USF Libraries
USF Digital Collections

Ocean-atmosphere interactions on the West Florida shelf

MISSING IMAGE

Material Information

Title:
Ocean-atmosphere interactions on the West Florida shelf
Physical Description:
Book
Language:
English
Creator:
Virmani, Jyotika I
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla.
Publication Date:

Subjects

Subjects / Keywords:
Coastal ocean observations
Surface fluxes
Relative humidity
Climatologies
One-dimensional temperature balance
Dissertations, Academic -- Marine Science -- Doctoral -- USF   ( lcsh )
Genre:
government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: Ocean-atmosphere fluxes on the West Florida Shelf (WFS) coastal ocean region are investigated using observations and derived surface fluxes from an array of buoys deployed between 1998 and 2003. The observed annual cycle shows that water column temperatures increase and are stratified when heat flux is positive, and they decrease and are well mixed when it is negative. Water temperature is minimum (maximum) when heat flux switches sign from negative (positive) to positive (negative) in early spring (autumn). Tropical and extra-tropical events help define the seasonal characteristics of the water temperature. Despite considerable daily and synoptic variability in relative humidity, observations on the WFS show that the monthly mean values are nearly constant at about 75%. Winter relative humidity varies from less than 50% to over 100% (supersaturation values of up to 3% are recorded and coincide with fog on shore) as extra-tropical fronts move over the WFS.Sensor distribution shows small spatial variations in relative humidity in the coastal ocean environment that depends on high frequency variability in meteorological conditions and low-frequency variability in oceanic conditions. Comparisons with observations show that standard climatologies are unable to reproduce spatial variability on the WFS, especially in relative humidity and surface heat flux components that are dependent on sea surface temperature. Model experiments show that careful attention must be paid in calculating and applying surface heat fluxes. Observations and models are employed to assess the relative importance of surface fluxes and convergence of heat flux by the ocean circulation in controlling ocean temperature.
Thesis:
Thesis (Ph.D.)--University of South Florida, 2005.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Jyotika I. Virmani.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 242 pages.
General Note:
Includes vita.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001680979
oclc - 62493989
usfldc doi - E14-SFE0001141
usfldc handle - e14.1141
System ID:
SFS0025462:00001


This item is only available as the following downloads:


Full Text

PAGE 1

Ocean-Atmosphere Interactions on the West Florida Shelf by Jyotika I. Virmani A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy College of Marine Science University of South Florida Major Professor: Robert H. Weisberg, Ph. D. Gary T. Mitchum, Ph. D. Mark A. Luther, Ph. D. Duane E. Waliser, Ph. D. Robert. A. Weller, Ph. D. Date of Approval: April 14, 2005 Keywords: coastal ocean observations, surface fluxes, relative humidity, climatologies, one-dimensional temperature balance Copyright 2005, Jyotika I. Virmani

PAGE 2

Dedication To my parents, brother, and extended family for their love and support and to all those who made this journey a fun experience.

PAGE 3

Acknowledgements I thank my advisor, Dr. Robert Weisbe rg, for his scientif ic mentorship and support. He has provided me with opportunitie s to further my knowledge and confidence as a scientist. I thank my committee and anonymous reviewers of my papers for their comments and suggestions, which helped to make this a more robust piece of work. I thank the faculty of the College of Mari ne Science for their confidence in me. In particular I thank our Dean, Dr Peter Betzer for his support and for his role in obtaining endowed fellowships for the students, and Dr Edward Van Vleet, the graduate program director, for helping me with graduate school and immigration issues. I also thank Drs. Carol Steele and Robert Byrne for their advi ce. To the College of Marine Science and International Student and Scholar Services staff I owe many thanks for assisting me with assorted university and immigr ation paperwork. I owe a huge debt of gratitude to our computer manager, Jeff Donovan for doing his best to maintain the comput ers. Id also like to thank Patrick Smith for his computer support. The data could not have been collect ed without Rick Cole, ably assisted by Jay Law. I also thank the other support staff of the COMPS array: Cliff Merz, Vembu Subramanian, Jeff Scudder, Randy Russell and a ll those who have helped on cruises. I also thank the crews of the R/V Suncoast er and Bellows and st aff at FIO and Doug Myhre and Joe Vanderbloemen who overs ee the College computer networks. My friends and colleagues in the Ocea n Circulation Group have been wonderful people to work with and have provided many useful comments for this study. In particular, Id like to thank my friends a nd colleagues Elizabeth Singh, and Drs. Robert Helber and Christina Holland for their frie ndship, advice and proof-r eading. I also thank Dr. Ruoying He for his assistance with the POM 3D model. Financially, I have been supported by the Getting Fellowship, the Garrels Fellowship, the University Graduate Fellows hip and the Knight Fellowship. My thanks goes to the friends and families of Paul L. Getting, Robert M. Garrels, and Elsie and William Knight for the endowed fellowships. Th is work was supported by the Office of Naval Research (Grants N00014-98-10158 and N00014-02-0972) and the National Oceanic and Atmospheric Admini stration (Grant NA76RG0463).

PAGE 4

Note to Reader The maps in section 2.4 and the Appendix are in color and may be obtained as such from the University of South Florida. This dissert ation is available in electronic format on-line via the University of South Flor ida library at http://www.lib.usf.edu.

PAGE 5

i Table of Contents List of Tables iii List of Figures iv List of Acronyms xiii Abstract xv 1. Introduction 1 1.1 Introduction 1 1.2 Background on the Gulf of Mexico and the West Florida Shelf 2 1.3 Data collection and QA/QC process 7 1.4 Calculation of surface fluxes 11 1.4.1 Bulk parameterization 11 1.4.2 Net heat flux 13 1.5 Dissertation organization 17 2. Climatological aspects of the West Florida Shelf 19 2.1 Abstract 19 2.2 Introduction 20 2.3 Details of available climatologies 20 2.4 Comparison of standard climatol ogies over the Gulf of Mexico 24 2.5 West Florida Shelf observed climatologies 28 2.6 Comparison between standard and in situ climatologies 31 2.7 Discussion 55 3. Relative humidity over the West Florida Continental Shelf 62 3.1 Abstract 62 3.2 Introduction 63 3.3 Data 65 3.4 Annual cycle 68 3.5 Weatherpak and IMET/ASIMET offset 74 3.5.1 Instrument differences 74 3.5.2 Sensor height differences 76 3.5.3 Air-Sea regime differences 77 3.6 Winter relative humidity and supersaturation 79 3.7 Conclusions 86

PAGE 6

ii 4. Features of the observed annual ocean-atmosphere flux variability on the West Florida Shelf 89 4.1 Abstract 89 4.2 Introduction 90 4.3 Background 91 4.4 Observations 95 4.5 Description and discussion 102 4.6 Conclusions 110 5. The relative importance of surface heat flux and convergence of heat flux by ocean circulation in cont rolling ocean temperature 113 5.1 Abstract 113 5.2 Introduction 114 5.3 Data for Model Forcing, Init ialization, and Verification 115 5.4 Surface Flux Modifications 118 5.4.1 Cool-Skin, Warm-Layer effects 118 5.4.2 Rain sensible heat flux 119 5.4.3 Moisture flux 120 5.4.4 Waves 122 5.5 Models 122 5.5.1 The Price, Weller and Pinkel one-dimensional mixed layer model 122 5.5.2 The one-dimensional Princeton Ocean Model 124 5.5.3 The three-dimensional Princeton Ocean Model 124 5.6 Results 126 5.6.1 The Price, Weller and Pinkel one-dimensional mixed layer model 126 5.6.2 One-dimensional temperature balance 141 5.6.3 The one-dimensional Princeton Ocean Model 143 5.6.4 The three-dimensional Princeton Ocean Model 147 5.7 Discussion 153 6. Summary 157 References 161 Appendix 171 About the author End Page

PAGE 7

iii List of Tables Table 1. Sensors used to collect me teorological data on the WFS. 10 Table 2. Water Types and Corresponding Values of l1, l2, I1, and I2. 14 Table 3. Ten standard climatologies, th eir spatial coverage, base years, and grid spacing. 22 Table 4. Variables available for each climatology. 24 Table 5. Mean and standard deviation of relative humidity (%) from 1999-2003. 69 Table 6. Mean offset and correlation ma trix for relative humidity sensors. 76 Table 7. Monthly mean heat flux (W/m2) needed to produce the observed temperature (Qreq) obtained from basic heat flux calculation (Qbase), applying a cool skin, warm layer correction (Qcswl), and the addition of penetrative radiation (Qpen). 142

PAGE 8

iv List of Figures Figure 1. Gulf of Mexico SST monthly co mposite satellite images from AVHRR. 3 Figure 2. NCEP Sea Level Pressure Climatology (mb). 4 Figure 3. April-June mean wind field fo r 1998 (top) and 1999 (bottom) calculated from NCEP daily-averaged reanalysis. 5 Figure 4. OCG moorings on th e West Florida Shelf. 7 Figure 5. Schematic (left) and photograph (r ight) of a surface air-sea measurement buoy on the WFS. 8 Figure 6. 36 hour lowpass filtered data and calculated fluxes from NA2, on the 25m isobath approximately 50km from Tampa Bay. 12 Figure 7. Qpen values for an initial Qsw of 1000 Wm-2 over 50m water column depth for water types IA and IB. 15 Figure 8. Autospectra of observed me teorological data at NA2 from May-July 2000. 16 Figure 9. Latent Heat Flux (W/m2) Climatologies: January to June. 26 Figure 10. Anomalies of Latent Heat Flux (W/m2) Climatologies: January to June. 27 Figure 11. WFS observations of AT, SST, BP, RH, downward SW and LW averaged from 1998-2003. 29 Figure 12. WFS observations of wind sp eed and direction, and east and north components. 30 Figure 13. Mooring locations and grid points of seven cl imatologies on the WFS. 31 Figure 14. Observed and standard climat ologies of air temperature at NA2, CM2, CM3 and over all moorings (WFS). 33 Figure 15. Observed and standard climat ologies of sea surface temperature at NA2, CM2, CM3 and over all moorings (WFS). 34 Figure 16. Observed and standard climatol ogies of barometric pressure at NA2, CM2, CM3 and over all moorings (WFS). 35

PAGE 9

v Figure 17. Observed and standard climatol ogies of relative humidity at NA2, CM2, CM3 and over all moorings (WFS). 36 Figure 18. Observation derived and standa rd climatologies of net shortwave radiation at NA2, CM2, CM3 and over all moorings (WFS). 38 Figure 19. Observation derived and sta ndard climatologies of net longwave radiation at NA2, CM2, CM3 and over all moorings (WFS). 39 Figure 20. Observation derived and standard climatologies of latent heat flux at NA2, CM2, CM3 and over all moorings (WFS). 40 Figure 21. Observation derived and standard climatologies of sensible heat flux at NA2, CM2, CM3 and over all moorings (WFS). 41 Figure 22. Observation derived and standard climatologies of net heat flux at NA2, CM2, CM3 and over all moorings (WFS). 42 Figure 23. Observed and standard clim atologies of wind speed at NA2, CM2, CM3 and over all moorings (WFS). 44 Figure 24. Observed and standard climat ologies of east component of wind at NA2, CM2, CM3 and over all moorings (WFS). 45 Figure 25. Observed and standard climatol ogies of north component of wind at NA2, CM2, CM3 and over all moorings (WFS). 46 Figure 26. Observation derived and standa rd climatologies of east momentum flux at NA2, CM2, CM3 and ove r all moorings (WFS). 47 Figure 27. Observation derived and standa rd climatologies of north momentum flux at NA2, CM2, CM3 and ove r all moorings (WFS). 48 Figure 28. Comparison of WFS interpolated long-term NCEP climatology (dash) and 5-year NCEP climatology (solid ), calculated with monthly means from 1998-2003, and WFS in situ observati ons (solid with squares) for air temperature, sea surface temperature, relative humidity and sea level pressure. 49 Figure 29. Comparison of WFS interpolated long-term NCEP climatology (dash) and 5-year NCEP climatology (solid), ca lculated with monthly means from 1998-2003, and WFS in situ observations (solid with squares) for shortwave radiation, longwave radiation, sensible heat flux and latent heat flux. 50

PAGE 10

vi Figure 30. Comparison of WFS interpolated long-term NCEP climatology (dash) and 5-year NCEP climatology (solid ), calculated with monthly means from 1998-2003, and WFS in situ observat ions (solid with squares) for wind speed, east and north compone nt of wind, east and north momentum flux. 51 Figure 31. Box-and-whisker plots for the NCEP climatology and WFS observations (solid) for air temperature, sea surf ace temperature, relative humidity, and sea level pressure. 52 Figure 32. Box-and-whisker plots for the NCEP climatology and WFS observations (solid) for shortwave radi ation, longwave radiati on, sensible heat flux, and latent heat flux. 53 Figure 33. Box-and-whisker plots for the NCEP climatology and WFS observations (solid) for wind speed, east and north components of wind, east and north momentum flux. 54 Figure 34. WFS in situ climatology and mean and standard deviation of ensemble of standard climatologies for AT, SST, BP, and RH. 56 Figure 35. WFS in situ climatology and mean and standard deviation of ensemble of standard climatologies for net SW, ne t LW, SH, LH, and net heat flux. 58 Figure 36. WFS in situ climatology and mean and standard deviation of ensemble of standard climatologies for wind spee d, east and north component of wind and momentum flux. 60 Figure 37. Moorings on the West Florida Shelf. 66 Figure 38. Monthly mean relative humidity (RH), barometric pressure (BP), air temperature (AT) and sea surface te mperature (SST) calculated from Weatherpak (solid) and IMET/ASIMET (dash) data. 68 Figure 39. Relative humidity (RH), barometr ic pressure (BP), air temperature (AT) and sea surface temperature (SST) at NA2 in January (left) and June (right) 2001. 70 Figure 40. Monthly mean specific humidity calculated from Weatherpak (solid) and IMET/ASIMET (dash) data. 71 Figure 41. Climatologies of relative humidity (RH) and air temperature (AT) calculated from in situ measurements (dark solid) with one standard deviation (dark dash). 72

PAGE 11

vii Figure 42. Climatologies of air temperatur e (AT), dew point temperature (DPT) and relative humidity (RH) at NDBC Venice C-MAN Station, Florida calculated using data between 1998-2002. 73 Figure 43. Relative humidity (RH) in February 2001 from CM3 (dark) and NA2 (light). 75 Figure 44. Climatology of optimal interp olated cloud-free sea surface temperature (oC) for February, derived from AVHRR and TMI satellites from 19982003 produced by Liu et al. (2005) with WFS air-sea mooring locations overlaid. 79 Figure 45. Relative humidity (RH), barometr ic pressure (BP), air temperature (AT, dark), sea surface temperature (SST, light) and 36-hour lowpass filtered winds at CM3 (left) and NA2 (right ) in January and February 2001. 80 Figure 46. Synoptic weather maps from the National Climatic Data Center Archived NCEP Charts for four days in February 2001 when high RH values were observed. 81 Figure 47. 15-minute observations of relative humidity (R H), barometric pressure (BP), air temperature (AT, dark), s ea surface temperatur e (SST, light) and winds during February 1st-3rd 2001. 82 Figure 48. Specific humidity in air (at 3 m; dark solid), at the sea surface (light solid) and the difference (dash) ca lculated at CM3 in January and February 2001. 84 Figure 49. Specific humidity versus air temp erature at various moorings in February 2001 (top) and February 2003 (bottom). 85 Figure 50. Annual distribution of number of days with observed relative humidity greater than 99% per year. 86 Figure 51. NCEP climatological wind field (arrows) overlaying (left) the skin temperature and (right) sea level pres sure over the western Atlantic Ocean and the Gulf of Mexico for (top) Janu ary, (second) April, (third) July, and (bottom) October. 91 Figure 52. NCEP climatological heat fl ux components averaged over the Gulf of Mexico. 93 Figure 53. The University of South Flor idas array of subsurface and surface moorings on the West Florida Shelf. 95

PAGE 12

viii Figure 54. Low-pass (36-hour Butterworth) filtered meteorological time series and calculated heat fluxes from variables a nd calculated heat fluxes from NA2 between June 1998 and March 2001. 97 Figure 55. Same as figure 54, but from EC3. 98 Figure 56. Same as figure 54, but from CM2. 99 Figure 57. Comparison of meteorological measurements from NA2 (IMET package black) and EC3 (Weatherpak grey) during May 2000. 100 Figure 58. Downward LW during May 2000 (t op) from NA2 (black) and EC3 (grey) and the difference (lower). 101 Figure 59. WFS measurements from June 1998-March 2001. 102 Figure 60. Top panel shows hourly averaged NA2 water temperatures from May to July 2000 at 1m, 4m, 7m, 10m, 13m, 16m, and 19m depths. 103 Figure 61. Components of the one-dimensi onal temperature equation from May to July 2000 at NA2. 105 Figure 62. Same as for Figure 60, but during September-October 2000. 106 Figure 63. Same as for Figure 61, but during September-October 2000. 107 Figure 64. Hourly averaged time series of measured meteorological variables and calculated heat fluxes from NA2 for September and October 2000. 108 Figure 65. Surface fluxes of heat an d momentum in 2000 from EC3. 116 Figure 66. Surface fluxes of heat and momentum in 2000 from NA2. 117 Figure 67. Difference between in situ SST and cool skin, warm layer corrected SST at EC3 (top) and NA2 (bottom) during 2000. 119 Figure 68. Rain sensible heat flux at EC3 (top) and NA2 (bottom) during 2000. 120 Figure 69. Evaporation, NCEP daily m ean precipitation, and EMP during 2000 at EC3 (top three panels) and NA2 (bottom three panels). 121 Figure 70. The differences between the dept h-averaged initial in situ and POM3d model (forced by EDAS winds and rela xed heat flux) temperatures in 2000, at EC3 (star) and NA2 (triangle). 125

PAGE 13

ix Figure 71. EC3 January temperatures (oC) from observations and PWP model experiments. 131 Figure 72. EC3 May temperatures (oC) from observations and PWP model experiments. 132 Figure 73. NA2 May temperatures (oC) from observations and PWP model experiments. 133 Figure 74. NA2 June temperatures (oC) from observations and PWP model experiments. 134 Figure 75. NA2 July temperatures (oC) from observations and PWP model experiments. 135 Figure 76. EC3 September temperatures (oC) from observations and PWP model experiments. 136 Figure 77. NA2 September temperatures (oC) from observations and PWP model experiments. 137 Figure 78. EC3 October temperatures (oC) from observations and PWP model experiments. 138 Figure 79. NA2 October temperatures (oC) from observations and PWP model experiments. 139 Figure 80. EC3 November temperatures (oC) from observations and PWP model experiments. 140 Figure 81. The depth-averaged local rate of change of te mperature (solid) and the net heat flux (dash) computed using the cool skin corrected basic heat flux with the penetrative radiation te rm for each mooring and month. 143 Figure 82. Observed (top), PWP model (mi ddle), and POM1d (bottom) temperature in May 2000 at EC3. 145 Figure 83. For May 2000, the 36-hour lowpass filtered net heat flux (W/m2; a), winds (m/s; b), along and across shelf curren t (cm/s, contour interval 10cm/s; c and d), temperature (oC) from observations (e), PWP model results (f), POM1d model results (g), PO M3d model results (h). 150 Figure 84. For June 2000, the 36-hour lowpass filtered net heat flux (W/m2; a), winds (m/s; b), and along and across shelf current (cm/s, contour interval 10cm/s; c and d), temperature (oC) from observations (e), PWP model results (f), POM1d model results (g), POM3d model results (h). 151

PAGE 14

x Figure 85. For October 2000, the 36-hour lowpass filtered net heat flux (W/m2; a), winds (m/s; b), and along and across sh elf current (cm/s, contour interval 10cm/s; c and d), temperature (oC) from observations (e), PWP model results (f), POM1d model results (g), POM3d model results (h). 152 Figure 86. Air Temperature (oC) Climatologies: January to June. 172 Figure 86. (Continued) Air Temperature (oC) Climatologies: July to December. 173 Figure 87. Sea Surface Temperature (oC) Climatologies: January to June. 174 Figure 87. (Continued) Sea Surface Temperature (oC) Climatologies: July to December. 175 Figure 88. Relative Humidity (%) Climatologies: January to June. 176 Figure 88. (Continued) Relative Humidity (%) Climatologies: July to December. 177 Figure 89. Specific Humidity (kg/kg) Climatologies: January to June. 178 Figure 89. (Continued) Specific Humid ity (kg/kg) Climatologies: July to December. 179 Figure 90. Sea Level Pressure (mb) Climatologies: January to June. 180 Figure 90. (Continued) Sea Level Pre ssure (mb) Climatologies: July to December. 181 Figure 91. Shortwave Radiation (W/m2) Climatologies: January to June. 182 Figure 91. (Continued) S hortwave Radiation (W/m2) Climatologies: July to December. 183 Figure 92. Longwave Radiation (W/m2) Climatologies: January to June. 184 Figure 92. (Continued) Longwave Radiation (W/m2) Climatologies: July to December. 185 Figure 93. Sensible Heat Flux (W/m2) Climatologies: January to June. 186 Figure 93. (Continued) Sensible Heat Flux (W/m2) Climatologies: July to December. 187 Figure 94. Latent Heat Flux (W/m2) Climatologies: January to June. 188

PAGE 15

xi Figure 94. (Continued) Latent Heat Flux (W/m2) Climatologies: July to December. 189 Figure 95. Wind (m/s2) Climatologies: January to June. 190 Figure 95. (Continued) Wind (m/s2) Climatologies: July to December. 191 Figure 96. Wind Speed (m/s2) Climatologies: January to June. 192 Figure 96. (Continued) Wind Speed (m/s2) Climatologies: July to December. 193 Figure 97. Taux (N/m2) Climatologies: January to June. 194 Figure 97. (Continued) Taux (N/m2) Climatologies: July to December. 195 Figure 98. Tauy (N/m2) Climatologies: January to June. 196 Figure 98. (Continued) Tauy (N/m2) Climatologies: July to December. 197 Figure 99. Anomalies of Air Temperature (oC) Climatologies: January to June. 198 Figure 99. (Continued) Anomalies of Air Temperature (oC) Climatologies: July to December. 199 Figure 100. Anomalies of Sea Surface Temperature (oC) Climatologies: January to June. 200 Figure 100. (Continued) Anomalies of Sea Surface Temperature (oC) Climatologies: July to December. 201 Figure 101. Anomalies of Relative Humidity (%) Climatologies: January to June. 202 Figure 101. (Continued) Anomalies of Relative Humidity (%) Climatologies: July to December. 203 Figure 102. Anomalies of Specific Humid ity (kg/kg) Climatologies: January to June. 204 Figure 102. (Continued) Anomalies of Speci fic Humidity (kg/kg) Climatologies: July to December. 205 Figure 103. Anomalies of Sea Level Pre ssure (mb) Climatologies: January to June. 206

PAGE 16

xii Figure 103. (Continued) Anomalies of Sea Level Pressure (mb) Climatologies: July to December. 207 Figure 104. Anomalies of Shortwave Radiation (W/m2) Climatologies: January to June. 208 Figure 104. (Continued) Anomalie s of Shortwave Radiation (W/m2) Climatologies: July to December. 209 Figure 105. Anomalies of Longwave Radiation (W/m2) Climatologies: January to June. 210 Figure 105. (Continued) Anomalie s of Longwave Radiation (W/m2) Climatologies: July to December. 211 Figure 106. Anomalies of Sensible Heat Flux (W/m2) Climatologies: January to June. 212 Figure 106. (Continued) Anomalie s of Sensible Heat Flux (W/m2) Climatologies: July to December. 213 Figure 107. Anomalies of Latent Heat Flux (W/m2) Climatologies: January to June. 214 Figure 107. (Continued) Anomalie s of Latent Heat Flux (W/m2) Climatologies: July to December. 215 Figure 108. Anomalies of Wind Speed (m/s) Climatologies: January to June. 216 Figure 108. (Continued) Anomalies of Wind Speed (m/s) Climatologies: July to December. 217 Figure 109. Anomalies of Taux (N/m2) Climatologies: January to June. 218 Figure 109. (Continued) Anomalies of Taux (N/m2) Climatologies: July to December. 219 Figure 110. Anomalies of Tauy (N/m2) Climatologies: January to June. 220 Figure 110. (Continued) Anomalies of Tauy (N/m2) Climatologies: July to December. 221

PAGE 17

xiii List of Acronyms ADCP Acoustic Doppler Current Profiler ALPEX Alpine Experiment ASIMET Air-Sea Interaction METeorological Package AT Air Temperature AVHRR Advanced Very High Resolution Radiometer BP Barometric Pressure CES Coastal Environmental System C-MAN Coastal-Marine Automated Network COADS Comprehensive Ocean-Atmosphere Data Set COARE Couple Ocean Atmosphere Response Experiment COOS Coastal Ocean Observing Systems ECMWF European Centre for Medium-Range Weather Forecasting EDAS Eta Data Assimilation System EMP Evaporation-Minus-Precipitation ERA-15 ECMWF Reanalysis-15 year data FGGE First GARP Global Experiment GARP Global Atmosphere Research Program GISST Global Ice and Sea Surface Temperature Analyses GMS Geostationary Mete orological Satellite GOES Geostationary Operational Environmental Satellite GPCP Global Precipita tion Climatology Project GTS Global Telecommunications System HIRS High-Resolution Infrared Sounder HOAPS Hamburg Ocean Atmos phere Parameters and Fluxes from Satellite Data ICOADS International Comprehensive Ocean-Atmosphere Data Set IMaRS Institute of Marine Remote Sensing (USF) IMET Improved METeorological Package ITCZ Inter-Tropical Convergence Zone JMA Japan Meteorological Agency LW Longwave Radiation NASA National Aeronautics and Space Administration NCAR National Center for Atmospheric Research NCEP National Centers fo r Environmental Prediction NDBC National Data Buoy Center NOAA National Oceanic and Atmospheric Administration OCG Ocean Circulation Group (USF) OI Optimal Interpolation OSUSFC Oregon State University Climat e Research Institutes Surface Fluxes POM Princeton Ocean Model PRC Precipitation PWP Price, Weller, and Pinkel Model

PAGE 18

xiv RH Relative Humidity S Supersaturation SIRS Satellite Infrared Spectrometer SI International System of units SOC Southampton Oceanography Centre SSM/I Special Sensor Microwave Imager SST Sea Surface Temperature SW Shortwave Radiation TIROS Television Infrared Observation Satellite TMI TRMM Microwave Imager TOGA Tropical Ocean Global Atmosphere TOVS TIROS Operational Vertical Sounder TRMM Tropical Rainfall Measuring Mission USF University of South Florida VTPR Vertical Temperatur e and Pressure Radiometer WFS West Florida Shelf WHOI Woods Hole Oceanographic Institution WHWP Western Hemisphere Warm Pool WMO World Meteorological Organization WND Winds

PAGE 19

xv Ocean-Atmosphere Interactions on the West Florida Shelf Jyotika I. Virmani ABSTRACT Ocean-atmosphere fluxes on the West Flor ida Shelf (WFS) coastal ocean region are investigated using observations and de rived surface fluxes from an array of buoys deployed between 1998 and 2003. The observed a nnual cycle shows that water column temperatures increase and are stratified when heat flux is positive, and they decrease and are well mixed when it is negative. Wate r temperature is minimum (maximum) when heat flux switches sign from negative (positive) to positive (negative) in early spring (autumn). Tropical and extra-tropical events help define the seasona l characteristics of the water temperature. Despite considerable daily and synoptic vari ability in relative humidity, observations on the WFS show th at the monthly mean values are nearly constant at about 75%. Winter relative humidity varies from less than 50% to over 100% (supersaturation values of up to 3% are recorded and coincide with fog on shore) as extra-tropical fronts move over the WFS. Sensor distribution shows small spatial variations in relative humidity in the co astal ocean environment that depends on highfrequency variability in meteorological conditions and low-frequency variability in oceanic conditions. Comparisons with observati ons show that standard climatologies are unable to reproduce spatial vari ability on the WFS, especially in relative humidity and surface heat flux components that are de pendent on sea surface temperature. Model experiments show that careful atte ntion must be paid in calculating and applying surface heat fluxes. Observations and models are employed to assess the relative importance of surface fluxes and convergence of heat flux by the ocean circulation in controlling ocean temperature. In spring and autumn, seasonal change in water temperature is mainly controlled by surface h eat flux with smaller contributions by ocean convergence, but synoptic scale variability is controlled by both surface heat flux and ocean circulation. Surface fl uxes are of primary importance in determining water temperature during the passage of tropi cal storms or extr a-tropical fronts.

PAGE 20

xvi The coastal ocean temperature balance is fu lly three-dimensional. Models must be supported by adequate surface heat flux bounda ry conditions. These require sufficient numbers of in situ measurement points for constraining atmospheric models. The number of observations will depend on the spatial scales of SST variability.

PAGE 21

1 Chapter 1 Introduction 1.1 Introduction In this dissertation various aspects of ocean-atmosphere fluxes over a coastal ocean, namely the West Florida Shelf (WFS), ar e investigated using in situ observations between 1998 and 2003. Interactions between the land, ocean, and atmosphere make the WFS an interesting and dynamic system to study. The ability to model and predict various aspects of such a system requires a comprehensive knowledge of surface fluxes, and how they are affected on synoptic, seasona l and interannual timescales by the larger scale atmospheric and oceanic circulations. Su ch information is also important from a social aspect as increasing numbers of peopl e, living near the coast, are affected by coastal ocean-atmosphere interactions; often conditions offshore determine the weather onshore. Climatological data are useful in provid ing insights about the long-term mean annual cycle of the ocean and atmosphere However, climatologies differ greatly depending on the data, model, or method used to produce them and in some cases do not encompass the coastal regions. Identifying variables that are poorly represented, and which climatology best reproduces the observe d coastal marine atmosphere annual cycle is an important step towards improving coupled ocean-atmosphere models. Sea surface temperature defines the inte rface between the ocean and atmosphere, therefore knowledge of ocean temperature is needed to accurately predict climate and weather. Determining the relative importan ce between surface heat flux and heat flux convergence by ocean circulation in cont rolling ocean temperatures on the WFS throughout the year will benefit oceanic and at mospheric models in this region and their supporting observing systems. Through a combin ation of observations and models, these issues are addressed.

PAGE 22

2 1.2 Background on the Gulf of Mexico and the West Florida Shelf The Gulf of Mexico (Gulf) is an almost entirely enclosed basin, with connections to the Caribbean Sea and the Atlantic Ocean through the Yucatan Channel and Florida Straits, respectively. A net wa ter mass transport of between 23.8 and 28 Sv (e.g., Maul et al. 1985; Schmitz and Richardson 1991; Ochoa et al. 2001; Sheinbaum et al. 2002) flows into the Gulf via the Yucatan Channel and forms the Loop Current. However, 30 Sv flows out of the Gulf as the Florida Curre nt (Larsen 1992), becoming the Gulf Stream along the east coast of the United States. This discrepanc y may be due to mass transport through channels in the Antill ies (Wilson and Johns 1997; Johns et al. 1999; Sheinbaum et al. 2002). The Mississippi Ri ver to the north is the larg est freshwater source to the Gulf, and has a drainage basin which cove rs ~1/3 of the contiguous United States. The bathymetry of the Gulf varies grea tly. Located on the eastern edge of the Gulf, the continental shelf off the west Florid a coast is wide and shallow with the shelf break approximately 200km from the coast. Conversely, the shel f off most of the Mexican coast is much narrower. There is also a broad shelf off the Yucatan Peninsula, at the Campeche Banks. This variation in bathymetry impacts the ocean circulation, which dynamically affects sea surface temperature (SST). Understanding causes of SST variability over this entire region is important because the Gulf, through oceanatmosphere interactions, is a major source of moisture flux to the U.S. Heartlands (Rasmusson 1967; Helfand and Schubert 1995; Higgins et al. 1996) and east of the Continental Divide (Sch mitz and Mullen 1996). The annual cycle of SST in the Gulf is fairly well defined. The Loop Current brings warmer Caribbean waters into the eastern Gulf, however during boreal winter (hereafter seasons will refer to boreal seasons ) the waters along the western, northern and eastern coasts are approximately 10oC cooler than mid-Gulf waters (Figure 1a). In summer the water is much warmer throughout the Gulf, but seasonal upwelling results in cold water on the shallow Campeche Banks along the Mexican coast, and along the northern Gulf coast east of the Mississippi River Delta (Figure 1b). Upwelling also

PAGE 23

3 occurs on the West Florida Shelf (WFS) from fall to spring, with surface manifestations evident in the spring transition period in response to synoptic wind fields (Weisberg et al. 1996). These SST variations across the Gulf of Mexico set up SST gradients that contribute towards driving th e climate over the Gulf and its surrounding landmasses. Meteorological patterns over the WFS s how marked seasonal variability in addition to episodic events such as hurricanes. The atmospheric circulation over this region is greatly influenced by the Bermuda High in the western North Atlantic and the Atlantic ITCZ, which migrates seasonally fr om the southern Caribbean during summer to Figure 1. Gulf of Mexico SST monthly composite satellite image from AVHRR. (a) January. Note: the color bar is different than in Figure 1(b) in order to allow the features to be seen. Figure 1. (b) June. Images provided by IMaRS, USF. SST (OC) 20 30 SST (OC) 20 30

PAGE 24

4 northern South America during winter. NCEP Sea Level Pressure climatology shows annual variations in the Bermuda High (Figure 2). Figure 2. NCEP Sea Level Pressure Climatology (mb). In summer, southeasterly trade wi nds are dominant (Cooper 1987) as the Bermuda High moves northward and is confined to the western Atlantic. This pressure system is integral in driving wind and mois ture across the Gulf and U.S. (Schmitz and Mullen 1996). In autumn and winter a zonal ridge of high pressure extends westwards from the Atlantic across the southern U.S. and northern Gulf of Mexico. The winds are predominantly northeasterly and are accompanied by synoptic extratropical systems (Fernandez-Partagas and Mooers 1975). Sp ring is a transitional period (Cooper 1987). There is also an interannual variability over the Gulf. For example, the NCEP daily reanalysis wind field averaged from April-June shows variations between 1998 and 1999 (Figure 3). Generally during this time of year, the winds are south-easterlies

PAGE 25

5 blowing onto the northern Gulf shore. In 1998, however, westerly winds blew onshore over Florida. 0 2 4 2 5 2 6 2 7 2 8 2 9 3 0 3 1 1998 0 2 4 2 5 2 6 2 7 2 8 2 9 3 0 3 1 1999 Figure 3. April-June mean wind field for 1998 (top) and 1999 (bottom) calculated from NCEP daily-averaged reanalysis. The spring 1998 winds resulted in an a nomalously strong cold tongue over the WFS, in which the colder waters extended mu ch further south and were a result of local air-sea interactions and anomalous upwelli ng caused both by anomalous winds and deep ocean influences (Weisberg et al. 2001). Observations on the WFS have shown that the ocean circulation is affected by a combination of factors. The Loop Current intr udes into the northeast Gulf (Huh et al. 1981), shedding eddies every 6-17 months (Vukovich 1988; St urges 1994; Sturges and Leben 2000) and inducing low-frequency variati ons on the southern a nd outer portions of the WFS (Niiler 1976; Maul 1977). Both observational and modeling studies have shown that landward of the shelf break the current s on the WFS are primarily forced by winds (He and Weisberg 2003a). Over the shelf, seas onal changes in the ci rculation have been detected by two sets of in situ measurements: drifters (Williams et al. 1977) and moored buoys (Weisberg et al. 1996). These measurem ents suggest that dur ing the winter, the

PAGE 26

6 along-shore mid-shelf currents are southwar d, whereas during the summer they are northward. Studies show that the trailing edges of synoptic wintertime frontal systems are generally upwelling favorabl e (Fernandez-Partagas and Mooers 1975). Atmospheric forcing also results in synop tic-scale variability on the middle and inner shelf regions (Blaha and Sturges 1981; Marmorino 1982; Mitchum and Clarke 1986). Satellite data gives evidence of the rapid re sponse of the WFS to synoptic -scale atmospheric forcing. Weisberg et al. (2001) show AVHRR imagery on April 10th, 1998 of the WFS following a period of downwelling. By April 13th, 1998, the AVHRR image shows a welldeveloped cold tongue, following three days of upwelling due to synoptic wind fluctuations. On smaller time-scales, the in teractions between tides and continental shelves are important (Clarke and Battisti 1981). Tidal observations on the WFS have shown that it is dominated by mixed semi -diurnal and diurnal tides (Koblinksy 1981; Marmorino 1983; Weatherly and This tle 1997; He and Weisberg 2002b). Model studies are helpful in understandi ng the circulation on the WFS. Yang and Weisberg (1999) investigated the seasona l circulation of the WFS by forcing the Princeton Ocean Model (POM) with the Hellerman and Rosenstein (1983) climatological wind field. They found that the climatological winds alone we re insufficient to account for the seasonal circulation of the WFS implying that it must also depend on other factors. Comparing model resu lts to drifter data, Yang et al. (1999) concluded that observed circulation features on the WFS, such as the seasonal upwelling resulted from wind forcing, coastal geometry, bottom topogr aphy, and synoptic weather systems. He and Weisberg (2002a) ran the POM using NCEP daily reanalysis winds and heat flux as inputs. Using a surface heat flux correction, based on the difference between modeled and observed SST, the spring transition was acc ounted for by a combination of winds and heat flux. Although reanalysis fields and climatologi es incorporate observations, they alone cannot account for the full variability in th e Gulf. For example, the NCEP sea surface temperature reanalysis fields do not capture the WFS cold tongue feature partly because the reanalysis grid spacing is too far apart. Yet in terms of climate over this region, and the gradient fields and flux estimates, it is a very important feature and needs to be

PAGE 27

7 resolved. Problems with the reanalysis fields suggest that both model improvements and in situ measurements are important. Surface fluxes are important because they are key ingredients of climate. Their reach extends beyond the immediate points of contact between the ocean and atmosphere, into the dynamics and thermodynamics of the Atmospheric Boundary Layer and the Oceanic Mixed Layer. Determining what is happening in these two layers improves our understanding of exchanges between the ocean and atmosphere. Our climate is further complicated by interactions with land, and the west Florid a continental shelf provides a natural tapestry in which to study land -ocean-atmosphere interaction processes. 1.3 Data collection and QA/QC process Since 1998, the Ocean Circulation Group (OCG ), USF, has had an array of up to 14 buoys on the West Florida Shelf (Figure 4). The array was initially designed to resolve ocean processes over the inner shelf and provi de observations of inner and outer shelf interactions. Surface flux moorings were situ ated to provide a broader distribution of observed winds over the WFS, when supplem ented by wind measurements from other moorings and coastal stations ma intained by other agencies. Figure 4. OCG moorings on the West Florida Shelf.

PAGE 28

8 In addition to measuring currents, salin ity and temperature in the water column, four of these buoys are equipped with meteor ological packages: Coastal Environmental Systems Weatherpak or th e Woods Hole Oceanographic Institution (WHOI) designed Improved METeorological/Air-Sea Interaction METeorological (IMET /ASIMET) sensor suites (Figure 5; Hosom et al. 1995), m easuring Air Temperature (AT), Relative Humidity (RH), Barometric Pressure (BP), Wind Speed and Direction (WND), Precipitation (PRC) and Sea Surface Temperature (SST). Figure 5. Schematic (left) and photograph (r ight) of a surface air-sea measurement buoy on the WFS.

PAGE 29

9 The IMET/ASIMET systems and one of the Weatherpaks also measure downward Longwave (LW) and Shortwave (SW) radiatio n. The salinity and temperature data are measured using SeaBird Inc.s SeaCats (S BE-16) or MicroCats (Mcats; SBE-37). The SeaCat observations are sampled once every 20 minutes and the Mcat observations are sampled once every 10 minutes. Currents, m easured using R.D.I. Acoustic Doppler Current Profilers (ADCP), are collected ev ery hour. The samples are collected every second during the first six minutes of each hour and then averaged to provide an hourly value. The Weatherpak data are collected every second for 15 minutes and averaged to provide a 15-minute average value. The IMET/ASIMET data are averaged from onesecond samples over the last minute of a 20minute sampling interval to provide a 20minute average value. In addition to data st ored on the buoys, data are also telemetered real-time via satellite from some of th ese buoys. The sensors used to collect the meteorological data and their spec ifications are given in Table 1. The data are retrieved from the instrume nt and anomalous values outside the general range for good data are removed. Default fl ag values are inserted in lieu of data gaps created by the general cut-off process a nd from other causes such as sensor failure, parity-bit errors, calibration issues, assorted hardware and software issues, and satellitetransmission issues for the real-time data The data are hourly averaged (except for currents) and inspected further for anomalous values. Care is taken to ensure that data during special events (e.g., hurri canes) are retained. As we gear up towards an automated coastal ocean observing system which delivers da ta in near real-time on the web, the data processing codes developed for the meteorological data have been integrated into existing OCG data processing codes. Stored data, supplemented by real time data, from 19982003 are used as the primary source for this study. To identify changes in the time series due to instrument replacements, mooring and instrumentation logs have been meticulously assembled and maintained for the entire array since 1998. There is a pot ential source of errors associ ated with instrumentation: instrument set-up, calibration and drift. Some of these errors have been identified by comparisons between moorings, and in the case of SST between sensors on the same

PAGE 30

10 mooring. This shows the benefit of having re dundant sensors and tw o moorings in close proximity. Data Pack Sensor Range Accuracy Resolutio n Sensor Height* AT W YSI 44203 or 44018 -50oC to +50oC or -40oC to 125oC .15oC or 0.3oC 3.0m AT IA Rotronic MP-100F-20o to 55oC 0.1oC 0.01oC 2.3m RH W Hygrometrics 1020SHT 0% to 100% 4% 3.0m RH IA Rotronic MP-101A 0% to 100% % 0.1% 2.3m BP W Setra 270 800 to 1100 mb 0.5 mb 0.1 mb 3.0m BP IA AIR S2B 850 to 1050 mb 1 mb 0.1 mb 2.3m WND W,IA RM Young 5103 0 to 60 ms-1 0 to 360o 2o 1o 3.2m or 2.8m PRC W RM Young 502030 to 50 mm mm 0.1 mm 2.5m PRC IA RM Young 502010 to 50 mm mm 0.1 mm 2.5m SST W YSI 44034 -50 oC to +50 oC0.15oC -0.9m SW W Eppley 8-48 0.285 to 2.8 m 3-5% 0.1W/m2 2.5m SW IA Eppley PSP 0.285 to 2.8 m 1% 0.1W/m2 2.6m LW W,IA Eppley PIR 3.5 to 50 m 3% 0.1W/m2 2.5m or 2.6m Table 1. Sensors used to collect meteorol ogical data on the WFS. The designation W and IA in the Pack column indicates whic h meteorological packag e the sensor was on. The range, accuracy and preci sion are obtained from the se nsor specifications provided by the manufacturers. *Height relative to mean sea level. Additional sources of errors include th e methods used to calculate the surface fluxes (bulk parameterizations). Errors in the measurement of each variable used to

PAGE 31

11 calculate the surface fluxes will compound th e flux errors. The following are some examples. The in situ SST measurement is ma de below the surface so it is not the skin temperature, which is ideally what is re quired by the bulk formulae. Air temperature measurements may contain biases due to ra diative heating of th e sensor. Radiative heating errors increase with increasing solar radiation and decreasing winds such that a mean daytime error of 0.27oC has been observed in a na turally ventilated sensor (Anderson and Baumgartner 1998). Wind measur ements will contain systematic and random errors. In high sea state conditions, buoy winds may be biased low and sea spray may contaminate the relative humidity measurements. The shortwave and longwave sensors may be coated with salt and aerosols (e.g., Waliser et al 1999; Medovaya et al. 2002). Additionally, ungimbaled radiometer se nsors are subject to errors from buoy tilting and depend on the time of year a nd day (e.g., MacWhorter and Weller 1991), although these may be reduced when hourly av eraged values are used. Although there are many sources of errors, these are minimized as much as possible by preand postdeployment calibrations at the appropriate facilities. 1.4 Calculation of surface fluxes 1.4.1 Bulk parameterization The corrected in situ meteorological and SST data are used to calculate the heat and momentum fluxes (e.g., Figure 6) usi ng versions 2.5 and 3.0 of the TOGA COARE algorithm (Payne 1972; Dickey et al. 1994; Liu 1994; Fairall et al. 1996; Fairall et al. 2003). The version number used will be indicated in each relevant chapter. This algorithm uses scalar bulk parameteri zations to calculate the fluxes.

PAGE 32

12 0 10 20 Wind Speed (m s) 0 20 AT (oC) 0 20 SST (oC) 1000 1020 1040 BP (mb) 0 50 100 RH (%) 0 50 Precipitation (mm) 0 Qnlw(W m) 0 500 Qnsw(W m) 0 0.5 Albedo 0 Qlh(W m) 0 100 Qsh(W m) J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D 0 500 Qnet(W m) 19981999200020012002 Figure 6. 36 hour lowpass filtered data and calculated fluxes from NA2, on the 25m isobath approximately 50km from Tampa Bay. Ga ps indicate when data was unavailable. The sensible heat flux, Qsh, is calculated using Qsh = a cp Ch (Ts Ta) U (1.1) where a is the density of air (1.22 kg m-3), cp is the specific heat of dry air (1004.67 J kg1 K-1), Ch is the Stanton number which is the tr ansfer coefficient of sensible heat, Ts is the

PAGE 33

13 sea surface temperature, Ta is the air temperature, and U is the magnitude of the wind vector relative to the surface current vector. The latent heat flux, Qlh, is calculated using Qlh = a Le Ce (qs qa) U (1.2) where Le is the latent heat of vaporization (2.5 x 106 J kg-1), Ce is the Dalton number which is the transfer coe fficient of latent heat, qs is the saturation specific humidity, and qa is the specific humidity of air. The net shortwave radiation, Qsw, is calculated using Qsw = (1) dsw (1.3) where is the albedo, and dsw is the observe d downward shortwave radiation. The albedo is computed using the atmospheric tr ansmittance and sun alt itude following Payne (1972). The net longwave radiation, Qlw, is calculated using Qlw = (dlw Ts 4) (1.4) where is the longwave emissivity (0.97), dlw is the observed downward longwave radiation, and is the Stefan-Boltzmann constant (5.67e-8 m2 K-4). The momentum flux, is calculated using i = a Cd U ui (1.5) where Cd is the drag coefficient and ui is the east or north horizontal wind component where i is east or north. 1.4.2 Net heat flux The net heat flux at the surface of the ocean, Qnet is calculated using: Qnet = Qsw + Qlw + Qlh + Qsh + Qpen (1.6) where Qpen is the Penetrative Radiation and is an additional term needed in shallow waters such as the coastal ocean where radi ation may be absorbed and reflected by the ocean floor. The amount reflected depends on th ree parameters: water turbidity, type of ocean floor, and depth of the water column. This reflected radiation adds a source of heating to the water column, but upon reaching the surface any remaining reflected

PAGE 34

14 radiation becomes a source of heat loss. Qsw is positive because it is a heat flux into the ocean, Qlw, Qlh, Qsh and Qpen are negative because they are generally heat fluxes out of the ocean. An estimate of the Penetrative Radiation, Qpen, can be obtained using the parameterization determined by Paulson and Simpson (1977): Qpen = I0 [I1 exp (-z/l1) + I2 exp (-z/l2)] (1.7) where I0 is the net insolation at the sea surface, l1 and l2 are attenuation lengths, and I1 and I2 are irradiance constants. Values of l1, l2, I1, and I2 have been determined empirically for various water ty pes, ranging from very clear to very murky as defined by Jerlov (1968) (Table 2). The W FS would normally be Type IA or Type IB water, unless the near-shore environment or Green Ri ver regions (Gilbes et al. 1996) are being considered. The first term in this equation ca lculates the absorption of the red spectral components of solar insolation in the upper few meters of the ocean. The second term calculates the absorption of the remaining insolation. Water Type l1 (m) l2 (m) I1 I2 Very Clear (Type I) 0.35 23 0.58 0.42 Fairly Clear (Type IA) 0.60 20 0.62 0.38 Medium Clarity (Type IB) 1.00 17 0.67 0.33 Fairly Murky (Type II) 1.50 14 0.77 0.23 Very Murky (Type III) 1.40 7.9 0.78 0.22 Table 2. Water Types and Co rresponding Values of l1, l2, I1, and I2. If Qsw is 1000 Wm-2, assuming no absorption by the sea floor, the amount of radiation reaching the ocean floor at the 25m isobath is approximately 100Wm-2 for type IA waters (Figure 7) and the reflected radiation, Qpen at the ocean-atmosphere boundary is 31.2 Wm-2. The difference in the surface Qpen at this depth between type IA and IB waters is 13.8 Wm-2.

PAGE 35

15 0 100 200 300 400 500 600 700 800 900 100 0 0 5 10 15 20 25 30 35 40 45 50 Depth (m)Qpen (W m) IB IA Figure 7. Qpen values for an initial Qsw of 1000 Wm-2 over 50m water column depth for water types IA and IB. The net heat flux will play a larger role in influencing water buoyancy and vertical mixing in shallow regions such as the coastal WFS, than in deeper shelf waters which are affected by the Loop Current. Observations from NA2 for May-July s how red frequency spectra with higher energy at lower frequencies. Peaks at the di urnal and semi-diurnal periods dominate the signal (e.g., Figure 8) and are in response to solar and tidal influences. An increase in energy at synoptic time scales is only discer nable in the temperature fields. The data spans three months, therefore the synoptic variability is not necessarily completely resolved. The time period chosen contained the longest continuous set of meteorological observations.

PAGE 36

16 10e 10e 10e 10e0 10 100 102 104 Air TemperatureoC2/cph 10e 10e 10e 10e 0 10 10 100 102 104 Sea Surface TemperatureoC2/cph 10e 10e 10e 10e0 10 100 102 104 Relative Humidity%2/cph 10e 10e 10e 10e 0 10 10 100 102 104 Barometric Pressuremb2/cph 10e 10e 10e 10e0 10 100 105 1010 Downward SWW2/m4/cph 10e 10e 10e 10e 0 10 100 102 104 106 Downward LWW2/m4/cph 10e 10e 10e 10e0 10 100 102 104 East Wind Componentm2/s2/cphcph 10e 10e 10e 10e 0 10 100 101 102 103 North Wind Componentm2/s2/cphcph Figure 8. Autospectra of observed meteorol ogical data at NA2 from May-July 2000. A 10% cosine window was applied an d frequency band averaging (over B = 9) was performed yielding 17 degrees of freedom. Lines on the left indicate 90% confidence interval.

PAGE 37

17 1.5 Dissertation organization Through a combination of observations a nd models, this disser tation investigates various aspects of the air-sea fluxes on the We st Florida Shelf and how they affect SST. Chapters 3 and 4 are peer-reviewed papers: Virmani, J. I., and R. H. Weisberg, 2005: Relative Humidity over the West Florida Continental Shelf. Mon. Wea. Rev. in press. Virmani, J. I., and R. H. Weisberg, 2003: Features of the Observed Annual OceanAtmosphere Flux Variability on the West Florida shelf. J. Climate 16 734-745. Chapters 3 and 4 are almost identical to the publications: Figure, table, and equation numbers have been changed to agree with the remainder of the dissertation. Abstracts from the papers are included. To maintain co nsistency, abstracts for chapters 2 and 5 are also included. The components of this work are divided into chapters as follows: Chapter 2: Climatological Aspects of the West Florida Shelf Climatologies of ocean-atmosphere variables are compared for the Gulf of Mexico. The difference between these climat ologies and observations are determined by comparing them to monthly mean in situ data from the West Florida Shelf. Variations between the relative humidity climatologies are investigated further in chapter 3. Considerable variability al so exists amongst standard climatologies for heat flux components on the WFS. Chapter 3: Relative Humidity over the We st Florida Continental Shelf Relative humidity variations on the We st Florida Shelf are examined. The monthly mean values are nearly constant at about 75%. Winter has the greatest relative humidity variability; values range from le ss than 50% to over 100% as extratropical

PAGE 38

18 fronts move over the WFS. An offset between values recorded by the Weatherpaks and those recorded by the IMET/ASIMET packages is investigated. In addition to sensor differences, a contributing cause to this offset appears to be the locations chosen for sensor deployment. The coarse NCEP reanalys is grid does not capture RH variability on the WFS, which depends not only on the high-frequency variability in meteorological conditions, but also on the low-frequency variability in oceanic conditions. Chapter 4: Features of the Observed Annual Oce an-Atmosphere Flux Variability on the West Florida Shelf The annual cycle of sea surface temperat ure and ocean-atmosphere fluxes on the West Florida Shelf is described using in situ measurements and climatology. Seasonal reversals in water temperature tendency o ccur when the net surface heat flux changes sign in spring and fall. In spring, surface flux variations result in su ccessive stratification and de-stratification of the water column. Fall is characterized by de-s tratification of the water column and a series of step-like decrea ses in the temperature, in response to both tropical storms and extra-tropical fronts. The surface heat flux is primarily responsible for the spring and fall seasonal ocean temperatur e changes but synoptic scale variability is also controlled by the ocean circulation dynamics. Chapter 5: The relative importance of surface heat flux and convergence of heat flux by ocean circulation in c ontrolling ocean temperature In situ data and the Price, Weller and Pinkel (PWP), and oneand threedimensional versions of the Princeton Ocean Model (POM) are used to investigate the relative importance of surf ace fluxes and heat flux convergence in controlling water temperature on the WFS at diffe rent times of the year. Su rface fluxes are of primary importance in determining the water temperatur e during the passage of tropical storms or extra-tropical fronts. The 3D POM results show improvements when forced by observed surface fluxes over Eta Data Assimilation Sy stem (EDAS) winds and a relaxed heat flux. Although not conclusive due to experimental design flaws, the results suggest that in addition to the heat flux ocean dynamics are requ ired to determine the temperature field.

PAGE 39

19 Chapter 2 Climatological Aspects of the West Florida Shelf 2.1 Abstract Ten standard climatologies of ocean-at mosphere variables are compared for the Gulf of Mexico. Five years of moored data on the West Florida Shelf (WFS) are monthly averaged and used to determine how the sta ndard climatologies di ffer from observations at these limited number of locations. NCEP be st represents the winter WFS air and sea surface temperatures, and da Silva, Oberhube r, and SOC best represent the WFS summer temperatures. The observed barometric pressure in winter deviates from any standard climatology because of extratr opical fronts. In September th e observed pressure is lower than the standard climatologies due to a bias in the monthly mean data. There is a large variation between all relative humidity climat ologies. Considerable variability exists amongst standard climatologies for the co mponents of heat flux on the WFS: the difference between the maximum and minimum values may vary for shortwave radiation during the summer by up to 80W/m2, for longwave during winter the difference may be up to 40 W/m2, for sensible heat flux the difference may be up to 50W/m2, and for the latent heat flux the difference may be up to 140 W/m2. From spring to fall, the Hastenrath and Lamb and moored net longwave climatologi es are comparable, whereas in winter the NCEP and ECMWF longwave fluxes agree more closely with observed values. The WFS latent heat flux climatology in fall and winter is greater than is shown by the standard climatologies. All standard climatologies overestimate the observed wind speeds in winter. There is no standard climatology that completely captures variability on the WFS, suggesting that observations are needed to identify the basic annual coastal oceanatmosphere variability.

PAGE 40

20 2.2 Introduction Climatological data sets are important in providing insights about the long-term mean annual cycle of the ocean and atmosphere. In order to develop climatologies, long time series of observations are needed. But sp atial and temporal data coverage may be sparse, resulting in climatological fields either being constrained by regions and time periods over which data are available, or be ing produced by a combination of models and observations. This introduces an element of uncertainty in the climatological fields because data measurement tech niques have changed over time, and models use different parameterization and data assimilation schemes. A comparison between ten commonly used climatologies (hereafter standard climatologies) in the climate community is conducted for the Gulf of Mexico. Using almost five years of meteorological obser vations from moorings West Florida Shelf monthly means of the measured variables a nd derived surface fluxes are averaged over the five years to produce a in situ climatology Here, the term in s itu is used in the strictest sense in that the data is directly from the locations of the WFS moorings, as opposed to the data used to create the sta ndard climatologies, which are averaged over spatial and temporal scales and provided on re gular grids. The in s itu climatologies are used to determine which standard climat ology, if any, best reproduces the observed coastal marine atmosphere annual cycle and to identify those vari ables that are poorly represented by the climatologies. 2.3 Details of available climatologies The ten standard climatologies listed in Table 3 are chos en because their data are easily available via Internet access as 12-month climatologies, or monthly means from which the climatology can be computed. A monthly mean is the average value of data over a period of one month, whereas a climatol ogical mean is the average of the monthly means over a period of a number of years. Th e spatial and temporal coverage over which

PAGE 41

21 the climatologies are formed varies. The grid spacing also varies between these climatologies from 0.5o latitude by 0.5o longitude in the da Silva and HOAPS to 4o latitude by 5o longitude in the OSUSFC. The la rge grid spacing of the OSUSFC climatology provides limited coverage of the Gulf (e.g., Figure 9). The ICOADS climatology also has limited coverage over the Gulf and does not include the WFS. These climatologies are plotted for completeness a nd to show how they compare with other standard climatologies, however they are not included in the subs equent discussion or comparison with in situ observations. The International Comprehensive Oc ean-Atmosphere Data Set (ICOADS) is based on surface marine observations (predom inantly ship-based) collected since 1854 (Slutz et al. 1985; Woodruff et al. 1987). COADS Release 1a, which underwent stricter quality control procedures, is used in th is study (Woodruff et al. 1993). For ICOADS the monthly means were used to calculate the climatology. The University of Wisconsin-Milwaukee and National Oceanographic Data Center/NOAA da Silva Atlas of surface marine data (Da Silva et al. 1994) is based on COADS release 1 data (Slutz et al. 1985). Th e heat and momentum fluxes were modified, moisture and radiation fluxes were added, and a Beaufort equivalent scale was created to correct wind speed measurements. The European Centre for Medium-Range Weather Forecasts (ECMWF) Reanalysis-15 data (ERA-15) is a data assim ilation using 15 years of data from COADS, FGGE, ALPEX, WMO GTS, TOGA COARE, TOVS, and meteorological data from Japan Meteorological Agency (JMA) and Aust ralian Bureau of Meteorology, as well as the Hadley Centre GISST and NCEP SST analyses. For ECMWF the monthly means were used to calcula te the climatology. The Hastenrath and Lamb (HL) surf ace heat flux climatology is a higher resolution climatology based on ship-based surface meteorological observations from 1911-1970 and bulk parameterizations (Hastenr ath and Lamb, 1978). In addition to the heat fluxes, Hastenrath and Lamb also pr oduced climatologies of pressure, wind speed, and sea surface temperature. These are not available on the Internet, however values taken from the atlas (Hastenrat h and Lamb, 1977) will be used.

PAGE 42

22 Climatology Spatial Coverage Temporal Coverage Grid da Silva 89.75oN-89.5oS, 0.25oE-0.25oW 89.5oN-85.9oS, 0.5oE-359.5oE 1945-1993 1945-1989 0.5o lat x 0.5o lon 1o lat x 1o lon ECMWF Reanalysis-15 90oN-90oS, 0oE-357.5oE 1979-1993 2.5o lat x 2.5o lon Hastenrath & Lamb 29.5oN-29.5oS, 99.5oW-18.5oE1911-1970 1o lat x 1o lon Hellerman & Rosenstein 90oN-90oS, 180oW-180oE 1870-1976 2o lat x 2o lon HOAPS II 80oN-80oS, 180oW-180oE 1987-2002 0.5o lat x 0.5o lon ICOADS 89.5oN-85.9oS, 0.5oE-359.5oE 1960-2002 1o lat x 1o lon NCEP/NCAR Reanalyis 90oN-90oS, 0oE-357.5oE 1948 present 1.9o lat x 1.9o lon 2.5o lat x 2.5o lon Oberhuber 90oN-90oS, 180oW-180oE 1950-1979 2o lat x 2o lon OSUSFC 90oN90oS, 180oW-175oE 1850-1974 4o lat x 5o lon SOC 84.5oN84.5oS, 40.5oE-39.5oE 1980-1997 1o lat x 1o lon Table 3. Ten standard climatologies, their spat ial coverage, base year s, and grid spacing. The Hellerman and Rosenstein (HR) wind stress climatology is based on monthly averaged ship and buoy observations from 1870-1976. The wind stress is derived from observed vector eastward and northward components. The Hamburg Ocean Atmosphere Paramete rs and Fluxes from Satellite Data (HOAPS) data set (Grassl et al. 2000) is computed at the Max Planck Institute and is based on Special Sensor Microwave/Imag er (SSM/I) and Advanced Very High Resolution Radiometer (AVHRR) satellite measurements. The fluxes are calculated using bulk parameterizations. The HOAPS II release us ed here has an expanded time series and

PAGE 43

23 improved fields of moisture and heat fl ux. The HOAPS monthly means were used to calculate a climatology. The National Centers for Environmen tal Prediction/National Center for Atmospheric Research (NCEP/NCAR) Reanalysis monthly means and climatology are derived using an analysis system and data as similation (Kalnay et al. 1996). Data sources include global rawinsonde data, COADS GTS, FGGE, SIRS,VTPR, TOVS, HIRS, SSM/I, GMS. The 4 times daily data is aver aged to form monthly means, which are available from 1948-present, and the mont hly means from 19681996 are calculated to form the climatological means. The Oberhuber climatology is also com puted at the Max Planck Institute (Oberhuber 1988) and is based on an an alysis of the COADS data set. The Oregon State University Surface (OSUSFC) climatology (Esbensen and Kushnir 1981) is based on a global ocean climatology prepared by the National Climatic Center and the Berliand and St rokina (1980) atlas of cloudine ss. Fluxes were calculated using bulk formulae. The Southampton Oceanography Center (SOC ) climatology is based an analysis of the COADS release 1a data set (Woodruf f et al. 1993) enhanced with additional corrections of ship-based observations (Jos ey et al. 1996). The SOC monthly means were used to calculate the climatology. The data quality of all COADS based clim atologies is biased towards frequently traversed ship routes in the northern he misphere. Table 4 shows the climatological variables used. All variables were first conve rted to SI units before comparisons were made. The unavailability of certain variab les within a climatology are indicated by the absence of units. The height at which these surface variables are available may differ between climatologies, depending on whether the observations are predominantly ship, buoy, or satellite based. In cases where a refere nce height is provided, it is included in the discussion. The SOC climatology air temperatur e, specific humidity and winds are 10m above sea level.

PAGE 44

24 A T SST RHSpeHBPSW LW SH LH u,v w ind Wind S peed Tau Da Silva oC oC % g/kg mb W /m2 W /m2 W /m2 W /m2m/s m/s N/m2 ECMWF Reanalysis-15 K K % Pa W /m2 W /m2 W /m2 W /m2m/s N/m2 Hastenrath & Lamb oC mb W /m2 W /m2 W /m2 W /m2 m/s Hellerman & Rosenstein dynes/ cm2 HOAPS II g/kg W /m2 W /m2 W /m2 m/s ICOADS oC oC % k g/kgmb m/s m/s NCEP/NCAR Reanalyis oC oC % g/kg mb W /m2 W /m2 W /m2 W /m2m/s m/s N/m2 Oberhuber oC oC % mb W /m2 W /m2 W /m2 W /m2m/s m/s dynes/ cm2 OSUSFC oC oC % k g/kgmb W /m2 W /m2 W /m2 W /m2 m/s SOC oC oC g/kg mb W /m2 W /m2 W /m2 W /m2 m/s N/m2 Table 4. Variables available for each climatology. If no units are presented, then no data was available for that variable. 2.4 Comparison of standard climatol ogies over the Gulf of Mexico Maps of the Gulf have been produced fo r all standard climatologies. An example is given in Figure 9. This shows the latent heat flux from January to June. The highest latent heat flux (LH) loss occurs in the eas tern Gulf because the warmer waters of the Loop Current in winter promote evaporati on. There is some variability between the climatologies. The NCEP LH loss is the greatest, with values exceeding -200 W/m2 over most of the Gulf in winter. Despite this, wintertime values on the WFS range from less than -100 W/m2 in the NCEP, ECMWF and HL, to about -200 W/m2 in the da Silva and

PAGE 45

25 Oberhuber. In spring and summer, WFS values are consistent amongst climatologies at approximately -100 W/m2. A complete set of climatological maps for all the variables in Table 4, including the latent heat fl ux, may be found in the Appendix. To facilitate comparisons between sta ndard climatologies, companion maps of anomalies from the ensemble mean of the standard climatologies are also shown. The climatological anomaly map for the latent heat flux from January to June is given as an example (Figure 10). Relative to the ensemble mean, NCEP shows the greatest latent heat loss over the Gulf, and SOC shows the least. The largest difference between the two occurs in winter in the eastern Gulf, and is associated with the Loop Current. A complete set of anomaly maps may be found in the Appendix.

PAGE 46

26 Figure 9. Latent Heat Flux (W/m2) Climatologies: January to June.

PAGE 47

27 Figure 10. Anomalies of Latent Heat Flux (W/m2) Climatologies: January to June.

PAGE 48

28 2.5 West Florida Shelf observed climatologies Monthly means are calculated and aver aged (in situ climatology) from observations at NA2, CM2, and CM3 because th ese moorings have more than two years of averaged hourly observations. Any month w ith greater than 50% of missing data is excluded. An overall WFS in situ climatology is also calculated using measurements from all the moorings, includi ng those with less than two years of data (EC3, CMP2, CMP4). The WFS radiation measurements are limited to data collected at EC3, NA2, and CMP4. However due to an instrument error in the pyrogeometer at CMP4, the WFS SW climatology is derived from only two moorings. The AT, SST, BP, and downward SW and LW climatologies are in close agreement at all locations on the WFS (F igure 11). The small exception is the southernmost mooring, CM3, which has warmer AT and SST in the winter. The annual cycle on the WFS shows an incr ease in AT and SST in the summer, with temperatures approaching 30oC. High BP is present over the WFS in winter, which decreases during the summer. A minimum BP in September is a result of tropical storms that annually impact the WFS at this time. The maximu m downward SW is in early summer. As summer progresses the downward SW decreases in magnitude, which coincides with an increase in downward LW radiation. These radiation changes are a consequence of increased cloud cover over the WFS in summer. There is a large variation in RH climatology between moorings. The annual cy cle shows constant relative humidity throughout the year but the range of values varies by up to 10%. RH over the WFS is discussed in greater detail in Chapter 3. Average wind speeds over the WFS are ge nerally about 3-4 m/s the summer and increase to about 6 m/s in the winter at all moorings (Figur e 12). In the southern portion of the WFS (CM3) the wind speeds in spri ng and summer are lower than the central WFS. The climatology of the east component of wind shows that the winds are almost always easterlies at all locatio ns. The north component of th e wind has a distinct annual

PAGE 49

29 cycle. The winter winds are northerlies and the summer winds are southerlies, with a rapid transition between the two wind regimes occurring across the WFS in the fall. 15 20 25 30 oC InSitu Air Temperature climatologies NA2 CM2 CM3 WFS 15 20 25 30 oC InSitu Sea Surface Temperature climatologies 1010 1015 1020 1025 mb InSitu Barometric Pressure climatologies 60 70 80 90 % InSitu Relative Humidity climatologies 100 200 300 W/m2 InSitu Downward Shortwave climatologies J F M A M J J A S O N D 300 350 400 450 MonthW/m2 InSitu Downward Longwave climatologies Figure 11. WFS observations of AT, SST, BP RH, downward SW and LW averaged from 1998-2003.

PAGE 50

30 2 4 6 8 m/s InSitu Wind Speed Climatologies NA2 CM2 CM3 WFS 0 100 200 300 o InSitu Wind Direction climatologies 0 2 4 m/s InSitu East Wind Component climatologies J F M A M J J A S O N D 0 2 4 Monthm/s InSitu North Wind Component climatologies Figure 12. WFS observations of wind speed and direction, and east and north components.

PAGE 51

31 2.6 Comparison between standard and in situ climatologies To best represent conditions on the W FS, seven standard climatologies were linearly interpolated to the loca tion of each mooring (Figure 13). 25 26 27 28 29 30 LongitudeLatitude20501002001000 CM2 CM3 EC3 NA2 CMP2 CMP4 Mooring da Silva half da Silva one, HL, SOC ECMWF, NCEP HR,Oberhuber NCEP Figure 13. Mooring locations and grid points of seven climatologies on the WFS. The ICOADS and OSUSFC data coverage over the Gulf was minimal and HOAPS did not reach the coast so these climatologies were excluded. The SOC climatology coverage did not extend to the CM P4 location, but was available for all other mooring locations. The standard climatolog ical values at NA2, CM2, and CM3 were compared with the corresponding in s itu climatology. Additionally, standard climatological values using linearly interpolated values at all moorings were calculated to correspond to the in situ WFS climatology. Idea lly, the standard climatologies would be recalculated using the monthly mean values for 1998-2003 in order to match the observed

PAGE 52

32 climatology. Monthly means over this pe riod were only available for the NCEP reanalysis so the climatologies were used as they were obtained. The annual cycle of air temperature (F igure 14) and sea surface temperature (Figure 15) between the standard and in situ climatologies are simila r. In winter, the in situ air temperature is colder than any sta ndard climatology and the in situ sea surface temperature is cooler than most of the sta ndard climatologies. At all locations, the NCEP climatology provides the best match to in situ winter temperatures. However during summer and fall the NCEP climatology is colder than the other standard and in situ temperatures. The largest difference between th e NCEP and other clim atologies occurs at NA2, and the smallest difference occurs at CM3. In summer, the da Silva, Oberhuber and SOC climatologies are closest to the in situ temperatures. The annual cycle in the standard pressure matches the in situ pressure apart from two time periods (Figure 16). The in situ barometric pressure in winter is higher than any standard climatology and in September the in situ pressure is lower than the standard climatologies. Although tropical storms occur a nnually in September, these are specific events with extremely low pressure that bias th e in situ value and are not indicative of the background monthly mean pressure. Therefor e, the deviation in September may be explained by a low-bias in the in situ barometric pressure. All climatologies are within a 20% range in relative humidity (Figure 17). A detailed comparison between the in situ and NCEP relative hu midity climatologies is in Chapter 3. It was noted in the previous sect ion that the relative humidity climatologies between moorings on the WFS did not agree. Ho wever, it is interesting to note that the standard climatologies do not agree either. The SOC relative humidity field is not provided with the climatology, and was calcu lated using the SOC specific humidity and air temperature, which are both provided but are given at 10m above sea level. The closest standard climatologies to the in situ relative humidity are the da Silva and Oberhuber at NA2 only. The summer, fall, and early winter values are not well represented by the standard climatologies.

PAGE 53

33 15 20 25 30 oC NA2InSitu Da Silva Oberhuber NCEP ECMWF SOC 15 20 25 30 oC CM2 15 20 25 30 oC CM3 J F M A M J J A S O N D 15 20 25 30 oC WFS Figure 14. Observed and standa rd climatologies of air te mperature at NA2, CM2, CM3 and over all moorings (WFS).

PAGE 54

34 15 20 25 30 oC NA2InSitu Da Silva Oberhuber NCEP ECMWF SOC HL 15 20 25 30 oC CM2 15 20 25 30 oC CM3 J F M A M J J A S O N D 15 20 25 30 oC WFS Figure 15. Observed and standard climatologi es of sea surface temperature at NA2, CM2, CM3 and over all moorings (WFS).

PAGE 55

35 1010 1015 1020 1025 1030 mb NA2InSitu Da Silva Oberhuber NCEP ECMWF SOC HL 1010 1015 1020 1025 1030 mb CM2 1010 1015 1020 1025 1030 mb CM3 J F M A M J J A S O N D 1010 1015 1020 1025 1030 mb WFS Figure 16. Observed and standard climatologies of barometr ic pressure at NA2, CM2, CM3 and over all moorings (WFS).

PAGE 56

36 60 70 80 90 % NA2InSitu Da Silva Oberhuber NCEP ECMWF SOC 60 70 80 90 % CM2 60 70 80 90 % CM3 J F M A M J J A S O N D 60 70 80 90 % WFS Figure 17. Observed and standard climatologies of relative humidity at NA2, CM2, CM3 and over all moorings (WFS).

PAGE 57

37 The annual cycle in net shortwave radiation is similar for all climatologies with a maximum in early summer (Figure 18), how ever the range is approximately 80 W/m2. The da Silva is highest at 300 W/m2 and NCEP is the lowest at about 220 W/m2. At NA2 the in situ and NCEP, Hast enrath and Lamb, ECMWF, a nd Oberhuber climatologies are comparable. There were no in situ radiat ion measurements at CM2 and CM3, but the WFS in situ climatology also includes data from EC3. In this case, th e in situ climatology agrees with the SOC climatology for most of the year. Except in winter, the Hastenrath and Lamb net longwave climatology is the closest to the in situ net longwave climatology (Figure 19). Both of these climatologies have a greater net longwave heat loss than any other climatology fr om spring to autumn. In winter the Hastenrath and Lamb longwave is about 10-15 W/m2 greater than the in situ values, which more closely agree with th e NCEP and ECMWF values at this time instead. The net longwave radiation range am ongst the standard climatologies is 15-40 W/m2. The latent and sensible heat fluxes were calculated from in situ observations using the COARE 3.0 algorithm (Fairall et al. 2003). At NA2 the NCEP and in situ latent heat flux climatologies are in good agreement from March until September (Figure 20). The in situ WFS climatology in fall and winter shows a greater latent heat flux loss than is achieved by the standard climatologies. Th e greatest range between the standard climatologies occurs in winter; at NA2 th e standard climatologies differ by about 140 W/m2. The in situ sensible heat flux (Figure 21) is well represented by the da Silva, Hastenrath and Lamb, SOC, and Oberhuber c limatologies (except in March, when there appears to be an anomalous value in th e Oberhuber). The NCEP and ECMWF values show greater sensible heat flux loss than the other climatologies in spring. The net heat flux varies by up to 100 W/m2 between the standard climatologies (Figure 22). In spring and summer, NCEP ECMWF, Oberhuber, and Hastenrath and Lamb are the closest comparison to the obser ved net heat flux on the WFS. In fall, the WFS heat flux from the ocean to the atmosphe re is higher than any standard climatology.

PAGE 58

38 100 200 300 400 W/m2 NA2InSitu Da Silva Oberhuber NCEP ECMWF SOC HL 100 200 300 400 W/m2 CM2 100 200 300 400 W/m2 CM3 J F M A M J J A S O N D 100 200 300 400 W/m2 WFS Figure 18. Observation derived and standard c limatologies of net s hortwave radiation at NA2, CM2, CM3 and over all moorings (WFS).

PAGE 59

39 W/m2 NA2InSitu Da Silva Oberhuber NCEP ECMWF SOC HL W/m2 CM2 W/m2 CM3 J F M A M J J A S O N D W/m2 WFS Figure 19. Observation derived and standard c limatologies of net l ongwave radiation at NA2, CM2, CM3 and over all moorings (WFS).

PAGE 60

40 0 W/m2 NA2InSitu Da Silva SOC Oberhuber NCEP ECMWF HL 0 W/m2 CM2 0 W/m2 CM3 J F M A M J J A S O N D 0 W/m2 WFS Figure 20. Observation derived and standard c limatologies of latent heat flux at NA2, CM2, CM3 and over all moorings (WFS).

PAGE 61

41 0 W/m2 NA2InSitu Da Silva SOC Oberhuber NCEP ECMWF HL 0 W/m2 CM2 0 W/m2 CM3 J F M A M J J A S O N D 0 W/m2 WFS Figure 21. Observation derived and standard c limatologies of sensible heat flux at NA2, CM2, CM3 and over all moorings (WFS).

PAGE 62

42 0 200 400 W/m2 NA2InSitu Da Silva SOC Oberhuber NCEP ECMWF HL 0 200 400 W/m2 CM2 0 200 400 W/m2 CM3 J F M A M J J A S O N D 0 200 400 W/m2 WFS Figure 22. Observation derived and standard c limatologies of net heat flux at NA2, CM2, CM3 and over all moorings (WFS).

PAGE 63

43 The annual cycle in wind speed is simila r for all climatologies, with higher wind speeds in the winter and lower wind speeds in the summer (Figure 23). However the magnitude between the in situ climatology and the standard climatologies differ, especially in the winter when all the sta ndard climatologies overestimate the in situ climatology. From spring to early winter, at CM2 and CM3 the NCEP and in situ values are comparable and at NA2 the Oberhuber and in situ values are comparable. The Hastenrath and Lamb climatology differs from all the others. The standard and in situ climatological annual cycles in the east (Figure 24) and north (Figure 25) wind components are similar. The winds are predominantly south-easterlies from early spring to late fall, and north-easterlies in winter. The largest diffe rence between the in situ and standard climatologies occurs at CM3, the southern most mooring, where the weakening of the easterlies in spring is not capture d by the standard climatologies. The in situ winds were used to calculat e the east and north momentum flux using the COARE 3.0 algorithm. Apart from NCEP in the winter and Oberhuber throughout the year, all standard climatologies are comparable to the in situ east momentum flux (Figure 26). Likewise, the north momentum flux (Figure 27) is reproduced by all the climatologies except for the NCEP and Oberhuber, which have a greater north momentum flux than in situ values in winter.

PAGE 64

44 0 2 4 6 8 10 12 m/s NA2InSitu Da Silva Oberhuber SOC NCEP HL 0 2 4 6 8 10 12 m/s CM2 0 2 4 6 8 10 12 m/s CM3 J F M A M J J A S O N D 0 2 4 6 8 10 12 m/s WFS Figure 23. Observed and standard climatol ogies of wind speed at NA2, CM2, CM3 and over all moorings (WFS).

PAGE 65

45 0 2 4 m/s NA2InSitu Da Silva Oberhuber NCEP ECMWF 0 2 4 m/s CM2 0 2 4 m/s CM3 J F M A M J J A S O N D 0 2 4 m/s WFS Figure 24. Observed and standard climatol ogies of east component of wind at NA2, CM2, CM3 and over all moorings (WFS).

PAGE 66

46 0 2 4 m/s NA2InSitu Da Silva Oberhuber NCEP ECMWF 0 2 4 m/s CM2 0 2 4 m/s CM3 J F M A M J J A S O N D 0 2 4 m/s WFS Figure 25. Observed and standa rd climatologies of north component of wind at NA2, CM2, CM3 and over all moorings (WFS).

PAGE 67

47 .2 .1 0 0.1 N/m2 NA2InSitu Da Silva SOC Oberhuber NCEP ECMWF HR .2 .1 0 0.1 N/m2 CM2 .2 .1 0 0.1 N/m2 CM3 J F M A M J J A S O N D .2 .1 0 0.1 N/m2 WFS Figure 26. Observation derived and standard climatologies of east momentum flux at NA2, CM2, CM3 and over all moorings (WFS).

PAGE 68

48 .2 .1 0 0.1 N/m2 NA2InSitu Da Silva SOC Oberhuber NCEP ECMWF HR .2 .1 0 0.1 N/m2 CM2 .2 .1 0 0.1 N/m2 CM3 J F M A M J J A S O N D .2 .1 0 0.1 N/m2 WFS Figure 27. Observation derived and standard climatologies of north momentum flux at NA2, CM2, CM3 and over all moorings (WFS).

PAGE 69

49 A 5-year NCEP climatology, interpolated to the mooring locations, was created using the monthly means from 1998-2003 in order test the validity of comparing climatologies derived from longer time periods with the in situ 5-year climatology, For most variables, the NCEP climatology diffe rences from observations were the same, regardless of whether the long-term or 5-ye ar NCEP climatology was used (Figures 2830). The exception to this is the pressure field, where the 5-year NCEP had the same winter and fall variations (due to extratro pical and tropical events ) as the observations. These events are smoothed out in th e longer standard climatologies. 15 20 25 30 oCAir Temperature 15 20 25 30 oCSea Surface Temperature 60 70 80 90 %Relative Humidity J F M A M J J A S O N D 1010 1015 1020 1025 mbSea Level Pressure Figure 28. Comparison of WFS in terpolated long-term NCEP climatology (dash) and 5year NCEP climatology (solid ), calculated with mont hly means from 1998-2003, and WFS in situ observations (solid with s quares) for air temperature, sea surface temperature, relative humid ity and sea level pressure.

PAGE 70

50 100 200 300 400 W/m2Shortwave Radiation W/m2Longwave Radiation 0 W/m2Sensible Heat Flux J F M A M J J A S O N D 0 W/m2Latent Heat Flux Figure 29. Comparison of WFS in terpolated long-term NCEP climatology (dash) and 5year NCEP climatology (solid ), calculated with mont hly means from 1998-2003, and WFS in situ observations (solid with s quares) for shortwave radiation, longwave radiation, sensible heat fl ux and latent heat flux.

PAGE 71

51 2 4 6 8 m/sWind Speed 0 2 4 m/sEast Component of Wind 0 2 4 m/sNorth Component of Wind .2 .1 0 0.1 N/m2East Momentum Flux J F M A M J J A S O N D .2 .1 0 0.1 N/m2North Momentum Flux Figure 30. Comparison of WFS in terpolated long-term NCEP climatology (dash) and 5year NCEP climatology (solid ), calculated with mont hly means from 1998-2003, and WFS in situ observations (solid with squa res) for wind speed, east and north component of wind, east and north momentum flux. The median, upper and lower quartile, 1.5 times the inter quartile range and outliers of the NCEP climatologies are comp ared to the WFS observations (Figures 3133). The upper and lower quartiles are the upper and lower edges of the box, with the median value indicated by the line within th e box. The whiskers mark values within 1.5 times the interquartile range, and outliers fr om this are indicted by the + signs. In summer, the observed air and sea surface temp eratures, and shortwave radiation do not fall within the range of values in the NCEP climatology. In spring, the observed sensible heat flux is outside the NCEP climatology range, and in summer and fall the relative humidity, latent heat flux and longwave radiation values are outside the NCEP

PAGE 72

52 climatology range. These results are the same when statistics for the 5-year NCEP climatology are calculated (not shown). Th erefore for these va riables, the NCEP climatology is unable to reproduc e the observed monthly means at these specific times of the year. 15 20 25 30 35 Air TemperatureoC 15 20 25 30 35 Sea Surface TemperatureoC 60 70 80 90 100 Relative Humidity% J F M A M J J A S O N D 1010 1015 1020 1025 Sea Level PressurembMonth Figure 31. Box-and-whisker plots for the NCEP climatology and WFS observations (solid) for air temperature, sea surface te mperature, relative humidity, and sea level pressure. The box represents median and upper and lower quartiles. The whiskers are 1.5 times the interquartile range. Outliers are indicated by +.

PAGE 73

53 100 200 300 400 Shortwave RadiationW/m2 Longwave RadiationW/m2 0 Sensible Heat FluxW/m2 J F M A M J J A S O N D 0 Latent Heat FluxW/m2Month Figure 32. Box-and-whisker plots for the NCEP climatology and WFS observations (solid) for shortwave radiati on, longwave radiation, sensible heat flux, and latent heat flux. The box represents median and upper a nd lower quartiles. The whiskers are 1.5 times the interquartile range. Outliers are indicated by +.

PAGE 74

54 2 4 6 8 Wind Speedm/s 0 5 East Component of Windm/s 0 5 North Component of Windm/s .2 .1 0 0.1 0.2 East Momentum FluxN/m2 J F M A M J J A S O N D .2 .1 0 0.1 0.2 North Momentum FluxN/m2Month Figure 33. Box-and-whisker plots for the NCEP climatology and WFS observations (solid) for wind speed, east and north co mponents of wind, east and north momentum flux. The box represents median and upper a nd lower quartiles. The whiskers are 1.5 times the interquartile range. Outliers are indicated by +. The remainder of this dissertation uses th e NCEP reanalysis data set because the years covered by the observations (1 998-2003) are readily available.

PAGE 75

55 2.7 Discussion Comparisons between ten commonly used climatologies show major differences in relative humidity and heat flux over the Gu lf of Mexico and West Florida Shelf. These are further compared with climatological m onthly averages from five years of moored ocean-atmosphere data and derived surf ace fluxes, collected on the WFS. The climatological averages of observed air temper ature, sea surface temperature, barometric pressure, downward shortwave and longwave radiation, and wind are in close agreement at all locations on the WFS but there are rela tive humidity differences between moorings. In situ averages are compared to the en semble mean and standa rd deviation of the standard climatologies to determine which standard climatology, if any, best reproduces this coastal marine environment. The standa rd deviations are smaller than expected because they have been computed about th e monthly means, therefore variability on synoptic and shorter time scales have already been removed. The annual cycles of air and sea surface temperature between the standard an d in situ climatologies are similar. In winter the in situ air and sea surface temperat ures are colder than an ensemble of the standard climatologies (Figur e 34). This suggests a deep-oce an bias in the ensemble mean, which may result because many of the climatologies are based on ship observations, and standard ship tracks avoi d coastal regions other than ports. The standard and in situ pressure is in good agr eement over most of the year except in winter, when the in situ barometric pressure is higher than any standard climatology, and in September, when the in situ pressure is lowe r than the standard climatologies. Extremely low pressures associated with tropical storms that occu r annually bias the observed September monthly mean; therefore the low in s itu pressure may not be representative of the average long-term background barometric pressure. The 5-year NCEP climatological comparison (Figure 28) shows that the re latively short time period over which the pressure monthly means are averaged is responsible for the observed annual cycle. Late winter and spring in situ relative humidity values on the WFS fall within one standard deviation of the ensemble mean standard climatology but the summer, fall, and early

PAGE 76

56 winter values are not well represented. A dditionally, figure 28 shows that the observed relative humidity is outside the NCEP range of values in summer and fall. There is large variance between the standard climatologies in the relative humidity. There is also considerable variability in relative humid ity between moorings on the WFS. Chapter 3 will consider, in greater detail, the differences in relative humidity observed over the shelf and why the standard climatologies may differ from in situ measurements. 15 20 25 30 oC Air Temperature InSitu Ensemble Mean S.D +1 S.D. 15 20 25 30 oC Sea Surface Temperature 1010 1015 1020 1025 mb Barometric Pressure J F M A M J J A S O N D 60 70 80 90 % Relative Humidity Figure 34. WFS in situ climatology and mean and standard deviation of ensemble of standard climatologies for AT, SST, BP, and RH.

PAGE 77

57 The annual cycle in net shortwave radiation is similar for all climatologies with a maximum in early summer. Observations fa ll within one standard deviation of the ensemble standard climatological m ean (Figure 35), however depending on the climatology used, long-term monthly mean va lues on the WFS may vary by as much as 80W/m2; the da Silva is highest at 300 W/m2 and NCEP is lowest at about 220 W/m2 (Figure 18). The maximum downward SW is in early summer (Figure 11). As summer progresses the downward SW decreases in magnitude, coincidi ng with increased downward LW radiation. These radiation chan ges are a consequence of increased cloud cover over the WFS in summer which also re sult in lower net LW radiation heat loss from the ocean. Except in winter, the in situ observations show a la rger heat loss due to net longwave radiation than th e standard climatologies (Fig ure 35). In winter, the amount of water vapor in the atmosphere over the W FS is less than at other times of the year (Figure 40, Chapter 3). Difficulties in modeli ng atmospheric water vapor may account for the departures of standard climatology l ongwave radiation from in situ values. The observed sensible heat flux on the WFS agrees with the ensemble m ean of the standard climatologies to within one standard deviation. The major variability in the latent heat flux across the Gulf is a consequence of the warm SST associated with the Loop Cu rrent and higher values are seen over the eastern Gulf. In autumn, the latent heat flux lo ss from the ocean is greater in the in situ WFS climatology than the standa rd climatology ensemble mean. At this time of year, the WFS experiences tropical systems and extratropical fronts, which both increase the latent heat flux from the ocean to the atmo sphere (Figures 64, Chapter 4). Similarly, in spring the low latent heat loss on the WFS co mpared to the standa rd climatologies may be a result of local weather events: Predomin antly in February and March, central Florida and the adjacent coastal ocean experiences very high relati ve humidity (Chapter 3), occasionally accompanied by fog, associated with extra-tropical systems. The evaporative loss during these times is very low. Seasonal weather events on the WFS (with extreme influences on the heat fluxes) may be filtered from the standard climatologies thereby accounting for the diffe rences between the in situ and standard

PAGE 78

58 latent heat flux. Despite diffe rences between various components of the net heat flux (primarily net longwave radiation and latent he at flux), the net heat flux in spring is the same in the in situ and standard climatol ogies, and the variance between the standard climatologies is small. However, the WFS in situ heat flux falls outside one standard deviation from the ensemble standard climatology for most of the rest of the year, so in situ observations leading to surface heat fluxes are necessary in determining the correct net heat flux over the coastal ocean. 100 200 300 400 W/m2 Shortwave Radiation InSitu Ensemble Mean S.D +1 S.D. W/m2 Longwave Radiation 0 W/m2 Sensible Heat Flux 0 W/m2 Latent Heat Flux J F M A M J J A S O N D 0 200 W/m2 Net Heat Flux Figure 35. WFS in situ climatology and mean and standard deviation of ensemble of standard climatologies for net SW, net LW, SH, LH, and net heat flux.

PAGE 79

59 All climatologies show the same annual cycle in winds: winter north easterlies become easterlies in the sp ring, shifting to south easterl ies in the summer before becoming easterlies again in the fall and wind speeds are higher in winter than summer. The wind speeds do not agree in winter or spring, when the standard climatologies overestimate the in situ winds. This suggests a deep-ocean bias since winds offshore tend to be larger than winds at the coast (e.g., Weisberg and Pietrafesa 1983). Small differences between the in situ and standard barometric pressure in winter and spring (Figure 34) may result in differences in atmo spheric isobaric gradients which could result in differences between in situ and standard wind speeds, as computed by the models used in some climatologies; or the observed valu es may be too low because wind observations are affected by waves. Two ot her reasons for this discre pancy involve the method in which the ensemble mean, standard deviati on, and WFS climatologies are calculated. As pointed out earlier, the ensemble mean and standard deviation are calculated from monthly means, but the in situ climat ology is the monthly mean from hourly observations. So the synoptic and diurnal scal e variability has been smoothed out of the standard climatologies, artifici ally creating a smaller variance. Secondly, the WFS in situ climatology is from 5 years of observations. The ensemble mean a nd standard deviation includes many decades of observations. Figure 30 shows that the NCEP reanalysis is closer to observations in m onthly mean wind speeds when 5 years of NCEP monthly means are used. The north component of the wi nd has a distinct annual cycle: The winter winds are northerlies and the summer winds are southerlies, with a rapid transition between the two wind regimes that occurs ac ross the WFS in autumn. The in situ east component agrees with the ensemble sta ndard climatology except during summer, when it is weaker than the climatologies and beco mes westerly. This may be a result of landsea breezes, which may be misrepresented in the standard climatologies because of observational biases in from ships and from satellites: scatterometer winds may be contaminated by proximity to land (Chao et al 2003). The in situ and ensemble standard climatology momentum flux agrees, although ther e is some variance in the ensemble mean so care must be taken in choosing a climatology.

PAGE 80

60 2 4 6 8 m/s Wind Speed InSitu Ensemble Mean S.D +1 S.D. 0 2 4 m/s East Component of Wind 0 2 4 m/s North Component of Wind .2 .1 0 0.1 N/m2 East Momentum Flux J F M A M J J A S O N D .2 .1 0 0.1 N/m2 North Momentum Flux Figure 36. WFS in situ climatology and mean and standard deviation of ensemble of standard climatologies for wind speed, east and north component of wind and momentum flux. The largest variability between the sta ndard climatologies in the heat flux components occurs in the net longwave radi ation, which depends on water vapor and

PAGE 81

61 cloud cover. There are also differences be tween the standard climatologies and the observations during seasons when individual weat her events lead to large changes in the ocean-atmosphere fluxes. These events occur an nually and usually in the same months anomalous years are when these events are ab sent. Therefore the question arises: are the monthly means from the climatologies misr epresenting the observed climatological monthly mean, or are these events biases in the observations? An additional problem with many of the standard climatologies is the bi as towards ship-based observations and ship tracks. Apart from major ports, ships making of ficial meteorological observations tend to steer clear of the coastal regions, and climatol ogies have to rely on the ability of models (and parameterizations), data assimilation sche mes or other observations. This highlights the need for long-term in situ coastal ocean observations. The results shown here suggest that, apart from the relative humidity, little spatial variability wa s found across the WFS suggesting that a small number of moori ngs, deployed over a period of years, were sufficient for obtaining a long-te rm monthly mean. However, this approach to coastal ocean observing is deceptively simple because in order to produce accurate models leading to forecasts of the coastal oceans, th e in situ data must resolve the diurnal, synoptic and interannual variability, which are not included in the climatological annual cycle.

PAGE 82

62 Chapter 3 Relative Humidity over the West Florida Continental Shelf 3.1 Abstract Observed relative humidity variations on the coastal ocean of the West Florida continental shelf (WFS) are examined over the five-year period 1998-2003. Despite considerable daily variability within seas ons, the monthly mean values are nearly constant at about 75%. Summertime specific humi dity is twice that during winter, so high air temperatures are responsible for the lo w summer monthly mean relative humidities. Winter has the greatest relative humidity vari ability; values range from less than 50% to over 100% as extratropical fronts move over the WFS. Saturation (and fog) occurs as warm moist air passes over colder water. Two different sensors, mounted on multiple moorings, were used to make these obser vations. Monthly mean values from the Rotronics MP-100F are higher than the Hygr ometrix 1020SHT. In addition to sensor differences, a contributing cause to this offset appears to be the locations chosen for sensor deployment. NCEP reanalysis clim atology over the WFS and land-based coastal data both show an annual cycle in monthly m ean relative humidity, w ith higher values in summer, suggesting that the reanalysis fiel d is influenced by land. Air-sea fluxes over the WFS are sensitive to small spatial variability in the coastal ocean and atmosphere. The large grid spacing of the NCEP reanalysis does not capture this variability. The lack of coastal ocean data for assimilation biases the NCEP reanalysis fields towards land-based measurements. Increased spatial coverage via evolving Coastal Ocean Observing Systems should remedy this problem by provi ding required information for describing and understanding the complicated ocean-atmosphere interactions that occur on continental shelves.

PAGE 83

63 3.2 Introduction Latent heat flux from the ocean to the atmosphere provides the primary linkage between the Earths climate engi ne and its solar energy source, and humidity is a factor in determining this flux. Here we describe re lative humidity observations made from an array of moorings on the West Florida Con tinental Shelf (WFS), located on the east side of the Gulf of Mexico (her eafter Gulf). The WFS is gently sloping and extends about 200km from the coastline to th e shelf slope. Weather systems affecting the eastern Gulf come from the north (continental Polar ai r) throughout the year, and also from the western Gulf in the winter (maritime Polar air originally from the Pacific), with the maximum number of fronts arriving duri ng December and March (Henry 1979). The average number of frontal systems over Florida and the northern Gulf in the winter is 7-8 (DiMego et al. 1976), occurring on timescales of 4-10 days. Large heat fluxes on the WFS occur during the passage of extratropi cal fronts (Price et al. 1978; Virmani and Weisberg 2003). The Loop Current advects warm Caribbean waters into the eastern Gulf, resulting in large latent heat fluxes east of the Mississippi River delta in the winter. Coastal waters on the shelf can be up to 10oC lower than mid-Gulf waters as a result of coastal upwelling on the WFS and surface heat flux over shallow water in winter and spring. These temperature differences produ ce sea surface temperat ure (SST) gradients and air-sea flux variations that impact the cl imate of the Gulf. In turn, moisture fluxes from the Gulf influence the climate of adj acent landmasses, especially the central U.S. (Rasmusson 1967; Higgins et al. 1996). Locall y, the surface heat flux is responsible for seasonal transitions in WFS ocean temper ature in spring and autumn with ocean dynamics playing a role in synoptic scale va riability (He and Weis berg 2002a; Virmani and Weisberg 2003; He and Weisberg 2003b). There are few observations of relative humidity (RH) in subtropical coastal environments. Breaker et al. (1998a, 1998b) describe some hu midity data collected from two National Data Buoy Center buoys in the northwestern Gulf of Mexico (between 25oN-28oN), in water depths of 120m and 3200m. Th e dominant variabi lity in specific

PAGE 84

64 humidity is on synoptic timescales and is associated with passing fronts between September-November. Complementing this, the present study provide s a description of the annual cycle of RH on the WFS, a subt ropical coastal ocean, with emphasis on the winter season when fog may occur. Many studies of RH in the lower tropos phere and coastal regions pertain to understanding fog in the extra-tropics (e.g., Telford and Chai 1993; Roach 1995). Coastal ocean fog can be formed in many ways. Fog formed on land may be advected offshore by nocturnal land breezes (Pili et al. 1979). Stra tus cloud lowering over the water (Pili et al. 1979) or the onshore movement of marine stratocumulus clouds via sea breeze or orographic effects from coastal mounta in ranges (Cereceda and Schemenauer 1991) result in fog. Radiative cooling of nea r-surface air may also produce fog (Emmons and Montgomery 1947; Roach 1995). In winter, cool surface air over warmer water promotes evaporation and convective mixing and create s fog (Pili et al. 1979). In summer warm moist air gets cooler as it passes over cold wa ter and produces large re gions of persistent, dynamically stable, dense fog (Stone 1936; Noonkester 1979; Roach 1995; Lewis et al. 2003). The location and features of the coas t determine the season and method of fog formation (e.g., Leipper 1994). Marine fogs (visibility <1km) form in 100% RH by condensation on salt nuclei and continue to grow in supersaturated c onditions (Woodcock et al. 1981). Fog may form without supersaturated conditions (Woodcoc k 1978) as a mixture of haze and fog particles (Gerber 1981). The tim e it takes for fog to grow depends on the number and size of nuclei available (Woodcock 1978), but satura ted air needs to persist for at least 103 seconds in order to allow salt nuclei to grow to fog droplet size. Fog layers are isothermal (Goodman 1977), have low wind speeds (Gerber 1981) and very little turbulent mixing (Lala et al. 1975). The concept of supersatur ation has been problematic because humidity is expressed as a percentage that, by de finition, cannot exceed saturation at 100%. Additionally, land-based U. S. radiosondes ha ve had problems reco rding high relative humidity values, returning 96%-98% measur ements when they should have returned 100% or higher values (Golden et al. 1986; Liu et al. 1991; Garand et al. 1992). Despite this, supersaturation has been observed in fog. The nomenclature used is that

PAGE 85

65 supersaturation (S) is expressed as a pe rcentage value above 100%. Hudson (1980) recorded 1% S off Oregon. Meyer et al. ( 1980) measured 0.12% S in a radiation fog. Gerber (1981), using a saturation hygrometer specifically designed to measure relative humidity between 95% and 105%, observed 0.4% S. Early maps of fog frequency around the U. S. show a winter maximum, with 1015 days annually, of heavy fog in west central Florida (Stone 1936). Due to a lack of data over water, Stone infers a January maximu m over the WFS. These numbers differ from subsequent maps because of the number and ty pes of weather stations used, and because the definition of heavy fog has changed. More recent estimates show 20-30 days of heavy fog per year over central Florid a (Court and Gerston 1966; Peace 1969). In addition to describing the observed a nnual cycle in RH over the WFS, we made two observations regarding RH m easurements that warranted further investigation: a) the two different meteorological packages used to collect data showed an offset on a monthly average; and b) the winter months showed relative humidities exceeding 100%. The following section describes the data. Sec tion 3.4 then describes the observed annual cycle. The difference between the two meteorological packages and details of the observed high RH are discussed in sec tions 3.5 and 3.6, respectively. Section 3.7 summarizes our findings. 3.3 Data Since 1998 the Ocean Circulation Group (OCG) in the College of Marine Science, University of South Florida, has maintained an array of up to 14 moorings on the WFS. The deepest and shallowest moor ings were at the 150 m and 10m isobaths, respectively. All the moorings measured current velocities, water temperature, and salinity. Six moorings carried meteorological packages that measured wind velocity, air temperature (AT), SST, RH, and barometric pr essure (BP); some al so had rainfall and shortwave and longwave radiation sensors. Figure 37 shows the location of these moorings and indicates wh ich had either Coastal E nvironmental Systems (CES) Weatherpaks (triangles) or Woods Hole Oceanographic Institution (WHOI) designed

PAGE 86

66 Improved METeorological/Air-Sea Interaction METeorological (IMET /ASIMET) sensor suites (squares). The Weatherpaks used a Hygrometrics 1020SHT Relative Humidity Sensor and YSI 44018 Air Temperature Sensor. The relative humidity sensor operates by measuring the hygromechanical st ress of cellulose crystallit e structures that absorb moisture. The IMET sensor suite used an ALDEN Relative Humidity Module (Model 7030-A), which has a Rotronics MP-100F Humi dity-Temperature Probe. This uses a C80 HYGROMER humidity sensor that measures changes in capacitance as a thin polymer film absorbs water vapor. Figure 37. Moorings on the West Florida Shelf. These two sensors operate under different principles. The Hygrometrics is classified as an organically based sensor, while the Rotronics is a thin film capacitive sensor (Crescenti et al. 1990). The ASIMET p ackage uses the same sensors as the IMET but is manufactured by Star Engineering instead of ALDEN. All RH sensors were located

PAGE 87

67 within an aspirated solar radiation shield to protect them from the effects of solar radiation and precipita tion. All sensors have a protective cover to reduce the effects of salt which are cleaned when the sensors ar e sent for post-deployment calibration. The Weatherpak sensor was mounted 3.0m above sea level and the IMET/ASIMET sensor was 2.3m above sea level. The Weatherpak ca librations were conducted at CES, and the IMET/ASIMET calibrations were carried out at WHOI. The Weatherpak data are collected every second for 15 minutes and averaged to provid e a 15-minute average value. The IMET/ASIMET data are averaged from one-second samples over the last minute of a 20-minute sampling interval to provide a 20-minute average value. In addition to information being stored in the in struments, these data were transmitted back to the OCG via the GOES East satellite. Long-term RH observations in a marine environment are difficult because of potential contamination by sea-spray, heating by solar radiation, length of the deployment and sensor calibration issues. Muller and B eekman (1987) tested eight RH sensors for reliability at various humidities and temper atures, and for long-term endurance. One of the sensors tested was the Rotronics Hygromer, the humidity sensor in the IMET/ASIMET. They found these sensors to be reliable, except in temperatures of 20oC, with no hysteresis effect s occurring after saturation. Although there has been one report of hysteresis effects at high humidities and very low temperatures for the Rotronics MP-100 based on data collected during the Humidity Exchange over the Sea (HEXOS) experiment in 1986 (Katsaros et al. 1994), ot her studies have not found any evidence of this (e.g., Breaker et al. 1998a). Subseque nt manufacturing improvements to the MP-100 sensor may have eliminated this problem (Breaker et al. 1998b), making these sensors more reliable at high humidities. Crescenti et al. (1990) also tested various sensors prior to developing the IMET package. Amongst those tested was the Hygrometrix 8503A, which is a cellulose crystallite sensor and is therefore a kin of the Hygrometrics 1020SHT (manufactured by Hygrometrix) currently us ed in the Weatherpaks. The Hygrometrix sensor they tested was insensitive to humi dities above 90% and exhi bited hysteresis after exposure to high humidities. They concluded th at cellulose crystall ite sensors were not appropriate for use at sea. However, exte nsive developmental te sting by Hygrometrix

PAGE 88

68 (Fenner 1973) and numerous oceanographic appl ications have shown the reliability of this type of sensor (R. Fenner, personal communication). Our measurements returned similar values between both se nsor types (within a few %). 3.4 Annual cycle Monthly means of the observed RH, BP, AT and SST from 1999-2003 for all moorings are shown in Figure 38. Heights of se nsors are given in Table 1. Generally the values between the moorings agree fairly we ll. Between 2000-2003, there is an offset in the monthly mean RH measured by the W eatherpaks (solid) and the IMET/ASIMET (dash) sensors, with the IMET/ASIMET values being 4-8% higher than the Weatherpaks. This offset was present despite calibration efforts so it was difficult to determine which value was correct and this issue wi ll be addressed in section 3.5. 60 80 100 RH (%) 1010 1020 1030 BP (mb) 10 20 30 AT (oC) 1999 2000 2001 2002 2003 10 20 30 SST (oC) Figure 38. Monthly mean relative humidity (RH), barometric pressure (BP), air temperature (AT) and sea surface temperatur e (SST) calculated from Weatherpak (solid) and IMET/ASIMET (dash) data. Individual moor ings are represented by varying grayscale lines.

PAGE 89

69 There was also a calibration problem with the BP sensors at two moorings (NA2 and EC3) in 2000 that was subsequently rectif ied. Hurricanes and tropic al storms result in a yearly pressure minimum in early autu mn, suggesting that the monthly means are biased by a few strong events. The temperatur e sensors on all moori ngs agree with each other. The CMP4 mooring temperatures are lower in the winter of 2003, but as this is seen in both air and sea surface temperatur es it is a genuine feature at the mooring location on the WFS. Weatherpak IMET/ASIMET All Moorings Winter (Jan-Mar) 74.0 13.7 78.5 12.0 75.2 13.2 Spring (Apr-Jun) 71.7 9.5 78.1 8.3 73.4 9.2 Summer (Jul-Sep) 69.9 6.4 77.4 5.3 72.0 6.1 Autumn (Oct-Dec) 69.9 11.2 76.9 10.6 71.8 11.1 Annual 71.5 10.1 77.7 8.7 73.2 9.4 Table 5. Mean and standard deviation of relative humidity (%) from 1999-2003. Table 5 shows the mean and standard de viations of RH per season and on an annual basis for the Weatherpaks, the IMET/ASIMET, and for all moorings combined. The values are calculated using all the availa ble data (real time and stored). The average annual RH on the WFS measured by the Weatherpaks is ~72%, measured by the IMET/ASIMET is ~78%, and for all moorings combined it is about 73%. The later value is biased towards the lower Weatherpak read ings because the array has a larger number of Weatherpak sensors compared to the IMET/ASIMET. The approximate instrument error range given by the manufacturers is 4% for the Hygrometrix and 2% for the Rotronics. Our results indicate that the Weathe rpak is not as unreliable at sea as some studies might suggest. The annual in situ values are slightly lower than COADS climatological estimates for the Gulf of Me xico (Peixoto and Oort 1996). All of our in

PAGE 90

70 situ sensors show greatest va riability during the winter and least variability during summer (Figure 39). The highest RH values occur during winter and the lowest during autumn. 1 7 14 21 28 40 60 80 100 RH (%) 1 7 14 21 28 1010 1020 1030 BP (mb) 1 7 14 21 28 10 20 30 AT (oC) 1 7 14 21 28 15 20 25 30 SST (oC)January 1 7 14 21 28 40 60 80 100 1 7 14 21 28 1010 1020 1030 1 7 14 21 28 10 20 30 1 7 14 21 28 15 20 25 30 June Figure 39. Relative humidity (RH), barometric pressure (BP), air te mperature (AT) and sea surface temperature (SST) at NA2 in January (left) and June (right) 2001. Relative humidity, a measur e of the amount of water vapor in air at a given temperature, is not the best indicator of the true water vapor content in air because of its temperature dependence. Low relative humidity values are a result of less water vapor in the air or high temperatures, and conversely fo r high relative humid ity values. Specific humidity, the mass of water vapor per unit mass of air, is a better indicator. The monthly mean specific humidity, calculated between 1999 -2003 from the in situ WFS data (Figure

PAGE 91

71 40), shows an annual cycle with twice as mu ch water vapor in the summer than the winter. This suggests that the low (~75%) m onthly mean RH during the summer over the WFS is a consequence of high air temperatures. 1999 2000 2001 2002 2003 0.006 0.008 0.01 0.012 0.014 0.016 0.018 0.02 0.022 Specific Humidity (kg/kg) Figure 40. Monthly mean specific humidity ca lculated from Weatherpak (solid) and IMET/ASIMET (dash) data. Individual moori ngs are represented by varying gray-scale lines. The winter RH values are more complicat ed. The AT and water vapor are low so the RH may not necessarily be any higher than in the summer, but la rge fluctuations are seen as a result of synoptic fronts that bri ng cold dry air over Florida and the Gulf, varying the RH from around 100% to less than 50%. On a monthly average these extremes cancel out to give the observed mean values. The wintertime RH fluctuations are discussed further in section 3.6. Climatologies of the multi-year in situ observations of RH and AT (Figure 41) were computed by taking the means of the hourly values for each month over all years. The mean standard deviation of the hourly values per month over the time record shows

PAGE 92

72 that the largest variations occur in January and the smallest in Ju ly. Also shown are the NCEP reanalysis RH and AT climatologies calculated using mont hly means from 19982003 to coincide approximately with our observation period. J F M A M J J A S O N D 55 60 65 70 75 80 85 90 RH(%) J F M A M J J A S O N D 10 15 20 25 30 35 AT(o C) Figure 41. Climatologies of relative humidity (RH) and air temperature (AT) calculated from in situ measurements (dark solid) with one standard deviati on (dark dash). NCEP reanalysis RH and AT climatologies (light solid), calculated usi ng monthly means from 1998-2003 interpolated to each mooring loca tion and then averaged. The line with triangles is the RH calculated us ing NCEP AT and in situ SH (calculated from in situ RH and AT). The NCEP reanalysis grid is 2.5o latitude by 2.5o longitude so only six grid points frame our observation area. The NCEP reanalysis data has been interpolated to each mooring location and then averaged. The in situ RH climatology does not exhibit a noticeable annual cycle unlike the NCEP RH climatology, which is higher during summer. The NCEP climatology has a larger annua l cycle at the grid point closest to land (27.5o N, 82.5oW) and a smaller annual cycle at the shelf break (27.5oN, 85oW). To

PAGE 93

73 determine if the difference between the RH climatologies were due to differences between the in situ and reanalys is AT or water vapor content the in situ RH and AT were used to calculate specific humidity, which was then used with the NCEP AT to calculate RH (Figure 41). The difference between the obs erved and re-calculated RH climatologies is due to a difference in the water vapor c ontent between the observed and reanalysis fields. During the summer months (June-Augus t) over 60% of the difference between the observed and NCEP RH climatologies can be accounted for by differences in water vapor. During winter and spring, the RH difference is almost entirely due to differences between reanalysis and observed AT. Climatologies of on-shore RH were calcu lated using hourly observations of air and dew point temperatures from the Na tional Data Buoy Centers Venice C-MAN (VENF1) coastal station betw een 1998-2002 (Figure 42). The dew point temperature, a measure of air moisture, is the air temperature at which saturation is reached (assuming constant pressure and water vapor). A hi gh dew point temperat ure indicates more moisture in the air. 5 10 15 20 25 30 AT(oC) 5 10 15 20 25 30 DPT(oC) J F M A M J J A S O N D 65 70 75 80 85 RH(%) Figure 42. Climatologies of air temperatur e (AT), dew point temperature (DPT) and relative humidity (RH) at NDBC Venice C-MAN Station, Florida calculated using data between 1998-2002.

PAGE 94

74 The annual cycle in this coastal RH is similar to the NCEP climatology and indicates that NCEP reanalysis values ove r the WFS are influenced by land. NCEP RH from the closest NCEP ocean grid point to the moorings (seaward of the moorings) also exhibits an annual cycle (not shown). The NC EP reanalysis field is the product of data assimilation in an atmospheric model with one-way coupling to an ocean model (Kalnay et al. 1996). Given the paucity of availabl e data over the coastal oceans, it is not surprising that the NCEP reanalysis over th e WFS is biased towards land measurements. Additionally, the large grid spacing of the r eanalysis does not allow it to capture true variations over coastal ocean regions, which poses a problem for coastal oceanatmosphere models. 3.5 Weatherpak and IMET/ASIMET offset The monthly averaged IMET/ASIMET RH values are higher than the Weatherpak RH values (Figure 38; Table 5). There are thr ee possible explanations for this offset: (a) there is a problem in the way one meteorol ogical package measures RH; (b) the sensors were at different heights, wh ich lead to different relative humidities being measured; (c) the sensors were located in different ai r-sea regimes on the WFS. This section investigates these three explanations. We conclude that the primary reason for the Weatherpak-IMET/ASIMET offset is that the sensors are located in different air-sea regimes on the WFS, although we do not discount the fact that a difference in the sensor design may also be a factor and would require further testing in the future. 3.5.1 Instrument differences The most likely cause for the offset between the two meteorological packages would, at first, appear to be a problem in the design or calibration of one. Multiple instruments have been used and all the sens ors have been individually calibrated at various times throughout the five years so it seems unusual that there would be a

PAGE 95

75 persistent offset as observed. Additionally, although the monthly means show an offset, the hourly values do not; on some days the Weatherpak values ar e larger than the IMET/ASIMET values (Figure 43). The temporal o ffset is a result of spatial and temporal variability in meteorological conditions acr oss the WFS, particularly the passage of extratropical fronts. 1 7 14 21 28 0 20 40 60 80 100 RH(%)February 2001 Figure 43. Relative humidity (RH) in February 2001 from CM3 (dark) and NA2 (light). The mean offsets between four moorings (two with Rotronics sensors and two with Hygrometrix sensors) were calculate d over a period of 1 06 days (2544 hourly samples) in 2002 during which time the RH data return was good (Table 6). The offset between the two Rotronics sensor s is greater than the offset between one of the Rotronics (CMP4) and one of the Hygrometrix (CM2) se nsors. Likewise, the offset between the two Hygrometrix sensors is gr eater than the offset between one of the Rotronics (NA2) and one of the Hygrometrix (CM3) sensor s. The greatest offsets are between the southernmost mooring (CM3) and all the inst ruments located to the north. The offsets between the northern moorings are smaller. Th is indicates that al though there may be an instrument bias, it is not easil y discernable with this data. It also suggests that the location of the instruments is important. The correlation coefficients between moor ings are also shown in Table 6. The lowest correlation is between the southern and northern mooring (a Rotronics, CMP4,

PAGE 96

76 and Hygrometrix, CM3, sensor). The highest correlation is between the two closest moorings (a Rotronics, NA2, a nd Hygrometrix, CM2, sensor). This further suggests that the observed RH differences depend more on location than on sensor. The variance was also calculated at each m ooring and increased with incr easing distance offshore (not shown). Rotronics Sensor Hygrometrix Sensor CMP4 NA2 CM2 CM3 CMP4 -2.6050 2.4746 8.2395 Rotronics Sensor NA2 0.6596 5.0874 10.6818 CM2 0.6637 0.7183 5.6864 Hygrometrix Sensor CM3 0.3932 0.5489 0.5473 Table 6. Mean offset and correlation matrix fo r relative humidity sensors. Upper right of the diagonal represents the offsets between sensor pairs. Lower left of the diagonal represents correlation coeffici ents between sensor pairs. 3.5.2 Sensor height differences The Weatherpak sensors were mounted 3.0m above sea level and the IMET/ASIMET sensors were mounted 2.3m above se a level. To investigate the effect of this height difference, observations fro m the IMET/ASIMET sensors were used to calculate the specific humidity and air temp erature at the height of the Weatherpak sensors. We assume that water vapor and te mperature varies with height according to a logarithmic profile:

PAGE 97

77 qwp = q0 + (q*/ ) ln (zwp/zqr) (1a) and twp = t0 + (t*/ ) ln (zwp/ztr) (1b) where q is specific humidity, t is temperature an d z is height. The subscripts wp refer to the Weatherpak sensor and to the value at the sea surface. is the von Karman constant (0.4), q* is the humidity scale (kg/kg), t* is the temperature scale (oC) and zqr, ztr are the specific humidity and temperature r oughness lengths, respec tively. The scales q*and t* are calculated from q* = CE (|Uz|/u*) (qz q0) (2a) and t* = CH (|Uz|/u*) (tz t0) (2b) where CE and CH are the dimensionless Dalton and Stanton numbers, respectively. The subscript z refers to observed va lues, U is the wind velocity, and u* is the surface friction velocity scale given by u* = |Uz| (CD)0.5 (2c) where CD is the neutral drag coefficient. Assuming that the transport of moisture and heat near the surface is dominated by wind shear over buoyancy effects, the parameterizations in 2a and 2b account for the effect of wi nd on the vertical profile of moisture and temperature. The mean difference between the values at the IMET/ASIMET sensor height and the values scaled to the W eatherpak sensor height were 5.25e-5 kg/kg in specific humidity and 0.03oC in AT. These were at least two or ders of magnitude smaller than the measured values and correspond to a 1.4% change in RH, which is within the instrument error range. 3.5.3 Air-Sea regime differences Near the coast SSTs are generally colder in winter and warmer in summer. The northern Gulf waters get colder during winter starting at the coast first and progressing farther offshore with each passing polar contin ental front. In spri ng, as solar radiation

PAGE 98

78 increases, the near-coastal waters warm up first (Huh et al. 1978; He and Weisberg 2002a). The WFS temperature field is furthe r complicated by the shelfs north-south orientation, shallow cross-shelf gradient, o cean dynamics and atmospheric forcing. For example, a seasonal mid-shelf cold tongue deve lops in spring as baroclinic pressure gradients are induced in response to temperat ure gradients formed by differential along and across shelf heating (Weisberg et al. 1996; He and Weisberg 2002a). Model studies have demonstrated a climat ological wind-driven preferen ce for wintertime upwelling and summertime downwelling in the northeaste rn shelf (Yang and Weisberg 1999). A review of the distribution of our inst rumentation over the WFS shows that our IMET/ASIMET sensors are situated closest to the coast and farthest north, while the Weatherpak sensors are farther offshore and south. A climatology of optimal interpolated cloud-free SST maps for the WFS have been derived from AVHRR and TMI satellites from 1998-2003 (He et al. 2003 and Liu et al. 2005). During the winter the IMET/ASIMET sensors are usually in colder wa ters than the Weatherpak sensors (Figure 44). The cooler water aids in cooling th e marine boundary layer immediately above the sea surface relative to moorings located in warmer waters. This is conducive to generally higher RH observations at the moori ngs located in colder waters. The exception to this occurs during the passa ge of synoptic fronts in th e winter when warm air is advected from the south, as discussed in section 3.6. The differ ing coastal air-sea environments in which the sensors are locate d provide the most compelling reason for the observed monthly RH offset between the Weatherpak-IMET/ASIMET data.

PAGE 99

79 .5 .5 .5 .5 .5 25 25.5 26 26.5 27 27.5 28 28.5 29 29.5 30 30.5 LongitudeLatitude NA2 CM3 CM2 CMP4 CMP2 EC3 1 6 1 7 1 8 1 9 2 0 2 1 2 2 2 3 2 4 2 5 Figure 44. Climatology of optimal interpol ated cloud-free sea surface temperature (oC) for February, derived from AVHRR and TMI satellites from 1998-2003 produced by Liu et al. (2005) with WFS air-sea mooring locations overlaid. 3.6 Winter relative humidity and supersaturation Synoptic weather systems over the WFS in winter cause large fluctuations in RH and therefore in latent heat flux from the o cean to the atmosphere (Virmani and Weisberg 2003). High RH results in small latent heat loss from the ocean and conversely for low RH. Occasionally values of 100% RH and higher (supersaturation) have been recorded at our moorings, followed by a drop of 40%-50% Supersaturation commonly occurs in clouds or fog and supersaturation values of 2% have been observed in the marine environment (Breaker et al 1998a). NOAA Storm Data Repor ts summarize information on storms and unusual weather phenomenon co llected by the Nationa l Weather Service

PAGE 100

80 and the National Climatic Data Center. Most of the information is from land-based observations. Reports from 1998-2003 show the presence of fog over west-central and southwestern Florida coasta l counties during the days we observe supersaturation offshore. The closest atmospheric soundings to our moorings are from the Ruskin/Tampa Bay National Weather Service station. On f oggy days these show that the atmospheric boundary layer is approximately 200m thick. In January and February 2001, RH ex ceeded 100% at CM3 (Weatherpak) and almost reached 100% at NA2 (ASIMET) (Figure 45). Figure 45. Relative humidity (RH), barometric pr essure (BP), air temperature (AT, dark), sea surface temperature (SST, light) and 36-hour lowpass filtered winds at CM3 (left) and NA2 (right) in January and February 2001. At both locations the RH approaches 100% when AT equals or exceeds SST over a period of a few days. This occurs when nor therly winds change to southerlies with 0 50 100 RH(%) 1000 1010 1020 1030 BP(mb) 5 10 15 20 25 AT & SST (oC) 14 21 28 1 7 14 0 10 Winds (ms)JanuaryFebruaryCM3 0 50 100 RH(%) 1000 1010 1020 1030 BP(mb) 5 10 15 20 25 AT & SST (oC) 14 21 28 1 7 14 0 10 Winds (ms)JanuaryFebruaryNA2

PAGE 101

81 00Z February 2, 2001 06Z February 12, 2001 18Z February 17, 2001 18Z February 27, 2001 50 48 57 42 66 61 50 43 48 56 55 50 63 50 73 68 74 67 58 50 62 60 66 66 70 66 73 70 73 69 66 64 44 36 43 36 48 39 47 45 51 48 49 52 65 63 58 57 64 66 64 53 73 67 61 61 72 71 69 70 74 73 56 49 71 56 72 74 35 60 30 60 61 36 63 40 66 46 68 50 70 63 85 59 82 70 85 56 73 69 77 70 81 79 64 70 59 65 51 57 53 44 67 61 72 67 71 41 77 40 79 47 75 52 78 61 83 63 79 57 71 67 77 67 77 66 63 83 72 81 65 80 80 76 70 78 70 71 55 52 passing extratropical fronts (Fernandez-Part agas and Mooers 1975), advecting warm air over colder waters. In advance of an appr oaching cold front, cl ockwise turning winds bring warm air from the south (Price et al. 1978). In the wake of the front, northerly winds bring cold dry polar ai r and the RH rapidly drops to around 50%. The cycle repeats itself; warm air, advected by southerlies, lead to higher RH in advance of the next front. Synoptic weather maps from the National Clim atic Data Center Archived NCEP Charts (Figure 46) show that high RH observations (e.g., Figure 43) gene rally occur in the presence of stationary or slow moving co ld fronts. A rapidly moving front on January 10th th (not shown) ensures th at, although there are sout herly winds, AT is only warmer than SST for a few hours and RH remains low. A closer examination on smaller timescales shows that very high RH may also be observed when SST is slightly warmer than AT (Figure 47). Figure 46. Synoptic weather maps from the National Climatic Data Center Archived NCEP Charts for four days in February 2001 when high RH values were observed.

PAGE 102

82 80 100 RH(%) 1015 1020 1025 BP(mb) 15 20 25 AT & SST (oC) 0 10 1 Feb 6am noon 6pm 2 Feb 6am noon 6pm 3 Feb 6am noon 6pm Winds (ms) Figure 47. 15-minute observations of relative humidity (RH), barometric pressure (BP), air temperature (AT, dark), sea surface temperature (SST, light) and winds during February 1st-3rd 2001. At midnight (all times are local time) on February 2, 2001 CM3 recorded values of 100% RH following a few hours of SST be ing warmer than AT. BP and SST are decreasing and AT is increasing at this time. Initial supersat uration (S) is observed at 2am when SST and AT are both 19.9oC. This is warmer than temperatures recorded in fog off the California coast by Goodman (1977). Values of 1% S persist whilst AT is greater than SST (0.2oC) and the winds are weak southerlies (maximum 3m s-1). Supersaturation increases to 2% at 5.15am, as AT begins to decrease and become cooler than SST. These conditions remain for the following 9 hours dur ing which time the winds are weak southeasterlies (less than 4m s). At 1pm, although SST is still warmer than AT, the winds begin rotating clockwise and become nor therlies by 3pm. The maximum S value recorded is 3%, which occurs at 4pm when AT begins to decrease and the wind speed increases, and again between 6-6.45pm imme diately following the coldest AT. From 6.45-9pm BP steadily increases and humidity decreases from 3% S to 99% RH. On

PAGE 103

83 February 2 the air at NA2 remained unsaturat ed, but high values of 99.3% are recorded. Although NA2 recorded similar fluctuations in the wind and BP, one difference between these two sites was that at NA2 AT never equaled or exceeded SST. The smallest SSTAT value (0.07oC) occurred at the same time that 2% S was recorded at CM3. The NOAA Storm Data Report for February 2, 2001 reports widespread dense fog over the west-central Florida coastal c ounties. It is possible that s upersaturation was not observed at NA2 because of its proximity to land. We do not have the ability to measure the Cloud Condensation Nuclei (CCN) at these moorings, but it is possible that a greater number of land-based CCN at NA2 prevente d supersaturation at that lo cation than farther offshore, at CM3. Saturation was also observed at CM3 on February 11th -12th and 17th (not shown) and coincided with observations of patchy dense fog over west-cen tral and southwest Florida (NOAA Storm Data Report). RH valu es at NA2 reached a maximum of 99.4% on these days. Supersaturation wa s only observed on February 11th -12th. In all cases SST was usually slightly greater than or equal to AT and wi nd speeds were low (less than 0.5m s-1 on the 17th). Observations of fog in wi nd speeds smaller than 0.5m s-1 have also been recorded by Gerber (1981). The highest value of 3% S measured by our sensor only existed for less than an hour. The ASIMET sens or used in February 2001 never recorded values greater than 99.4% but, as stated ea rlier, this may be due to its location. From Gill (1982) the observed BP (mb) and SST (oC) can be used to calculate the saturation specific humidity at the ocean surface, qw, and in air, qa: qw,a = (0.62197ew,a)/(BP-0.378ew,a) (3) where ew and ea are the saturation vapor pressures at the ocean surface and in the air, respectively. There are many formulae for calculating ew,a, but the difference between them is negligible (<0.05%) at temperatures of 20-30oC (Elliott and Gaffen 1991). We use the COARE algorithm (Fairall et al 1996), which is based on Buck (1981): ew,a = 6.1121 x (1.0007 + 3.46e-6 BP) exp{17.502 Ts,a /(240.97 + Ts,a)} (4) where Ts is SST and Ta is AT. The specific humidity at the sea surface, q0, is q0 = Kqw. (5)

PAGE 104

84 K is a reduction factor for the saturation vapor pressure over salt water. Assuming salinity is 35ppt, K= 0.98. The sp ecific humidity of air, q, is q = qa(RH/100). (6) The specific humidity at the sea surface depe nds on SST and BP, and is independent of in situ RH, whereas the specific humidity of air depends on AT, BP, and RH. During times of high RH in January and February 2001, the specific humidity of air (q), from observations at 3 m, is greater than th e specific humidity at the sea surface (q0) at CM3 (Figure 48). 14 21 28 1 7 14 .01 .005 0 0.005 0.01 0.015 0.02 Specific Humidity (kg/kg)JanuaryFebruary Figure 48. Specific humidity in air (at 3 m; dark solid), at the sea surf ace (light solid) and the difference (dash) calculated at CM3 in January and February 2001. The plots in Figure 49 show sp ecific humidity values as a function of AT for February 2001 (upper panel) and February 2003 (lower panel). The solid black lines are the saturation specific humidity of air, qa, calculated using (3) with in situ AT and BP from

PAGE 105

85 each mooring. The saturation specific humidity lines from each mooring overlap, as shown in the figure, and they denote the specifi c humidity at which air is saturated at a given AT. The dots are the specific humidity va lues of air, q, at each mooring calculated using (6). There are two things to note from this figure. The distri bution of dots shows the humidity-air temperature range at each m ooring site varies, especially at lower humidities. For example, a specific humidity of 0.006 kg kg-1 occurs between 10oC 14oC in February 2003 at CMP4, however at CM2 the same specific humidity is only found between 13oC 16oC. Conversely, if the AT is between 13oC-16oC the specific humidity is higher at CMP4 than at CM2. Secondly, based on the definition of relative humidity (6), supersaturation (values above the saturation line) was observed in February 2001 at CM3, and the air at NA2 was almost saturated. In February 2003 the air at all moorings was near-saturation at some poi nt during the month and saturation was observed at CMP4. Interestingly, during thes e near-saturation events the RH value remained constant over many hours. This o ccurred simultaneously at multiple moorings using both ASIMET and Weatherpak sensors. On e possible explanation for this is that there are small fluctuations in humidity that the instruments are incapable of resolving. 10 12 14 16 18 20 22 24 26 2 8 0 0.005 0.01 0.015 0.02 Specific Humidity (kg/kg)February 2001 NA2 CM3 Saturation 10 12 14 16 18 20 22 24 26 2 8 0 0.005 0.01 0.015 0.02 Air Temperature (oC)Specific Humidity (kg/kg)February 2003 CMP4 CM3 CM2 Saturation Figure 49. Specific humidity versus air temperat ure at various moori ngs in February 2001 (top) and February 2003 (bottom). Saturation spec ific humidity (solid) is also calculated from data.

PAGE 106

86 February 2001 was used to show exam ples of observations of high RH. The number of days with RH values greater than 99% was compiled to show the annual distribution of high RH (Fi gure 50). The number of days in 1998 may be low because data was only collected for the later part of that year. If there was only one measurement of high RH in the day it was not included. Ve ry high RH values are observed in all years except 2000 and are a recurring feature during winter and early spring on the WFS. February and March have the great est occurrence of high RH days. Figure 50. Annual distribution of number of days with observed relative humidity greater than 99% per year. 3.7 Conclusions Four years of meteorological measurem ents on the WFS have been used to describe the annual cycle in RH in a subtropical coastal ocean environment. The monthly mean values are approximately cons tant at about 75% th roughout the year, but there is considerable daily and synoptic variability between seasons; winter has the

PAGE 107

87 greatest variability, summer has the least. Un like these in situ data, NCEP reanalysis climatology and monthly mean land-based coas tal data for the region show a summer maximum and winter minimum in RH, suggesting that the reanalysis field is influenced by land. The in situ, NCEP reanalysis and land-based coastal annual cycle in monthly mean AT are similar, implying that the differe nces in RH are because of the water vapor. There are two problems in using NCEP reanalys is fields over coastal oceans: a) the large grid spacing does not allow it to capture coastal ocean variability; and b) the reanalysis fields are produced by a model with data assim ilation, but the paucity of in situ data in coastal environments results in an intrinsic land or deep-ocean bias. Both of these issues need to be addressed in order to improve upon coastal ocean results from coupled oceanatmosphere models. This provides added ju stification for emergent Coastal Ocean Observing Systems. Although the monthly mean RH is consta nt during the year, the monthly mean specific humidity, calculated from in situ m easurements, has an annual cycle with twice as much water vapor in the summer than th e winter. This suggests that the monthly average summer RH of 75% is a consequence of high AT. In winter, RH varies according to synoptic fronts. High values are observed ah ead of slow moving or stationary fronts, as southerly winds advect warm moist air over co lder water. During this time the RH can exceed 100% and we observe supersaturation values of up to 3%. Over the WFS in winter dense fog may be formed in a dynami cally stable atmospheric boundary layer as the warm overlying air is cooled by the s ea surface. These types of fogs are more commonly observed in summer in the extr a-tropics. Our observations show that, generally, high RH occurs when AT is clos e to or exceeds SST. However during a fog event SST may be slightly greater than AT and RH will still remain high or continue to increase. From our data we cannot determine if this increase is because of increased evaporation and convective mixing, or because of radiative cooling of the air. A more detailed study under these conditions would be required to determine the exact cause. As the front passes, clockwise rotating winds become northerly and RH decreases by 4050% as cold, dry air is brought into the region. These high/low RH fluctuations are observed every year in winter and early spring on the WFS, with February and March

PAGE 108

88 having the most days with RH greater th an 99%. Occasionally, during high RH, many hours of constant humidity are recorded simu ltaneously at multiple moorings. These may be because RH fluctuations are small and th e sensors are incapable of detecting them. Further work needs to be done to determin e the exact reason fo r the constant RH. Two sensors are used to measure RH. Although the monthly average RH values from the IMET/ASIMET (Rotronics) sensors are higher than from the Weatherpak (Hygrometrix) sensors, regardless of the time of year, they agree to within a few percent suggesting that they are both cap able of making measurements at sea. Part of the offset may be due to the different sensors used, cal ibration and sensor height above sea level, however our analyses suggest that the most important factor is the location of the sensors on the WFS. The IMET/ASIMET sensors are pos itioned farther north and closer to shore than the Weatherpak sensors, and are theref ore in different air-sea regimes. This study has shown the sensitivity of RH to small spatial variations in the coastal ocean environment. RH depends not only on the hi gh-frequency variability in meteorological conditions, but also on the lowfrequency variability in oceanic conditions, specifically SST, which is controlled by both surface heat flux and ocean circulation dynamics (He and Weisberg 2002a; Virmani and Weisbe rg 2003; He and Weisberg 2003b). People who live near the coast are affected by coastal ocean-atmosphere interactions; often conditions such as f og offshore determine the weather onshore. Careful attention must be paid to the sp atial distribution of resources to measure meteorological conditions. More observations in the coastal ocean are required to fully understand air-sea interactions over these la nd-to-deep sea transition regions, which further justifies the need for improved coastal ocean observing systems.

PAGE 109

89 Chapter 4 Features of the observed annual ocean-atmos phere flux variability on the West Florida Shelf 4.1 Abstract The annual cycle of sea surface temperat ure and ocean-atmosphere fluxes on the West Florida Shelf is described using in situ measurements and climatology. Seasonal reversals in water temperature tendency o ccur when the net surface heat flux changes sign in boreal spring and fall. Synoptic-scale variability is also important. Momentum and heat flux variations result in suc cessive water column stratification and destratification events, particularly at shallowe r depths during spring. Fall is characterized by de-stratification of the water column and a series of step-like decreases in the temperature. These are in response to both tropical storms and extratropical fronts. Tropical storms are responsible for the larges t momentum fluxes, but not necessarily for the largest surface heat fluxes. A one-dimen sional analysis of the temperature equation suggests that surface heat flux is primarily responsible for the spring and fall seasonal ocean temperature changes, but that synoptic scale variability is also controlled by the ocean circulation dynamics. During summer, the situation is reve rsed and the major influence on water temperature is ocean dynami cs, with the heat flux contributing to the synoptic-scale variability. There is also evidence of interannual variability: the wintertime temperatures get increasingly co lder from 1998 to 2000; and the greatest stratification and coldest subsurface temperatures occur in 1998. NCEP-NCAR reanalysis fields do not reprodu ce the high spatial flux variability observed in situ or with satellite measurements. Reconciling these di fferences and their impacts on the climate variability of this region provides challenge s to coupled ocean-atmosphere models and their supporting observing systems.

PAGE 110

90 4.2 Introduction Sea surface temperature (SST) within the Gulf of Mexico (h ereafter Gulf) shows considerable spatial and temporal variability. Some of this variab ility is due to ocean dynamics associated with coastal upwelli ng and the Loop Current, but an equally important amount is due to local ocean-atmosph ere fluxes. In all but the summer months the coastal regions of the Gulf are generally cooler than the midbasin, where Loop Current intrusion provides a relatively uni form delivery of warm water. These temperature contrasts produce basinwide SST gr adients and variations in the surface heat, moisture, and momentum fluxes that influen ce the climate of the Gulf and surrounding landmasses. Understanding causes of SST vari ability in this region is important because the Gulf is a major source of moisture flux to the U.S. Heartland (Rasmusson 1967). The West Florida continental Shelf (WFS) o ccupies the eastern si de of the Gulf. Except for the Florida Panhandle in the north where the shelf narrows to a minimum width at DeSoto Canyon, the WFS is broad and gently sloping. It supports a highly diverse and productive ecosystem, and it ha s a major influence on the climate of the surrounding landmasses. The coastline and is obath geometries greatly impact the WFS circulation and the heat and moisture fluxe s of this region. Observation and modeling efforts are underway to investigate the inte ractions between the WFS and the climate of the Gulf. To date, most observational st udies have concentrated on the WFS ocean circulation (e.g., Niiler 1976; Williams et al. 1977; Weisberg et al. 1996), the influence of the Loop Current (e.g., Huh et al. 1981; Sturges and Leben 2000), its tides (e.g., Koblinsky 1981; Weatherly and Thistle 1997; He and Weisberg 2002b), the effects of winds (e.g., Fernandez-Partagas and M ooers 1975; Mitchum and Sturges 1982; Marmorino 1982; Cragg et al. 1983; Clarke and Brink 1985; Mitchum and Clarke 1986; Weisberg et al. 2001), and the impact of tr opical storms on the WFS ocean mixed layer (Price et al. 1978). Less attenti on has been given to the effect of the annual cycle of heat fluxes on the WFS and its adjacent landmass.

PAGE 111

91 7 92424232325 272020212122 January 90 W 75 W 30 N 20 N 1019101971016 20 N 90 W 75 W 30 N 0 226252 2522222222424 April 30 N 20 N 90 W 75 W 1017101610171018 012 114 20 N 30 N 90 W 75 W 29982929 2828July 30 N 75 W 20 N 90 W 101810181017110191020 116 10151 20 N 75 W 90 W 30 N 2428282727229 254252626 October 30 N 75 W 20 N 90 W 101710161015 3014 30 N 75 W 20 N 90 W This paper describes the annual cycle of observed atmospheric fluxes and ocean temperatures at three locati ons on the WFS. The following section gives an overview of the annual cycle over the Gulf region, and previous observational and modeling work over the WFS. Details of the observations used in this work are gi ven in section 4.4. A description and discussion of the annual cy cle of observed ocean temperatures and associated air-sea fluxes for this region of the WFS is in section 4.5. Finally, section 4.6 contains a summary of the salient features. 4.3 Background In broad terms, the annual cycle of SST in the Gulf is fairly well defined. The Loop Current advects relatively warm Caribbe an waters into the eastern Gulf year-round (Figure 51). Figure 51. NCEP climatological wind fiel d (arrows) overlaying (left) the skin temperature and (right) sea le vel pressure over the western Atlantic Ocean and the Gulf of Mexico for (top) January, (s econd) April, (third) July, and (bottom) October. Contour intervals for temperature are 1oC. Contour intervals for pressure are 0.5mb.

PAGE 112

92 Combined with local heating, SST tends to be uniformly warm in boreal summer, and when SST exceeds 28.5oC the Gulf is part of the We stern Hemisphere Warm Pool (WHWP; Weisberg 1996; Wang and Enfield 2001). However, seasonal upwelling often results in cold water on the shallow Campech e Banks, along the Mexican coast, along the northern Gulf coast east of the Mississippi River delta, and on the WFS. Therefore the coastal waters tend to be much cooler with as much as a 10oC difference in SST between the Loop Current and the coastal wa ters, particularly during the winter and spring seasons. These temperature differe nces produce basinwide SST gradients and variations in the surface heat, moisture, and momentum fluxes that influence the climate of the Gulf of Mexico and its surroundings. There are also seasonal atmospheric varia tions that affect the climate of this region (Figure 51). In boreal winter, a zona l ridge of high pressu re extends westwards from the Atlantic across the southern United States and the northern Gulf. In boreal summer, the Bermuda High is better defined, wi th the high pressure system confined to the western Atlantic only. This pressure sy stem is integral in driving the wind field across the Gulf of Mexico and United States. The National Centers for Environmental Prediction (NCEP) climatological heat flux components averaged over the Gulf (Figur e 52) show an annual cycle, with heating of Gulf waters approximately from March to September and cooling at other times. The largest annual variation is in the net s hortwave radiation (SW), which peaks in May. There is also a large annual cycle in the la tent heat flux with a range of almost 100W m-2, and maximum loss during winter. With the a ddition of the net longwave radiation (LW) and sensible heat flux, the annua l cycle in the net heat flux is obtained and varies from about 80W m-2 during the summer to W m-2 during the winter (positive and negative values being fluxes into and out of the ocea n, respectively). Geogr aphically, the largest annual variability in the latent heat flux clim atology occurs east of the Mississippi River delta in the winter and is associated w ith the warm waters of the Loop Current.

PAGE 113

93 J F M A M J J A S O N D 200 150 100 50 0 50 100 150 200 250 300 Month[W m 2] Net Shortwave Net Longwave Latent Heat Flux Sensible Heat Flux Net Heat Flux Figure 52. NCEP climatological heat flux compon ents averaged over the Gulf of Mexico. Using the Da Silva et al. (1994) data, Wang and Enfield (2001) reported a similar annual cycle of fluxes for the entire WHWP region. Averaging the NCEP flux climatology components over the same region fo r direct comparison with the Da Silva climatology shows that although the overall f eatures of the annual cycle are the same between the two climatologies, there is a disc repancy in the magnitudes. From Wang and Enfield (2001), the maximum net SW radia tion in the Da Silva data is ~30W m-2 greater than NCEP during boreal spring and summer. The next largest component, the latent heat flux has ~10W m-2 larger loss in NCEP than in the Da Silva data. The sensible heat flux and net LW radiation losses are also sli ghtly larger in the NCEP climatologies. Therefore, the net heat flux averaged ove r the WHWP during the spring and summer from the NCEP climatology is ~60W m-2 less than that produced by the Da Silva climatology. This is an example of the larg e discrepancies that cu rrently exist between different climatologies. Previous observations on the WFS show th at the ocean circulation is affected by a combination of factors. The Loop Current intr udes into the northern Gulf (e.g., Huh et al. 1981), shedding eddies every 6-17 months (e.g., Vukovich 1988; Sturges 1994; Sturges

PAGE 114

94 and Leben 2000) and inducing low-frequency vari ations on the outer po rtions of the WFS (Niiler 1976; Maul 1977; Meyers et al. 2001). On the shelf, seasonal changes in the circulation have been detected by two sets of in situ measurements: drifters (Williams et al. 1977) and moored buoys (Weisberg et al. 199 6). These measurements suggest that the strongest along-shore midshelf currents occu r during the transition seasons, such that during spring they are southeastward and in fall they are northwes tward. Synoptic-scale variability on the middle and inner shelf regions can result from atmospheric forcing (e.g., Blaha and Sturges 1981; Mitchum and Sturges 1982; Ma rmorino 1982). Satellite data show that SST can respond rapidly to synoptic-scale atmospheric forcing, in the form of well-developed upwelling-induced cold tongues (e.g., Weisberg et al. 2000). The trailing edges of subsynoptic wintertime frontal systems are also generally upwelling favorable (Fernandez-Partagas and Mooers 1975). On smaller timescales, diurnal variability plays an important part on the WFS: tidal observations have shown that the WFS is dominated by mixed semi-diurnal and diurnal tides (e.g., Koblinksy 1981; Marmorino 1983; Weatherly and Thistle 1997; He and Weisberg 2002b). One of the few studies of ocean-atmosphere in teractions on the WFS concerne d the impact of tropical storms on the ocean mixed layer and air-sea heat exchange. Price et al. (1978) used observations and a model for this investigati on and found that entrainment at the base of the mixed layer was the primary mechanism fo r observed deepening of the mixed layer and ocean cooling. Model studies provide insights on the seasonal WFS circulation. Using an adaptation of the Princeton Ocean Model (POM ) forced by the Hellerman and Rosenstein (1983) monthly climatological wind fields, Yang and Weisberg (1999) diagnosed the monthly mean circulation patterns of the WF S. Climatological winds alone were found to be insufficient, suggesting that heat flux a nd baroclinicity must also be important. He and Weisberg (2002a) applied the POM usi ng NCEP-National Center for Atmospheric Research (NCAR) reanalysis winds and heat flux as inputs. Because of low spatial resolution in the NCEP reanal ysis grid, a heat flux rela xation to SST was required to capture the WFS cold tongue. With the fl ux correction, the model baroclinic field accounted for the spring season circulation. As an essentia l contributor to understanding

PAGE 115

95 86 85 84 83 82 81 25 26 27 28 29 30 LongitudeLatitude10002001005020 CM2 EC3 NA2 sub surface surface the seasonal circulation on the WFS, adequa te surface heat (and momentum) flux fields are required, necessitating improvements to bo th models and in situ measurements for this region. 4.4 Observations Since 1998, and in contribution to multidisci plinary studies, University of South Florida personnel have maintained an array of up to 14 buoys on the WFS (Figure 53). At the time the work for this chapter was done an array of up to 12 buoys were being maintained. In addition to measuring currents, temperature, and salinity, four buoys were equipped with meteorological sensors for air temperature, relative humidity, precipitation, wind speed and di rection, and barometric pres sure. Two moorings also measured downward longwave and shortwave radiation. The surface meteorological measurements were made by a combination of Coastal Climate Weatherpaks and Woods Hole Oceanographic Institution (WHOI) design ed Improved Meteorology (IMET) or AirSea Interaction Meteorology (ASIMET) systems. Figure 53. The University of South Floridas array of subsurface and surface moorings on the West Florida Shelf. The data used in th is study are from moorings marked by solid triangles and labeled.

PAGE 116

96 Data from the moorings CM2, EC3, and NA2 (Figure 53) are used. These contain the most complete sets of data (both subsur face temperatures and surface meteorological, Figures 54-56) from June 1998 to March 2001. Mi ssing data are indicated by flag values. NA2 was 37 km offshore at the 25 m isobat h. EC3, approximately 9 km from NA2, was 46km offshore near the 30 m isobath. CM2 was the farthe st offshore mooring to be considered here, and was 102 km offshore in water depth of about 50 m. The water salinity and temperature data, measured using Sea-Bird Electronics SeaCATs are sampled once every 20 minut es, and the MicroCATs and TSKA WaDaRs were sampled once every 10 minutes. The Weat herpak data are collected every second for 15 minutes and averaged to provide a 15-minute average value. The IMET/ASIMET data are averaged from one-second samples over the last minute of a 20-minute sampling interval to provide a 20-minute average value. Hourly averages of these data were used to make the surface heat flux calculations. Most of the data analysis and description of seasons and events are based on these hourly averages. Smoothing using the 36-hour low-pass Butte rworth filter was used to identify features within the entire time series. Th e analysis of the one-dimensional temperature equation required further data processing. Using th e vertically integrated temperature, extrapolated to the depth of the mooring, dT/dt was calculated using a fast Fourier transform, and then low-pass filtered as above.

PAGE 117

97 0 10 20 Wind Speed (m s 1) 0 20 AT (oC) 0 20 SST (oC) 1000 1020 1040 BP (mb) 0 50 100 RH (%) 0 50 Precipitation (mm) 200 100 0 Qnlw(W m 2) 0 500 Qnsw(W m 2) 0 0.5 Albedo 600 400 200 0 Qlh(W m 2) 200 100 0 100 Qsh(W m 2) J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M 1000 500 0 500 2001 19992000 Qnet(W m 2) Figure 54. Low-pass (36-hour Butterworth) fi ltered meteorological time series and calculated heat fluxes from variables and calcu lated heat fluxes from NA2 between June 1998 and March 2001. Gaps due to missing data.

PAGE 118

98 0 10 20 Wind Speed (m s 1) 0 20 AT (oC) 0 20 SST (oC) 1000 1020 1040 BP (mb) 0 50 100 RH (%) 0 50 Precipitation (mm) 200 100 0 Qnlw(W m 2) 0 500 Qnsw(W m 2) 0 0.5 Albedo 600 400 200 0 Qlh(W m 2) 200 100 0 100 Qsh(W m 2) J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M 1000 500 0 500 2001 19992000 Qnet(W m 2) Figure 55. Same as figure 54, but from EC3.

PAGE 119

99 0 10 20 Wind Speed (m s 1) 0 20 AT (oC) 0 20 SST (oC) 1000 1020 1040 BP (mb) 0 50 100 RH (%) 0 50 Precipitation (mm) 200 100 0 Qnlw(W m 2) 0 500 Qnsw(W m 2) 0 0.5 Albedo 600 400 200 0 Qlh(W m 2) 200 100 0 100 Qsh(W m 2) J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M 1000 500 0 500 2001 19992000 Qnet(W m 2) Figure 56. Same as figure 54, but from CM2.

PAGE 120

100 20 25 30 SST (oC) 250 300 350 400 450 Downward LW (W m 2) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 0 500 1000 May 2000 Downward SW (W m 2) Before using the data to make any calculations, some measurement error mitigation was done by comparing data from corresponding instruments on the NA2 and EC3 moorings (8.85 km apart). An example is given in figure 57, which shows the SST, and downward LW and SW measurements at NA2 (IMET) and EC3 (Weatherpak) during May 2000. Figure 57. Comparison of meteorological measurements from NA2 (IMET package black) and EC3 (Weatherpak grey) dur ing May 2000. SST (top), downward longwave (middle) and downward shor twave (bottom) radiation. Note that the SST is the shallowest measur ement of temperature at a depth of 1m, and not the skin temperature. The SST and dow nward SW compare favorably; however there is an average offset in th e downward LW of about 50W/m2 (Figure 58). At first glance, it is not obvious which sensor is correct, b ecause both values fall within previously observed ranges (Weller and Anderson 1996; Jo sey et al. 1997). Parameterizations of downward LW radiation (e.g., Berliand and Be rliand 1952; Anderson 1952; Clark et al. 1974; Bunker 1976) were used with observed SST, air temperature, and barometric pressure to calculate the expected downward LW radiation. These s howed that regardless of the amount of cloud cover, the values of calculated downward LW agree more closely with those measured at EC3.

PAGE 121

101 250 300 350 400 450 W/m2Downward Longwave NA2 EC3 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 0 20 40 60 80 100 W/m2Downward Longwave difference (EC3NA2) Mean: 55.5 SD: 8.4 Figure 58. Downward LW during May 2000 (top) from NA2 (black) and EC3 (grey) and the difference (lower). It was since determined that the NA2 IM ET LW sensor from May to December 2000 was incorrectly calibrated and was rectified for subsequent deployments. A discrepancy was also found in the pressure sensors, with readings at EC3 being about 6mb higher than NA2. Results from chapter 3 showed which sens ors were correct. At the time of this work the discrepancy in the pressure sensor s had not been resolved; however, as this results in a net heat flux di fference of less than 1 W m-2, it was not a major source of error. Hourly averaged values were used to calculate the surfac e heat and momentum fluxes using the COARE 2.5 flux algorithm.

PAGE 122

102 J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M 10 15 20 25 30 35 199920002001 Water Temperatures ( oC) J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M 300 200 100 0 100 200 300 2001 19992000 Net Heat Flux (W m 2) 4.5 Description and discussion As expected in the Northern Hemisphere, the temperatures in the water column at all locations showed warming during spring and early summer, peaking in August, and cooling during fall, usually reaching a minimum in February (Figure 59). Figure 59. WFS measurements from June 1998-March 2001. (top) The10-day low-passfiltered CM2 water temperatures at 1m, 5m, 10m, 15m, 20m, 25m, 30m, 35m, and 40m depths. The line thickness thins with increasing depth: the thickest is at 1m; the thinnest is at 40m. (bottom) The10-day low-pass-filter ed net heat flux calculated from measured parameters. Data from all three moorings we re combined to get the net heat flux series. There is an approximate 90o phase shift between the annual cycles of temperature and net heat flux, with heat flux leading temper ature; that is, waters begin their seasonal warming or cooling when the net surface heat flux switches sign in the spring and fall and the largest rate of change of temperature occu rs when the net surface heat flux is largest. During the spring transition season, temperatures are increasing and th e heat gain to the ocean is greatest. Conversely, during the fall transition season, decreasing temperatures correspond to a time of maximum heat loss from the ocean. During spring and early summer (Febru ary-July), in addition to gradually increasing water temperatures, the subsurf ace temperature profile shows considerable

PAGE 123

103 22 23 24 25 26 27 28 29 30 31 32 Water Temperatures (oC) 0 0.2 0.4 0.6 Wind Stress (N m 2) 1000 500 0 500 1000 Daily Mean Qnet (W m 2) 1 7 14 21 28 1 7 14 21 28 1 7 14 21 28 1 1000 500 0 500 1000 July June May Qnet (W m 2) stratification. At NA2 and EC3, variations in stratification occur at synoptic timescales (Figure 60). Decreases in wind stress and incr eases in surface heat flux cause increases in stratification. Decreased wi nd stress leads to decreased turbulent mixing, and increased heat (buoyancy) flux leads to incr eased water column stability. Figure 60. Top panel shows hourly averaged NA2 water temperatures from May to July 2000 at 1m, 4m, 7m, 10m, 13m, 16m, and 19 m depths. The line thickness thins with increasing depth. Middle panel shows hourly averaged wind stress. Bottom panel shows the net heat flux (thin line) and its daily mean (thick line). Ocean current data collected at NA2 during this time period (not shown) shows that coastal upwelling, in response to increased wind stress, is coincident with all periods of stratification during May and June, however in July stratification occurs without upwelling. Conversely, downwelling results pr edominantly in destratification, however, there are times when destratification occurs without any associat ed downwelling. This shows the complicated interplay between ocean dynamics and surface heat flux in determining the temperature of the water colu mn. Such synoptic-scale variability is not as evident at the farthest offshore CM2 site where the stratification is larger, requiring proportionately larger flux variations to affect change.

PAGE 124

104 To determine the relative importance of the surface heat flux and ocean dynamics in controlling SST, we analyzed the one -dimensional depth-averaged temperature equation using the corrected data from NA2: H c Q t Tp net (1) where T is the vertically integrated temperature, cp is the heat capacity of seawater (3998 J kg-1 K-1), is the density of seawater (1023.34 kg m-3), H is the water depth, and Qnet is the net surface heat flux from Qnet = Qsw + Qlw Qlh Qsh Qpen (2) where Qsw is the net shortwave radiation, Qlw is the net longwave radiation, Qlh is the latent heat flux, Qsh is the sensible heat flux, and Qpen is the penetrative radiation over twice the depth of the water column (Cha pter 1). The assumption made is that no radiation is absorbed at the coastal ocean floor, instead it is entirely reflected and therefore contributes to the heat loss from the ocean. Th is assumption is a simplification, however information is currently unavailabl e to determine this more precisely. An estimate of Qpen, with maximum values of 31 W m-2, at NA2 was obtained using the Paulson and Simpson (1977) parameteri zation for type-1A water (Jerlov 1968). Absorption of Qpen through the water column follows an exponential decay. The residual, obtained by subtracting the ri ght side of (1) from the left, accounts for the threedimensional effects of ocean circulation dynamics on dT/dt i.e. advection, and any remaining errors. The change in the vertically integrat ed water temperature at NA2 during MayJuly is 6.92oC, of which 3.28oC (47.5%) results from surface heat flux, and 3.63oC (52.5%) results from ocean dynamics. The cu rrents along the WFS during this time are predominantly northwestward, advecting warmer water from the south, so ocean dynamics are marginally more important in affecting temperature change over this 3month period than the surface heat flux. However, on a month-by-month basis, the relative importance of these tw o factors changes because this period covers the transition of the surface heat flux from positive to negative.

PAGE 125

105 dT/dt ( oC day 1) 0.4 0.2 0 0.2 0.4 Mean:0.0752Std:0.0951 (dQ/dz)/ *cp ( oC day 1) 0.4 0.2 0 0.2 0.4 Mean:0.0357Std:0.0658 0.4 0.2 0 0.2 0.4 residual ( oC day 1) 1 7 14 21 28 1 7 14 21 28 1 7 14 21 28 Mean:0.0395Std:0.0614MayJuneJuly Figure 61. Components of the one-dimensional temperature equation from May to July 2000 at NA2. Top panel shows depth-averaged temperature change; middle panel shows surface heat flux; bottom panel shows the re sidual, which combines ocean dynamics and errors. The largest depth-averaged temperature increase (3.86oC) is during May; 61% of this is due to the surface heat flux (2.35oC) and the other 39% is due to ocean dynamics (1.51oC). Although heat flux is the largest c ontributor to temperature change during May, ocean dynamics are responsible for the sy noptic-scale variability (shown in Figure 61 and obtained from the standard deviation). The relative influences of surface heat flux and ocean dynamics are almost equal in June which has a total vertically integrated temperature increase of 2.32oC. In this case, the ocean dynamics account for 55.5% (1.31oC), and the heat flux accounts for 43.5% (1.01oC). The daily mean net heat flux begins to change sign from positive to negative during June. This is evident in July when the surface heat flux has a cooling effect (-0.08oC) and ocean dynamics (0.81oC) are mostly responsible for the change in th e vertically averaged temperatures (0.74oC). No upwelling events are observed in July and th e only downwelling event observed serves to

PAGE 126

106 24 25 26 27 28 29 30 31 32 Water Temperatures ( oC) 0 0.2 0.4 0.6 Wind Stress (N m 2) 1000 500 0 500 1000 Daily Mean Qnet (W m 2) 1 7 14 21 28 1 7 14 21 28 1000 500 0 500 1000 September October Qnet (W m 2) destratify and increase the overall water co lumn temperature, thereby increasing the vertically averaged temperature. Contrary to May, the synoptic-sca le variability during July was mostly in response to variations in the surface heat flux. The relative influences of heat flux and ocean circulation dynamics on the spring transition period were also investigated by He and Weisberg (2002a) using a regional adaptation of the Princeton Ocean Model for ced by NCEP reanalysis wind and heat flux, and by river inflows. Consis tent with our results from May 2000, they found that heat flux largely controls the spring season transi tion of water temperature. Large synoptic scale variations were attributed to a co mbination of ocean circulation dynamics and surface heat flux. With shoaling water depth, the convergence of heat flux by the ocean circulation plays a proportio nately larger role. Figure 62. Same as for Figure 60, but during September-October 2000. Hurricane Gordon passed the mooring between 16-19 September. An extratropical weather system with cool, dry air passed the moor ing between 9-15 October. Annual cooling and de-stra tification of the water column begins almost concurrently at all locations during Septem ber of each year, and continues throughout the

PAGE 127

107 dT/dt ( oC day 1) 1.5 1 0.5 0 0.5 1 Mean: 0.0774Std:0.3160 (dQ/dz)/ *cp ( oC day 1) 1.5 1 0.5 0 0.5 1 Mean: 0.0857Std:0.1067 1.5 1 0.5 0 0.5 1 residual ( oC day 1) 1 7 14 21 28 1 7 14 21 28 Mean:0.0083Std:0.2718October September fall and winter. The daily mean surface heat flux continues to decrease and becomes almost entirely negative as the cooling continues. During September and October 2000 (Figure 63), the mean vertically integrated temperature changes by .72oC. Over this period, ocean dynamics continues to ha ve a marginal warming effect (0.51oC), but the major factor in the observed heat loss is due to the surface heat flux (-5.23oC). Analogous with the spring season transition, the heat flux largely controls the fall season transition of water temperature change, but the ocean dynamics are mainly responsible for the synoptic-sca le variability. Figure 63. Same as for Figure 61, but during September-October 2000. The cooling of the water column is not gradual during the fall transition period. Instead it occurs in a series of steplike decrea ses in temperature at all depths (Figure 62). This is the peak of the hurricane season a nd some of the observed decreases in water temperature are due to the passage of tropi cal storms or hurricanes. Hurricane Gordon, 16-19 September 2000, for example, results in a maximum daily net heat loss of almost 300W m-2, and a depth-average temperature change of -0.75oC at NA2 (Figure 62). The surface heat flux accounts for 80% of this decrease (-0.6oC) and ocean circulation is responsible for the other 20% (-0.15oC). The fall cooling and de stratification begins in earnest once the first major tropical storm pa sses by. The high winds lead to a well-mixed

PAGE 128

108 10 20 30 AT (oC) 990 1000 1010 1020 1030 BP (mb) 0 50 100 RH (%) 0 20 40 Precipitation (mm) 300 200 100 0 Qnlw (W m 2) 0 500 1000 Qnsw (W m 2) 600 400 200 0 Qlh (W m 2) 1 7 14 21 28 1 7 14 21 28 200 100 0 100 September October Qsh (W m 2) water column (even at CM2, which is twice as deep as NA2). The temperature decrease during the tropical storm, and th e generally negative net heat flux during this time of year ensures that water temperatures cannot warm up again following a storm. Therefore the water continues to cool and the water column remains destratified. Figure 64. Hourly averaged time series of measured meteorological variables and calculated heat fluxes from NA2 for September and October 2000. We need to be cautious in attributing all the steplike decreases in temperature to tropical storms during this season. Extratropical weather systems are also very important. For example, the temperature decreases by ~3oC between the 9 and 15 October 2000 (Figure 62). This decrease coincides with an increase in wind speed and barometric pressure, a decrease in air temperature a nd relative humidity, an increase in the net longwave radiation loss from the ocean, and unus ually large increases in both latent and

PAGE 129

109 sensible heat losses (Figure 64) resulting in an increase in th e net heat loss (Figure 62). The latent and sensible heat losses are partia lly a result of the increased wind speed, but that alone cannot account for the large va riation. Air temperature decreases by 7oC from ~25oC to a minimum of 18oC, which also increases the sensible heat flux. Relative humidity decreases substantially from about 78% to less than 60%, which contributes to the increase in the latent heat flux. These observations suggest that the passing of an extratropical front over this region, with a dr ier, cooler air mass, results in increased evaporation and cooling, which in turn resu lts in cooler SSTs. Some of these sudden decreases in water temperatur e are accompanied by a rapid lo ss of heat from the ocean, so for example, during this October even t the net surface heat flux approached W m-2. Heat flux in this case accounts for 69% of the large temperature cooling during the event; ocean dynamics accounts for the other 31%. We see that tropical storms do not necessarily have the largest impact on WFS ocean-atmosphere fluxes. There are other large cha nges in the vertically av eraged temperature during October, especially around the 17th and 27th, which cannot be accounted for by passing storms or large heat flux changes, but are a consequence of ocean dynamics. Observations show that, following periods of sustained northerly winds, strong southward sub-surface currents advect cooler wate r into the region from the north. The influence of ocean dynamics and heat flux on the depth-averaged temperature change differs during September and October. In September, the temperature cools by 0.96oC as a result of surface cooling (-1.52oC) offset by ocean dynamical warming (0.56oC). In October however, the combined cooling effects due to net heat flux (3.71oC) and ocean dynamics (-0.06oC) lead to a temperature change of .76oC. Constructive and destructive interference be tween ocean circulati on and surface heat flux influences on temperature can either accentu ate or mitigate water column temperature changes. The minimum temperature is usually in February of each year, except during 1999 when a warming of the water column at all locations occurred from mid-January to mid-February. This reversed the normal coo ling trend and ensured th at temperatures did not get as cold as in other years; at CM2 they remained above 20oC (Figure 59). This

PAGE 130

110 warming was partly attributable to very low wind speeds that resulted in smaller latent and sensible heat losses. The latent heat flux was close to zero for most of this time. This was also accompanied by extraordin arily high relative humidity, reaching supersaturation (maximum values of 3%) at NA2 and almost 100% re lative humidity at CM2 for a few days in both January and February. Although there are no in situ measurements of precipitation available, the salinity showed a large decrease (>1.5psu) at the surface during this time, coinciding with a period of heavy rainfall as seen in the Huffman et al. (2001) Global Precipitation Climatology Projects (GPCP) One-Degree Daily Precipitation Data Set. The coldest temperatures in 1998 at NA2 (minimum of 18oC) were in January, whereas at EC3 (minimum of 20oC) they were in March, showing that even though the moorings were only 8.8 km apart, water temperature variations exist on that scale. These data also reveal an interannual variability in ocean temperatures: the wintertime temperatures got increasingly co lder from 1998 to 2000. In 1998, at CM2 the subsurface temperatures were cooler, but the surface temperatures were warmer (>28.5oC) for a longer period of time than in subs equent years or at other locations. The CM2 data also showed interannual variability in the stratificati on, which decreased in magnitude and temporal extent from 1998 to 2000. 4.6 Conclusions In situ measurements of oceanic and at mospheric variables from the west Florida shelf from 1998-2001 are used to describe feat ures of the annual cycle of SST and heat flux. Generally, when the heat flux switches sign from negative to positive, SST is at its minimum and the water column begins to warm up, and conversely when the heat flux switches sign from positive to negative. Observations during the spring season tran sition show a series of synoptic-scale momentum and heat flux varia tions that result in stratification of the water column. These synoptic variations are most evident in shallow water where smaller fluxes are

PAGE 131

111 required to affect change. Model studies show similar behaviors (He and Weisberg 2002a). Analyses of the one-dimensional temper ature equation show how the relative importance between surface heat flux and ocean dynamics changes seasonally in controlling the water temperature. Observ ations from the end of the spring 2000 transition season show that the depth-averaged temperature change is predominantly due to surface heat flux, and the synoptic-scale va riability is primarily by ocean dynamics. During summer, this situation is reversed a nd the major influence on the depth-averaged temperature is ocean dynamics, with heat flux being responsible for the synoptic variability. Water temperature increases due to advection of warm water from the south, but is partially offset by synoptic-scale ocean cooling events when the net surface heat flux changes from positive to negative. Cooling and destratification of the wate r column characterizes the fall season. This occurs as a series of steplike decreases in the water temperature. Some of these are attributable to tropical storms but others are a consequence of surface heat flux change due to extratropical fronts, or ocean dynamics as cool water is advected into the region. The passage of the first tropical storm of th e season heralds the subsequent decline in water temperature: the storm-induced temper ature decrease coupled with the generally negative net heat flux during this time ensu res that the water column cannot warm up again. Although the largest momentum flux is associated with tropical storms, they do not necessarily have the largest impact on ocean-atmosphere heat fluxes. The largest observed change in surface heat flux (and water temperature decrease) on the WFS was in response to an extratropical weather syst em. As in spring, surface heat flux is the predominant driver of the fall water temperat ure transition, and synop tic scale variability is primarily due to ocean circulation dynamics. There is also evidence of interannual vari ability: the wintertim e temperatures got increasingly colder from 1998 to 2000, and the stratification was greatest and the subsurface temperatures were coldest in 1998 when compared to subsequent years. There was also an unusual warming in early 1999, which prevented the water

PAGE 132

112 temperatures from getting as cold as observed in other years. The data suggest that this was due to low winds and minimal latent heat loss. Heat fluxes from different climatologies differ greatly (chapter 2). The NCEP fields are useful in providi ng a large-scale picture, but they are unable to reproduce important regions of spatial flux variability s hown by in situ and sate llite measurements. Driving ocean models with the reanalysis fiel ds, without flux corrections, fails to produce observed seasonally varying features on the WFS. Reconciling these differences and their impacts on regional climate variability studies provides challenges to coupled ocean-atmosphere models and their supporting observing systems.

PAGE 133

113 Chapter 5 The relative importance of surface heat flux and convergence of heat flux by ocean circulation in cont rolling ocean temperature 5.1 Abstract In situ data are used in an attempt to determine the relative importance of surface heat fluxes and heat flux convergence by ocean circulation in controlling water temperature on the WFS throughout the year The models employed are the quasianalytical one dimensional PWP and the numeri cal one and three dimensional versions of POM. Additional experiments are conducted to investigate the importa nce of cool skin, warm layer effects, rain sensible heat flux and moisture flux on PWP model results. Incorporation of a bottom re flection term improves the results in many cases. The 3D POM forced by Eta Data Assimilation System (EDAS) winds and a heat flux, relaxed to observed SST, produces better temperature fiel ds in the transitional seasons than in summer or winter. Despite experimental de sign limitations, the 3D POM results are improved when observed surface fluxes are used to force the model. Results show that in spring and summer, surface forcing is a criti cal determinant of the temperature field, however the inclusion of ocean dynamics, even during a season when surface heat fluxes dominate, are important on synoptic timescales. In autumn and winter, the surface fluxes and ocean dynamics are both required in determining the temperature field, however surface fluxes are of primary importance duri ng the passage of tropical storms or extratropical fronts. Although the relative importan ce of ocean circulation and surface fluxes vary in controlling water temperature throughout the year, both are necessary. Without in situ observations of surface fluxes to constrai n atmospheric models, models of the coastal ocean will be unable to accu rately reproduce ocean temperature variability unless other techniques such as heat flux correction by da ta assimilation of sea surface temperature are used. It is recognized that modeled surf ace heat flux will never be perfect and hence

PAGE 134

114 data assimilation will always be necessary; neve rtheless a goal should be to minimize that requirement. 5.2 Introduction A component common to atmospheric and oceanic studies is sea surface temperature. Atmospheric models use sea surface temperature as a boundary condition and to determine heat exchange between the ocean and atmosphere. Ocean models use water temperature to determine the density field and hence the circulation. Therefore knowledge of the controls on ocean temperatur e is one of the kernels of information needed to predict climate and weather. This chapter investigates the relative importance of surface heat flux and heat flux convergence by ocean circulation in controlling ocean temperatures on the WFS throughout the year. Observed water column temperature is simulated during different times of the year using observed fluxes and models. There are three goals. The first is to use the onedimensional Price, Weller and Pinkel (PWP; Price et al. 1986) mixed layer model to: (a) assess how well a one-dimensiona l model can account for temperature variations with depth; and (b) use the one-dimensional model discrepancies to refine the net surface heat flux. The second goal is to use the one-dime nsional version of the Princeton Ocean Model (POM1d; Blumberg and Mellor 1987) to see how a more sophisticated mixing parameterization affects the results. Fina lly, the three-dimensional Princeton Ocean Model (POM3d), adapted for the WFS (He and Weisberg 2002a), is used to see if improvements can be made to the modeled water temperature by adding the convergence of heat flux by the ocean circulation. Howe ver it should be noted that restrictive assumptions limited the success of achieving this third goal. In situ temperature and salinity, used to initialize the models and for model output comparison, and meteorological data, used to calculate the surface flux model forcing, are described in the next section. There are various factors that ma y affect the surface fluxes and some background information on thes e is provided in se ction 5.4. Details of the PWP, the POM1d, and the POM3d adapte d for the WFS (Blumberg and Mellor 1987;

PAGE 135

115 He and Weisberg, 2002a) are in section 5.5. Mode l results and in situ temperature field comparisons are given in section 5.6, which al so includes a brief description of the wind and ocean circulation on the WFS. Section 5.7 contains a discussion. 5.3 Data for Model Forcing, Init ialization, and Verification Surface fluxes of heat and momentum for January, May, June, July, September, October, and November 2000 are calculated usin g hourly averaged in situ meteorological data from the EC3 (Figure 65) and NA2 (F igure 66) moorings and the COARE 3.0 flux algorithm (Fairall et al. 2003). These months represent differe nt phases of the seasonal cycle. The fluxes provide surface forcing for th e three models. Hourly averaged in situ temperature and salinity (T/S) observations at discrete depths, from each mooring, are interpolated onto a 1m vertical grid, and th ese data are used for model initialization and verification. The PWP and POM1d models ar e initialized with the first T/S profile from each month. Unlike the one-dimensional models, the POM3d model requires a background gradient field in order to produce a dynamica lly consistent ocean circulation. Therefore the initial T/S profile for each month, when the POM3d model is used, differs from the observed T/S profile. This is discussed fu rther in section 5.5.3. Along and across shelf currents are produced by ro tating the hourly nor th and east curr ent vectors by 27o, respectively.

PAGE 136

116 0 W mLatent Heat Flux 0 W mSensible Heat Flux 0 500 1000 W mNet Shortwave 0 W mNet Longwave 0 1000 W mNet Radiation 0 1 N mEast Wind Stress J F M A M J J A S O N D 0 1 N mNorth Wind Stress Figure 65. Surface fluxes of heat and momentum in 2000 from EC3.

PAGE 137

117 0 W mLatent Heat Flux 0 W mSensible Heat Flux 0 500 1000 W mNet Shortwave 0 W mNet Longwave 0 1000 W mNet Radiation 0 1 N mEast Wind Stress J F M A M J J A S O N D 0 1 N mNorth Wind Stress Figure 66. Surface fluxes of heat and momentum in 2000 from NA2.

PAGE 138

118 5.4 Surface Flux Modifications 5.4.1 Cool-Skin, Warm-Layer effects Surface flux parameterizations are ideally based on temperature at the oceanatmosphere interface; however, observed sea surface temperature is measured approximately 1m below the surface. This introduces an error into the surface flux calculations. The correct surface temperature can be calculated using: Ts = Tm (z) Tc + Tw (z) (5.1) where Tm (z) is the in situ SST measurement at depth z, Tc is the cool-skin correction and Tw is the warm-layer correction (Fairall et al. 1996). Cooling due to net longwave, sensible heat, and latent heat loss from the ocean creates a cool-ski n layer that occurs within a few millimeters of the ocean-atmos phere interface (Saunders 1967). Shortwave radiation may result in a warmer surface laye r overlaying a cooler mixed layer. In this case, the upper ocean is stra tified by solar radiation to th e extent that wind-induced mixing is inhibited. The observed warm-layer e ffect can amount to surface temperatures being as much as 3oC warmer than the mixed layer temp erature (Fairall et al. 1996). The cool-skin and warm-layer effects are com puted using the COARE 3.0 algorithm (Fairall et al. 2003). The modified SST on average is 0.31oC lower than in situ SST with a standard deviation of 0.12oC at EC3 and 0.09oC at NA2 in 2000 (Figure 67). The greatest variability is in fall and winter.

PAGE 139

119 .5 0 0.5 1 o CEC3 SSTCSSST Mean: 0.31oC S.D.: 0.12oC J F M A M J J A S O N D .5 0 0.5 1 o CNA2 SSTCSSST Mean: 0.31oC S.D.: 0.09oC Figure 67. Difference between in situ SST and cool skin, warm layer corrected SST at EC3 (top) and NA2 (bottom) during 2000. 5.4.2 Rain sensible heat flux An additional source of sensible heat flux may occur during precipitation. If rain drop and sea surface temperatures differ there is an exchange of sensible heat (Gosnell et al. 1995). The rain sensible heat flux, Qrsh, may be calculated using: Qrsh = cw R (To Tr) (5.2) where cw is the specific heat of water ( 4186 J/kg/K), R is the rain rate, To is the bulk SST, and Tr is the mean temperature of rain at the surface (Gosnell et al. 1995). In situ precipitation measurements are not always av ailable, so NCEP da ily mean precipitation data are used for consistency. The rain sensible heat flux (Figure 68) is calculated in the

PAGE 140

120 COARE 3.0 algorithm, which uses an empirical formula. The greatest rain sensible heat loss, over 10W m-2, is associated with Hurricane Gordon. Increased precipitation in summer and early autumn (Figur e 69) may result in over 5W m-2 rain sensible heat loss. 0 5 W mEC3 rain sensible heat flux J F M A M J J A S O N D 0 5 W mNA2 rain sensible heat flux Figure 68. Rain sensible heat flux at EC3 (top) and NA2 (bottom) during 2000. 5.4.3 Moisture flux In addition to heat and momentum fluxes, the moisture flux is also an important component of the surface fl uxes and alters the buoyancy. The evaporation, E, is calculated using E = Qlh / ( w Le) (5.3) where Qlh is the latent heat flux, w (1025 kg m-3) is the density of seawater and Le (2.5 x 106 W m-2) is the heat of vaporization. The cal culated evaporation, NCEP daily mean precipitation, and EMP during 2000 are s hown in Figure 69. High amounts of water

PAGE 141

121 vapor in summer result in low evaporation. Larg e evaporation events are associated with tropical and extra-tropical systems in fall a nd winter; for example, Hurricane Gordon in September and the passage of an extra-tropic al front in October. Daily local convective activity and tropical storms produce higher pr ecipitation in summer and early fall; the highest single precipitation event in the 2000 record is Hurricane Gordon. 0 2 10 m sEC3 Evaporation 0 2 4 6 10 m sEC3 Precipitation 0 5 10 m sEC3 EMP 0 2 10 m sNA2 Evaporation 0 2 4 6 10 m sNA2 Precipitation J F M A M J J A S O N D 0 5 10 m sNA2 EMP Figure 69. Evaporation, NCEP daily mean precipitation, and EMP during 2000 at EC3 (top three panels) and NA2 (bottom three panels).

PAGE 142

122 5.4.4 Waves Ocean waves at the ocean-atmosphere boundary play an important role in modifying the wind profile and hence the momentum flux (e.g., Large et al. 1995). The moorings did not have a wave instrument on board so wave observations were not available. However, wave data was availa ble for September, October and November 2000 from the closest NDBC buoy located on the 50m isobath at 28.51oN, 84.51oW. This was approximately 215km from NA2 and EC3, so it was unlikely to be an accurate representation of the waves at the mooring locations. A test was done to investigate how including wave effects would impact the model results by including wave height and wave period in the COARE 3.0 algorithm. Ther e was very little difference between these model results and previous model results so th e effect of waves will not be shown. It is expected that the impact will be greater during autumn and winter, however, until wave data is available at the mooring location it is difficult to determine this. 5.5 Models 5.5.1 The Price, Weller and Pinkel one -dimensional mixed layer model The PWP model (Price et al. 1986) is a one-dimensional numerical model that simulates the upper ocean diurnal response to a given set of input conditions. It is initialized using a T/S profile and is for ced by surface heat, moisture, and momentum fluxes. A Richardson number mixing scheme is employed. The model determines the absorption of incident solar insolation in the water column using the Paulson and Simpson (1977) exponential parameterization described in secti on 1.4.2 (equation 1.7). Basic surface fluxes, calculated using COARE 3.0 without the cool-skin, warmlayer effects, evaporation-minus-precipitati on (EMP), and rain sensible heat flux, are used to force the PWP. A vertical grid reso lution of 0.5m and a time step of 1800s were used. The model results are compared to the observed temperature field to identify time

PAGE 143

123 periods when surface forcing dominates wit hout associated ocean advection. However, being cognizant that surface fluxes also have er rors, the model is then used to try to identify potential deficiencies and to f ine-tune the surface fluxes. The surface flux modifications (section 5.4) ar e applied to the basic flux fo rcing and model results are compared to observations to determine which modifications improve the output. The PWP model is run for water type I, IA, IB, and II (section 1.4) and the results are compared to observations to see which water type best reflects the water at the mooring locations. The results show that th e closest are type IA and IB. Type IB produces temperatures closer to the observations and is us ed in all model runs unless otherwise indicated. More advanced satellite algorithms also show that WFS water clarity is predominantly type IB (Enfield and Lee 2005). In addition to altering the fl uxes, one modification is also made to the model to adapt it for a coastal ocean environment. Th e PWP model already in cludes an exponential absorption of the incident shortwave radiation an d any radiation that is present at the base of the mixed-layer is entrained into the d eeper ocean and is lost from the system. However in coastal oceans with shallow shelve s, such as the WFS, radiation is absorbed and reflected by the ocean floor. The amount reflected depends on three parameters: water turbidity, type of ocean floor, and de pth of the water column. Given a fair-tomedium clarity water type (section 1.4.2), a reflective sandy bottom, and a shallow water depth (25-30 m), an assumption is made th at no absorption occurs at the bottom and 100% of the radiation that re aches the bottom is reflected. Figure 13 (chapter 1) shows the absorption profile of in cident radiation of 1000Wm-2 to depths of 50m. The NA2 mooring is at the 25m isobath, so at the surface the amount lost (Qpen in equation 1.7) is about 30Wm-2. This reflected radiation also adds a source of heating to the water column, which is not accounted for in the PWP model because this model continues mixing with the deeper ocean until the stability conditions are satisfied. The model is modified to incorporate the attenuated reflected heating te rm in each grid cell before the stability criteria are calculated.

PAGE 144

124 5.5.2 The one-dimensional Princeton Ocean Model The POM1d was adapted from the thr ee-dimensional Princeton Ocean Model (Blumberg and Mellor 1987). Th e principal difference between the POM1d and the onedimensional PWP is the mixing scheme; the POM1d contains a level 2.5 turbulence closure scheme (Mellor and Yamada 1974, 1982) As a sigma coordinate model, the vertical coordinate is scaled relative to the water depth. Th ere are 54 sigma levels and the time step is 900s. The first in situ T/S prof ile for each month is used to initialize the model. The derived surface fluxes are used as the model forcing. This model is also modified, as outlined in section 5.5.1, to include the ocean bottom reflection. 5.5.3 The three-dimensional Princeton Ocean Model The POM3d, a coastal ocean circulation model (Blumberg and Mellor 1987), was adapted for the WFS (He and Weisberg 2002a). It contains a second moment turbulence closure scheme (Mellor and Yamada 1974, 1982). Large variations in topography can be handled with the sigma coordinate, whic h makes adaptation to different coastal bathymetries easier. The sigma coordinate has 21 layers. The horizontal grid is in curvilinear orthogonal coor dinates; the smallest grid spacing, closest to the coast, is 2km, and the largest grid spacing, farthest from th e coast, is 6km. The external and internal time steps are 18s and 900s, respectively. There are two limitations in the POM3d experiments. Unlike the one-dimensional models, the 3D model requires background gradie nt fields in order to produce and be dynamically consistent with the ocean circula tion. The model is run for two months prior to each case study in order to set up these gradient fields, so although the initial T/S profiles for the 3D model runs differ from the observed T/S profiles they are dynamically consistent with the circulation. The surface forcing for the runs that generate the gradient fields are EDAS winds and heat fluxes with the heat flux corrected by relaxing the model

PAGE 145

125 SST to an observed SST product produced from satellite data by op timal interpolation. The OI SST is produced as described in He et al. (2003). The initial temperature difference makes it difficult to evaluate the performance of the model so an attempt to compensate for this discrepancy is made by looking at the temper ature tendency of the model and in situ temperature fields. The diffe rence between the depth-averaged initial in situ and model temperature profiles is applied to the model output, and then the temperature change with time in the m odel and observations are compared. This difference is greater in summer and winter than spring or fa ll (Figure 70) suggesting that the WFS POM3d model temperature fields, forced by EDAS winds and heat fluxes relaxed to observed SST, are closer to observations in the transitional seasons. J F M A M J J A S O N D 1.5 0.5 0 0.5 1 1.5 2 Figure 70. The differences between the depth-av eraged initial in si tu and POM3d model (forced by EDAS winds and relaxed heat flux) temperatures in 2000, at EC3 (star) and NA2 (triangle). The second limitation is the scarcity of in situ surface flux data. The POM3d model surface forcing field needs to en compass the model domain. However the oC

PAGE 146

126 objective of this study is to determine how well it performs when forced with surface fluxes derived from in situ observations; ther efore forcing it with modified reanalysis heat fluxes, relaxed to the observed SST, will defeat the purpose. The surface forcing for the case studies are obtained fr om observations at a single location, and are taken to be uniform at all points in the model domain. This artificial uniform surface forcing will inherently introduce some circulation irregularities. This model is also modified, as outli ned in section 5.5.1, to include the ocean bottom reflection. 5.6 Results The model experiments are conducted for di fferent phases of the seasonal cycle. Spring (May) and fall (October, and Novembe r) are transitional seasons when the net heat flux tendency (dQ/dt) is increasing or decreasing, respectively. In winter (January) and summer (June and July), the net heat flux tendency is steady. To provide direct comparison with observations, the model temp erature fields do not extend to the ocean floor. Water type IB is used is all m odel experiments unless otherwise indicated. 5.6.1 The Price, Weller and Pinkel one -dimensional mixed layer model Results from the PWP model experime nts are shown in Figures 71-80. Each figure has the same format and represents the temperature field for one month at one location. The top panel shows observed temperat ure. Subsequent panels show the model output temperature under different conditions as follows. The second panel shows results from forcing the model with the basic fluxes. These include the momentum and heat flux only. The heat flux is calculated as given in equation 1.6 (Chapter 1), without the Qpen term. The third panel shows results from the model forced with fluxes that are modified to include a cool skin, warm layer SST correc tion. This is included in the fluxes used to produce the remaining panels. The fourth panel has results fro m two simultaneous modifications. A bottom reflection term is adde d to the model as outlined in the previous

PAGE 147

127 section. Although this leads to an additional heating of the wa ter column, there is a heat loss from the surface as the radiatio n reflected from the ocean floor (Qpen) reaches the airsea interface. So Qpen is included in the net heat fl ux (equation 1.6). These modifications are also included in model experiments us ed to produce the subsequent panels. The results in the fifth panel are from the mode l forced by the net heat flux modified to include the rain sensible heat term. This is a small heat loss from the ocean (Figure 68). The moisture flux is added to the surface forc ing, and the results are shown in panel six. The last two panels in each figure show the model and surface fluxes (Qpen) modified for water type IA and II, respectively. During January water te mperature cools by ~3oC (Figure 71). Two periods of warming occur in the first two weeks, followed by a 1oC decrease in temperature over two days on 14-15 January. A downwelling even t occurs on 23-24 Ja nuary, with cold water at depth and warmer surface waters. Th e water warms up slightly at the end of January. All model results s how temperature cooling during January. They also all capture the cold event in the middle of th e month. The output using basic fluxes (second panel) is colder than observations and s ubsequent model outputs; however all model results are cooler than the observations. The addition of ra in sensible heat flux or moisture flux does not change the model re sults appreciably from the bottom reflection case. The bottom reflection model output produce s results similar to the observations in type IB water, although the type II model r un (final panel) shows a further improvement. The inability of the model to fully produce the structure of the observed temperature field, including the downwelling on the 23-24 Janu ary and the warming at the end of the month indicates that there are influences other than one-dimensional thermodynamics that affect the temperature. During May the water temperature is warm ing (Figures 72 and 73). Initially the water column is well-mixed. Stratification occurs between 10-18 May and 25-30 May as the heat flux increases and wind stress d ecreases (section 4.5). Between May 15-17 and May 28-30, upwelling at depth and surface hea ting inhibit mixing and create a strong thermocline. Following this on the 18 and 31 May, respectively, a decrease in heat flux and mixing of colder waters from depth (i n response to increased momentum flux)

PAGE 148

128 combine to rapidly erode the stratification a nd leave the water column well-mixed again. Values at NA2 are warmer than EC3 in the observed and modeled temperatures. All model results show the synoptic -scale stratification. Upwelling is not in the model output at either location but the st rong thermocline and subsequently well-mixed water column are in the results, suggesting that fluxes are dominant in producing the observed temperature field at this time of year. With the exception of the water type II experiments, all model results are cooler than observations. At NA2, the model temperatures at the surface of the type II results are warmer than observations. Unrealistic stratification at depth is seen in all model results and is especially pr onounced in the type II results, suggesting that inhibited mixing at depth is increasing the model output temperature closer to the surface. This unr ealistic stratification is because local wind mixing is insufficient to mix the heat down to the bottom and the bottom Ekman layer is under-represented in the PWP. This lower stratif ication is ameliorated if type IA water is used, and more realistic isotherms are produced. The diurnal cycle from the surface heat fl ux is clearly observed in June (Figure 74). Surface heating in conjunc tion with upwelling at depth forms a thermocline at the beginning of the month, which mixes as heat flux decreases. This is the same process observed in May and is part of the WFS transition from winter to summer. Model results show weaker stratification and are cooler than observations Using the cool skin, warm layer correction, botto m reflection and Qpen, the model temperature is closer to observations with type II wate r, although the thermocline is considerably shallower than observed at the end of the month. The structur e of the temperature field is better in the type IB output, but heat is dispersed over a greater depth and the te mperature is cooler. During July the water is warming (Fi gure 75) and three synoptic-scale warm events occur. The models capture these to varying degrees, but ove rall the temperatures are colder than observed. As seen in previ ous months, the two cases that use the cool skin, warm layer SST correction, bottom reflection, Qpen, and water type IB or type II produce the closest model outpu t to observations; the trade off is between an improved temperature profile or improved temperatures.

PAGE 149

129 The water temperature cools in Septembe r (Figures 76 and 77). On September 17 the entire water column cools suddenly by about 0.5oC in response to Hurricane Gordon. In all cases the model output reproduces this sudden temperature drop and this is clearly a surface flux related cooling. However the model results, forced with basic fluxes, show cooling about 10 days earlier, which is not seen in the observations. Also, following the hurricane the model temperature decreases more than observations, especially at EC3. An observed surface warming on 24-29 September is partially captured in all model results showing that surface fluxes are responsible. This is verified by the t ype II model output at NA2, which is remarkably similar to the obser vations, both in temper ature and structure. However, the temperature field prior to 17 Se ptember in the type II case is not as well represented, suggesting that the wate r type changes after the hurricane. Temperatures decrease during October (Figures 78 and 79). A sharp drop in temperature of over 1oC throughout the water column on 9-10 October is due to the passage of an extra-tropical system. This is reproduced in all model results and is primarily a response to surface fluxes. Features such as an upwelling on 20-22 October and a cooling on 23-26 October, punctuat ed by a small warming on the October 24 followed by rapid cooling on October 26, are no t fully captured in the model results. Although the model output from the first part of the month improves when forced by fluxes with a cool skin warm laye r correction, bottom reflection and Qpen, the temperatures in the latter half are warmer than observations, especially at NA2, which is contrary to all other months. This suggests th at water temperatures in the last two weeks of October are strongly dependent on ocean circulation dynamics as opposed to surface fluxes. Temperatures decrease during November (Figure 80). There are two large warm events on 4-12 November and 26-27 November the entire water column warms by 0.5oC on 13-14 November and cools by over 1oC on 21-23 November, and there is another warming on 25-27 November. All model temperat ure fields are cooler than observations. The only feature reproduced is the water co lumn cooling on 21-23 November; therefore ocean dynamics must play a large role in de termining the observed temperature field at this time.

PAGE 150

130 The PWP experiments show that modifyi ng the fluxes to include a cool skin, warm layer correction and a Qpen term, and modifying the model to reflection from the ocean floor improved the one-dimensional model results. There is also considerable sensitivity on water type, which varies depe nding on the location and time of year. Not all observed features can be accounted for by this one-dimensional approach, especially in fall, which suggests that ocean circulation is important. The fluxes modified to include a cool skin, warm layer SST correction will be used for the base runs in the POM1d and POM3d experiments.

PAGE 151

131 5 1 0 1 5 2 0 2 5 Observed Temperature 5 1 0 1 5 2 0 2 5 PWP Temperature: Base Run 5 1 0 1 5 2 0 2 5 PWP Temperature: Cool Skin 5 1 0 1 5 2 0 2 5 PWP Temperature: Bottom Reflection 5 1 0 1 5 2 0 2 5 PWP Temperature: Rain Sensible Heat 5 1 0 1 5 2 0 2 5 PWP Temperature: EMP 5 1 0 1 5 2 0 2 5 PWP Temperature: Type 1A 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 1 19 2 20 21 2 22 3 23 24 25 26 2 27 2 28 29 0 30 1 31 5 1 0 1 5 2 0 2 5 PWP Temperature: Type 2 17 18 19 20 21 22 Figure 71. EC3 January temperatures (oC) from observations and PWP model experiments.

PAGE 152

132 5 1 0 1 5 2 0 2 5 Observed Temperature 5 1 0 1 5 2 0 2 5 PWP Temperature: Base Run 5 1 0 1 5 2 0 2 5 PWP Temperature: Cool Skin 5 1 0 1 5 2 0 2 5 PWP Temperature: Bottom Reflection 5 1 0 1 5 2 0 2 5 PWP Temperature: Rain Sensible Heat 5 1 0 1 5 2 0 2 5 PWP Temperature: EMP 5 1 0 1 5 2 0 2 5 PWP Temperature: Type IA 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 8 18 9 19 0 20 21 22 2 23 2 24 2 25 2 26 27 28 29 30 31 5 1 0 1 5 2 0 2 5 PWP Temperature: Type 2 22 23 24 25 26 27 28 Figure 72. EC3 May temperatures (oC) from observations and PWP model experiments.

PAGE 153

133 5 1 0 1 5 2 0 2 5 Observed Temperature 5 1 0 1 5 2 0 2 5 PWP Temperature: Base Run 5 1 0 1 5 2 0 2 5 PWP Temperature: Cool Skin 5 1 0 1 5 2 0 2 5 PWP Temperature: Bottom Reflection 5 1 0 1 5 2 0 2 5 PWP Temperature: Rain Sensible Heat 5 1 0 1 5 2 0 2 5 PWP Temperature: EMP 5 1 0 1 5 2 0 2 5 PWP Temperature: Type 1A 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 1 19 2 20 21 2 22 3 23 24 25 26 2 27 2 28 29 0 30 1 31 5 1 0 1 5 2 0 2 5 PWP Temperature: Type 2 22 23 24 25 26 27 28 Figure 73. NA2 May temperatures (oC) from observations and PWP model experiments.

PAGE 154

134 5 1 0 1 5 2 0 2 5 Observed Temperature 5 1 0 1 5 2 0 2 5 PWP Temperature: Base Run 5 1 0 1 5 2 0 2 5 PWP Temperature: Cool Skin 5 1 0 1 5 2 0 2 5 PWP Temperature: Bottom Reflection 5 1 0 1 5 2 0 2 5 PWP Temperature: Rain Sensible Heat 5 1 0 1 5 2 0 2 5 PWP Temperature: EMP 5 1 0 1 5 2 0 2 5 PWP Temperature: Type 1A 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 1 19 20 2 21 22 2 23 4 24 25 6 26 27 28 29 30 5 1 0 1 5 2 0 2 5 PWP Temperature: Type 2 23 24 25 26 27 28 29 30 Figure 74. NA2 June temperatures (oC) from observations and PWP model experiments.

PAGE 155

135 5 1 0 1 5 2 0 2 5 Observed Temperature 5 1 0 1 5 2 0 2 5 PWP Temperature: Base Run 5 1 0 1 5 2 0 2 5 PWP Temperature: Cool Skin 5 1 0 1 5 2 0 2 5 PWP Temperature: Bottom Reflection 5 1 0 1 5 2 0 2 5 PWP Temperature: Rain Sensible Heat 5 1 0 1 5 2 0 2 5 PWP Temperature: EMP 5 1 0 1 5 2 0 2 5 PWP Temperature: Type 1A 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 2 20 2 21 2 22 2 23 2 24 25 6 26 7 27 8 28 9 29 30 31 5 1 0 1 5 2 0 2 5 PWP Temperature: Type 2 27 28 29 30 31 Figure 75. NA2 July temperatures (oC) from observations and PWP model experiments.

PAGE 156

136 5 1 0 1 5 2 0 2 5 Observed Temperature 5 1 0 1 5 2 0 2 5 PWP Temperature: Base Run 5 1 0 1 5 2 0 2 5 PWP Temperature: Cool Skin 5 1 0 1 5 2 0 2 5 PWP Temperature: Bottom Reflection 5 1 0 1 5 2 0 2 5 PWP Temperature: Rain Sensible Heat 5 1 0 1 5 2 0 2 5 PWP Temperature: EMP 5 1 0 1 5 2 0 2 5 PWP Temperature: Type IA 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1 17 8 18 19 2 20 21 22 2 23 24 25 26 27 28 29 30 5 1 0 1 5 2 0 2 5 PWP Temperature: Type 2 28 29 30 31 Figure 76. EC3 September temperatures (oC) from observations and PWP model experiments.

PAGE 157

137 5 1 0 1 5 2 0 2 5 Observed Temperature 5 1 0 1 5 2 0 2 5 PWP Temperature: Base Run 5 1 0 1 5 2 0 2 5 PWP Temperature: Cool Skin 5 1 0 1 5 2 0 2 5 PWP Temperature: Bottom Reflection 5 1 0 1 5 2 0 2 5 PWP Temperature: Rain Sensible Heat 5 1 0 1 5 2 0 2 5 PWP Temperature: EMP 5 1 0 1 5 2 0 2 5 PWP Temperature: Type 1A 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 7 17 18 9 19 20 1 21 22 3 23 24 5 25 2 26 7 27 2 28 9 29 3 30 5 1 0 1 5 2 0 2 5 PWP Temperature: Type 2 28 29 30 31 Figure 77. NA2 September temperatures (oC) from observations and PWP model experiments.

PAGE 158

138 5 1 0 1 5 2 0 2 5 Observed Temperature 5 1 0 1 5 2 0 2 5 PWP Temperature: Base Run 5 1 0 1 5 2 0 2 5 PWP Temperature: Cool Skin 5 1 0 1 5 2 0 2 5 PWP Temperature: Bottom Reflection 5 1 0 1 5 2 0 2 5 PWP Temperature: Rain Sensible Heat 5 1 0 1 5 2 0 2 5 PWP Temperature: EMP 5 1 0 1 5 2 0 2 5 PWP Temperature: Type IA 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 5 1 0 1 5 2 0 2 5 PWP Temperature: Type 2 23 24 25 26 27 28 29 30 31 Figure 78. EC3 October temperatures (oC) from observations and PWP model experiments.

PAGE 159

139 5 1 0 1 5 2 0 2 5 Observed Temperature 5 1 0 1 5 2 0 2 5 PWP Temperature: Base Run 5 1 0 1 5 2 0 2 5 PWP Temperature: Cool Skin 5 1 0 1 5 2 0 2 5 PWP Temperature: Bottom Reflection 5 1 0 1 5 2 0 2 5 PWP Temperature: Rain Sensible Heat 5 1 0 1 5 2 0 2 5 PWP Temperature: EMP 5 1 0 1 5 2 0 2 5 PWP Temperature: Type 1A 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 8 18 9 19 0 20 21 22 2 23 2 24 2 25 2 26 27 28 29 30 31 5 1 0 1 5 2 0 2 5 PWP Temperature: Type 2 23 24 25 26 27 28 29 30 31 Figure 79. NA2 October temperatures (oC) from observations and PWP model experiments.

PAGE 160

140 5 1 0 1 5 2 0 2 5 Observed Temperature 5 1 0 1 5 2 0 2 5 PWP Temperature: Base Run 5 1 0 1 5 2 0 2 5 PWP Temperature: Cool Skin 5 1 0 1 5 2 0 2 5 PWP Temperature: Bottom Reflection 5 1 0 1 5 2 0 2 5 PWP Temperature: Rain Sensible Heat 5 1 0 1 5 2 0 2 5 PWP Temperature: EMP 5 1 0 1 5 2 0 2 5 PWP Temperature: Type IA 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1 16 7 17 1 18 19 2 20 21 22 23 4 24 25 6 26 27 8 28 2 29 0 30 5 1 0 1 5 2 0 2 5 PWP Temperature: Type 2 20 21 22 23 24 25 26 Figure 80. EC3 November temperatures (oC) from observations and PWP model experiments.

PAGE 161

141 5.6.2 One-dimensional temperature balance To quantify the error in the surface heat fluxes used in the preceding section, the one-dimensional depth-averaged temper ature equation (5.4) was analyzed. NET p HQ dz t T c 0 (5.4) where T is the observed temperature, cp is the heat capacity of seawater (3998 J kg-1 K-1), is the density of seawater (1023.34 kg m-3), H is the water depth, and QNET is the net surface heat flux which ma y or may not include Qpen. Table 7 shows the value of the observed dz t T cp H 0, which indicates the net heat flux required (Qreq) to produce the monthly mean depth-averaged observed temperature and Qnet with varying surface fluxes: the second colu mn contains values of the basic net heat flux used to force the PWP base run ( Qbase); the third column c ontains values of the net heat flux with the cool skin, warm layer modified SST (Qcswl); and the fourth column contains values of the net heat flux with th e corrected SST and the penetrative radiation term (Qpen). The final three columns contain values of the residual between Qreq and Qbase, Qcswl, and Qpen, respectively. The residual between Qreq and the net heat fluxes include advective terms and errors; however these cannot be quantified in a one-dimensional case. A small residual value in one of the fina l three columns indicates that version of the heat flux best accounts for local temperature ch ange compared to the other versions of the heat flux. Comparing values from the last three co lumns, the smallest (absolute) residual value in almost all cases is between Qreq and Qcswl. The exceptions to this are in October when the smallest residual is found using Qbase, and at NA2 in July and September when the smallest residual is found using Qpen. This suggests that in mo st cases, the addition of the cool skin correction is be neficial. In October, when Qreq Qbase is smallest, the results from the PWP experiments could not reproduce the temperature field in the latter half of the month, indicating that horizont al advection is important.

PAGE 162

142 Previous studies by He and Weisberg (2002a, 2003b) using th e POM 3D model, adapted for the WFS, have been able to quant ify the contribution to the local temperature change due to advection during the spring a nd fall seasons. The contribution by heat flux from these model results are on the same orde r of magnitude as those obtained from the 1D temperature analysis in chapter 4, whic h are based on observations alone. Similarly, the contribution by the advectiv e terms is on the same order of magnitude as those given by the residuals in Table 7. Therefore it may be suggested that the differences between the observed temperature and the PWP mode l output is because of the missing threedimensional effects. Qreq Qbase Qcswl Qpen Qreq Qbase Qreq Qcswl Qreq Qpen Jan -74.44 -120.13 -107.63 -108.94 45.69 33.19 34.50 May 185.44 103.87 114.30 111.67 81.57 71.14 73.77 Sep -45.21 -79.93 -60.89 -62.76 34.73 15.68 17.55 Oct -178.78 -179.31 -156.75 -158.67 0.53 -22.02 -20.10 EC3 Nov -99.40 -174.00 -156.46 -157.79 74.59 57.05 58.38 May 167.23 129.92 139.01 134.09 37.32 28.22 33.14 Jun 101.13 68.08 81.63 76.94 33.06 19.50 24.20 Jul 31.43 22.13 36.99 32.99 9.29 -5.56 -1.55 Sep -37.94 -47.45 -30.51 -33.98 9.51 -7.43 -3.95 NA2 Oct -144.13 -131.81 -111.96 -115.44 -12.32 -32.17 -28.69 Table 7. Monthly mean heat flux (W/m2) needed to produce the observed temperature (Qreq), obtained from basic heat flux calculation (Qbase), applying a cool skin, warm layer correction (Qcswl), and the addition of penetrative radiation (Qpen). The monthly mean values from Table 7 do not give an indication of the synoptic variability in the one-dimensional temperatur e balance. The times during each month for each mooring when the 1-D temperature equation is well-balanced may by seen in figure

PAGE 163

143 81, which shows the depth-averaged local rate of change of temperat ure (solid line) and the surface heat flux, Qpen (dash line). 0 500 January (EC3)W/m2 0 200 400 600 May (EC3)W/m2 0 200 400 600 May (NA2)W/m2 0 200 June (NA2)W/m2 0 200 400 July (NA2)W/m2 0 200 September (EC3)W/m2 0 200 September (NA2)W/m2 0 500 October (EC3)W/m2 0 500 October (NA2)W/m2 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 0 500 November (EC3)W/m2 Figure 81. The depth-averaged local rate of change of temperat ure (solid) and the net heat flux (dash) computed using the cool skin co rrected basic heat flux with the penetrative radiation term for each mooring and month.

PAGE 164

144 Deviations between the solid and dash lines indicate that something in addition to heat flux is affecting the local water temp erature. These deviations occur on synoptic time scales in fall, winter, and spring. As in the case of the mont hly mean residuals, the value of these deviations are quantitatively comparable to numbers found for heat convergence by ocean circulation in the tw o model studies of the spring and fall transition season by He and Weisberg (2002a, 2003b). 5.6.3 The one-dimensional Princeton Ocean Model For each month, the POM1d model is run for two cases: a base flux case which includes the cool skin, warm layer corrected SST only, and a bottom reflection modified model with surface heat flux forcing th at includes the SST correction and a Qpen term. The purpose of running a second one-dimensional model is to determine how the model output changes with a different mixing scheme Not all results are shown here because overall they agree with PWP as follows. Th e POM1d model temperatures are generally cooler than observations. M odel outputs at NA2 are generally better than at EC3. The model output forced with cool skin corrected fluxes and with bottom reflection and Qpen shows an improvement over the base run. The greatest differences between the tw o mixing schemes are seen in months when surface heat fluxes are important and stra tification occurs. An example is shown in the EC3 May results (Figure 82) using the m odels modified to include bottom reflection and forced by fluxes that include the cool skin, warm layer correction and Qpen. The top panel is the observed temperature, the mi ddle panel is the PWP model output and the lower panel is the POM1d output There is a well-mixed water column in early May in the observations and PWP, whereas the POM 1d continues stratification at depth for almost a week. The surface heat flux is not warming the water column in POM1d as much as shown in the observations or PWP temperatures. However, there is a more disperse thermocline in the POM1d. The obs erved temperature field has the disperse thermocline shown in POM1d, but are have hi gher temperatures which are closer to

PAGE 165

145 PWP. Previous PWP model outputs show that th ese results may improve if type II waters are used instead of type IB. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 5 1 0 1 5 2 0 2 5 3 0 Observed Temperature 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1 17 18 9 19 20 21 2 22 23 24 25 2 26 27 8 28 29 3 30 31 5 1 0 1 5 2 0 2 5 3 0 PWP Temperature 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 5 1 0 1 5 2 0 2 5 3 0 1D POM Model Output 21 22 23 24 25 26 27 28 Figure 82. Observed (top), PWP model (middl e), and POM1d (bottom) temperature in May 2000 at EC3. A bottom boundary layer is not included in the one-dimensional PWP mixed layer model. The extent to which the Ekman laye r extends at the surface may be estimated by the Monin-Obukhov Length Scale, Lm:

PAGE 166

146 ) / (2 / 3wT kg Lm (5.5) where is the stress, is the density of water (~1000 kg/m3), is the coefficient of thermal expansion (~2 x10-4 K-1), k is the von Karman constant (0.4), g is gravity (~10m/s) and wTis the heat flux. This length give s the depth at which the buoyancy inhibition equals the shear produc tion of turbulence. We can estimate this for the WFS in cases of stratification and de stratification. From Figure 60 (c hapter 4) if the wind stress during stratification in spri ng is approximately 0.1 N/m2 and the daily averaged heat flux is 100W/m2, the Monin-Obukhov Length scale at th e top of the water column is approximately 4m. Similarly from Figure 62, if the wind stress during destratification in autumn is approximately 0.2 N/m2 and the heat flux is 100 W/m2, the Monin-Obukhov Length scale is 5.5m. Similar arguments can be made for the bottom mixed layer based on a given buoyancy gradient and work by bottom stress. Alternatively, the surface or bottom Ekman boundary layer may be estimated based on previous observations or from model re sults. Observations of turning of currents with depth indicate the influen ce of surface wind stress to depths of about 10m at the 50m isobath (Weisberg et al. 2000). From a stratif ied model, the ageost rophic portion of the momentum balance (i.e. the portion due to fr iction) gives a surface Ekman layer of about 5m at the 20m isobath (Weisberg et al. 2001). During upwelling, modeled values of the vertical eddy coefficient (Li and Weisberg 1999) result in Ekman layer depths of approximately 10-20m at the 25 m isobath. The Ekman boundary layers (bottom and surface) vary depending on the stratification and ocean dynamics. The one-dimensional model experiments cover many time periods and should theref ore encompass a range of bounda ry layer depths, including cases where the bottom and surface Ekman laye rs overlap but these effects were not accounted for when the one-dimensional experiments were conducted.

PAGE 167

147 5.6.4 The three-dimensional Princeton Ocean Model For each month, the POM3d model is run for two cases: a base flux case which includes the cool skin, warm layer corrected SST only, and a bottom reflection modified model with surface heat flux forcing th at includes the SST correction and a Qpen term. The purpose of running a three-dimensional mode l is to determine if the inclusion of advection accounts for variations in the obser ved temperature field that are not due to surface fluxes, as suggested by the one-dimensional model output. Two major problems are identified with th ese three-dimensional experiments; the initial T/S profiles for the 3D model runs differ from the observed T/S profiles and the surface forcing is uniform at all points in the model domain. An attempt to rectify the first problem is made by removing the differe nce between the model and observed initial temperature from the model output and comp aring the temperature tendency between the two (section 5.5.3). The ocean circulation irre gularities introduced by the second problem remain in the results. The results from th e bottom reflection case only will be shown. The figures in this section have the same form at. The top panel are the winds, the second and third panels are the across a nd along shelf observed currents, the fourth panel is the PWP model result, the fifth panel is the POM1d m odel result, and the sixth panel is the POM3d model result. In May the surface fluxes dominate (Fi gure 83), and there are synoptic-scale stratifications in the observed temperatur e field. There are upwe lling events on 16-18 May and again 28-30 May that corresponds w ith northerly winds, as indicated by the upward bending isotherms and colder water at depth. Neither event is seen in the onedimensional models because it is a result of ocean circulation. The POM3d model does show upwelling at both these times. The upwe lling at the end of the month is stronger in the model than in the observations. The wa ter temperature is warmer than observed, which may be a consequence of using the same forcing over the entire model domain. Despite this, the POM3d result suggests th at the inclusion of ocean dynamics, even during a season when heat fluxes dominate SST are important. This importance pertains to both the synoptic and longer time scales. Th e synoptic scale effect is evident in the

PAGE 168

148 May 16-18 response. The longer time scale effect is evident by the model being too warm at the end of May. In nature (and in the pres ence of spatially varyi ng heat flux) advection of cooler water from the north w ould have mitigated the over-warming. In June, the initial temperature profile in the model is colder and highly stratified compared to observations (Figure 84). This stra tification persists at depth for almost three weeks, however after the firs t two weeks in mid-June, the water column becomes better mixed and the results rapidly improve. Once the initial stratification has been eliminated, the POM3d model results are an improvement over the one-dimensional model fields. Additionally, this figure sugge sts that forcing the POM3d model with observed surface fluxes in the summer rather than EDAS based fluxes may improve the model performance. In October, northerly winds persist throughout the month and are strong during the passage of an extra-tropical system on 9 October, resulting in a well-mixed water column and temperature decrease due to a la rge heat flux loss from the ocean into the atmosphere (figure 85). This event is well simulated by all models. On 18-23 October the water column is stratified with warmer water at the surface and upwelling-induced cooling at depth. The POM3d model results pr oduce a highly stratif ied water column at this time, with an upwelling event and a ssociated cooler wate rs at depth. Model upwelling is stronger than observations, probabl y due to artificially enhanced circulation. The 1-D models continue to remain well-mixe d, and the upwelling is not seen. There is a small warming of the water column on 24 October, which is not seen in the onedimensional models, however it is in the POM 3d results, suggesting that this warming is a result of ocean circulation. The along shel f currents show upwelling which results in warm water being brought to the surface from depth, as explained by He and Weisberg (2003b). The negative net surface heat flux cool s the shallower water closer to the coast faster than farther off-shore, therefore upw elled water may be warmer in the fall. The temperature field from the POM3d model is much colder than observations from 24-31 October and the isotherms are unrealistic comp ared to observations. The entire shelf is being forced with the same negative heat flux during this time and the persistent northerly winds are advecting cold water from the nor th, as indicated by the strong southward

PAGE 169

149 current. Although the models and observations in October are not as cl ose as in May (for example), they do suggest the possibility for improvements of the temperature field when the three-dimensional ocean circ ulation on the WFS is included.

PAGE 170

150 1 7 14 21 28 200 0 200 5 10 15 20 25 Across Shelf Component 5 10 15 20 25 Along Shelf Component 1 7 14 21 28 0 4 8 5 10 15 20 25 Observed Temperature 5 10 15 20 25 PWP with bottom reflection 5 10 15 20 25 POM 1D with bottom reflection 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 5 10 15 20 25 POM 3D with bottom reflection 21 22 23 24 25 26 27 28 29 (a) (b) (c) (d) (e) (f) (g) (h) Figure 83. For May 2000, the 36-hour lo wpass filtered net heat flux (W/m2; a), winds (m/s; b), along and across shelf current (c m/s, contour interval 10cm/s; c and d), temperature (oC) from observations (e), PWP model results (f), POM1d model results (g), POM3d model results (h). All model results in clude bottom reflection and are forced with surface fluxes modified to include a c ool skin warm layer correction and Qpen. Qnet (W/m2) m/s Depth (m) Depth (m) Depth (m) Depth (m) Depth (m) Depth (m) Winds

PAGE 171

151 Depth (m) Depth (m) 1 7 14 21 28 200 0 200 5 10 15 20 25 Across Shelf Component 5 10 15 20 25 Along Shelf Component 1 7 14 21 28 0 4 8 5 10 15 20 25 Observed Temperature 5 10 15 20 25 PWP with bottom reflection 5 10 15 20 25 POM 1D with bottom reflection 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 5 10 15 20 25 POM 3D with bottom reflection 22 23 24 25 26 27 28 29 30 (a) (b) (c) (d) (e) (f) (g) (h) Figure 84. For June 2000, the 36-hour lowpass filtered net heat flux (W/m2; a), winds (m/s; b), and along and across shelf current (c m/s, contour interval 10cm/s; c and d), temperature (oC) from observations (e), PWP model results (f), POM1d model results (g), POM3d model results (h). All model results in clude bottom reflection and are forced with surface fluxes modified to include a c ool skin warm layer correction and Qpen. (W/m2) m/s Qnet Winds Depth (m) Depth (m) Depth (m) Depth (m)

PAGE 172

152 Depth (m) Depth (m) Depth (m) Depth (m) Depth (m) Depth (m) 1 7 14 21 28 800 600 400 200 0 200 5 10 15 20 25 Across Shelf Component 5 10 15 20 25 Along Shelf Component 1 7 14 21 0 4 5 10 15 20 25 Observed Temperature 5 10 15 20 25 PWP with bottom reflection POM 1D with bottom reflection 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 5 10 15 20 25 POM 3D with bottom reflection 18 20 22 24 26 28 30 (a) (b) (c) (d) (e) (f) (g) (h) Figure 85. For October 2000, the 36-hour lowpass filtered net heat flux (W/m2; a), winds (m/s; b), and along and across shelf current (c m/s, contour interval 10cm/s; c and d), temperature (oC) from observations (e), PWP model results (f), POM1d model results (g), POM3d model results (h). All model results in clude bottom reflection and are forced with surface fluxes modified to include a c ool skin warm layer correction and Qpen. m/s (W/m2) Qnet Winds

PAGE 173

153 5.7 Discussion In situ data and three models are used to determine the relative importance of ocean circulation dynamics and surface fluxe s in controlling water temperature on the WFS throughout the year. Results from the PW P and the one-dimensional version of the POM model show how well the surface fluxes al one are able to reproduce the observed temperature field. Results from the three-di mensional POM are used in an attempt to verify the importance of ocean circulation in determining water temperature at certain times of the year. Additional experiment s are conducted with the PWP model to investigate the importance of cool skin, warm layer effects, rain sensible heat flux, moisture flux, bottom reflection, and water ty pe on model results. Th e model results are improved when the model is forced with basi c surface fluxes that have been modified to include cool skin, warm layer effect. In corporating a bottom reflection term also improved the PWP results in many cases, alt hough the assumption of 100% reflection at the bottom is a gross simplification. The optim al water type appeared to vary depending on the location and season. Water type IB was used as an average. There is a lack of data on water type on the WFS, and work needs to be done to determine the spatial and temporal variability. A cursory investigation of this could be conduc ted in the future by correlating observed salinity changes with wate r turbidity and incorporating that into the models. Analyses of the depth-averag ed one-dimensional temperature equation quantitatively show that surface fluxes com puted using net shortwave, net longwave, latent and sensible heat fluxes, without an SST correction for cool skin, warm layer effects, is least effective in closing the one-dimensional heat budget. Synoptic scale variability in local temperat ure in autumn, winter and sp ring cannot be accounted for by heat flux alone. Generally all one-dimensional model out put temperatures are colder than observations. This could also be due to a variety of reasons: (a ) there is almost always an advection of warmer water at these locations throughout the year; (b ) the derived surface fluxes are too low; (c) there is a problem w ith the mixing schemes used in the models.

PAGE 174

154 Warm advection, introduced in the three-dimensional model, did improve the results in spring and summer and previous model resu lts also show it to contribute to water temperatures on the WFS in spring (He and Weisberg 2002a). The POM3d also reproduced accurate temperature fields in su mmer when solar heating was strong, despite problems associated with forcing the entire domain with fluxes from a point measurement. Differences between the mode l and observed thermocline and isotherm structure, with the POM1d showing a more diffuse thermocline than the PWP. Although part of the difference is because the PW P model does not include bottom mixing, the mixing schemes are also contributing to the differences between the model and observed temperature fields. A thorough investigation into this is outside the scope of this work, but will be looked at in the future. In an attempt to produce better gradient fields and hence the advective contribution to water column temperature th e POM3d is run for two months (forced by EDAS winds and a heat flux re laxed to match the OI SST) prior to the month that it is forced using in situ derived surface fluxe s. The initial T/S profiles for the POM3d experiments are produced by the two-month runs and differ from the initial observed T/S profiles. Calculating the depthaveraged difference between th e initial model and in situ temperatures shows that the difference is smaller in the transitional seasons and larger in summer and winter. Therefore the WFS POM 3d model temperature fields forced by EDAS are closer to observations in spring a nd fall. In all seasons except autumn, the POM3d model initial temperature profile is stratified more than observations, with stratification being worse in summer. This suggests that while the POM3d may simulate currents with a quantifiable degree of accuracy under moderate wind forcing, it is more difficult to maintain the density field. Howe ver, despite experimental design limitations, once the initial T/S profile is mixed the POM3d results improve in the summer when observed surface fluxes are used to force th e model. Accurate surface flux forcing is required in order to reproduce the observed fields using a coastal ocean circulation model. Given uncertainties in surface fluxes (momentum and buoyancy) and open boundary values, data assimilation becomes of increasing importance in maintaining the internal density structure.

PAGE 175

155 Late autumn, winter, and spring on the WFS are characterized by the synopticscale passage of extra-tropical fronts (chapt ers 3 and 4), with cl ockwise turning winds (Fernandez-Partagas and Mooers 1975). Water te mperatures decrease with the passage of each front in response to surface heat flux, a nd the water column becomes well-mixed in response to an increased surface momentum flux. In spring, heat flux into the ocean begins to warm the water temperature. The WFS transition from winter to summer occurs when the isotherms begin to stratify in response to an increase in heat flux and a decrease in wind stress. Upwelling at depth and surface heating inhibit mixing and create a strong thermocline. Stratification is eroded by a decr ease in net surface heat flux into the ocean due to the passage of synoptic scale extratropical fronts, which also bring increased clockwise turning winds that promote mixi ng and produce a well-mixed water column. The signature of these synoptic variations is most evident in shallow water where smaller fluxes are required to affect change. Mode l studies show simila r behaviors (He and Weisberg 2002a). In summer, the diurnal cycl e from the surface heat flux is clearly seen. The cooling of water temperature in autumn is punctuated by tropical storms and extratropical fronts. In both cases, the entire wa ter column cools suddenly due to associated surface fluxes. In fall, the three-dimensional o cean circulation is important in determining the temperature field on the WFS. Observati ons verify previous model results (He and Weisberg 2003b) and show upwelling may result in warm water at the surface in fall when the heat flux out of the ocean cools th e near-shore surface waters and warmer water is advected towards land. Warm water is also observed by advection from the south. Chapter 4 and POM model studie s by He and Weisberg (2002 a, 2003b) show that surface heat flux is primarily responsible for the tran sitional seasonal ocean temperature changes, but synoptic scale variability is also controlled by the c onvergence of heat flux by the ocean circulation. The systematic exploration of the flux corrections using the PWP show the sensitivity of water temperature to surface fl uxes and water type. Care must be taken in calculating surface fluxes; the standard calcula tion of net heat flux does not provide the best model rendition of observed temperatur es. Recognizing the spatial and temporal variability of fluxes further complicates their applicability to ocean models. It was hoped

PAGE 176

156 that the POM3d results would demonstrate an improvement in the modeled temperature field by the inclusion of ocean circulation; however this was not always the case because of the limitations of the experiment. Despite this, the changes that it did introduce to the temperature field relative to the one-dimen sional model results, indicates that with improved heat fluxes the density may eventually be better modeled. The one-dimensional temperature analysis results (here and in chapter 4) show that throughout the year the relative importance of surface fluxes and the convergence of energy via ocean circulation va ry in controlling the water temperature. In spring and summer, surface forcing is important in dete rmining the temperature field, however the inclusion of the fully three dimensional o cean circulation, even during a season when heat fluxes dominate, are important on synoptic timescales. In autumn and winter, the surface fluxes and ocean dynamics are both requ ired in determining the temperature field, however surface fluxes are of primary importa nce during the passage of tropical storms or extra-tropical fronts. Although the relative importance may vary, the inclusion of ocean circulation and surface fluxes are nece ssary. Given the difficulty in measuring surface fluxes and related parameters at sea, especially with a dense spatial distribution, perhaps the most prudent method to improving the surface forcing for coastal ocean models, beyond heat flux relaxation to observe d SST, will be to use point observations of in situ data in the coastal region as part of a larger data assimilation.

PAGE 177

157 Chapter 6 Summary This dissertation investigat es various aspects of the ocean-atmosphere fluxes on the West Florida Shelf (WFS) and how they affect ocean temperature. A coastal ocean region, such as the WFS, is the nexus of interactions between the land, ocean, and atmosphere, which makes it an interesting a nd dynamic system to study. The ability to model and predict various aspects of such a system, including th e ecology, requires a comprehensive knowledge of surface fluxes, and how they are affected on synoptic, seasonal and interannual timescales by the larger scale atmospheric and oceanic circulations. Such information is also impor tant from a social aspect as increasing numbers of people, living near the coast, are affected by coastal ocean-atmosphere interactions. In particular, conditions o ffshore may affect the weather onshore. Chapter 1 sets the stage for this work by describing the large-scale influences on the WFS, namely the ocean circulation of the Gulf of Mexico and the Bermuda High, located over the Atlantic. A review of prev ious observations on the WFS include the ocean circulation, the influence of the Loop Cu rrent, its tides, the effects of winds, and the impact of tropical storms on the ocean mixe d layer. In the liter ature, less attention has been given to the effect of the annual cy cle of heat fluxes on the WFS and its adjacent landmass. This dissertation aims to fill this gap. A review of previous model experiments using models of increasing complexity have shown that WFS ocean circulation de pends on many factors in cluding wind forcing, coastal geometry, bottom topography, and synop tic weather systems: surface fluxes are also important. The ocean response to lo cal wind forcing induced upwelling and downwelling events on the WFS is asymmetric and highlights the importance of the density field to modeling ocean circulation on the shelf (Weisberg et al. 2001). Baroclinic effects via surface heat flux are important in reproducing the seasonal ocean circulation. Specifically, NCEP/NCAR daily reanalysis wi nds and heat fluxes alone are insufficient to produce the seasonal ocean ci rculation unless an SST correc tion is applied to the heat

PAGE 178

158 flux (He and Weisberg 2002a, 2003b). In order to refine the heat flux relaxation to SST, an observed SST product produced from satell ite data by optimal interpolation was developed (He et al. 2003). This dissert ation determines, via a combination of observations and models, the relative importa nce of surface fluxes and ocean circulation dynamics in controlling ocean temperature. Observations and derived surface fluxes from an array of buoys deployed on the WFS, between 1998 and 2003, are described. Heat and momentum fluxes are calculated using versions 2.5 and 3.0 of the TOGA COARE algorithm (Fairall et al. 1996, 2003) which uses bulk parameterizations to compute th e fluxes. Because the shelf is shallow, an additional heat flux, the penetr ative radiation, is included in the net heat flux by assuming 100% reflection of shortwave radiati on that reaches the ocean floor. Climatological data provide insights about the long-term mean annual cycle of the ocean and atmosphere. In Chapter 2, a comparison of ocean-atmosphere variables between ten standard climatologies show diffe rences in relative humidity and heat flux over the Gulf of Mexico and WFS, and although they are useful in pr oviding a large-scale picture, they are unable to re produce spatial flux variability s hown by in situ and satellite measurements. The largest variability occurs in parameters that depend on water vapor and cloud cover, which is known to be problem atic for atmospheric models (Cess et al. 1996). Maps of these standard climatologies and their anomalies are given in the Appendix. WFS observations and derived surface fluxes are m onthly averaged and used to identify (a) differences between the sta ndard climatologies a nd observations and (b) which standard climatology, if any, best re produces this coastal marine atmosphere because this is an important step toward s improving coupled ocean-atmosphere models and their supporting observing systems. The la nd-sea transition is poorly resolved and no standard climatology completely captures th e variability on the WFS, suggesting that long-term in situ observations, which may also eventually be used for data assimilation, are needed to identify the basic annual coastal ocean-atmosphere variability. Observed relative humidity variations on the coastal ocean of the WFS are examined in detail over the four-year period 1999-2003 in Chapter 3. Despite considerable daily and synoptic variability within seasons, the monthly mean values are

PAGE 179

159 nearly constant at about 75% Unlike observations, NCEP re analysis climatology over the WFS and land-based coastal data both show an annual cycle in monthly mean relative humidity, suggesting that the reanalysis field is influenced by land. There are two problems in using NCEP reanalysis fields ove r coastal oceans: a) th e large grid spacing does not allow it to capture coastal ocean vari ability; and b) the r eanalysis fields are produced by a model with data assimilation, but the paucity of in situ data in coastal environments results in an intrinsic land or deep-ocean bias. On synoptic timescales winter has the greatest relati ve humidity variability; values range from less than 50% to over 100% (supersaturation values of up to 3% are recorded and co incide with fog on shore) as extra-tropical fronts move over the WFS. The two different meteorological packages used to collect data show an offs et on a monthly average which, after further investigation, was primarily attributed to the location of the se nsors on the WFS. The IMET/ASIMET moorings are positioned farthe r north and closer to shore than the Weatherpak moorings, and are therefore in different air-sea regimes showing that RH values are sensitive to small spatial variations in the coastal ocean environment and depend not only on the high-frequency variabi lity in meteorological conditions, but also on the low-frequency variability in oceanic conditions. The annual cycle of sea surface temperat ure and ocean-atmosphere fluxes on the WFS are described in Chapter 4 using in s itu measurements and climatology. Generally, when the heat flux switches sign from ne gative to positive in spring, SST is at a minimum and the water column begins to warm up, and conversely when the heat flux switches sign from positive to negative in fall. In spring, synoptic scale mome ntum and heat flux vari ations, associated with extra-tropical fronts, result in successive wate r column stratification and de-stratification events. Fall is characterized by de-stratification of the water column and a series of steplike decreases in the temperature in response to both tropical storms and extra-tropical fronts. The passage of the first tropical storm of the season heralds th e subsequent decline in water temperature: the st orm-induced temperature decrease coupled with the generally negative net heat flux during this time ensu res that the water column cannot warm up again. A one-dimensional analysis of the temp erature equation shows that in spring and

PAGE 180

160 autumn, the seasonal change in water temperat ure is due to surface heat flux and synoptic scale variability is primarily du e to ocean circulation dynamics. In Chapter 5, the WFS in situ data and the Price, Weller, Pinkel one-dimensional mixed layer model, and oneand three-dime nsional versions of the Princeton Ocean Model are employed to address the question of the relative importance of surface heat flux versus energy convergence via ocean circ ulation in controlling water temperatures. Additional experiments are conduc ted with the PWP model to investigate the impact of cool skin, warm layer effects, rain sensib le heat flux, moisture flux, bottom reflection, and water type on model results. These show that careful attention must be paid in calculating and applying the surface heat fluxes. Results from the PWP and POM model experiments show that throughout the year th e relative importance of ocean circulation dynamics and surface fluxes vary in controlli ng the water temperature. As in chapter 4, the seasonal transitions are mainly controll ed by heat flux with smaller contributions from convergence by the ocean circulation but synoptic scale variability is controlled by both the ocean circulation and surface heat flux. In spring and summer, surface forcing is important in determining the temperatur e field, however the inclusion of ocean circulation, even during seasons when heat fluxes dominate, are important on synoptic timescales. In autumn and winter, surface fluxe s and ocean circulation are both required in determining the temperature field, howev er surface fluxes are of primary importance during the passage of tropical stor ms or extra-tropical fronts. The coastal ocean temperature balance is 3D, so it requires 3D models. But these must be supported by adequate surface heat flux boundary conditions, which in turn require a sufficient number of in situ m easurement points for c onstraining atmospheric models. How many measurement points are ne cessary is a trickier question and the answer will depend on the spatial scales of SST variability and atmosphere model resolution. Presently the coastal oc ean is under-sampled and under-modeled.

PAGE 181

161 References Anderson, E. R., 1952: Energy budget studies, wa ter-loss investigati ons: Lake Hefner Studies. U. S. Geo. Surv. Circ., 229 71-88. Anderson, S. P., and M. F. Baumgartner, 1998: Radiative heating e rrors in naturally ventilated air temperature measurements made from buoys. J. Atmos. and Oceanic Tecnhol., 15 157-173. Berliand, M. E., and T. G. Berliand, 1952: Meas urements of the effective radiation of the Earth with varying cloud amounts (in Russian). Izv. Akad. Nauk SSSR, Ser. Geofiz., 1 Berliand, M. E., and T. G. Strokina, 1980: Global distribu tion of the total amount of clouds (in Russian). Hydrometeorological. Leningrad, Russia. 71pp. Blaha, J., and W. Sturges, 1981: Evidence fo r wind-forcing circulation in the Gulf of Mexico. J. Mar. Res., 39, 711-733. Blumberg, A. F., and G. L. Mellor, 1987: A description of a three-dimensional coastal ocean circulation model. Three-Dimensional Coasta l Ocean Models, Vol. 4., N. Heaps, Ed., Amer. Geophys. Union, Washington DC, 208-233. Breaker, L. C., D. B. Gilhousen, and L. D. Burroughs, 1998a: Preliminary results from long-term measurements of atmospheric mo isture in the marine boundary layer in the Gulf of Mexico. J. Atmos. Ocean. Tech., 15 661-676. Breaker, L. C., D. B. Gilhousen, H. L. Tolman, and L. D. Burroughs, 1998b: Initial results from long-term measurements of atmospheric humidity and related parameters in the marine boundary layer at two locations in th e Gulf of Mexico. J. Mar. Sys., 16 199-217. Buck, A. L., 1981: New equations for computi ng vapor pressure and enhancement factor. J. Appl. Meteor., 20 1527-1532. Bunker, A. F., 1976: Computations of surf ace energy flux and annual air-sea interaction cycles of the North Atlantic Ocean. Mon. Wea. Rev., 104 1122-1140. Cereceda, P., and R. Schemenauer, 1991: The occurrence of fog in Chile. J. Appl. Meteor., 30 1097-1105. Cess, R. D., and co-authors, 1996: Cloud fee dback in atmospheric general circulation models: An update. J. Geophys. Res., 101 12791-12794.

PAGE 182

162 Chao, Y., Z. Li, J. C. Kindle, J. D. Padua n, and F. P. Chavez, 2003: A high-resolution surface vector wind product for coastal o ceans: Blending satellite scatterometer measurements with regional mesoscale atmospheric model simulations. Geophys. Res. Lett., 30 (1), 1013. Clark, N. E., L. Eber, R. M. Laurs, J. A. Re nner, and J. F. T. Saur, 1974: Heat exchange between ocean and atmosphere in th e Eastern North Pacific for 1961-1971. NOAA Tech. Rep. NMFS SSRF-682, U. S. Dept. of Commerce, Washington, D.C.. Clarke, A. J. and D. S. Battisti, 1981: Th e effect of continental shelves on tides. Deep-Sea Res., 28A 665-682. Clarke, A. J., and K. H. Brink, 1985: The res ponse of stratified, fric tional flow of shelf and slope waters to fluctuating larg e-scale low frequency wind forcing. J. Phys. Oceanogr., 15 439-453. Cooper, C. K., 1987: A numerical modeling st udy of low-frequenc y circulation on the West Florida Shelf. Coastal Engineering, 11 29-56. Court, A., and R. D. Gerston, 1966: Fog frequency in the United States. Geog. Rev., 56 543-550. Cragg, J., G. Mitchum, and W. Sturges, 1983: Wind-induced seasurface slopes on the West Florida Shelf. J. Phys. Oceanogr., 13 2201-2212. Crescenti, G. H., R. E. Payne, and R. A. Weller, 1990: Improved Meteorological measurements from buoys and ships (I MET): Preliminary comparisons of humidity sensors. WHOI Tech. Rep., WHOI-90-18, 57pp. Da Silva, A. M., C. C. Yang, and S. Levitus, 1994: Atlas of Surface Marine Data, 1994. Vol. 1: Algorithms and Procedures, NOAA Atlas NESDIS 6, U.S. Dept. of Commerce, Washington, DC, 83pp. Dickey, T. D., D. V. Manov, R. A. Weller, and D. A. Siegel, 1994: Determination of longwave heat flux at the air-sea in terface using measurements from buoy platforms. J. Atmos. Oceanic Tech., 11 1057-1078. DiMego, G. J., L. F. Bosart, and G. W. Ende rsen, 1976: An examination of the frequency and mean conditions surrounding frontal incursions into the Gulf of Mexico and Caribbean Sea. Mon. Wea. Rev., 104 709-718. Elliott, W. P., and D. J. Gaffen, 1991: On the utility of radiosonde humidity archives for climate studies. Bull. Amer. Met. Soc., 72 1507-1520.

PAGE 183

163 Emmons, G., and R. B. Montgomery, 1947: Note on the physics of fog formation. J. Meteor., 4 206-209. Enfield, D. B., and S.-K. Lee, 2005: The heat balance of the Western Hemisphere Warm Pool. J. Climate, accepted. Esbensen, S. K., and Y. Kushnir, 1981: Th e heat budget of the global ocean: An atlas based on estimates from surface marine observations. Tech. Rep. 29 Climate Research Institute, Oregon State University, Corvallis. Fairall, C. W., E. F. Bradley, D. P. Roge rs, J. B. Edson, and G. S. Young, 1996: Bulk parameterization of air-sea fluxes fo r Tropical Ocean-Global Atmosphere Coupled-Ocean Atmosphere Response Experiment. J. Geophys. Res., 101 37473764. Fairall, C. W., E. F. Bradley, J. E. Hare A. A. Grachev, and J. B. Edson, 2003: Bulk parameterization of air-sea fluxes: up dates and verification for the COARE algorithm. J. Climate, 16 571-591. Fenner, R. L., 1973: Use of cellulose crystallite structures with solid state strain gages for humidity and moisture measurement. ISA Convention, ASI 73249, 251-252. Fernandez-Partagas, J., and C. N. K. Mooe rs, 1975: A subsynoptic study of winter cold fronts in Florida. Mon. Wea. Rev., 103 742-744. Garand, L., C. Grassotti, J. Hall, and G. L. Klein, 1992: On differences in radiosonde humidity-reporting practices and their implications for numerical weather prediction and remote sensing. Bull. Amer. Met. Soc., 73 1417-1423. Gerber, H. E., 1981: Microstr ucture of a radiation fog. J. Atmos. Sci., 38 454-458. Gilbes, F., C. Tomas, J. J. Walsh, and F. E. Muller-Karger, 1996: An episodic chlorophyll plume on the West Florida Shelf. Cont. Shelf Res., 16 1201-1224. Gill, A. E., 1982: Atmosphere-Ocean Dynamics. Academic Press, Inc., 662 pp. Golden, J. H., R. Serafin, V. Lally, and J. Facundo, 1986: Atmospheric sounding system. Mesoscale Meteorology and Forecasting, P.S. Ray, Ed., Amer. Meteor. Soc., 5070. Goodman, J., 1977: The microstructure of California coastal fog and stratus. J. Appl. Meteor., 16 1056-1067. Gosnell, R., C. W. Fairall, and P. J. Webste r, 1995: The sensible he at of rainfall in the tropical ocean. J. Geophys. Res., 100 18437-18442.

PAGE 184

164 Grassl, H., V. Jost, R. Kumar, J. Schulz, P. Bauer, P. Schluessel, 2000. The Hamburg Ocean-Atmosphere Parameters and Fluxe s from Satellite Data (HOAPS): A climatological atlas of sate llite-derived airsea interaction parameters over the oceans. Report No. 312, ISSN 0937-1060, Max Plank Institute for Meteorology, Hamburg. Hastenrath, S., and P. J. Lamb, 1977: Climatic Atlas Of The Tropical Atlantic And Eastern Pacific Oceans. The University of Wisconsin Press, 97 pp. Hastenrath, S., and P. J. Lamb, 1978: Heat Budget Atlas Of The Tropical Atlantic And Eastern Pacific Oceans. The University of Wisconsin Press, 90 pp. He, R., and R. H. Weisberg, 2002a: West Fl orida Shelf circulation and temperature budget for the 1999 spring transition. Cont. Shelf Res., 22 719-748. He, R., and R. H. Weisberg, 2002b: Tides on the West Florida Shelf, J. Phys. Oceanogr., 32 3455-3473. He, R., and R. H. Weisberg, 2003a: A Loop Current intrusion case study on the West Florida Shelf. J. Phys. Oceanogr., 33 465-477. He, R., and R. H. Weisberg, 2003b: West Fl orida Shelf circulation and temperature budget for the 1998 fall transition. Cont. Shelf Res., 23 777-800. He, R., R. H. Weisberg, H. Zhang, F. E. Muller-Karger, and R. W. Helber, 2003: A cloud-free, satellite-derived, sea surface temperature analysis for the West Florida Shelf. Geophys. Res. Lett., 30 (15), 1811. Helfand, H. M., and S. D. Schubert, 1995: C limatology of the Great Plains low-level jet and its contribution to the continental moisture budge t of the United States. J. Climate, 8 784-806. Hellerman, S., and M. Rosenstein, 1983: No rmal monthly wind stress over the world ocean with error estimates. J. Phys. Oceanogr., 13 1093-1104. Henry, W. K., 1979: Some aspects of the fate of cold fronts in the Gulf of Mexico. Mon. Wea. Rev., 107 1078-1082. Higgins, R. W., K. C. Mo, and S. D. Schube rt, 1996: The moisture budget of the central United States in spring as evaluate d in the NCEP/NCAR and the NASA/DAO reanalyses. Mon. Wea. Rev., 124 939-963.

PAGE 185

165 Hosom, D. S., R. A. Weller, R. E. Payne and K. E. Prada, 1995: The IMET (Improved Meteorology) ship and buoy systems. J. Atmos. And Oceanic Technol., 12 527540. Hudson, J. G., 1980: Relationship betw een fog condensation nuclei and fog microstructure. J. Atmos. Sci., 37 1854-1867. Huh, O. K., W. J. Wiseman, Jr., and L. J. Rouse, Jr., 1978: Winter cycle of sea surface thermal patterns, northeastern Gulf of Mexico. J. Geophys. Res., 83 4523-4529. Huh, O. K., W. J. Wiseman, Jr., L. J. Rouse, Jr., 1981: Intrusion of Loop Current waters onto the West Florida Continental Shelf. J. Geophys. Res., 86 4186-4192. Huffman, G. J., R. F. Adler, M. Morrissey, D. T. Bolvin, S. Curtis, R. Joyce, B. McGavock, and J. Susskind, 2001: Global precipitation at one-degree daily resolution from multi-satellite observations. J. Hydrometeor., 2(1) 36-50. Jerlov, N. G., 1968: Optical Oceanography. Elsevier, 199pp. Johns, E., W. D. Wilson, and R. L. Molinari, 1999: Direct observations of velocity and transport in the passages between the Intr a-Americas Sea and the Atlantic Ocean, 1984-1996. J. Geophys. Res., 104-C11 25805-25820. Josey, S. A, E. C. Kent, D. Oakley, and P. K. Taylor, 1996: A new global air-sea heat and momentum flux climatology. International WOCE Newsletter, 24 3-5. Josey, S. A., D. Oakley, and R. W. Pascal, 1997: On estimating the atmospheric longwave flux at the ocean surface fro m ship meteorological reports. J. Geophys. Res., 102 27961-27972. Kalnay, E., and co-authors, 1996: The NC EP/NCAR 40-year reanalysis project. Bull. Amer. Met. Soc., 77 437-471. Katsaros, K. B., and co-authors, 1994: Measur ements of humidity and temperature in the marine environment during the HEXOS main experiment. J. Atmos. Oceanic Technol., 11 964-981. Koblinsky, C. J., 1981: The M2 tide on the West Florida Shelf. Deep-Sea Res., 28A 1517-1532. Lala, G. G., E. Mandel, and J. E. Jiusto, 1975: A numerical evalua tion of radiation fog variables. J. Atmos. Sci., 32 720-728.

PAGE 186

166 Large, W. G., J. Morzel, and G. B. Crawford, 1995: Accounting for surface wave distortion of the marine wind profile in low-level ocean storms wind measurements. J. Phys. Oceanogr., 25 2959-2971. Larsen, J. C., 1992: Transport and heat flux of the Florida Current at 27oN derived from cross-stream voltages and profiling data: Theo ry and observations. Phil. Trans. R. Soc. London Ser. A., 338 169-236. Leipper, D. F., 1994: Fog on the U.S. west coast: A review. Bull. Amer. Met. Soc., 75 229-240. Lewis, J., D. Koracin, R. Rabin, and J. Busi nger, 2003: Sea fog off the California coast: Viewed in the context of transient weather systems. J. Geophys. Res., 108 4457, doi:10.1029/2002JD002833. Li, Z., and R. H. Weisberg, 1999: West Flor ida Continental Shelf response to upwelling favorable wind forcing: 2. Dynamics. J. Geophys. Res., 104 23427-23442. Liu, W. T., W. Tang, and P. P. Niiler, 1991: Humidity profiles over the ocean. J. Climate, 4 1023-1034. Liu, Y., R. H. Weisberg, and R. He, 2005: S ea surface temperature patterns on the West Florida Shelf using Growing Hier archical Self-Organizing Maps. J. Atmos. Oceanic Technol., accepted. MacWhorter, M. A., and R. A. Weller 1991: Error in measurements of incoming shortwave radiation made from ships and buoys. J. Atmos. Oceanic Technol., 8 108-117. Marmorino, G. O., 1982: Wind-forced sea level va riability along the West Florida Shelf. J. Phys. Oceanogr., 12, 389-404. Marmorino, G. O., 1983: Variability of current temperature, and bottom pressure across the West Florida Continental Shelf, winter 1981-1982. J. Geophys. Res., 88 4439-4457. Maul, G., 1977: The annual cyclone of the Gulf Loop Current. Part I: Observations during a one year time series. J. Mar. Res., 35, 29-47. Maul, G. A., D. A. Mayer, and S. R. Ba ig, 1985: Comparisons be tween a continuous 3year current-meter observation at the sill of the Yucatan Strait, satellite measurements of Gulf Loop Current area, and regi onal sea-level. J. Geophys. Res., 90 9089-9096.

PAGE 187

167 Medovaya, M., D. E. Waliser, R. A. Weller, and M. J. McPhaden, 2002: Assessing ocean buoy shortwave observations using clear-sky model calculations. J. Geophys. Res., 107 doi:10.1029/2000JC000558. Mellor, G. L., and T. Yamada, 1974: A hi erarchy of turbulence closure models for planetary boundary layers. J. Atmos. Sci, 13 1791-1806. Mellor, G. L., and T. Yamada, 1982: Development of a turbulence closure model for geophysical fluid problems. Rev. Geophys., 20 851-875. Meyer, M. B., J. E. Jiusto, and G. G. La la, 1980: Measurements of visual range and radiation-fog (haze) microphysics. J. Atmos. Sci., 37 622-629. Meyers, S. D., E. M. Siegel, R. H. Weisber g, 2001: Observations of currents on the West Florida Shelf break. Geophys. Res. Lett., 28 2037-2040. Mitchum, G. T., and W. Sturges, 1982: Wind-dr iven currents on the West Florida Shelf. J. Phys. Oceanogr., 12,1310-1317. Mitchum, G. T., and A. J. Clarke, 1986: The frictional nearshore response to forcing by synoptic scale winds. J. Phys. Oceanogr., 16 934-946. Muller, S. H., and P. J. Beekman, 1987: A test of commercial humidity sensors for use at automatic weather stations. J. Atmos. Oceanic Technol., 4 731-735. Niiler, P. P., 1976: Observations of low-frequency currents on the West Florida Continental Shelf. Memoires Societe Royale des Sciences de Leige, 6, X 331358. NOAA Storm Data Reports, 1998-2003. Natio nal Oceanographic and Atmospheric Administration. ISSN: 0039-1972. Noonkester, V. R., 1979: Coastal mari ne fog in southern California. Mon. Wea. Rev., 107 830-851. Oberhuber, J. M., 1988. An atlas based on COADS data set. Tech. Rep. 15 MaxPlanck-Institut fr Meteorologie. Ochoa, J., J. Sheinbaum, A. Badan, J. Candela, and D. Wilson, 2001: Geostrophy via potential vorticity inversion in the Yucatan Channel. J. Marine Res., 59 725-747. Paulson, C. A., and J. J. Simpson, 1977: Irradiance measurements in the upper ocean. J. Phys. Oceanogr., 7 952-956. Payne, R. E., 1972: Albedo of the sea surface. J. Atmos. Sci., 29 959-970.

PAGE 188

168 Peace, R. L., Jr., 1969: Heavy-fog regions in the conterminous United States. Mon. Wea. Rev., 97 116-123. Peixoto, J. P., and A. H. Oort, 1996: The climatology of relative humidity in the atmosphere. J. Climate, 9 3443-3463. Pili, R. J., E. J. Mack, C. W. Rogers, U. Katz, and W. C. Kocm ond, 1979: The formation of marine fog and the development of fog-stratus systems along the California coast. J. Appl. Meteor., 18 1275-1286. Price, J. F., C. N. K. Mooe rs, and J. C. Van Leer, 1978: Observation and simulation of storm-induced mixed-layer deepening. J. Phys. Oceanogr., 8 582-599. Price, J. F., R. A. Weller, and R. Pinkel, 1986: Diurnal cycling: Observations and models of the upper ocean response to diur nal heating, cooling, and wind mixing. J. Geophys. Res., 91 8411-8427. Rasmusson, E. M., 1967: Atmospheric water va por transport and th e water balance of North America: Part I. Characterist ics of the water vapor flux field. Mon. Wea. Rev., 95 403-426. Roach, W. T., 1995: Back to basics: Fog. Part 3 The formation and dissipation of sea fog. Weather, 50 80-84. Saunders, I. S., 1967: The relationship between thermally stimulated luminescence and thermally stimulated conductivity. British J. Appl. Physics, 18 ,1219-1220. Schmitz, J. T., and S. Mullen, 1996: Water vapor transport associated with the summertime North American Monsoon as depicted by ECMWF analyses, J. Climate, 9, 1621-1634. Schmitz, W. J., Jr., and P. L. Richardson, 1991: On the sources of the Florida Current. Deep-Sea Res., 38 1: S389-S409 (Suppl.). Sheinbaum, J., J. Candela, A. Baden, and J. Ochoa, 2002: Flow structure and transport in the Yucatan Channel. Geophys. Res. Lett., 29(3) doi:10.1029/2001GL013990. Slutz, R. J., S. J. Lubker, J. D. Hiscox, S. D. Woodruff, R. L. Jenne, D. H. Joseph, P. M. Steurer, and J. D. Elms, 1985. Compre hensive Ocean-Atmosphere Data Set; Release 1. NOAA Environmental Research Laboratories, Climate Research Program, Boulder, CO 268pp. (NTIS PB86-105723) Stone, R. G., 1936: Fog in the United States and adjacent regions. Geog. Rev., 26 111134.

PAGE 189

169 Sturges, W., 1994: The frequency of ri ng separations from the Loop Current. J. Phys. Oceanogr., 24 1647-1651. Sturges, W., and R. Leben, 2000: Frequency of ring separations from the Loop Current in the Gulf of Mexico: A revised estimate. J. Phys. Oceanogr., 30 1814-1819. Telford, J. W., and S. K. Chai, 1993: Marine fog and its dissipation over warm water. J. Atmos. Sci., 50 3336-3349. Virmani, J. I., and R. H. Weisberg, 2 003: Features of the observed annual oceanatmosphere flux variability on the West Florida shelf. J. Climate, 16 734-745. Virmani, J. I., and R. H. Weisberg, 2005: Relative humidity over the West Florida Continental Shelf. Mon. Wea. Rev., in press. Vukovich, F. M., 1988: Loop Cu rrent boundary variations. J. Geophys. Res., 93 1558515591. Waliser, D. E., R. A. Weller, and R. D. Cess, 1999: Comparisons between buoyobserved, satellite-derived, and mode led surface shortwave flux over the subtropical North Atlantic dur ing the Subduction Experiment. J. Geophys. Res., 104 31301-31320. Wang, C., and D. Enfield, 2001: The tropical Western Hemisphere Warm Pool. Geophys. Res. Lett., 28 1635-1638. Weatherly, G. L., and D. Thistle, 1997: On the wintertime current s in the Florida Big Bend Region. Cont. Shelf Res., 17 1297-1319. Weisberg, R. H., 1996: On the evolution of SST over the PACS region. Abstracts of 76th AMS Annual Meeting, Atlanta, Georgia, Amer. Meteor. Soc., 378. Weisberg, R. H., B. D. Black, and H. Yang, 1996: Seasonal modulation of the West Florida Continental Shelf circulation. Geophys. Res. Lett., 23 2247-2250. Weisberg, R. H., B. D. Black, and Z. Li, 2000: An upwelling case study on the Floridas West coast. J. Geophys. Res., 105 11459-11469. Weisberg, R. H., Z. Li, and F. E. Muller-K arger, 2001: West Florida Shelf response to local wind forcing: April 1998. J. Geophys. Res., 106 31239-31262. Weisberg, R. H. and L. J. Pietrafesa, 1983: Kinematics and correlation of the surface wind field in the South Atlantic Bight. J. Geophys. Res., 88 4593-4610.

PAGE 190

170 Weller, R. A., and S. Anderson, 1996: Surf ace Meteorology and air-sea fluxes in the Western Equatorial Pacific Warm Pool during the TOGA Coupled OceanAtmosphere Response Experiment. J. Climate, 9 1959-1990. Williams, J., W. F. Grey, E. B. Murphy, and J. J. Crane, 1977: Memoirs of the Hourglass Cruises, Rep. IV(III). Mar. Res. Lab., Fl a. Dept. of Nat. Res., St. Petersburg. Wilson, W. D. and W. E. Johns, 1997: Velocity structure and transport in the Windward Islands passages. Deep-Sea Res., 44 487-520. Woodcock, A. H., 1978: Marine fog dr oplets and salt nuclei Part I. J. Atmos. Sci., 35 657-664. Woodcock, A. H., D. C. Blanchard, and J. E. Jiusto, 1981: Marine f og droplets and salt nuclei Part II. J. Atmos. Sci, 38 129-140. Woodruff, S.D, R. J. Slutz, R. L. Jenn e, and P.M. Streuer, 1987. A Comprehensive Ocean-Atmosphere Data Set. Bull. Amer. Meteor. Soc., 68 1239-1250. Woodruff, S.D., S. J. Lubker, K. Wolter, S. J. Worley, and J. D. Elms, 1993. Comprehensive Ocean-Atmosphere Data Set (COADS) Release 1a: 1980-92. Earth System Monitor, 4 (1), 1-8. Yang, H., and R. H. Weisberg, 1999: Response of the West Florida Shelf circulation to climatological wind stress forcing. J. Geophys. Res., 104 5301-5320. Yang, H., R. H. Weisberg, P. P. Niiler, W. Sturges, and W. Johnson, 1999: Lagrangian circulation and Forbidden Zone on the West Florida Shelf. Cont. Shelf Res., 19 1221-1245.

PAGE 191

171 Appendix A: Maps of Standard Climat ologies over the Gulf of Mexico This appendix contains maps of the st andard climatologies and companion maps of anomalies from the ensemble mean of th e standard climatologies as described in section 2.4 of this dissertation. Details of th e standard climatologies are available in section 2.3. The SOC climatology air temperatur e, specific humidity and winds are 10m above sea level. Relative humdity is not provided in the SOC climatology, so it was calculated using the specific humidity and air temperature at 10m above sea level. Specific humidity was not provided in the EC MWF or Oberhuber climatologies, so it was calculated using relative humidity and air temperature. The variables the maps are available for include: air temperature, sea su rface temperature, relative humidity, specific humidity, sea level pressure, shortwave radia tion, longwave radiation, sensible heat flux, latent heat flux, wind speed, winds, a nd the east and north wind stress.

PAGE 192

172 Appendix A: (Continued) Figure 86. Air Temperature (oC) Climatologies: January to June.

PAGE 193

173 Appendix A: (Continued) Figure 86. (Continued) Air Temperature (oC) Climatologies: July to December.

PAGE 194

174 Appendix A: (Continued) Figure 87. Sea Surfa ce Temperature (oC) Climatologies: January to June.

PAGE 195

175 Appendix A: (Continued) Figure 87. (Continued) Sea Surface Temperature (oC) Climatologies: July to December.

PAGE 196

176 Appendix A: (Continued) Figure 88. Relative Humidity ( %) Climatologies: January to June.

PAGE 197

177 Appendix A: (Continued) Figure 88. (Continued) Relative Humidity ( %) Climatologies: July to December.

PAGE 198

178 Appendix A: (Continued) Figure 89. Specific Humidity (kg/kg) Climatologies: January to June.

PAGE 199

179 Appendix A: (Continued) Figure 89. (Continued) Specific Humidity (kg/kg) Climatologies: July to December.

PAGE 200

180 Appendix A: (Continued) Figure 90. Sea Level Pressure (mb) Climatologies: January to June.

PAGE 201

181 Appendix A: (Continued) Figure 90. (Continued) Sea Level Pressure (mb) Climatologies: July to December.

PAGE 202

182 Appendix A: (Continued) Figure 91. Shortwave Radiation (W/m2) Climatologies: Ja nuary to June.

PAGE 203

183 Appendix A: (Continued) Figure 91. (Continued) S hortwave Radiation (W/m2) Climatologies: July to December.

PAGE 204

184 Appendix A: (Continued) Figure 92. Longwave Radiation (W/m2) Climatologies: January to June.

PAGE 205

185 Appendix A: (Continued) Figure 92. (Continued) L ongwave Radiation (W/m2) Climatologies: July to December.

PAGE 206

186 Appendix A: (Continued) Figure 93. Sensible Heat Flux (W/m2) Climatologies: January to June.

PAGE 207

187 Appendix A: (Continued) Figure 93. (Continued) Se nsible Heat Flux (W/m2) Climatologies: July to December.

PAGE 208

188 Appendix A: (Continued) Figure 94. Latent Heat Flux (W/m2) Climatologies: January to June.

PAGE 209

189 Appendix A: (Continued) Figure 94. (Continued) La tent Heat Flux (W/m2) Climatologies: July to December.

PAGE 210

190 Appendix A: (Continued) Figure 95. Wind (m/s2) Climatologies: January to June.

PAGE 211

191 Appendix A: (Continued) Figure 95. (Continued) Wind (m/s2) Climatologies: July to December.

PAGE 212

192 Appendix A: (Continued) Figure 96. Wind Speed (m/s2) Climatologies: January to June.

PAGE 213

193 Appendix A: (Continued) Figure 96. (Continued) Wind Speed (m/s2) Climatologies: July to December.

PAGE 214

194 Appendix A: (Continued) Figure 97. Taux (N/m2) Climatologies: January to June.

PAGE 215

195 Appendix A: (Continued) Figure 97. (Continued) Taux (N/m2) Climatologies: July to December.

PAGE 216

196 Appendix A: (Continued) Figure 98. Tauy (N/m2) Climatologies: January to June.

PAGE 217

197 Appendix A: (Continued) Figure 98. (Continued) Tauy (N/m2) Climatologies: July to December.

PAGE 218

198 Appendix A: (Continued) Figure 99. Anomalies of Air Temperature (oC) Climatologies: January to June.

PAGE 219

199 Appendix A: (Continued) Figure 99. (Continued) Anoma lies of Air Temperature (oC) Climatologies: July to December.

PAGE 220

200 Appendix A: (Continued) Figure 100. Anomalies of S ea Surface Temperature (oC) Climatologies: January to June.

PAGE 221

201 Appendix A: (Continued) Figure 100. (Continued) Anomalies of Sea Surface Temperature (oC) Climatologies: July to December.

PAGE 222

202 Appendix A: (Continued) Fi g ure 101. Anomalies of Relative Humidit y ( % ) Clim a tolo g ies: Januar y to June.

PAGE 223

203 Appendix A: (Continued) Figure 101. (Continued) Anomalie s of Relative Humidity (%) Climatologies: July to December.

PAGE 224

204 Appendix A: (Continued) Fi g ure 102. Anomalies of S p ecific Humidit y ( k g /k g) Climatolo g ies: Januar y to June.

PAGE 225

205 Appendix A: (Continued) Figure 102. (Continued) Anomalies of Specific Humid ity (kg/kg) Climatologies: July to December.

PAGE 226

206 Appendix A: (Continued) Fi g ure 103. Anomalies of Sea Level Pressure ( mb ) Climatolo g ies: Januar y to June.

PAGE 227

207 Appendix A: (Continued) Figure 103. (Continued) Anomalies of Sea Level Pressure (mb) Clim atologies: July to December.

PAGE 228

208 Appendix A: (Continued) Fi g ure 104. Anomalies of Shortwave Radiation ( W/m2 ) Climatolo g ies: Januar y to June.

PAGE 229

209 Appendix A: (Continued) Figure 104. (Continued) Anomalies of Shortwave Radiation (W/m2) Climatologies: July to December.

PAGE 230

210 Appendix A: (Continued) Fi g ure 105. Anomalies of Lon g wave Radiation ( W/m2 ) Climatolo g ies: Januar y to June.

PAGE 231

211 Appendix A: (Continued) Fi g ure 105. ( Continued ) Anomalies of Lon g wave Radiation ( W/m2 ) Climatolo g ies: Jul y to December.

PAGE 232

212 Appendix A: (Continued) Fi g ure 106. Anomalies of Sensible Heat Flux ( W/m2 ) Climatolo g ies: Januar y to June.

PAGE 233

213 Appendix A: (Continued) Fi g ure 106. ( Continued ) Anomalies of Sensible Heat Flux ( W/m2 ) Climatolo g ies: Jul y to December.

PAGE 234

214 Appendix A: (Continued) Fi g ure 107. Anomalies of Latent Heat Flux ( W/m2 ) Climatolo g ies: Januar y to June.

PAGE 235

215 Appendix A: (Continued) Fi g ure 107. ( Continued ) Anomalies of La tent Heat Flux ( W/m2 ) Climatolo g ies: Jul y to December.

PAGE 236

216 Appendix A: (Continued) Fi g ure 108. Anomalies of Wind S p eed ( m/s ) Climatolo g ies: Januar y to June.

PAGE 237

217 Appendix A: (Continued) Fi g ure 108. ( Continued ) Anomalies of Wind S p eed ( m/s ) Climatolo g ies: Jul y to December.

PAGE 238

218 Appendix A: (Continued) Fi g ure 109. Anomalies of Taux ( N/m2 ) Climatolo g ies: Januar y to June.

PAGE 239

219 Appendix A: (Continued) Fi g ure 109. ( Continued ) Anomalies of Taux ( N/m2 ) Climatolo g ies: Jul y to December.

PAGE 240

220 Appendix A: (Continued) Fi g ure 110. Anomalies of Tau y ( N/m2 ) Climatolo g ies: Januar y to June.

PAGE 241

221 Appendix A: (Continued) Fi g ure 110. ( Continued ) Anomalies of Tau y ( N/m2 ) Climatolo g ies: Jul y to December.

PAGE 242

About the Author Jyotika Itee Virmani, born in Manchester United Kingdom, earned her Bachelor of Physics degree with honours from Imperial College, University of London in 1991. After working for two years as a Research Scientis t at the GEC-Marconi Research Center in Chelmsford, Essex, U.K., she won a Rotary Foundation Ambassadorial Graduate Scholarship in 1993 which allowed her to pursue her M.S. degree in Marine Environmental Science with Dr. Marvin Gell er at the ITPA, Marine Science Research Center, SUNY at Stony Brook, U.S.A. In 1995 she entered the Ph.D. program in the College of Marine Science, USF, to study tr opical Pacific ocean-atm osphere interactions with Dr. Robert Weisberg. An experimental problem resulted in her switching to studying WFS coastal ocean-atmosphere interactions in 2000. During her Ph.D. she has received the Getting Fellowship, Garrels Fellowship, Un iversity Graduate Fellowship, and Knight Fellowship and has been active in educa tional outreach and college governance.


xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001680979
003 fts
005 20060215071050.0
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 051207s2005 flu sbm s000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001141
035
(OCoLC)62493989
SFE0001141
040
FHM
c FHM
049
FHMM
090
GC11.2 (Online)
1 100
Virmani, Jyotika I.
0 245
Ocean-atmosphere interactions on the West Florida shelf
h [electronic resource] /
by Jyotika I. Virmani.
260
[Tampa, Fla.] :
b University of South Florida,
2005.
502
Thesis (Ph.D.)--University of South Florida, 2005.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 242 pages.
Includes vita.
520
ABSTRACT: Ocean-atmosphere fluxes on the West Florida Shelf (WFS) coastal ocean region are investigated using observations and derived surface fluxes from an array of buoys deployed between 1998 and 2003. The observed annual cycle shows that water column temperatures increase and are stratified when heat flux is positive, and they decrease and are well mixed when it is negative. Water temperature is minimum (maximum) when heat flux switches sign from negative (positive) to positive (negative) in early spring (autumn). Tropical and extra-tropical events help define the seasonal characteristics of the water temperature. Despite considerable daily and synoptic variability in relative humidity, observations on the WFS show that the monthly mean values are nearly constant at about 75%. Winter relative humidity varies from less than 50% to over 100% (supersaturation values of up to 3% are recorded and coincide with fog on shore) as extra-tropical fronts move over the WFS.Sensor distribution shows small spatial variations in relative humidity in the coastal ocean environment that depends on high frequency variability in meteorological conditions and low-frequency variability in oceanic conditions. Comparisons with observations show that standard climatologies are unable to reproduce spatial variability on the WFS, especially in relative humidity and surface heat flux components that are dependent on sea surface temperature. Model experiments show that careful attention must be paid in calculating and applying surface heat fluxes. Observations and models are employed to assess the relative importance of surface fluxes and convergence of heat flux by the ocean circulation in controlling ocean temperature.
590
Adviser: Dr. Robert H. Weisberg.
653
Coastal ocean observations.
Surface fluxes.
Relative humidity.
Climatologies.
One-dimensional temperature balance.
690
Dissertations, Academic
z USF
x Marine Science
Doctoral.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.1141