3.38 Near surface air temperatures 18:00 85 Figure 3.39 Near surface air temperatures 18:30 85 Figure 3.40 Near surface air temperatures 19:00 86 Figure 3.41 Near surf ace air temperatures 19:30 86 Figure 3 .42 Near surface air temperatures 20:00 87 Figure 3.43 Near surface air temperatures 20:30 87 Figure 3.44 Near surface air temperatures 21:00 88 Figure 3.45 Near surface air temperatures 21:30 88 PAGE 11 ix Figure 3.46 Near surface air temperatures 22:00 89 Figure 3.47 Near surface air temperatures 22:30 89 Figure 3.48 Near surface air temperatures 23:00 90 Figure 3.49 Near surface air temperatures 23:30 90 Figure 3.50 Rural urban delta temperatures 00:00 93 Fi gure 3.51 Rural urban delta temperatures 00: 30 93 Figure 3.52 Rural urban delta temperatures 01:00 94 Figure 3.53 Rural urban delta temperatures 01:30 94 Figure 3.54 Rural urban delta temperatures 02:00 95 Figure 3.55 Rural urban delta temp eratures 02:30 95 Figure 3.56 Rural urban delta temperatures 03:00 96 Figure 3.57 Rural urban delta temperatures 03:30 96 Figure 3 .58 Rural urban delta temperatures 04:00 97 Figure 3.59 Rural urban delta temperatures 04:30 97 Figure 3. 60 Rural urban delta temperatures 05:00 98 Figure 3. 61 Rural urban delta temperatures 05:30 98 Figure 3. 62 Rural urban delta temperatures 06:00 99 Figure 3.63 Rural urban delta temperatures 06:30 99 Figure 3. 64 Rural urban delta temperatures 07:0 0 100 Figure 3. 65 Rural urban delta temperatures 07:30 100 Figure 3. 66 Rural urban delta temperatures 08:00 101 Figure 3. 67 Rural urban delta temperatures 08:30 101 Figure 3. 68 Rural urban delta temperatures 09:00 102 PAGE 12 x Figure 3. 69 Rural urban delta temperatures 09:30 102 Figure 3. 70 Rural urban delta temperatures 10:00 103 Figure 3. 71 Rural urban delta temperatures 10:30 103 Figure 3. 72 Rural urban delta temperatures 11:00 104 Figure 3. 73 Rural urban delta temperatures 11:30 1 04 Figure 3. 74 Rural urban delta temperatures 12:00 105 Figure 3. 75 Rural urban delta temperatures 12:30 105 Figure 3. 76 Rural urban delta temperatures 13:00 106 Figure 3. 77 Rural urban delta temperatures 13:30 106 Figure 3. 78 Rural urban delta t emperatures 14:00 107 Figure 3. 79 Rural urban delta temperatures 14:30 1 07 Figure 3. 80 Rural urban delta temperatures 15:00 108 Figure 3. 81 Rural urban delta temperatures 15:30 108 Figure 3. 82 Rural urban delta temperatures 16:00 109 F igure 3 83 Rural urban delta temperatures 16:30 109 Figure 3. 84 Rural urban delta temperatures 17:00 110 Figu re 3. 85 Rural urban delta temperatures 17:30 110 Figure 3. 86 Rural urban delta temperatures 18:00 111 Figure 3. 87 Rural urban delta temperatu res 18:30 111 Figure 3. 88 Rural urban delta temperatures 19:00 112 Figure 3. 89 Rural urban delta temperatures 19:30 112 Figure 3. 90 Rural urban delta temperatures 20:00 113 Figure 3. 91 Rural urban delta temperatures 20:30 113 PAGE 13 xi Figure 3. 92 R ural urban delta temperatures 21:00 114 Figure 3. 93 Rural urban delta temperatures 21:30 1 14 Fi gure 3.94 Rural urban delta temperatures 22:00 11 5 Figure 3. 95 Rural urban delta temperatures 22:3 0 11 5 Figure 3. 96 Rural urban delta temperatures 2 3:00 116 Fi gure 3. 97 Rural urban delta temperatures 23:30 116 Figure 3. 98 o C 120 Figure 3. 99 o C 122 Figure 3. 100 o C 123 Figure 3.101 24 hour 124 Figure 4.1 Correlation val u es at differing radii values 126 Figure 4.2 Correlation values of 50 meter impervious surface and period 127 Figu re 4.3 Histog ram of r egression model with constant standardized residuals 130 Figure 4. 4 Plot of the r egression model with constant standardized residual s 130 Figure 4.5 Histogram of r egression model without constant standar dized residuals 132 Figure 4.6 Morans I for various distance weighting 1 36 Figure 4.7 Lag residuals Morans I calculation 137 Figure 4.8 Lag residuals Morans I permutation run 137 Figure 4.9 r egression model witho ut constant R 2 and 1 values 140 Figure 5.1 (2007) 143 Figure 5.2 Curve fit p lots 144 PAGE 14 xii Figure 5 .3 Tampa International Airport wind speed and direction 150 6/1/2 007 to 7/17/2007 Figure 5.4 MacDill AFB wind speed and dir ection 6/1/2007 to 7/17/2007 151 Figu re 5.5 Fort Howard P ark wind speed and direct ion 6/1/2007 to 7/17/2007 1 51 Figure 5.6 2008 Tampa International Airpor t wind speeds and direction 152 PAGE 15 xiii Characterization of an Urban Heat Island (UHI) in the Tampa Region of Florida JoAnn Sullivan ABSTRACT Numerous research studies have been conducted on the modification of weather and local climate by the urba n environment. In studying the urban e nvironment effects, researchers have investigated the urban heat island (UHI), anthropogenic cloud condensation nuclei, anthropogenic heat emissions and other factors. Many of these studies used data sampling networks, while other studies relied solely on computer modeling. This research has taken an approach between the sampling network studies (which were often limited in spatial density) a nd the pure computer model studies (which lacked the benefits of observational data) t o investigate the Tampa Bay R egion UHI The research utilized inexpensive commercially availa ble temperature logging sensors with in a 525 km 2 study area One hundred temperature logging sensors, deployed within the study area, generated in excess of 250,000 time and temperature data points for analysis. The large number of temperature sensors enabled the generation of detailed spatiotem poral maps of the Tampa Bay R egion UHI. A nalysis of the data revealed a significant relationship between the percentage of impervious surface in the st udy area and the intensity of the local UHI delta temperature s. In addition the analysis identi fied the existence of micro UHI s within residential areas. T hese micro UHIs affected readings within the residential areas. PAGE 16 xiv In conjunction with the investigati on of the relationship between the percentage of impervious surface and the generation of a UHI, wind speeds role as a moderating factor was also investigated. It was found that increases in wind speed are correlated with a lessoning of the observed UHI W ind speeds above approximately 2 ms 1 exhibit a significant negative relationship to the develo pment of a UHI The results of this study add to the field of UHI research in subtropical environment s PAGE 17 1 1.0 INTRODUCTION For centuries mankind has be en trying to influence the weather, particularly precipitation. In the 1800s many so called rainm akers plied their trade through out the parched lands of the Great Plains of the United States (Clark 1980) Coming into towns with boasts of their rain making prowess, they pointed their cannons and rockets at the heavens and assaulted the sky. It was not until the dawn of the 20 th C entury when early researche r s began to notice areas of increased precipitation near several larger cites, that it was hypothesize d that humans, with their propensity to congregate in urb an areas, were indeed influencing the weather. What the rainmakers of the plains tried to do was already inadvertently bein g accomplished by the cities that followed the plow. Within the last severa l years there has been an increased effort to better understand the relationship of the ever increasing urban landscape and its potential effect on climate (Grimmond 2006). As recent ly as the summer of 2008, scientists, in conjunction with the European Sp ace Agency (ESA), conducted the Desirex 2008 study to evaluate the usefulness of satellite and aircraft borne instrumentation to help reduce heat wave deaths and potentially mitigate the urban effect on summertime heat waves (ESA 2008) In June of 2008 th e National Weather Service (NWS) in Phoenix Arizona investigated the possible effects of the urban landscape on the summer monsoon season (NWS 2008). Increasingly the international study o f urban induced heat islands is taking on added importance. China, with a population of over 1 billion, has seen a rapid growth in its urban areas. Recent studies by Wang and Hu (2006), Hau et al (2007), Dan et al. (2009) PAGE 18 2 and Li et al. (2009) have investigated and document ed the development and expansion of urban heat islands ( UHI s ) in China. Studies by governmental agencies and independent researchers demonstrate the increased attention that is being paid to the potential effects of the urban environment on the climate. With the continued influx of people into urban ar eas (which according to the UN Population Division (2006) has been occurring in the United States at a rate of 1 percent per year since 1975) and the resultant expansion of the urban landscape the modification of local climate by the urban land form is li kely to continue to expand as well. A case in point is the state of Florida. Figure 1.1 depict s the change in population between the period 1900 to 2000 in the state of Florida D uring the latter half of the twentieth century between the years 1950 to 200 0 the population of the state of Florida increased by 475 percent with much of this growth occurring in counties that border the Gulf of Mexico or the Atlantic Ocean. The counties in the Tampa Bay Region grew at similar rates. Table 1.1 lists the populati on growth rates of the counties in the Tampa Bay Region for the period between 1950 and 2000. PAGE 19 3 Year 2000 1990 1980 1970 1960 1950 1940 1930 1920 1910 1900 Florida Population 20,000,000 15,000,000 10,000,000 5,000,000 0 Figure 1.1 Florida population growth between the year s 1900 and 2000 (U.S. Census 2000 ) Counties of the Tampa Bay Region Population growth 1950 2000 in per cent Hillsborough Pinellas Pasco Polk 300 480 1575 290 Table 1.1 Percentage of growth in population of the counties in the Tampa Bay Region from 1950 to 2000 (U.S. Census 2000 ) A vast majority of the people who moved to Florida during this period resi ded in coastal urban area s What effects might this large influx of people have had on rural lands? An examination of t he period between 1990 and 2000 reveals that the overall population of Florida increased by 25.53 percent. The population of the four c ounties of the Tampa B ay Region increased at similar rate s Utilizing United States Geologic PAGE 20 4 Survey (USGS) impervious surface maps of the Tampa Bay Region for the period between 1991 and 2001 the area of pristine rural land having no impervious surface wi thin the Tampa Bay Region declined by 25.43 percent closely mirroring the increase in population The expansion of the population in Florida and the Tampa Bay Region clearly impacts the impervious surface in the state. How this general e xpansion of urban areas and Florida urban areas in particular affect the climate in and around these urban areas are of critical importance in light of recent projections of gl obal warming and climate change (Cotton and Pielke 2007, Changnon 1992, Titus 1992). In the case of Florida Roth (2007) reviewed the lite rature on urban climate change and found that less than 20 percent of urban climate research has been conducted in tropic al or sub tropic al areas and those studies that were conducted were mostly descriptive in natur e. Not only are large shifts in population occurring in the sub tropic al regions within the United States, worldwide the tropic al and sub tropic al regions are seeing tremendous urban expansions (Roth 2007). There is a need to more fully understand the effec ts of the urban environment on local climate patterns and the resultant changes in local temperatures and precipitation particularly in sub tropic al regions such as Florida. 1.1 Early studies detailing urban environment effects on weather and climate The detailed study of the urban environment e ffects on weather and climate did not really begin until the start of the 20 th century. Early studies by Fassig ( 1907), Saucier ( 1949), and Lig da and Bigler ( 1956 ) helped to est ablish the hypothetical down wind eff ect of cities while s tudies by Woollum and Canfield ( 1968) and Loose and Bornstein PAGE 21 5 (1977) helped in localizing the hypothesized down wind e ffect of cities. It is generally accepted within the scientific community that the first usage of the term urban he at isla nd can be attributed to Gordon Manley (Manley 1958). Manley examined the frequency of snowfall in metropolitan England. He noted that there were differences in the recorded temperatures between the rural and nearby metropolitan areas The metropoli tan areas were warmer which affected the frequency of snow fall with decreased events To describe these differences between rural and metropolitan temperatures he coined the term urban heat island Within the literature, the UHI of a study area is comm only reported as a single maximum delta temperature ( ) value (Yow and Carbone 2006) or in some cases as semi concentric circles emanating from the Central Business District ( CBD ) (Unger et al 2001). As the evidence for the modification of weather by urban areas began to amass Changnon ( 1968) produced a report on the La Porte precipitation anomaly down wind of Chicago Changnon was able to correlate the relationship between anthropogenic cloud condensation nuclei (CCN) and prevailing wind conditions with an increase in precipitation in La Porte, India na Changnon examined the historical meteorological records of the La Porte weather station in Northwest Indian a for the period between 1935 and 1965 He found the existence of unusual weather conditions which were poten tially indicative of a modification of the atmosph ere and precipitation due to it s proximity to the Chicago urban area. Starting in 1935 and continuing until the mid 1960 s the La Porte weather station precipitation and thunder values remained 30 to 40 percent higher than the surrounding ar eas. Changnon found a good correlation between the anthropogenic induced smoke haze days in Chicago and the precipitation at the La Port e PAGE 22 6 weather station. The existence of a relationship between anthropogenic atmospheric pollutants and weather modification caused considerable concern in the scientific community and the public in general The report by Changnon (1968) The La Porte Anomaly Fact or Fiction focused research interests on the effects of the urban environment on weather modification by cities. T he Changnon La Porte study brought about a new understanding that urban areas were having an effect on the weather. The Changnon study helped spur subsequent research in this area and was a driving force in the development of the ground breaking St. Lo uis METROMEX study. In t he wake of the Changnon report three years later four research teams came together in St. Louis, Missouri to study the weather modification induced by the urban industrial environment (Changnon 1971). As a result t he St. Louis METROMEX field project was undertaken. Prior to the METR OMEX project, urban related weather changes had been studied only to a limited degree The St. Louis METROMEX field project was more ambitious and gat hered data from 1971 until 1976. In o rder to mor e fully document the effects that urban areas were having on weather modification the METROMEX study employed a dedicated netwo rk of sensors over a period of five years. Figure 1 .2 is reconstructed from Changnon and Huff ( 1971) and highlights the instru mentation employed during the METROMEX study As part of the METROMEX study 220 rain gauges and hail pads were distributed evenly over an area of 5700 km 2 in and around St. Louis, Missouri. Three radar sites were employed to gather first ech o rain data in the study area while r adiosondes were deployed to gather PAGE 23 7 data on the upper atmospheric conditions In addition aerosol samplers were used to de tect CCN concentrations and air pollutants in and around the city. Figure 1 .2 METROMEX instrumentation The St. Louis METROMEX study was the first research effort to try and adequately document the weather modification induced by an urban area Some of the key findings Surface Operations Ground Based Vertical Operations Aircraft Operations Fixed Networks Mobile Operations Radar Measurement Other Rain gages Rain water samplers Hygrothermo graphs Microbaro graphs Wind sets Raindrop spectrometer Hail sensors Thunder detectors Atmospheric electricity samplers Aerosol samplers Temperature Precipitation Humidity Aerosols Wind Rain Samplers FPS 18 TPS 10 TPS 10 Cloud cameras Radiosonde stations Pibal stations Instrumented aircraft Cloud drop spectra Precipitation Vertical air motion Water content Temperature Humidity Radiation Nuclei Aerosol samples Chaff St Louis METROMEX Instumentatuion PAGE 24 8 of the study were validated by the work of Loose and Bornstein (1977) that showed a convergenc e of surface winds as they passed around an urban area. METROMEX also identified changes in the surface energy budget when rural lands are converted to urban impervious surface s The rural areas, with their moist soils and vegetation, convert a large porti on of the surface absorbed heat into latent heat of evaporation and transpiration while the urban areas with much less vegetation convert more of the absorbed surface heat in to sensible heat. This results in an urban area that is warmer than the surroun ding rural area. During METROMEX St. Louis was found to have a well developed UHI centered over the commercial district with the maximum expression of the UHI occurring between the hours of 00:00 and 06:00. It was also found that increased sensible heat in the urban area caused an increase in the depth of the urban boundary layer. This deepening of the urban boundary layer set s up conditions similar to the sea breeze experienced in Florida H owever in St Louis a country breeze developed with surface a ir coming in from the surrounding rural area resulting in convergence and uplifting of air near the city center. This uplifting of moist air from the rural area lead s to increased cloud depth and increased precipitation over and downwind of St. Louis. It i s interesting to note that while earl ier research by Changnon (1968) correlated the increase in precipitation in La Porte, Indiana with the smoke/haze days in nearby Chicago, METROMEX concluded that CCN did not play a major role in the St Louis urban prec ipitation anomaly. METROMEX proved to be a watershed moment in the study of the urban environment. The following are the St. Louis METROMEX study findings as summarized by Changnon et al ( 1977, p i ) PAGE 25 9 Key climatic effects are increased cloudiness (+10%), in creased total summer rainfall (+30%), and increased severe storm activity (+100%). These increases occur over the city and 10 to 25 miles beyond (east) the urban industrial areas. The urban induced anomalies occur most often with squall lines and cold fron ts; they maximize in the afternoon and again at night (2100 2400); they appear to be as active in dry periods as in wet periods. Impacts include more runoff, but also more local flooding, soil erosion, silting, and water pollution. The effect of altered we ather leads to a 3 to 4 percent average increase in local crop yields I t should be noted that t he METROMEX study initially employed just 7 weather stations to record temperature data. This was increased to 25 temperature reporting weather stations later in the study period While the METROMEX study concentrated on the modification of precipitation patterns attributable to the St. Louis urban area, the low number of temperature recording weather stations in the large survey area may have limited their abi lity to accurately identify the complex spatiotemporal patterns of the UHI 1.2 Underlying physical processes Since the METROMEX study in the 1970 s, numerous other studies have been undertaken examining the effects of urban environme nts on weather, with s ignificant attention being given to the concept of the UHI and the underlying physical processes According to Kalnay and Cai ( 2003) u rban surfaces can have significant effects on land surface temperatures (LSTs) The buildings, roads, and other paved sur faces in urban areas usually have higher solar radiation absorption, a greater thermal conductivity and greater capacity for releasing heat stored during the day at night. This generally leads to a modified climate that is warmer than the surrounding rura l area. PAGE 26 10 A UHI is defined where the near surface temperatures in the urban area are higher than temperatures in the rural area ( NWS 2010) Figure 1.3 is a cross sectional view of the temperature profile of a typical UHI. One should note that there are fl uctuations in the surface air temperature s across the profile. These can be attributed to the differing land cover types that are present in the area. One should also note that the greatest temperature differential exists between the rural and downtown ar ea. Fig ure 1.3 Cross section view of a UHI ( U.S. EPA www.epa.gov/hiri/about/index.htm ) In order to better understand what is driving the temperature differences bet ween the rural and urban areas, it is necessary to examine th e underlying physical pro cesses. Referencing Pielke (198 4) and Cotton and Pielke ( 2007 ), surface energy budgets are examined as they relate to land use changes that occur w hen rural areas are converted for urban uses. The modification of the variables in the surface energy budget equation ( Figure 1.4) provides a means of identifying the mechanism for the development of the PAGE 27 11 UHI. According to Cotton and Pielke (2007) t he basic surface energy (Equations 1 and 2) and moisture budget (Equation 3) equations can be written as: ) ( TR E L H Qg Rn (1) w here Rn = net radiant fluxes at the Earths surface Qg = soil heat flux H= turbulent sensible heat flux L(E+TR) = latent heat of vaporization flux E= evaporation flux of water TR= transpiration flux of wate r Rn can be further expanded: LW LW Q Q A Qs Rn ) 1 ( (2) w here Rn = net radiant fluxes at the Earths surface Qs = solar insolation flux A= surfa ce albedo LW Q = downward long wave radiation flux LW Q = upward long wave radiation flux I RO TR E P (3 ) w here P= precipitation E= evaporation TR= transpiration RO= run off I= infiltration Advection & Convection H Qg Conduction / Heat storage L(E+T) Evaporation & Transpiration R n Physical processes Radiation Fi gure 1.4 Graphic representation of surface energy balance. PAGE 28 12 Equations (1) and ( 3 ) do not act independent ly of each other. I f the amount of precipitation is h eld constant, an increase in RO ( as might be expected when rural soils are replaced with pave ment in the urban landscape ) leads to a decrease in ( E+TR). Working back into Equation (1), with Rn and Qg held constant, the decrease in (E+TR) will result in a low er ing of the latent heat of vaporization flux with a resultant increase in the sensible heat flux In addition, with increased RO there can be a decrease in I which affects Qg With decreases in moisture content the ability of the soil to retain heat is lowered which lowers Qg also resulting in an increase in sensible heat flux The increase in imper vious surfaces associated with urban development of rural lands, results in increased RO. This is one of the mechanism s that can lead to the development of a UHI. Another mechanism involved in the generation of a UHI is the decrease in surface albedo as sociated with urban surfaces (asphalt, concrete, roofing materials etc ). If for the purposes of this examination we assume that LW Q and LW Q remain constant and that the solar insolation flux Qs is constant, it can be seen that a decrease in albedo will result in a net increase in Rn Again working with Equation (1) holding Qg and (E+TR) constant the increase in Rn associated with the change in albedo will result in an increase in the sensible heat flux It shoul d be noted that there can be a significant difference between the daytime rural and daytime urban temperatures. H owever empirically it has been found by Oke ( 1982 ) Bornstein and Lin ( 2000 ) and Dixon and Mote ( 2003) that the maximum UHI is not expressed u ntil the nocturnal hours. This temporal shift of the UHI can be attributed to the differing heat capacities of the rural and urban environments. It has also been shown PAGE 29 13 that the UHI is most prevalent in the summer months and much less so in the winter month s One of the mechanisms that lead to the development of a UHI is the differing values of thermal inertia (heat storage) of the materials that make up the fabric of the urban and rural environments. Referencing Chen et al (2008 p 1724 ) T hermal inertia i s one of the typical parameters for describing the thermal characteristics of a material, it is a physical parameter representing the ability of a material to conduct and store heat, and in the context of planetary science, it is a measure of the sub surfa ces ability to store heat during the day and reradiate it during the night Thermal inertia is defined as : c P (4 ) w here P = thermal inertia = thermal conductivity of the material = material density c = specific h eat capacity Table 1.2 lists the specific thermal properties for concrete and dry sand. Material Th ermal Conductivit y k Wm 1 K 1 Density kg m 3 Specific Heat Capacity c Jkg 1 K 1 Concrete 1.4 2300 880 Dry Sand 0.42 16 02 840 Tab le 1.2 Thermal properties Examining two materials, one characteristic of the urban environment, concrete, and one characteristic of the rural environment in the Tampa study area, dry sand material heat storage can be compared. Concrete P = ) 880 )( 2300 )( 4 1 ( = 1683 W m 2 K 1 s (5 ) Dry Sand P = ) 840 )( 1602 )( 42 0 ( = 751 W m 2 K 1 s (6 ) PAGE 30 14 It can be seen that concre te has a greater thermal inertia value than dry sand. The above thermal inertia values indicate that the rural sand will release its stored heat quicker and hence will cool quicker than the urban concrete. During the nocturnal hours the greater thermal in ertia of concrete means that the urban environment will be warmer than the rural environment resulting in a nocturnal UHI. The configuration of impervious surfaces in the urban environment can also contribute to the magnitude of the UHI. The vertical sur faces in the urban canyon add additional sensible heat to the UHI in the CBD. It should be noted that urban canyons with their vertical development can present a surface area that is larger than seen by remote sensing satellites and depicted on the USGS im pervious surface maps used in this study. Additionally anthropogenic urban heat sources such as air conditioning can also provide an input to the UHI. All of these additional factors make the calculation of urban surface energy budget s difficult. As note d by Roberts et al (2006 ) due to the complex array of materials and the three dimensionality of the urban environment there is no accepted standard to determine urban heat storage, because there is no instrument or technique available to give correct values against which to calibrate other methods 1.3 Density of sensors in UHI studies Using a more traditional sampling approac h to investigate the UHI as noted earlier in the METROMEX study Bornste in and Lin ( 2000 ) studied summertime thunderstorms in Atlanta, Georgia and found that the Atlanta UHI induced a convergence zone that initiated thunderstorms. Dixon and Mote ( 200 3) also investigated Atlanta, Georgia and found similar results. One limiting factor in the more traditional sampling method u sed PAGE 31 15 in these earlier studies is the size of the temperature sampling network The Dixon and Mote study utilized 10 recording weather stations situated o ver a 21 county area while Bornstein and Lin utilized 42 recording weather stations spread over a 105,00 0 km 2 study area. The number of temperature sensors employed in these two studies allowed the researchers to establish a general perspective on the Atlanta UHI and its effects However i n order to obtain a more detailed understanding of the spatial distri bution of the UHI of Atlanta, Georgia a larger temperature sampling network would be desirable. The low number of temperature recording stations that seem to characterize many UHI studies may be attributable to the generally high cost of recording station s and the logistic complexity of deploying the stations in the field. Recently, researchers have deploy ed denser sensor networks to study the UHI. A case in point, as detailed by Bassara et al. (2009), is the new Oklahoma City micronet. Thirty five weat her monitoring stations were deployed atop light fixtures and traffic signals in downtown Oklahoma City and the surrounding suburbs. While an advancement in sensor density, the high cost of advanced weather stations and the logistics required to install a nd monitor them does not make this a viable approach which can b e easily applied to other areas. 1.4 Increases in computing power and resources Numerous other research methods have been undertaken in an effort to quantify the UHI of the urban environment With increases in computing power and resources, researchers have looked towards modeling as a means of expanding their understanding of urban environments. Utilizing a numerical model to parameterize the effects of the PAGE 32 16 urban environment, Grimmond and Ok e (1999) were able to predict the one dimensional temporal variability of heat fluxes in urban areas. In a study by Mihalakakou et al (2001), the researchers accurately modeled urban characteristics such as the surface energy budget, daytime heat island i ntensity, and surface roughness of the urban environment. In a similar vein, Marshall et al (2004) used a numerical model to evaluate the impact of anthropogenic land cover change in Florida. Marshall ran his model with reconstructed pre 1900 natural lan d cover and then replaced this with 1993 land use data. Marshall (2004 p 28) found that, The simulated spatial patterns of the surface sensible and latent heat flux were altered significantly, resulting in changes in the structure and strength of climatol ogically persistent, surface forced mesoscale circulations. Utilizing computer models to investigate the UHI, several researchers such as Myrup (1969), Bottyan and Unger (2003), Mason (2005) and Roberts et al (2006) developed multiple linear statistic al computer models to estimate the mean maximum UHI, while Otte et al (2004) developed an urban canopy parameterization that could be used in the MM5 atmospheric model to improve boundary layer behavior estimation in weather forecasting. The increases i n computing power over the past three decades have also enabled the development of sophisticated Geographic Information Systems (GIS) and imaging software that allow a researcher to examine his or her data spatially. Sophisticated algorithms within the GI S software allow for the routine comparison of temporal data for change detection. This change detection capability has allowed researchers to investigate the expansion of urban areas and compare changes in the UHI with changes in the urban environment. Utilizing GIS and remote sensing data, Iino and Hoyano (1996) examined PAGE 33 17 the surface temperature distributions within an urban area, while Lo et al (1997) developed a spatial heating and cooling model based upon differing land use classifications within a G IS. 1.5 Satellite imaging of the environment Beginning with the April 1 st 1960 launch of TIROS I a new form of instrumentation became available, satellite remote sensing. TIROS provided the first set of atmospheric data from space. As a follow on to TIROS, beginning in August of 1964, the NIMBUS series of satellites were launched. The NIMBUS series of satellites employed multi spectral sensors for the measurement of atmospheric temperatures, critical for the forecasting of hurricanes. Because of th eir radiometric and spatial resolution constraints data from these ea rly satellites did not lend themselves to the study of features on the scale of urban environments. The follow on series of NOAA Geostationary Operational Environmental Satellites (GOES ) first launched in 1975, provide increased spectral, spatial and radiometric resolution compared to the NIMBUS series However the 4 by 4 kilometer ground pixel resolution in the thermal infrared band is a limiting factor in UHI studies. Starting i n 197 2 the LANDSAT series of satellites were launched. The first LANDSAT satellite was equipped with a multi spectral se nsor (MSS) that could image in five discrete bands of the electromagnetic spectrum, 3 bands in the visible spectrum, 1 band in the near infra red spectrum and 1 band in the thermal infrared spectrum. The 79 by 79 meter ground pixel resolution of the visible and near infrared band and the 240 by 240 meter ground pixel resolution of the thermal infrared band enabled the study of urban PAGE 34 18 features o n a scale that was useful for UHI investigation. After the initial success of the first three MSS LANDSAT satellites, four additional LANDSAT satellites were launched with increased spectral, spatial and radiometric resolution. The latest satellite, LANDS AT 7 has a 30 by 30 meter ground pixel resolution for the visible and near infrared band and a 60 by 60 meter ground pixel resolution for the thermal infrared band with additional bands in the shortwave infrared spectrum ( Jensen 2000). Landsat images have been utilized by researches such as Xian and Crane (2005, 2006) and Yuan and Bauer (2007) in the study of UHI s I n addition Landsat images are readily available on the World Wide Web at little or no cost to the user With the increased availability of satellite resources researchers have turned to satellite thermal imaging systems in an effort to gather more detailed information on the UHI. Initial work by Price (1979) assessed the UHI effect through the use of satellite data Roth et al. (1989) later examined the use of satellite data to derive the UHI of several co a stal cities and examined their utility in urban climatology. Additionally Aniello et al (1995) examined the use of LANDSAT data in evaluating micro UHIs. Looking to quantify the urban cha racteristics contributing to UHI s Xian and Crane ( 2005, 2006) and Xian et al. (2007) utilized LANDSAT satellite thermal data to examine the urban characteristics and associated land cover in Tampa Bay, Florida. Similar work was conducted by Yuan and Bau er (2007) utilizing LANDSAT images to compare the normalized difference vegetative index (NDVI) and impervious surfaces as an indicator of surface UHI effects. The Xian and Crane studies utilized the thermal infrared (TIR) capabilities of LANDSAT in their examination of the urban heat signature in their respective study areas. Results from both studies suggest that the percentage of PAGE 35 19 impervious surface in an urban area is a useful source of data to interpret urban thermal patterns and Land Surface Temperatu re ( LST ) While satellite data analysis can provide useful information on components of the UHI, it is limited in three areas. First, over flight times of LANDSAT and other satellite platforms for the continental United States occur in the early morning h ours, well prior to the maximum heating effects of the day, and provide no coverage during the maximum UHI night time period. Second, LST recorded by the satellites do not directly reflect the air temperature in the urban environment, which is a major for cing function on the generation of a UHI. Third, the temporal frequency of imaging by LANDSAT and other satellite platforms only allows for observations every 2 to 16 days, assuming no periods of cloud cover, limiting the amount of data that can be collec ted. The technological advances in the past 50 years since the coining of the term urban heat island have been tremendous; however, remote sensing and computer modeling all rely on empirical data to validate their results. 1.6 Land Use Land Cover ( LULC) in the study of UHI s The conversion of rural lands, with their vegetation and bare soils, to urban lands with their impervious surface s and sparse vegetation is a major factor in the development of a UHI. Properly identifying areas of rural and urban land s, with their accompanying characteristics, is of critical importance to UHI researchers. In the United States federal and state agencies have assumed the task of developing Land Use Land Cover ( LULC ) maps. As a federal agency t he USGS produces a series of LULC maps that includes the USGS percent impervious surface map used in this study. Likewise, the state of Florida PAGE 36 20 produces the Florida Land Use Land Cover Classification System (FLUCCS) series of maps. The goal of these maps is to classify LULC based u pon standardized methods. According to the state of Florida, Department of Transportation l and use, cover and forms classification system manual (1999 p8 ), The various categories and subcategories listed and defined herein reflect the types of data and in formation which can be extracted from aerial photography of various types (panchromatic, natural color or false color infrared) and scales (large, medium and small) and from the current generation of airborne and satellite multispectral imaging systems. Co lor, shade, shape, size, texture, shadows, context and, in the case of non photographic imagery, multispectral and multitemporal characteristics are some of the features used to implement land use/cover classification. While the FLUCCS utilizes multiple s ources to assign land to a particular classification, of the 74 sub categories of urban and built up areas none of these categories take into account the impervious surface. Many researchers use the standardized LULC as a bas is for their investigation of the UHI. Lo and Quattrochi (2003) utilized changes in LULC maps of the Atlanta Metropolitan Area in the analysis of the UHI and its health implications. Wend et al. (2007) utilized LULC maps in assessing the effects of LULC patterns on thermal conditions in Indianapolis, Indiana. Internationally Zhou et al. (2004) and He et al. (2007) examined the effect of LULC change on the change in UHI intensity in Chinese cities. In addition Stathopoulou and Cartalis (2007) examined UHIs in Greece utilizing LANDSAT and Corine Land Cover (CLC) data. While many valid results have been obtained using current LULC maps Lu and Weng (2006) note that, in the study of UHIs most of the LULC maps do not include impervious surface in their classification. They suggest that the lack of impervious surface percentages in many LULC maps limits the investigation of UHIs. Lu and Weng (2006 p 155 ) note that PAGE 37 21 I n urban studies land use is more useful than land cover information, because land use is directly related to social economic activ ities. A uniform definition of urban land use classes is not available yet. Most previous research classified urban built up areas into residential areas and commercial and industrial areas, or their subcategories when high resolution data were used. Lu and Weng assert that the incorporation of impervious surface in the LULC system will result in improved accuracy when studying UHIs. In their study Lu and Weng classi fied land use into five categories based on population density and impervious surfa ce. Four of these were subcategories of residential ( very high, high, medi um and low intensity) and the fifth category was comme rcial/industrial/transportation grouped together While four residential categories may be appropriate for Indianapol is, Indiana due to the structure of their residential areas the structure of residential areas in the Tampa study area exhibits less diversity and therefore is more closely align ed with Lu and Wengs low intensity residential areas. In addition Lu and Weng combined commercial/industrial/transportation area s into one category. In comparison to Indianapolis, Indiana the Tampa study area exhibits a less well defined commercial and industrial area (instead there are several areas). T herefore in the Tampa study area i t would seem appropriate to de link the commercial and industrial areas. While the Lu and Weng study promotes the inclusion of impervious surface in land use classifications as noted above, their classification system may not be appropriate for all urban e nvironments. In light of these issues the current study utilized a generalized in situ observation method to delineate commercial, industrial, and residential areas. For the purposes of this study an area was deemed to be in the commercial impervious su rface category if the observed surrounding area consisted of high rise office buildings, a large shopping center or larger strip mall. An area was deemed to be in the industrial PAGE 38 22 impervious surface category if the observed surrounding area was primarily us ed for manufacturing or shipping with building heights of 2 stories or less. An area was deemed to be in the residential impervious surface category if the observed surrounding area was composed of single family or duplex residences. 1.7 Wind speeds and the UHI Another factor that can have a moderating effect on the development of a UHI i n the urban area is the speed of surface winds When examining the role of wind speed on UHI development the configuration of the E arths surface can have a significant influence on wind speed Oke (1987) characterized the wind variation over cities by defining two sub layers; the unobstructed layer and the urban canopy layer In the unobstructed layer the wind is not influenced by surface friction while in the urban can opy layer the wind speed is infl uenced by the surface roughness of the urban area. Surface roughness is commonly expressed as the height value z 0 in meters. Table 1.3 lists some common surface roughness values reproduced from Oke (1987) Terrain z 0 meter s Scattered settlement (farms, villages, trees, hedges) Suburban Low density residences High density residences Urban Commercial PAGE 39 23 speed within the urban environment as compare d to the unobstruct ed wind speed commonly associated with the rural environment. Santamouris et al (1999) utilized simultaneous w ind measurements in the urban canyon s of Athens, Greece and found t hat the wind speed above the canyons reached as high as 5 ms 1 while the wind speed in the canyons never exceeded 1 ms 1 Earlier studies (mentioned previously) in the United States by Fassig (1907), Saucier (1949), and Ligda and Bigler (1956) found this same retardation of surface w inds by the urban environment. In examining wind speeds affect o n the UHI intensity, Escourrou (1991) found that the heat island intensity decreas es with increasing wind speed I n addition the temperature difference was insig nificant for wind speeds > 5ms 1 Table 1.4 is reconstructed from the Escourrou study relating heat i sland intensity and wind speeds. Wind speed in rural areas ms 1 Heat island intensity K 1 2 3 4 5 4.5 3.9 3.4 2.6 2.2 Table 1.4 Heat island intensity and the corresponding wind speeds in the G reater Paris area (Escourrou 1991) Utilizing data from cites in the United States Oke (1973 ) developed an equat ion that relates the wind speed, over the range from 0 to 4.5 ms 1 a nd the populations to the delta te mperature (UHI). dT=P 0.25 /4U 0.5 (7 ) where dT= change in temperature in o C P= population of a city in millions U= wind speed in ms 1 PAGE 40 24 Escourrou (1991) noted that there appeared to be an increase in the rural to urban temperature differences when the wind speed was < 2 ms 1 and that wind speeds > 5 ms 1 had a rapidly decreasing effect on the delta temperature. When considering the study of UHI s in subtropical climates such as Florida one must consid er the unique weather cond itions that prevail in these climates Because of its latitudinal location, peninsular form and its geographical relationship to the rest of the continental Unites States, Florida experiences a unique set of weather phenomenon. T he state of Florida extends from approximately 25 o n orth l atitude to 30 o n orth l atitude. The climate of the state is categorized primarily as humid subtropical with the extreme southern portion classified as tropical savanna. The peninsula width of Florida is approximately 240 kilometers and it is bounded on the west by the Gulf of Mexico and on the east by the Atlantic Ocean. Floridas location between these two bodies of water plays a major role in the development of Florida weather a s is noted by Mogil and Seaman ( 2008) and Stowers and Tab b ( 1987 ) Florida is the lightning capita l of the United States with over 50 lightning flashes per square mile per year ( Mogi l and Seaman 2008) and w hile not generally considered by the public an area of high tornadic activity, Florida actually experiences more tornados per square mile than any other state. The weather phenomenon most associated with Florida is the hurricane with t he state of Florida experiencing twice as many land falling hurricanes as the next clos est state Texas While hurricanes may be Floridas most well know n weather phenomenon, b etween the months of May and September Floridas weather is primarily influenced by the development and convergence of the west coast and east coast sea breeze s. Accor din g to Byers and Rodebush (1948) d aytime heating of the land surface in the central regions of PAGE 41 25 Florida sets up a temperature gradient between the warming interior and the relatively cooler humid air over the Gulf of Mexico and the Atlantic Ocean. The sea breeze normally begins to develop several hours after sunrise as the interior begins to warm, and does not reach its maximum until the maximum day time tempe ratures are achieved at around 15:00 to 16:00 local time. Sea breeze development is most pronounc ed during relatively calm synoptic conditions. They can develop in both humid and dry conditions however their greatest effect on Florida weather is during periods when humid conditions are present in the upper levels of the atmosphere. To better unders tand the development of the Florida sea breezes the following description is provided. Air pressure is defined as the weight of the air molecules pressing down from above at any particular point. Prior to the heating of the day the Atlantic, Gulf of Mexi co and the peninsular interior air columns have similar vertical profiles with decreasing air pressures with increases in altitude. As the daylight hours begin the land surface in the interior of the peninsula begins to warm which causes an expansion of t he interior air column. W ith the expansion of the interior air column there are now more air molecules aloft in the interior air column than there are in the Atlantic or Gulf of Mexico air columns. Therefore the pressure of the interior air column at the r eference altitude is now higher than the pressure of the Atlantic or Gulf of Mexico air columns at the same reference altitude As air flows from areas of high pressure to areas of low pressure the upper air higher pressure of the interior air column beg ins to diverge and flow to wards the upper air low er pressure areas over the Atlantic and Gulf of Mexico air columns. Over the Florida peninsula this normally occurs at an altitude of approximately 1500 meters T w o things now begin to happen S ince air is flowing out of the interior air PAGE 42 26 column in the upper atmosphere there are fewer air molecules pressing down on the interior surface T his causes the pressure to drop and a surface low form s At the same time air flows into the Atlantic and Gulf of Mexico air columns to form surface highs at their base. The final phase of the initiation of the Atlantic and Gulf of Mexico sea breezes is the movement of air from the ocean surface high pressure areas to the interior surface low pressure area. The sea breeze circulation is now complete and will continue as long as the interior surface heating continues. Depending on surface conditions and sea surface temperatures as the land cools after sunset, a land breeze may develop as the air column characteristics are re versed. Figure 1.5 is a graphic depiction of a daytime sea breeze flow. Figure 1. 5 Florida sea breeze flow The classic Florida afternoon thunderstorm usually occur s between the mo nths of May and September when a semi permanent high pressure system has settled over the Bermuda region of the Atlantic Ocean and synoptic conditions over Florida are relatively PAGE 43 27 calm. Daytime heating of the peninsula interior begins to setup the sea breeze s from both coasts. Moist marine air is drawn into the interior where it begins to rise. Latent heat is released to the atmosphere as the moist sea breeze air condenses This intensifies the vertical movement of the interior air column which in turn intensifies the sea breeze to result in the classic Florida afternoon thunde rstorm When examining severe weather in southwest Florida Collins et al (2009 ) noted that the most intense thunderstorms occur where the west coast and east coast sea breeze s collide Additional conditions can complicate the actual development regions o f th e sea breeze Depending on the outline of the coast, sea breeze movement can be modified ( Baker et al 2001). Sea breeze s will tend to develop orthogonal to the coast line, allowing for the development of convex, concave, as well as straight line sea breeze fronts. Convex coastlines can present an interesting situation because of the orthogonal development of the sea breeze i n relationship to the coastline. In this case a focusing effect can be generated. One additional facet that eff ects the develo pment of the sea breeze has only developed in the last 100 years anthropogenic modification of the land cover ( Pielke et al 1999). During an investigation of anthropogenic land cover changes Pielke found that when simulated 1900 land cover characteristic s were used as a baseline and actual 1973 and 1993 land cover conditions were input into the Regional Atmospheric Modeling System (RAMS) the model showed a modification of the sea breezes. The model also showed that there was a respective 9 and 11 percent decrease in precipitation along with a redu ction in deep cumulus rainfall. This would tend to support the theory that anthropogenic land use changes may also influence Floridas weather. In reference to PAGE 44 28 another subtropical location Freitas et al (200 6 ) documented the interaction of the UHI with sea breezes in S o Paulo Brazil. They noted that i ncreases in the lateral dimension of the urban environment can slow the transition of the sea breeze through the urban environment. As was noted by Oke (1987 ), Escourrou (1991) and Freitas et al (2006 ) surface winds tend to have a moderating effect on the development of the UHI I t can therefore be postulated that the speed of the sea breezes of the Florida peninsula should have an effect on the development of a UHI in Florida cities. G iv en the periodicity of sea breeze conditions in Florida one can also postulate that there is a temporal component to the moderation of UHI development in Florida cities caused by sea breezes. 1.8 UHI s ocietal i mpacts There has been an increased awar eness within the research community and the public in general of the potential changes, in local climate associated with an expanding urban environment. The increases in temperatures associated with a UHI can have an economic as well a s a climate impact The economic impact is felt by consumers in the form of higher electric bills associated with the increased use of air conditioning to counteract higher UHI temperatures. Within the United States the primary energy source s used in the generation of elect ricity are fossil fuels in the form of coal, natural gas, and oil. The burning of these fossil fuels generates carbon dioxide, a green house gas The increases in electricity usage associated with UHI increases in air conditioning use, helps to reinforce t he positive feedback mechanism of global warming. PAGE 45 29 Currently there are numerous methods available to reduce UHIs. As discussed by Akbari and Konopacki (2005), selection of building materials can have a substantial impact on UHI temperature. Referri ng back to the urban energy balance equations of Section 1.2, if the albedo of the impervious surfaces is increased the net result will be a lowering of the sensible heat. The choice of lighter colored building and roofing materials along with lighter col ored pavement would have a direct effect on lowering the urban environment sensible heat and the consequent UHI. Another simple and cost effective method of decreasing the UHI temperature is the planting of trees and vegetation in the urban areas with high percentages of impervious materials (Emmanuel et al. 2007). Trees provide a direct shading effect and thus reduce solar insolation impinging on the impervious surfaces. In addition, trees and vegetation increase the transpiration rate in the energy balan ce equation also resulting in a lowering of sensible heat. This idea is promoted by Rosenzweig et al. (2009). In a proactive manner r esearchers such as Coutts et al (2008) are actively trying to develop tools that will assist urban planners in reducing th e impact of future urban development on the generation of a UHI. The simple act of forethought in the selection of building and paving materials and ground cover when planning the urban environment of cities can have a large scale impact on UHI, energy us age, and greenhouse gas emissions. 1. 9 Research q uestions With advances in technology, research into the effects of the urban environment on weather and climate modification appears to be moving towards the use of remote sensing and computer modeling in lieu of the empirical studies of the past. This may be PAGE 46 30 due to the high costs associated with the large sampling networks required for empirical studies or just the convenience of remote sensing and computer models. What appears to be lacking is research in the region straddling empirical studies and remote sensing / computer modeling of urban environments. Having reviewed the literature several research questions have emerged. First, the UHI of a study area may commonly be reported as a range of value s or a (Yow and Carbone 2006) or in some cases as semi concentric circles emana ting from the CBD (Unger et al. 2001). It is believed that the spatial and temporal values of the UHI are actually more complex. The question then is within th e Tampa study area what is the actual spatial and temporal variability of the UHI ? Second can it be experimentally established if there is a significant relationship between the percentage of impervious surface and increases in urban temperatures? Third it has been shown by other researchers t hat the speed of surface winds can have an effect on the development of a UHI. Utilizing the data that has been collected during this research what is the significance of the relationship between the mean measured rural to urban temperature differences and the mean recorded surface wind speeds ? 1. 10 Hypothese s Having identified inferences within the literature that suggest that impervious surfaces and surface winds play a role in the generation of the UHI it is h ypothe sized that; PAGE 47 31 1. The meas ured UHI delta temperature i n the Tampa study area will show both spatial and temporal variability. 2 There is a significant relationship between the percentage of impervious surface and the intensity of the local UHI in Tampa, Florida 3 There is a significant relationship between the speed of the surface winds and the intensity of the UHI in Tampa, Florida. I n addition the temporal characteristics of the Florida sea breezes contribute to a modified temporal UHI profile unique to the Tampa, Florida study area. PAGE 48 32 2 .0 METHODS As detailed in the introduction o ver the last 20 years Tampa has experienced a period of rapid population growth along with a rapid conversion of rural lands to urban use. Accompanying t his land cover change has been an increase in impervious surfaces, buildings, roads, and other paved surfaces common in the urban environment This research examines the relationship between impervious surfaces and the generation of the UHI as well as con tributions by surface winds to the moderation of the UHI in Tampa Bay R egion In addition this research looks at the spatial and temporal characteristics of the UHI in the Tampa study area Figure 2.1 provides a graphic representation of the basic methods and flow of this research. Utilizing the research design flow diagrams as a template, the methods employed in this research will be discussed. PAGE 49 33 Tampa Florida Urban Heat Island Research Design Flow Examination of the mean delta temperature in the Tampa Florida study area spatially and temporally Location Preliminary work Data plotting Evaluation Hillsborough County FL study area Evaluate the generated plots for the existence of UHI spatiotemporal patterns in the study area Prepare spatial and temporal GIS outputs for each 30 minute sensor measurement Process the 2007 sample period sensor data to arrive at calculated temperature values for the study Calculate mean delta temperature values Proceed to impervious surface analysis Figure 2.1 Research design flow ( spatial and temporal analysis ) PAGE 50 34 Figure 2.1 (continued) Research design flow (i mpervious surface) PAGE 51 35 Figure 2.1 (continued) Research design f low (wind s peed) PAGE 52 36 2 .1 Research a rea The Tampa Bay R egion and mor e specifically the urban area in and a round the city of Tampa was chosen due to its l ocation in the sub tropic al region of the Northern H emisphere and its ease of access by this researcher. Figure 2.2 is an aerial image of the Tampa Bay R egion looking from the downtown are a to wards the north. T he four lane highway in the lower center of the image defines a portion of the southern boundary of the study area. The exact location of the study area is shown in Figure 2.3 Figure 2.2 Aerial image of the Tampa Bay Region ( Image reprinted with permiss ion from www.bigstockphoto.com ) To delineate the study area for this research topographic and impervious surface data layer files of the Tampa Bay R egion were downloaded from the U SGS seamless data PAGE 53 37 server. To aid in the delineation of the study area the topographic and impervious surface layer files were imported in to the Environmental Systems Research Institute s ( ERSI ) GIS program ArcMap version 9.2. Examining t he impervious surface layer boundaries of the study area were defined to include a large portion of the metropolitan area of the city of Tampa as well as the surrounding suburbs and rural lands. Several physical boundaries were identified that helped to define the study area. Interstate 75 serv ed as a convenien t eastern boundary while County L ine R oad on the border between Hillsborough and Pasco counties served as a convenient northern boundary. The southern boundary was chosen such that it encompassed the CBD and lands to t he we st of I nterstat e 75 and east of Tampa Bay The western boundary was chosen to skirt the upper regions of Tampa B ay commonly referred to as O ld Tampa Bay to the point where it contacts the northern boundary at County L ine R oad. Figure 2.3 depicts the boundary line over laying the impervious surfac e layer of the Tampa Bay Region while Figure 2.4 depicts the study area within the context of the G reater Tampa Bay Region and the state of Florida PAGE 54 38 Figure 2.3 Tampa study area boundaries PAGE 55 39 Figure 2.4 Study area in relat ion to the G reater Tampa Bay Region PAGE 56 40 2.2 Tempe rature data logging sensor p lacement In a review of UHI studies between 1950 and 2006 Stewart (2007) found that the classification of a site as rural at times appeared arbitrary and in general followed no d efined quantitative rules. Stewart also found a very wide variation, within UHI studies, in the physical characteristics of the rural reference sites as described by the various study authors. As a result of this, Stewart was concerned about differing micr o climates that might exist at the different rural sites of the studies. Lacking any clear guidance from the literature at the start of this research as to the selection of rural sites an average value of less than 15 percent impervious surface was set as criteria for classification as rural for this research. Figure 2.5 and 2.6 are photographs of the TECO utility pole and the surrounding landscape near rural sensor number 99. The location of s ensor number 99 is a good example of the site characteristics of the rural sensor locations of this study. Figure 2.5 Rural sensor number 99 mounting pole PAGE 57 41 Fi gure 2.6 Area near rural sensor number 99 The study initially occurred in 2007. As part of this research a follow up visual examination of the rural s ensor locations was conducted in 2008 and 2009. During 2008 a visual inspection of the area around rural sensor number 1 revealed the beginnings of significant construction near the site, therefore rural sensor number 1 was not included in 2008 validation sampling. Figure 2.7 is a 2009 photograph of the surrounding area near rural sensor number 1. PAGE 58 42 Figure 2.7 2009 image of the rural sensor number 1 surrounding area A visual inspection in 2009 also revealed that the area surrounding the location of rura l sensor number 100 was being modified by construction. This construction was not present in 20 08. Figure 2.8 is a photograph of the area near rural sensor site number 100. Due to the rapid expansion of urban areas in the Tampa study area future duplicat ion of the study will not be possible without the establishment of new rural sensor locations. PAGE 59 43 Figure 2.8 2009 image of the area surrounding rural sensor number 100 A visual examination of the 2002 USGS impervious surface raster image of the study a rea revealed that the spatial distribution of impervious surface in the study was not homogenous and in fact exhibited a varying spatial distribution. Since the study area contained large areas of lower percentages of impervious surface it was initially de cided that a sampling scheme would be employed that biased the sample points towards the areas of higher percentages of impervious. A search of automated sample point generation GIS tools revealed what was believed to be a routine that could take this int o consideration It was found that Hawths a nalysis tool set for ArcGIS ( Beyer 2006) contained a sample point generation routine that would allow the use of a GIS raster layer to serve as a probability matrix in the selection of sample points. Since the US GS PAGE 60 44 Figure 2.9 Hawths analysis toolset for ArcGIS. percentage of impervious surface map contained values of percentages of impervious surface ranging from 0 to 100 this layer was chosen to serve as the probability matrix raster layer in the sample point generation routine. According to Beyer (2006), whe n using a raster layer as a weighted probability distribution matrix the values in the raster layer are used as probabilities of placement such that raster cells with larger values are more likely to have a point placed in them. The probability for each c ell is calculated by: (value of cell / maximum value). Figure s 2.9 and 2.10 represent the tools available with the Hawths a nalysis tool set The tool of interest to this research is the Generate Random Points tool. Figure 2.10 depicts an example of the r aster and vector data sets that are required by this tool. Figure 2.10 Hawths analysis tool random point generation. PAGE 61 45 The desired outcome of Hawths sample point generation routine was a biasing of sample points towards areas of higher percentages of impervious surfaces. Havin g previously defined the boundar y that represents the Tampa study area and having made the appropriate extraction of impervious surface data within the study area t hese parameters were entered in to the random point generation routine. The number of points to generate was set at 100. The value of 100 was chosen for three reasons Firstly so that the sample size of the data sensor locations would be large in comparison to other studies found in the literature ; secondly, to e nsure that the problems associated with small number sets woul d not be encountered during the st atistical analysis of the data; and thirdly due to the cost constraints on the research since 100 sensors was the absolute maximum that could be purchased within the research budget. Utilizing the USGS percentages of impe rvious surface raster as the weighted probability distribution raster layer a random point set of 100 points was generated where the cells having a higher percentage of impervious surface were more likely to have a sample point in that location. At the co mpletion of the point generation routine placement effectiveness was evaluated. Overall the study area exhibits a mean percentage of impervious surfaces of approximately 25 percent. Evaluating the percentage of impervious surface at the sample points revea ls a mean percentage of impervious surfaces of approximately 45 percent. It would appear that the input parameters utilized in the Hawths sample point generation routine had desired effect of biasing the sensor placement towards areas of higher percentage s of impervious surfaces. It should be noted that five northerly locations of the 100 generat ed sample points were in the most rural setting s and were selected to serve as rural reference sample PAGE 62 46 points. Because the TECO utility poles did not exactly corr espond to the designated locations sensor s were relocated in the field to a TECO pole within 25 meters of the generated point and the new Global Positioning System ( GPS ) coordinates were recorded. This resulted in the average impervious surface of the fi ve rur al sensors being approximately 15 percent The calculated range of values was from 4.6 to 26 percent impervious surface. It should be noted that when calculating the final value of impervious surface near the rural sites, two sites exhibited values g reater than 15 percent. It is believed that these values may be caused by an edge artifact of the calculating algorithm related to the resolution of the 2002 USGS impervious surface map. T hrough in field examination of the points it was concluded that the points c ould still be considered rural as all of the ru ral sites w ere near open fields ( Figure s 2.5 and 2.6 ) T he only impervious surface in their vicinity was a two lane road. V isually all five rural sites appea red to be similar The selection of an aver age of less than 15 percent impervious surface as criteria for rural status is not without question H owever a s stated in Section 2.2 a review of the UHI literature revealed no definitive criteria for a location to be classified as rural. To minimize th e differences in siting characteristics of the rural sensors, during the analysis phase, a mean temperature value of the five rural sensors was calculated for each 30 minute sample period. The sensor placement output generated by the Hawths random point g eneration routine was input into ArcMap and combined with the raster and vector layers of the study area. Figure 2.11 depicts the temperature data logging sensor spatial placement in the study area. PAGE 63 47 Figure 2.11 Temperature data logging sensor p lac ement During the 2007 field installation of the temperature sensors, observations were made as to the age of the surrounding locale. Based upon the observations of landscape and building age all structures, landscape and impervious surfaces were deemed to be in existence prior to 2002 and as such were accounted for in the 2002 USGS impervious surface image. Rural Sensor Location s PAGE 64 48 Postscript; in hind sight the biasing of sampling sites towards areas of higher percentages of impervious surface was not appropriate, a more appro priate method of sampling would have been the selection of equal numbers of sample s within discrete binning g roups of percentage of impervious surface 2.3 Temperature d ata l ogging s ensors During the initial stages of the research method s design seve ral types of temperature logging equipment were investigated. Some of the key requirements of the temperatur e data logging equipment were that it had to be low in cost, robust in construction ( as it was being left unattended in the field ) and be able to l og temperature reading s every half hour for a minimum period of thirty days before requiring downloading. The first to be evaluated were commercially available weather stations with data logging capabilities Figure 2.12 is an example of one of the severa l lower cost data logging weath er station s that are available for sale. Figure 2.12 commercially available data logging weather station Zephyr Instruments PWS 1000TD TZ PAGE 65 49 The unit in F igure 2.12 retailed for $374 bu t could be found on sale for $299. It is indicative of the cost of unit s with similar capabilities. It has the capability of measuring and logging temperature, wind speed, wind direction, and rain amounts. Electronic s of t his type need to be housed in a water tight enclosure and equipped with a battery pack to run for at le ast thirty days without re charging. With the additional required component s the per unit price is approximately $450. A compliment of 100 units would have been required to fully populate the sensor locations of the study with a result ant cost of $45,000 In addition to the initial purchase price the ongoing logistics cost of supporting this device is expensive An example of the use and logistics required to support a network of 35 commercial weather station data logger s can be found in Bassara et al. ( 2009). This option was not feasible for this study Given the costs, a custom weather station that could record and log temperature, wind speed and wind direction was built and tested. I t was determined that the unit could be built for aproximately $100 a nd would require approximately 400 man hours t o construct 100 unit s. The total cost for this approach was in excess of $10,000. This option was not feasible either. In addition to the high cost of implementation, both of the above options had problems wi th site location constraints. T he above options require a fixture to attach the measurement sensors and logging equipment Many of the designated sensor locations were in residential neighborhoods, including economically depressed areas. E nsuring security and an unobstructed view for the wind sensor require s mounting the units on the top of many buildings or houses, necessitating obtaining permission from building and PAGE 66 50 land owners. An additional concern with mounting the unit on the top of buildings is the possibility of temperature effect s of the structure itself. A third approach was needed. A review of sensor manufacturers led the search for an appropriate low cost temperature data logging sensor to the Dallas Semiconductor company. The author had previo usly used a line of Dallas Semiconductor environmentally sealed sensor components for a different project A search through the Dallas Semiconductor data catalogue revealed a small form factor temperature data logging sensor which contained an internal po wer source. The device is the Dallas DS1920x Thermo chron series of data loggers. This device is available in several different temperature ranges. The device chose n for this study was the DS1621H which has a temperature measurement range of 15 to 46 o C w hich is a good match to the expected summertime temperature range of 20 to 32 o C in the Tampa study area as obtained from the NWS website ( www.srh.noaa.gov ) The Dallas DS1621H is capable of logging temperature data at user defined rates of 1 to 255 minutes between measurements and can store 2048 such measurements. The DS1621H has a resolution of 0.125 o C and an accuracy of better than 1 C over its entire range (Dallas Semiconductor data book 2010). I n addition each DS1621H sensor has a unique serial num ber embedded within its firmware that allows for the tracking of each deployed sensor. A search of the literature revealed that other researchers have used the Dallas Thermo chron sensor in their studies. Hubbart et al (2005 p. i) evaluated the performa nce of the Thermochron sensor and their results indicated that, The Thermochron IButton is an accurate, inexpensive alternative to more expensive temperature data logging systems, and is well suited for obtaining quality spatially distributed data ... . Hubbart also ran a PAGE 67 51 series of accuracy checks on 61 Thermochron sensors and found that collectively the devices exhibited an accuracy of 0.21 C This is much better than the published specification noted above It should be noted that each Thermochron unde rgoes final trimming ca l ibration in the factory utilizing National Institute of Standards and Technology (NIST) traceable temperature standards. Johnson et al. (2005) used the Thermochron sensor to document groundwater and river interactions, and Hartman and Oring (2006) used the Thermochron sensor to remotely monitor bird nesting. The HOBO series of data loggers was also investigated; however, the least expensive HOBO outdoor temperatur e sensor ha d a cost of $36 (in quantity purchase), over double the cost of the Thermochron sensor. In addition, it did not have the capability to synchronize recording start times or have variable sampling rate s. I ts 90 percent temperature settling time was 10 minut es, compared to the two minute settling time of the Thermochron sensor. Utilizing the Thermochron sensor w ith a price of $16 per sensor in quantities of 100 the total cost of implementing the 100 sensor network of this study was $1,600 (significantly le ss than the weather stations deployed by Bassara et al. (2009) mentioned earlier) Therefore the DS1921H temperature logging sensor was chosen for the Tampa area UHI study. Detailed specifications for the Dallas DS1621H can be found in Appendix A. Figure 2 .13 shows the relative size of the Thermochron sensor. While other local weather characteristics such as wind speed and pressure, were not able to be measured with this device, needed temperature data were obtained. PAGE 68 52 Figure 2.1 3 Thermo chron sensor relative size Owing to its small size, self contained power source and environmentally sealed case, the Thermochron sensor was ideally suited for installation at all of the designated sensor sites. Prior to installation in the fiel d, each sensor was affixed to a wooden support structure via double sided foam tape. A thin structure of wood was chosen for its relative thermal insulation, environmental stability and its ease of mounting During the initial round of on site physical in spections within the study area it was noted that utility poles were ubiquitous. Utility poles seemed to be a good mounting structure for the temperature data logging sensors T herefore the Tampa Electric Company (TECO) was contacted to obtain permission to attach the Thermochron sensor s to their utility poles. Permission was granted in April of 2007. Figure 2.14 depicts a typical installation of a sensor on a TECO utility pole. PAGE 69 53 Figure 2.13 Typical sensor installation on a TECO utility pole It is ac knowledged that there are times of the day when the sensor could be in direct sunlight; however, to minimize the exposure, d uring installation each sensor was positioned on the north side of the utility pole. The sensor was affixed at a height of approxima tely 2 meters as recommended by the World Meteorological Organization height stand ards for temperature monitoring ( Barnett et al. 1998) The sensor was attached to the utility pole by means of a base 36 inch plastic cable wrap on to which the wooden suppo rt structure with its sensor was attached by a small cable wrap and position ed in a manner so that it would hang freely and not contact the utility pole GPS coordinates were collected for each installed sensor location along with ancillary data includi ng the utility pole number, the surrounding topography and the urban classification PAGE 70 54 ( commercial industrial residential and rural). Urban classification types are defined in Section 1.6 Upon return ing from the field, installed sensor GPS coordinates we re differentially corrected and imported in to ESRI ArcMap to update the sensor location layer. All 100 sensors we re placed in the field initially between M ay 30 th and 31 st 2007. All sensors had been previous programmed to start recording at midnight on J une 1 st 2007 Sensors were again deployed in the summer of 2008 to serve as a validation sample set. Data gathering during the 2008 validation period utilized a subset of the 2007 TECO pole sam ple locations. In 2008, thirty one sample locations, 4 rural and 27 randomly selected urban sites were chosen from the 2007 sample sites to serve as the validation set. In order to not have to employ small sample size statistics a minimum of 30 samples were required. Thirty one sites were selected to ensure that the failure of a temperature sensor would not drop the sample size below 30. As noted in Section 2.2, the area near sensor site #1 was undergoing commercial development T herefore rural site #1 was excluded form the 2008 data run resulting in just four r ural sensors used in the 2008 sample period. At the completion of the 2008 sample period, data were download ed and formatted into an SPSS database as was the case for the data obtained in the 2007 sampling period. Data analysis on the 2008 data was conduct ed in the same manner as the 2007 data analysis 2.4 Temperature data logger p rogram setup Prior to installation of the temperature data lo gging sensors in the field the sensor s were programmed with the current local time, the required sampling rate, re quired PAGE 71 55 sampling start date and time, and location identification number. A commercial software package offered by Scanning Devices Inc. was purchased to manage the setup and to later download and format the collected temperature data. Figure 2.15 depicts a sample programming setup and data download page of the Scanning Devices software. In F igur e 2.15 it can be seen that this particular sensor was assigned the identification number of Sensor #13, and logged temperature data were read. F igure 2.15 Scann ing Devices programming sample p age The same data page is used for uploading to and downloading from the device. File output from the program is in comma delimited format. After completion of the sampling 30 PAGE 72 56 run, the data were downloaded from the devices and exported into Excel format for data reduction. Prior to the start of device setup the correct time for the programming computer was obtained from the U.S. N aval observatory atomic clock ensuring that all devices were synchronized to a reference time s tandard. During programming the device s on board clock was synchronized with the programming computers clock. For the purposes of this research all devices were set at a sampling rate of 30 minutes. In addition all devices were set to start recordin g at the same identical time. In the case of the 2007 sample period this start time was 12 oclock midnight on June 1 st For the 2008 sample pe riod the devices were set to start temperature logging at 12 oclock midnight on May 14 th The earlier startin g date in 2008 was necessitated by this authors prior scheduled medical needs. Synchronizing the devices to a common clock and starting the devices at the same time ensured that temperature readings were obtained at the same time by all devices. As no ted in Section 2.2 t his is one advantage of the method used in this study over the mobile transect method used in some other studies (e.g. Hedquist and Brazel 2006, Yi Sun et al. 2009) In the other studies there were inherent time differences in the tempe rature measurements These time differences induce errors when attempts are made to temporally normalize the temperature measurements (Peterson 2003) 2.5 Sensor data reduction At the end of the sampling period all sensors were returned to the Universit y of South Floridas meteorology laboratory from the fi eld and the temperature data were downloaded from each sensor and entered into Excel T ime period means for rural and PAGE 73 57 urban sensors were calculated along with rural to urban s for each time period To obtain a mean rural value of temperature at each 30 minute sa mple point the temperature data from all five rural sensors were aggregated over the sample period and a mean rural temperature was calculated for each 30 minute per iod to serve as the rural temperature in subsequent analysis. In addition rural data were aggregated in an effort to minimize any differenc es in nearby impervious surface percentages and differences in land cover. As noted earlier, rural sites were simila r. During the processing of the site specific urban sensor data a mean value of temperature for each 30 minute period over the sample period was calculated. T he sam ple point daily temperature data of each individual urban sensor was then aggregated by 30 minute periods and a mean value was calculated for each 30 minute period for each sensor. Delta temperature values were then calculated by subtracting the mean rural temperature from the site specific urban temperature at each 30 minute s ample point. Res ultant data were reformatted into an SPSS database for further analysis. 2.6 Sensor proximate percent impervious surface For the analysis of the relationship between the percentage of impervious surface and the rural to urban temperature differences it w as necessary to calculate the values of percentage of impervious surface at varying radii, surrounding each temperature logging sensor site. ESRI ArcMap software was utilized to calculate the percent age of impervious surface at each sensor location. A re view of the literature did not reveal a method to determine a value of distance at which impervious surface has a measurable effect on the temperature therefore a simple PAGE 74 58 method was devised. An arbitrary value of 25 meters was chosen as a starting point. Th is 25 meter value was first doubled and then doubled again to arrive at distances of 25, 50 and 100 meters. The selection of 100 meters as a maximum value was not predicated on prior knowledge rather it was a subjective decision made by this researcher. S ince the resolution of the 2002 USGS percentage of impervious surface raster image was 30 by 30 meters, a 30 meter point was also included. Details of the analysis at these differing distances are presented in Section 4. It should be noted that though USG S impervious surface maps have a resolution of 30 by 30 m eters the generation of the maps entails combining LANDSAT images with training data obtained from 1 m eter resolution orthographic maps enabling differentiation of mixed pixel information. A complet e description of the methods used to generate the percentage of impervious surfac e map is described in Homer et al ( 2004). The same 2002 USGS impervious surface raster that was used in the Hawth s tools calculation of sensor locations was also used to obt ain the value of the percentage of impervious surface in proximity to each sensor location. The USGS im pervious surface raster layer was first converted to vector form, with each generated polygon value representing the percentage of impervious surface. A selection by distance to centroid was utilized to identify polygons within the pre defined distance s (noted above) of each sensor location A selection by circular sectional areas could have also been used Having identified the polygons that were within the prescribed distances of the sensor location, the mean impervious surface value s were calculated for each sensor location at the four distance values. These calculated values were then imported into an SPSS database to serve as the value of the indepe ndent variable in the PAGE 75 59 later Ordinary Least Squares ( OLS ) regression analysis. Figure 2.16 depicts several of the different centroid selection areas surrounding temperature sensor locations Figure 2.16 Percentage of impervious surface at temperature log ging sensor locations PAGE 76 60 2.7 Wind speed data collection To enable the analysis of the effect of wind speed on the moderation of a UHI in Tampa Bay R egion hourly wind speed and direction data were obtained for the Tampa I nternational Airport weather stati on. Data for th e Tampa International Airport were downloaded from the National Climatic Data Center (NCDC) historical data site for the period from June 1 st 2007 to July 17 th 2007. It was again downloaded from May 14 th to July 1 st for the 2008 validatio n sampling period. Daily hourly wind speed data for the period were aggregated and a mean hourly wind speed for the sample period was calculated. The hourly wind data were subsequently interpolated to derive half hour wind speed values. In addition to the Tampa International Airport data, w ind speed and direction data were also gathered for MacDill Air Force base (AFB) to the south of the Tampa study area and F ort Howard Park to the northwest of the Tampa study area. The same process of aggregation and cal culation of mean values and half hour values was performed on the MacDill AFB and Fort Howard Park data sets. Based on the wind speed and direction data from these sites the magnitude and onset time of the sea breeze was determined It should be noted th a t a pronounced onset of a nocturnal land breeze was not revealed in the data, rather there was a slow shifting of wind direction to the sou th during the night time hours. The wind speed data along with the calculated study area mean delta temperature valu es were incorporated into an SPSS database for subsequent regression analysis. Results of the regression ana lysis are presented in Section 5.0 PAGE 77 61 2.8 Analysis m ethods Various components of the impervious surface, rural to urban temperature differences, a nd wind speed data were evaluated using correlation analysis, d escriptive stati stics, tests for normality, f tests and t tests as a measure of significance of a relationship, and OLS regression analysis to quantify a relationship Relationships that were identified were considered significant if the two tailed t test had a p value of < 0.10 Results of this analysis can be found in the analysis results of Sections 4 and 5 Rural and urban temperature data and the associated descripti ve statistics can be found in data collection Section 3. In addition to examining the overall spatial and temporal make up of the Tampa study area UHI, individual impervious surface categories were also examined. In addition ESRI ArcMap 9.2 software was utilized for basic spatial analysis and interpolation of variable fields into raster graphic display; w hile GeoDa version 0.95 software was utilized for spatial autocorrelation analysis and spatial autocorrelation correction of regression analysis. PAGE 78 62 3.0 D ETAILS OF THE URBAN AND RURAL TEMPERATURES AND UHI 3.1 Introduction Having gathered the rural and urban temperature data and having processed the data accomplished vi a the three separate pathways which were defined in Section 2.0 These pathways are as follows: an evaluation of the spatial and temporal characteristics. 2. An examination of the relationship betw een the percentage of impervious surface with potential adjustments for clustering. 3. An examination of the relationship between the sample period mean wind speed values for the Tampa Bay R egion study Bay R egion study area 3.2 Data collection At the comple tion of the sampling period s the raw temperature data were downloaded from all of the sensors. Table 3.1 provides a sample E xcel listing of a raw data down load ed from a sensor The fourteen digit hexadecimal number in the fourth column is the embedded device serial number with all sensors having a unique serial number. PAGE 79 63 Sensor #98 Running Temp o F 4f2000011fb121 1 Tue, Jun, 05, 05:29, AM 73 2 Tue, Jun, 05 05:59, AM 73.2 3 Tue, Jun, 05, 06:29, AM 73.2 5 Tue, Jun, 05, 07:29, AM 77 6 Tue, Jun, 05, 07:59, AM 79.9 7 Tue, Jun, 05, 08:29, AM 79.5 8 Tue, Jun, 05, 08:59, AM 82 9 Tue, Jun, 05, 09:29, AM 82.8 10 Tue, Jun, 05, 09:59, AM 85.3 11 Tue, J un, 05, 10:29, AM 82.8 12 Tue, Jun, 05, 10:59, AM 84.9 13 Tue, Jun, 05, 11:29, AM 87.8 14 Tue, Jun, 05, 11:59, AM 90.5 15 Tue, Jun, 05, 12:29, PM 88 16 Tue, Jun, 05, 12:59, PM 88.7 17 Tue, Jun, 05, 01:29, PM 90.1 : : : : : : : 2046 : Tue, Jul, 17, 07:59, PM : 70.3 2047 Tue, Jul, 17, 08:29, PM 71.2 2048 Tue, Jul, 17, 08:59, PM 71.2 Table 3.1 Sample raw data download. Of the 100 sensors initially dep loyed two sensors stopped recording prior to the end of the sample period and four sensors were removed by unknown parties. For the purpose of the analysis data from the two stopped sensors were discarded. The remaining 94 sensors provided data over the entire sampling period. A fter downl oading the 2007 data a setup error was noted. During the 2007 sampling period t he sensors were not programmed to stop recording at the completion of 2048 sample s T heref ore some of the initial data were overwritten. This did not affect the accuracy of the data I t just meant that the data from Jun e 1 st to June 4 th 2007 were overwritten by later data. This programming error was corrected for the 2008 sampling period and data recording stopped at the programmed t ime Therefore no data were overwritten. Dur ing the PAGE 80 64 sampling period each temperature sensor logged the 2048 temperature readings allowed by the device With a compliment of 94 valid sensors in 2007 and 31 valid sensors in 2008 this resulted in respectively a total of 192,512 and 63,488 temperature r eadings recorded during the sampling period s Initial data reduction was accomplished using the Excel statistical package. The cal culation of mean rural temperature value s involved calculating a daily sampling period mean of the combined rural sensors (Ta ble 3.2). Week Day Time 24hr Sensor #1 Sensor #2 Sensor #98 Sensor #99 Sensor #100 Mean Period_mean Tue 0:00 77.7 76.6 75.7 75.7 75.9 76.32 75.88 Wed 0:00 73.2 72.5 73.6 72.7 71.4 72.68 Thu 0:00 75 73.4 73.4 73.6 72.3 73.54 Fri 0:00 70.3 69. 8 70.3 68.9 69.3 69.72 Sat 0:00 77.5 75.7 76.1 75.9 76.8 76.4 Sun 0:00 78.8 77.7 77.9 77.5 77.2 77.82 Mon 0:00 79.5 77 76.8 76.1 76.6 77.2 Tue 0:00 77.7 76.8 78.1 76.6 76.8 77.2 Wed 0:00 68.5 67.1 67.1 66.4 65.3 66.88 Thu 0:00 73.8 73.8 74.1 73 .4 72.1 73.44 Fri 0:00 77.5 77.2 77.9 77 76.8 77.28 Sat 0:00 77.7 77 77.5 76.6 75.2 76.8 Sun 0:00 78.3 77.7 77.2 77.2 75 77.08 Mon 0:00 74.8 75.7 74.5 73.4 73.2 74.32 Tue 0:00 75.9 74.1 74.5 74.1 73.4 74.4 Wed 0:00 77 73.4 74.8 74.8 75.2 75.04 Thu 0:00 79 76.8 77.9 77.5 75.4 77.32 Fri 0:00 75 73 73.2 73 73.6 73.56 Sat, 0:00 79.5 77.2 76.6 77 76.6 77.38 Sun 0:00 79.3 76.3 76.1 75.9 77.2 76.96 Mon 0:00 78.3 79 77.9 77 77.5 77.94 Tue 0:00 78.1 78.1 78.1 77.5 76.8 77.72 Wed 0:00 75.2 7 4.1 74.8 73.8 73 74.18 Thu 0:00 75.4 72.7 75 74.5 72.7 74.06 Fri 0:00 77.2 74.3 76.3 75.9 75.9 75.92 : : : : : : : : : : : : : : : : : : : : : : : : Tue 23:30 77.7 76.6 75.7 75.9 76.1 76.4 76.27 Wed 23:30 73 73 73 72.1 71.6 72.54 Thu 23:30 7 5.7 74.3 74.5 74.8 73.4 74.54 Fri 23:30 69.8 69.8 70.3 69.1 69.6 69.72 Table 3.2 E xample of c alculated 30 minute sample mean temperatures o F PAGE 81 65 From these daily period means an overall rural mean temperature was calculated for each daily time period. These rural thirty minute means were later used in the calculation of the rural to urban temperature s Figure 3.1 depicts a 24 hr plot of the mean urban and rural near surface air temperatures within the study area. It should be noted that all times within this document are expressed as local Eastern Daylight Time (EDT) Figure 3.1 24 hour study area mean urban and rural near surface air temperatures As shown in Figure 3. 1 the daily maximum and minimum temperature values exhibit good alignment with the historic climatic data for this time of year as stated in Section 2.3 The minimum temperat ure, as expected, occurs around 06:30 just prior to sunrise Temperature o C PAGE 82 66 and the maximum temperature occurs at around 15:00 displaying the thermal lag effect of displacement from the maximum solar insolation at solar noon occurring as a result of the atmosphere being h eated from below. Table 3.3 lists the descriptive statistics of the values found in Figure 3.1 Examining Figure 3.1 it is interesting to note the steep slope of the curves between the hours of 06:30 and 11:00. This is the period of time when surface win ds are diminishing to their lowest value. Range Minimum Maximum Mean Std. Deviation Variance Urban Air Temperatures o C 10.93 23.16 34.10 28.66 3.76 14.21 Rural Air Temperatures o C 10.39 22.63 33.01 27.65 3.85 14.87 Table 3.3 Urban and rural daily temp erature descriptive statistics Figures 3.2 through 3.49 depict the spatiotemporal distribution of near surface air temperatures obtained by a spline interpolation of the point temperature data into a raster format. Distinctive spatial patter n s can be o bserved in Figures 3.2 through 3.49 Of interest is the rapid development of temperature differences shortly after sunrise at appro ximately 06:30 and continuing into the early after noon. In the afternoon the patterns become more muted possibly due to the influence of the sea bre eze. After sun set at approximately 20:30 nocturnal patterns become evident and continue through the night hours with a slow muting of the patterns until the following sunrise. PAGE 83 67 Figure 3.2 Near surface air temperatures 00: 00 Figure 3.3 Near surface air temperatures 00: 30. PAGE 84 68 Figure 3.4 Near surface air temperatures 01: 00. Figure 3.5 Near surface air temperatures 01: 30. PAGE 85 69 Figure 3.6 Near surface air temperatures 02: 00. Figure 3.7 Near surface air temperatures 02: 30 PAGE 86 70 Figure 3.8 Near surface air temperatures 03: 00. Figure 3.9 Near surface air temperatures 03: 30. PAGE 87 71 Figure 3.10 Near surface air temperatures 04: 00. Figure 3.11 Near surface air temperatures 04: 30. PAGE 88 72 Figure 3.12 Near surface air temperatures 0 5: 00. Figure 3.13 Near surface air temperatures 05: 30. PAGE 89 73 Figure 3.14 Near surface air temperatures 06: 00. Figure 3.15 Near surface air temperatures 06: 30. PAGE 90 74 Figure 3.16 Near surface air temperatures 07: 00. Figure 3.17 Near surface air temperat ures 07: 30. PAGE 91 75 Figure 3.18 Near surface air temperatures 08: 00. Figure 3.19 Near surface air temperatures 08: 30. PAGE 92 76 Figure 3.20 Near surface air temperatures 09: 00. Figure 3.21 Near surface air temperatures 09: 30. PAGE 93 77 Figure 3.22 Near surface air temp eratures 10: 00. Figure 3.23 Near surface air temperatures 10: 30. PAGE 94 78 Figur e 3.24 Near surface air temperatures 11: 00. Figure 3.25 Near surface air temperatures 11: 30. PAGE 95 79 Figure 3.26 Near surface air temperatures 12: 00. Figure 3.27 Near surface air temperatures 12: 30. PAGE 96 80 Figure 3.28 Near surface air temperatures 13: 00. Figure 3.29 Near surface air temperatures 13: 30. PAGE 97 81 Figure 3.3 0 Near surface air temperatures 14: 00. Figure 3.3 1 Near surface air temperatures 14: 30. PAGE 98 82 Figure 3.3 2 Near surface air temperatures 15: 00. Figure 3.3 3 Near surface air temperatures 15: 30. PAGE 99 83 Figure 3.3 4 Near surface air temperatures 16: 00. Figure 3.3 5 Near surface air temperatures 16: 30. PAGE 100 84 Figure 3.3 6 Near surface air temperatures 17: 00. Figure 3.3 7 Near surf ace air temperatures 17: 30. PAGE 101 85 Figure 3.3 8 Near surface air temperatures 18: 00. Figure 3.39 Near surface air temperatures 18: 30. PAGE 102 86 Figure 3.40 Near surface air temperatures 19: 00. Figure 3.41 Near surface air temperatures 19: 30. PAGE 103 87 Figure 3.42 Near surface air temperatures 20: 00. Figure 3. 43 Near surface air temperatures 20: 30. PAGE 104 88 Figure 3.44 Near surface air temperatures 21: 00. Figure 3.45 Near surface air temperatures 21: 30. PAGE 105 89 Figure 3.46 Near surface air temperatures 22: 00. Figure 3.47 N ear surface air temperatures 22: 30. PAGE 106 90 Figure 3.48 Near surface air temperatures 23: 00. Figure 3.49 Near surface air temperatures 23: 30. PAGE 107 91 3.3 Calculation of delta temperature ( A procedure similar to calculating the rural mean values was used to calculate the was then calculated by subtracting the rural 30 minute period mean from each individual urban 30 minute period mean. These val ues were then joined in a GIS database with ancillary location data and previously calculated centroid impervious surf ace percentage values. Table 3.4 provides a partial listing of this GIS sensor database (see Appendix B for a full listing) It should be noted that temperature s recorded by the sensors are in o F and were later converted to o C. Sensor Pole # Observed elevation 00:00 Delta T 03 :00 Delta T 23:00 Delta T 50m 100m 1 110718 Rural 0 20 feet 0.00 0.00 0.00 26.00 12.50 2 095513 Rural 21 40 fe et 0.00 0.00 0.00 11.33 5.50 3 2296646916 Residential 21 40 feet 1.48 1.62 1.34 32.67 40.00 4 2353746885 Residential 21 40 feet 1.10 1.13 1.14 33.67 37.00 6 2373347337 Residential 0 20 feet 0.02 0.03 0.12 48.67 60.50 7 2410947170 Residential 0 20 fee t 1.38 1.39 1.43 52.33 58.50 8 382847 Residential 0 20 feet 1.71 1.77 1.82 39.50 46.00 9 2593547923 Residential 21 40 feet 0.68 0.69 0.69 42.00 48.50 10 2729847384 Residential 0 20 feet 0.16 0.09 0.06 39.00 39.00 11 2766247545 Residential 0 20 feet 1.69 1.61 1.80 40.67 65.00 12 2279034542 Commercial 0 20 feet 1.62 1.57 1.68 29.00 39.00 13 2792245743 Commercial 0 20 feet 1.14 1.17 1.22 20.33 19.00 14 2764044450 Residential 0 20 feet 0.44 0.49 0.44 70.50 70.50 15 2817543955 Residential 0 2 0 feet 1.05 0.93 1.14 45.33 44.50 16 2786243398 Industrial 0 20 feet 0.91 1.04 0.89 36.00 36.00 17 signpost Industrial 0 20 feet 0.50 0.73 0.33 39.67 39.67 18 power pole Industrial 0 20 feet 0.28 0.36 0.06 31.00 30.00 19 2769243660 Industrial 0 20 feet 0.31 0.40 0.10 32.33 40.00 20 2703344016 Residential 0 20 feet 0.25 0.32 0.20 45.00 39.00 21 105408 Residential 0 20 feet 1.16 1.27 1.12 6.33 5.00 : : : : : : : : : 97 lamp in pa Commercial 0 20 feet 2.31 2.37 2.34 65.25 64.00 98 sun coast Rural 0 2 0 feet 0.00 0.00 0.00 30.00 30.00 99 1137 1 Rural 0 20 feet 0.00 0.00 0.00 5.25 8.50 100 2967048444 Rural 0 20 feet 0.00 0.00 0.00 4.67 4.00 Table 3.4 ArcMap attribute table including ( o F ) and i mpervious surface values ( percent ) PAGE 108 92 As part of the data analysis spatiotemporal maps were constructed of generated from the sample data. Figure s 3.50 to 3 .97 depict values in the Tampa study are a for each 30 minute time slot during a 24 hour day An examination of F igures 3.50 to 3.97 reveals the spatial and tempora l structure of the UHI in Tampa Bay R egion to be complex. As noted earlier, t he UHI of a study area may commonly be reported as a ran ge of values or a maximum (Yow and Carbone, 2006) or in some cases as semi concentric circles emanating from the CBD (Unger et al 2001) In reality, as exhibited by the Tampa Bay R egion UHI spatiotemporal plots, the UHI of a city or area is complex and nuanced and can not be expressed as a single number, range of numbers or even a simple concentric plot. As noted in Section 1.3 w hat many of the other studies may be lacking is a sufficient number of sensors or temperature readings to fully express the complexity of the UH I in their individual study area. This study has implemented a sufficiently dense network of temperature recording sensors such that complexities of the Tampa UHI may be observed. PAGE 109 93 Figure 3.50 rural urban delta temperatures 00:00. Figure 3.51 rural urban del ta temperatures 00:30. PAGE 110 94 Figure 3.52 rural urban delta temperatures 01:00. Figure 3.53 rural urban delta temperatures 01:30. PAGE 111 95 Figure 3.54 rural urban delta temperatures 02:00. Figure 3.55 rural urban delta temperatures 02:30. PAGE 112 96 Figure 3.56 rural urban delta temperatures 03:00. Figure 3.57 rural urban delta temperatures 03:30. PAGE 113 97 Figure 3.58 rural urban delta temperatures 04:00. Figure 3.59 rural urban delta temp eratures 04:30. PAGE 114 98 Figure 3.60 rural urban delta temperatures 05:00. Figure 3.61 rural urban delta temperatures 05:30. PAGE 115 99 Figure 3.62 rural urban delta temperat ures 06:00. Figure 3.63 rural urban delta temperatures 06:30. PAGE 116 100 Figure 3.64 rural urban delta temperatures 07:00. Figure 3.65 rural urban delta temperatures 07:30. PAGE 117 101 Figure 3.66 rural urban delta temperatures 08:00. Figure 3.67 rural urban delta temperatures 08:30. PAGE 118 102 Figure 3.68 rural urban delta temperatures 09:00. Figure 3.69 rural urban delta temperatures 09:30. PAGE 119 103 Figure 3.70 rural urban delta temperatures 10:00. Figure 3.71 rural urban delta temperatures 10:30. PAGE 120 104 Figure 3.72 rural urban delta temperatures 11:00. Figure 3.73 rural urban delta temperatures 11:30. PAGE 121 105 Figure 3.75 rural urban delta temperatures 12:30. Figure 3.74 rural urban delta temperatures 12:00. PAGE 122 106 Figure 3.76 rural urban delta temperatures 13: 00. Figure 3.77 rural urban delta temperatures 13:30. PAGE 123 107 Figure 3.78 rural urban delta temperatures 14:00. Figure 3.79 rural urban delta temperatures 14:30. PAGE 124 108 Figure 3.81 rural urban delta temperatures 15:30. Figure 3.80 rural urban delta temperatures 15:00. PAGE 125 109 Figure 3.82 rural urban delta temperatures 16:00. Fi gure 3.83 rural urban delta temperatures 16:30. PAGE 126 1 10 Figure 3.85 rural urban delta temperatures 17:30. Figure 3.84 rural urban delta temperatures 17:00. PAGE 127 111 Figure 3.87 rural urban delta temperatures 18:30. Figure 3.86 rural urban delta temperatures 18:00. PAGE 128 112 Figure 3.89 rural urban delta temperatures 19:30. Figure 3.88 rural urban delta temperatures 19:00. PAGE 129 113 Figure 3.91 rural urban delta temperatures 20:30. Figure 3.90 rural urban delta temperatures 20:00. PAGE 130 114 Figure 3.93 rural urban delta temperatures 21:30. Figure 3.92 rural urban delta temperatures 21:00. PAGE 131 115 Figure 3.94 rural urban delta temperatures 22:00. Figure 3.95 rural urban delta temperatures 22:30. PAGE 132 116 Figure 3.96 rural urban delta temperatures 23:00. Figure 3.9 7 rural urban delta temperatures 23:30. PAGE 133 117 An examination of the spatial and temporal patterns of the near surface air temperature and the delta temperature maps reveals what appear to be the existence of micro urban heat is lands (M U HI) with in the study area. Arising shortly after sunrise, the MUHI s are initially several kilometers in size and tend to exhibit spatial growth unti l approximately 12:30 EDT After 12:30 EDT the MUHI s begin to diminish corresponding to the increasing speed of th e local sea breeze. Sea breeze response is detailed in Section 5.0. The sea breeze muted afternoon MUHI patterns continue until sunset at 20:30 EDT at which time the structure of the noctu rnal UHI begins to emerge in the areas of higher impervious surface. The existence of large MUHI s during the day runs contrary to published literature which indicates that the maximum UHI is expressed during nocturnal hours. It should be noted that within the study area there are unique sets of local features that affect the spatial development of the MUHI s. Referring back to Figure 2.3, it can be seen that there are a series of small lakes orient ed in a north to south direction through the center portion of the study area. Because o f the very high thermal inertia of water, these lakes tend to exhibit stable temperatures on a diurnal basis. In addition the Tampa study area exhibits a cl assic example of urban sprawl with distinct pockets of residential housing and their associated strip malls and shopping centers. These pockets of urban sprawl contain areas of locally high percentages of impervious surface. These local features are beli eved to enable the development of isolated MUHI s. The emergence and expansion of the local MUHI s can be seen by an examination of Figures 3.67 through 3.74. Two hours after sunrise (Figure 3.67), seven distinct local MUHI s have emerged. PAGE 134 118 While several of th e MUHI s have developed in the more commercial and industrial urban areas the two northern and the one eastern MUHI have developed in areas that are predominantly residential. It is believed that these three residential MUHI s are being generated in respon se to localized areas of high impervious surface. Of particular note is the MUHI in the northeastern section of the study area. The two sensors within this MUHI were located in the median of a four lane boulevard that transected this residential area. Th e finding of an area of relatively high within a residential area was unexpected. It should be noted that while all sensors functioned properly during the study period, in the future it would be desirable to have additional sensors within the identified MUHI s. While the large number of senso rs deployed in this study enabled the identification of MUHI s within the study area, additional research will be required to more fully quantify the MUHI s and local factors that are influencing their development. In a more traditional general form, tempora l mean values for residential, industrial and commercial areas are presented in the following subsection. 3.4 Commercial, industrial and residential UHI comparisons As outlined in Section 1.6 different impervious surface categories were selected for evaluat ion. The designation of a commercial, industrial or residential impervious surface area was based on in situ observations taken during installation of the sensors. For the purposes of this study an area was deemed to be in the commercial impervious surface category if the observed surrounding area consisted of high rise office buildings, a large shopping center or larger strip mall. An area was deemed to be in the industrial impervious surface category if the observed surrounding area was primarily used fo r PAGE 135 119 manufacturing or shipping with building heights of 2 stories or less. An area was deemed to be in the residential impervious surface category if the observed surrounding area was composed of single family or duplex residences. The following figures (Fig ures 3.98 to 3.100) depict the commercial, industrial, and residential areas. A discussion of each figure will now be commercial impervious surface areas considered temporally. J ust after sunrise at 06:30 there is a marked drop in the rural to urban temperature difference. This drop can most likely be attributed to the greater thermal mass and heat capacity of the commercial area as compared to the rural area. Grimmond and Oke (199 9 pp 922) found that, Results indicate the storage heat flux is a significant component of the surface energy balance at all sites and is the greatest at downtown and light industrial sites. In addition, they found that there is a distinctive time lag be tween delta thermal storage and net radiative flux. Urban areas, with their greater thermal mass, have a slower response time (thermal hysteresis), requiring a greater amount of energy input for a corresponding rise in temperature as compared to a rural l ocation. PAGE 136 120 o C. As detailed in Oke (1982), rural areas will tend to express residual energy flux as increased latent heat of evapotranspiration, and a smaller value as sensible heat, whereas the urban area with its drier environment will express residual energy flux primarily as sensible heat. After initial thermal capacitance charge, the difference in generation of sensible heat will result in the rural to urban temperature differences being greater. Commercial areas, with higher percent age of impervious surface areas will show a higher sensible heat value as compared to rural areas. This results in higher temperature values in commercial areas as compared to rural areas. After sunset there is another increase in 24 hour mean delta temperature for high density sensor locations PAGE 137 121 values. This is due to the greater heat storage of the commercial area as compared to the rural area. The heat stored during the day is released during the nocturnal hours. The nocturnal maximum in commercial areas is approximately 1. 12 o C. areas considered temporally. An examination of Figure 3.99 reveals that the rural to ( in a similar fashion to commercial area s) s This is most likely due to a lower thermal mass and heat capacity as compared to commercial area s With a lower thermal mass and heat capacity, temperature s in in dustrial area s recover quicker than was exhibited by temperatures in commercial area s U sing Equation 1 as compared to the rural area, the evapotranspiration rate in the industrial area is lower resulting in a higher sensible heat and a higher recorded te mperature compared to the rural area. However, the lower thermal mass and heat capacity of industrial area s compared to commercial area s results in less stored heat being released in the nocturnal 19:30 remain unexplained at this time. There is a possibility that they are related to a reduction in anthrop ogenic heat at the end of a work day. Further study will be required. The nocturnal maximum in the industrial area is approximately 0.72 o C. PAGE 138 122 o C. alues for residential impervious surface areas values decline shortly after sunrise ( in a similar fashion to the commercial and industrial areas ) With a lower thermal mass and heat capacity the residential area temperatures tend to recover even quicker than was exhibited by the commercial and industrial areas. The nocturnal maximum in residential area s is approximately 0.81 o C. It would normally be expected that a residential area would exhibit a lower nocturnal UHI value 24 hour mean delta temperature for medium density sensor locations PAGE 139 123 than an industrial area. However it is postulated that the previously identified MUHI s in the residential areas are affecting the mean residential UHI values. Figure 3.100 Mean residential imp e o C. Figur e 3.101 Tampa study area for each 30 minute period over twenty four hours It can be seen from F he study area is approximately 0.824 o C. Expressed another way, the daily mean UHI in the Tampa study area is approximately 0.824 o C 24 hour mean delta temperature for low density sensor locations PAGE 140 124 Figure 3.101 o C 3.5 Summary The spatial/temporal plots of t values show that the expression of the UHI in the Tampa study area varies values corresponding to areas of higher percentages of impervious surface. In contrast nd to correspond to areas of lower percentages of impervious surface and areas of vegetation and lakes The discovery of MUHI s within the study area and the tend to support the first hypot hesis that, The measured UHI delta temperature in the Tampa study area will show both spatial and temporal variability. o C PAGE 141 125 4.0 RELATIONSHIPS OF IMPERVIOUS SURFACE TO THE UHI A review of the relevant literature suggests that the percentage of impervious su rface in t he urban environment may be a good predictor of UHI intensity. Studies by Xian and Crane ( 2005, 2006) and Yuan and Bauer (2007) both suggest that the percentage of impervious surface in an urban area is a useful source of data as it relates to u rban thermal patterns and LST It is hypothesized that there is a significant relationship between the percentage of impervious surface and the intensity of the UHI in Tampa, Florida. As a first method of analysis, the statistical package SPSS was used to perform a basic correlation analysis of the percentage of impervious surface calculated at the different sample radii as defined in Section 2 .6 e sensor locations. As was stated in Section 2 .6 within the literature no definitive influential distance relating impervious surface to temperature was found. Individual correlation analysis was performed on the four chosen distance s in an effort to identify the proper distance for analysis. It should be noted that all temperature values were logged by the data sensors in o F T o F. Conversion from the Fahrenheit to Cel sius was performed for the final outputs. Correlation values were calculated for all 48 half hour time periods. The 00:30 time was found to have the highest correlation values. Table 4.1 lists the r esults for the 00:30 time A complete listing of the t ime varying correlation v alues can be found in Appendix C PAGE 142 126 Radius of sample area near sensor location (m) n Pearsons r Correlation P value Two tailed 100 94 .544 <0.001 50 94 .579 <0.001 30 94 .550 <0.001 25 94 .442 <0.001 Table 4.1 Correlation r A simple plot of the correlation values at 0 0:30 (Figure 4.1 ) would tend to indicate that the temperature correlation with percentage of impervious surface decline s either side of the 50 meter distance value. This is the case for all calcu lated correlations for all time periods as shown in Appendix C While not definitive the 50 meter value of percent impervious surface may reflect the extent of influence that impervious s urface has on temperature and therefore is utilized in subsequent calculations. Figure 4.1 Correlation v alues at differing radii values. Utilizing the 50 me ter impervious surface values a correlation calculation was performed to com pare specifically the 50 meter percent impervious surface va lues and all PAGE 143 127 of the 30 minute time pe graphically depicts the result of this analysis. Examining Figure 4. 2 it is interesting to note the rapid drop in correlation values shortly after sunrise at 06:30 and the continued low corre lation values until approximately solar noon at 13:30 at which time the correlation values begin rising. Figure 4.2 Correlation values of 50 meter impervious surface and period A similar trend was no ted in the Section 3. 2 T he rapid decline in the correlation values after sunrise may be attributed to a lessening of the contribution of stored heat in the impervious surface and an increase in heating attributable to solar insola tion. As the solar insolation begins to decrease after solar noon, the stored heat energy in the impervious surfaces would tend to contribute more to atmospheric heat ing. After sunset at 20:30 solar insolation is absent while the correlation Sample_time 1: 00: 00 0: 00: 00 23: 00: 00 22: 00: 00 21: 00: 00 20: 00: 00 19: 00: 00 18: 00: 00 17: 00: 00 16: 00: 00 15: 00: 00 14: 00: 00 13: 00: 00 12 : 00: 00 11: 00: 00 10: 00: 00 9: 00: 00 8: 00: 00 7: 00: 00 6: 00: 00 5: 00: 00 4: 00: 00 3: 00: 00 2: 00: 00 1: 00: 00 0: 00: 00 23: 00: 00 Correlation 0.600 0.500 0.400 0.300 0.200 0.100 0.000 Sunrise 06:30 Sunset 20:30 Solar noon 13:30 Sample time PAGE 144 128 values contin ue to rise until they reach a maximum at approximately 00:30 It should be noted that the correlation values are fairly stable during the nocturnal hours. Working w ith the 00:30 and the 50 meter percent impervious surface values, a regression model was generated The regression model o utputs are shown in Tables 4.2 through Table 4.5 A complete listing of the regression model outputs for all time per iods can be found in Appendix D Mean Std. Deviation N T_0_30 1.437 1.044 94 percent impervious 42.970 18.720 94 Table 4.2 r egression model with constant d escriptive statistics Model R R Square Adjusted R Square Std. Error of the Estimate 1 .579 .335 .327 .856 Tab le 4.3 r egression model with constant summary Model Sum of Squares df Mean Square F Sig. 1 Regression 33.974 1 33.974 46.282 .000 Residual 67.534 92 .734 Total 101.508 93 Table 4.4 r egression model with constant ANOVA Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Std. Error 1 (Constant) .050 .222 .227 .821 Imp_percent_50m .032 .005 .579 6.803 .000 Table 4. 5 r egression model with constant coefficients Examining Tabl e 4.4 Anova results the calculated global F test has a value of 46 .282 indicating that there is a significant relationship between the percentage of impervious surface and the temperature difference between the rural and urban measured temperatures. In t he case of the 00:30 time period the two tailed p value is < 0.001 An examination of the coefficients in Table 4.5 reveals that while the constant 0 of the PAGE 145 129 regression equation is very close to zero its p value is high 0 is not 1 coefficient is highly significant with a two tailed p value of < 0.001 As a check for normality the residuals from the regre ssion analysis were investigated. Table 4.6 lists the r esidual statistics of the 00:30 time period regression analysis while Figure 4.3 is a histogram of the residuals and Figure 4.4 is a P P plot of the residuals. From Table 4.6 it can be seen that the mean value of the residual s the standard predicted value and the standard residual value are all zero which would tend to indicate a normal distribution of the residuals. This can be confirmed visually by examining th e histogram plot of Figure 4.3 which also depicts a fairly normal distribution of the residuals Furthermore, an examination of Figure 4.4 which shows a P P plot of the residuals, indicates that the expected and observed cumulative probabilities approximate a linear function. The values in these figures and table tend to support the validity of the derived regression equation. Dependent Variable: T_0_30 Minimum Maximum Mean Std. Deviation N Predicted Value .270 2.879 1.437 .604 94 Residual 2.631 1.775 .000 .852 94 Std. Predicted Value 1.932 2.385 .000 1.000 94 Std. Residual 3.071 2.072 .000 .995 94 Table 4.6 r egression model with cons tant residual statistics PAGE 146 130 Regression Standardized Residual 2 0 2 4 Frequency 15 10 5 0 F igure 4.3 Histogram of r egression model with constant standardized residuals Observed Cum Prob 1.0 0.8 0.6 0.4 0.2 0.0 Expected Cum Prob 1.0 0.8 0.6 0.4 0.2 0.0 Figure 4.4 P P p lot of the r egression model with constant standardized residual s PAGE 147 131 In an effort to more fully investigate the relationshi p between the percentage of results of t he first regression model contained a 0 constant that was very close to zero but was not significant T herefore the second regressi on model was run with a zero crossing. This model did not contain a 0 term. Tables 4.7 to 4.10 list the results of the second regression model run. A complete listing of the second regression model outputs for all time per iods can be found in Appendix E Figure 4.5 is a histogram of the residual values of regression model 2. Model R R Square Adjusted R Square Std. Error of the Estimate 2 .878(b) .772 .769 .852 Table 4.7 2007 r egression model without constant summary Model Sum of Squares df Mea n Square F Sig. 2 Regression 228.265 1 228.265 314.164 .000 Residual 67.572 93 .727 Total 295.837 94 Table 4.8 2007 r egression model without constant ANOVA values Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta B Std. Error 2 Imp_percent_50m .033 .002 .878 17.725 .000 Table 4.9 2007 r egression model without constant coefficients Minimum Maximum Mean Std. Deviation N Predicted Value .226 2.915 1.429 .622 94 Residual 2.611 1.788 008 .852 94 Std. Predicted Value 1.932 2.385 .000 1.000 94 Std. Residual 3.064 2.098 .009 1.000 94 Table 4.10 2007 r egression model without constant residual statistics PAGE 148 132 Regression Standardized Residual 2 0 2 4 Frequency 12.5 10.0 7.5 5.0 2.5 0.0 Figure 4.5 Histogram of r egression model without constant standardiz ed residuals Because of the differing methods in calculating the model 1 and model 2 regression values a direct comparison of the R 2 values of the first and second regression models is not valid However the residual statistics and histogram plots of the second model tend to indicate a slightly better fit with a normal distribution of residuals T his might be explained by the fact that the rural sensors in the regression m odel percentage of imperv ious was a small non zero value. T 0 constant of the first regression model. Ideally zero percent impervious surface surrounding it. Given that the 0 constant in the first regression model is not significant and should not be included in the regression equation and given that the 1 values are nearly identical in both regression models, t he second regression model PAGE 149 133 would appear to be the better fit. The resultant regression equation from the second model can be written as follows. o F = 0.033 .002 ( 50meter impervious perce ntage) (8 ) o C = 0.018 .0011 ( 50meter impervious perc entage) (9 ) w here As a validation of the 2007 regression results 00:30 values and percentage of impervious surface values were run through an identi cal OLS regressi on evaluation. Tables 4.11, 4.12 and 4.13 list the results of the regression analysis of the 2008 data. Model Summary R R Square Adjusted R Square Std. Error of the Estimate .878 .771 .763 1.199 Table 4.11 2008 r egression model witho ut constant summary ANOVA Sum of Squares df Mean Square F Sig. Regression 140.271 1 140.271 97.571 .000 Residual 41.691 29 1.438 Total 181.962 30 Table 4.12 2008 r egression model without constant Coefficients Coefficients Unstandardiz ed Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50 m eter .044 .004 .878 9.878 .000 Table 4. 1 3 2008 r egression model without constant coefficients PAGE 150 134 The equation for the regression analysis of the 2008 data is shown below. o C = 0.02 4 .004 (50meter impervious percentage) (10 ) The resulting equation for the 2008 data regression analysis compares quantitatively with the equation for the 2007 regression analysis. Complete resu lts can be found in Appendix F Finally t percentage of impervious values were subjected to a similar OLS regression analysis T he results are shown below in Tables 4.14 4.15, and 4.16 Model Summary a R R Square Adjusted R Square Std. Error of the Estimate .877 .769 .767 .966 Table 4.14 2007 and 2008 c ombined r egression model without constant summary ANOVA a Sum of Squares df Mean Square F Sig. Regression 366.293 1 366.293 392.456 .000 Residual 110.133 118 .933 Total 476.426 119 Table 4.15 2007 and 2008 c ombined r egression model without constant ANOVA values. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50 m eter .036 .002 .877 19.811 .000 Table 4.16 2007 and 2008 c ombined r egression model without constant coefficients PAGE 151 135 The resultant equation for the regression analysis of the combination of the 2007 and 2008 data is shown below. Complete results can be found in Appendix G o C = 0.020 .002 (50meter impervious percentage) (11 ) The results presented in this s ection show that there is a significant relationship between percentage of impervious surface at the sample points i n the Tampa study area. In an effort to examine possible autocorrelation within the data that might affect the regression equations, the software package GeoDa version 0.95 was utilized to look for evidence of autocorrelation within the data. Given that Morans I is widely accepted as a measure of autocorrelation values. Utilizing the GeoDa weights file creation routine it was found that the minimum lag distance which en sured that every sample point would have a neighbor in the weight table, was 4572 me ters This was rounded to a lag distance of 5000 meters and a weights file was created for use in the Morans I calculation and subsequent spatial regression analysis. Additionally Morans I values were calculated at differing distance values to verify t hat the selected 5000 meter s value indicated the greatest degree of autoc orrelation ( Figure 4. 6). Running a Morans I calculation on the 00:30 Morans I value which ranges from 1 to +1, was 0.4125 indicating a moderate degree of autocorrelation exists within the data. In an effort to further identify any autocorrelation a GeoDa classic spatial regression meter impervious surface percentage. These are the same time period and percentage of impervious surface values that were used in the SPSS regression analysis earlier in this PAGE 152 136 section The results of the Geo Da spatial regression indicated that there was a significant spatial dependence in the lag distance. Figure 4.6 Morans I for various distance weighting A GeoDa spatial lag model regression ana lysis was run to correct for spatial lag dependencies. T he spatially corrected regression equation for the 2007 meter impervious su rface percentage is shown below, p < 0.001 o C = 0.014 .004 (50meter impervious percentage) (12 ) Utilizing the same process the resultant spatially corrected regression equation f or the 2008 da ta is shown below, p < 0.001 o C = 0.017 .00 4 (50meter impervious percentage) (13 ) PAGE 153 137 As a verification of the spatial lag model regression analysis of the 2007 and 2008 data, the resultant residuals of the spatial lag model re gression analysis were examined for continued evidence of auto correlation. A univariate Morans I was calculated for the lag model regression residuals Figure 4. 7 depict s the results for the 2007 data with 2008 results being nearly identical. To verify that the Morans I which was calculated for the lag model regression residuals was not significant, 999 permutations were run; the results are shown in Figure 4.8 Figure 4. 7 Lag residuals Morans I calculation Figur e 4. 8 Lag residuals Morans I permutation run PAGE 154 138 An analysis of the spatial lag model regression residuals in which the permutation p value is not significant would t end to indicate that Equation 12 does not exhibit spatial autocorrelation T he same conclu sion can be said of Equation 13 To arrive at an approximation of the equations representing the combined 2007 and 2008 data the mean of the equation coefficients was calculated. The resultant equation is, o C = 0.016 .004 (50meter impervious percentage) (14 ) It should be noted that the 2007 and 2008 data could not be combined and a spatial regression model run to calculate a combined 2007 and 2008 equation because the sensor locations were coincident between the years and would have shown an extreme amount of autocorrelation that would incorrectly bias any resulting regression equation. Equations 12 and 13 should be considered to represent an accurate re presentation of the re percentage of impervious surface in the study area for their corresponding sample period Because of the coincidence of the data points in the 2007 and 2008 data E quation 1 4 is believed to be accurate but cannot be independen tly verified with the data at hand. As to what is the nature of the autocorrelation within the study data, the SPSS percentage of impervious surface at the sample point, wh ile the GeoDa Morans I and ed that some of the influence on attributed to surface winds within the study area tr ansferring he at between adjacent locations. It is postulated that as surface wind speeds decrease the autocorrelation ( as PAGE 155 139 expressed by Morans I ) will also decrease Likewise t he coefficient values of the spatially corrected regression equation will appr oach the uncorrected coefficient values of the SPSS regression equation. Further study, with wind speed sensors incorporated with temperature sensors at sample sites, will be required to validate this postulation. Additional examination of the collected an d calculated data revealed other interesting comparisons Figure 4 9 is a plot of the twenty four hour regression model R 2 and 1 values An examination of Figure 4. 9 reveals that there is a rapid drop in R 2 values shortly after sunrise that mirror s the Pe arsons r co rrelation response of Figure 4.2 and commercial industrial and residential locations. The inference of the lower R 2 values after sunrise is that the impervious surface is providing a smaller contribution to the ove rall would be expected given that solar insolation is the major contributor to temperature values during the daylight hours. After a short lag period the R 2 values do begin to rise but never to the level that is seen in the noct urnal hours. PAGE 156 140 Time 23: 00 22: 00 21: 00 20: 00 19: 00 18: 00 17: 00 16: 00 15: 00 14: 00 13: 00 12: 00 11: 00 10: 00 09: 00 08: 00 07: 00 06: 00 05: 00 04: 00 03: 00 02: 00 01: 00 00: 00 Second regression model R squared values 0.800 0.600 0.400 0.200 0.000 Second regression model Beta 1 values 0.100 0.080 0.060 0.040 0.020 0.000 0.472 Mean R squared value Mean Beta 1 value 0.0295 Figure 4.9 r egression model without constant R 2 and 1 values T he lag and then rise in R 2 v alues might be attributed to the initial thermal lag of the impervious surface Beginning at sunrise the incoming solar radiation provides sh ort wave energy to the impervious surfaces and ground. After the thermal lag is overcome the impervious surface s begin releasing long wave heat energy in the atmosphere as latent heat flux This is indicated in Figure 4.9 by the rise in R 2 values. One feat ure of note in Figure 4. 9 is the drop in the R 2 value between 15:00 and 17:30 In the study area it is common during the summer months to have daily thunderstorms that occur between the hours of 15:00 and sunset. This may be causing the drop in R 2 values d uring this time period. EDT PAGE 157 141 4.1 Summary A significant relationship ( p < 0.001 ) between the percentage of impervious surface was found. This relationship holds for the 2007 and 2008 sample periods as well as a combined 2007 and 2008 data set. C omplete l isting s of the separate 2007 and 2008 regression va lues and the combined 2007 and 2008 data set regression results are provided in Appendices D, E, F and G Additionally w hile the literature ( Oke 1982, Bornstein and Lin 2000, Dixon and Mote 2003 ) tends t o indicate that the maximum UHI is normally expressed in the nocturnal hours the rise in 1 values of Figure 4.9 tend to indicate that an impervious surface related UHI is also exhibited in the day time hours Within the Tampa study area there appears to be a daytime UHI that is promoted by the percentage of impervious surface in the study are a and for certai n time periods is a value greater than the nocturnal UHI value. This runs contrary to some of the published literature. This daytime UHI i s examined further in Section 5 1. The results of the impervious surface regression models with the re lationships showing significance at p <0.001 indicate s that at the height of the nocturnal heat island the relationship between the percentage of impervious surface and the rural to urban temperature difference can account for 34 percent and 77 perce nt of t he nocturnal UHI in the Tampa study area depending on the regression model (with or without a 0 constant). Accepting the zero crossing regression model (R. Chandler University College of London personnel communication May 12 th 2008) the percentage of impervious surface within the Tampa study area can account for approximately 77 percent PAGE 158 142 [uncorrected] or 55 percent [spatially corrected] of the nocturnal UHI in the Tampa study area. Based on the 2007 and 2008 data sets the equation for the impervious surface contribution to the Tampa study area UHI is shown below. o C = 0.020 .002 (50 meter radius impervious percentage) [uncorrected] (15 ) o C = 0.016 .002 (50 meter radius impervious percentage) [spatially corrected] (16 ) The results of the analysis on the percentage of within the study area would tend to support the second hypothesis that, There is a significant relationship between the percentage of impervious surface and the intensity of the UHI in Tampa, Florida. While the relationship between the percentage of imp the UHI in the Tampa study area is valid, due to the unique features of Florida ( discussed in Chapter 1 .0 ) the coefficient values of the relationship equation may need to be modified if the equation is employed in other regio ns and times PAGE 159 143 5.0 R ELATIONSHIP OF WIND SPEED TO UHI During a review of the relevant literature it was found that increasing wind speeds in the urban environment can have a negative effect on the development of the UHI Escourrou (1991) no ted that the re appeared to be a decrease in the rural to urban temperature differences with increasing wind speed s with wind speeds > 5 ms 1 having a rapidly decreasing effect on the delta temperature. In addition r esearchers Morris and Simmonds (2002) showed that w ind speeds over 2 ms 1 resulted in a statistically significant reduction of the UHI magnitude. Based on this prior research the values for the entire Tampa study area w ere examined and compared with the wind s peeds recorded at the Tampa Internatio nal Airport weather station Figure 5.1 is a dual axis plot of the 2007 mean Figure 5.1 Tampa study area (2007) o C PAGE 160 144 It should be noted that the topography at the Tampa Internatio nal A irport is relatively flat with few obstructions. Because of this flatness and lack of obstructions the recorded wind speeds tend to be higher at the Tampa International Airport site than wind speeds found in the more vegetated and built environment of other parts of the Tampa study area where frictional influences would tend to reduce the wind speed In an effort to identify a relationship bet t and wind speed values were subjected to regression analysis Examining possible relationships, Figure 5.2 s the mean wind speed with the linear, logarithmi c, inverse, and quadratic regression equation plots superimposed on the data. Figure 5.2 Curve fit p lots PAGE 161 145 A n inspection of Figur e 5.2 w ould tend to indicate that of the four fit equations, the quadratic rel ationship appears to be a better fit to the da ta over the given range of wind speeds A closer examination of the quadratic model s ummary statistics of Tables 5.1 and 5.2 reveals that the 2 term has a p value of 0.064 W hile this is slightly higher than a p <0.05 the p value is still within a 10 percent significance which was defined as the significant relationship criteria in Section 2.8 Model Summary R R Square Adjusted R Square Std. E rror of the Estimate .659 .435 .410 .684 Table 5.1 Quadratic m odel s ummary Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Wind Speed 5.267 2.319 3.736 2.272 .028 Wind Speed ** 2 .658 .347 3.119 1. 897 .064 (Constant) 11.356 3.754 3.025 .004 Table 5.2 Quadratic c oefficients According to Mc Clave and Sincich (2006) data collected in the field sometimes exhibits non normal distributions T hey suggest that because of this as a minimum, a natural l og transform should be evaluated Therefore p rior to selecting the quadratic model as the best fit for the relationship between the UHI mea and wind speeds, a PAGE 162 146 natural log transform of the wind speed was implemented. Tables 5.3 to 5.5 list the results of this analysis. Model Summary Model R R Square Adjusted R Square Std. Error of the Estimate 1 .639 a .408 .395 .69174 Table 5.3 Natural l og model s ummary ANOVA b Model Sum of Squares df Mean Square F Sig. 1 Regression 15.181 1 15.181 31.727 .000 Residual 22.011 46 .479 Total 37.193 47 Table 5.4 Natural l og ANOVA values Coefficients Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta 1 (Constant) 4.856 .612 7.932 .000 LN (Wind Speed) 2.939 .522 .639 5.633 .000 Table 5.5 Natural log c oefficients It can be seen that the F values and t values of the natural log transform regression are greater than the quadratic regression fit with the 0 and 1 having a p value <0.001. The question then arises as to which regression equation better describes the relationship and the mean wind speed. The R 2 value of the quadratic PAGE 163 147 regression equation is 0.435 while the value of R 2 for the linear natural log wind speed transform regression equation is 0.408. This indicate s that the quadratic equation accounts for more of the relationship compared to the natural log transform In addition the quadratic equation fit line appears to conform better to the data. Given the fact that previous res earch for other areas has shown the hip to be non linear the quadratic regression equation mean wind speed relationship in the research area. Accounting for approximately 44 percent of the relationship mean wind speed relationship can therefore be defined as, (17 ) where in o C ( windspeed ) = the recorded wind speed in ms 1 As a validation of E q uation 17 a similar regression analysis was run on the values and the mean wind speeds at the Tampa International Airport for the sample period 5/14/2008 to 7/1/2008. In order to make a direct comparison to the 2007 regression analysis (due to the range of wind speeds in 2007 being 2.1 to 4.3 ms 1 ), and due to the prior research that suggests ( Section 1.7 ) a significant relationship exists at >2.0 ms 1 ( though these researchers used binned groups of data rather than a continuous data set ) wind speeds < 2.1 were not considered in the 2008 data (despit e the range of wind speeds in 2008 being 2.0 ms 1 to 4.33 ms 1 ). Again as in 2007 the quadratic model resulted in the highest R 2 and the most significant t values Tables 5.6 to 5.8 list the regression values for the 2008 quadratic regression model. PAGE 164 148 M odel Summary R R Square Adjusted R Square Std. Error of the Estimate .533 .284 .244 .462 Table 5.6 2008 Q uadratic model summary ANOVA Sum of Squares df Mean Square F Sig. Regression 3.054 2 1.527 7.145 .002 Residual 7.694 36 .214 Total 10.748 38 5.7 2008 Anova values Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta ms 1 4.521 1.240 5.943 3.646 0.001 ms 1 ** 2 .691 .195 5.780 3.545 0.001 (Constant) 9.009 1.897 4.749 .000 5.8 2008 Coeffi cients summary T he 2008 mean (18 ) While similar to the 2007 the on accounts for a lower percentage (29 percent for 2008 versus 44 percent for 2007) and mean wind speed. In addition there are differences in the coefficient values. Both the 2007 and the 2008 relationship have a p va lue of <0.064 and as such are significant. It is possible that some of the differences can be attributed to the differences in sample PAGE 165 149 period starting dates. It should be noted that due to the data constraints of not being able to evaluate wind speed data p oints below 2.1 m s 1 or greater than 4.3 m s 1 for both 2007 and 2008 t he model s are only valid for wind speed s between 2.1 and 4.3 m s 1 The increasing wind speed range is 2.3 o C which compares favorably with r esearch conducted by Escourrou (1991) When examining the relationship between UHI and wind speed Morris and Simmonds (2002 p 175 ) f ound that, T here was a statistically significant difference (at the 95 percent confidence interval) between the mean UHI of each of the four wind speed groups >2.0 ms 1 Their results are supported by the finding s of this study. In future research it would be high ly desir able to have wind speed data available from each sensor site. In addition it would be valuable to have wind data in the range of 0 to 2 ms 1 as this is the wind speed range that O ke (1987) and Escourrou (1991) reported the greatest increase in UHI Lacking this data Equation s 1 7 and 1 8 must be considered as only partial solution s to the relationship and wind speed in the study area 5 .1 Sea breeze winds affect on UHI As was detailed in Section 1.7 a common summertime phenomenon of the Florida pe ninsula is the daily sea breeze The daily west coast sea breeze brings unique wind conditions to the Tampa study area. As was shown in Section 5.0 there is a significant An examination of the mean 24 hour wi nd profile from 6/1/2007 to 7/17 /2007, in the Tampa study area reveals an interesting wind speed and direction profile Figure 5.3 is a dual axis plot of the wind speed and wind dir ection during the above mentione d time period PAGE 166 150 Figure 5.3 Tampa International Airport wind speed and direction 6/1/2007 to 7/17/2007 Beginning at midnight it can be seen that the wind speed is diminishing while the wind direction is shifting towards the south. This trend continues after sunrise at 06:30 until approximately 10:30 at which time there is an abrupt change in wind speed and direction indicating the onset of the sea breeze The wind speed begins to increase and the wind direction begins to shift to a more westerly direct ion. As seen in Figure 5.3 the sea breeze maximum spee d is reached just before sunset After sunset the wind speed begins to decline and the wind direction begins its shift to a more southerly direction. It is of interest that there is not an abrupt onset of a land breeze during the nocturnal hours as was the case of the sea breeze rather there is a slow shifting of wind direction towards the south. Degrees PAGE 167 151 The Tampa International Airport recording site is located with in the Tampa study area and is situated close to the western boundary of the study area. To obtain a broader perspective w ind speed and direction data were obtained from the NCDC for two sites outside of the study area. The first site was at MacDill Airforce Base located directly to the south of t he study area and the second site was at Fort Howard Park which is located northwest of the study area Figure 5.4 depicts a d ual axis plot of the MacDill Airforce Base wind speed and direction during the same period and Figure 5.5 is a plot of the wind s peed and direction recorded at Fort Howard Park. It can be seen that the wind speed and d i rection profiles of Figures 5.4 and 5.5 closely match the wind speed and direction profile of the Tampa International Airport site. It should be noted that there is an approximately one hour time shift in the Fort Howard Park profile. Based upon the above data it would appear that the west coast sea breeze near the study area begins to manifest itself at approximately 10:30 The low wind speed of the day coincides with the time when the thermal lag has been overcome as shown in Figure 3.101. The result is the generation of the highest recorded Figure 5.4 Ma cDill AFB wind speed and direction 6/1/2007 to 7/17/2007. Figure 5.5 Fort Howard Park wind speed and direction 6/1/2007 to 7/17/2007. Degrees Degrees PAGE 168 152 spatiotemporal patterns of Figure s 3.70 to 3.75 window the wind speeds beg alues is evident in Figures 3.79 to 3.84 After 13:00 This is most likely the result of the smearing effect of the wind speed as the wind transfers heat from one place to another The 2008 wind speed profiles (Figure 5.6) were similar to the 2007 wind speed profiles with sim ilar It appears that the unique topographic features and meteorological conditions of Florida, along with the impervious surface in the study area, allow for the brief generation of a large daytime UHI just prior to the recorded onset of the sea breeze. Figure 5.6 2008 Tampa International Airport wind speeds and direction Degrees PAGE 169 153 5.2 Summary The results of the wind speed regression models with a significance of p <0.064 indicates that there is a significant relationship between wind speeds and the rural to urban temperature difference in the Tampa study area. In 2007 t his relationship would tend to account for up to an approximately 44 percent reduction of the UHI Tampa study area. Shown below is the equation for the 2007 wind speed contribution to the Tampa study area UHI (19 ) Results of the 2008 sample period analy sis indicate that in 2008 the relationship would tend to account for approximately 29 percent reduction of the UHI Tampa study area. Shown below is the equation for the 2008 wind speed contribution to the Tampa study area UHI (20 ) T he 2007 and 2008 results support the findings of other studies by Escourrou (1991) and Morris and Simmonds (2002) as to the significance of wind speed in UHI modification. The results of the analysis on the Tampa study area wind speeds and the within th e study area would tend to support the third hypothesis that, Surface winds tend to moderate the intensity of the UHI in Tampa, Florida. It should be noted as was previously mentioned the model is only valid for wind speed s between 2.1 and 4.3 m s 1 Ov er this wind speed range an increase in wind speed Furthermore due to the fact that one location is used to provide the wind value in the analysis these equations for wind can only be used to des cribe the data obtained and can not be used for predictive purposes. PAGE 170 154 6 .0 CONCULSIONS T his research has shown that the overall Tampa Bay Region study area exhibits a m oderate UHI H owever unlike other studies; this research has shown the actual spatial/temporal characteristics of the UHI and has found them to be complex in nature and dependent in a large part on the percentage of impervious surface present at a location. This complexity is evident by the existence of MUHI s within the study area. These findings support the first hypothesis that The measured UHI delta temperature in the Tampa study area will show both spatial and temporal variability. This study has shown that there is a significant relationship between the percent age of impervious surface at a location at that location It has also shown that the study area exhibits a daytime UHI which is greater than the nocturnal UHI. T he unique topographic features and meteorological conditions of Florida along with the impervious surface in the study area allow for the brief generation of a large daytime UHI just prior to the recorded onset of the sea breeze. This runs con trary to work by others (Oke 1982, Bornstein and Lin 2000 and Dixon and Mote 2003 ) who indicate that the maximum UHI is expressed in the nocturnal hours. Additionally regression model was utilized to correct for the autocorrelation with the result showing a non significant Morans I value These findings support the second hypothesis that There PAGE 171 155 is a significant relationship between the percentage of impervious surface and the intensity of the UHI in Tampa, Florida. In addition to the investig ation of the relationship between the percentage of impervi it has been shown that there is a significant relationship between the wind speed in the Tampa Bay Region study area and the moderation of the Tampa Bay Region UHI values. Based upon the analysis of 2007 and 2008 wind s peed s recorded at the Tampa International Airport and the rural to urban calculated for each sensor location in the study area, it was found that a significant relationship exists between wind speeds in the range of 2.1 ms 1 and 4.3 ms 1 The rel ationship between wind speeds and to confirm the findings of other authors such as Escourrou (1991) and Morris and Simmonds (2002) that wind speeds > 2 ms 1 have a significant moderating effect on the development of a UHI This research has also d ocumented the temporal profile of the sea breeze and wind speed moderation of the Tampa UHI. These findings support the third hypothesis that There is a significant relationship between the speed of the surface winds and the intensity of the UHI in Tampa, Florida. I n addition the temporal characteristics of the Florida sea breezes contribute to a uniquely modified temporal UHI profile in the Tampa, Florida study area. While these results are valid for the Tampa study area the myriad of conditions that ar e unique to Florida and the Tampa area may preclude the direct transfer of the coefficients of the relationships to other areas However, it is believed that the underlying re lationships should exist and may be valid in other urban areas PAGE 172 156 In regards to win d speed, a recent article by Pryor (2009) examined the trend in wind speed s across the contiguous United States in response to climate change Pryor (2009) found that wind speeds have been decreasing from 0.5 percent to 1.0 percent a year since 1973. While these findings can have an impact on the wind energy industry which was the focus of their paper, they can also have an impact on UHI. As has been shown in this study higher wind speeds have a moder ating effect on the intensity of the UHI If the Pryor (2009) research is correct lower expected future wind speeds would have a lower moderating effect on UHI resulting in an overall higher UHI value. A key question is then whether the results of this research can be utilized by urban planners and politic al leaders in the re development of the current urban environment and the development of future urban landscapes. As discussed in Section 1.8 t he choice of lighter colored building and roofing materials along with lighter colored pavement would have a dire ct effect on lowering the urban environment sensible heat and the consequent UHI T here are asphalt and concrete blends that are porous and allow for the infiltration of rain water T raditional paving materials such as cobblestone and brick also allow rai n water to seep in to the subsurface If these materials are utilized they will decrease surface runoff and increase infiltration. The porous characteristics of these materials will allow for the evaporation of infiltrated surface water thus increasing the l atent heat of evaporation in the energy balance equation resulting in decreased sensible heat. A nother simple and cost effective method of decreasing the UHI temperature is the planting of trees and vegetation in the urban areas with high percentages of impervious materials ( Emmanuel et al. 2007). Trees provide a dir ect shading effect and thus reduce solar insolation impinging on the imperv ious surfaces. I n addition tree s and vegetation PAGE 173 157 increase the transpiration rate in the energy balance equation als o resulting in a lowering of sensible heat. This idea is promoted by Rosenzweig et al. (2009). Another factor to consider is the economic consequence of using the current low albedo impervious surfaces in the urban en vironment. With the reliance on air co nditioning for space cooling, particularly in southern latitude s there is an associated increased energy cost. Along with the increased energy cost there is also an increased climatic cost In the United States a very large percentage of electrical ener gy is produced by burning fossil fuels. The increases in electrical energy usage required for air conditioning to counteract the increases in temperature in the urban environment results in increased production of carbon dioxide and other green house gass es In addition the expected lower future wind speeds (noted by Pryor 2009) will mean increased UHI temperature adding to the higher energy needs for air conditioning. Multiply this over the tho usands of urban environments and the generation of increased greenhouse gasses becomes appreciable. Researchers such as Coutts et al (2008) are actively trying to develop tools that will assist urban planners in reducing the impact of future urban dev elopment on the generation of a UHI The simple act of forethoug ht in the selection of building and paving materials when planning the urban environment of cities can have a large scale impact on the UHI energy usage, and greenhouse gas emissions 6.1 Perceived weaknesses of this study Having completed the investiga tions described in this dissertation, a number of weaknesses have become apparent, and are described below. PAGE 174 158 The rapid urban development in the Tampa study area, particularly in the suburban fringe near the rural sensor locations pose s a challenge to repli cate or continue further analysis in coming years. I n an attempt to ensure that the rural sensor locations can be maintained as a base line, a ny future study will require that rural sensor locations be remote enough such that development is unlikely to oc cur in the near term. The wind analysis of this study relied heavily on a single location for wind speed data. This limited the evaluation to a general study area analysis instead of a localized detailed analysis. In addition while wind data can be obtain ed at resolutions of less than 30 minutes the corresponding requirement to set a more frequent temperature sensor sampling rate would significantly decrease the survey period that could be stored in the device. F uture studies should attempt to deploy wind speed sensors along with greater sample storage capacity temperature sensors thus enabling a higher degree of precision in the definition of the UHI relationship. This will come at a much greater expense, but with the increased precision one may be able to assess the relationship of wind speeds and UHI at wind speeds less than 2.1 ms 1 Additionally wind speed sensors in direct proximity to temperature sensors will allow the use of multiple regression analysis (not possible with only one wind speed site) to define a more encompassing equation of the relationship between the percentage of the generation of a UHI. Lastly the need to down load the data from 100 sensors after 2048 samples were obtained caused logistical difficulties due to the 525 km 2 size of the study area. It would be desirabl e to deploy sensor arrays that can stay in the field unattended for extended PAGE 175 159 periods of time reporting data via telemetry as is being done with the Oklahoma mic ro net system (Bassara 2009). 6.2 Areas of potential future work It was mentioned in Section 2. 2 that a universally accepted defi nition for classifying urban and rural sites does not exist. Recent literature by Stewart and Oke (2009) mentioned a proposed methodology for researchers to classify rural and urban land types. According to Stewart and Ok e, they are working on validating the new method of classification. A validation of their findings will remove some of the uncertainty in land use classification in UHI research and hopefully provide a more concrete method for comparing the results of UHI studies. With the details of this work, which should be available in the peer reviewed literature in the near future, this may provide guidance for future UHI studies While this research can be seen as a significant advancement towards improving the dens ity of temperature m easurements in the study of UHI s in a sub tropical region, additional work is still required to investigate MUHIs. Additionally it would be desirable to include the measurement of wind speeds at the temperature measurement sites. The inc lusion of wind speed measurements will allow for better precision and a greater understanding of the role of wind speed i n moderating the UHI. Freitas et al. (2006 ) noted that the sea breeze interaction with the UHI is also dependent on the lateral extent of the urban environment. Increases in the lateral dimension of the urban environment can slow the transition of the sea breeze through the urban e nvironment. It is possible in the future that the expanding Tampa urban environment may delay the west coas t sea PAGE 176 160 breeze to the extent that it would allow the east coast sea breeze to migrate further across the state This would have the effect of shifting the sea breeze convergence zone (with associated thunderstorms) closer to the west coast. A shift in the co nvergence zone could modify the precipitation patterns near the west coast. It would be desirable to initiate a multi year study of the sea breeze interaction with the Tampa UHI and any long term movement in the sea breeze convergence zone. A related area of interest is the interaction of precipitation patterns with changes in the spatial extent of impervious surfaces. Utilization of advanced satellite imaging for the analysis of changes in the extent and density of impervious surfaces in conjunction with a spatially dense rain gauge system might enable the development of a model relating impervious surface and precipitation patterns. The study area of Florida also presents a myriad of research opportunities for the examination of the various interactions and interrelationships of anthropogenic modification s to the land surface Further research into these in teractions, particularly in sub tropical regions (noted by Roth 2007), is critical if we are to understand how human kind is affecting the local and gl obal climate. The cloud seeders of today are much like the rain makers of the past century assaulting the heavens for that precious drop of rain. Hopefully through continued research we are not destined to continually repeat the past inadvertent modifi cation of our weather and climate. Only time will tell. PAGE 177 161 7 .0 REFERENCES Akb ari, H. and S. Konopacki 2005 Calculating energy saving potentials of heat island reduction strategies Energy Policy 33 721 756 Aniello C., K. Morgan, A Busbey and L. N ewland 1995 Mapping micro urban heat i slands using LANDSAT TM and GIS Computers and Geoscience 73 (5) 965 969 Anselin, L., I. Syabri, and Y. Kho 2010 GeoDa: An Introduction to spatial d ata a nalysis h andbook of applied spatial a nalysis Springer Berlin Heidelberg 73 89 Baker R.D., B.H. Lynn, A Boone, W.K. Tao, and J. Simpson 2001 The influence of soil moisture, coastline c urvature, and land breeze circulations on sea breeze initiated p recipitation Journal of Hydrometeorology 2 193 211 Barnett A ., D.R. Hatton and D.W. Jones 1998 Recent changes in thermometer screen design and their impact World Meteorological Organization WMO/TD 871 Bassara, J.B., B.G. Illston, C.A. Fiebrich, R.A. McPherson, J.P. Bostic, P. Browder, D.B. Demko, C. Morgan, and K. Kesler. 2009. Eighth Symposium on the Urban En vironment AMS Conference. J1.1 Beyer H.L. 2006 Hawths analysis t ools for ArcGIS, version 3.26 [software]. http://www.spatialecology.com Bornstein R. 19 87 Mean diurnal circulation and thermodynamic evolution of urban boundary layers. In Modelling the Urban Boundary Layer. Bulletin American Meteorological Society, 53 93 Bornstein R., and Q. L. Lin 2000 Urban heat i slands a nd summertime convective t hunders torms in Atlanta: three case studies Atmospheric Environment 34 (3) 507 516 Bottyan Z. and J. Unger 2003 A multiple linear statistical model for estimating the mean maximum urban heat i sland Theoretical and Applied Climatology 75 233 243 Byer s H.R. an d H.R. Rodebush 1948 Causes of thunderstorms of the Florida p enisula Journalof Meterology 5 275 280 Cenedese A. and P. Mon ti 2003 Interaction between an inland urban heat island and a sea breeze flow: A laboratory s tudy Journal of Applie d Meteorology 42 1569 1583 PAGE 178 162 Changnon S.A. 1968 The La Porte weather anomaly fact or f iction Bulletin Ameri can Meteorology Societ y 49 4 11 Changnon S.A. 1992 Inadvertent weather modification in urban a reas: Lessons for global climate c hange Bulleti n of the American Meteorological Societ y 73(5) 619 627 Changnon S.A an d A.H. Huff 1971 METROMEX : an investigation of inadvertent weather m odification Bulleti n American Meteorological Societ y 52 ( 10 ), 958 967 Changnon, S.A., F.A. Huff, P.T. Schicked anz, and J. L.Vogel 1977 METROMEX, Volume 1 & 2: Weather anomalies and i mpacts National Technical Information Service Bulletin 62, 264 Chen Y., M. DU and R Dong 2008 Correlation between urban heat island e ffect and the thermal inertia using Aster d ata in Beijind, China The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B7, 1723 1728 Clar k C. S 1980 The Rain Maker University of Nebraska Press Collins, J.M., C. Paxton, A. Williams and D. Noah 2009 Southwest Florida warm season tornado development. National Weather Associations Electronic Journal of Operational Meteorology EJ12 Comarazamy D.E., J.E. Gonzales, J. Luvall, D. Rickman, and A. Picon 2006 A validation study of the u rban heat island in the tropical coastal c ity of San Juan, Puerto Rico Sixth Symposium on the Urban Environment 7 Cotton W.R. and R.A. Pielke Sr. 2007 Human Impacts on Weather and Climate Cambridge University Press Coutts A.M., J. Beringer, and N.J. Tapper 2008 Investigating the climatic impact of urban planning strategies through the use of regional modeling; a case study for Melbourne Australia International Journal of Climatology 28 1943 1957 Dallas Semiconductor http://www.maxim ic.com/quick_view2.cfm/qv_pk/3246 last access 6/3/2010 Dan S, H. An, B. Dan, H. Xu, L. Yang and G. Chen 2009 An analysis of urban heat island effects in Chongqing based on AVHRR and DEM Resou rces and Environment in the Yangtze Basin 18, 680 685 Dixon P. G., an d T. L. Mote 2003 Patterns and c a uses of Atlanta's urban h eat island initiated p recipitation Journal of Applied Meteorology 42 (9) 1273 1284 PAGE 179 163 Emmanuel R. H. Roselend and E. Johans son 2007 Urban shading a design option for the tropics? A study in Colombo, Sri Lanka International Journal of Climatology 27 1995 2004 European Space Agency 2008 Observing the Earth www.esa.int/esa Last acces s 08/12/2009 Escourrou G. 1991 Le Climat et la Ville Nathan University Editions, Paris Fassig, O.L. 1907 The climate and w eather of Baltimore Maryland Weather Service 276 283 Freitas E.D., C.M. Rozoff, W.R. Cotton, and P.L. Silva Dias 2006 Interact ions of the urban heat island and sea breeze circulations during winter over the metropolitan a rea of San Paulo Brazil Boundary Layer Meteorology 122, 43 65 Gedzelman, S.D. S. Austin, R. Cermak, N. Stefano, S. Partridge, S. Quesenberry, and D. A. Robins on 2003 Mesoscale aspects of the urban heat i sland around New York City Theoretical and Applied Climatology 75 29 42 Grimmond, C.S.B 2006 Progress in measuring and observing the urban atmosphere Theoretical and Applied Climatology 84 3 22 Grimmon d C.S.B., and T. R. Oke 1999 Aerodynamic properties of urban areas derived, from analysis of surface f orm Journal of Applied Meteorology 38(9) 1262 1292 Hartman C.E. and L.W. Oring 2006 An inexpensive method for remotely monitoring nest activity Jou rnal of Field Ornithology 77(4) 418 424 H au L. Z. G. Ma and W. D. Guo 2008 The impact of urbanization on air temperature across China Theoretical and Applied Climatology 93 179 194 He J.F., J.Y. Liu, D. F. Zhuang, W. Zhang, and M. L. Liu 2007 Assess ing the effect of land use land cover change on the change of urban heat island intensity Theoretical and Applied. Climatology 90, 217 226 Hedquist B.C., and A. J. Brazel 2006 Urban, residential, and rural climate comparisons from mobile transects and fixed s tations: Phoenix, Arizona. Journal of the Arizona Nevada Academy of Science 38, 77 87 Homer C., C. Huang, L. Yang, B. Wylie, and M. Coa n 2004 Development of the 2001 national l and cover d atabase for the United States Photogrammetric Engineering and Remote Sensing 70, 829 840 PAGE 180 164 Hubbart J., T. Link, C. Campbell, and D. Cobos 2005 Evaluation of a low cost t em perature measurement system for environmental a pplica tions Hydrological Processes 19 1517 1523 Iino A. and A. Hoyano 1996 Development of a method to predict the heat island potential using remote s ensing and GIS data Energy and Building 23 199 205 Jensen J. R. 2000 Remote Sensing of the Environme nt Prentice Hall publishers Johnson A. N., B.R. Boer, W .W. Woessner, J.A. Stanford, G. C. Poole, S.A. Thomas, and S. J. ODaniel 2005 Evaluation of inexpensive small diameter temperature logger for documenting ground water river i nteractions Ground Water Monitoring and Remediation 25 68 74 Kalnay, E and M. Cai 2003 Impact of urbanization and land use change on c limate Nature 423(29) 528 531 Li J.J, X.R. Wang, X.J. Wang, W. Ma and H. Zhang 2009 Remote sensing evaluation of urban heat island and its spatial pattern of the Shanghai metropolitan area, China Ecological Complexity 6, 413 420 Ligda, M., and S. Bigler 1956 Use of radar in severe s torm detection: Hydrology and c limatology Final Report Texas A&M University Lo C.P., D.A. Quattrochi, and J .C. Luvall 1997 Application of high resolution t hermal infrared remote sensing and GIS to assess the urban heat e ffect International Journal of Remote Sensing 18 (2), 287 304 Lo C.P., an d D. A. Quattrochi 2003 Land use and land cover change, urban heat island phenomenon, and health i mplications: A remote sensing a pproach Photogrammetric Engineering & Remote Sensing 69(9), 1053 1063. Loose T. and R.D. Bornstein 1977 Observations of mesoscale effects on frontal movement through an urban a rea Mo nthly Weather Review 105 (5), 563 571 Lu D ., and Q. Weng 2006 Use of impervious surface in urban land use classification Remote Sensing of Environment 102 146 160 Manley G. 1958 On the frequency of snowfall in metropolitan England Quarterly Journa l of the Royal Meteorological Society 84(359), 70 72 Marshall C.H., R.A. Pielke, L.T. Steyaert, and D. A. Willard. 2004 The Impact of anthropogenic land cover change on the Florida peninsula sea breezes and warm season sensible w eather Monthly Weather R eview 132 (1) 28 52 McClave J.T ., and T. Sincich 2006 Statistics (10 th ed.) Pearson Prentice Hall PAGE 181 165 Mihalakakou G., M. Santamouris, N. Papnikolaou, C. Cartalis and A. Tsangra ssoulis 2001 Simulation of the urban heat island phenomenon in Mediterranean c limates. Pure and Applied Geophysics 161 429 451 Mogil H.M. and K.L. Seaman 2008 Floridas climate and w eather Weatherwise 61 ( 6 ), 14 19 Morris C.J.G. and I. S immons 2002 Quantification of influences of wind and cloud on the nocturnal urban heat island of a l arge c ity Journal of Applied Meteorology 40 169 182 National Climatic Data Center 1998 Climatic Wind Data for the United States National Weather Service 2008 Does the urban heat island affect rainfall variability across the Phoenix, AZ m etropolitan a r ea during the monsoon s eason National Weather Service Publication 13 June 2008 National Weather Service 2010 www.nws.noaa.gov/glossary/ Last access 3/21/2010 Oke T.R. 1973 City size an d the urban heat island Atmospheric Environment, 7, p769 779 Oke T.R. 1982 The Energetic basis of the urban heat i sland Quarterly Journal of the Royal Meteorological Society 108 1 24 Oke, T.R. 1987 Boundary Layer Climates Cambridge University Press Otte T.L., L. Avrahan, D. Sylvain and J. C hing 2004 Implementation of an urban canopy parameterization in a mesoscale m eteorol ogical m odel American Meteorological Society 43 1649 1665 Pielke, R.A. Sr., R.L. Walko, L.T. Steyaert, P.L. Vidale, G.E. Liston, W.A. Lyons, and T. N. Chase, 1999: The influence of anthropogenic landscape changes on weather in south Florida. Monthly Weather Review 127 1663 1673 Peterson T.C. 2003 Assessment of urban versus rural in situ surface temperatures in the contigu ous United States: No difference f ound Journal of Climate 16 (18), 2941 2959 Pielke R.A. Sr. 1984 Mesoscale Meteorological Modeling International Geophysical Series 78 661 Price J.C. 1979 Assessment of the urban heat island effect through the use of satellite d ata Monthly Weather Review 107,1554 1557 PAGE 182 166 Pryor, S.C., R.J. Barthelmie, D.T. Young, E. S. Takle, R.W. Arritt, D. Flory, W. J. Gutowski, A. Nunes, and J. Roads 2009 Wind speed trends over the contiguous USA Journal of Geophysical Research 114 Roberts S.M., T.R. Oke, C.S.B. Grimmond, and J.A. Voogt 2006 Comparison of four methods to estimate urban heat s torage Journal of Applied Meteorology and Climatology 45 1766 1781 Roth M. 2007 Review of urban climate research in (sub)tropical region s International Journal of Climatology 27 1859 1873 Roth M., T.R. Oke, and W.J. Emery 1989 Satellite derived urban heat islands from three coastal cities and the utilization of such data in urban c limatology International Journal of Remote Sensing 1 0, 1699 1720 Rosenzweig C., W.D. Solecki, L. Parshall, B. Lynn, J. Cox, R. Goldberg, S. Hodges, S. Gaffin, R.B. Slosberg, P. Savio, F. Dunstan, and M. Watson 2 009 Mitigating New York Citys heat i sland Bulletin American Meteorological Society 1297 1312 Santamouris m, S.E., N. Papanikolaou I Livada I. Koronakis, C Georgakis A. Argiriou, and D.N. Assimakopoulo s 1999 On the Impact of urban c limate on the energy consumption of b uildings Solar Energy 70(3), 201 216 Saucier, W. J. 1949 Texas west gul f c yclones Monthly Weather Review 77, 219 231 State of Florida, Department of Transportation 1999 Florida land use, cover and form classification system. [Handbook ] Department of Transportation surveying and mapping geographic mapping section Stathop oulou M., and C. Cartalis 2007 Daytime urban heat islands from Landsat ETM+ and Corine land cover data: An application to major cities in Greece Solar Energy 81 358 368 Stewart I.D. 2007 Landscape representation and the urban rural dichotomy in e mper ica l urban heat island l iterature, 1950 2006 Acta Climatologica Et Chorologica 40( 41 ) 111 121 Stewart I.D ., and T. Oke 2009 Newly developed thermal climate zones for d efining and measuring urban heat island magnitude in the canopy l ayer Timothy R. Oke Symposium 11 15 January, 2009 Stowers, D.M. and N.D. Tabb 1987 An investigation of the variances from the traditional summer precipitation in the west central Florida region Florida Scientist 50 177 183 PAGE 183 167 Titus, J.G. 1992. The costs of climate c han ge to the United States Originally published in: Majumdar, S.K., L.S.Kalkstein, B. Yarnal, E.W. Miller, and L.M. Rosenfeld (eds). Global Climate Change: Implications, Challenges, and Mitigation Measure s Unger, J., Z. Smeghy, . Gulys, Z. Bottyn and L Mucsi 2001 Land use and meteorological aspects of the urban heat island Applied Meteorol o gy 8 189 194 United Nations, Department of Economic and Social Affairs, Population Division 2006 World Urbanization Prospects: The 2005 Revision Working Paper No ESA/P/WP/200 U.S. Census Bureau 2000 Tiger line and block group data files USGS 2007 Gulf of Mexico Integrated Service Data Information Management System (MIMS) http://gulfsci.usgs.gov/index.html Wang Y. and F. Hu 2006 Variations of the urban heat island in summer of the recent 10 year over Beijing and its environment effect Diqui Wull Xuebao 49, 61 68 Weng Q., H. Liu, and D. Lu 2007 Assessing the effects of land use and land cover patterns o n thermal conditions using landscape metrics in city of Indianapolis, United States Urban Ecosystem 10 203 219 Woollum C. A. an d N.L. Canfield 1968 Washington metropolitan area p recipi tation and temperature patterns t echnical Memorandum WBTM ER 28, U .S. Department of Commerce ESSA Xian G. a nd M. Crane 2005 Assessment of urban growth in the Tampa Bay watershed using remote sensing d ata Remote Sensing of the Environment 97 203 215 Xian G. and M. Crane 2006 An analysis of urban thermal characte ristics and associated land c over in Tampa B ay and Las Vegas using Landsat satellite d ata Remote Sensing of the Environment 104 146 156 Xian G., M. Crane and J. Su 2007 An a nalysis of urban d evelopment and its e nvironmental impact on the Tampa Bay w at ershed Journal of Environmental Management 85 965 976 Yi Sun C., A.J. Brazel, T.L. Chow, B. C. Hedquist and L. Prashad 2009 Desert heat island study in winter by mobile transect and remote sensing techniques. Theoretical and Applied Climatology 98 323 335 Yow D. M. and G. J. Carbone 2006 The urban heat island and local t emperatur e v ariations in Orlando, Florida Southeastern Geographer 46(2) 297 321 PAGE 184 168 Yuan F. an d M. Bauer 2007 Comparison of impervious surface area and normalized difference vegetation index as indicators of surface urban h eat island effects in Landsat i magery Remote Sensing of the Environment 106 375 38 Zhou L., R. E. Dickinson, Y. Tian, J. Fang, Q. Li, R. K. Kaufmann, C. J. Tucker, and R. B. Myneni 2004 Evidence for a significant urbanization effect on climate in China Papers of the National Academy of Sciences 101 (26) 9540 9544 PAGE 185 169 APEENDIX A THERMOCHRON DG1921H SPECIFICATIONS PAGE 186 170 DS1921H, DS1921Z High Resolution Thermochron i Button Range H: +15C to +46C; Z: 5C to +26C Description The DS1921H/Z Thermochron i Buttons are rugged, self sufficient systems t hat measure temperature and record the result in a protected memory chapter. The recording is done at a user defined rate, both as a direct storage of temperature values as well as in the form of a histogram. Up to 2048 temperature values taken at equidist ant intervals ranging from 1 to 255 minutes can be stored. The histogram provides 64 data bins with a resolution of 0.5C. If the temperature leaves a user programmable range, the DS1921H/Z will also record when this happened, for how long the temperature stayed outside the permitted range, and if the temperature was too high or too low. Additional 512 bytes of read/write NV memory allow storing information pertaining to the object to which the DS1921H/Z is associated. Data is transferred serially via the 1 Wire protocol, which requires only a single data lead and a ground return.Every DS1921H/Z is factory lasered with a guaranteed unique electrically readable 64 bit registration number that allows for absolute traceability. The durable stainless steel packa ge is highly resistant to environmental hazards such as dirt, moisture, and shock. Accessories permit the DS1921H/Z to be mounted on almost any object, including containers, pallets, and bags. Key Features Digital thermometer measures temperature 1/8C increments with 1C accuracy Built in real time clock (RTC) and timer has accuracy of 2 minutes per month from 0to 45C Automatically wakes up and measures temperature at user programmable i ntervals from 1 to 255 minutes Logs up to 2048 consecutive temperature measurements in protected nonvolatile (NV) random access memory Records a long term temperature histogram with 1/2C resolution Programmable temperature high and tempe rature low alarm trip points Records up to 24 time stamps and durations when temperature leaves the range specified by the trip points 512 bytes of general purpose read/write NV random access memory Communicates to host with a single digi tal signal at 15.4kbits or 125kbits per second using 1 Wire protocol Fixed range: H: +15C to +46C; Z: 5C to +26C PAGE 187 171 APPENDIX B DELTA TEMPERATURE 0 F VALUES AT SENSOR LOCATIONS PAGE 188 172 APPEN DIX B continued Sensor # Time 00:00 00:30 01:00 01:3 0 02:00 02:30 03:00 03:3 0 1 .00 .00 .00 .00 .00 .00 .00 .00 2 .00 .00 .00 .00 .00 .00 .00 .00 3 1.48 1.62 1.73 1.81 1.85 1.98 2.06 2.10 4 1.10 1.13 1.28 1.28 1.16 1.10 1.11 1.06 6 .02 .03 .08 .11 .18 .13 .21 .16 7 1.38 1.39 1.32 1.28 1.14 1.14 1.03 .94 8 1.71 1.77 1.77 1.79 1.63 1.70 1.63 1.59 9 .68 .69 .61 .55 .40 .35 .19 .17 10 .16 .09 .24 .25 .38 .49 .56 .63 11 1.69 1.61 1.55 1.47 1.44 1.32 1.13 1.07 12 1.62 1.57 1.59 1.55 1.68 1.66 1.69 1.68 13 1.14 1.17 1.18 1.16 1.26 1.30 1.27 1.33 14 .4 4 .49 .49 .41 .56 .54 .56 .56 15 1.05 .93 .81 .73 .48 .44 .22 .11 16 .91 1.04 1.02 1.03 .94 .79 .71 .65 17 .50 .73 .76 .95 .81 .88 .83 .83 18 .28 .36 .40 .43 .32 .29 .31 .33 19 .31 .40 .44 .49 .37 .40 .35 .34 20 .25 .32 .34 .35 .20 .17 .14 .13 21 1.16 1.27 1.37 1.42 1.34 1.37 1.39 1.35 22 1.24 1.39 1.47 1.58 1.54 1.58 1.50 1.54 23 .81 .84 .84 .85 .72 .71 .65 .66 25 2.70 2. 85 2.80 2.88 2.73 2.74 2.69 2.65 26 1.61 1.72 1.87 1.95 1.85 1.88 1.88 1.89 27 1.23 1.31 1.31 1.34 1.19 1.18 1.12 1.14 28 .52 .61 .59 .68 .5 5 .49 .40 .46 29 1.17 1.29 1.23 1.23 1.19 1.13 1.02 .95 30 2.11 2.12 2.14 2.18 2.04 1.92 1.86 1.74 31 .95 1.03 1.04 1.11 1.02 .99 .83 .86 32 1.36 1.44 1.42 1.43 1.60 1.40 1.43 1.30 33 .68 .69 .61 .55 .40 .35 .19 .17 34 1.19 1.31 1.35 1.30 1.22 1.22 1.05 1.14 35 1.11 1.15 1 .20 1.24 1.13 1.14 .99 .97 36 1.04 1.26 1.20 1.26 1.25 1.20 1.05 .99 38 1.50 1.63 1.70 1.82 1.71 1.69 1.57 1.49 39 2.17 2.29 2.29 2.28 2.19 2. 15 2.09 2.06 40 2.10 2.31 2.37 2.34 2.28 2.29 2.31 2.30 PAGE 189 173 APPENDIX B continued Sensor # Time 00:00 00:30 01:00 01:3 0 02:00 02:30 03:00 03:3 0 41 1.30 1.38 1.45 1.30 1.29 1.3 7 1.32 1.35 42 1.60 1.79 1.89 1.84 1.86 1.95 1.97 1.94 43 2.84 3.04 3.21 3.22 3.21 3.18 3.18 3.06 44 1.83 1.99 2.09 2.16 2.22 2.25 2.26 2.25 45 3.28 3.40 3.49 3.49 3.46 3.42 3.46 3.42 46 2.85 2.93 3.11 3.09 3.06 3.07 3.05 3.01 47 2.68 2.90 3.11 3.30 3.26 3.29 3.27 3.37 49 2.29 2.44 2.62 2 .70 2.66 2.68 2.75 2.85 50 1.41 1.58 1.56 1.50 1.37 1.32 1.34 1.43 51 2.71 2.76 2.82 2.75 2.62 2.57 2.57 2.58 52 1.99 2.06 2.15 2.12 2.00 1.90 1. 86 1.87 53 2.01 2.08 2.16 2.08 1.99 1.95 1.84 1.88 54 1.70 1.82 1.94 1.87 1.77 1.68 1.68 1.72 55 3.78 3.88 3.96 3.94 3.87 3.81 3.88 3.88 56 1.40 1.53 1.60 1.57 1.49 1.51 1.60 1.66 57 2.28 2.27 2.33 2.30 2.14 2.13 2.09 2.17 58 1.03 1.11 1.20 1.18 1.05 1.01 1.01 1.03 59 1.65 1.77 1.96 1.86 1.79 1.80 1.83 1.82 60 2.11 2.15 2.25 2.24 2.19 2.24 2.18 2.19 61 2.06 2.18 2.34 2.39 2.33 2.27 2.27 2.32 62 1.63 1.72 1.80 1.84 1.78 1.78 1.73 1 .77 63 1.35 1.47 1.50 1.56 1.51 1.47 1.50 1.57 64 2.88 3.04 3.12 3.11 3.14 3.06 3.17 3.16 65 2.38 2.46 2.36 2.30 2.20 2.12 2.11 2.07 66 2.25 2.31 2.24 2.33 2.37 2.39 2.41 2.45 67 1.58 1.66 1.56 1.56 1.51 1.49 1.49 1.53 68 2.93 2.76 2.65 2.67 2.54 2.44 2.35 2.31 69 2.33 2.30 2.22 2.26 2.26 2.18 2.07 2.05 70 1.98 2.01 1.90 1.91 1.74 1.72 1.61 1.62 71 3.96 3.85 3.69 3.73 3.46 3.44 3.25 3.19 73 1.67 1.75 1.66 1.78 1.78 1.71 1.80 1.90 74 2.04 2.10 2.13 2.18 2.09 2.04 2.04 2.07 75 2.57 2.59 2.57 2.47 2.38 2.25 2.18 2.18 76 .42 .42 .31 .42 .32 .30 .28 .23 77 1.27 1.30 1.37 1.29 1.19 1.13 1.07 1.14 78 2.70 2.77 2.80 2.79 2.68 2.62 2.57 2.55 PAGE 190 174 APPENDIX B continued Sensor # Time 00:00 00:30 01:00 01:3 0 02:00 02:30 03:00 03:3 0 80 .35 .50 .50 .60 .53 .52 .52 .67 81 .17 .19 .23 .21 .16 .13 .06 .16 82 2.05 2.13 2.11 2.12 2.05 1.97 1.98 1.95 83 .79 .82 .89 .91 .74 .71 .67 .71 84 1.53 1.53 1.52 1.56 1.41 1.29 1.20 1.19 85 1.03 1.14 1.17 1.21 1.13 1.07 1.00 .97 86 1.31 1.49 1.60 1.54 1.50 1.45 1.49 1.50 87 1.33 1.49 1.63 1.64 1.50 1.45 1.34 1.38 88 .53 .56 .61 .58 .44 .38 .27 .25 89 1.21 1.39 1.47 1.50 1.38 1.34 1.33 1.30 90 1.44 1.58 1.67 1.59 1.51 1.52 1.50 1.42 91 .44 .51 .56 .66 .60 .58 .51 .52 92 .85 .84 .77 .88 .74 .74 .69 .70 93 1.24 1.47 1.50 1.55 1.53 1.51 1.53 1.55 94 1.35 1.40 1.33 1.28 1.27 1.27 1.23 1.26 95 2.22 2.15 2.04 2.01 1.85 1.76 1.62 1.61 96 1.83 1.74 1.61 1.62 1.52 1.41 1.34 1.27 97 2.31 2.37 2.36 2.39 2.26 2.13 2.10 2.07 98 .00 .00 .00 .00 .00 .00 .00 .00 99 .00 .00 .00 .00 .00 .00 .00 .00 100 .00 .00 .00 .00 .00 .00 .00 .00 PAGE 191 175 APPENDIX B continued Sensor # Time 04:00 04:30 05:00 05:30 06:00 06:30 07:00 07:3 0 1 .00 .00 .00 .00 .00 .00 .00 .00 2 .00 .00 .00 .00 .00 .00 .00 .00 3 2.14 2.24 2.28 2.29 2.32 2.28 1.64 .35 4 1.07 1.08 1.10 .98 .96 .92 .38 1.63 6 .11 .03 .04 .02 .06 .02 .62 2.89 7 .87 .92 .91 .79 .68 .65 .40 3.15 8 1.59 1.65 1.63 1.56 1.48 1.51 2.33 4.12 9 .04 .00 .11 .15 .25 .31 .63 2.31 10 .65 .68 .81 .89 .94 .98 1.08 .74 11 1.05 1.03 .95 .87 .82 .75 .35 1.39 12 1.65 1.60 1.59 1.54 1.58 1.62 1.61 1.08 13 1.22 1.24 1.23 1.17 1.20 1.12 1.25 2.82 14 .55 .46 .52 .52 .53 .44 .40 3.19 15 .02 .01 .12 .23 .35 .49 .51 1.53 16 .65 .69 .63 .57 .52 .43 .15 1.17 17 91 .90 .93 .86 .87 .85 .72 1.88 18 .31 .35 .40 .39 .38 .38 .21 1.28 19 .40 .43 .45 .48 .45 .47 .44 2.13 20 .24 .24 .28 .2 5 .19 .16 .19 1.53 21 1.40 1.43 1.41 1.32 1.26 1.21 .64 1.27 22 1.56 1.62 1.61 1.52 1.50 1.52 1.41 .80 23 .66 .65 .58 .52 .53 .44 .33 .84 25 2.68 2.69 2.58 2.53 2.44 2.27 2.01 .60 26 1.99 2.04 1.96 1.97 1.87 1.79 1.28 .55 27 1.23 1.25 1.25 1.20 1.15 1.11 .82 .64 28 .53 .52 .58 .47 .46 .48 .09 1.36 29 1.00 1.15 1.18 1.24 1.22 1.32 1.35 .22 30 1.79 1.78 1.71 1.67 1.59 1.52 1.21 .70 31 .87 .88 .86 .76 74 .76 1.08 2.40 32 1.30 1.22 1.30 1.27 1.22 1.29 1.21 .02 33 .04 .00 .11 .15 .25 .31 .63 2.31 34 1.16 1.09 1.03 1.04 .95 .85 .73 .6 3 35 .99 .97 .86 .81 .77 .71 .68 2.33 36 1.00 1.07 .98 .99 1.00 .97 .92 .48 38 1.50 1.51 1.44 1.36 1.33 1.35 1.06 .69 39 2.05 2.06 1.88 1.87 1.79 1.75 1.54 .06 PAGE 192 176 APPENDIX B continued Sensor # Time 04:00 04:30 05:00 05:30 06:00 06:30 07:00 07:3 0 40 2.28 2.33 2.22 2.18 2.17 2.17 2.05 .72 41 1.42 1.43 1. 39 1.23 1.19 1.21 1.51 2.15 42 2.02 2.01 1.95 1.82 1.80 1.71 1.43 .30 43 3.12 3.11 3.10 3.11 3.15 3.26 3.54 3.16 44 2.27 2.31 2.28 2.25 2.24 2.2 3 2.14 .54 45 3.40 3.39 3.35 3.31 3.24 3.25 3.07 1.13 46 3.01 3.06 2.94 2.95 2.88 2.84 2.35 .27 47 3.37 3.41 3.41 3.30 3.37 3.41 3.21 1.44 49 2.86 2.94 2.89 2.90 2.83 2.75 2.45 .43 50 1.53 1.53 1.44 1.40 1.38 1.26 1.04 2.02 51 2.68 2.68 2.67 2.58 2.48 2.39 2.05 .56 52 1.87 1.92 1.81 1 .75 1.65 1.53 1.41 3.03 53 1.92 1.98 1.88 1.79 1.74 1.63 1.58 3.74 54 1.77 1.81 1.69 1.72 1.63 1.54 1.41 .10 55 3.83 3.79 3.79 3.78 3.76 3.71 3. 68 2.51 56 1.68 1.65 1.68 1.70 1.62 1.51 1.61 .91 57 2.07 2.10 2.06 2.05 1.94 1.77 1.54 .33 58 1.01 1.09 1.06 1.09 1.01 .86 .53 .77 59 1.88 1.95 1.96 1.97 1.99 1.96 1.87 2.57 60 2.28 2.32 2.29 2.28 2.22 2.23 2.48 3.21 61 2.40 2.46 2.45 2.42 2.35 2.24 1.98 .95 62 1.86 1.90 1.84 1.84 1.91 1.71 1.55 .49 63 1.72 1.78 1.62 1.56 1.59 1.32 1.00 .56 64 3.28 3.31 3.33 3.33 3.24 3.15 3.03 1.86 65 2.08 2.10 2.00 1.93 1.86 1.72 1.71 1 .62 66 2.48 2.42 2.46 2.46 2.40 2.35 2.20 2.39 67 1.67 1.60 1.56 1.49 1.31 1.18 .92 1.08 68 2.23 2.21 2.15 2.09 1.89 1.69 1.48 .09 69 2.10 2.09 2.06 2.03 1.81 1.76 2.11 3.01 70 1.74 1.80 1.62 1.59 1.50 1.45 1.63 5.12 71 3.13 3.16 3.01 2.92 2.81 2.70 2.24 .03 73 1.99 2.02 2.09 2.02 2.10 2.00 1.89 .82 74 2.13 2.25 2.20 2.18 2.19 2.21 2.02 3.29 75 2.18 2.22 2.11 1.97 1.91 1.81 1.96 2.77 76 .25 .30 .22 .18 .12 .05 .16 1.47 77 1.18 1.25 1.13 1.07 .99 .92 .59 .89 78 2.56 2.59 2.58 2.58 2.62 2.75 2.57 1.76 PAGE 193 177 APPENDIX B continued Sensor # Time 04:00 04:30 05:00 05:30 06:00 06:30 07:00 07:3 0 80 .74 .84 .73 .70 .64 .57 .08 1.01 81 .18 .24 .16 .11 .06 .03 .52 2.33 82 1.96 1.93 1.90 1.87 1.89 1.88 1.76 3.26 81 .18 .24 .16 .11 .06 .03 .52 2.33 82 1.96 1.93 1.90 1.87 1.89 1.88 1.76 3.26 83 .77 .89 .82 .83 .88 .95 1.50 2.48 84 1.18 1.21 1.15 1.08 1.04 .98 .8 4 1.48 85 .96 1.04 1.03 .96 .89 .88 .58 2.14 86 1.59 1.60 1.70 1.55 1.48 1.38 1.08 4.11 87 1.39 1.48 1.51 1.44 1.48 1.40 1.71 .19 88 .25 .34 .31 .20 .28 .32 .30 .84 89 1.37 1.43 1.50 1.51 1.64 1.78 1.64 .82 90 1.47 1.57 1.55 1.51 1.63 1.75 1.56 1.77 91 .52 .57 .55 .47 .42 .39 .17 .66 92 .75 .78 .72 .69 .69 .82 .70 .91 93 1.67 1.75 1.64 1.71 1.74 1.76 1.33 .60 94 1.29 1.29 1.21 1.24 1.19 1.13 .79 74 95 1.56 1.55 1.47 1.42 1.33 1.27 .81 1.07 96 1.31 1.31 1.16 1.07 1.00 .89 .59 .14 97 2.12 2.14 2.01 2.02 2.00 1.93 1.71 2.45 98 .00 .00 .00 .00 .00 .00 .00 .00 99 .00 .00 .00 .00 .00 .00 .00 .00 100 .00 .00 .00 .00 .00 .00 .00 .00 PAGE 194 178 APPENDIX B continued Sensor # Time 08:00 08:3 0 09:00 09:30 10:00 10:30 11:00 11:3 0 1 .00 .00 .00 .00 .00 .00 .00 .00 2 .00 .00 .00 .00 .00 .00 .00 .00 3 1.57 3.06 3.55 2.74 2.67 4.3 9 5.82 1.74 4 2.82 3.96 4.09 3.33 2.28 4.87 5.03 4.73 6 2.94 4.75 6.23 6.67 7.73 6.72 5.77 5.28 7 .12 .99 1.67 1.91 2.07 5.84 .58 1.68 8 5.80 7.27 8.59 9.22 9.93 8.67 7.47 7.68 9 2.92 1.04 2.20 3.85 4.33 4.66 4.09 3.48 10 3.31 5.19 6.81 8.47 8.74 8.70 6.38 5.90 11 .64 3.87 4.44 1 .80 .65 1.45 .07 .34 12 5.38 8.27 9.75 13.26 13.42 12.47 9.27 9.63 13 3.22 3.25 .22 6.58 8.34 7.15 5.39 4.68 14 6.21 7.81 8.91 10.37 11.92 10.63 9.21 8.39 15 1.76 1.51 1.52 3.68 4.15 4.93 3.55 2.85 16 1.72 2.32 2.29 .92 .05 .98 .46 .10 17 3.93 4.56 5.81 7.12 8.19 8.44 4.71 1.69 18 2.06 1 .95 1.28 3.82 5.06 5.73 3.76 2.85 19 5.14 5.70 7.18 7.97 9.06 8.51 7.29 5.40 20 2.50 2.06 1.69 7.47 11.32 11.02 9.96 8.53 21 1.90 3.08 1.10 .64 .37 .54 2.02 2.55 22 .34 .62 .75 .02 .44 .27 .65 1.86 23 1.46 3.64 5.10 7.33 8.08 9.04 6.63 5.54 25 6.99 9.25 10.74 11.82 14.08 13.92 12.04 9.42 26 4.11 2.32 1.15 1.44 1.93 2.41 3.30 4.02 27 1.39 1.84 2.67 7.48 7.66 7.85 6.54 5.19 28 2.07 .48 4.67 6.84 7.92 6.83 5.73 5.05 29 .41 4.92 7.38 9.55 10.35 10.24 9.43 7.93 30 1.97 2.66 2.12 3.87 4.54 5.86 4.36 3.84 31 3.92 4.75 5.23 6.45 8.19 7.07 6.09 5.19 32 1.76 1.19 .83 1.76 2.51 2.14 .60 1.68 33 2.92 1.04 2.20 3.85 4.33 4.66 4.09 3.48 34 .41 2.15 2.36 3.66 3.85 4.35 3.21 2.85 35 4.24 4.63 4.54 5.13 6.32 5.78 3.87 4.11 36 1.19 .12 .26 1.31 1.99 2.27 2.18 .47 38 2.01 3.21 3.54 2.78 .09 .16 2.47 3.08 39 .82 1.59 1.82 4.90 5.98 6.26 5.43 3.66 PAGE 195 179 APPENDIX B continued Sensor # Time 08 :00 08:30 09:00 09:30 10:00 10:30 11:00 11:3 0 40 .09 .70 1.01 .46 .30 .81 .12 .97 41 2.98 2.15 2.27 3.76 4.52 4.38 3.10 1.69 42 1.43 2.56 2.09 .57 1.06 1.55 .45 .49 43 3.34 2.55 2.47 3.35 3.44 3.92 3.01 1.06 44 1.64 .60 .57 1.34 2.51 2.52 1.44 .30 45 3.73 2.94 3.11 3.90 5.00 5.31 4.11 3.00 46 1.07 2.42 2.57 2.03 1.38 .72 1.44 .97 47 .39 .50 .92 .08 2.22 2.05 1.20 .40 49 1.02 2.50 2.64 1.99 2.37 1.51 1.02 2.13 50 3.01 3.08 2.5 2 3.97 4.89 5.20 3.26 2.71 51 .12 .33 .91 1.85 2.07 1.94 .22 .54 52 3.70 3.09 2.90 3.71 3.60 4.05 3.24 2.21 53 5.30 6.05 6.51 6.91 7.73 8.54 6.70 5.38 54 3.11 4.50 5.32 6.48 6.84 7.40 5.89 5.70 55 5.63 6.18 6.41 7.32 8.14 7.92 6.17 5.84 56 .55 .78 8.85 11.27 16.19 15.46 13.09 11.47 57 .18 3.77 5.10 5.61 7.11 7.02 5.67 5.33 58 1.35 2.51 2.96 2.25 .74 3.10 2.99 2.81 59 3.62 3.18 2.93 3.64 2.88 3.63 2.33 1.63 60 3.23 .09 .07 2.57 3.42 3.80 3.04 2.10 61 .58 2.02 2.41 2.01 1.96 1.46 2.02 1.56 62 3.29 3.86 3.78 4.20 5.33 5.48 4.07 3.55 63 1.59 2.22 2.23 1.65 .96 .28 1.35 1.88 64 3.45 3.49 3.64 1.02 2.27 1.11 .40 .62 65 4.08 3.78 4.48 4.94 5.60 5.54 4.83 3.61 66 2.42 1.35 .97 1.57 1.52 1.62 .27 .11 67 1.16 2.13 2.46 1.91 1.66 1.22 1.63 2.25 68 .39 .75 .47 .86 1.54 2.59 2.54 1.71 69 3.75 3.15 3.42 5.02 5.06 5.19 4.73 3.04 70 9.20 11.64 12.50 14.96 15.59 15 .10 13.47 10.75 71 1.26 1.86 2.65 2.12 1.64 1.10 1.48 2.24 73 .55 1.68 2.00 3.32 4.57 3.90 3.12 1.94 74 5.34 6.16 6.41 4.37 2.88 1.87 .89 .17 75 3.33 3.17 3.06 3.69 4.99 5.00 4.56 2.69 76 .89 1.20 2.09 3.46 2.86 3.44 2.30 2.96 77 1.72 2.46 2.49 .07 2.14 2.93 2.40 2.39 78 .25 .32 .69 .36 2.20 3.50 3.13 2.23 PAGE 196 180 APPENDIX B continued Sensor # Time 08:00 08:30 09:00 09:30 10:00 10:30 11:00 11:3 0 80 2.70 3.36 2.33 1.88 1.28 .53 1.45 1.70 81 3.36 4.31 4.43 3.53 3.31 2.83 3.45 3.87 82 3.99 4.11 4.52 5.44 5.37 5.65 3.96 2.26 83 2.81 2.35 2.35 3.16 3.28 2.82 .84 .25 84 3.78 3.48 3.48 5.22 5.78 5.38 4.50 4.03 85 2.43 .41 2.02 .21 .27 2.03 .63 .73 86 5.29 5.51 2.22 2.04 5.82 2.01 .34 1.05 87 .18 1.07 .61 5.57 7.42 7.38 5.72 5.74 88 1.38 1.19 1.38 2.61 2.96 3.28 1.67 1.37 89 1.13 1.15 1.40 2.94 3.54 3.90 2.75 2.26 90 1.94 1.73 2.48 3.13 3.77 3.97 2.16 1.48 91 1.64 1.39 .67 2.08 3.16 2.90 1.67 .40 92 1.91 3.11 3.01 1.86 .76 .02 .83 1.67 93 1.98 3.20 3.38 1.22 5.07 4.23 3.25 1.22 94 .39 3.81 4.52 6.00 6.99 6.50 4.80 2.92 95 2.29 3.32 3.46 2.54 1.77 3.30 2.58 .38 96 2.67 2.69 2.60 5.03 5.50 4.68 4.43 2.89 97 2.57 3.12 3.40 4.51 4.64 5.24 3.01 1.86 98 .00 .00 .00 .00 .00 .00 .00 .00 99 .00 .00 .00 .00 .00 .00 .00 .00 100 .00 .00 .00 .00 .00 .00 .00 .00 PAGE 197 181 APPENDIX B continued Sensor # Time 12:00 12:30 13 :00 13:30 14:00 14:30 15:00 15:3 0 1 .00 .00 .00 .00 .00 .00 .00 .00 2 .00 .00 .00 .00 .00 .00 .00 .00 3 .86 4.82 3.44 .91 .26 1.04 2.06 2.69 4 4.13 4.19 3.59 3.26 1.27 .39 .44 .64 6 3.63 3.84 1.39 .86 .62 .87 .41 .20 7 .35 .44 .91 .36 .39 .02 .62 1.22 8 6.20 4.11 4.87 2.00 1.87 1.26 .14 .54 9 3.41 2.14 2.65 1.46 1.95 .89 .86 1.11 10 5.29 4.80 2.73 .62 .66 .14 .48 2.22 11 3.44 5.42 4.74 3.29 2.07 1.31 .84 .24 12 7.85 7.76 7.02 3.97 2.31 2.01 2.71 2.08 13 3.96 4.00 3.19 .94 .12 .08 .09 .71 14 5.92 5.69 3.81 2.42 1.50 .51 .58 .04 15 2.71 2.17 2.68 3.15 2.00 1.16 1.16 1.48 16 .38 1.46 2.68 3.34 3.47 3.50 3.95 3.43 17 2.36 3.36 4.48 3.16 1.67 .71 1.20 .43 18 3.07 3.33 4.09 2.13 1.00 .39 .93 .58 19 5.06 4.71 3.54 2.33 .99 .22 .43 .06 20 6.61 5.94 4.86 3.28 2.48 1.49 1.74 1.78 21 2.48 2.26 1.57 1.50 2.10 2.46 2.32 2.58 22 1.67 .98 .04 .03 .60 .38 1.62 .88 23 5.72 4.47 4.57 2.81 2.09 1.78 1.32 1.37 25 8.91 6.39 7.26 4.87 4.28 3.50 3.38 3.47 26 4.23 3.48 2.32 2.22 2.01 2.26 1.68 1.53 27 3.95 4.80 3.18 2.48 1.86 1.34 1.04 .55 28 4.55 3.99 4.26 1.85 .83 .28 .41 .18 29 6.17 6.86 6.6 9 4.51 3.92 2.76 3.22 2.17 30 3.30 3.01 3.32 2.61 2.66 1.82 1.57 .78 31 3.80 4.18 3.61 2.27 1.41 .94 1.01 .41 32 .80 .18 .46 .01 .35 .34 .28 .61 33 3.41 2.14 2.65 1.46 1.95 .89 .86 1.11 34 3.08 1.89 2.87 2.59 3.83 3.29 3.23 2.47 35 3.04 2.30 3.06 1.79 2.26 1.29 1.49 1.48 36 .55 .93 1.67 .81 .96 .06 1.49 .98 38 2.65 2.59 1.31 1.66 1.45 1.31 .19 .84 39 3.54 2.78 3.56 3.08 3.54 3.26 4.04 4.76 PAGE 198 182 APPENDIX B continued Sensor # Tim e 12:00 12:30 13:00 13:30 14:00 14:30 15:00 15:3 0 40 1.15 .26 .81 .71 1.14 1.31 2.52 1.71 41 1.64 1.84 2.16 1.17 1.34 1.45 1.60 1.73 42 1.25 1.89 .66 .58 .31 .52 1.37 .82 43 .56 1.57 2.36 1.84 1.69 1.39 1.64 1.62 44 .23 .66 1.00 .27 .51 .19 .89 .25 45 2.50 1.52 2.54 2.31 1.67 1.58 2.61 2.24 46 .67 .72 .95 .98 .68 .34 1.16 .91 47 .01 .93 .93 1.96 1.47 1.56 2.85 2.03 49 2.22 1.49 .66 1.15 1.14 1.72 1.21 1.52 50 1.78 1.86 1.62 1.81 1.47 2.15 2.48 1.98 51 .91 .33 .60 .38 .00 .08 2.56 3.04 52 1.26 1.29 1.73 1.33 1.42 1.44 1.50 2.08 53 2.90 1.89 2.71 2.49 1.93 1.54 1.64 1.65 54 5.34 4.97 2.84 2.81 2.57 1.88 3.39 3.11 55 5.00 5.01 4.13 3.83 3.93 3.32 3.79 4.19 56 10.27 10.48 8.41 6.50 3.94 4.01 4.16 3.85 57 4.03 4.14 2.85 3.00 2.75 2.72 2.85 3.38 58 1.81 1.13 1.26 1.24 1.23 .97 1.40 1.28 59 .08 .70 .06 .31 .65 1.08 .06 .21 60 1.90 .98 1.58 .71 .60 .90 1.38 1.12 61 .74 .77 .45 .68 .37 .67 1.86 1.90 62 2.84 2.72 3.57 1.57 .57 .15 1.44 .93 63 1.93 1.78 .82 1.08 .79 1.03 .32 .43 64 .25 .70 .39 .85 .60 .03 .09 .82 65 3.19 2.16 2.44 1.95 2.76 3.24 3.99 3.79 66 .71 .99 .36 1.06 .24 1.12 1.22 .58 67 .66 .49 .87 .83 1.49 .99 1.32 .73 68 1.46 2.28 3.47 3.28 5.20 5.28 6.10 6.03 69 3.07 3.51 3.67 1.44 2.36 2.24 2.73 1.85 70 9.82 9.21 8.51 3.83 3.33 2.41 2.52 1.52 71 1.50 .98 .39 .54 1.01 1.09 1.34 1.01 73 1.75 1.05 1.84 1.77 1.90 2.91 3.66 3.33 74 .32 1.30 2.26 2.03 2.31 2.01 2.38 1.69 75 1.27 1.76 2.68 2.58 2.64 3.12 3.94 2.64 76 1.81 1.12 1.62 1.42 1.54 2.30 2.27 2.30 77 .40 .28 2.08 .75 .76 1.05 1.21 1.77 78 .12 .69 1.88 1.78 1.59 2.41 3.54 3.47 PAGE 199 183 APPENDIX B continued Sensor # Time 12:00 12:30 13:00 13:30 14:00 14:30 15:00 15:3 0 80 .38 1.25 2.42 1.04 .49 1.34 2.21 2.60 81 2.75 .40 .87 .75 1.05 .97 .23 .47 82 2.33 2.00 3.44 3.03 2.89 3.22 4.47 4.29 83 .55 .28 .52 .46 1.28 .27 1.45 .61 84 3.84 3.83 4.54 2.63 3.70 3.5 7 4.86 3.58 85 1.00 .27 2.31 .16 .43 .81 .85 1.31 86 1.51 1.21 .46 .91 .58 1.07 .94 1.69 87 4.47 4.06 3.76 3.33 3.37 2.64 2.34 1.01 88 .28 2.29 1.99 .97 1.43 1.15 .61 .03 89 1.65 2.71 3.07 1.94 1.46 1.57 1.63 2.18 90 .10 .84 1.08 .00 .00 .25 .57 1.07 91 1.31 .61 .55 .43 .60 .99 .55 1.39 92 1.28 .16 1.57 2.32 3.00 3.65 4.87 3.47 93 1.14 1.60 .86 .65 .91 1.15 1.55 2.62 94 2.58 2.93 1.26 .29 .15 .12 04 1.08 95 1.32 2.27 3.00 1.50 1.40 .78 .73 .55 96 3.04 3.41 2.84 2.04 2.22 1.96 2.95 2.19 97 .35 .26 .32 .24 .04 .68 .49 .93 98 .00 .00 .00 .00 .00 .00 .00 .00 99 .00 .00 .00 .00 .00 .00 .00 .00 100 .00 .00 .00 .00 .00 .00 .00 .00 PAGE 200 184 APPENDIX B continued Sensor # Time 16:00 16:30 17:00 17:30 18:00 18:30 19:00 19:3 0 1 .00 .00 .00 .00 .00 .00 .00 .00 2 .00 .00 .00 .00 .00 .00 .00 .00 3 2.89 2.22 1.79 2.29 2.16 1.96 2.25 1.47 4 1.54 .88 .84 1.18 .95 .42 .82 .16 6 .46 .94 .16 .36 .18 .75 .26 .55 7 1.38 .67 .54 .94 .35 .33 .18 .47 8 1.08 .67 .61 1.29 .75 .19 .23 .53 9 1.56 1.63 .79 .63 .42 .06 .02 .10 10 1.61 1.44 1.80 2.19 1.80 1.04 1.24 .69 11 .36 .07 .11 .49 .05 .54 .54 .86 12 1.41 1.34 1.08 .44 .65 1.34 1.83 1.73 13 .75 .74 .93 2.62 2.33 1.98 2.10 1.74 14 .43 .48 .92 1.59 1.19 .42 .74 .42 15 1.52 1.26 1.47 .81 1.57 1.46 1.53 1.77 16 4.14 3.27 3.25 2.28 2.90 2.87 2.65 1.23 17 .20 .47 .39 1.32 1.34 .92 1.12 .88 18 .44 .04 .19 1.21 1.28 .82 .64 .38 19 .55 1.05 1.19 1.99 2.07 1.60 1.60 1.27 20 .36 .07 .43 .70 .84 .80 1.24 .69 21 3.06 3.90 3.46 3.45 2.58 2.21 2.47 1.51 22 1.04 .36 .29 2.88 2.98 2.30 2.29 1.16 23 .27 .40 .32 .81 .10 .11 1.14 .47 25 1.83 1.15 1.11 .91 .43 .66 .65 1.15 26 1.95 2.56 2.24 2.17 1.44 .67 .56 .05 27 .15 .34 .47 1.17 1.23 .19 .32 .70 28 1.06 1.33 1.29 1.99 2.11 1.68 1.97 1.18 29 1.9 7 1.70 1.71 .64 .37 1.15 1.08 1.32 30 1.12 1.50 1.95 1.74 1.88 2.47 2.50 2.51 31 .52 .75 .55 .63 1.01 .09 .02 .04 32 .12 .69 .94 1.63 .65 1.59 .13 .08 33 1.56 1.63 .79 .63 .42 .06 .02 .10 34 2.35 2.04 3.05 3.02 2.72 2.59 2.50 .45 35 1.35 1.66 1.82 1.55 1.61 1.30 1.08 .89 36 1.53 .62 1.38 .21 .51 .45 1.58 .89 38 .27 .03 .84 1.08 .80 1.27 .93 .86 39 3.88 3.16 4.47 3.35 3.22 3.85 3.19 3.28 PAGE 201 185 APPENDIX B continu ed Sensor # Time 16:00 16:30 17:00 17:30 18:00 18:30 19:00 19:3 0 40 1.38 .76 1.40 1.16 1.48 1.49 1.35 1.41 41 1.61 1.42 2.19 1.30 2.08 1.68 1.16 1.21 42 .71 .0 6 .65 .77 1.44 .48 1.58 .62 43 .49 .33 .05 .26 1.00 .55 .83 1.17 44 .43 .36 .30 .45 .96 .20 .32 .07 45 1.73 1.41 1.35 2.66 3.21 2.80 2.73 2.61 46 .94 1.28 1.83 1.86 2.54 .71 .07 1.62 47 .78 1.92 2.29 2.11 2.41 .78 1.78 .16 49 2.07 2.40 1.80 .85 .57 .50 .13 .95 50 1.26 .09 .05 .03 .41 .09 1.25 .46 51 2.30 .73 2.30 1.70 2.24 .30 .67 .53 52 2.07 2.09 1.50 1.68 2.23 2.63 2.62 2.60 53 1.25 1.98 2. 23 1.13 .98 2.13 1.74 .69 54 2.36 1.62 1.88 1.75 2.24 2.15 .57 .48 55 2.79 3.01 4.06 4.00 3.42 4.03 4.12 3.55 56 3.27 2.18 2.78 2.94 3.25 2.3 0 .17 .71 57 3.80 2.74 2.76 3.09 1.18 .93 .50 1.73 58 1.89 .82 .83 .48 2.07 .36 2.06 .97 59 .54 .65 .04 .27 .03 .65 .39 .02 60 .48 .81 1.70 1.12 1.11 2.13 1.31 1.84 61 1.65 1.02 1.00 1.09 1.62 2.15 1.50 1.42 62 .24 .17 .51 .60 .84 1.83 1.26 1.47 63 .97 1.24 .72 1 .09 1.24 .43 .78 .21 64 1.30 1.76 1.32 1.47 1.56 .35 .90 .15 65 3.84 3.17 3.36 2.95 2.76 3.88 3.57 3.32 66 .05 .43 .30 .10 .02 1.42 96 1.39 67 .50 .72 .74 .26 .01 .69 .34 .80 68 3.80 2.00 2.24 2.64 2.29 1.91 .79 1.68 69 1.29 1.04 .78 1.15 1.28 1.79 1.42 .96 70 1.10 .56 .18 .26 .31 .78 .52 1.14 71 1.10 1.78 1.79 1.90 2.78 4.02 3.91 4.09 73 3.08 1.66 1.15 1.99 2.08 2.95 1.97 .22 74 2.00 1.85 1.89 2.11 1.94 2.48 1.79 .55 75 2.67 2.84 3.27 2.96 3.22 3.91 3.23 1.42 76 1.63 1.54 .76 1.10 .35 .85 .17 .69 77 .75 .14 1.24 1.14 1.93 1.44 1.52 .46 78 3.08 2.16 2.60 1.75 2.17 2.62 2.94 2.83 PAGE 202 186 APPENDIX B continued Sensor # Time 16:00 16:30 17:00 17:30 18:00 18:30 19:00 19:3 0 80 1.84 1.67 2.25 .92 1.10 1.42 1.26 1 .29 81 .14 1.20 2.28 3.68 3.91 3.17 3.08 2.01 82 3.13 3.62 3.61 4.00 3.64 2.78 .20 .90 83 .27 .84 1.22 1.60 1.81 .76 .61 .01 84 3.45 2.06 2.54 1.51 1.50 2.35 .65 .72 85 1.91 1.92 1.10 1.27 2.27 1.88 2.08 1.13 86 2.16 2.42 1.81 2.73 2.73 1.73 1.71 1.05 87 .97 .43 1.00 .90 1.20 .82 .82 .12 88 .79 .92 .62 .43 .41 .93 .33 .14 89 2.46 2.11 1.85 1.94 1.74 2.73 2.03 1.63 90 .56 1.38 1.71 2.27 2.30 1.66 1.72 .90 91 .63 1.77 2.42 3.21 3.22 2.41 2.58 1.45 92 2.42 .88 1.85 2.69 2.55 1.90 2.11 1.18 93 2.78 2.43 2.55 3.32 3.14 2.38 2.54 1.20 94 1.36 2.54 2.23 1.75 .76 .69 1.67 .98 95 .92 1.33 1.02 1.15 .98 .10 .67 1.66 96 2.18 1.37 1.46 .97 1.23 2.57 1.80 2.21 97 1.07 1.01 1.89 2.14 2.06 1.15 1.17 .14 98 .00 .00 .00 .00 .00 .00 .00 .00 99 .00 .00 .00 .00 .00 .00 .00 .00 100 .00 .00 .00 .00 .00 .00 .00 .00 PAGE 203 187 APPENDIX B continued Sensor # Time 20:00 20:30 21:00 21:30 22:00 22:30 23:00 23:3 0 1 .00 .00 .00 .00 .00 .00 .00 .00 2 .00 .00 .00 .00 .00 .00 .00 .00 3 .46 .06 .64 .91 1.02 .93 1.02 1.34 4 1.02 1.41 1.61 1.56 1.57 1.18 .82 1.14 6 .23 .10 .16 .18 .15 .00 .02 .12 7 1.28 1.61 2.00 1. 92 1.91 1.69 1.36 1.43 8 1.42 1.63 2.05 2.02 2.05 1.91 1.69 1.82 9 .71 .93 1.29 1.20 .98 .73 .82 .69 10 .39 .50 .76 .42 .34 .00 .0 9 .06 11 2.00 2.31 2.50 2.40 1.97 1.81 1.94 1.80 12 .92 .94 1.04 1.32 1.46 1.89 1.74 1.68 13 .78 .62 .69 .95 1.01 1.39 1.23 1.22 14 .00 .08 .03 .14 .35 .53 .36 .44 15 2.44 2.19 2.06 1.96 1.40 1.25 1.21 1.14 16 1.59 1.36 1.11 1.04 .90 .78 .87 .89 17 .30 .00 .12 .01 .10 .06 .19 .33 18 .14 .01 .10 .03 .24 .15 .02 .06 19 .29 .01 .01 .05 .10 .14 .02 .10 20 .25 .37 .45 .28 .06 .06 .09 20 21 .20 .38 .69 .60 .65 .61 .93 1.12 22 .12 .30 .62 .59 .60 .64 .87 1.09 23 .33 .43 .46 .44 .42 .29 .56 .75 25 2.15 2.37 2.34 2.25 2.25 2.09 2.43 2.67 26 .90 .93 1.17 1.01 1.08 .90 1.14 1.43 27 .96 1.15 1.22 .98 .89 .75 1.01 1.14 28 .13 .32 .38 .43 .13 .04 .23 .46 29 1.59 1.27 1.37 1.15 .83 .59 .81 .95 30 2.92 2.59 2.83 2.59 2.37 1.95 2.02 1.99 31 .35 .46 .70 .66 .54 .34 .56 .69 32 .66 .89 1.00 .99 1.10 1.42 1.27 1.32 33 .71 .93 1.29 1.20 .98 .73 .82 .69 34 .89 .68 .89 .97 .88 .77 .95 1.14 35 .85 .66 .73 .82 .76 .68 .86 .99 36 .44 .17 .51 .63 .69 .63 .89 1.06 38 .70 .89 .90 .68 .76 .78 1.13 1.32 39 2.41 2.10 2.01 2.03 1.74 1.61 1.90 2.12 40 1.72 1.48 1.40 1.56 1.28 1.37 1.72 2.01 PAGE 204 188 APPENDIX B continued Sensor # Time 20:00 20:30 21:00 21:30 22:00 22:30 23:00 23:3 0 40 1.72 1.48 1.40 1.56 1.28 1.37 1.72 2.01 41 .67 .72 .78 .91 .73 .60 .97 1.14 42 .45 .91 .81 1.04 .95 .93 1.31 1.43 43 1.82 1.99 2.14 2.32 1.96 2.07 2.44 2.68 44 .56 .62 .93 1.25 1.08 .93 1.23 1.55 45 2.51 2.62 2.89 3.02 2.81 2.63 2.97 3.11 46 1.96 2.34 2.62 2.55 2.43 2.24 2.53 2.75 47 1.14 1.51 1.83 1.92 2.00 1.96 2.26 2.45 49 1.04 1.26 1.48 1.72 1.63 1.56 1.91 2.13 50 .73 1.01 1.28 1.35 1.12 .95 1.29 1.38 51 1.64 1.88 2.35 2.56 2.28 2.13 2.38 2.46 52 2.54 1.61 1.82 2.04 1.74 1.49 1.74 1.84 53 1.47 1.46 1.82 1.96 1.72 1.54 1.79 1.79 54 1.40 1.30 1.43 1.51 1.33 1.14 1.41 1.49 55 3.63 3.35 3.53 3.54 3.41 3.30 3.51 3.63 56 1.32 1.23 1.26 1.16 1.09 .87 1.18 1.22 57 1.97 2.07 2.12 2.15 1.99 1.94 2.24 2.16 58 .27 .49 .72 .85 .67 .68 .91 .88 59 .87 .86 1.06 1.39 1.21 1.16 1.44 1.43 60 1.62 1.40 1.67 1.91 1.83 1.70 1.88 1.91 61 1.84 1.17 1.49 1.73 1.66 1.58 1.89 1.84 62 1.68 1.14 1.19 1.41 1.29 1.19 1.45 1.34 63 .19 .54 .70 1.02 .93 .90 1.17 1.07 64 1.39 1.90 2.00 2.45 2.52 2.45 2.59 2.70 65 2.91 2.40 2.21 2.36 2.41 2.19 2.27 2.29 66 1.42 1.52 1.47 1.93 2.03 1.98 2.10 2.06 67 1.54 1.09 1.23 1.41 1.66 1.53 1.51 1.64 68 2.88 2.89 3.04 3.09 3.20 2.86 2.83 2.90 69 2.14 2.10 2.14 2.20 2.41 2.23 2.16 2.37 70 1.94 1.81 1.84 1.93 2.04 1.92 1.90 1.98 71 4.66 4.42 4.44 4.43 4.45 4.06 4.01 4.04 73 .9 1 .85 1.05 1.29 1.47 1.29 1.50 1.54 74 1.24 1.30 1.38 1.57 1.64 1.45 1.80 1.86 75 2.18 1.93 2.09 2.35 2.42 2.20 2.40 2.44 76 .08 .16 .45 .54 .42 .01 .22 .39 77 .58 .67 1.04 1.21 1.04 .85 1.15 1.14 PAGE 205 189 APPENDIX B continued Sensor # Time 20:00 20:30 21:00 21:30 22:00 22:30 23:00 23:3 0 78 3.21 2.64 2.93 2. 95 2.65 2.38 2.61 2.67 80 .22 .01 .19 .41 .11 .05 .21 .39 81 .71 .34 .02 .23 .08 .07 .06 .16 82 1.83 2.00 2.21 2.09 1.85 1.55 1.8 3 2.04 83 .55 .42 .76 .67 .55 .25 .54 .67 84 1.50 1.64 1.84 1.64 1.27 1.05 1.29 1.43 85 .18 .74 1.04 .98 .63 .51 .79 1.00 86 .20 .77 1.20 1.04 .82 .60 .90 1.10 87 .83 1.16 1.40 1.25 1.18 .88 1.02 1.18 88 .37 .13 .52 .43 .47 .03 .05 .25 89 .75 .83 1.10 1.12 1. 17 .71 .94 1.07 90 .27 .85 1.35 1.33 1.39 .92 1.14 1.31 91 .37 .16 .40 .50 .54 .09 .20 .34 92 .21 .30 .68 .80 .77 .42 .65 .8 1 93 .14 .40 .98 1.15 1.27 .93 .94 1.11 94 .32 .51 .99 1.22 1.30 1.04 1.07 1.25 95 2.15 2.47 2.72 2.80 2.58 2.11 2.26 2.23 96 2.57 2.25 2.37 2.39 2.31 1.93 2.00 1.97 97 1.29 1.81 2.15 2.24 2.34 2.05 2.20 2.34 98 .00 .00 .00 .00 .00 .00 .00 .00 99 .00 .00 .00 .00 .00 .00 .00 .00 100 .00 .00 .00 .00 .00 .00 .00 .00 PAGE 206 190 APPENDIX C CORRELATION VALUES OF IMPERVIOUS PERCENT TO TIME PERIOD DELTA TEMPERATURES PAGE 207 191 APPENDIX C continued Imp_ percent _25m Imp_ percent _30m Imp_ percent _50m Imp_ percent _100m K_0_00 Pearson Correlation .442(**) .547(**) .573(**) .532(**) Sig. (2 tailed) 8.0730E 06 1.1442E 08 1.6021E 09 3.3682E 08 N 94 94 94 94 K_0_30 Pearson Correlation .442(**) .550(**) .579(**) .544(**) Sig. (2 tailed) 8.1483E 06 9.44 139E 09 1.01586E 09 1.4259E 08 N 94 94 94 94 K_1_00 Pearson Correlation .436(**) .548(**) .578(**) .553(**) Sig. (2 tailed) 1.14803E 05 1.08545E 08 1.02409E 09 7.40529E 09 N 94 94 94 94 K_1_30 Pearson Correlation .430(**) .543(**) .579(**) .555(**) Sig. (2 tailed) 1.55085E 05 1.60716E 08 9.48885E 10 6.30364E 09 N 94 94 94 94 K_2_00 Pearson Correlation .424(**) .540(**) .577(**) .553(**) Sig. (2 tailed) 2.02264E 05 1.89403E 08 1.18542E 09 7.23072E 09 N 94 94 94 94 K_2_30 Pearson Correlation .426(**) .541(**) .578(**) .557(**) Sig. (2 tailed) 1.84352E 05 1.82418E 08 1.02637E 09 5.33994E 09 N 94 94 94 94 K_3_00 Pearson Correlation .409(**) .531(**) .569(**) .551(**) Sig. (2 tailed) 4.20435E 05 3.69512E 08 2.26952E 09 8.85044E 09 N 94 94 94 94 K_3_30 Pearson Correlation .398(**) .518(**) .560(**) .548(**) Sig. (2 tailed) 7.19052E 05 8.83447E 08 4.41906E 09 1.04718E 08 N 94 94 94 94 K_4_00 Pearson Correlation .389(**) .512(**) .555(**) .543(**) Sig. (2 tailed) 0.000105057 1.33234 E 07 6.25724E 09 1.50969E 08 N 94 94 94 94 K_4_30 Pearson Correlation .389(**) .509(**) .553(**) .544(**) Sig. (2 tailed) 0.000108619 1.58845E 07 7.63678E 09 1.48296E 08 N 94 94 94 94 K_5_00 Pearson Correlation .385(**) .508(**) .550(**) .542(**) Sig. (2 tailed) 0.00012856 1.7165E 07 9.28162E 09 1.72093E 08 N 94 94 94 94 K_5_30 Pearson Correlation .386(**) .509(**) .550(**) .542(**) Sig. (2 tailed) 0.000124447 1.58882E 07 9.48331E 09 1.69504E 08 N 94 94 94 94 K_6_00 Pearson Correlation .38 5(**) .508(**) .549(**) .546(**) Sig. (2 tailed) 0.000126699 1.73686E 07 9.94825E 09 1.27402E 08 N 94 94 94 94 PAGE 208 192 APPENDIX C continued K_6_30 Pearson Correlation .400(**) .524(**) .564(**) .560(**) Sig. (2 tailed) 6.62035E 05 5.95306E 08 3.14157E 09 4.54073E 09 N 94 94 94 94 K_7_00 Pearson Correlation .429(**) .539(**) .563(**) .561(**) Sig. (2 tailed) 1.56362E 05 2.06248E 08 3.41117E 09 3.99554E 09 N 94 94 94 94 K_7_30 Pearson Correlation .222(*) .244(*) .247(*) .2 98(**) Sig. (2 tailed) 0.03161257 0.017760206 0.016378119 0.00354674 N 94 94 94 94 K_8_00 Pearson Correlation 0.025484305 0.024546357 0.003976266 0.036330668 Sig. (2 tailed) 0.807375515 0.814335327 0.969659248 0.728109991 N 94 94 94 94 K_8_30 Pea rson Correlation 0.020619737 0.027447474 0.081094555 0.065023051 Sig. (2 tailed) 0.843622623 0.792857341 0.437157614 0.533516979 N 94 94 94 94 K_9_00 Pearson Correlation 0.060984931 0.008730818 0.046085059 0.033413374 Sig. (2 tailed) 0.55928519 5 0.933440136 0.659164213 0.749188575 N 94 94 94 94 K_9_30 Pearson Correlation 0.07991663 0.014320459 0.029066485 0.017347371 Sig. (2 tailed) 0.443867195 0.891037664 0.780936995 0.868195428 N 94 94 94 94 K_10_00 Pearson Correlation 0.082730608 0. 034556901 0.016239252 0.017936885 Sig. (2 tailed) 0.427934471 0.740903184 0.876546186 0.863758686 N 94 94 94 94 K_10_30 Pearson Correlation 0.067222254 0.027117606 0.000821472 0.042829233 Sig. (2 tailed) 0.519736008 0.79529202 0.993730365 0.6818961 53 N 94 94 94 94 K_11_00 Pearson Correlation 0.060249338 0.032448931 0.001797713 0.033497757 Sig. (2 tailed) 0.564042359 0.756198668 0.986280033 0.748576199 N 94 94 94 94 K_11_30 Pearson Correlation 0.037454338 0.005515875 0.034375794 0.001695927 Sig. (2 tailed) 0.720043711 0.957920545 0.742213453 0.987056786 N 94 94 94 94 K_12_00 Pearson Correlation 2.50574E 05 0.038024118 0.074877029 0.038865725 Sig. (2 tailed) 0.999808755 0.715965161 0.47321726 0.709955421 N 94 94 94 94 PAGE 209 193 A PPENDIX C continued K_12_30 Pearson Correlation 0.01378706 0.041972691 0.065690287 0.027258419 Sig. (2 tailed) 0.895072684 0.687925004 0.529316726 0.794252469 N 94 94 94 94 K_13_00 Pearson Correlation 0.022590638 0.027499139 0.038106117 0.017012 347 Sig. (2 tailed) 0.828893129 0.792476192 0.715378849 0.870718675 N 94 94 94 94 K_13_30 Pearson Correlation 0.033823313 0.02747725 0.0013241 0.054722791 Sig. (2 tailed) 0.746215049 0.792637667 0.989894371 0.600382106 N 94 94 94 94 K_14_00 Pears on Correlation 0.145862244 0.128059297 0.100906308 0.135636818 Sig. (2 tailed) 0.160682479 0.218682011 0.333191322 0.192409102 N 94 94 94 94 K_14_30 Pearson Correlation 0.141536987 0.125535949 0.113715644 0.138544327 Sig. (2 tailed) 0.173591922 0.22 7968959 0.275137857 0.182959864 N 94 94 94 94 K_15_00 Pearson Correlation 0.169733023 0.16486993 0.16668724 0.180694955 Sig. (2 tailed) 0.101945366 0.112292368 0.10833438 0.081361724 N 94 94 94 94 K_15_30 Pearson Correlation 0.160123726 0.150631948 0.159219163 0.170730504 Sig. (2 tailed) 0.123158455 0.147289217 0.12531808 0.099918373 N 94 94 94 94 K_16_00 Pearson Correlation 0.179106076 0.14234335 0.131989438 0.154624347 Sig. (2 tailed) 0.084120966 0.171129134 0.204755398 0.13673895 N 94 94 94 94 K_16_30 Pearson Correlation .221(*) 0.196699056 0.200943083 .238(*) Sig. (2 tailed) 0.031918034 0.057412183 0.0521364 0.021012126 N 94 94 94 94 K_17_00 Pearson Correlation .259(*) .260(*) .266(**) .311(**) Sig. (2 tailed) 0.011790378 0.01152 3353 0.009560733 0.002304344 N 94 94 94 94 K_17_30 Pearson Correlation .261(*) .284(**) .288(**) .302(**) Sig. (2 tailed) 0.011018906 0.005612945 0.004855132 0.003052641 N 94 94 94 94 K_18_00 Pearson Correlation .327(**) .352(**) .357(**) .357(**) Sig. (2 tailed) 0.001315128 0.000493768 0.000416584 0.000404636 N 94 94 94 94 PAGE 210 194 APPENDIX C continued K_18_30 Pearson Correlation .339(**) .374(**) .347(**) .345(**) Sig. (2 tailed) 0.000836006 0.000208076 0.000619101 0.00066191 N 94 94 94 94 K_19_00 Pearson Correlation .410(**) .443(**) .422(**) .404(**) Sig. (2 tailed) 4.00374E 05 7.67925E 06 2.29473E 05 5.38632E 05 N 94 94 94 94 K_19_30 Pearson Correlation .491(**) .543(**) .505(**) .458(**) Sig. (2 tailed) 4.96189E 07 1.53694E 08 2.09466E 07 3.40486E 06 N 94 94 94 94 K_20_00 Pearson Correlation .472(**) .533(**) .518(**) .456(**) Sig. (2 tailed) 1.54334E 06 3.16568E 08 8.92578E 08 3.93823E 06 N 94 94 94 94 K_20_30 Pearson Correlation .436(**) .522(**) .509(**) .442(**) S ig. (2 tailed) 1.10517E 05 6.8851E 08 1.60042E 07 8.16058E 06 N 94 94 94 94 K_21_00 Pearson Correlation .447(**) .528(**) .529(**) .468(**) Sig. (2 tailed) 6.25373E 06 4.44775E 08 4.35383E 08 1.95741E 06 N 94 94 94 94 K_21_30 Pearson Correlation .4 32(**) .524(**) .530(**) .478(**) Sig. (2 tailed) 1.40311E 05 6.04002E 08 3.83447E 08 1.13316E 06 N 94 94 94 94 K_22_00 Pearson Correlation .421(**) .518(**) .532(**) .477(**) Sig. (2 tailed) 2.35482E 05 8.89994E 08 3.44409E 08 1.19663E 06 N 94 94 94 94 K_22_30 Pearson Correlation .413(**) .512(**) .528(**) .474(**) Sig. (2 tailed) 3.48792E 05 1.35945E 07 4.44831E 08 1.36599E 06 N 94 94 94 94 K_23_00 Pearson Correlation .431(**) .527(**) .545(**) .496(**) Sig. (2 tailed) 1.44665E 05 4.94333 E 08 1.31897E 08 3.7068E 07 N 94 94 94 94 K_23_30 Pearson Correlation .430(**) .532(**) .557(**) .510(**) Sig. (2 tailed) 1.53505E 05 3.48455E 08 5.68665E 09 1.56661E 07 N 94 94 94 94 ** Correlation is significant at the 0.01 % level (2 tailed). Correlation is significant at the 0.05 % level (2 tailed). PAGE 211 195 APPENDIX D 2007 PERIOD OLS REGRESSION RESULTS WITH CONSTANT VALUE PAGE 212 196 APPENDIX D continued Time 00:00 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .573 .328 .321 .847 The independent variable is Imp : percent : 50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 32.265 1 32.265 44.946 .000 Residual 66.043 92 .718 Total 98.308 93 The independent variable is Imp : percent : 50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp : percent : 50m .031 .005 .573 6.704 .000 (Constant) .017 .220 .077 .939 PAGE 213 197 APPENDIX D continued Time 00:30 Model Summary R R Square Adju sted R Square Std. Error of the Estimate .579 .335 .327 .857 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 33.974 1 33.974 46.282 .000 Residual 67.534 92 .734 Total 101.508 93 The independe nt variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .032 .005 .579 6.803 .000 (Constant) .050 .222 .227 .821 PAGE 214 198 APPENDIX D continued Time 01:00 Model Summa ry R R Square Adjusted R Square Std. Error of the Estimate .578 .335 .327 .873 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 35.216 1 35.216 46.258 .000 Residual 70.040 92 .761 Total 105.256 9 3 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .033 .005 .578 6.801 .000 (Constant) .046 .226 .202 .841 PAGE 215 199 APPENDIX D continued T ime 01:30 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .579 .336 .328 .870 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 35.172 1 35.172 46.483 .000 Residual 69.614 92 757 Total 104.786 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .033 .005 .579 6.818 .000 (Constant) .058 .226 .259 .796 PAGE 216 200 APP ENDIX D continued Time 02:00 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .577 .333 .325 .879 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 35.411 1 35.411 45.828 .000 R esidual 71.087 92 .773 Total 106.498 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .033 .005 .577 6.770 .000 (Constant) .032 .2 28 .140 .889 PAGE 217 201 APPENDIX D continued Time 02:30 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .578 .335 .327 .869 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 34.929 1 34.929 46.252 .000 Residual 69.478 92 .755 Total 104.407 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .033 .005 .578 6.801 .00 0 (Constant) .047 .225 .206 .837 PAGE 218 202 APPENDIX D continued Time 03:00 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .569 .323 .316 .887 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 34.595 1 34.595 43.935 .000 Residual 72.441 92 .787 Total 107.036 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .03 3 .005 .569 6.628 .000 (Constant) .081 .230 .351 .726 PAGE 219 203 APPENDIX D continued Time 03:30 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .560 .314 .306 .892 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 33.411 1 33.411 42.022 .000 Residual 73.149 92 .795 Total 106.560 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .032 .005 .560 6.482 .000 (Constant) .056 .231 .242 .810 PAGE 220 204 APPENDIX D continued Time 04:00 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .555 .308 .301 .900 The independent variable is Imp_percent_50m. AN OVA Sum of Squares df Mean Square F Sig. Regression 33.226 1 33.226 41.034 .000 Residual 74.493 92 .810 Total 107.719 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .032 .005 .555 6.406 .000 (Constant) .025 .233 .107 .915 PAGE 221 205 APPENDIX D continued Time 04:30 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .553 .306 .298 .900 The independent variable i s Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 32.815 1 32.815 40.472 .000 Residual 74.594 92 .811 Total 107.409 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardize d Coefficients t Sig. B Std. Error Beta Imp_percent_50m .032 .005 .553 6.362 .000 (Constant) .012 .234 .051 .960 PAGE 222 206 APPENDIX D continued Time 05:00 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .550 .303 .295 .899 The ind ependent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 32.275 1 32.275 39.924 .000 Residual 74.373 92 .808 Total 106.648 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coeff icients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .031 .005 .550 6.319 .000 (Constant) .021 .233 .091 .928 PAGE 223 207 APPENDIX D continued Time 05:30 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .550 .3 02 .295 .896 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 32.029 1 32.029 39.864 .000 Residual 73.918 92 .803 Total 105.947 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .031 .005 .550 6.314 .000 (Constant) .053 .233 .226 .821 PAGE 224 208 APPENDIX D continued Time 06:00 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .549 .302 .294 .892 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 31.643 1 31.643 39.730 .000 Residual 73.274 92 .796 Total 104.917 93 The independent variable is Imp_percent_5 0m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .031 .005 .549 6.303 .000 (Constant) .082 .232 .354 .724 PAGE 225 209 APPENDIX D continued Time 06:30 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .564 .319 .311 .877 The independent variable is Imp : percent : 50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 33.097 1 33.097 42.998 .000 Residual 70.816 92 .770 Total 103.913 93 The independent vari able is Imp : percent : 50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp : percent : 50m .032 .005 .564 6.557 .000 (Constant) .155 .228 .681 .498 PAGE 226 210 APPENDIX D continued Time 07:00 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .563 .317 .310 .884 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 33.385 1 33.385 42.761 .000 Residual 71.828 92 .781 Total 105.213 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .032 .005 .563 6.539 .000 (Constant) .299 .229 1.304 .196 PAGE 227 2 11 APPENDIX D continued Time 07:30 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .247 .061 .051 1.663 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 16.541 1 16.541 5.979 .016 Residual 254.502 92 2.766 Total 271.043 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .023 .009 .247 2.445 .016 (Constant) .309 .431 .717 .475 PAGE 228 212 APPE NDIX D continued Time 08:00 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .004 .000 .011 2.734 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression .011 1 .011 .001 .970 Residu al 687.861 92 7.477 Total 687.872 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .001 .015 .004 .038 .970 (Constant) 1.169 .709 1.648 .103 PAGE 229 213 APPENDIX D continued Time 08:30 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .081 .007 .004 3.325 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 6.732 1 6. 732 .609 .437 Residual 1016.881 92 11.053 Total 1023.612 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .014 .018 .081 .780 .4 37 (Constant) 1.916 .862 2.221 .029 PAGE 230 214 APPENDIX D continued Time 09:00 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .046 .002 .009 3.640 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F S ig. Regression 2.594 1 2.594 .196 .659 Residual 1218.922 92 13.249 Total 1221.517 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .009 .020 .046 .443 .659 (Constant) 2.240 .944 2.372 .020 PAGE 231 215 APPENDIX D continued Time 09:30 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .029 .001 .010 3.901 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 1.184 1 1.184 .078 .781 Residual 1399.812 92 15.215 Total 1400.996 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Err or Beta Imp_percent_50m .006 .022 .029 .279 .781 (Constant) 3.390 1.012 3.350 .001 PAGE 232 216 APPENDIX D continued Time 10:00 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .016 .000 .011 4.108 The independent variable is Imp_p ercent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression .410 1 .410 .024 .877 Residual 1552.820 92 16.878 Total 1553.230 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coeffi cients t Sig. B Std. Error Beta Imp_percent_50m .004 .023 .016 .156 .877 (Constant) 4.183 1.066 3.925 .000 PAGE 233 217 APPENDIX D continued Time 10:30 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .001 .000 .011 3.731 The inde pendent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression .001 1 .001 .000 .994 Residual 1280.991 92 13.924 Total 1280.992 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coeffic ients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .000 .021 .001 .008 .994 (Constant) 4.271 .968 4.412 .000 PAGE 234 218 APPENDIX D continued Time 11:00 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .002 .000 .011 3.489 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression .004 1 .004 .000 .986 Residual 1120.070 92 12.175 Total 1120.073 93 The independent variable is Imp_percent_50m. Coefficients U nstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .000 .019 .002 .017 .986 (Constant) 3.059 .905 3.379 .001 PAGE 235 219 APPENDIX D continued Time 11:30 Model Summary R R Square Adjusted R Square Std. Error of t he Estimate .034 .001 .010 3.278 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 1.169 1 1.169 .109 .742 Residual 988.313 92 10.743 Total 989.482 93 The independent variable is Imp_percent_5 0m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .006 .018 .034 .330 .742 (Constant) 2.460 .850 2.893 .005 PAGE 236 220 APPENDIX D continued Time 12:00 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .075 .006 .005 2.827 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 4.146 1 4.146 .519 .473 Residual 735.289 92 7.992 Total 739.435 93 The independent va riable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .011 .016 .075 .720 .473 (Constant) 2.293 .733 3.127 .002 PAGE 237 221 APPENDIX D continued Time 12:30 Model Summar y R R Square Adjusted R Square Std. Error of the Estimate .066 .004 .007 2.553 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 2.599 1 2.599 .399 .529 Residual 599.580 92 6.517 Total 602.178 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .009 .014 .066 .631 .529 (Constant) 2.330 .662 3.518 .001 PAGE 238 222 APPENDIX D continued Time 13:00 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .038 .001 .009 2.108 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression .595 1 .595 .134 .715 Residual 408.854 92 4.44 4 Total 409.448 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .004 .012 .038 .366 .715 (Constant) 2.356 .547 4.308 .000 PAGE 239 223 A PPENDIX D continued Time 13:30 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .001 .000 .011 1.599 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression .000 1 .000 .000 .990 Res idual 235.346 92 2.558 Total 235.347 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .000 .009 .001 .013 .990 (Constant) 1.394 .41 5 3.359 .001 PAGE 240 224 APPENDIX D continued Time 14:00 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .101 .010 .000 1.458 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 2.013 1 2.013 .946 .333 Residual 195.682 92 2.127 Total 197.695 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .008 .008 .101 .973 .333 (Constant) .944 .378 2.496 .014 PAGE 241 225 APPENDIX D continued Time 14:30 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .114 .013 .002 1.461 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. R egression 2.574 1 2.574 1.205 .275 Residual 196.493 92 2.136 Total 199.067 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .009 .0 08 .114 1.098 .275 (Constant) .679 .379 1.791 .077 PAGE 242 226 APPENDIX D continued Time 15:00 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .167 .028 .017 1.628 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df M ean Square F Sig. Regression 6.973 1 6.973 2.629 .108 Residual 243.977 92 2.652 Total 250.950 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp _percent_50m .015 .009 .167 1.621 .108 (Constant) .794 .422 1.879 .063 PAGE 243 227 APPENDIX D continued Time 15:30 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .159 .025 .015 1.719 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 7.073 1 7.073 2.393 .125 Residual 271.950 92 2.956 Total 279.023 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B St d. Error Beta Imp_percent_50m .015 .010 .159 1.547 .125 (Constant) .420 .446 .942 .349 PAGE 244 228 APPENDIX D continued Time 16:00 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .132 .017 .007 1.652 The independent variable is Imp_pe rcent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 4.450 1 4.450 1.631 .205 Residual 250.984 92 2.728 Total 255.434 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coeffic ients t Sig. B Std. Error Beta Imp_percent_50m .012 .009 .132 1.277 .205 (Constant) .181 .428 .422 .674 PAGE 245 229 APPENDIX D continued Time 16:30 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .201 .040 .030 1.510 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 8.831 1 8.831 3.871 .052 Residual 209.874 92 2.281 Total 218.705 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients S tandardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .016 .008 .201 1.968 .052 (Constant) .269 .392 .686 .494 PAGE 246 230 APPENDIX D continued Time 17:00 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .266 .071 .061 1 .626 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 18.514 1 18.514 7.005 .010 Residual 243.150 92 2.643 Total 261.664 93 The independent variable is Imp_percent_50m. Coefficients Unstan dardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .024 .009 .266 2.647 .010 (Constant) .559 .422 1.326 .188 PAGE 247 231 APPENDIX D continued Time 17:30 Model Summary R R Square Adjusted R Square Std. Error of the Es timate .288 .083 .073 1.746 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 25.386 1 25.386 8.331 .005 Residual 280.322 92 3.047 Total 305.708 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .028 .010 .288 2.886 .005 (Constant) 1.089 .453 2.405 .018 PAGE 248 232 APPENDIX D continued Time 18:00 Model Summary R R Square Adjusted R Sq uare Std. Error of the Estimate .357 .127 .118 1.716 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 39.492 1 39.492 13.416 .000 Residual 270.820 92 2.944 Total 310.311 93 The independent var iable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .035 .010 .357 3.663 .000 (Constant) 1.305 .445 2.932 .004 PAGE 249 233 APPENDIX D continued Time 18:30 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .347 .120 .111 1.588 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 31.709 1 31.709 12.568 .001 Residual 232.108 92 2.523 Total 263.817 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .031 .009 .347 3.545 .001 (Constant) .798 .412 1.936 .056 PAGE 250 234 APPENDIX D continued Time 19:00 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .422 .178 .169 1.484 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 43.875 1 43.875 19.910 .000 Residual 202.736 9 2 2.204 Total 246.611 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .037 .008 .422 4.462 .000 (Constant) 1.464 .385 3.802 .00 0 PAGE 251 235 APPENDIX D continued Time 19:30 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .505 .255 .247 1.111 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 38.886 1 38.886 31.49 6 .000 Residual 113.585 92 1.235 Total 152.471 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .035 .006 .505 5.612 .000 (Constan t) 1.084 .288 3.762 .000 PAGE 252 236 APPENDIX D continued Time 20:00 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .518 .268 .260 .920 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regress ion 28.582 1 28.582 33.743 .000 Residual 77.927 92 .847 Total 106.508 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .030 .005 .5 18 5.809 .000 (Constant) .272 .239 1.141 .257 PAGE 253 237 APPENDIX D continued Time 20:30 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .509 .259 .251 .827 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 22.018 1 22.018 32.200 .000 Residual 62.908 92 .684 Total 84.926 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_per cent_50m .026 .005 .509 5.675 .000 (Constant) .048 .215 .223 .824 PAGE 254 238 APPENDIX D continued Time 21:00 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .529 .279 .272 .830 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 24.565 1 24.565 35.669 .000 Residual 63.360 92 .689 Total 87.924 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. E rror Beta Imp_percent_50m .027 .005 .529 5.972 .000 (Constant) .059 .215 .274 .785 PAGE 255 239 APPENDIX D continued Time 21:30 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .530 .281 .274 .850 The independent variable is Imp_percent _50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 26.033 1 26.033 36.013 .000 Residual 66.505 92 .723 Total 92.538 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .028 .005 .530 6.001 .000 (Constant) .067 .221 .302 .764 PAGE 256 240 APPENDIX D continued Time 22:00 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .532 .283 .275 .852 The independent varia ble is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 26.324 1 26.324 36.305 .000 Residual 66.708 92 .725 Total 93.033 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standar dized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .028 .005 .532 6.025 .000 (Constant) .033 .221 .149 .882 PAGE 257 241 APPENDIX D continued Time 22:30 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .528 .279 .271 .845 T he independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 25.420 1 25.420 35.611 .000 Residual 65.671 92 .714 Total 91.090 93 The independent variable is Imp_percent_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .028 .005 .528 5.967 .000 (Constant) .187 .219 .855 .395 PAGE 258 242 APPENDIX D continued Time 23:00 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .5 45 .297 .290 .840 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 27.499 1 27.499 38.944 .000 Residual 64.964 92 .706 Total 92.463 93 The independent variable is Imp_percent_50m. Coefficien ts Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .029 .005 .545 6.240 .000 (Constant) .063 .218 .291 .772 PAGE 259 243 APPENDIX D continued Time 23:30 Model Summary R R Square Adjusted R Square Std. Error of the Estimate .557 .310 .302 .846 The independent variable is Imp_percent_50m. ANOVA Sum of Squares df Mean Square F Sig. Regression 29.590 1 29.590 41.305 .000 Residual 65.907 92 .716 Total 95.497 93 The independent variable is Imp_percen t_50m. Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta Imp_percent_50m .030 .005 .557 6.427 .000 (Constant) .018 .220 .080 .936 PAGE 260 244 APPENDIX E 2007 PERIOD OLS REGRESSION RESULTS WIT HOUT CONSTANT VALUE PAGE 261 245 APPENDIX E continued Time 00:00 Model Summary a .871 .759 .757 .843 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 208.425 1 208.425 293.479 .000 66.047 93 .710 274.472 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .032 .002 .871 17.131 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 3.00 2.00 1.00 0.00 1.00 2.00 K_0_00 Linear Observed PAGE 262 246 APPENDIX E continued Time 00:30 Model Summary a .878 .772 .769 .852 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 228.265 1 228.265 314.164 .000 67.572 93 .727 295.837 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .033 .002 .878 17.725 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 3.00 2.00 1.00 0.00 1.00 2.00 T_0_30 Linear Observed PAGE 263 247 APPENDIX E continued Time 01:00 Model Summary a .878 .770 .768 .868 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 235.040 1 235.040 311.953 .000 70.071 93 .753 305.111 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .034 .002 .878 17.662 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 3.00 2.00 1.00 0.00 1.00 2.00 K_1_00 Linear Observed PAGE 264 248 APPENDIX E continued Time 01:30 Model Summary a .880 .774 .771 .865 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 238.271 1 238.271 318.085 .000 69.665 93 .749 307.936 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .034 .002 .880 17.835 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 3.00 2.00 1.00 0.00 1.00 2.00 K_1_30 Linear Observed PAGE 265 249 APPENDIX E continued Time 02:00 Model Summary a .867 .752 .749 .874 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 215.540 1 215.540 281.920 .000 71.102 93 .765 286.642 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .032 .002 .867 16.790 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 3.00 2.00 1.00 0.00 1.00 2.00 K_2_00 Linear Observed PAGE 266 250 APPENDIX E continued Time 02:30 Model Summary a .866 .750 .748 .865 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 208.812 1 208.812 279.376 .000 69.510 93 .747 278.323 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .032 .002 .866 16.715 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 3.00 2.00 1.00 0.00 1.00 2.00 K_2_30 Linear Observed PAGE 267 251 APPENDIX E continued Time 03:00 Model Summary a .856 .732 .729 .883 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 198.063 1 198.063 253.932 .000 72.538 93 .780 270.602 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .031 .002 .856 15.935 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 3.00 2.00 1.00 0.00 1.00 2.00 K_3_00 Linear Observed PAGE 268 252 APPENDIX E continued Time 03:30 Model Summary a .854 .729 .726 .887 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 197.128 1 197.128 250.466 .000 73.195 93 .787 270.323 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .031 .002 .854 15.826 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 3.00 2.00 1.00 0.00 1.00 2.00 K_3_30 Linear Observed PAGE 269 253 APPENDIX E continued Time 04:00 Model Summary a .856 .732 .729 .895 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 203.762 1 203.762 254.353 .000 74.502 93 .801 278.265 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .031 .002 .856 15.948 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 3.00 2.00 1.00 0.00 1.00 2.00 K_4_00 Linear Observed PAGE 270 254 APPENDIX E continued Time 04:30 Model Summary a .859 .738 .736 .896 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 210.613 1 210.613 262.574 .000 74.596 93 .802 285.209 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .032 .002 .859 16.204 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 3.00 2.00 1.00 0.00 1.00 2.00 K_4_30 Linear Observed PAGE 271 255 APPENDIX E continued Time 05:00 Model Summary a .853 .728 .725 .894 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 198.820 1 198.820 248.593 .000 74.379 93 .800 273.199 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .031 .002 .853 15.767 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 3.00 2.00 1.00 0.00 1.00 2.00 K_5_00 Linear Observed PAGE 272 256 A PPENDIX E continued Time 05:30 Model Summary a .848 .719 .716 .892 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 189.493 1 189.493 238.278 .000 73.959 93 .795 263.452 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .030 .002 .848 15.436 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 3.00 2.00 1.00 0.00 1.00 2.00 K_5_30 Linear Observed PAGE 273 257 APPENDIX E continued Time 06:00 Model Summary a .843 .710 .707 .888 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 180.061 1 180.061 228.225 .000 73.374 93 .789 253.435 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .030 .002 .843 15.107 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 3.00 2.00 1.00 0.00 1.00 2.00 K_6_00 Linear Observed PAGE 274 258 APPENDIX E continued Time 06:30 Model Summary a .841 .707 .703 .875 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 171.376 1 171.376 223.935 .000 71.172 93 .765 242.548 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .029 .002 .841 14.964 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 3.00 2.00 1.00 0.00 1.00 2.00 K_6_30 Linear Observed PAGE 275 259 APPENDIX E continued Time 07:00 Model Summary a .811 .658 .655 .887 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 140.985 1 140.985 179.232 .000 73.155 93 .787 214.140 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .026 .002 .811 13.388 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 3.00 2.00 1.00 0.00 1.00 2.00 K_7_00 Linear Observed PAGE 276 260 APPENDIX E continued Time 07:30 Model Summary a .423 .179 .171 1.659 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 55.928 1 55.928 20.324 .000 255.923 93 2.752 311.851 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .016 .004 .423 4.508 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 6.00 4.00 2.00 0.00 2.00 4.00 K_7_30 Linear Observed PAGE 277 261 APPENDIX E continued Time 08:00 Model Summary a .372 .138 .129 2.759 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 113.674 1 113.674 14.928 .000 708.169 93 7.615 821.844 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .023 .006 .372 3.864 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 10.00 7.50 5.00 2.50 0.00 2.50 K_8_00 Linear Observed PAGE 278 262 APPEN DIX E continued Time 08:30 Model Summary a .306 .094 .084 3.394 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 110.570 1 110.570 9.598 .003 1071.407 93 11.521 1181.977 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .023 .007 .306 3.098 .003 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 15.00 10.00 5.00 0.00 5.00 K_8_30 Linear Observed PAGE 279 263 APPENDIX E continued Time 09:00 Model Summary a .404 .163 .154 3.729 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 252.047 1 252.047 18.122 .000 1293.485 93 13.908 1545.531 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .035 .008 .404 4.257 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 15.00 10.00 5.00 0.00 5.00 K_9_00 Linear Observed PAGE 280 264 APPENDIX E continued Time 09:30 Model Summary a .569 .324 .316 4.109 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 751.730 1 751.730 44.514 .000 1570.549 93 16.888 2322.280 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .060 .009 .569 6.672 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 15.00 10.00 5.00 0.00 5.00 K_9_30 Linear Observed PAGE 281 265 APPENDIX E continued Time 10:00 Model Summary a .642 .412 .405 4.415 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 1267.797 1 1267.797 65.037 .000 1812.892 93 19.493 3080.689 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .078 .010 .642 8.065 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 20.00 15.00 10.00 5.00 0.00 5.00 K_10_00 Linear Observed PAGE 282 266 APPENDIX E continued Time 10:30 Model Summary a .695 .483 .477 4.085 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 1449.350 1 1449.350 86.844 .000 1552.088 93 16.689 3001.438 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .084 .009 .695 9.319 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 20.00 15.00 10.00 5.00 0.00 5.00 K_10_30 Linear Observed PAGE 283 267 APPENDIX E continued Time 11:00 Model Summary a .611 .373 .366 3.679 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 748.618 1 748.618 55.295 .000 1259.091 93 13.539 2007.709 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .060 .008 .611 7.436 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 15.00 10.00 5.00 0.00 5.00 K_11_00 Linear Observed PAGE 284 268 APPENDIX E continued Time 11:30 Model Summary a .504 .254 .246 3.405 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 367.170 1 367.170 31.670 .000 1078.218 93 11.594 1445.388 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .042 .007 .504 5.628 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 15.00 10.00 5.00 0.00 5.00 K_11_30 Linear Observed PAGE 285 269 APPENDIX E continued Time 12:00 Model Summary a .472 .223 .215 2.957 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 233.441 1 233.441 26.689 .000 813.431 93 8.747 1046.872 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .034 .007 .472 5.166 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 15.00 10.00 5.00 0.00 5.00 K_12_00 Linear Observed PAGE 286 270 APPENDIX E continued Time 12:30 Model Summary a .539 .290 .282 2.705 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 277.907 1 277.907 37.995 .000 680.233 93 7.314 958.140 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .037 .006 .539 6.164 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 10.00 5.00 0.00 K_12_30 Linear Observed PAGE 287 271 APPENDIX E continued Time 13:00 Model Summary a .651 .424 .418 2.298 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 361.693 1 361.693 68.462 .000 491.328 93 5.283 853.021 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .042 .005 .651 8.274 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 10.00 8.00 6.00 4.00 2.00 0.00 2.00 K_13_00 Linear Observed PAGE 288 272 APPENDIX E continued Time 13:30 Model Summary a .608 .370 .363 1.686 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 154.963 1 154.963 54.547 .000 264.206 93 2.841 419.170 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .027 .004 .608 7.386 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 7.50 5.00 2.50 0.00 2.50 K_13_30 Linear Observed PAGE 289 273 APPENDIX E continued Time 14:00 Model Summary a .638 .407 .400 1.499 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 143.246 1 143.246 63.762 .000 208.932 93 2.247 352.178 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .026 .003 .638 7.985 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 6.00 4.00 2.00 0.00 2.00 4.00 K_14_00 Linear Observed PAGE 290 274 APPENDIX E co ntinued Time 14:30 Model Summary a .577 .333 .326 1.479 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 101.502 1 101.502 46.422 .000 203.342 93 2.186 304.844 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .022 .003 .577 6.813 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 6.00 4.00 2.00 0.00 2.00 4.00 K_14_30 Linear Observed PAGE 291 275 APPENDIX E continued Time 15:00 Model Summary a .652 .426 .419 1.650 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 187.758 1 187.758 68.925 .000 253.340 93 2.724 441.098 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .030 .004 .652 8.302 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 7.50 5.00 2.50 0.00 2.50 K_15_00 Linear Observed PAGE 292 276 APPENDIX E continued Time 15:30 Model Summary a .533 .284 .276 1.718 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 108.676 1 108.676 36.810 .000 274.570 93 2.952 383.246 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .023 .004 .533 6.067 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 7.50 5.00 2.50 0.00 2.50 K_15_30 Linear Observed PAGE 293 277 APPENDIX E continued Time 16:00 Model Summary a .400 .160 .151 1.644 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 47.827 1 47.827 17.688 .000 251.471 93 2.704 299.298 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .015 .004 .400 4.206 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 6.00 4.00 2.00 0.00 2.00 4.00 K_16_00 Linear Observed PAGE 294 278 APPENDIX E continued Time 16:30 Model Summary a .330 .109 .100 1.506 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 25.833 1 25.833 11.389 .001 210.948 93 2.268 236.782 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .011 .003 .330 3.375 .001 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 2.00 0.00 2.00 4.00 K_16_30 Linear Observed PAGE 295 279 APPENDIX E continued Time 17:00 Model Summary a .348 .121 .112 1.632 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 34.179 1 34.179 12.828 .001 247.799 93 2.665 281.978 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .013 .004 .348 3.582 .001 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 6.00 4.00 2.00 0.00 2.00 4.00 K_17_00 Linear Observed PAGE 296 280 APPENDIX E contin ued Time 17:30 Model Summary a .170 .029 .019 1.790 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 8.901 1 8.901 2.778 .099 297.948 93 3.204 306.849 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .007 .004 .170 1.667 .099 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 2.00 0.00 2.00 4.00 K_17_30 Linear Observed PAGE 297 281 APPENDIX E continued Time 18:00 Model Summary a .237 .056 .046 1.784 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 17.604 1 17.604 5.529 .021 296.128 93 3.184 313.732 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .009 .004 .237 2.351 .021 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 2.00 0.00 2.00 4.00 K_18_00 Linear Observed PAGE 298 282 APPENDIX E continued Time 18:30 Model Summary a .414 .171 .162 1.612 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 49.941 1 49.941 19.227 .000 241.562 93 2.597 291.503 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .016 .004 .414 4.385 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 6.00 4.00 2.00 0.00 2.00 4.00 K_18_30 Linear Observed PAGE 299 283 APPENDIX E continued Time 19;00 Model Summary a .231 .053 .043 1.588 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 13.211 1 13.211 5.237 .024 234.591 93 2.522 247.802 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .008 .003 .231 2.289 .024 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 6.00 4.00 2.00 0.00 2.00 4.00 K_19_00 Linear Observed PAGE 300 284 APPENDIX E continued Time 1930 Model Summary a .466 .218 .209 1.187 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 36.433 1 36.433 25.852 .000 131.062 93 1.409 167.495 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .013 .003 .466 5.085 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 2.00 0.00 2.00 K_19_30 Linear Observed PAGE 301 285 APPENDIX E continued Time 20:00 Model Summary a .778 .606 .602 .922 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 121.508 1 121.508 142.988 .000 79.029 93 .850 200.537 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .024 .002 .778 11.958 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 5.00 4.00 3.00 2.00 1.00 0.00 1.00 K_20_00 Linear Observed PAGE 302 286 APPENDIX E continued T ime 20:30 Model Summary a .820 .673 .669 .823 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 129.434 1 129.434 191.245 .000 62.942 93 .677 192.376 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .025 .002 .820 13.829 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 5.00 4.00 3.00 2.00 1.00 0.00 1.00 K_20_30 Linear Observed PAGE 303 287 APPENDIX E continued Time 21:00 Model Summary a .853 .727 .724 .826 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 168.723 1 168.723 247.453 .000 63.411 93 .682 232.134 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .029 .002 .853 15.731 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 2.00 0.00 2.00 K_21_00 Linear Observed PAGE 304 288 APPENDIX E continued Time 21:30 Model Summary a .855 .730 .727 .846 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 180.212 1 180.212 251.759 .000 66.571 93 .716 246.783 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .030 .002 .855 15.867 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 2.00 0.00 2.00 K_21_30 Linear Observed PAGE 305 289 APPENDIX E continued Time 22:00 Model Summary a .839 .704 .701 .847 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 159.049 1 159.049 221.681 .000 66.725 93 .717 225.773 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .028 .002 .839 14.889 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 2.00 0.00 2.00 K_22_00 Linear Observed PAGE 306 290 APPENDIX E continued Time 22:30 Model Summary a .804 .647 .643 .844 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 121.291 1 121.291 170.412 .000 66.193 93 .712 187.483 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .024 .002 .804 13.054 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 2.00 0.00 2.00 K_22_30 Linear Observed PAGE 307 291 APPENDIX E continued Time 23:00 Model Summary a .843 .710 .707 .836 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 159.382 1 159.382 227.957 .000 65.024 93 .699 224.406 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .028 .002 .843 15.098 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 2.00 0.00 2.00 K_23_00 Linear Observed PAGE 308 292 APPENDIX E continued Time 23:30 Model Summary a .857 .735 .732 .842 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 182.911 1 182.911 258.083 .000 65.912 93 .709 248.823 94 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. PAGE 309 293 APPENDIX F 2008 PERIOD OLS REGRESSION RESULTS ZERO CROSSING PAGE 310 294 APPENDIX F continued Time 00: 00 Model Summary a .880 .775 .766 1.282 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 141.146 1 141.146 85.912 .000 41.073 25 1.643 182.219 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .044 .005 .880 9.269 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 6.00 4.00 2.00 0.00 2.00 K_0_00 Linear Observed PAGE 311 295 APPENDIX F continued Time 00: 30 Model Summary a .884 .781 .772 1.262 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 142.130 1 142.130 89.206 .000 39.832 25 1.593 181.962 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .044 .005 .884 9.445 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 6.00 4.00 2.00 0.00 2.00 K_0_30 Linear Observed PAGE 312 296 APPENDIX F co ntinue d Time 01: 00 Model Summary a .875 .765 .756 1.278 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 132.935 1 132.935 81.405 .000 40.825 25 1.633 173.760 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .043 .005 .875 9.022 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 6.00 4.00 2.00 0.00 2.00 K_1_00 Linear Observed PAGE 313 297 APPENDIX F continued Time 01: 30 Model Summary a .878 .770 .761 1.274 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 136.267 1 136.267 83.892 .000 40.608 25 1.624 176.875 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .043 .005 .878 9.159 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 6.00 4.00 2.00 0.00 2.00 K_1_30 Linear Observed PAGE 314 298 APPENDIX F continued Time 02: 00 ANOVA a 134.164 1 134.164 84.916 .000 39.499 25 1.580 173.664 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .043 .005 .879 9.215 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 6.00 4.00 2.00 0.00 2.00 K_2_00 Linear Observed PAGE 315 299 APPENDIX F continued Time 02: 30 Model Summary a .876 .767 .757 1.230 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 124.228 1 124.228 82.107 .000 37.825 25 1.513 162.052 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .042 .005 .876 9.061 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 6.00 4.00 2.00 0.00 2.00 K_2_30 Linear Observed PAGE 316 300 APPENDIX F continued Time 03: 00 Model Summary a .874 .763 .754 1.242 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 124.363 1 124.363 80.636 .000 38.557 25 1.542 162.920 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .042 .005 .874 8.980 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 6.00 4.00 2.00 0.00 2.00 K_3_00 Linear Observed PAGE 317 301 APPENDIX F continued Time 03: 30 Model Summary a .863 .745 .735 1.257 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 115.358 1 115.358 72.986 .000 39.513 25 1.581 154.871 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .040 .005 .863 8.543 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 6.00 4.00 2.00 0.00 2.00 K_3_30 Linear Observed PAGE 318 302 APPENDIX F continued Time 04: 00 Model Summary a .851 .724 .713 1.262 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 104.341 1 104.341 65.512 .000 39.817 25 1.593 144.159 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .038 .005 .851 8.094 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 6.00 4.00 2.00 0.00 2.00 K_4_00 Linear Observed PAGE 319 303 APPENDIX F continued Time 04: 30 Model Summary a .852 .726 .715 1.243 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 102.329 1 102.329 66.264 .000 38.607 25 1.544 140.935 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .038 .005 .852 8.140 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 6.00 4.00 2.00 0.00 2.00 K_4_30 Linear Observed PAGE 320 304 APPENDIX F continued Tim e 05: 00 Model Summary a .845 .713 .702 1.264 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 99.386 1 99.386 62.214 .000 39.938 25 1.598 139.324 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .037 .005 .845 7.888 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 6.00 4.00 2.00 0.00 2.00 K_5_00 Linear Observed PAGE 321 305 APPENDIX F continued Time 05: 30 Model Summary a .842 .709 .697 1.200 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 87.700 1 87.700 60.884 .000 36.011 25 1.440 123.711 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .035 .004 .842 7.803 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 6.00 4.00 2.00 0.00 2.00 K_5_30 Linear Observed PAGE 322 306 APPENDIX F continued Time 06: 00 Model Summary a .622 .387 .362 1.480 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 34.496 1 34.496 15.754 .001 54.741 25 2.190 89.236 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .022 .006 .622 3.969 .001 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 4.00 2.00 0.00 2.00 K_6_00 Linear Observed PAGE 323 307 APPENDIX F continued Time 06: 30 Model Summary a .621 .386 .361 2.232 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 78.179 1 78.179 15.687 .001 124.589 25 4.984 202.768 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .033 .008 .621 3.961 .001 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 7.50 5.00 2.50 0.00 2.50 K_6_30 Linear Observed PAGE 324 308 A PPENDIX F continued Time 07: 00 Model Summary a .584 .341 .314 2.991 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 115.595 1 115.595 12.919 .001 223.688 25 8.948 339.283 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .040 .011 .584 3.594 .001 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 7.50 5.00 2.50 0.00 2.50 K_7_00 Linear Observed PAGE 325 309 APPENDIX F continued Time 07: 30 Model Summary a .671 .451 .429 3.209 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 211.304 1 211.304 20.515 .000 257.500 25 10.300 468.804 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .054 .012 .671 4.529 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 10.00 7.50 5.00 2.50 0.00 2.50 K_7_30 Linear Observed PAGE 326 310 APPENDIX F continued Time 08: 00 Model Summary a .685 .470 .448 3.490 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 269.762 1 269.762 22.144 .000 304.554 25 12.182 574.316 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .061 .013 .685 4.706 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 10.00 7.50 5.00 2.50 0.00 2.50 K_8_00 Linear Observed PAGE 327 311 APPENDIX F continued Time 08: 30 Model Summary a .747 .558 .540 3.413 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 367.131 1 367.131 31.526 .000 291.129 25 11.645 658.260 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .071 .013 .747 5.615 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 10.00 7.50 5.00 2.50 0.00 2.50 K_8_30 Linear Observed PAGE 328 312 APPENDIX F continued Time 09: 00 Model Summary a .780 .609 .593 3.108 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 376.065 1 376.065 38.928 .000 241.515 25 9.661 617.580 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .072 .012 .780 6.239 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 10.00 7.50 5.00 2.50 0.00 K_9_00 Linear Observed PAGE 329 313 APPENDIX F continued Time 09: 30 Model Summary a .702 .493 .472 2.636 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 168.600 1 168.600 24.273 .000 173.649 25 6.946 342.249 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .048 .010 .702 4.927 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 7.50 5.00 2.50 0.00 2.50 K_9_30 Linear Observed PAGE 330 314 APPENDIX F continued Time 10: 00 Model Summary a .509 .259 .230 2.440 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 52.082 1 52.082 8.746 .007 148.882 25 5.955 200.964 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .027 .009 .509 2.957 .007 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 7.50 5.00 2.50 0.00 2.50 K_10_00 Linear Observed PAGE 331 315 APPENDIX F continued Time 10: 30 Model Summary a .201 .040 .002 2.095 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 4.613 1 4.613 1.051 .315 109.725 25 4.389 114.338 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .008 .008 .201 1.025 .315 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 6.00 4.00 2.00 0.00 2.00 4.00 K_10_30 Linear Observed PAGE 332 316 APPENDIX F continued Time 11: 00 Model Summary a .090 .008 .032 1.751 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a .624 1 .624 .204 .656 76.660 25 3.066 77.284 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .003 .007 .090 .451 .656 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 2.00 0.00 2.00 4.00 K_11_00 Linear Observed PAGE 333 317 APPENDIX F continued Time 11: 30 Model Summary a .546 .298 .269 1.387 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 20.390 1 20.390 10.592 .003 48.127 25 1.925 68.517 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .017 .005 .546 3.254 .003 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 4.00 3.00 2.00 1.00 0.00 1.00 2.00 3.00 K_11_30 Linear Observed PAGE 334 318 APPENDIX F continued Time 12: 00 Model Summary a .610 .372 .347 1.407 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 29.326 1 29.326 14.818 .001 49.476 25 1.979 78.801 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .020 .005 .610 3.849 .001 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 4.00 2.00 0.00 2.00 K_12_00 Linear Observed PAGE 335 319 APPENDIX F continued Time 12: 30 Model Summary a .757 .573 .556 1.110 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 41.343 1 41.343 33.576 .000 30.783 25 1.231 72.125 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .024 .004 .757 5.795 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 5.00 4.00 3.00 2.00 1.00 0.00 1.00 K_12_30 Linear Observed PAGE 336 320 APPENDIX F continued Time 13: 00 Model Summary a .746 .556 .538 1.268 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 50.327 1 50.327 31.281 .000 40.222 25 1.609 90.549 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .026 .005 .746 5.593 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 5.00 4.00 3.00 2.00 1.00 0.00 1.00 K_13_00 Linear Observed PAGE 337 321 APPENDIX F continued Time 13: 30 Model Summary a .710 .504 .484 1.462 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 54.329 1 54.329 25.419 .000 53.433 25 2.137 107.762 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .027 .005 .710 5.042 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 6.00 4.00 2.00 0.00 2.00 K_13_30 Linear Observed PAGE 338 322 APPENDIX F continued Time 14: 00 Model Summary a .613 .376 .351 1.641 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 40.590 1 40.590 15.072 .001 67.328 25 2.693 107.918 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .024 .006 .613 3.882 .001 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 6.00 4.00 2.00 0.00 2.00 K_14_00 Linear Observed PAGE 339 323 APPENDIX F continued Time 14: 30 Model Summary a .634 .402 .378 1.716 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 49.413 1 49.413 16.790 .000 73.575 25 2.943 122.987 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .026 .006 .634 4.098 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 6.00 4.00 2.00 0.00 2.00 K_14_30 Linear Observed PAGE 340 324 APPENDIX F continued Time 15: 00 Model Summary a .651 .423 .400 1.850 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 62.781 1 62.781 18.346 .000 85.552 25 3.422 148.334 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .030 .007 .651 4.283 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 6.00 4.00 2.00 0.00 2.00 K_15_00 Linear Observed PAGE 341 325 APPENDIX F continued Tim e 15: 30 Model Summary a .652 .425 .402 2.001 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 74.001 1 74.001 18.473 .000 100.149 25 4.006 174.151 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .032 .007 .652 4.298 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 6.00 4.00 2.00 0.00 2.00 K_15_30 Linear Observed PAGE 342 326 APPENDIX F continued Time 16: 00 Model Summary a .677 .458 .437 2.308 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 112.605 1 112.605 21.142 .000 133.151 25 5.326 245.756 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .040 .009 .677 4.598 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 8.00 6.00 4.00 2.00 0.00 2.00 K_16_00 Linear Observed PAGE 343 327 APPENDIX F continued Time 16: 30 Model Summary a .629 .396 .372 2.376 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 92.533 1 92.533 16.386 .000 141.179 25 5.647 233.712 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .036 .009 .629 4.048 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 7.50 5.00 2.50 0.00 2.50 K_16_30 Linear Observed PAGE 344 328 APPENDIX F continued Time 17 : 00 Model Summary a .644 .414 .391 2.519 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 112.189 1 112.189 17.675 .000 158.679 25 6.347 270.867 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .039 .009 .644 4.204 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 7.50 5.00 2.50 0.00 2.50 K_17_00 Linear Observed PAGE 345 329 APPENDIX F continued Time 17 : 30 Model Summary a .576 .332 .305 2.322 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 66.899 1 66.899 12.408 .002 134.785 25 5.391 201.683 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .030 .009 .576 3.523 .002 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 8.00 6.00 4.00 2.00 0.00 2.00 K_17_30 Linear Observed PAGE 346 330 APPENDIX F continued Time 18 : 00 Model Summary a .653 .426 .404 1.826 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 62.017 1 62.017 18.591 .000 83.397 25 3.336 145.415 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .029 .007 .653 4.312 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 4.00 2.00 0.00 2.00 K_18_00 Linear Observed PAGE 347 331 APPENDIX F continu ed Time 18 : 30 Model Summary a .813 .661 .648 1.105 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 59.592 1 59.592 48.834 .000 30.508 25 1.220 90.100 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .029 .004 .813 6.988 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 4.00 3.00 2.00 1.00 0.00 1.00 2.00 K_18_30 Linear Observed PAGE 348 332 APPENDIX F continued Time 19 : 00 Model Summary a .855 .730 .719 1.154 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 90.073 1 90.073 67.662 .000 33.281 25 1.331 123.353 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .035 .004 .855 8.226 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 4.00 3.00 2.00 1.00 0.00 1.00 2.00 K_19_00 Linear Observed PAGE 349 333 APPENDIX F continued Time 19 : 30 Model Summary a .858 .737 .726 1.252 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 109.683 1 109.683 70.009 .000 39.168 25 1.567 148.851 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .039 .005 .858 8.367 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 4.00 2.00 0.00 2.00 K_19_30 Linear Observed PAGE 350 334 APPENDIX F continued Time 20 : 00 Model Summary a .864 .747 .737 1.245 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 114.297 1 114.297 73.683 .000 38.780 25 1.551 153.077 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .040 .005 .864 8.584 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 4.00 2.00 0.00 2.00 K_20_00 Linear Observed PAGE 351 335 APPENDIX F continued Time 20 : 30 Model Summary a .873 .762 .753 1.265 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 128.215 1 128.215 80.127 .000 40.004 25 1.600 168.219 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .042 .005 .873 8.951 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 4.00 2.00 0.00 2.00 K_20_30 Linear Observed PAGE 352 336 APPENDIX F continued Time 21 : 00 Model Summary a .880 .774 .765 1.225 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 128.611 1 128.611 85.670 .000 37.531 25 1.501 166.142 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .042 .005 .880 9.256 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 4.00 2.00 0.00 2.00 K_21_00 Linear Observed PAGE 353 337 APPENDIX F conti nued Time 21 : 30 Model Summary a .881 .776 .767 1.191 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 122.455 1 122.455 86.370 .000 35.445 25 1.418 157.900 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .041 .004 .881 9.294 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 4.00 2.00 0.00 2.00 K_21_30 Linear Observed PAGE 354 338 APPENDIX F continued Time 22 : 00 Model Summary a .884 .781 .773 1.166 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 121.344 1 121.344 89.312 .000 33.966 25 1.359 155.310 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .041 .004 .884 9.451 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 6.00 4.00 2.00 0.00 2.00 K_22_00 Linear Observed PAGE 355 3 39 APPENDIX F continued Time 22 : 30 Model Summary a .885 .783 .774 1.166 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 122.825 1 122.825 90.289 .000 34.009 25 1.360 156.834 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .041 .004 .885 9.502 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 6.00 4.00 2.00 0.00 2.00 K_22_30 Linear Observed PAGE 356 340 APPENDIX F continued Time 23 : 00 Model Summary a .882 .779 .770 1.187 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 123.934 1 123.934 88.012 .000 35.204 25 1.408 159.137 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .041 .004 .882 9.381 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 6.00 4.00 2.00 0.00 2.00 K_23_00 Linear Observed PAGE 357 341 APPENDIX F continued Time 23 : 30 Model Summary a .885 .783 .775 1.218 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 133.977 1 133.977 90.373 .000 37.062 25 1.482 171.039 26 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .043 .005 .885 9.506 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 80.00 60.00 40.00 20.00 6.00 4.00 2.00 0.00 2.00 K_23_30 Linear Observed PAGE 358 342 APPENDIX G AGGREGATED 2007 AND 200 8 PERIOD OLS REGRESSION RESULTS WITHOUT CONSTANT PAGE 359 343 APPENDIX G continued Time 00 : 00 Model Summary a .869 .755 .753 .973 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 343.647 1 343.647 362.869 .000 111.749 118 .947 455.396 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .035 .002 .869 19.049 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 6.00 4.00 2.00 0.00 2.00 K_0_00 Linear Observed PAGE 360 344 APPENDIX G continued Time 00 : 30 Model Summary a .877 .769 .767 .966 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 366.293 1 366.293 392.456 .000 110.133 118 .933 476.426 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .036 .002 .877 19.811 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 6.00 4.00 2.00 0.00 2.00 K_0_30 Linear Observed PAGE 361 345 APPENDIX G continued Time 01 : 00 Model Summary a .876 .767 .765 .972 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 366.002 1 366.002 387.418 .000 111.477 118 .945 477.479 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .036 .002 .876 19.683 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 6.00 4.00 2.00 0.00 2.00 K_1_00 Linear Observed PAGE 362 346 APPENDIX G continued T ime 01 : 3 0 Model Summary a .877 .770 .768 .971 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 372.257 1 372.257 394.992 .000 111.208 118 .942 483.465 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .037 .002 .877 19.874 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 6.00 4.00 2.00 0.00 2.00 K_1_30 Linear Observed PAGE 363 347 APPENDIX G continued Time 02 : 00 Model Summary a .869 .754 .752 .977 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 346.023 1 346.023 362.294 .000 112.701 118 .955 458.724 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .035 .002 .869 19.034 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.0 0 80.0 0 60.0 0 40.0 0 20.0 0 0.00 6.0 0 4.0 0 2.0 0 0.0 0 2.00 K_2_00 Linea r Observed PAGE 364 348 APPENDIX G continued Time 02 : 30 Model Summary a .868 .754 .752 .957 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 330.587 1 330.587 360.910 .000 108.086 118 .916 438.673 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .035 .002 .868 18.998 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp : percent : 50m 100.00 80.00 60.00 40.00 20.00 0.00 6. 00 4.00 2.00 0.00 2.00 K_2_30 Linear Observed PAGE 365 349 APPENDIX G continued Time 03 : 00 Model Summary a .859 .738 .736 .978 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 318.941 1 318.941 333.132 .000 112.973 118 .957 431.914 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .034 .002 .859 18.252 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 6.00 4.00 2.00 0.00 2.00 K_3_00 Linear Observed PAGE 366 350 APPENDIX G continued Time 03 : 30 Model Summary a .856 .733 .731 .978 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 310.576 1 310.576 324.723 .000 112.859 118 .956 423.435 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .033 .002 .856 18.020 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.0 0 80.0 0 60.0 0 40.0 0 20.0 0 0.0 0 6.0 0 4.0 0 2.0 0 0.0 0 2.00 K : 3 : 30 Linea r Observe d PAGE 367 351 APPENDIX G continued Time 04 : 00 Model Summary a .856 .732 .730 .978 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 308.108 1 308.108 322.265 .000 112.816 118 .956 420.925 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .033 .002 .856 17.952 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.0 0 80.0 0 60.0 0 40.0 0 20.0 0 0.0 0 6.0 0 4.0 0 2.0 0 0.0 0 2 .00 K_4_00 Linea r Observed PAGE 368 352 APPENDIX G continued Time 04 : 30 Model Summary a .859 .739 .736 .970 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 313.589 1 313.589 333.341 .000 111.008 118 .941 424.597 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .034 .002 .859 18.258 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 6.00 4.00 2.00 0.00 2.00 K_4_30 Linear Observed PAGE 369 353 APPENDIX G continued Time 05 : 00 Model Summary a .852 .726 .724 .976 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 298.535 1 298.535 313.201 .000 112.474 118 .953 411.009 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .033 .002 .852 17.697 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp : percent : 50m 100.00 80.00 60.00 40.00 20.00 0.00 6.00 4.00 2.00 0.00 2.00 K : 5 : 00 Linear Observed PAGE 370 354 APPENDIX G continued Time 05 : 30 Model Summary a .849 .721 .719 .955 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 278.205 1 278.205 305.146 .000 107.582 118 .912 385.787 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .032 .002 .849 17.468 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp : percent : 50m 100.0 0 80.0 0 60.0 0 40.0 0 20.0 0 0.0 0 6.0 0 4.0 0 2.0 0 0.0 0 2.00 K_5_30 Linea r Obs erve d PAGE 371 355 APPENDIX G continued Time 06 : 00 Model Summary a .791 .625 .622 1.041 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 213.256 1 213.256 196.627 .000 127.980 118 1.085 341.235 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .028 .002 .791 14.022 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 2.00 0.00 2.00 K_6_00 Linear Observed PAGE 372 356 APPENDIX G continued Time 06 : 30 Model Summary a .751 .564 .561 1.280 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 250.584 1 250.584 152.835 .000 193.470 118 1.640 444.055 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .030 .002 .751 12.363 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 7.50 5.00 2.50 0.00 2.50 K_6_30 Linear Observed PAGE 373 357 APPENDIX G continued Time 07 : 00 Model Summary a .671 .450 .445 1.604 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 248.342 1 248.342 96.547 .000 303.523 118 2.572 551.865 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .030 .003 .671 9.826 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 7.50 5.00 2.50 0.00 2.50 K_7_00 Linear Observed PAGE 374 358 APPENDIX G continued Time 07 : 30 Model Summary a .502 .252 .246 2.213 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 195.044 1 195.044 39.843 .000 577.642 118 4.895 772.686 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .027 .004 .502 6.312 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 10.00 7.50 5.00 2.50 0.00 2.50 K_7_30 Linear Observed PAGE 375 359 APP ENDIX G continued Time 08 : 00 Model Summary a .475 .226 .219 3.015 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 312.831 1 312.831 34.404 .000 1072.945 118 9.093 1385.776 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .034 .006 .475 5.866 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.0 0 80.0 0 60.0 0 40.0 0 20.0 0 0.0 0 10.0 0 7.5 0 5.0 0 2.5 0 0.0 0 2.50 K_8_00 Linea r Observe d PAGE 376 360 APPENDIX G continued Time 08 : 30 Model Summary a .443 .196 .190 3.530 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 359.311 1 359.311 28.836 .000 1470.363 118 12.461 1829.675 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .036 .007 .443 5.370 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.0 0 80.0 0 60.0 0 40.0 0 20.0 0 0.0 0 15.0 0 10.0 0 5.0 0 0.0 0 5.00 K_8_30 Linea r Observed PAGE 377 361 APPENDIX G continued Time 09 : 00 Model Summary a .506 .256 .250 3.692 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 554.436 1 554.436 40.670 .000 1608.626 118 13.632 2163.062 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .045 .007 .506 6.377 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 15.00 10.00 5.00 0.00 5.00 K_9_00 Linear Observed PAGE 378 362 APPENDIX G continued Time 09 : 30 Model Summary a .585 .342 .337 3.822 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 897.595 1 897.595 61.450 .000 1723.613 118 14.607 2621.208 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .057 .007 .585 7.839 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.0 80.0 0 60.0 0 40.0 0 20.0 0 0.0 15.0 0 10.0 0 5.0 0 0.0 0 5.00 K_9_30 Linea r Observe PAGE 379 363 APPENDIX G continued Time 10 : 00 Model Summary a .600 .360 .355 4.174 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 1156.237 1 1156.237 66.364 .000 2055.877 118 17.423 3212.114 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .065 .008 .600 8.146 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50 m 100.00 80.00 60.00 40.00 20.00 0.00 20.00 15.00 10.00 5.00 0.00 5.00 K_10_00 Linear Observed PAGE 380 364 APPENDIX G continued Time 10 : 30 Model Summary a .607 .368 .363 4.051 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 1128.612 1 1128.612 68.787 .000 1936.062 118 16.407 3064.674 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .064 .008 .607 8.294 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 20.00 15.00 10.00 5.00 0.00 5.00 K_10_30 Linear Observed PAGE 381 365 APPENDIX G continued Time 11 : 00 Model Summary a .524 .274 .268 3.556 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 564.054 1 564.054 44.614 .000 1491.856 118 12.643 2055.910 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .045 .007 .524 6.679 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 15.00 10.00 5.00 0.00 5.00 K_11_0 0 Linear Observed PAGE 382 366 APPENDIX G continued Time 11 : 30 Model Summary a .482 .232 .226 3.116 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 346.447 1 346.447 35.685 .000 1145.601 118 9.708 1492.048 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .035 .006 .482 5.974 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 15.00 10.00 5.00 0.00 5.00 K_11_3 0 Linear Observed PAGE 383 367 APPENDIX G continued Time 12 : 00 Model Summary a .473 .224 .217 2.702 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 248.248 1 248.248 33.993 .000 861.733 118 7.303 1109.981 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .030 .005 .473 5.830 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 15.00 10.00 5.00 0.00 5.00 K_12_00 Linear Observed PAGE 384 368 APPENDIX G continued Time 12 : 30 Model Summary a .549 .301 .295 2.451 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 305.210 1 305.210 50.791 .000 709.079 118 6.009 1014.289 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .033 .005 .549 7.127 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 10. 00 5.00 0.00 K_12_30 Linear Observed PAGE 385 369 APPENDIX G continued Time 13 : 00 Model Summary a .650 .423 .418 2.137 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 394.633 1 394.633 86.431 .000 538.776 118 4.566 933.409 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .038 .004 .650 9.297 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 10.00 8.00 6.00 4.00 2.00 0.00 2.00 K_13_00 Linear Observed PAGE 386 370 APPENDIX G continued Time 13 : 30 Model Summary a .630 .396 .391 1.640 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 208.523 1 208.523 77.491 .000 317.531 118 2.691 526.053 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .027 .003 .630 8.803 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 7.50 5.00 2.50 0.00 2.50 K_13_30 Linear Observed PAGE 387 371 APPENDIX G contin ued Time 14 : 00 Model Summary a .632 .399 .394 1.531 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 183.654 1 183.654 78.397 .000 276.428 118 2.343 460.081 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .026 .003 .632 8.854 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.0 0 80.0 0 60.0 0 40.0 0 20.0 0 0.0 0 6.0 0 4.0 0 2.0 0 0.0 0 2.00 4.00 K_14_00 Linea r Observed PAGE 388 372 APPENDIX G continued Time 14 : 30 Model Summary a .593 .352 .346 1.533 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 150.411 1 150.411 63.979 .000 277.413 118 2.351 427.825 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .023 .003 .593 7.999 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 6.00 4.00 2.00 0.00 2.00 4.00 K_14_30 Linear Observed PAGE 389 373 APPENDIX G continued Time 15 : 00 Model Summary a .653 .426 .421 1.693 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 251.073 1 251.073 87.562 .000 338.351 118 2.867 589.424 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .030 .003 .653 9.357 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 7.50 5.00 2.50 0.00 2.50 K_15_00 Linear Observed PAGE 390 374 APPENDIX G continued Time 15 : 30 ANOVA a 179.377 1 179.377 56.067 .000 377.519 118 3.199 556.895 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .025 .003 .568 7.488 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Model Summary a .568 .322 .316 1.789 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 7.50 5.00 2.50 0.00 2.50 K_15_30 Linear Observed PAGE 391 375 APPENDIX G continued Time 16 : 00 Model Summary a .488 .238 .232 1.875 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 129.841 1 129.841 36.949 .000 414.654 118 3.514 544.495 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .022 .004 .488 6.079 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 1 00.0 0 80.0 0 60.0 0 40.0 0 20.0 0 0.0 0 7.5 0 5.0 0 2.5 0 0.0 0 2.50 K_16_00 Linear Observed PAGE 392 376 APPENDIX G continued Time 16 : 30 Model Summary a .429 .184 .178 1.802 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 86.675 1 86.675 26.686 .000 383.264 118 3.248 469.939 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .018 .003 .429 5.166 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 7.50 5.00 2.50 0.00 2.50 K_16_30 Linear Observed PAGE 393 377 APPENDIX G continued Time 17 : 00 Model Summary a .446 .199 .192 1.936 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 109.641 1 109.641 29.248 .000 442.342 118 3.749 551.983 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .020 .004 .446 5.408 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.0 0 80.0 0 60.0 0 40.0 0 20.0 0 0.0 0 7.5 0 5.0 0 2.5 0 0.0 0 2.50 K_17_0 Linea r Observe d PAGE 394 378 APPENDIX G continued Time 17 : 30 Model Summary a .306 .093 .086 1.963 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 46.875 1 46.875 12.162 .001 454.798 118 3.854 501.673 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .013 .004 .306 3.487 .001 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.0 0 80.0 0 60.0 0 4 0.0 0 20.0 0 0.0 0 7.5 0 5.0 0 2.5 0 0.0 0 2.50 K_17_3 Linea r Observe d PAGE 395 379 APPENDIX G continued Time 18 : 00 Model Summary a .363 .131 .124 1.827 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 59.660 1 59.660 17.866 .000 394.049 118 3.339 453.709 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .015 .003 .363 4.227 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 6.00 4.00 2.00 0.00 2.00 4.00 K_18_00 Linear Observed PAGE 396 380 APPENDIX G continued Time 18 : 30 Model Summary a .520 .270 .264 1.528 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 102.090 1 102.090 43.714 .000 275.579 118 2.335 377.669 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .019 .003 .520 6.612 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 6.00 4.00 2.00 0.00 2.00 4.00 K_18_30 Linear Observed PAGE 397 381 APPENDIX G continued Time 19 : 00 Model Summary a .421 .177 .170 1.599 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 64.865 1 64.865 25.356 .000 301.866 118 2.558 366.731 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .015 .003 .421 5.035 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.0 0 80.0 0 60.0 0 40.0 0 20.0 0 0.0 0 6.0 0 4.0 0 2.0 0 0.0 0 2.00 4.00 K_19_0 Linea r Observe d PAGE 398 382 APPENDIX G continued Time 19 : 30 Model Summary a .599 .359 .354 1.305 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 112.523 1 112.523 66.120 .000 200.812 118 1.702 313.335 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .020 .002 .599 8.131 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 2.00 0.00 2.00 K_19_30 Linear Observed PAGE 399 383 APPENDIX G continued Time 20 : 00 Model Summary a .797 .635 .632 1.044 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 224.268 1 224.268 205.571 .000 128.733 118 1.091 353.001 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .028 .002 .797 14.338 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 2.00 0.00 2.00 K_20_00 Linear Observed PAGE 400 384 APPENDIX G continued Time 20 : 30 Model Summary a .822 .675 .672 .996 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 243.224 1 243.224 245.337 .000 116.984 118 .991 360.208 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .030 .002 .822 15.663 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.0 0 80.0 0 60.0 0 40.0 0 20.0 0 0.0 0 4.0 0 2.0 0 0.0 0 2.00 K_20_30 Linea r Observe d PAGE 401 385 APPENDIX G continued Time 21 : 00 Model Summary a .852 .726 .724 .961 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 288.883 1 288.883 312.958 .000 108.923 118 .923 397.806 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .032 .002 .852 17.691 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 4.00 2.00 0.00 2.00 K_21_00 Linear Observed PAGE 402 386 APPENDIX G continued Time 21 : 30 Model Summary a .858 .736 .734 .950 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 297.296 1 297.296 329.419 .000 106.493 118 .902 403.789 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .033 .002 .858 18.150 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.0 0 80.0 0 60.0 0 40.0 0 20.0 0 0.0 0 4.0 0 2.0 0 0.0 0 2.00 K_21_30 Linea r Observe d PAGE 403 387 APPEN DIX G continued Time 22 : 00 Model Summary a .847 .718 .716 .953 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 272.872 1 272.872 300.361 .000 107.200 118 .908 380.072 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .031 .002 .847 17.331 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.0 0 80.0 0 60.0 0 40.0 0 20.0 0 0.0 0 6.0 0 4.0 0 2.0 0 0.0 0 2.00 K_22_00 Linea r O bserve d PAGE 404 388 APPENDIX G continued Time 22 : 30 Model Summary a .821 .674 .671 .972 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 230.801 1 230.801 244.079 .000 111.581 118 .946 342.382 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .029 .002 .821 15.623 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.00 80.00 60.00 40.00 20.00 0.00 6.00 4.00 2.00 0.00 2.00 K_22_30 Linear Observed PAGE 405 389 APPENDIX G continued Time 23 : 00 Model Summary a .849 .721 .719 .950 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 275.532 1 275.532 305.282 .000 106.501 118 .903 382.033 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .031 .002 .849 17.472 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. Imp_percent_50m 100.0 0 80.0 0 60.0 0 40.0 0 20.0 0 0.0 0 6.0 0 4.0 0 2.0 0 0.0 0 2.00 K_23_00 Linea r Observe d PAGE 406 390 APPENDIX G continued Time 23 : 30 Model Summary a .860 .740 .738 .960 R R Square Adjusted R Square Std. Error of the Estimate The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. ANOVA a 309.734 1 309.734 336.413 .000 108.642 118 .921 418.376 119 Regression Residual Total Sum of Squares df Mean Square F Sig. The independent variable is Imp_percent_50m. The equation was estimated without the constant term. a. Coefficients .033 .002 .860 18.342 .000 Imp_percent_50m B Std. Error Unstandardized Coefficients Beta Standardized Coefficients t Sig. PAGE 407 391 About The Author JoAnn Sullivan was born and raised in Omaha, Nebraska. After completing high school JoAnn attended the University of Nebraska Lincoln for two years. Needing a break from studies, JoAnn enlisted in the United States Army and from 1972 u ntil 1980 served as a nuclear, chemical and biological warfare specialist stationed in West Germany. After leaving the Army she returned to school and in 1982 she received a BS degree in Electronic Engineering from the University of Nebraska Omaha. JoAnn spent the next 20 some years in various engineering assignments throughout the United States and Europe. A job layoff was the impetus for JoAnn to return to school and in 2006 she received as MA in Geography from the University of Nebraska Omaha. JoAnn en rolled in the PhD program at the University of South Florida in 2006 and completed a 1 year visiting instructor position at the University of Central Arkansas in 2009. JoAnn continues to be interested in urban climatology and finding alternate uses for co mmercial electronic technology in geography. 