USF Libraries
USF Digital Collections

Enhancement of predictive capability of transit boardings estimation and simulation tool (tbest) using parcel data :

MISSING IMAGE

Material Information

Title:
Enhancement of predictive capability of transit boardings estimation and simulation tool (tbest) using parcel data : an exploratory analysis
Physical Description:
Book
Language:
English
Creator:
Rana, Tejsingh
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Transit demand modeling
Trip rates
Special generator
Trip attraction
Dissertations, Academic -- Civil & Environmental Engineering -- Masters -- USF   ( lcsh )
Genre:
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: TBEST is a comprehensive third generation transit demand forecasting model, developed by the FDOT Public Transit Office (PTO) to help transit agencies in completing their Transit Development Plans (TDPs). The on-going project funded by FDOT, related to TBEST, aims at further enhancing the capabilities of the TBEST model based on additional opportunities identified by the research team. The project focuses on enhancing TBEST's capabilities in following areas: 1) Improving the precision of socio- demographic data by using property appraisal data (parcel data) and, 2) Improving the quality of data regarding trip attraction. Based on the improvement areas, this study aims at performing an exploratory analysis to 1) Identify the differences in activity levels (population and employment) within transit stop buffers due to change in input data i.e. from aggregate census data to disaggregate parcel data. 2) Explore various strategies (development of employment based trip attraction and, parcel land use based trip attraction and exploring how special generators are dealt with in the past studies) to enhance the trip attraction capability of the TBEST model. The results obtained from this analysis provide insights on the strategies and helps define suggestions to further enhance the precision of TBEST model. The results show that use of parcel level data improves the accuracy in capturing the activity levels within the catchment area of each stop. The results also suggest use of parcel land use based trip attraction for stops with special generators or use of interaction variable (interaction between special generator dummy and size (square footage etc.) of the special generator) to enhance the trip attraction capability of the TBEST model.
Thesis:
Thesis (MSCE)--University of South Florida, 2010.
Bibliography:
Includes bibliographical references.
System Details:
Mode of access: World Wide Web.
System Details:
System requirements: World Wide Web browser and PDF reader.
Statement of Responsibility:
by Tejsingh Rana.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains X pages.

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:
usfldc doi - E14-SFE0004628
usfldc handle - e14.4628
System ID:
SFS0027943:00001


This item is only available as the following downloads:


Full Text

PAGE 1

Enhancement of Predictive Capability of Transit Boardings Estimation and Simulation Tool (TBEST) U sing Parcel Data: An Exploratory Analysis b y Tejsingh A. Rana A thesis submitted in partial fulfillment of the requirements for the degree of Maste r of Science in Civil Engineering Department of Civil and Environmental Engineering College of Engineering University of South Florida Major Professor: Abdul Pinjari, Ph.D. Steven E. Polzin, Ph.D. Xuehao Chu, Ph.D. Date of Approval: Ju ly 7, 2010 Keywords: transit demand modeling trip rates, special generator trip attraction Copyright 2010, Tejsingh Rana

PAGE 2

ACKNO WLEDGMENTS This research project would not have been possible without the support of many people. Foremost, I would like to expres s my sincere gratitude to my advisor Dr. Abdul Pinjari who was abundantly helpful and offered invaluable a ssistance, support and guidance throughout the masters program Besides my advisor, I would like to thank the rest of my thesis committee: Dr s Steven E. Polzin and Xuehao Chu for their enc ouragement, insightful comments which helped me learn better and enhance the contents of this work. I would like to thank Mr. Rodney Bunner from Geodecisions, Inc., Tampa for his constant guidance and suggestions. I would like to thank Drs. Jian J. Lu, Chanyoung Lee and Yu Zhang for the courses they have taught. I am thankful to Ingrid hall, Barbara Johnson, Rafael Urena and all the people of the Civil and Environmental Engineering Department for their help. I thank a ll my friends: Meeta Saxena, Sujan Sikder, Amir Shaikh, Mohammed Ahmed Rehmatullah, Nagesh Nayak, Arjun Chauhan, Samira Obied, Anita Kumari, Sharad Malvade and Lalit Patil for support and company. Last but not the least, I would like to thank my family: my parents and my sisters for their love, support and nurturing.

PAGE 3

i TABLE OF CONTENTS LIST OF TABLES ............................................................................................................. iii LIST OF FIGURES ............................................................................................................ v ABSTRACT ....................................................................................................................... vi CHAPTER 1 INTRODUCTION ........................................................................................ 1 1.1 Background ................................................................................................1 1.2 TBEST ........................................................................................................1 1.3 Objectives ...................................................................................................3 1.4 Methodology ..............................................................................................4 1.5 Outline of the Thesis ..................................................................................4 CHAPTER 2 DATA DESCRIPTION ................................................................................ 6 2.1 Census 2000 Data .......................................................................................6 2.2 2009 Property Appraisal Data (Parcel Data) ..............................................6 2.3 2007 InfoUSA Employment Data ............................................................10 2.4 Jacksonville Transportation Authority (JTA) Transit Network Data .........................................................................................................10 2.5 Institute of Transportation Engineers (ITE) Trip Generation Manual, 8th Edition ...................................................................................13 2.6 2001 National Household Travel Survey (NHTS) Database ...................13 CHAPTER 3 TBEST MODEL ......................................................................................... 15 3.1 Feat ures of TBEST ...................................................................................15 3.2 Enhancements in TBEST Model ..............................................................18 3.2.1 Parcel Level Data Capability ........................................................ 18 3.2.2 Trip Attraction Capability ............................................................ 18 CHAPTER 4 EXPLORING PARCEL LEVEL DATA CAPABILITY ........................... 20 4.1 Methodol ogy ............................................................................................20 4.1.1 Block Group Level Demographic Disaggregation ....................... 21 4.1.2 Stop Level Analysis ..................................................................... 23 4.1.3 Route Level Analysis ................................................................... 27 4.2 Results ......................................................................................................30 4.2.1 Results of Stop Level Analysis .................................................... 30 4.2.2 Results of Route Level Analysis .................................................. 49

PAGE 4

ii CHAPTER 5 TRIP ATTRACTION CAPABILITY ENHANCEMENT ......................... 54 5.1 Employment B ased Trip Attraction .........................................................54 5.2 Development of Parcel Land Use Based Trip Attraction/Production ...............................................................................55 5.3 Special Generator Enhancement ..............................................................68 5.3.1 Literature Review .......................................................................... 68 5.4 Analysis ....................................................................................................77 5.5 Results ......................................................................................................79 CHAPTER 6 CONCLUSIONS ........................................................................................ 85 6.1 Conclusions ..............................................................................................85 6.2 Suggestions for Enhancement of TBEST Model .....................................87 REFERENCES ................................................................................................................. 89 APPENDICES .................................................................................................................. 93 Appendix A: Summary of Studies and Datasets Reviewed for Special Generator Enhancement ..................................................................94 Appendix B: Results of the Exploratory Analysis Performed for the Enhancemen t of Trip Attraction Capability ..................................128

PAGE 5

iii LIST OF TABLES Table 1 Description of Variables in Property Appraisal Data (Parcel Data) ...................... 7 Table 2 List of Land Uses Available in Parcel Data ........................................................... 8 Table 3 Details of Routes Selected for the Exploratory Analysis .................................... 11 Table 4 Definitions of Time Periods in TBEST ............................................................... 16 Table 5 Strategies for Assigning Block Group Level Census Population to Parcels ....... 21 Table 6 Frequency Distribution of Num ber of Dwelling Units in MultiF amily Parcels ...................................................................................................................22 Table 7 Aggregate and Disaggregate Level Single Family Population Computed for Diffe rent Si zes of Catchment Area (Buffer) A round Route R5 Stops ...........32 Table 8 Aggregate and Disaggregate Level Multi Family Population Computed for Different Si zes of Catchment Area (Buffer) A round R oute R5 Stops ...........35 Table 9 Aggregate and Disaggregate Level Total Employment Computed for Different Si zes of Catchment Area (Buffer) A round Route R5 Stops .................38 Table 10 D istribution of Number of Stops Within E ach Percentage Difference Category of Single Family P opulation for Different Sizes of Catchment A rea .....................................................................................................................46 Table 11 Di stribution of N umber of Stops Within E ach Percentage Difference C ategory of MultiFamily P opulation for Different Sizes of Catchment A rea .....................................................................................................................46 Table 12 Distribution of Number of Stops Within Eac h Percentage Difference Category of Total Employment fo r Different Sizes of Catchment A rea ............47 Table 13 Parameter Estimates (t stats) of the Line ar Regression Analysis B etween Total Boarding, P opulation and Employment Within Each Stop Buffer ............48 Table 14 Absolute Percentage Difference B etween Ag gregate Level and Disaggregate L evel Single Family Population, MultiFa mily Population and Employment Obtained U sing Route Level Analysis for Different Sizes of Catchment Area (Buffer) .......................................................................53

PAGE 6

iv Table 15 Trips Attraction per Employee for Each Type of Employment and Each TBEST Time Period ...........................................................................................54 Table 16 Trip Rates of Parcel Level Land Use for TBEST Time Periods Using ITE Trip Generation Manual and NHTS 2001 Database ...........................................58 Table 17 Temporal Distribution of Weekday Trips in 2001 NHTS D ata ......................... 64 Table 18 Parcel Land Uses Hav ing Peak Hour Period Different f rom TBEST Time Period .........................................................................................................65 Table 19 Temporal Distribution of Trips in 2001 NHTS D ata ......................................... 66 Table 20 Ta bulation of Special Generators W ith ITE Trip Rates, Relevant Studies and Corresponding Vari ables Used ....................................................................70 Table 21 List of Various Special Generators and the Options (Best, Next Best and Other) for the Explanatory Variables of Each Generator ...................................76 Table 22 Potential Special Generators in the Parcel Data ................................................ 77 Table 23 Results of the Linear Regression Analysis Between Total Boarding, Total Employment, Trips Generate d by Employment and Trips Generated Using Trip Rates for Stops With and W ithout Special Generator .............................................................................................................82 Table 24 Results of the Levene's Test and T Test for Total Boarding and Area of the Land Us es in the Stop Buffer Between Stops With and Without Special Generator ................................................................................................83 Table 25 Results of the Linear Regression Analysis Between Total Boar ding, Special Generator Dummy V ariable and Speci al Generator Area. .....................84 Table 26 List of Special Generators and Variables Used in Laredo Travel Demand Model ....................................................................................................97 Table 27 List of S pecial Generators and Variables Used in Lincoln MPO Travel Demand Model ........................................................................................98 Table 28 List of Special Generators and Variables Used in Texas Travel Demand Model ....................................................................................................99 Table 29 List of Special Generators and Variables Used in DFWRTM ......................... 100 Table 30 Total Employment, Trips Generated by Employme nt, Trips Generated Using Trip Rate s, Area of Non Residential Land U ses and Special Gener ators for the Stops in all the F our Routes ................................................128

PAGE 7

v LIST OF FIGURES Figure 1 Selected Routes and their Respective Stops ....................................................... 12 Figure 2 Stop Level Analysis at Aggregate Level ............................................................ 25 Figure 3 Stop Level Analysis at Disaggregate Level ........................................................ 26 Figure 4 Route Level Analysis at Aggregate Level .......................................................... 28 Figure 5 Route Level Analysis at Disaggregate Level ..................................................... 29 Figure 6 Examples E xpla ining the Difference B etween Aggregate Level and Disaggregate Level .............................................................................................30 Fi gure 7 Graph S howing Aggregate and Disaggregate Level Sin gle Famil y Population Computed U sing Stop Level Analysis for Different Si zes of Catchment Area (Buffer) A round Route R5 Stops .............................................42 Figure 8 Graph S howing Aggregate and Disaggregate Level MultiFamily Population Computed U sing Stop Level Analysis for Different Si zes of Catchment Area (Buffer) A round Route R5 Stops .............................................43 Figure 9 Gr aph S howing Aggregate and Disaggregate L evel Total Employment Computed U sing Stop Level Analysis for Different Si zes of Catchment Area (Buffer) A round Route R5 Stops ...............................................................44 Figure 10 Graphs S howing Aggregate and Disaggregate Level Single Family and Mu ltiFamily Population Computed U sing Route Level Analysis for Different Sizes of Catchment Area (Buffer) .....................................................51 Figure 11 Graphs Showing A ggre gate and Disaggregate Level Total Employment Computed U sing Route Level Analysis for Different Sizes of Catchment Area (Buffer) .....................................................................52 Figure 12 Graphs Showing Differences B etween Trip G enerated by Employment and Trip G enerated U sing Parcel Land U se Based Trip Rates for Route P7 Stops With and W ithout Special Generator .......................80

PAGE 8

vi Enhancement of Predictive Capability of Transit Boardings Estimati on and Simulation Tool (TBEST) U sing Parcel Data: An Exploratory Analysis Tejsingh Rana A B ST RACT TBEST is a comprehensive third generation t ransit demand forecasting model, developed by the FDOT Public Transit Office (PTO) to help transit agencies in completing their Transit Development Plans (TDPs) The ongoing project funded by FDOT related t o TBEST, aims a t further enhancing the capabilities of the TBEST model based on additional opportunities identified by the research team. The project focuses on enhancing TBEST s capabilities in following areas: 1) Improving the p recision of socio demogra phic data by using property appraisal data (parcel data ) and, 2) I mproving the quality of data regarding trip attraction Based on the improvement areas, t his study aims at performing an exploratory analysis to 1) Identify the differences in activity level s (population and employment) within transit stop buffers due to change in input data i.e. from aggregate census data to disaggregate parcel data. 2) Explore various strategies ( development of employment based trip attraction and, parcel land use based tri p attraction and exploring how special generators are dealt with in the past studies ) to enhance the trip attraction capability of the TBEST model. The results obtained from this analysis provide insights on the strategies and helps define suggestions to f urther enhance the precision of TBEST model. The results show that use of parcel level data improves the accuracy in capturing the activity levels within the c atchment area of each stop The results also suggest use of parcel land use based trip attraction for stops with special

PAGE 9

vii generators or use of interaction variable ( interaction between special generator dummy and size (square footage etc.) of the special generator ) to enhance the trip attraction capability of the TBEST model.

PAGE 10

1 CHAPTER 1 INTRODUCTI ON 1.1 Background The 2008 American Community Survey (ACS) shows that public transportation is used as the main mode of travel to work by only 5 % of the population of age greater than 16. But still, public transportation serves millions of people in the United States as the only means of transportation. Given the demand for public transportation, Transit agencies strive to benefit every segment of American society individuals, families, communities, and businesses by providing efficient and convenient t ransit services. As per the statute detailed in Public Transit 1473.001, all transit agencies in Florida are required to provide Transit Development Plans (TDPs). TDP is a planning document which includes ridership forecasts for the following ten years using the transit demand estimation tool that is either approved or provided by the Florida Department of Transportation (FDOT ). Transit Boardings Estimation and Simulation Tool (TBEST) is the travel demand forecasting tool for public transportation developed by the FDOT Public Transit Office (PTO) to help transit agencies in completing their TDPs 1.2 TBEST TBEST is a comprehensive third generation t ransit demand forecasting model which provides forecasts of ridership at each stop specific to route and di rection, thus making it more accurate and detailed as compared to other existing transit planning models. Stoplevel ridership can also be aggregated to route, segment, and system level. TBEST is truly user friendly as it is interfaced with ArcGIS which a llow s a user to easily

PAGE 11

2 change or edit the route and stop configuration. TBEST is capable of evaluating the impact of service span, frequency, fare pricing and speed on the transit ridership. T BEST accounts for spatial accessibility by considering circular buffers around individual stops to identify the market for the transit system. More details on TBEST m odel and its methodology are provided in Chapter 3. The ongoing project funded by FDOT related to TBEST, aims a t further enhancing the capabilities of the TBEST model based on additional opportunities identified by the research team. The project focuses on enhancing TBEST s capabilities in two specific areas. The f irst area includes enhancing the precision of socio demographic data by using disaggregate parcel level spatial representation to capture the activity levels in transit stop buffers. Currently the TBEST model uses the 2000 Census data at the block group level (aggregate) with an assumption of uniform spatial distribution of population over an en tire block group to capture the socioeconomic characteristics within the stop buffer. This aggregate level spatial representation does not completely capture the variation in land use within a transit stop buffer which could lead to the inaccurat e estima tion of activity levels. Disaggregating the block group level soci odemographic data to the parcel level should enhance the stop level predictive capability of TBEST as the parcel data gives a more realistic spatial distribution of each land use The other modification to TBEST involves improving the quality of data regarding trip attraction. At present employment and special generator1 1 Special generators are defined as land uses that do not generate or attract trips at the same rate as other land uses dummy variable are the only variables used in the TBEST model to measure transit trip attractiveness. We know that emplo yment may account for workers accessing a particular land use, but employmen t

PAGE 12

3 does not take into account customer s or visitor s accessing that land use as they vary depending on the activity levels at that land use. Similarly, the special generator dummy va riable does not take into account the activity levels at each special generator. Thus, we can say that employment and special generator dummy variable does not completely explain the activity levels and total trip attraction to a destination Strategies fo r enhancement of the data supporting trip attraction can be developed by exploring a better way to handle special generators such that they are defined in terms of trip attraction rather than as a dummy variable in the model. E mployment based trip rates an d trip rates obtained using the Institute of Transportation Engineers (ITE) trip generation manual and disaggregate parcel level data would help in developing strategies to improve the trip attractiveness of the TBEST model 1.3 Objectives Demographics an d socioeconomic characteristics such as population and employment are the primary inputs for the TBEST model (and practically for all travel models) to estimate potential transit users. Lack of precision in such input datasets would result in biased and inaccurate forecast s As the ongoing project on TBEST model enhancement aims at moving to parcel level data, one of the objectives of this paper is to disaggregate block group census d ata to the parcel level and to identify the differences in activity level s (population and employment) around transit stop s due to change in input data i.e. from aggregate census data to disaggregate parcel data. The second objective of this research is to explore possible options to improve trip attraction capabilities In or der to meet this objective, the study examines the

PAGE 13

4 e mployment based trip attraction and the develop ment of parcel land use based trip attraction using the ITE trip generation manual parcel level data and the 2001 National Household Travel Survey (NHT S) da ta for all the time periods used in the TBEST model. Th e study also focuses on improving the predictive capability of the TBEST model by exploring how special generators are dealt with in various regional travel demand models and transit analysis studies and develops strategies on how to deal with special generators in the TBEST model. 1.4 Methodology The current research study will perform exploratory analysis on the objectives listed above. The results obtained from the analysis performed will then be us ed to define suggestions that may b e implemented to further enhance the precision of TBEST model in the future. The explorative analysis will be performed on Duval County, Florida. ArcGIS 9.3 will be used to apply the disaggregate census data to parcel lev el and to capture the differences in activity level due to change in input data. T he exploratory analysis of the strategies for trip attraction capability enhancement will be performed using ArcGIS. 2000 census data, parcel data, InfoUSA employment data, t ransit network data, ITE trip generation manual and 2001 NHTS data will be used in this research. All these datasets will be discussed in detail in Chapter 2. 1.5 Outline of the Thesis This thesis contains six chapters. Chapter 1 provides the introduction of the TBEST Model and the ongoing TBEST enhancements. The first section introduces the importance of transit ridership forecasting tool such as TBEST. Following that is a brief introduction of the TBEST Model and identified opportunities for the enhancement of

PAGE 14

5 TBESTs capabilities. Chapter 2 provides the description of all the datasets used in this paper. Chapter 3 discusses the elements, methodology and the ongoing enhancements of the TBEST m odel in detail. Following the TBEST Model chapter, Chapter 4 pro vides the description of strategies used for disaggregating zonal social demographic data to the parcel level. It also discusses the methodology and results of the explorat ory analysis to capture the differences in activity levels (population and employment) using aggregate census data, disaggregate parcel data and InfoUSA employment data. Chapter 5 discusses the possible ways ( development of employment based trip attraction and, parcel land use based trip attraction and exploring how special generators are dealt with in the past studies ) of improving the trip attraction capability of the TBEST model and the results of the exploratory analysis Chapter 6 provides general conclusions based on the explorative analysis and suggestions for the TBEST model enhanc ement.

PAGE 15

6 CHAPTER 2 DATA DESCRIPTION To achieve the objectives mentioned in the introduction section, the following datasets were used in this study 2.1 Census 2000 Data Summary File 1 (SF 1) and Summary File 3 (SF 3) of the Census 2000 data made availab le by U.S. Census Bureau were used for this analysis. SF 1 contains data on age, sex, race, households, families, owned or rented and housing units collected from all people and housing units (100percent data). Whereas, SF 3 is a sample data collected fro m about 1 in 6 households and weighted to represent the total population. It consists of 813 detailed tables of Census 2000 social, economic and housing characteristics like education, employment status, income, value of housing unit, year structure built. Demographic and socioeconomic characteristics such as single family population, multifamily population, household size, median income etc were obtained from the detailed tables in SF 1 and SF 3 at the block group level. This data was then joined to the block group shape file obtained from Census 2000 TIGER/Line Data using the unique ID for each block group present in both datasets. The Duval County block group shape file consists of 423 block groups with their respective areas. 2.2 2009 Property Apprais al Data (Parcel Data) The 2009 Property Appraisal data for Duval County, FL was obtained from the Florida Department of Revenue (DOR). The data includes about 100 land uses broadly classified in to residential, industrial, commercial, agricultural, institu tional, government

PAGE 16

7 and miscellaneous categories based on the activity or use of the property. Property appraisal contains information on land use, property type, area, physical address, sale value, book value. Since the data describe properties based on their land use, each property can be called as a parcel2Table 1 Description of Variables in Property Appraisal Data (Pa rcel Data) Table 1 gives the list of variables available in the property appraisal data (parcel data) used in the analysis. The dataset also consist of many variables on sales value, just value, assessed value and property tax which were not required for this analysis. The dataset does not include any information on demographics and socioeconomic characteristics in each parcel. Parcel data can help in obtaining a more realistic spatial distribution of population around the transit stops a s the location of each land use is known. The dataset includes data for 90,742 parcels of which 75,342 are single family parcels, 2156 are multi family parcels and 13244 are nonresidential parcels. Parcels which were coded as vacant residential and the ones with missing information on land us e or property type were deleted for the analysis. Table 2 shows the list of land uses available in the parcel data 2 Parcel is defined as piece of land described based on the ownership or land use. Variable Name Variable Description PARCEL_ID Unique ID given to each Property (parcel) DOR_UC DOR Land Use Code LND_SQFOOT Land Square Footage EFF_YR_BLT Effective Year Built TOT_LVG_AREA Total Living or Usable Area NO_RES_UNITS Number of Buildings & Residential Units PHY_ADDR1, PHY_ADDR2, PHY_CITY and PHY_ZIP Physical Address of the Property CENSUS_BK Census Block Group

PAGE 17

8 Table 2 List of Land Uses Available in Parcel Data DOR Land Use Code PROPERTY TYPE Property Type Residential 000 Vacant Residential 001 Single Family 002 Mobile Home 003 Multi family 10 units or more 004 Condominiums 005 Cooperatives 006 Retirement Homes 007 Miscellaneous Residential (migrant camps, boarding homes, etc.) 008 Multi family less than 10 units 009 Undefined Rese r ved for Use by Department of Revenue Property Type Commercial 010 Vacant Commercial 011 Stores, one story 012 Mixed use store and office or store and residential or residential comb ination 013 Department Stores 014 Supermarkets 015 Regional Shopping Centers 016 Community Shopping Centers 017 Office buildings, non professional service buildings, one story 018 Office buildings, non professional service buildings, multi story 019 Professional service buildings 020 Airports (private or commercial), bus terminals, marine terminals, piers, marinas. 021 Restaurants, cafeterias 022 Drive in Restaurants 023 Financial institutions (banks, saving and loan companies, mortgage companies credit services) 024 Insurance company offices 025 Repair service shops (excluding automotive), radio and T.V. repair, refrigeration service, electric repair, laundries, Laundromats 026 Service stations 027 Auto sales, auto repair and storage, auto s ervice shops, body and fender shops, commercial garages, farm and machinery sales and services, auto rental, marine equipment, trailers and related equipment, mobile home sales motorcycles, construction vehicle sales. 028 Parking lots (commercial or patro n) mobile home parks 029 Wholesale outlets, produce houses, manufacturing outlets 030 Florist, greenhouses 031 Drive in theaters, open stadiums 032 Enclosed theaters, enclosed auditoriums 033 Nightclubs, cocktail lounges, bars 034 Bowling alleys, sk ating rinks, pool halls, enclosed arenas 035 Tourist attractions, permanent exhibits, other entertainment facilities, fairgrounds (privately owned). 036 Camps 037 Race tracks; horse, auto or dog 038 Golf courses, driving ranges 039 Hotels, motels Pro perty Type Industrial 040 Vacant Industrial 041 Light manufacturing, small equipment manufacturing plants, small machine shops, printing plants 042 Heavy industrial, heavy equipment manufacturing, large machine shops, foundries, steel fabricating plan ts, auto or aircraft plants 043 Lumber yards, sawmills, planing mills 044 Packing plants, fruit and vegetable packing plants, meat packing plants 045 Canneries, fruit and vegetable, bottlers and brewers distilleries, wineries 046 Other food processing, candy factories, bakeries, potato chip factories 047 Mineral processing, phosphate processing, cement plants, refineries, clay plants, rock and gravel plants. 048 Warehousing, distribution terminals, trucking terminals, van and storage warehousing 049 Open storage, new and used building supplies, junk yards, auto wrecking, fuel storage, equipment and material storage

PAGE 18

9 Table 2 Continued DOR Land Use Code PROPERTY TYPE Property Type Agricultural 050 Improved agricultural 051 Cropland soil capabi lity Class I 052 Cropland soil capability Class II 053 Cropland soil capability Class III 054 Timberland site index 90 and above 055 Timberland site index 80 to 89 056 Timberland site index 70 to 79 057 Timberland site index 60 to 69 058 Tim berland site index 50 to 59 059 Timberland not classified by site index to Pines 060 Grazing land soil capability Class I 06 1 Grazing land soil capability Class I1 062 Grazing land soil capability Class I11 063 Grazing land soil capability Class IV 064 Grazing land soil capability Class V 065 Grazing land soil capability Class VI 066 Orchard Groves, Citrus, etc. 067 Poultry, bees, tropical fish, rabbits, etc. 068 Dairies, feed lots 069 Ornamentals, miscellaneous agricultural Property Type In stitutional 070 Vacant 0 71 Churches 072 Private schools and colleges 073 Privately owned hospitals 074 Homes for the aged 075 Orphanages, other non profit or charitable services 076 Mortuaries, cemeteries, crematoriums 077 Clubs, lodges, union hal ls 078 Sanitariums, convalescent and rest homes 079 Cultural organizations, facilities Property Type Government 080 Undefined Reserved for future use 081 Military 082 Forest, parks, recreational areas 083 Public county schools 084 Colleges 0 85 Hospitals 086 Counties (other than public schools, colleges, hospitals) 087 State, other than forests, parks, recreational areas, colleges, hospitals 088 Federal, other than forests, parks, recreational areas, hospitals, colleges 089 Municipal, oth er than parks, recreational areas, colleges, hospitals Property Type Miscellaneous 090 Leasehold interests (government owned property leased by a non governmental lessee) 091 Utility, gas and electricity, telephone and telegraph, locally assessed rai lroads, water and sewer service, pipelines, canals, radio television 092 Mining lands, petroleum lands, or gas lands 093 Subsurface rights 094 Right of way, streets, roads, irrigation channel, ditch, etc. 095 Rivers and lakes, submerged lands 096 Sew age disposal, solid waste, borrow pits, drainage reservoirs, waste land, marsh, sand dunes, swamps 097 Outdoor recreational or parkland, or high water recharge subject to classified use assessment. Centrally Assessed (Unclassified) 098 Centrally assesse d Non Agricultural Acreage 099 Acreage not zoned agricultural

PAGE 19

10 2.3 2007 InfoUSA Employment Data The 2007 address based (disaggregate) employment data provided by InfoUSA was obtained for the entire state of Florida in a point layer shapefile format ( each point corresponds to an employer or business). The InfoUSA employment database is a comprehensive data base of around 14 million U.S. businesses and is continuously updated using public sources. For each address, InfoUSA provides the information on busine ss name, location, franchise code, industry classification code (Standard Industrial Classification (SIC) System and North American Industry Classification System (NAICS)), the sales volume Industrial Employment (SIC Code 1 to 39), Commercial Employment ( SIC Code 50 to 59), Service Employment (SIC Code 40 to 49, 60 to 99) and Total Employment (SIC Code 1 to 99). The 2007 InfoUSA employment data with 847,108 records in the entire state of Florida was used as the employment data for the analysis. The busine sses in the Duval County were selected using the variable County_Code in the InfoUSA data. The Duval County includes 39,649 employer or businesses which were used to obtain the total employment by type associated with each transit stop. 2.4 Jacksonville Transportation Authority (JTA) Transit Network Data The transit agency selected for this analysis is Jacksonville Transportation Authority (JTA). JTA operates transit service in the city of Jacksonville, Duval County, Florida and the surrounding area with 45 routes and about 60 39 stops. The Automatic Passenger Count (APC) data, Schedule data and TBEST model for public transit in Jacksonville was obtained from the JTA for the time period of five months from 5th May 2009 to 4th October 2009 (May Pick). The APC data contains stop arrival times and stop

PAGE 20

11 boardings per trip and per route. Since not all the buses in Jacksonville are equipped with APC, the daily boardings for different time periods of the day used in the TBEST model cannot be obtained directly The APC data was used to derive average boardings per vehicle arrival for different time periods of the day used in the TBEST model. In order to get the total daily boa rding, data on total number of vehicle arrivals for each stop was obtained from the JTA TBEST model. T here were no matching variables between the APC data and the JTA TBEST mode l and also t he transit network (routes and their respective stops) in JTA TBEST model did not match with t he APC data and Schedule data. Therefore, JTA TBEST model was digitized3Table 3 D etails of Routes Selected for the Exploratory Analysis based on the APC data and Schedule data and simultaneously a lookup table (one to one mapping) was prepared to match the stops in APC data with correspondin g stops in the JTA TBEST model Once all the routes and stops were digitized, transit route segments were created in a line layer format and their respective transit stops were created in point layer format. The lookup table was then used to get the total boarding at each stop for different time periods of the day used in the TBEST model. The analysis in this paper will be performed for four routes shown in Table 3 as this study aims at exploring the enhancement strategies 3 Stops and routes were added and removed from the JTA TBEST model to match the APC data and Schedule data. No Route No Route Description Number of Stops 1 R5 Murray Hill Regency FCCJ UNF 90 2 P7 Dunn FCCJ North/Normandy 125 3 U2 University Boulevard Connector 74 4 F1 Broadway Detroit/Florida Ave 80

PAGE 21

12 The routes shown in Figure 1 were selected such that none of this routes lies in a single land use (residential, commercial, industrial etc) zone i.e. there are different types of land uses along these routes. Also, population density and emplo yment density (shown in Figure 1) at census block level were taken into consideration while deciding the routes. Figure 1 Selected Routes and their Respective Stops

PAGE 22

13 2.5 Institute of Transportation Engineers (ITE) Trip Generatio n Manual 8th Edition The Trip Generation Manual 8th Edition made available by ITE includes a user's guide as well as two data volumes with land use descriptions, vehicle trip generation rates, equations and data plots. Volume I contains the trip generation rates, plots and equations for land uses 000 through 499. These include the categories Port and Terminal; Industrial/Agricultural; Residential; Lodging; and Recreational. V olume 2 contains the trip generation rates, plots and equations for land uses 500 through 999, which include Institutional; Medical; Office; Retail; and Services categories. These volumes include data f rom more than 4800 sites In this manual most of the trip rates are available for one or more of: (1) a weekday, (2) weekday AM peak one hour42.6 2001 National Household Travel Surve y (NHTS) Database (3) weekday PM peak one hour, (4) Saturday and (5) Sunday. The trip rates for each land use are available for various independent variable s like area (square foot, acres), employees, attendees etc. The 2001 National Household Travel Survey (NHTS) data made available by U.S. Department of Transportation/Federal Highway Administration, Bureau of Transportation Statistics (BTS) and National Highway Tra ffic Safety Administration (NHTSA) was used in this study. NHTS collects d ata on daily trips taken in a 24hour period and is organized into five different data files namely household file, person file, vehicle file, travel day trip file and long trip file Records from each data file can be linked to one another using the Household ID number. The 2001 NHTS contain data on the 69,817 households 160,758 persons 139,382 vehicles and 642,292 trips. The travel 4 The peak one hour trip rates ( for AM and PM peaks) are defined as the weighted average vehicle trip rate during the hour of highest volume of traffic entering and exiting the site (during the AM and PM hours).

PAGE 23

14 day trip file was used in this study which includes information on purpose of the trip (work, shopping, etc.), means of transportation used (car, bus, subway, walk, etc.) travel time, time of day and day of we ek when the trip took place. This information will be used to develop the trip rates for different time periods of the day used in the TBEST model

PAGE 24

15 CHAPTER 3 TBEST M ODEL TBEST modeling software was briefly introduced in Chapter 1. This chapter will describe the features and methodology of the TBEST model ( Xuehao et al., 2007 and TBEST 3.2 User Guide, 2009) It will also discuss the opportunities for the enhancement of the TBEST model. TBEST is a third generation transit planning tool developed by the Florida Department of Transportation which, provides forecasts of ridership at each stop specific to route and direction. The featu res of TBEST model are presented and described below. 3.1 Features of TBEST 1) Direct and Transfer Boardings 2) Time of Day Based Analysis 3) Spatial Accessibility (Socio Economic Characteristics) 4) Time Space Network Connectivity 5) Competing and Complementary System Effects 6) GIS Based Software Tool 7) Performance Measures One of the distinctive features of the TBEST Model is the fact that it distinguishes between direct and transfer boardings. Transit passengers are either transferring or boarding directly at any given st op. Distinguishing between these two is important because it provides users better understanding of the trip linking that is occurring. Methodology for distinguishing between direct and transfer boardings is as follows.

PAGE 25

16 First ly stops are categorized into following options, one with transfer opportunity and one without any transfer opportunity Using the data from the nontransfer stops, TBEST estimates the direct boardings model, then that model is applied to the transfer stops to estimate the boardings at the transfer opportunity stops. To estimate the transfer boardings, estimated direct boardings are subtracted from the total boardings. TBEST includes separate ridership estimation equations for each time of day and day of week. These times of day incorp orat ed in TBEST are shown in Table 4. Table 4 Definitions of Time Periods in TBEST Period No Name of the Time Period Time Interval 1 Weekday AM peak period 6:00 8:59 AM 2 Weekday off peak period 9:00 AM 2:59 PM 3 Weekday P M peak period 3:00 5:59 PM 4 Weekday night period 6:00 PM 5:59 AM (next day) 5 Saturday 12 midnight 11:59 PM 6 Sunday 12 midnight 11:59 PM To account for spatial accessibility, TBEST considers various characteristics such as age, income, auto a vailability, work status, race et c of the people in the circular buffer area around each stop. This in formation is used to determine ridership at each stop. TBEST considers the overall connectivity and time space accessibility of the transit system by meas uring the activity opportunities (population and employment) that can be reached within a certain tim e frame and number of transfers. As t he network connectivity i.e. schedule of the transit system may vary with the time of the day TBEST computes network accessibility for the temporal dimension The ability of the T BEST model to account for time space network connectivity and accessibility makes it the ideal tool for transit ridership forecasting The competing and complementary effects of the transit

PAGE 26

17 sys tem may affect and enhance the ridership at each stop. T BEST clearly accounts for both of these effects in computing stoplevel ridership. T BEST is interfaced with ArcGIS 9.3 which allows the user to change and edit the s ocio economic scenarios, supply at tributes, and route and stop configurations. This freedom makes T BEST a truly user friendly transit ridership forecasting tool. The output of the T BEST model gives estimates of several performance measures such as route miles, service miles, service hour s, boardings per service mile or hour, and average boardings per service run at the individual route level and for the whole system. These performance measures can be used to assess the impacts of various socio economic and supply sc enarios on system performance. Appendix A of the TBEST 3.2 User Guide available at http://www.tbest.org/ provides more detailed and complete description of the framework and TBEST methodology.

PAGE 27

18 3.2 Enhancements in TBEST Model To further improve the predictive capabilities of the TBEST model, following areas of improvement were identified by the research team 3.2.1 Parcel L evel Data Capability The first improvement area focuses on improving the precision of the input information that the TB EST model uses to determine the activity levels in the transit stop buffers. This can be achieved by using address level data at parcel level of geography instead of the currently used block group level data. Since, t here is strong relationship between transit uses and walking distance to transit stops (Sullivan, 1996; Neilson, 1972) using parcel level data will help in capturing actual accessibility of population and activities to the transit stop location. T his effort aims at developing a methodology for disaggregating block group level soci odemographic data to the parcel level and using the demographic data the disaggregate parcel level data can be used to enhance the stop level predictive capability of TBEST 3.2.2 Trip Attraction Capability This task focuses on enhancing the predictive capability of TBEST by improving the quality of data regarding trip attraction. Currently, employment and special generator dummy variable are the only vari ables used in TBEST to measure transit trip attractiveness. As d iscussed earlier these variables do not completely explain the activity levels and the trip attraction at a destination. The two possibilities for improving the trip att raction quantification are as follows:

PAGE 28

19 1) E xploring a better way to treat special generators such that they are defined in terms of trip attraction rather than as a dummy variable in the model 2) Using employment based trip rates or parcel land use based vehicle trip rates obtained using ITE trip generation manual 2001 NHTS data and the disagg regate parcel level data

PAGE 29

20 CHAPTER 4 EXP L ORING PARCEL LEVEL DATA CAPABILITY The first effort on improving the forecasting capability of the TBEST model is focused on enhancing the precision of input data by moving from aggregate block group level data to disaggregate parcel level data. The possible benefit of moving to parcel level data would be increase in the accuracy of capturing population distributions relative to the transit system. This chapter aims at disaggregat ing block group census data to the parcel level and identify ing the differences in activity levels (population and employment) within transit stop buffers due to change in input data i.e. from aggregate census data to disaggregate parcel data. To capture the differences in activity levels, single family population, multifamily population and total employment were obtained at the following two levels: 1) At the aggregate level, block group level census data and InfoUSA employment data aggregated at block group level w ere used with an assumption of uniform spatial distribution of population and employment over the entire block group. 2) At the disaggregate level, parcel level data with population for each parcel (assigned based on the strategies discussed in t able 5) and address level InfoUSA emplo yment data w ere used. 4.1 Methodology As discussed in the data section, parcel data does not have information on the number and characteristics of the population. As the main objective of this analysis is to compare the demographics and employment captured by using aggregate block group

PAGE 30

21 census data and disaggregate parcel data, demographics in each residential parcel is required to determine the population within the catchment area of the transit system (routes and stops). Therefore, population from the 2000 Census data at the block group level of geography is applied to each parcel54.1.1 Block Group Level Demographic Disaggregation Table 5 gives the strategies used for assigning block group level census population to each parcel. As the assignment strat egies are different for single family parcels ( l and use code 001 to 005), multi family parcels ( land use code 003 & 008) and retirement homes and miscellaneous residential parcels ( land use code 006 & 007), three point layer files for single family (75,342 parcels), multifamily (2,156 parcels) and retirement homes parcels (2 parcels) were created. Table 5 Strategies for Assigning Block Group Level Census Population to Parcels Land Use Code Residential Use Basis of Allocation Assi gnment formula 000 Vacant Residential Dwelling unit 0 001 Single Family Dwelling unit Block group single family population divided by sum of parcels per block group 002 Mobile Home Dwelling unit 004 Condominiums Dwelling unit 005 Cooperatives Dwelli ng unit 003 & 008 Multi family Dwelling unit Block group multi family population divided by total number of multifamily dwelling units in each block group times number of dwelling units in that parcel 006 & 007 Retirement Homes and Miscellaneous Resid ential (migrant camps, boarding homes, etc.) Square footage Block group quarters population divided by square footage times parcel square footage. Single family population and multi family population for each block group were obtained from 2000 Census dat a. Based on the uniform spatial distribution concept, single family 5 There is almost a decade difference between the datasets: 2000 census data and 2009 parcel data. This difference can be elim inated by using the information on the built year of the parcel. The information on built year is missing for almost 10,000 parcels in Duval county parcel data. Therefore, the results of this analysis need to be interpreted with caution.

PAGE 31

22 and multi family population for each block group was assigned to the single family and multifamily parcels within that block group. For single family parcels, spatial join analysis was performed between the block group census data and single family parcels to obtain the number of parcels in each block group. The population for each single family parcel was obtained by dividing the single family population of the block group in which the pa rcel is spatially distributed by the total number of single family parcels in that block group. Table 6 Frequency Distribution of Number of Dwelling Unit s in Multi F amily Parcels Number of Dwelling Units Frequency Percent Distributi on Average Area (Sq.ft.) 2 955 44.3 1920.13 3 335 15.5 2883.04 4 618 28.7 3571.49 5 64 3.0 4382.02 6 53 2.5 5100.13 7 13 .6 6239.23 8 49 2.3 5680.08 9 2 .1 6983.50 10 to 50 38 1.8 15026.58 Greater than 50 29 1.3 106460.69 Total 2156 100.0 For multifamily parcels, the frequency distribution of number of dwelling units and average area shown in table 6 were reviewed to see if there are large developments which might cross multiple block groups Based on this review, t he number of dwelling (resi dential) units for each parcel was used to obtain the population for each multi family parcel. For multifamily parcels, spatial join analysis was performed between the block group census data and multi family parcels to obtain the total number of multifa mily residential units in each block group. The population for each multi family parcel was

PAGE 32

23 obtained by dividing the multi family population of block group in which the parcel is spatially distributed by the total number of multifamily residential units in that block group and then multiplying the ratio with number of residential units in each parcel. These strategies can only be used to assign population to the parcels; as they do not differentially distribute the other social demographic characteristics of the household. To achieve the objectives of understanding how well the parcel data represents the demographics for the transit boarding models, the analysis was carried out at two levels: Stop level and Route level 4.1.2 Stop Level Analysis In stop le vel analysis, socio economic data is computed for each stop in the transit system. Stop level analysis was considered because transit use is highly related to accessibility of the population and activities to the transit stop. In this analysis, point layer file of transit stops is used and buffers are generated around each stop to capture population and employment in the catchment area of each stop which generally represent the market to the transit system. Three catchment areas (buffer) of 200 meters (1/8t h of a mile), 400 meters (1/4th of a mile) ( Murray, 1998; Murray, 2001and Xuehao et.al, 2007) and 800 meters (1/2 of a mile) ( Zhao 2003; McDonnell 2006 and Reese, 2007) were used for this analysis. In each of these buffers, single family population, mult i family population and total employment were obtained at both aggregate level and disaggregate level6At aggregate level, it is assumed that single family population, multi family population and employment are uniformly distributed over entire block group. For this 6 The overlappin g of the transit stop buffers is not considered in the computation of population and employment within each stop buffer. As the overlapping issue is very important for the transit demand modeling, the overall implication from this analysis may change.

PAGE 33

24 analysis, buffers were generated around each stop and union was performed between the stop buffers and the underlying block group level census polygon layer file. This divides the stop buffers into parts (fractions) based on the census block gr oup boundary (shown in Figure 2). The population and employment for each part (fraction) is calculated based on the areabased fractional computation and then aggregated for each stop to determine the single family population, multi family population and t otal employment captured by each stop buffer.

PAGE 34

25 At disaggregate level, the assumption of uniform spatial distribution of population and employment over entire block group is relaxed as the parcel data when used in conjunction with census data provides the population location and address level InfoUSA data provides the employment location. In this analysis, buffers were generated for each stop and the stop buffers were spati ally joined to the single family and multifamily parcel point layer files to determine the single family population and multifamily population in the stop buffer. Point layer file of 2007 InfoUSA employment data was also spatially joined to the stop buff ers to determine the total employment for each stop buffer. Figure 3 shows an example of the stop buffer and spatial distribution of single family parcels, multifamily parcels and InfoUSA employment data with in the stop buffer. Figure 2 Stop Level Analysis at Aggregate Level

PAGE 35

26 Figure 3 Stop Level Analysis at Disaggregate Level

PAGE 36

27 4.1.3 Route Level Analysis In route level analysis, socioeconomic data is computed for each route in the transit system. The route level analysis was also performed for three catchment areas (buffer) of 200 meters (1/8th of a mile), 400 meters (1/4th of a mile) and 800 meters (1/2 of a mile) of the routes In this analysis, line layer file of transit route segments is used and buffers are generated for each route to determine the population and employment in the catchment area of each route. Similar to the stop level analysis, single family population, multi family population and total employment within route buffers were obtained at aggregate level and disaggregate level. As the assumption of uniform spatial distribution exist at aggregate level, buffers were generated around each route segment and union was performed between the route buffers and the underlying block group level census polygon layer file. This divides the route buffers into parts (fractions) based on the census block group boundary (shown in Figure 4). Single family population, multifamily population and total employment falling within each route buffer was then determined based on the area based fractional computation.

PAGE 37

28 Figure 4 Route Level Analysis at Aggregate Level The assumption of uniform spatial distribution is relaxed at disaggregate level. In this analysis, buffers were generated for the route segments and were spatially joined to the single family and multifamily parcel point layer files to determine the single family population and multi family population within each route buffer. Point layer file of 2007 InfoUSA employment data was also spatially joined to the route buffers t o determine the total employment within each route buffer. Figure 5 shows an example of the route buffer and spatial distribution of single family parcels, multifamily parcels and InfoUSA employment data within the route buffer.

PAGE 38

29 Figure 5 Route Level Analysis at Disaggregate Level

PAGE 39

30 4.2 Results 4.2.1 Results of Stop Level Analysis This section describes the results of stop level analysis which determines population and employment in the catchment area of each stop at aggregate level (block group) a nd disaggregate s level (parcel) Case (a) Case (b) Figure 6 Exam ples E xplaining the Difference B etween Aggregate Level and Disaggregate Level Figure 6 explains the difference between aggregate and disaggr egate level with the help of two examples The selected stop in case (a) does not have any residential parcel within its buffer and therefore determines zero population at the disaggregate level. Whereas at aggregate level, the stop buffer will show some population as all the block groups overlapping the stop buffer will contribute population based on the assumption of uniform spatial distribution of population in each block group. In case (b), single family population obtained using the parcel data will be much higher than that

PAGE 40

31 obtained using block group census data as large number of single family parcels lie in the selected stop buffer. For simplicity, stop level analysis results for only route R5 are discussed in detail. Tables 7, 8 and 9 pre sent the single family population, multifamily population and employment at aggregate and disaggregate level for each stop buffer of route R5. The low value of single family population, multi family population and total employment at disaggregate level wh en compared to aggregate level can be explained using case (a) of figure 6. Similarly, the high value of single family population, multi family population and total employment at disaggregate level when compared to aggregate level can be explained using case (b) of figure 6. The absolute percent differences between the aggregate and disaggregate population and employment computed for each stop are shown in the tables 7 8 and 9. It was observed that absolute percent difference decreases with an increase in the size of the catchment area. The absolute percent differences are higher for single family population when compared to multifamily population. This indicates that the single family population is more affected as compared to mu l tifamily population by t he use of disaggregate parcel data. Also, the absolute percent differences indicate that the total employment is more affected as compared to population when population and employment within each stop buffer are captured at disaggregate level i.e. using pa rcel data and address level InfoUSA employment data.

PAGE 41

32 Table 7 Aggregate and Disaggregate Level Single Family Population Computed for Different Sizes of Catchment Area (Buffer) A round Route R5 Stops No Stop Name Single Family Popula tion Buffer 1 (1/8 th mile) Buffer 2 ( 1/4 th mile) Buffer 3 ( 1/2 mile) Aggregate level Disaggregate level Absolute % Diff Aggregate level Disaggregate level Absolute % Diff Aggregate level Disaggregate level Absolute % Diff 1 F.C.C.J. Kent Campus 145.96 6.41 95.60 597.59 414.28 30.67 2,551.77 2,557.66 0.23 2 Park St. & Glendale St. 166.12 236.52 42.38 661.60 681.24 2.97 2,510.87 2,368.47 5.67 3 Park St. & Pinegrove Ave. 175.48 251.89 43.54 630.27 832.62 32.11 2,725.09 2,911.83 6.85 4 Park St. & Van Wert Ave. 111.05 134.59 21.20 640.26 859.69 34.27 2,884.23 3,335.60 15.65 5 Park St. & Ingleside Ave. 222.60 232.26 4.34 799.90 974.67 21.85 2,871.20 3,798.79 32.31 6 Park St. & Talbot Ave. 241.49 224.30 7.12 854.62 1,159.66 35.69 2,705.77 3,5 70.86 31.97 7 Park St. & Edgewood Ave. 213.65 271.29 26.98 833.83 1,143.26 37.11 2,538.41 3,413.33 34.47 8 Park St. & Valencia Rd. 162.63 356.26 119.05 695.80 1,085.24 55.97 2,402.37 3,285.81 36.77 9 Park St. & Seminole Rd. 101.94 104.86 2.87 442.67 888 .48 100.71 2,150.32 3,081.25 43.29 10 Park St. & Aberdeen St. 46.52 159.75 243.43 216.80 612.51 182.52 1,413.13 2,204.38 55.99 11 Park St. & McDuff Ave. 78.30 103.23 31.84 339.79 394.37 16.06 1,815.57 2,925.24 61.12 12 Park St. & Willow Branch Ave. 83.4 0 53.78 35.51 341.76 504.86 47.72 1,695.51 3,080.42 81.68 13 Park St. & Cherry St. 88.82 93.75 5.55 358.14 638.25 78.21 1,663.42 2,558.49 53.81 14 Park St. & James St. 92.53 278.52 201.00 384.31 688.01 79.03 1,692.03 2,609.71 54.24 15 Park St. & King St 84.69 138.12 63.08 357.33 591.23 65.46 1,757.97 2,678.74 52.38 16 King St. & Oak St. 47.05 126.09 168.01 253.79 569.41 124.36 1,429.56 2,271.26 58.88 17 Riverside Ave. & Barrs St. 26.88 76.90 186.15 152.21 340.45 123.68 1,116.52 1,770.75 58.59 18 Rive rside Ave. & Stockton St. 24.09 104.42 333.47 113.82 258.72 127.30 994.53 1,355.61 36.31 19 Riverside Ave. & Osceola St. 19.81 94.31 376.09 114.74 334.21 191.26 829.17 1,172.11 41.36 20 Riverside Ave. & Copeland St. 18.08 93.15 415.22 99.87 230.95 131.24 698.09 973.54 39.46 21 Riverside Ave. & Goodwin St. 13.86 34.64 149.96 74.46 197.79 165.64 569.48 889.77 56.24 22 Riverside Ave. & Margaret St. 7.74 8.77 13.26 43.70 169.91 288.84 444.94 660.03 48.34 23 Riverside Ave. & Lomax St. 5.01 67.23 1,241.43 20 .91 102.31 389.23 300.05 503.92 67.94 24 Riverside Ave. & Post St. 5.01 0.00 100.00 20.81 102.31 391.69 292.74 388.75 32.80 25 Riverside Ave. & Riverside Park Pl 5.21 0.00 100.00 32.11 23.38 27.18 264.55 299.75 13.31 26 Riverside Ave. & Roselle St. 21. 33 0.00 100.00 60.19 0.00 100.00 196.96 182.46 7.36 27 Riverside Ave. & Edison Ave. 22.39 0.00 100.00 75.19 23.25 69.08 230.18 219.92 4.46 28 Riverside Ave. & Jackson St. 22.39 23.25 3.86 89.01 46.50 47.76 235.40 219.92 6.58 29 Riverside Ave. & Ston ewall St. 22.39 0.00 100.00 80.99 46.50 42.58 218.67 219.92 0.57 30 Pearl St. & Bay St. 0.00 0.00 0.00 1.50 0.00 0.00 64.62 0.00 100.00 31 Forsyth St. & Julia St. 0.00 0.00 0.00 0.93 0.00 0.00 45.78 0.00 100.00 32 Forsyth St. & Laura St. 0.00 0.00 0.00 0.13 0.00 0.00 41.10 17.00 58.64 33 Forsyth St. & Ocean St. 0.00 0.00 0.00 0.11 0.00 0.00 74.42 68.00 8.63 34 Newnan St. & Adams St. 0.00 0.00 0.00 8.31 0.00 100.00 149.30 80.25 46.25 35 Newnan St. & Duval St. 0.04 0.00 0.00 13.81 17.00 23.13 243.80 80. 25 67.08 36 Newnan St. & Ashley St. 1.90 0.00 100.00 19.98 17.00 14.90 386.31 407.81 5.57 37 Newnan St. & Beaver St. 2.61 0.00 100.00 24.48 17.00 30.56 453.08 504.56 11.36 38 F.C.C.J. Station 3.99 0.00 100.00 16.82 0.00 100.00 237.56 283.02 19.14 39 Re gency Square Hub 44.20 0.00 100.00 176.80 0.00 100.00 800.62 154.90 80.65 40 9451 S Regency Square Blvd. 51.15 0.00 100.00 204.03 0.00 100.00 1,028.19 263.81 74.34

PAGE 42

33 Table 7 C ontinued No Stop Name Single Family Population Buffer 1 (1/8 th mile) Buffer 2 ( 1/4 th mile) Buffer 3 ( 1/2 mile) Aggregate level Disaggregate level Absolute % Diff Aggregate level Disaggregate level Absolute % Diff Aggregate level Disaggregate level Absolute % Diff 41 9550 S. Regency Square Blvd. 51.15 0.00 100.00 204. 93 0.00 100.00 785.64 25.06 96.81 42 S. Regency Square Blvd. & Monument Rd. 51.15 0.00 100.00 204.93 0.00 100.00 791.67 824.99 4.21 43 355 Monument Rd. 51.15 0.00 100.00 204.93 0.00 100.00 811.95 889.47 9.55 44 445 Monument Rd. 51.15 0.00 100.00 204.93 0.00 100.00 820.41 889.47 8.42 45 514 Monument Rd. 51.15 0.00 100.00 204.93 25.06 87.77 820.41 826.83 0.78 46 544 Monument Rd. 51.15 0.00 100.00 204.93 12.53 93.89 821.04 87.69 89.32 47 989 Monument Rd. 51.15 0.00 100.00 204.93 0.00 100.00 822.32 87.69 89.34 48 Monument Rd. & Treddick Pkwy. 51.15 0.00 100.00 204.93 0.00 100.00 822.22 238.03 71.05 49 Monument Rd. & Lee Rd. 51.15 162.86 218.39 204.93 588.81 187.32 880.37 1,869.64 112.37 50 1431 Monument Rd. 51.15 187.92 267.38 219.63 400.89 82.53 962.97 1,971.18 104.70 51 1505 Monument Rd. 79.81 0.00 100.00 311.31 114.44 63.24 1,161.93 956.58 17.67 52 St. Johns Bluff Rd. & Monument Rd. 63.68 0.00 100.00 279.76 127.16 54.55 1,185.48 1,093.71 7.74 53 St. Johns Bluff Rd. & Causey Ln. 59.40 0.00 100.00 235.63 82.82 64.85 1,046.80 1,006.42 3.86 54 St. Johns Bluff Rd. & S. Akers Dr. 60.11 152.92 154.40 234.65 356.37 51.87 919.21 979.06 6.51 55 St. John's Bluff Rd. & Lone Star Rd. 60.33 146.40 142.66 236.11 483.53 104.79 913.47 1,042.64 14.14 56 850 St. Johns Bluff Rd. 60.55 76.29 26.01 236.34 343.32 45.26 899.17 1,029.74 14.52 57 St. Johns Bluff Rd. & Craig Industrial Dr. 59.93 38.15 36.35 235.73 152.59 35.27 906.09 584.92 35.45 58 St. Johns Bluff Rd. & Airport Terrace Dr. 60.42 38.15 36.87 236.92 101.72 57.06 1,205.97 355.85 70.49 59 St. Johns Bluff Rd. & Atlantic Blvd. 86.07 0.00 100.00 434.56 25.43 94.15 1,885.47 861.60 54.30 60 St. Johns Bluff Rd. & Theresa Dr. 192.79 0.00 100.00 757.10 344.33 54.52 2,626.81 2,035.48 22.51 61 St. Johns Bluff Rd. & Bradley Rd. 165.95 25.31 84.75 678.74 644.12 5.10 2,444.60 2,929.51 19.84 62 St. Johns Bluff Rd. & Lost Pine Dr. 71.20 14.01 80.32 369.17 93.30 74.73 1,783.01 2,132.50 19.60 63 St. Johns Bluff Rd. & Fraser Rd. 61.04 0.00 100.00 262.03 48.91 81.33 1,129.39 1,137.86 0.75 64 St. Johns Bluff Rd. & Alden Rd. 61.05 7.05 88.45 248.32 65.94 73.45 998.51 845.97 15.28 65 2656 St. Johns Bluff Rd. 53.09 0.00 100.00 223.72 7.05 96.85 944.76 723.41 23.43 66 St. Johns Bluff Rd. & Judicial Dr. 52.88 0.00 100.00 217.07 21.16 90.25 919.66 732.18 20.39 67 St. Johns Bluff Rd. & Saints Rd. 53.39 7.05 86.79 218.10 73.64 66.23 887.10 750.08 15.45 68 St. Johns Bluff Rd. & Beach Blvd. 35.33 0.00 100.00 147.20 0.00 100.00 652.52 118.23 81.88 69 Central Pkwy.& St. Johns Bluff Rd. 33.25 0.00 100.00 136.55 0.00 100.00 577.09 0.00 100.00 70 11655 Central Pkwy. 33.22 0.00 100.00 133.09 0.00 100.00 573.13 162.59 71.63 71 11710 Central Pkwy. 33.22 0.00 100.00 133.09 0.00 100.00 589.72 121.94 79.32 72 11818 Central Pkwy. 33.22 0.00 100.00 141.24 0.00 100.00 673.78 170.01 74.77 73 F.C.C.J. Southside Campus 68.42 0.00 100.00 269.92 80.02 70.36 965.09 927.14 3.93 74 Central Pkwy. & Beach Blvd. 62.28 0.00 100.00 227.28 60.97 73.17 849.91 951.65 11.97 75 Beach Blvd. & Central Pkwy. 47.64 0.00 100.00 197.06 0.00 100.00 786.97 503.53 36.02

PAGE 43

34 Table 7 C ontinued No Stop Name Single Family Population Buffer 1 (1/8 th mile) Buffer 2 ( 1/4 th mile) Buffer 3 ( 1/2 mile) Aggregate level Disaggregate level Absolute % Diff Aggregate leve l Disaggregate level Absolute % Diff Aggregate level Disaggregate level Absolute % Diff 76 12000 Beach Blvd. 42.02 31.66 24.65 173.63 379.77 118.73 724.68 1,078.77 48.86 77 Beach Blvd. & Sans Pareil St. 41.50 8.42 79.70 171.74 249.81 45.46 707.82 809.39 14.35 78 3694 Kernan Blvd. 33.22 0.00 100.00 145.12 67.28 53.64 649.44 1,630.94 151.13 79 Kernan Blvd. & Gehrig Dr. 33.22 116.63 251.09 133.09 975.48 632.94 576.14 2,775.30 381.71 80 Kernan Blvd. & Mantle Dr. 33.22 360.50 985.19 133.09 1,420.81 967. 54 547.70 2,933.57 435.61 81 Kernan Blvd. & Hunter's Haven Ln. 33.22 275.68 729.85 133.09 1,092.11 720.57 532.84 2,905.23 445.23 82 Kernan Blvd. & Blue Stream Dr. 33.22 106.03 219.17 133.09 561.96 322.24 532.82 2,216.03 315.90 83 Kernan Blvd. & First Co ast Technology Pkwy. 33.22 0.00 100.00 133.09 0.00 100.00 532.82 614.98 15.42 84 UNF Dr. & Alumni Dr. 33.22 0.00 100.00 133.09 0.00 100.00 532.82 0.00 100.00 85 U.N.F. Osprey Landing (U.N.F Dr.) 33.22 0.00 100.00 133.09 0.00 100.00 532.82 0.00 100.00 86 U.N.F. Library (U.N.F. Dr.) 33.22 0.00 100.00 133.09 0.00 100.00 534.90 0.00 100.00 87 U.N.F. Arena (U.N.F. Dr.) 33.22 0.00 100.00 133.09 0.00 100.00 532.82 0.00 100.00 88 Town Center & Brightman Bl 36.99 0.00 100.00 148.21 0.00 100.00 593.33 30.24 94.90 89 Town Crossing & Buckhead Branch 36.99 0.00 100.00 148.21 0.00 100.00 593.33 136.09 77.06 90 Town Center Mall 36.99 0.00 100.00 148.21 0.00 100.00 593.33 60.48 89.81 Average 56.30 55.88 129.37 231.68 255.18 111.98 992.28 1,138.02 60.98 Standard D eviation 51.24 90.04 200.15 349.51 727.83 1,102.17

PAGE 44

35 Table 8 Aggregate and Disaggregate Level Multi Family Population Computed for Different Siz es of Catchment Area (Buffer) A round Route R5 Stops No Stop Name Multi family pop ulation Buffer 1 (1/8 th mile) Buffer 2 ( 1/4 th mile) Buffer 3 ( 1/2 mile) Aggregate level Disaggreg ate level Absolute % Diff Aggregate level Disaggregate level Absolute % Diff Aggregate level Disaggregate level Absolute % Diff 1 F.C.C.J. Ke nt Campus 7.48 7.33 1.94 24.52 7.33 70.09 143.49 59.09 58.82 2 Park St. & Glendale St. 7.04 0.00 100.00 31.27 0.00 100.00 194.80 163.90 15.86 3 Park St. & Pinegrove Ave. 16.87 0.00 100.00 68.59 61.59 10.20 306.28 330.14 7.79 4 Park St. & Van Wert Ave. 3 4.73 32.14 7.44 116.80 153.34 31.28 393.79 478.93 21.62 5 Park St. & Ingleside Ave. 46.34 82.13 77.22 160.96 200.36 24.47 433.12 456.64 5.43 6 Park St. & Talbot Ave. 48.79 36.88 24.40 148.11 187.50 26.60 500.10 707.08 41.39 7 Park St. & Edgewood Ave. 36 .12 4.86 86.54 126.18 95.03 24.69 563.80 976.43 73.19 8 Park St. & Valencia Rd. 13.71 5.14 62.49 121.86 87.84 27.92 609.53 1,077.80 76.82 9 Park St. & Seminole Rd. 34.62 141.10 307.62 142.00 542.02 281.71 706.27 1,174.15 66.25 10 Park St. & Aberdeen St. 32.64 246.69 655.80 115.97 483.53 316.96 692.88 1,010.58 45.85 11 Park St. & McDuff Ave. 62.14 231.60 272.69 231.34 701.60 203.28 837.58 1,568.37 87.25 12 Park St. & Willow Branch Ave. 65.99 146.50 122.01 283.48 651.08 129.68 926.07 1,773.31 91.49 13 P ark St. & Cherry St. 79.97 108.21 35.30 313.50 837.19 167.05 1,061.32 1,942.98 83.07 14 Park St. & James St. 88.69 312.51 252.34 332.53 666.93 100.57 1,164.88 2,037.93 74.95 15 Park St. & King St. 85.60 180.31 110.64 338.95 704.27 107.78 1,269.25 2,192.4 6 72.74 16 King St. & Oak St. 59.36 172.21 190.13 281.26 622.56 121.35 1,185.41 2,017.23 70.17 17 Riverside Ave. & Barrs St. 15.19 16.65 9.62 160.47 437.65 172.73 1,117.63 1,964.52 75.78 18 Riverside Ave. & Stockton St. 14.30 47.79 234.19 93.98 285.44 2 03.72 1,066.39 1,981.81 85.84 19 Riverside Ave. & Osceola St. 23.20 172.63 644.00 116.83 391.89 235.42 916.76 1,698.04 85.22 20 Riverside Ave. & Copeland St. 26.80 210.22 684.35 128.95 376.62 192.06 795.94 1,539.37 93.40 21 Riverside Ave. & Goodwin St. 29.04 155.71 436.26 133.52 403.20 201.99 719.85 1,561.25 116.89 22 Riverside Ave. & Margaret St. 31.20 0.00 100.00 126.15 385.91 205.93 672.01 1,321.29 96.62 23 Riverside Ave. & Lomax St. 32.14 60.93 89.55 128.60 60.93 52.62 597.43 1,232.34 106.27 24 Ri verside Ave. & Post St. 32.14 60.93 89.55 127.69 60.93 52.28 562.09 1,202.09 113.86 25 Riverside Ave. & Riverside Park Pl 31.84 0.00 100.00 110.67 60.93 44.94 502.47 977.81 94.60 26 Riverside Ave. & Roselle St. 7.57 0.00 100.00 68.39 60.93 10.92 340.05 649.38 90.96 27 Riverside Ave. & Edison Ave. 5.98 0.00 100.00 45.80 0.00 100.00 279.45 179.86 35.64 28 Riverside Ave. & Jackson St. 5.98 0.00 100.00 24.99 49.71 98.92 208.12 58.00 72.13 29 Riverside Ave. & Stonewall St. 5.98 0.00 100.00 27.38 33.14 21.06 207.73 58.00 72.08 30 Pearl St. & Bay St. 25.89 0.00 100.00 96.69 0.00 100.00 349.05 0.00 100.00 31 Forsyth St. & Julia St. 25.89 0.00 100.00 99.38 0.00 100.00 437.52 0.00 100.00 32 Forsyth St. & Laura St. 25.89 0.00 100.00 103.11 0.00 100.00 574. 47 39.33 93.15 33 Forsyth St. & Ocean St. 25.89 0.00 100.00 103.20 0.00 100.00 628.23 118.00 81.22 34 Newnan St. & Adams St. 25.89 0.00 100.00 109.32 39.33 64.02 716.09 118.00 83.52 35 Newnan St. & Duval St. 25.71 0.00 100.00 169.34 39.33 76.77 963.67 1 18.00 87.76 36 Newnan St. & Ashley St. 63.01 0.00 100.00 264.91 39.33 85.15 1,201.06 882.13 26.55 37 Newnan St. & Beaver St. 78.30 0.00 100.00 300.83 39.33 86.92 1,255.98 903.90 28.03 38 F.C.C.J. Station 89.22 0.00 100.00 339.89 0.00 100.00 1,278.86 833 .03 34.86 39 Regency Square Hub 30.92 0.00 100.00 123.28 0.00 100.00 603.45 0.00 100.00 40 9451 S Regency Square Blvd. 45.85 0.00 100.00 181.76 0.00 100.00 702.82 0.00 100.00

PAGE 45

36 Table 8 Continued No Stop Name Multi family population Buffer 1 (1/8 th mi le) Buffer 2 ( 1/4 th mile) Buffer 3 ( 1/2 mile) Aggregate level Disaggregate level Absolute % Diff Aggregate level Disaggregate level Absolute % Diff Aggregate level Disaggregate level Absolute % Diff 41 9550 S. Regency Square Blvd. 45.85 0.00 100.00 183.69 0.00 100.00 660.70 0.00 100.00 42 S. Regency Square Blvd. & Monument Rd. 45.85 0.00 100.00 183.69 0.00 100.00 665.17 0.00 100.00 43 355 Monument Rd. 45.85 0.00 100.00 183.69 0.00 100.00 709.79 0.00 100.00 44 445 Monument Rd. 45.85 0.00 100.00 183.69 0.00 100.00 735.37 0.00 100.00 45 514 Monument Rd. 45.85 0.00 100.00 183.69 0.00 100.00 735.37 0.00 100.00 46 544 Monument Rd. 45.85 0.00 100.00 183.69 0.00 100.00 730.95 0.00 100.00 47 989 Monument Rd. 45.85 0.00 100.00 183.69 0.00 100.00 72 2.07 1,176.00 62.87 48 Monument Rd. & Treddick Pkwy. 45.85 0.00 100.00 183.69 0.00 100.00 722.73 1,176.00 62.72 49 Monument Rd. & Lee Rd. 45.85 0.00 100.00 183.69 0.00 100.00 639.42 0.00 100.00 50 1431 Monument Rd. 45.85 0.00 100.00 160.16 0.00 100.00 507.26 0.00 100.00 51 1505 Monument Rd. 0.00 0.00 0.00 0.73 0.00 0.00 131.91 0.00 100.00 52 St. Johns Bluff Rd. & Monument Rd. 1.02 0.00 0.00 4.95 0.00 100.00 33.88 0.00 100.00 53 St. Johns Bluff Rd. & Causey Ln. 0.96 0.00 0.00 4.00 0.00 100.00 33.56 0.00 100.00 54 St. Johns Bluff Rd. & S. Akers Dr. 0.92 0.00 0.00 3.99 0.00 100.00 35.75 0.00 100.00 55 St. John's Bluff Rd. & Lone Star Rd. 0.91 0.00 0.00 3.92 0.00 100.00 47.19 0.00 100.00 56 850 St. Johns Bluff Rd. 0.90 0.00 0.00 3.91 0.00 100.00 73.67 0 .00 100.00 57 St. Johns Bluff Rd. & Craig Industrial Dr. 0.93 0.00 0.00 3.94 0.00 100.00 98.09 0.00 100.00 58 St. Johns Bluff Rd. & Airport Terrace Dr. 0.91 0.00 0.00 3.88 0.00 100.00 210.97 0.00 100.00 59 St. Johns Bluff Rd. & Atlantic Blvd. 11.36 0.00 100.00 81.60 0.00 100.00 443.04 206.74 53.34 60 St. Johns Bluff Rd. & Theresa Dr. 57.15 0.00 100.00 214.11 0.00 100.00 689.49 454.83 34.03 61 St. Johns Bluff Rd. & Bradley Rd. 51.73 0.00 100.00 190.37 0.00 100.00 660.28 124.04 81.21 62 St. Johns Bluff Rd. & Lost Pine Dr. 9.11 0.00 100.00 62.85 0.00 100.00 392.62 0.00 100.00 63 St. Johns Bluff Rd. & Fraser Rd. 6.66 0.00 100.00 33.84 0.00 100.00 184.17 0.00 100.00 64 St. Johns Bluff Rd. & Alden Rd. 8.45 0.00 100.00 40.99 0.00 100.00 179.18 0.00 100.00 65 2656 St. Johns Bluff Rd. 11.76 0.00 100.00 50.50 0.00 100.00 199.92 0.00 100.00 66 St. Johns Bluff Rd. & Judicial Dr. 11.53 0.00 100.00 52.00 0.00 100.00 207.08 0.00 100.00 67 St. Johns Bluff Rd. & Saints Rd. 12.09 0.00 100.00 53.15 0.00 100.00 220.48 0.00 100.00 68 St. Johns Bluff Rd. & Beach Blvd. 12.24 0.00 100.00 50.97 0.00 100.00 202.63 0.00 100.00 69 Central Pkwy.& St. Johns Bluff Rd. 21.85 0.00 100.00 72.07 0.00 100.00 242.59 0.00 100.00 70 11655 Central Pkwy. 21.97 0.00 100.00 88.02 0.00 100.00 354.10 0.00 100.00 71 11710 Central Pkwy. 21.97 0.00 100.00 88.02 0.00 100.00 363.54 0.00 100.00 72 11818 Central Pkwy. 21.97 0.00 100.00 89.62 0.00 100.00 374.55 0.00 100.00 73 F.C.C.J. Southside Campus 28.88 0.00 100.00 108.51 0.00 100.00 357.81 0.00 100.00 74 Central Pkwy. & Beach Blvd. 27.67 0.00 100.00 106.51 0.00 100.00 395.04 0.00 100.00 75 Beach Blvd. & Central Pkwy. 24.80 0.00 100.00 100.57 0.00 100.00 379.26 0.00 100.00 76 12000 Beach Blvd. 15.62 0.00 100.00 62.10 0.00 100.00 277.85 0.00 100.00 77 Beach Blvd. & Sans Pareil St. 15.99 0.00 100.00 60.12 0.00 100.00 237.51 0.00 100.00

PAGE 46

37 Table 8 Continued No Stop Name Multi family population Buffer 1 (1/8 th mile) Buffer 2 ( 1/4 th mile) Buffer 3 ( 1/2 mile) Aggregate level Disaggregat e level Absolute % Diff Aggregate level Disaggregate level Absolute % Diff Aggregate level Disaggregate level Absolute % Diff 78 3694 Kernan Blvd. 21.97 0.00 100.00 95.45 0.00 100.00 414.41 0.00 100.00 79 Kernan Blvd. & Gehrig Dr. 21.97 0.00 100.00 88.02 0.00 100.00 379.13 0.00 100.00 80 Kernan Blvd. & Mantle Dr. 21.97 0.00 100.00 88.02 0.00 100.00 361.56 0.00 100.00 81 Kernan Blvd. & Hunter's Haven Ln. 21.97 0.00 100.00 88.02 0.00 100.00 352.39 0.00 100.00 82 Kernan Blvd. & Blue Stream Dr. 21.97 0.00 100.00 88.02 0.00 100.00 352.37 0.00 100.00 83 Kernan Blvd. & First Coast Technology Pkwy. 21.97 0.00 100.00 88.02 0.00 100.00 352.37 0.00 100.00 84 UNF Dr. & Alumni Dr. 21.97 0.00 100.00 88.02 0.00 100.00 352.37 0.00 100.00 85 U.N.F. Osprey Landi ng (U.N.F Dr.) 21.97 0.00 100.00 88.02 0.00 100.00 352.37 0.00 100.00 86 U.N.F. Library (U.N.F. Dr.) 21.97 0.00 100.00 88.02 0.00 100.00 342.79 0.00 100.00 87 U.N.F. Arena (U.N.F. Dr.) 21.97 0.00 100.00 88.02 0.00 100.00 352.37 0.00 100.00 88 Town Cente r & Brightman Bl 4.59 0.00 100.00 18.38 0.00 100.00 73.57 0.00 100.00 89 Town Crossing & Buckhead Branch 4.59 0.00 100.00 18.38 0.00 100.00 73.57 0.00 100.00 90 Town Center Mall 4.59 0.00 100.00 18.38 0.00 100.00 73.57 0.00 100.00 Average 28.17 27.03 1 16.60 115.71 97.41 102.77 504.15 428.23 85.12 Standard Deviation 21.48 65.27 82.53 202.32 322.57 657.40

PAGE 47

38 Table 9 Aggregate and Disaggregate Level Total Employment Computed for Different Si zes of Catchment Area (Buffer) A r ound Route R5 Stops No Stop Name Total Employment Buffer 1 (1/8 th mile) Buffer 2 (1/4 th mile) Buffer 3 (1/2 th mile) Aggregate level Disaggregate level Absolute % Diff Aggregate level Disaggregate level Absolute % Diff Aggregate level Disaggreg ate level Absolute % Diff 1 F.C.C.J. Kent Campus 40.52 225.00 455.26 141.47 271.00 91.56 506.20 378.00 25.33 2 Park St. & Glendale St. 37.69 0.00 100.00 147.71 236.00 59.77 590.94 603.00 2.04 3 Park St. & Pinegrove Ave. 24.35 2.00 91.79 120.18 20.00 8 3.36 544.90 454.00 16.68 4 Park St. & Van Wert Ave. 25.68 17.00 33.80 125.05 92.00 26.43 582.10 467.00 19.77 5 Park St. & Ingleside Ave. 58.94 77.00 30.65 187.63 104.00 44.57 526.64 586.00 11.27 6 Park St. & Talbot Ave. 68.53 27.00 60.60 192.88 138.00 2 8.45 549.98 602.00 9.46 7 Park St. & Edgewood Ave. 52.79 6.00 88.63 183.99 119.00 35.32 572.18 857.00 49.78 8 Park St. & Valencia Rd. 21.96 12.00 45.37 160.42 58.00 63.84 588.63 873.00 48.31 9 Park St. & Seminole Rd. 28.14 2.00 92.89 115.42 55.00 52.35 631.44 683.00 8.17 10 Park St. & Aberdeen St. 2.08 7.00 235.91 18.74 28.00 49.41 473.10 654.00 38.24 11 Park St. & McDuff Ave. 39.69 19.00 52.13 157.59 85.00 46.06 604.94 656.00 8.44 12 Park St. & Willow Branch Ave. 45.17 25.00 44.66 196.96 101.00 48.72 877.57 833.00 5.08 13 Park St. & Cherry St. 56.77 22.00 61.25 224.38 199.00 11.31 1692.21 5303.00 213.38 14 Park St. & James St. 64.21 115.00 79.11 331.74 432.00 30.22 2523.24 5867.00 132.52 15 Park St. & King St. 196.29 305.00 55.38 926.54 636.00 31.3 6 3267.69 6374.00 95.06 16 King St. & Oak St. 373.91 311.00 16.82 1357.69 5253.00 286.91 3801.06 6081.00 59.98 17 Riverside Ave. & Barrs St. 828.40 4707.00 468.20 2317.41 5358.00 131.21 4441.14 6591.00 48.41 18 Riverside Ave. & Stockton St. 741.06 4942. 00 566.88 2434.91 5340.00 119.31 4823.83 7189.00 49.03 19 Riverside Ave. & Osceola St. 330.58 443.00 34.01 1539.82 5379.00 249.33 5068.64 7539.00 48.74 20 Riverside Ave. & Copeland St. 164.69 213.00 29.33 942.09 1062.00 12.73 5039.77 7611.00 51.02 21 Ri verside Ave. & Goodwin St. 126.78 251.00 97.98 520.22 1255.00 141.24 4224.58 7785.00 84.28 22 Riverside Ave. & Margaret St. 131.78 383.00 190.63 461.48 1512.00 227.64 3086.60 3318.00 7.50 23 Riverside Ave. & Lomax St. 135.66 389.00 186.75 536.64 1379.00 156.97 2667.69 7337.00 175.03 24 Riverside Ave. & Post St. 135.66 503.00 270.78 570.14 1682.00 195.01 3528.72 7838.00 122.12 25 Riverside Ave. & Riverside Park Pl 143.02 462.00 223.03 983.98 5631.00 472.27 4459.32 8110.00 81.87 26 Riverside Ave. & Rose lle St. 733.13 4292.00 485.43 2011.70 5538.00 175.29 6421.19 8511.00 32.55 27 Riverside Ave. & Edison Ave. 771.65 4320.00 459.84 2560.93 5526.00 115.78 7693.26 9525.00 23.81 28 Riverside Ave. & Jackson St. 771.65 888.00 15.08 3066.86 3137.00 2.29 10865 .57 9683.00 10.88 29 Riverside Ave. & Stonewall St. 771.65 2362.00 206.10 3002.30 2940.00 2.08 13264.02 13843.00 4.37 30 Pearl St. & Bay St. 2179.65 5164.00 136.92 7803.53 12060.00 54.55 22427.47 32582.00 45.28 31 Forsyth St. & Julia St. 2179.65 1677. 00 23.06 8159.73 23225.00 184.63 23196.97 33806.00 45.73 32 Forsyth St. & Laura St. 2179.65 3359.00 54.11 8653.10 23559.00 172.26 26949.30 33555.00 24.51 33 Forsyth St. & Ocean St. 2179.65 863.00 60.41 8665.02 22952.00 164.88 28041.45 36578.00 30.44 34 Newnan St. & Adams St. 2179.65 1797.00 17.56 8394.41 8380.00 0.17 26972.60 32574.00 20.77 35 Newnan St. & Duval St. 2156.05 1915.00 11.18 7035.53 7978.00 13.40 24060.62 32594.00 35.47 36 Newnan St. & Ashley St. 1496.62 1663.00 11.12 5597.52 5878.00 5.01 20419.43 30716.00 50.43 37 Newnan St. & Beaver St. 1258.03 831.00 33.94 5155.92 5664.00 9.85 19258.92 28681.00 48.92 38 F.C.C.J. Station 589.74 735.00 24.63 3045.03 14486.00 375.73 15343.97 23948.00 56.07 39 Regency Square Hub 127.47 2365.00 1755.40 510 .83 3227.00 531.71 1998.29 6202.00 210.37 40 9451 S Regency Square Blvd. 123.44 0.00 100.00 495.07 384.00 22.43 1805.83 5994.00 231.92

PAGE 48

39 Table 9 Continued No Stop Name Total Employment Buffer 1 (1/8 th mile) Buffer 2 (1/4 th mile) Buffer 3 (1/2 th mile ) Aggregate level Disaggregate level Absolute % Diff Aggregate level Disaggregate level Absolute % Diff Aggregate level Disaggregate level Absolute % Diff 41 9550 S. Regency Square Blvd. 123.44 1534.00 1142.72 494.55 1874.00 278.93 1999.98 6228.0 0 211.40 42 S. Regency Square Blvd. & Monument Rd. 123.44 1033.00 736.85 494.55 1812.00 266.40 1995.41 6797.00 240.63 43 355 Monument Rd. 123.44 81.00 34.38 494.55 1609.00 225.35 1983.80 6543.00 229.82 44 445 Monument Rd. 123.44 235.00 90.38 494.55 855. 00 72.89 1979.85 3285.00 65.92 45 514 Monument Rd. 123.44 25.00 79.75 494.55 895.00 80.97 1979.85 3337.00 68.55 46 544 Monument Rd. 123.44 24.00 80.56 494.55 42.00 91.51 1967.20 1610.00 18.16 47 989 Monument Rd. 123.44 17.00 86.23 494.55 281.00 43.18 19 41.79 2275.00 17.16 48 Monument Rd. & Treddick Pkwy. 123.44 104.00 15.75 494.55 104.00 78.97 1943.69 1485.00 23.60 49 Monument Rd. & Lee Rd. 123.44 125.00 1.26 494.54 207.00 58.14 1814.32 302.00 83.35 50 1431 Monument Rd. 123.44 115.00 6.84 453.96 118.0 0 74.01 1586.30 531.00 66.53 51 1505 Monument Rd. 44.34 253.00 470.64 168.29 280.00 66.38 785.62 442.00 43.74 52 St. Johns Bluff Rd. & Monument Rd. 30.23 23.00 23.92 119.08 288.00 141.85 493.85 471.00 4.63 53 St. Johns Bluff Rd. & Causey Ln. 29.73 109.0 0 266.63 116.83 354.00 203.01 492.69 462.00 6.23 54 St. Johns Bluff Rd. & S. Akers Dr. 30.24 22.00 27.25 116.73 146.00 25.08 494.73 742.00 49.98 55 St. John's Bluff Rd. & Lone Star Rd. 30.40 48.00 57.91 117.77 260.00 120.76 515.93 536.00 3.89 56 850 St. Johns Bluff Rd. 30.55 238.00 679.04 117.94 296.00 150.98 563.46 430.00 23.69 57 St. Johns Bluff Rd. & Craig Industrial Dr. 30.11 31.00 2.96 117.50 96.00 18.30 597.66 515.00 13.83 58 St. Johns Bluff Rd. & Airport Terrace Dr. 30.46 22.00 27.78 118.36 48.0 0 59.44 637.95 1336.00 109.42 59 St. Johns Bluff Rd. & Atlantic Blvd. 34.81 139.00 299.35 147.05 812.00 452.20 686.48 1707.00 148.66 60 St. Johns Bluff Rd. & Theresa Dr. 51.23 73.00 42.51 196.64 865.00 339.90 730.84 1641.00 124.54 61 St. Johns Bluff Rd. & Bradley Rd. 53.24 90.00 69.03 197.41 122.00 38.20 752.05 342.00 54.52 62 St. Johns Bluff Rd. & Lost Pine Dr. 39.29 31.00 21.10 157.96 299.00 89.29 669.59 506.00 24.43 63 St. Johns Bluff Rd. & Fraser Rd. 39.16 313.00 699.22 155.62 350.00 124.91 713.75 558.00 21.82 64 St. Johns Bluff Rd. & Alden Rd. 44.05 14.00 68.22 191.49 210.00 9.67 802.62 661.00 17.65 65 2656 St. Johns Bluff Rd. 62.59 97.00 54.98 246.81 244.00 1.14 923.36 1678.00 81.73 66 St. Johns Bluff Rd. & Judicial Dr. 62.20 220.00 253.71 258. 86 415.00 60.32 972.84 1509.00 55.11 67 St. Johns Bluff Rd. & Saints Rd. 63.14 310.00 390.98 260.75 525.00 101.34 1048.28 1635.00 55.97 68 St. Johns Bluff Rd. & Beach Blvd. 73.01 298.00 308.16 284.90 1251.00 339.10 1146.26 3064.00 167.30 69 Central Pkwy .& St. Johns Bluff Rd. 43.27 42.00 2.93 221.26 680.00 207.34 1062.56 2420.00 127.75 70 11655 Central Pkwy. 42.90 160.00 272.92 171.89 1705.00 891.91 759.86 3075.00 304.68 71 11710 Central Pkwy. 42.90 1416.00 3200.36 171.89 2185.00 1171.16 765.94 3568.00 365.83 72 11818 Central Pkwy. 42.90 430.00 902.23 183.03 2545.00 1290.52 869.23 3777.00 334.52 73 F.C.C.J. Southside Campus 91.03 0.00 100.00 345.46 3.00 99.13 1110.54 1288.00 15.98 74 Central Pkwy. & Beach Blvd. 82.64 253.00 206.15 300.67 899.00 198.99 1080.12 3651.00 238.02 75 Beach Blvd. & Central Pkwy. 62.62 362.00 478.06 259.35 914.00 252.42 986.80 3667.00 271.61 76 12000 Beach Blvd. 37.78 152.00 302.35 155.38 520.00 234.67 712.24 1275.00 79.01 77 Beach Blvd. & Sans Pareil St. 38.08 203.00 433.10 149.37 347.00 132.31 604.15 568.00 5.98 78 3694 Kernan Blvd. 42.90 33.00 23.08 190.12 49.00 74.23 849.27 269.00 68.33 79 Kernan Blvd. & Gehrig Dr. 42.90 10.00 76.69 171.89 10.00 94.18 753.80 65.00 91.38 80 Kernan Blvd. & Mantle Dr. 42.90 0.00 100.00 171.89 4.00 97.67 710.70 27.00 96.20

PAGE 49

40 Table 9 Continued No Stop Name Total Employment Buffer 1 (1/8 th mile) Buffer 2 (1/4 th mile) Buffer 3 (1/2 th mile) Aggregate level Disaggregate level Absolute % Diff Aggregate level Disaggregate level Absolu te % Diff Aggregate level Disaggregate level Absolute % Diff 81 Kernan Blvd. & Hunter's Haven Ln. 42.90 4.00 90.68 171.89 6.00 96.51 688.18 78.00 88.67 82 Kernan Blvd. & Blue Stream Dr. 42.90 8.00 81.35 171.89 12.00 93.02 688.15 76.00 88.96 83 Kernan Blvd. & First Coast Technology Pkwy. 42.90 38.00 11.43 171.89 44.00 74.40 688.15 501.00 27.20 84 UNF Dr. & Alumni Dr. 42.90 0.00 100.00 171.89 55.00 68.00 688.15 515.00 25.16 85 U.N.F. Osprey Landing (U.N.F Dr.) 42.90 55.00 28.19 171.89 55.00 68.00 688. 15 55.00 92.01 86 U.N.F. Library (U.N.F. Dr.) 42.90 0.00 100.00 171.89 0.00 100.00 717.82 55.00 92.34 87 U.N.F. Arena (U.N.F. Dr.) 42.90 0.00 100.00 171.89 0.00 100.00 688.15 0.00 100.00 88 Town Center & Brightman Bl 96.71 42.00 56.57 387.47 90.00 76.77 1551.21 1709.00 10.17 89 Town Crossing & Buckhead Branch 96.71 84.00 13.15 387.47 773.00 99.50 1551.21 2803.00 80.70 90 Town Center Mall 96.71 223.00 130.57 387.47 1735.00 347.77 1551.21 2786.00 79.60 Average 297.65 609.28 220.61 1,132.28 2,318.31 155 .62 3,959.75 5,735.97 77.77 Standard Deviation 569.73 1173.30 2127.17 4710.90 6820.46 9228.99

PAGE 50

41 Figure 7 and 8 show graphs of aggregate and disaggregate level single family population and multi family population captured within 1/8th mile, 1/4th m ile and 1/2 mile stops buffer of route R5. The graphs show differences between aggregate and disaggregate population captured with each stop buffer. These differences between aggregate and disaggregate population can be clearly explained with the help of example s in f igure 6. Figure 9 shows graphs of aggregate and disaggregate level total employment captured within 1/8th mile, 1/4th mile and 1/2 mile stops buffer of route R5. The results indicate that aggregate level total employment is less than disaggrega te level total employment for most of the stops and for all sizes of the catchment area. Since high density of employment is observed along the route, the assumption of uniform spatial distribution of employment within each block group might be leading to the underestimation of employment at the aggregate level.

PAGE 51

42 Figure 7 Graph Showing Aggregate and Disaggregate Level Sin gle Family Population Computed U sing Stop Level Analysis for Different Si zes of Catchment Area (Buffer) A round Route R5 Stops

PAGE 52

43 Figure 8 Graph Showing Aggregate and Disaggregate Level Mu lti Family Population Computed U sing Stop Level Analysis for Different Si zes of Catchment Area (Buffer) A round Route R5 Stops

PAGE 53

44 Figure 9 Graph Showing Aggregate and Disaggregate L evel Total Employment Computed U sing Stop Level Analysis for Different Si zes of Catchment Area (Buffer) A round Route R5 Stops

PAGE 54

45 The percent differences between the aggregate and disaggregate level single fami ly population, multi family population and total employment for the stops of all four routes were also studied. This percent differences were classified into categorizes using the nested means classification technique7These tables help in understanding the variation in percentage difference between aggregate and disaggregate level single family population, multi family population and total employment. It is observed that the number of stops having percentage difference in negative range decreases with the increase in buffer size of the transit stops. Also, number of stops having 100% differences in multi family population are mor e as compared to single family population and total employment. As the multifamily parcels are few in number and are widely distributed, this variation can be explained. Table 10, 11 and 12 presents the dis tribution of number of stops within each percentage difference category of single family population, multi family population and total employment for different sizes of catchment area. In table 10, the percentage difference categories in negative range sig nify that the aggregate level single family population is higher than disaggregate level single family population. The category of 100% difference indicates that there is no single family population at disaggregate level. The category of % difference indicates that there is no difference between aggregate level and disaggregate level single family population. The percentage difference categories in positive range signify that the disaggregate level single family population is higher than aggregate level single family population. Similarly, the categories in table 11 and 12 can be explained. 7 In Nested Means Classification, mathematical mean of the attribute values is calculated and the data is separated into two classes based on the mean. Data is further classified by calculating the means of the values within these two categories.

PAGE 55

46 Table 10 D istribution of Number of Stops Within E ach Percen tage Difference C ategory of Single Family P opulation for Different Sizes of Catchment A rea Percent age Difference Categories (%) Buffer 1 (1/8 th mile) Buffer 2 (1/4 th mile) Buffer 3 (1/2 mile) Number of Stops Percent Distribution (%) Number of Stops Perce nt Distribution (%) Number of Stops Percent Distribution (%) 100 99 26.8 56 15.2 12 3.3 76 to 99 25 6.8 18 4.9 12 3.3 61 to 7 5 12 3.3 20 5.4 8 2.2 41 to 6 0 18 4.9 19 5.1 11 3.0 21 to 40 28 7.6 25 6.8 29 7.9 11 to 20 11 3.0 20 5.4 12 3.3 1 to 10 10 2.7 24 6.5 57 15.4 0 14 3.8 0 .0 0 .0 1 to 10 16 4.3 26 7.0 69 18.7 1 1 to 20 16 4.3 15 4.1 46 12.5 2 1 to 40 20 5.4 28 7.6 52 14.1 4 1 to 60 13 3.5 30 8.1 25 6.8 6 1 to 90 14 3.8 33 8.9 25 6.8 9 1 to 200 46 12.5 37 10.0 7 1.9 Greater than 2 00 27 7.3 18 4.9 4 1.1 Total 369 100.0 369 100.0 369 100.0 Table 11 Distribution of Number of Stops Within Each Percentage Difference C ategory of Mu lti Family P opulation for Different Sizes of Catchment A rea Percent age Difference Categories (%) Buffer 1 (1/8 th mile) Buffer 2 (1/4 th mile) Buffer 3 (1/2 mile) Number of Stops Percent Distribution (%) Number of Stops Percent Distribution Number of Stops Percent Distribution (%) 100 235 63.7 162 43.9 96 26.0 8 1 to 99 5 1.4 19 5. 1 23 6.2 6 1 to 80 4 1.1 20 5.4 14 3.8 36 to 60 14 3.8 14 3.8 24 6.5 1 to 3 5 6 1.6 13 3.5 54 14.6 0 18 4.9 15 4.1 1 .3 1 to 30 13 3.5 19 5.1 67 18.2 3 1 to 60 10 2.7 22 6.0 25 6.8 6 1 to 100 6 1.6 18 4.9 38 10.3 10 1 to 150 9 2.4 14 3.8 16 4.3 1 5 1 to 200 7 1.9 15 4.1 6 1.6 20 1 to 250 8 2.2 12 3.3 2 .5 Greater than 250 34 9.2 26 7.0 3 .8 Total 369 100.0 369 100.0 369 100.0

PAGE 56

47 Table 12 Distribution of Number of Stops Within Each Percentage Difference C ategory of Total Empl oyment fo r Different Sizes of Catchment A rea Percent age Difference Categories (%) Buffer 1 (1/8 th mile) Buffer 2 (1/4 th mile) Buffer 3 (1/2 mile) Number of Stops Percent Distribution (%) Number of Stops Percent Distribution Number of Stops Percent Distri bution (%) 100 18 4.9 5 1.4 1 .3 66 to 99 38 10.3 18 4.9 12 3.3 4 1 to 6 5 24 6.5 21 5.7 27 7.3 2 1 to 40 26 7.0 20 5.4 39 10.6 1 to 20 5 1.4 23 6.2 46 12.5 1 to 20 27 7.3 39 10.6 68 18.4 2 1 to 40 35 9.5 35 9.5 60 16.3 4 1 to 55 20 5.4 35 9.5 32 8.7 56 to 85 37 10.0 44 11.9 35 9.5 86 to 130 42 11.4 45 12.2 21 5.7 131 to 200 21 5.7 34 9.2 16 4.3 201 to 300 22 6.0 26 7.0 10 2.7 Greater than 300 54 14.6 24 6.5 2 .5 Total 369 100.0 369 100.0 369 100.0 The KS (Kolmogorove Smirnov) test8L inear regression analysis was also performed for each size of catchment area with total boarding at each stop as dependent variable and population and employment as independent variables. The results in table 13 show that disaggregate level population and employment have h igher t stats than aggregate level population and employment for 1/8th was performed to test if the population and employment obtained at both aggregate and disaggregate levels are statistically different. The KS test was carried out to test disaggregate and aggregate level single family population, multi family population and total employment for different sizes of catchment area. The results showed a lower pvalue of less than 0.05 resulting in rejection of the null hypothesis of no difference between aggregate and disaggregate level population and employment. 8 T he K olmogorovSmirnov test (KS test) is a non parametric test which tries to determine if two datasets differ significantly

PAGE 57

48 mile, 1/4th mile and 1/2 mile of catchment area. This indicates that population and employment captured at disaggregate level explain transit ridership at each stop more accurately as compared to population and employment at aggregate level. Infact the regression results at aggregate level for 1/8th mile, 1/4th mile and 1/2 mile of catchment area indicates a negative (but statistically insignificant) effect of population on transit boardings. Such unusual results can be avoided by capturing the population at disaggregate level. Table 13 Parameter Estimates (t stats) of the Linear Regr ession Analysis B etween Total Boarding, Population and Employment Within Each Sto p Buffer Independent Variable Buffer 1 (1/8th mile) Buffer 2 (1/4th mile) Buffer 3 (1/2 mile) Aggregate level Disaggregate level Aggregate level Disaggregate level Aggregate level Disaggregate level Constant 9.303 (2.20) 3.127(1.55) 8.712 (1.93) 6.46 8 (2.23) 5.550 (1.08) 3.939 (1.07) Population in stop buffer -0.056 ( -0.57) 0.066 (0.98) -0.033 ( -0.33) 0.054 (0.57) 0.051(0.50) 0.121 (1.22) Total Employment in stop buffer 0.496 (5.09) 0.794 (11.69) 0.480 (4.84) 0.538 (5.74) 0.436 (4.31) 0.468 (4.72) R2 0.27 0.61 0.24 0.53 0.18 0.21

PAGE 58

49 4.2.2 Results of Route Level Analysis This section describes the results of route level analysis based on the population (single family and multifamily) and employment obtained for each route for different sizes (1/8th mile, 1/4th mile and half mile) of catchment area at aggregate and disaggregate level. Figure 10 show graphs comparing aggregate and disaggregate level single family population and multi family population for each route buffer. The graphs show large diffe rences in the single family population and multi family population obtained at both levels. Multifamily population seems to be more affected as compared to single family population for all sizes of catchment area. Figure 11 shows the plot of total employm ent captured at aggregate and disaggregate level for each route buffer. The graphs indicate that the total employment at the aggregate level is less than the total employment at the disaggregate level for different sizes of catchment area of all the transi t route s As discussed in the results of stop level analysis, assumption of uniform spatial distribution at aggregate level and concentration of employment along the transit route leads to this underestimation. The graphs also show that the differences in aggregate and disaggregate level single family population, multifamily population and total employment reduces as the size of the catchment area increases. Table 14 presents the absolute percent difference between aggregate level and disaggregate level s ingle family population, multi family population and employment obtained for different sizes of catchment area (buffer). The results indicate that there is an average difference of about 20 % in aggregate level and disaggregate level single family populati on captured within 1/8th mile catchment area of a route. This average difference in single family population reduces to about 9 % and 5% as the catchment area of the

PAGE 59

50 transit route increases to quarter mile and half mile respectively. Similarly, the average absolute percent difference reduces for multi family population and total employment as the size of the catchment area increases. The average absolute percent differences also indicate that the total employment is more affected as compared to population ( Pascoe, 2007). These differences suggest the need of capturing the demographics at disaggregate level i.e. using parcel data in the catchment area of a transit service.

PAGE 60

51 Single Family Population in 1/8 th mile Buffer Multi Family Population in 1/8 th mile Buffer Single Family Population in 1/4th mile Buffer Multi Family Population in 1/4th mile Buffer Single Family Population in 1/2 mile Buffer Multi Family Population in 1/2 mile Buffer Figure 10 Graphs S howi ng Aggregate and Disaggregate Level Single Family and Mu lti Family Population Computed U sing Route Level Analysis for Different Sizes of Catchment Area (Buffer)

PAGE 61

52 Figure 11 Graph s S howing A ggregate and D isaggregate L evel Total Employment Computed U sing Route Level Analysis for Different Sizes of Catchment Area (Buff er) Total Employment in 1/8 th mile Buffer for all Routes Total Employment in 1/4 th mile Buff er for all Routes Total Employment in 1/2 mile Buffer for all Routes

PAGE 62

53 Table 14 Absolute Percentage Difference Between Aggregate Level and Disaggregate L evel Single Family Population, MultiFamily Population and Employ ment Obtained U sing Route Level Analysis for Different Sizes of Catchment Area (B uffer) Route No Route Description Buffer 1 (1/8th mile) Buffer 2 (1/4th mile) Buffer 3 (1/2th mile) Absolute Percent Difference in Single Family Population Absolute Percent Difference in Multi Family Population Absolute Percent Difference in Total Emp loyment Absolute Percent Difference in Single Family Population Absolute Percent Difference in Multi Family Population Absolute Percent Difference in Total Employment Absolute Percent Difference in Single Family Population Absolute Percent Difference in Mu lti Family Population Absolute Percent Difference in Total Employment R5 Murray Hill Regency -FCCJ UNF 29.73 65.21 78.46 9.65 58.69 62.95 0.04 35.07 21.17 P7 Dunn FCCJ North/Normandy 14.68 13.67 49.04 0.17 11.46 42.14 4.02 3.4 7 20.27 U2 University Boulevard Connector 6.15 56.88 128.01 9.13 3.31 60.69 9.23 25.52 21.74 F1 Broadway Detroit/Florida Ave 31.12 31.52 50.87 15.98 30.83 31.57 8.45 10.38 21.99 Average 20.42 41.82 76.59 8.73 26.07 49.34 5.43 18.61 21.29

PAGE 63

54 CHAPTER 5 TRIP ATTRACTION CAPABILITY ENHANCEMENT This chapter discusses the possibilities which were explored to improve the trip attraction capabilities of the TBEST model. 5.1 Em ployment B ased Trip Attraction This section discusses the strategy used to refine employment data for accurately capturing the trip attraction f rom nonresidential land uses. The strategy used aims at taking into account the trips due to employment by developing the trip attraction per employment type. Table 15 shows the trips attraction per employee for each type of employment and each TBEST time period. The values used in table 15 reflect project team judgment and knowledge of travel behavior and are not empirically derived. These va lues will be applied to the InfoUSA employment categ ories for each nonresidential land use (DOR land use code greater than 10 ) shown in table 2. Table 15 Trips Attraction per Employee for Each Type of Employment and Each TBEST Time Period Trip Attractions/Productions Assigned per Em ployee Industrial Commercial Service Am Peak .5 .4 .5 Midday .1 .2 .2 Pm Peak .5 .4 .5 Evening .3 .2 .2 Saturday .1 .2 .2 Sunday .1 .2 .2 Total

PAGE 64

55 5.2 Development of Parcel Land Use Based Trip Attraction/Production In addition to using the emplo yment based trip attraction noted above, TBEST can also be enhanced by adding information about parcel land use. Th e study aims at improving the predictive capability of the TBEST model by developing parcel land use based trip attraction (instead of employ ment) using the ITE trip generation manual and the NHTS 2001 database The nonresidential land use categories (DOR land use code greater than 10) in the parcel data ( shown in table 2 ) w ere used as a foundation to develop a strategy to match parcel land use classification with IT E land uses Table 16 gives the trip rates of parcel level land use for TBEST time periods using ITE s trip generation manual and the NHTS 2001 database The first part of the table with the heading TRIP RATE FROM ITE TRIP GENERA TION MANUAL gives the trip rate for each land use using the ITE trip generation manual, 8th Edition. In this manual most of the trip rates are available for one or more of: (1) a weekday, (2) weekday AM peak one hour, (3) weekday PM peak one hour, (4) Sat urday and (5) Sunday. The methodology used to obtain trip rates for each parcel land use category is as follows: 1) The trip rates were obtained by matching each land use category with the closest available ITE land use category (one to one mapping) 2) Sever al parcel level land use codes include multiple ITE landuse categories under one single (parcel level) land use code. In such cases of a one to many correspondence from parcel level land use co des to ITE land use categories, the ITE trip generation rates were averaged across the land uses. For example, f lorists

PAGE 65

56 and g reen houses are included within a single parcel level land use code (030)93) Under some parcel level land use codes, however, rather disparate types of landuses are clu bbed. For example, airports, marinas, and other water terminals were classified into a single parcel level land use code (020). These landuses are significantly different from each other in terms of their trip generation characteristics. In such cases, th e table provides separate trip rates for each of the land uses. The trip rate for this land use code was obtained by taking an average of the ITE trip rate for florists and the ITE trip rate for greenhouses. Same strategy has been used for several other landuse categories such as motels and hotels (039), and auto sales and auto parts (027). 4) For parcel level land use codes such as restaurants and parks which are classified into many types in the ITE trip generation manual (i.e., a many to one correspondence), the maximum value of the trip rates of the different ITE land uses is reported. For example, ITE trip rates are available for two types of restaurants (021) high quality restaurants and highturnover restaurants. The trip rate of high turnover restaurants (which is higher than that of the highquality restaurants) is reported in this case. 5) Several parcel level land use categories do not have trip rates available by square footage in the in ITE trip gener ation manual. For example, the a irports category (020) does not have tri p rates per square footage. Therefore, trip rates for such land use categories are given with respect to other variables available in the ITE Trip generation manual. Other land uses such as service stations (026), race tracks 9 Numbers in parentheses show the parcel level land use code

PAGE 66

57 (037), golf courses (038), hot els & motels (039), homes for the aged (074) and military base (081) have the same issue. 6) If the trip rate is not available for a particular time period, it is marked as NA Not Available in the cell corresponding to that land use and time period.

PAGE 67

58 Table 16 Trip Rates of Parcel Level Land Use for TBEST Time Periods Using ITE Trip Generation Manual and NHTS 2001 Database DOR code PROPERTY TYPE TRIP RATE FROM ITE TRIP GENERATION MANUAL TRIP RATE BASED ON ITE TRIP MANUAL AND NHTS 200 1 DATABASE FOR TBEST TIME PERIODS Unit (Indepen dent Variable) Week day Week day AM Peak Hour Week day PM Peak Hour Satur day Sun day Weekday AM Peak Period Weekday PM Peak Period Week day Off Peak Period Week day Night Period Satur day Sun day Commerc ial Using temporal distributi on of trips in NHTS Using Peak Factor s Using temporal distributi on of trips in NHTS Using Peak Factor s Using temporal distributi on of trips in NHTS Using Peak Factor s Using temporal distributi on of trips in NHTS Using Pea k Factor s 10 Vacant Commercial 11 Stores, one story 1000 Sq.ft GFA 22.88 2.14 2.81 25.4 NA 3.85 4.37 5.67 8.15 8.10 6.27 5.29 4.09 25.40 4.54 12 Mixed use store and office or store and residential or residential combination 1000 Sq.ft GFA 17.225 1.97 2.27 NA NA 2.90 4.02 4.27 6.58 6.10 4.01 3.98 2.61 3.05 3.42 13 Department Stores 1000 Sq.ft GFA 22.88 2.14 2.81 25.4 NA 3.85 4.37 5.67 8.15 8.10 6.27 5.29 4.09 25.4 4.54 14 Supermarkets 1000 Sq.ft GFA 102.24 10.05 11.85 177.59 166.4 4 17.19 20.53 25.36 34.37 36.19 28.66 23.62 18.69 177.59 166.44 15 Regional Shopping Centers 1000 Sq.ft GFA 42.94 1 3.73 49.97 25.24 7.22 2.04 10.65 10.82 15.20 18.21 9.92 11.87 49.97 25.24 16 Community Shopping Centers 1000 Sq.ft GFA 42.94 1 3.73 49.97 25.24 7.22 2.04 10.65 10.82 15.20 18.21 9.92 11.87 49.97 25.24 17 Office buildings, non professional service buildings, one story 1000 Sq.ft GFA 11.57 1.8 1.73 NA NA 1.94 3.68 2.87 5.02 4.10 1.74 2.67 1.13 2.05 2.30 18 Office buildings, nonprofessional service buildings, multistory 1000 Sq.ft GFA 23.14 3.6 3.46 NA NA 3.89 7.35 5.74 10.03 8.19 3.48 5.35 2.27 4.10 4.59 19 Professional service buildings 1000 Sq.ft GFA 11.01 1.55 1.49 2.37 0.98 1.85 3.17 2.73 4.32 3.90 2.13 2.54 1.39 2.37 0.98 20 Airports Employee s 13.4 1.21 1 12.2 14.7 2.25 2.47 3.32 2.90 4.74 4.86 3.10 5.29 12.2 14.7 Marine terminals, piers, marinas 1000 Sq.ft 0.48 NA NA 0.57 0.79 0.08 NA 0.12 NA 0.17 NA 0.11 NA 0.57 0.79 21 Restaurants, cafeterias 1000 Sq.ft GFA 127.15 13.53 18.49 158.37 131.8 4 21.37 27.64 31.53 53.62 45.01 27.78 29.37 18.11 158.37 131.84 22 Drive in Restaurants 1000 Sq.ft GFA 496.12 54.81 46.14 722.03 542.7 2 83.40 111.98 123.04 133.81 175.63 151.53 114.60 98.81 722.03 542.72 23 Financial institutions 1000 Sq.ft GFA 148.15 17.31 26.69 86.32 31.9 24.90 35.36 36.74 77.40 52.45 21.42 34.22 13.97 86.32 31.9 24 Insurance company offices 25 Repair service shops (excluding automotive) 1000 Sq.ft GFA 44.32 6.84 5.02 42.04 26.43 7.45 13.97 10.99 14.56 15.69 9.56 10.24 6.23 42.04 26.43

PAGE 68

59 Table 16 Continued DO R code PROPERTY TYPE TRIP RATE FROM ITE TRIP GENERATION MANUAL TRIP RATE BASED ON ITE TRIP MANUAL AND NHTS 2001 DATABASE FOR TBEST TIME PERIODS Unit (Indepen dent Variable) Week day Wee k day AM Peak Hour Week day PM Peak Hour Satur day Sun day Weekday AM Peak Period Weekday PM Peak Period Week day Off Peak Period Week day Night Period Satur day Sun day Commercial Using temporal distributio n of trips in NHTS Using Peak Factor s Using temporal distributi on of trips in NHTS Using Peak Factor s Using tempora l distributi on of trips in NHTS Using Peak Factor s Using temporal distributi on of trips in NHTS Using Peak Factor s 26 Service stations Vehicle Fueling Positions 168.56 12.58 15.65 NA NA 28.33 25.70 41.80 45.39 59.67 59.00 38.94 38.47 29.87 33.44 27 Aut o sales, auto repair and storage 1000 Sq.ft GFA 47.625 3.31 4.61 21.03 10.48 8.01 6.76 11.81 13.37 16.86 16.64 11.00 10.85 21.03 10.48 Auto service shops, commercial garages. 1000 Sq.ft GLA NA 3.22 4.01 15.86 2.59 2.19 6.58 3.23 11.63 4.62 NA 3.01 NA 15. 86 2.59 28 Parking lots (commercial or patron) mobile home parks 1000 Sq.ft 0. 91 0.08 0.105 0.83 0.74 0.15 0.16 0.23 0.30 0.32 0.27 0.21 0.17 0.83 0.74 29 Wholesale manufacturing outlets, produce houses, 1000 Sq.ft GFA 6.73 0.58 0.52 1.59 2.3 1.13 1. 18 1.67 1.51 2.38 2.44 1.55 1.59 1.59 2.3 30 Florist, greenhouses 1000 Sq.ft 40.2 5.63 4.99 57.38 39.45 6.76 11.50 9.97 14.47 14.23 8.61 9.29 5.62 57.38 39.45 31 Drive in theaters, open stadiums 1000 Sq.ft 0.765 NA NA NA NA 0.13 NA 0.19 NA 0.27 NA 0.18 NA 0.14 0.15 32 Enclosed theaters, enclosed auditoriums 1000 Sq.ft GFA NA NA 26.7 99.28 81.9 n/a10n/a NA 77.43 NA 110.64 NA 72.10 99.28 81.9 33 Nightclubs, cocktail lounges, bars 1000 Sq.ft GFA NA NA 15.49 NA NA n/a n/a NA 44.92 n/a n/a NA 41.83 32.14 35.99 34 Bowling alleys, pool halls 1000 Sq.ft GFA 33.33 3.13 3.54 NA NA 5.60 6.39 8.27 10.27 11.80 10.09 7.70 6.58 5.91 6.61 Enclosed arenas 1000 Sq.ft 0.765 NA NA NA NA 0.13 NA 0.19 NA 0.27 0.18 0.14 0.15 Skating rinks 1000 Sq.ft GFA NA NA 2.36 NA NA 4.65 NA NA 6.84 9.78 9.78 6.37 6.37 4.90 5.48 35 Tourist attractions, permanent exhibits 1000 Sq.ft 2.075 0.066 0.265 2.24 1.871 0.35 0.13 0.51 0.77 0.73 0.71 0.48 0.46 2.24 1.87 36 Camps 1000 Sq.ft NA 0.012 0.024 NA NA NA 0.02 NA 0.07 0.10 0.10 0.06 0.06 0.05 0.06 37 Race tracks; horse 1000 Sq.ft 0.987 NA NA NA NA 0.17 NA 0.25 NA 0.35 NA 0.23 NA 0.17 0.20 Race tracks; Auto Attendees NA NA NA 0.28 NA Race tracks; Dog Attendees NA NA 0.41 NA NA 38 Golf courses, driving ra nges Employee s 55.57 4.14 6.71 72 58.29 9.34 8.46 13.78 19.46 19.67 16.74 12.84 10.91 72 58.29 39 Hotels, motels Employee s 42.74 1.16 1.24 12.4 10.37 7.18 2.37 10.60 3.60 15.13 22.26 9.87 14.51 12.4 10.37 10 n/a (not applicable) is equivalent to a zero trip rate

PAGE 69

60 Table 16 Continued DOR code PROPERTY TYPE TRIP RATE FROM ITE TRIP GENERATION MANUAL TRIP RATE BASED ON ITE TRIP MANUAL AND NHTS 2001 DATABASE FOR TBEST TIME PERIODS Unit (Indepen dent Variable) Week day Week day AM Peak Hour Week day PM Peak Hour Satu rday Sun day Weekday AM Peak Period Weekday PM Peak Period Week day Off Peak Period Week day Night Period Satur day Sun day Industrial Using temporal distributio n of trips in NHTS Using Peak Factor s Using temporal distributi on of trips in NHTS Using Peak Factor s Using temporal distributi on of trips in NHTS Using Peak Factor s Using temporal distributi on of trips in NHTS Using Peak Factor s 40 Vacant Industrial 41 Light manufacturing, small equipment manufacturing plants, small machine shops, instrument manufacturing pr inting plants 1000 Sq.ft GFA 6.97 1.01 1.08 1.32 0.68 1.17 2.06 1.73 3.13 2.47 1.07 1.61 0.70 1.32 0.68 42 Heavy industrial, heavy equipment manufacturing, large machine shops, foundries, steel fabricating plants, auto or aircraft plants 1000 Sq.ft GFA 1. 5 0.69 0.68 NA NA 0.25 1.41 0.37 1.97 0.53 NA 0.35 NA 0.27 0.30 43 Lumber yards, sawmills, planing mills 1000 Sq.ft GFA 1.5 0.69 0.68 NA NA 0.25 1.41 0.37 1.97 0.53 NA 0.35 NA 0.27 0.30 44 Packing plants, fruit and vegetable packing plants, meat packing plants.1 1000 Sq.ft GFA 3.82 0.78 0.75 1.49 0.62 0.64 1.59 0.95 2.18 1.35 0.03 0.88 0.02 1.49 0.62 45 Canneries, fruit and vegetable, bottlers and brewers distilleries, wineries.1 1000 Sq.ft GFA 3.82 0.78 0.75 1.49 0.62 0.64 1.59 0.95 2.18 1.35 0.03 0.88 0.02 1.49 0.62 46 Other food processing, candy factories, bakeries, potato chip factories.1 1000 Sq.ft GFA 3.82 0.78 0.75 1.49 0.62 0.64 1.59 0.95 2.18 1.35 0.03 0.88 0.02 1.49 0.62 47 Mineral processing, phosphate processing, cement plants, refineries, clay plants, rock and gravel plants. 111000 Sq.ft GFA 3.82 0.78 0.75 1.49 0.62 0.64 1.59 0.95 2.18 1.35 0.03 0.88 0.02 1.49 0.62 48 Warehousing, distribution terminals, trucking terminals, van & storage warehousing 1000 Sq.ft GFA 3.56 0.42 0.45 1.23 0.78 0.60 0.86 0.88 1.31 1.26 0.85 0.82 0.55 1.23 0.78 11 Manufacturing facilities are areas where the primary activity is the conversion of raw materials or parts into finished products.

PAGE 70

61 Table 16 Continued DOR code PROPERTY TYPE TRIP RATE FROM ITE TRIP GENERATION MANUAL TRIP RATE BASED ON ITE TRIP MANUAL AND NHTS 2001 DATABASE FOR TBEST TIME PERIODS Unit (Independe nt Variable) Week day Week day AM Peak Hour Week day PM Peak Hour Satu r day Sun day Weekday AM Peak Period Weekday PM Peak Period Week day Off Peak Period Week day Night Period Satur day Sun day Agricultural Using temporal distributio n of trips in NHTS Using Peak Factor s Using temporal distributi on of trips in NHTS Using Peak Factor s Using temporal distributi on of trips in NHTS Using Peak Factor s Using temporal distributi on of trips in NHTS Using Peak Factor s 50 Improved agricultural 51 Cropland soil capability Class I 1000 Sq.ft GFA 6.97 1.01 1.08 1.32 0.68 1.17 2.06 1.73 3.13 2.47 1.07 1.61 0.70 1.32 0.68 52 Cropland soil capability Class II 53 Cropland soil capability Class III 5458 Timberland site index 50 and above We assume zero trip rates here. If any, the trip rates for these land uses can be expected to be rather small. n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a 59 Timberland not classified by site index to Pines n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a 6065 Grazing land soil capability Class I to Class VI n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a 66 Orchard Groves, Citrus, etc. 1000 Sq.ft GFA 6.97 1.01 1.08 1.32 0.68 1.17 2.06 1.73 3.13 2.47 1.07 1.61 0.70 1.32 0.68 67 Poultry, bees, tropical fish, rabbits, etc. 1000 Sq.ft GFA 6.97 1.01 1.08 1.32 0.68 1.17 2.06 1.73 3.13 2.47 1.07 1.61 0.70 1.32 0.68 68 Dairies, feed lots 1000 Sq.ft GFA 6.97 1.01 1.08 1.32 0.68 1.17 2.06 1.73 3.13 2.47 1.07 1.61 0.70 1.32 0.68 69 Ornamentals, miscellaneous a gricultural 1000 Sq.ft GFA 6.97 1.01 1.08 1.32 0.68 1.17 2.06 1.73 3.13 2.47 1.07 1.61 0.70 1.32 0.68

PAGE 71

62 Table 16 C ontinued DOR code PROPERTY TYPE TRIP RATE FROM ITE TRIP GENERATION MANUAL TRIP RATE BASED ON ITE TRIP MANUAL AND NHTS 2001 DATABASE FOR TBEST TIME PERIODS Unit (Independent Variable) Week day Week day AM Peak Hour Week day PM Peak Hour Satu rday Sunday Weekday AM Peak Period Weekday PM Peak Period Week day Off Peak Period Week day Night Period Satur day Sun day Institutional Using te mporal distributio n of trips in NHTS Using Peak Factor s Using temporal distributi on of trips in NHTS Using Peak Factor s Using temporal distributi on of trips in NHTS Using Peak Factor s Using temporal distributio n of trips in NHTS Using Peak Factor s 70 Va cant 71 Churches 1000 Sq.ft GFA 9.11 0.87 0.94 10.37 36.63 1.53 1.78 2.27 2.73 3.22 2.79 2.10 1.82 10.37 36.63 72 73 Private schools and colleges 1000 Sq.ft GFA NA NA 5.5 NA NA 10.83 NA NA 15.95 NA 22.79 NA 14.85 11.41 12.78 74 Privately owned hospitals 1000 Sq.ft GFA 16.5 1.25 1.46 10.18 8.91 2.77 2.55 4.09 4.23 5.84 5.88 3.81 3.83 10.18 8.91 74 Homes for the aged Dwelling Units 3.71 0.29 0.34 2.77 2.33 0.62 0.59 0.92 0.99 1.31 1.29 0.86 0.84 2.77 2.33 75 Orphanages, othe r non profit or charitable services 121000 Sq.ft GFA 28.6 NA NA NA NA 4.81 NA 7.09 NA 10.12 NA 6.61 NA 5.07 5.67 76 Mortuaries, cemeteries, crematoriums 1000 Sq.ft 0.108 0.017 0.037 0.136 0.175 0.02 0.03 0.03 0.11 0.04 0.00 0.02 0.00 0.136 0.175 77 Clubs 1000 Sq.ft GFA 43 3.19 5.84 38.46 36.77 7.23 6.52 10.66 16.94 15.22 11.83 9.93 7.72 38.46 36.77 Lodges, union halls Employee s 46.9 4.3 4.05 29.55 29.1 7.88 8.78 11.63 11.75 16.60 15.96 10.83 10.41 29.55 29.1 78 Sanitariums, convalescent and rest homes 1000 Sq.ft GFA 7.58 0.42 0.72 NA NA 1.27 0.86 1.88 2.09 2.68 2.80 1.75 1.83 1.34 1.50 79 Cultural organizations, facilities 13 1000 Sq.ft GFA 68 NA NA NA NA 11.43 NA 16.86 NA 24.07 NA 15.71 NA 12.05 13.49 12 San Francisco Interim Transportation Impact Analysis Guidelines for Environmental Review, Interim Edition, January 2000, The Planning Dep artment City and County of San Francisco. 13 Study on Jewish Cultural Center 2000

PAGE 72

63 Table 16 Continued DOR code PROPERTY TYPE TRIP RATE FROM ITE TRIP GENERATION MANUAL TRIP RATE BASED ON ITE TRIP MANUAL AND NHTS 2001 DATABASE FOR TBEST TIME PERIODS Unit (Independent Variable) Week day Week day AM Peak Hour Week day PM Peak Hour Satu rday Sunday Weekday AM Peak Period Weekday PM Peak Period Week day Off Peak Period Week day Night Period Satu rday Sunday Government Using temporal distributio n of trips in NHTS Using Peak Factor s Using temporal distributi on of trips in NHTS Using Peak Factor s Using temporal distributi on of trips in NHTS Using Peak Factor s Using temporal distributio n of trips in NHTS Using Peak Factor s 80 Undefined Reserved for future use 81 Military Employee 1.78 0.37 0.37 2.64 1.67 0.30 0.76 0.44 1.07 0.63 0.00 0.41 0.00 2.64 1.67 82 Forest 1000 Sq.ft 2.64 NA NA NA NA 0.44 NA 0.65 NA 0.93 NA 0.61 NA 0.47 0.52 Parks, recreational areas 1000 Sq.ft NA NA NA 4.14 NA 3.93 NA 5.79 NA 8.27 NA 5.39 NA 4.14 4.64 83 Public county schools include all property of Board of Public Instruct ion 1000 Sq.ft GFA 12.89 3.06 2.12 4.37 1.79 2.17 6.25 3.20 6.15 4.56 0.30 2.98 0.19 4.37 1.79 84 Colleges 1000 Sq.ft GFA 27.49 3.09 2.64 11.23 1.21 4.62 6.31 6.82 7.66 9.73 8.18 6.35 5.34 11.23 1.21 85 Hospitals 1000 Sq.ft GFA 16.5 1.25 1.46 10.18 8.91 2.77 2.55 4.09 4.23 5.84 5.88 3.81 3.83 10.18 8.91 86 Counties (other than public schools, colleges, hospitals) including non municipal government. 1000 Sq.ft GFA 27.92 2.21 2.85 NA NA 4.69 4.52 6.92 8.27 9.88 9.16 6.45 5.98 4.95 n/a 87 State, other than military, forests, parks, recreational areas, colleges, hospitals 141000 Sq.ft GFA 27.92 2.21 2.85 NA NA 4.69 4.52 6.92 8.27 9.88 9.16 6.45 5.98 4.95 n/a 88 Federal, other than military, forests, parks, recreational areas, hospitals, colleges 151000 Sq.ft GFA 27.92 2.21 2.85 NA NA 4.69 4.52 6.92 8.27 9.88 9.16 6.45 5.98 4.95 n/a 89 Municipal, other than parks, recreational areas, colleges, hospitals 1000 Sq.ft GFA 27.92 2.21 2.85 NA NA 4.69 4.52 6.92 8.27 9.88 9.16 6.45 5.98 4.95 n/a 91 Utility, gas and e lectricity, telephone and telegraph, locally assessed railroads, water and sewer service, pipelines, canals, radio television communication 1000 Sq.ft GFA NA 0.8 0.76 NA NA NA 1.63 NA 2.20 NA 3.15 NA 2.05 1.58 1.77 14 Motor vehicle department office is an exception with 166.02 trips per 1000 Sq.ft GFA for weekday. 15 Postal office is an exception with 108.19 trips per 1000 Sq.ft GFA for w eekday.

PAGE 73

64 As the ITE trip generation manual does not provide the trip rates for TBEST time periods (shown in Table 3), NHTS 2001 database along with ITE trip generation manual was used to obtain trip rates for TBEST time periods The other part of the table 16 with the heading TRIP RATE BASED ON ITE TRI P MANUAL AND NHTS 2001 DATABASE FOR TBEST TIME PERIODS gives the trip rate of each land use for all the TBEST t ime periods. The methodology used to obtain trip rates for each TBEST time period is as follows: 1) In this part of table, t he columns Weekday AM Peak Period and Weekday PM Peak Period are in turn split into two columns each Using temporal distribution of trips in NHTS and using Peak Factors16Method 1: By m ultiplying the weekday trip rate obtained from ITE manual to the temporal distributions of weekday trips in the NHTS 2001 database shown in the table 17. . The trip rates in these two columns have been computed using two different methods: Table 17 Temporal Distribution of Weekday T rips in 2 001 NHTS D ata Period No. Weekday Time period Percent 1 6am to 8:59 am (AM peak period) 16.8 2 9:00 am to 2:59 pm (Off peak period) 35.4 3 3:00 pm to 5:59 pm (PM peak period) 24.8 4 6:00 pm to 5:59 am (Night period) 23.1 Method 2: By multiplying the peak one hour trip rate from the ITE trip generation manual to the peak factor obtained from the NHTS 2001 database. 16 This peak factor was computed by taking the ratio of the number of trips in the peak period to number of trips in the peak one hour of the am or pm peak period.

PAGE 74

65 From the two methods mentioned above, trip rates obtained using peak factor (Method 2) should be used because this methodology is more specific to the various land uses as compared to using the temporal distribution of NHTS 2001 database. But some of the parcel land uses mentioned below have peak hour periods differ ent from the TBEST time periods Table 18 shows the parcel land uses which have peak hour period differe nt from TBEST time period. Table 18 Parcel Land Uses Having Peak Hour Period Different f rom TBEST Time Period DOR landuse code Property Type Peak Hour Period 12 Departmental Stores AM Peak Period = 11:00 a.m. to 12:00 p.m. PM Pe ak Period = 12:30 p.m. to 5:00 p.m. 20 Airports AM Peak Period = 11:00 a.m. to 12:00 p.m. PM Peak Period = 5:00 p.m. to 7:00 p.m. 23 Bank AM Peak Period = 8:00 a.m. to 12:00 p.m. PM Peak Period = 12:00 p.m. to 6:00 p.m. 71 Church AM Peak Period = 10:00 a.m. to 12:00 p.m. PM Peak Period = 7:00 p.m. to 11:00 p.m. 72 Private Schools PM Peak Period = 2:30 p.m. to 4:00 p.m. 73 & 85 Hospitals AM Peak Period = 8:00 a.m. to 10:00 a.m. PM Peak Period = 1:00 p.m. to 5:00 p.m. 77 Lodges AM Peak Period = 11:00 a. m. to 12:00 p.m. PM Peak Period = 3:00 p.m. to 4:00 p.m. 83 Public County Schools PM Peak Period = 2:00 p.m. to 4:00 p.m. For the above mentioned land uses, it is better to use the trip rates obtained from the temporal distribution of trips in NHTS 2001 database. 2) The columns Weekday Off Peak Period and Weekday Night Period in the table 16 are also split into following two columns Using temporal distribution of trips in NHTS and using Peak Factors. The trip rates in these two columns have been c omputed using two different methods:

PAGE 75

66 Method 1: By multiplying the weekday trip rate obtained from ITE manual to the above mentioned temporal distributions of weekday trips in the NHTS 2001 database (shown in t able 17) Method 2: By subtracting the sum of weekday AM peak and PM peak trip rates obtained using peak factors from the weekday trip rates and then multiplying this difference with the percentage distribution between weekday off peak period and weekday night period obtained from the NHTS 2001 databa se. 3) For some land uses such as service stations (26)17Table 19 Temporal Distribution of T rips in 2001 NHTS D ata night clubs (33), skating rinks, bowling alleys (34), race tracks (37), heavy industries (42), Private schools (72), trip rates for Saturday and Sunday are not available in ITE trip generation manual. The trip rates for Saturday and Sunday (highlighted as bold figures in table 16) are obtained using the distributions of trips in NHTS 2001 database s hown in Table 19. Period No. Time Period % Distribution 1 AM peak period (6am to 8:59 am ) 12.22 2 Off peak period ( 9:00 am to 2:59 pm ) 25.72 3 PM peak period (3:00 pm to 5:59 pm ) 18.00 4 Night period ( 6:00 pm to 5:59 am) 16.76 5 Saturday ( 12 midnight 11:59 PM ) 12.88 6 Sunday ( 12 midnight 11:59 PM ) 14.42 100.00 For some land uses the weekday trip rate is also not available in the ITE trip generation manual. For such land uses, the distribution table above can be used to obtain the trip rates of the weekday time periods. 17 Numbers in parentheses show the parcel level land use code

PAGE 76

67 4) For land uses such as theatres (32) which generally open only after 9 am, trip rate for weekday morning peak period is marked as n/a not applicable18 5) Since the trip rates for orphanages, other charitable servi ces (75), and for cultural organizations (79) are not available in the ITE trip generation manual, we used the following source: San Francisco Interim Transportation Impact Analysis Guidelines for Environmental Review, Interim Edition, January 2000, The Planning Department City and County of San Francisco. Similarly, for land uses such as nightclubs and bars (33) which generally operate in the even ing hours, the trip rates for weekday morning peak and weekday off peak periods are marked as n/a. These trip rates can be used to capture the activity l evels at each land use. But these trip rates should be used with caution as they are vehicle trip rates and do not completely represent transit trip making. 18 Please note that n/a (not applicable) is equivalent to a zero trip rate and this is different from NA (not available)

PAGE 77

68 5.3 Special Generator Enhancement This chapter is focused on improving the predictive capability of the TBEST model by exploring a better way to treat special generators rather than by just using it as a dummy variable in the model. Special generators are defined as land uses that do not generate or attract trips at the same rate as other land uses. To explore different ways of treating special generators, it is very important to understand how various regional travel demand models and transit analysis studies deal with special generators. Section 5.3.1 discusses how previous studies deal with special generators 5.3.1 Literature R eview To explore a better way to treat special generators various regional travel demand models and trans it analysis studies dealing with special generators were reviewed. A considerable exploration of various regional travel demand models reveal the following ways in which special generators are dealt with in the literature: 1) Separate production and attractio n models are developed using generation rates (Pickett, 2001; Wilbur Smith associates, 2008; Dallas Fort worth Regional Travel Model De scription, 2006; and Kikuchi et al ., 2004) specific to each generator. These rates are dependent on the location, activit y level of the generator and are either borrowed from other areas or developed from the survey data (traffic counts and characteristics of special generator s, etc.) on number of trips attracted. These separate models are mostly developed using linear regression analysis. 2) Special g enerators are assigned unique trip rates obtained from the ITE Trip Generation Manual (Lima & Associates, 2006; Pearson et al ., 2009) or other trip generation manuals like San Diego Municipal Code, 2003 to capture trip

PAGE 78

69 attraction. Trip attractions due to special generators are estimated using this trip rate per trip generation variable19Also, A review of various transit analysis studies show the following way of treating special generators : 1) Separate models are developed using trip rates obtained from the survey data [Onboard surveys ( Parsons Brinckerhoff, 2000) and site survey s or interviews (Kurth, 1997; The Duffey Company, 2000; Usvyat, 2009)]. Table 20 presents special generator categories with their specific ITE Trip Generation Manual recommended trip rates, relevant studies by different investigators and the corresponding variables evaluated with results. The spec ial generators mentioned in table 20 were selected as they are more likely to attract trans it trips. The first part of table 20 with the heading Trip Rate f rom ITE Trip Generation Manual gives the trip rate for each special generator category using the ITE trip generation manual, 8th Edition. The trip rates were obtained by matching each special generator category wit h the closest available ITE land use category. The trip rates are available for various independent variables on a weekday, Saturday an d Sunday. The other part of table 20 with the heading Variables Used to Explain Special Generator Trip Generation in the Literature gives the description of each study on how it deals with a specific special generator and includes the list of variables used in that study to explain special generator trip attraction. Summary of each study and dataset is given in the Appendix A. 19 Trip generation variable can be defined as Independent variable which best explains the trip attraction of that special generator

PAGE 79

70 Table 20 Tabulation of Special Generators W ith ITE Trip Rates, Relevant S tudies and Corresponding Variables U sed Sr. No Special Generators Trip Rate from ITE Trip Generation Manual Variables Used to Explain Special Generator Trip Generation in the Literature Unit [Independent Variable(X)] On a Average Trip rate / unit Fitted Curve Equation List of Variables Used Study Description 1 Commercial Airports Employees Weekday 13.4 -Number of Boardings (Enplanements). Hojong et al. (2008) developed a trip generation model to estimate number of person trip attracted by using number of enplanements as an independent variable in the regression analysis. Trip attraction is obtained for 66 international airports in U.S by using data from Bureau of Transportation Statistics T100 international segment database. Saturday 12.2 -Sunday 14.7 -Average Flights per Day Weekday 104.73 -Number of Deplaning Passengers. Number of Boardings In the Laredo Travel Demand Model prepared by Wilbur Smith Associates (2008), trip attraction for international airport is estimated based on the number of deplaning passengers and number of boardings. Trip attraction model was developed using linear regress ion analysis. Saturday 98.46 -Sunday 119.61 -No. of Employees In the Lincoln MPO Travel Demand Model, ITE Trip Generation Manual, 7th Edition is used to calculate the trip attraction. Trip attraction for airports is obtained based on the number of employees in the airport. Commercial Flights per Day Weekday 122.21 -Saturday 113.04 -Number of Boardings 2007 Passenger Boarding and All Cargo Data maintained by Federal Aviation Administration (FA A) can be used for the trip generation model. This dataset contains total number of boardings for the Commercial Service Airports (at least 2500 passenger boardings/year). This data only gives annual boarding at commercial service airports. Sunday 137. 71 -2 Water port / Marine Terminal Number of Berths Weekday 171.52 298.56(X) 417.4 Acreage of the port. City of San Diego has developed its own Trip Generation Manual. Trip rates for each land use were obtained by conducting detailed local surveys ( vehicle trips) at various sites of each land use type. Vehicle trip rate for Marinas is 20 trips /acre. Acres Weekday 11.93 18.01(X) 287.06 3 Major regional amusement parks Employees Weekday 8.33 -Acreage of the Park. No. of Visitors / day. In th e Laredo Travel Demand Model prepared by Wilbur Smith Associates (2008), trip attraction for regional parks is estimated based on the number of visitors/day and acreage of the park. Linear regression analysis was performed using data from the traffic count s. Saturday 22.08 -Sunday 20.96 -Acres Weekday 75.76 -Total attendance per day. Kurth et al. (1997) developed a four step model to estimate the annual transit trips attracted by amusement parks. Trip generation, trip distribution, mode choice and transit assignment models were used based on the data (attendance per day) obtained from the local surveys. Saturday 180.2 -Sunday 171.02 -

PAGE 80

71 Table 20 Continued Sr. No Special Generators Trip Rate from ITE Trip Generation Manual Variables Used to Explain Special Generator Trip Generation in the Literature Unit [Independent Variable(X)] On a Average Trip rate / unit Fitted Curve Equation List of Variables Used Study Description 4 Major sports facilities Stadia, Arena etc Employees Weekday 10 -Capacity of the Facility. In the Laredo Travel Demand Model prepared by Wilbur Smith Associates (2008), trip attraction for regional sports facilities is estimated based on the capacity of the facility. Trip attra ction model was developed using linear regression analysis. Acres Weekday 33.33 -Total Attendance per event Kurth et al. (1997) developed a four step model to estimate the annual transit trips attracted by stadiums. Trip generation, trip distributi on, mode choice and transit assignment models were used based on the data (attendance per event) obtained from the local surveys. 5 Recreational community center Members Saturday 0.07 -Area of the facility (1000 Sq.ft). City of San Diego has developed its own Trip Generation Manual. Trip rates for each land use were obtained by conducting detailed local surveys (vehicle trips) at various sites of each land use type. Vehicle trip rate for Recreational Building is 45 trips/1,000 sq. ft. Sunday 0.15 -Employees Weekday 27.25 -Saturday 18.34 -Sunday 12.03 -1000 Sq.ft Gross Floor Area Weekday 22.88 -Saturday 9.1 -Sunday 13.6 -6 High school Students Weekday 1.71 0.81 Ln(X) + 1.86 Number of students. Num ber of staff. In the Laredo Travel Demand Model prepared by Wilbur Smith Associates (2008), trip attraction for high school is estimated based on the number of students and number of staff. Linear regression analysis was performed using data from the survey of high schools. Saturday 0.61 -1000 Sq.ft Gross Floor Area Weekday 12.89 -Saturday 4.37 -Number of students enrolled. School Enrollment data is collected annually in the October Current Population Survey (CPS) and can be used for the trip attraction model. http://www.census.gov/population/www/socdemo/school.html Employees Weekday 19.74 1.13 Ln(X) + 2.31 Saturday 6.57 -7 College / University Students Weekday 2.38 2.23(X) + 440 Number of students. Number of staff. In the Laredo Travel Demand Model prepared by Wilbur Smith Associates (2008), trip attraction for college/university is estimated based on the number of stu dents and number of staff. Linear regression analysis was performed using data from the survey of colleges/universities. Saturday 1.3 -Number of Employees. In the Dallas Fort Worth Regional Travel Model, trips attracted by college/university is comp uted by applying the trip attraction rates to the employment and adding extra increment trips associated with college/university The number of incremental trips for college/university is obtained by taking the difference of cross classification model ge nerated trip rates and trip rates obtained from regional travel survey. Employees Weekday 9.13 0.74(X) + 3.92 Number of Employees. In the Lincoln MPO Travel Demand Model, ITE Trip Generation Manual, 7 th Edition is used to calculate the tr ip attraction. Trip attraction for university main campus is obtained based on the number of employees. Saturday 3.12 -

PAGE 81

72 Table 20 Continued Sr. No Special Generators Trip Rate from ITE Trip Generation Manual Variables Used to Ex plain Special Generator Trip Generation in the Literature Unit [Independent Variable(X)] On a Average Trip rate / unit Fitted Curve Equation List of Variables Used Study Description 8 Hospital Beds Weekday 11.81 7.42(X) + 1733.31 Number of Employees. N umber of Beds. In the Laredo Travel Demand Model prepared by Wilbur Smith Associates (2008), trip attraction for hospital/medical center is estimated based on the number of employees and number of beds. Linear regression analysis was performed using data f rom the survey of hospitals/medical centers. Saturday 8.14 0.58 Ln(X) + 4.65 Sunday 7.19 0.61 Ln(X) + 4.38 Employees Weekday 5.2 4.4(X) + 711.46 Number of Employees. In the Lincoln MPO Travel Demand Model, ITE Trip Generati on Manual, 7th Edition is used to calculate the trip attraction. Trip attraction for medical centers is obtained based on the number of employees. Saturday 3.78 2.95(X) + 691.43 Sunday 3.34 2.56(X) + 663.23 1000 Sq.ft Gross Floor Area+ Week day 16.5 10.13(X) + 2191.79 Number of Beds Americ an Hospital Association (AHA) collects data on number of beds for more than 6500 AHA registered hospitals throughout the United States. This dataset is available at stat e and regional geographic level and ca n be used. Saturday 10.18 0.43 Ln(X) + 5.79 Number of Employees. In the Dallas Fort Worth Regional Travel Model, trips attracted by hospital is computed by applying the trip attraction rates to the employment and adding extra increment trips associated with hospital. The number of incremental trips for hospital is obtained by taking the difference of cross classification model generated trip rates and trip rates obtained from regional travel survey. Sunday 8.91 3.53(X) + 1937.21 9 Sho pping Center (SC) 1000 Sq.ft Gross Leasable Area Weekday 42.94 0.65 Ln(X) + 5.83 Number of Parking Spaces. Number of Stores. Type of Stores. Floor area of SC. Kikuchi et al. (2 004) developed macroscopic and microscopic model to estimate the attraction rate of SC. In macroscopic approach, relationship between the listed variables & attraction rate was obtained using regression analysis and in the microscopic approach, attraction rate of SC was taken as weighted sum of attraction rates of individual stores. The data used in both the approaches was obtained by the surveys conducted at various shopping centers. Saturday 49.97 0.63 Ln(X) + 6.23 Number of Employees. In the Lincoln MPO Travel Demand Model, ITE Trip Generation Manual, 7th Edition is used to calculate the trip attraction. Trip attraction for malls is obtained based on the number of employees. Sunday 25.24 15.63(X) + 4214.46 Numb er of Employees. In the Laredo Travel Demand Model prepared by Wilbur Smith Associates (2008), trip attraction for shopping center is estimated based on the number of employees. Linear regression analysis was performed using data from the traffic counts done at various shopping centers.

PAGE 82

73 Table 20 Continued Sr. No Special Generators Trip Rate from ITE Trip Generation Manual Variables Used to Explain Special Generator Trip Generation in the Literature Unit [Independent Variable(X)] On a Average Trip rate / unit Fitted Curve Equation List of Variables Used Study Description 10 Free Standing Superstore 1000 Sq.ft Gross Floor Area Weekday 53.13 1.35 Ln(X) + 2.11 1000 Sq.ft Gross Floor Area. The Texas Transportation Institute (TTI) conducted a nationwide discount superstore trip generation study by collecting site and trip generation data for typical season, peak hours using manual or video counting. Trip rate (per 1000 Sq.ft GFA) is calculated for various time periods by plotting the best fi tted curve for the data collected. Saturday 64.07 1.45 Ln(X) + 1.74 Sunday 56.12 1.74 Ln(X) + 0.09 Area of the Facility (1000 Sq.ft). City of San Diego has developed its own Trip Generation Manual. Trip rates for each land use were obtained by co nducting detailed local surveys (vehicle trips) at various sites of each land use type. Vehicle trip rate for superstore is 40 trips /1000 Sq.ft. 11 Park and Ride Lot with Bus service Parking Spaces Weekday 4.5 4.04(X) + 117.33 Service area population. Ra tio of auto costs to transit costs. Distance from park and ride facility to major employment centers. Number of express buses during the morning (AM) peak. Best (not average) time between the park and ride facility and the CBD. Presence of nearby park and ride facilities. Availability of midday service. Robert Pillar (1997) developed the planning and design manual for park and ride facilities. The methodology used for estimating the park and ride demand is all about defining a service area (catchment area) for the park and ride facilities and then developing equations based on the lot attributes using multivariate regression analysis. The listed variables are u sed in the model based on availability of data from past surveys and existing database and the potential ease of developing similar data for the evaluation of future lots. Occupied Spaces Weekday 9.62 -Park and ride is also modeled using the traditional modeling technique, which is identifying the attraction and production zones and then determining the proportion of trip interchange for park and carpool and bus park and ride users. Acres Weekday 372.32 -12 Intermodal Terminals N umber of Employees. Building /Floor area. Trip rate was calculated using linear regression based on the variables number of employees and floor area. Data for the analysis is obtained by conducting surveys. http://praytorianguard.com/blog1/?p=456

PAGE 83

74 A project on Understanding Tr ansit: Basic Course M aterial on Public Transportation executed by the Center for Urban Transportation Studies, University of WisconsinMilwaukee suggest site interviews at speci al generators as an important source for transit analysis. Such interviews will be helpful in knowing the location, features (size in terms of visitors, employment, area etc.) and ridership developed by special generators. Also, a paper by Carter (1984) focuses on the importance of special generator information in transit and traffic analysis. The author presents detailed recommendations on questionnaire content and procedures. According to the special generators report by LSA Associates, Inc. (2008), revi ew and application of special generator developments is one of the important aspects in developing the t ravel demand model. Firstly, potential special generators are identified and categorized into broad categories like event centers, airports, stadiums, r esorts, theme parks, religious, tourist destinations etc. based on the type of development, establishment, or area. Secondly, special generators are evaluated based on the database, which includes the following information on special generators: 1) Descriptio n and location of activity 2) Duration and recurrence of activity (single event throughout the day, random vs. scheduled etc.) 3) Ca tegory of the special generator 4) Trip distribution (local vs. regional, et c.) and mode choice information 5) The seas onal variabilit y of trip making 6) Independent trip generation (activity) variables and their availability

PAGE 84

75 The most important data to develop separate models for each special generator are the independent trip generation (activity) variables as they define the trip attraction capability of a particular special generator. D ifferent data sources we re explored to obtain information on trip generation (explanatory) variable available for various generators. The datasets useful for defining the attraction capability of the follo wing special generators are: 1) For special generators like schools, colleges and universities, the number of enrollments best describes the trip generation. This data can be obtained from the following datasets: 2000 U.S Census data, Current Population Survey (CPS) and American Community Survey (ACS). 2) For airports, annual passenger enplanement for commercial service airports can be obtained from the Federal Aviation Administration (FAA) Passenger Boarding (Enplanement) and All Cargo Data. This dataset is easily available for the current year and the next fiscal year. 3) For hospitals, the American Hospital Association (AHA) annual survey database can be used. This dataset provides the number of beds for more than 6500 AHA registered hospitals throughout the United States. This dataset is available at the state and the regional geographic level. This dataset is not available online and can be ordered in the form of a CD and a book. Based on the literature review and the availability of information on the trip generation variable for each special generator category best, next best and other explanatory variables were stated for each special generator. Table 21 shows the options for explanatory variables of each special generator

PAGE 85

76 Table 21 List o f Various Special Generators and the Optio ns (Best, Next Best and Other) f or the Explanatory Variables of Each Generator To enhance the tri p attraction capability of the TBEST model, s pecial generators can be represented by the explanatory variables which best describe the activity levels at that generator. Special generators Options for Explanatory Variables Best Next Best Other Regional Airports Boardings (Enplanements) Plane Arrivals/depart ures Employees Water port /Marine Terminals Area (Acres) Employees Major Regional Amusement Parks Visitors/day Parking spaces Employees or acres Major Sports Facilities Total Attendance/event Capacity (seats) Parking spaces Recreational Community Center Visitors/day Parking spaces Area (1000 Sq.ft) High School Students Enrolled Employees College/University Students Enrolled Employees Hospitals Number of Beds Employees Shopping Centers (SCs) Employees Parking Spaces Floor Area o f SC Free Standing Superstore Area (1000 Sq.ft Gross Floor Area) Park and Ride Lots with Transit Service Number of Parking Spaces Service Area Population Ratio of Auto Costs to Transit Costs. Distance from Park and Ride Facility to Major Employment Centers Number of Express Buses during the Morning (AM) Peak Best (not average) Time Between the Park and Ride Facility and the CBD Presence of Nearby Park and Ride Facilities Availability of Midday Service Intermodal Terminals Employees Building /Floor Area

PAGE 86

77 5.4 A nal ysis E xploratory analysis has been performed to achieve the above mentioned possibilities of enhancing trip attraction of the TBEST model. It aims at capturing trip s generated by employment, trip s generated using parcel land use based trip rates, total weekday boarding, presence of special generator s and area of the special gen erator s and nonresidential land uses within the stop buffers for all the stops of the four routes shown in table 3. The analysis will give insights on the various strategies discussed above in this chapter and also help in defining suggestions for the improvement of the TBEST model. Table 22 Potential Special Generators in the Parcel Data DOR Land Use Code PROPERTY TYPE 013 Department Stores 014 Supermarkets 015 Regional Shopping Centers 016 Community Shopping Centers 020 A irports (private or commercial) 20 028 bus terminals, marine terminals, piers, marinas. Parking lots (commercial or patron) mobile home parks 072 Private schools and colleges 073 Privately owned hospitals 082 Forest, parks, recreational areas 083 Publi c county schools 084 Colleges 085 Hospitals Based on the special generators identified in the previous chapter parcel land uses shown in t able 22 were believed to generate or attract trips at a higher rate as compared to other land uses21 20 There is an ambiguity in the Duval Countys parcel data regarding how airports are defined as the Jacksonville airport is coded in counties (86) land use and not in land use code 20. The above m entioned l and uses were selected from Duval county s parcel data and a separate point layer file was created. Trips generated by each land use w ere calculated by multiplying the trip rates obtained using the ITE trip generation 21 To accurately define potential special generators, there is need to go through each land us e either using area or some other variable like trip rates

PAGE 87

78 manual (discussed in section 5.2) by the total useable area given in the parcel data. Quarter mile buffers for each stop were generated and spatially joined to the special generator layer file and nonresidential parcels to determine the number of special generators in each stop buff er and total trips generated within each stop buffer respectively Also, s top buffers were spatially joined to the InfoUSA employment data layer file to determine the employment by type within each stop buffer. The trips generated by employment were calcul ated using the trip attraction per employee shown in table 15. The stop buffers having zero special generators were studied separately as compared to stops with one or more special generators in the catchment area.

PAGE 88

79 5.5 Results This section describes the r esults of this analysis based on the trips generated by employment, trip s generated using parcel land use based trip rates total weekday boarding, presence of special generator s in the stop buffer and the useable area of the special generators and non re sidential parcels with in the stop buffer Figure 12 shows graphs capturing the difference between trips generated by employment and tr ips generated using parcel land use based trip rates for both stop buffer s with and without special generators. The graphs show that stops with special generators have large differences in the trips generated when compared to stops without special generators in their catchment area. The absolute percent difference between the trips obtained using employment and parcel land us e based trip rates was computed for each stop. It was observed that the absolute percent difference is higher for stops with special generators when compared to stops without them. The KS (Kolmogorove Smirnov) test was performed to test if the trips gener ated using employment and parcel land use based trip rates are st atistically different. The KS test showed a lower p value of 0.003 for stops with special generators resulting in the rejection of the null hypothesis of no statistical difference between the trips ; w hereas, a p value of 0.199 for stops without special generators resulted in the rejection of the alternative hypothesis of statistical difference between the trips. This shows that trips generated by employment and trips generated using parcel la nd use based trip rates are similar for stops without special generator s Detailed results for each stop of all four routes are presented in Appendix B.

PAGE 89

80 Figure 12 Graphs Showing Differences B etween T rip G e nerated by Employment and Trip G enerated U sing Parcel Land U se Based Trip Rates for Route P7 Stops With and Without Special Generator A linear regression analysis was performed with total weekday boarding as the dependent variable and total employment (currently used in the TBEST model) trips generated by both employment and parcel land use based trip rates as the independent variables for both approaches (stops with special generator s and stops without special Trips Generated within 1/4 th mile Buffer of Stops with Special Generator Trips Generated within 1/ 4 th mile Buffer of Stops without Special Generator

PAGE 90

81 generator s ) The result of this analysis given in t able 23 show s that for stops with special generators the coefficient and t stats of trips generated using parcel land use based trip rates (model 3) are higher than that of total employment (model 1) and trip generated by employment (model 2) Also, the coefficient and t stats of model 4 w here both the trips generated by employment and trip s generated using trip rates we re used in the same model show similar results. This indicates that trips generated using trip rates better explain total boarding at stops with special generator s as compared to the trips generated by employment. On the other hand, results of the stops wi thout special generator s show that total employment, trips generated by employment and parcel land use based trip rates are not statistically significant in explaining the total boarding. This clearly conveys that none of these independent variables are ab le to explain total boarding for stops without special generator s but for stops with special generator s trips generated using parcel land use based trip rates should be used22 22 The R2 values in this analysis are very low and therefore it is difficult to make strong conclusions or implications from this analysis

PAGE 91

82 Table 23 Results of the Linear Regression Analysis B etween Total Boarding, Total Employment, Trips G en erated by Employment and Trips Generated U sing Trip Rates for Stops With and Without Special Generator Independent Variable Parameter Estimates (t stats) Stops with Special Generator Stops without Special Generator Model 1 Model 2 Model 3 Model 4 Model 1 Model 2 Model 3 Model 4 Constant 4.175 (3.18) 4.156 ( 3.18 ) 3.415 ( 2.56 ) 3.449 (2.58) 3.674 (4.44) 3.742 ( 4.54 ) 3.963 ( 4.78 ) 3.841 (4.58) Total Employment 0.163 (2.27) 0.064 (0.84) Trips ge nerated by Employment 0.167 ( 2.33) 0.058 ( 0.46) 0.050 ( 0.66) 0.107 (0.95) Trips generated using Parcel Land use based Trip rates 0.225 (3.20 ) 0.274 (2.19) 0.003 (0.04 ) 0.077 ( 0.68) R2 0.027 0. 028 0. 051 0.052 0.004 0. 003 0. 000 0.005 N 192 177 The KS test was performed again between trips generated by employment and trips generated using trip rates based on the size (area in sq.ft.) of the special generators within the stop buffer. This helps in defining the threshold value for the siz e of the special generator within the stop buffer The stops with special generators above this threshold value will results in statistical difference s between the trips generated by employment and trips generated using parcel land use based trip rates, wh ile the special generators below this threshold value will not present any statistical differences. The results of this test show a threshold value of 7000 sq.ft. for the size of the special generator. To test variation of size (area in square foot) of the nonresidential land uses within the stop buffer and total boarding between stops with and without special generators, Levene's test for equality of variances and t test for equality of means were performed. Levene's test is used to evaluate the equality of variances in different samples.

PAGE 92

83 If the resulting p value of Levene's test is less than some critical value ( 0.05 in this case ), the obtained differences in sample variances are unlikely to have occurred based on random sampling Ta ble 24 shows the resul ts of th e Leve ne's test and t test for total boarding and area of the land uses in the stop buffer between stops with and without special g enerator s The results of th e Levenes test show a high p value (0.384) which resul ts in the rejection of the altern ative hypothesis of unequal variances of total boarding for stops with and without special generator s The area of the land uses within each stop buffer shows very low pvalue (0.00) Thus, the null hypothesis of equal variances is rejected and it is concl uded that there is a difference between the variances in the area. The t test assesses whether the means o f two groups are statistically different from each other. The results of the t test show that the means of the total boarding are equal for stops with and without special generators i.e. high pvalue (0.259) Whereas, the means of the area of land uses in the stop buffer are statistically different for both stops with and without special generator s .i.e. low pvalue (0.000 ) Table 24 Results of the Levene's Test and T Test for Total Boarding and Area of the Land U ses in the Stop Buffer Between S tops With and Without Special Generator Total Boarding Area of the Land Uses in the stop buffer Stops without Special Generator Stops w ith Special Generator Stops without Special Generator Stops with Special Generator N 177 192 177 192 Mean 3.975 5.575 37329.29 235956.75 Std. Deviation 9.916 16.249 59332.366 4.030 Levene's Test for Equality of Variances (p value) 0.384 0. 0 00 t test f or Equality of Means (p value) 0.259 0.000 Also, linear regression analysis was performed with total boarding at each stop as the dependent variable and total employment, special generator dummy and special

PAGE 93

84 generator area (interaction variable between spe cial generator dummy and area of the special generator) as inde pendent variables. Since at present employment and special generator dummy variable are the only variables used in the TBEST model to measure transit trip attractiveness t his analysis was perf ormed to test if the interaction variable is more effective in capturing the impact of special generators on trip attractio n as compared to employment and special generator dummy variable. The results in table 25 show a higher R2 value for the employment a nd special generator area variable as compared to the employment and special generator dummy variable23Table 25 Results of the Linear Regression Analysis B etween Total Boa rding, Special Generator Dummy V ariable and Special Generator Area This indicates that the interaction variable explains total boarding more accurately as compared to the special generator dummy variable. Independent Variable Parameter Estimates (t stats) Constant 3.508 ( 3.42) 3.158 (3.87) Total Employment 0.146 (2.70) 0.121 (2.30) Special generator dummy (1 if special generator is present within stop buffer, 0 otherwise) 0.027 ( 0.51) Special generator area in sq.ft 0.131 (2.49) R2 0. 023 0. 0 40 N 369 23 The R2 values in this analysis a re very low and therefore it is difficult to make strong conclusions or implications from this analysis.

PAGE 94

85 C HAPTER 6 CONCLUS IONS 6.1 Conclusions This study uses exploratory analysis to obtain insights on the strategies used to enhance the precision of input data and the trip attraction capabilities of the TBEST model. The results of the analysis to identify the differences in activity levels (population and e mployment) within transit stop buffers d ue to the change from aggregate census data to disaggregate parcel data showed statistical differences in disaggregate and aggregate level population and employment. The results of the linear regression analysis indi cate that disaggregate level population and employment explain total boarding at each stop more accurately when compared to the aggregate level population and employment. Based on these results, it can be said that the use of parcel level data can potentia lly improve the accuracy in capturing the activity levels within the catchment area of each stop. Other findings that surfaced from this analysis are: 1) t he differences between the aggregate and disaggregate population and employment decreases with an inc rease in the size of the catchment area of each stop .2) c hange in input data from disaggregate level to aggregate level affects total employment more when compared to population.3) a ggregate level analysis leads to the underestimation of total employment due to the assumption of uniform spatial distribution of employment within each block group. For the enhancement of trip attraction, possibilities such as employment trip attraction, parcel land use b ased trip attraction and special generator enhancement we re explored The results of this analysis show that the absolute percent difference between

PAGE 95

86 trips obtained by employment and trips obtained using parcel land use based trip rates is higher for stops with special generators when compared to stops without them. Also, the trips obtained by employment and trips obtained using trip rates are statistically different for stops with special generator s Further, the results of the linear regression analysis show that trips generated using parcel land use based trip rates explain total boarding at stops with special generator s better than total employment and trips generated by employment. The threshold value of 7000 sq.ft for size of the special generator was identified. The above mentioned findings suggest the us e of parcel land use based trip rates to capture trip attraction, specifically for stops with special generators of area greater than 7000 sq.ft. An e xtensive literature review shows that special generators are usually handled separately using the informatio n on location, category, and independent trip generation variable which best describes the attraction at that special generator. Based on the availability of data on the trip generation variable and literature review, the variables which best describe the activity levels at each special generator were identified (shown in table 21). The information on the explanatory variables can be obtained either using datasets mentioned in section 5.3.1 or by conducting site interviews or surveys at that special generat or. Also, defin ing special generators in terms of trip attraction rather than using a dummy variable in the model will help in improving the predictive capability of the TBEST model. Therefore, linear regression analysis was performed to explore the intera ction variable between the special generator dummy and area of the special generator compared to the employment and special generator dummy variable. The results of the analysis showed that the interaction variable can better explain special

PAGE 96

87 generator attr action when compared to the employment and special generator dummy variable. 6.2 Suggestions for Enhancement of TBEST Model Based on the findings from the analysis and literature review, the following suggestions can be made for the e nhancement of the TBES T model: 1) Strategies for disaggregating the block group level demographics to parcels stated in table 4 should be used. The u se of parcel data with disaggregated demographics relaxes the assumption of uniform spatial distribution of demographic data over block group level of geography resulting in the enhancement of predictive capability of the TBEST model 2) To better capture trip attraction in the TBEST model parcel land use based trip attraction should be considered only for stops with special generators of area greater than 7000 sq.ft.as the results show that trips captured by employment and trip rates are similar for stops without special generator s and stops with special generator s of area less than 7000 sq.ft. 3) Each special generator can be modeled separately using the explanatory variables which best describe the activity levels (shown in table 21) at that generator The information on the explanatory variables can be obtained by using datasets mentioned in section 5.3.1 specific to each type of generat or. Also, site interviews or surveys can be conducted at that special generator and data on 1) l ocation of the special generator 2) duration of the special generator 3) t rip generation variabl es (attendance, employees, area, etc.) and 4) t rip distributio n and modal share should be obtained to account for attraction in the TBEST model.

PAGE 97

88 4) Another way of treating special generators would be to use an interaction variable between special generator dummy and size (square footage, etc.) of the special generator i nstead of simply using special generator dummy variable in the TBEST model Using the interaction variable will definitely be more effective in capturing the impact of special generators on trip attraction.

PAGE 98

89 REFE RENCES Baik, H., A. A. Trani, N. Hinze, H. Swingle, S. Ashiabor, and A. Seshadri. Forecasting Model for Air Taxi, Commercial Airline, and Automobile Demand in the United States. Transportation Research Record: Journal of the Transport ation Research Board, No. 2052, Washington, D.C., 2008, pp. 920. Carter, M Transit and Traffic Analysis, Transportation Research Board Special Report 206, 1985, pp. 152157. Hull, Edwin, Edwin Hull Associates, Application of a Parkand Ride Forecasting Procedure in the Greater Vancouver Transportation Model pres ented in the 13th Annual International EMME/2 Users Group Conference, Houston, Texas, October,1998. Kikuchi S., M. Felsen, S. Mangalpally and A. Gupta. Trip Attraction Rates of Shopping Centers in Northern New Castle County, Delaware, prepared for the Delaware Department of Transportation, July, 2004. Kittelson & Associates, Inc., TCRP Report 88: A Guidebook for Developing a Transit Performance Measurement System, Transportation Research Board, Washington, D.C., 2003. Kurth, David L., B. V. Meter, S. Myung and M. C. Shaefer. Quantifying Special Generator Ridership in Transit Analysis. Proceedings of the Sixth TRB Conference on the Application of Transportation Planning Methods, Dearborn, Michigan, May 1997. Lima & Associates, Lincoln MPO Travel Deman d Model Documentation, prepared for the Lincoln Metropolitan Planning Organization, January, 2006. LSA Associates, Inc., Weekend Travel Demand Model: Technical Memorandum #9 Special Generators for the Weekend Model prepared for the Southern California A ssociation of Governments, June, 2008. http://www.lsa assoc.com/SCAG/

PAGE 99

90 McDonnell, S., S. Ferreira, and F. Convery : Impact of Bus Priority Attributes on Catchment Area Residents in Dublin, Ireland, Journal of Public Transportation, 9(3), 2006, 13762. Murray, A., R. Davis, R Stimson and L. Ferreira. Public transport access. Transportation Research D 3, 1998, 31928. Murray, A. Strategic Analysis of Public Transport Coverage. Socioeconomic Planning Sciences 35, 2001, pp. 175188. Neilson, G. K., and W. K. Fowler. Relation between Transit Ridership and Walking Distances in a Low Density Florida Retirement Area. Highway Research Record No. 403, 1972, pp. 2634. Parsons Brinckerhoff, WACOG Connector Program Trans it Feasibility and Implementation Plan, Task 1 # Transit Demand Model Summary, prepared for the Western Arizona Councils of Governments (WACOG), September, 2000. Pascoe, R. Updating Demographics Using Parcel Data for Transit Corridor Analyses. Presented at National Center for Transit Research (NCTRs) GIS in Transit Conference, Tampa, Florida, 2007. Passenger Boarding Statistics. FAA, U.S. Department of Transportation. http://www.faa.gov/airports_airtraffic/airports/planning_capacity/passenger_allcargo_stat s/ Updated September 26, 2008. Pearson, David F., B. S. Bochner, P. Ellis, M. I.Ojah. Discount Superstore Trip Generation. Institute of Transporta tion Engineers (ITE) Journal, June 2009. Pickett, W. Kirby. Traffic Data and Analysis Manual: Chapter 2 (Section 2) Texas Travel Demand Model Package. http: //onlinemanuals.txdot.gov/txdotmanuals/tda/urban_travel_demand_forecasting.htm Updated September 01, 2001. Pillar, Robert S., Parsons Brinckeroff Quade & Douglas, Inc., A Comprehensive Planning and Design Manual for Parkand Ride Facilities: Chapter 5 Suburban Parkand Ride Demand Estimation Techniques, October,1997.

PAGE 100

91 Reese, M. A GIS Analysis of the Bus Rapid Transit System in Curitiba, Brazil Student project for course PLAN 512 GIS for Planners, University of Virginia, Fall 2007. http://www.arch.virginia.edu/~dlp/Courses/P512F07/CuritibaGIS_presentationopt.pdf_ School Enrollment Data. Current Population Survey (CPS), U.S Census Bureau. http://www.census.gov/population/www/socdemo/school.html Updated October, 2007. Sullivan, S. O and J. Morrall, Walking distances to and from light rail transit sta tions, Transportation Research Record No. 1538, 1996, pp. 1926 T 100 International Segment (All Carriers) Database. Bureau of Transportation Statistics, U.S. Department of Commerce. www.transtats.bts.gov/Fields.asp?Table_ID=261 Updated December, 2008. T BEST (Transit Boardings Estimation and Simulation Tool), Florida Department of Transportation, Tallahassee. 2009. www.tbest.org Trip Generation Manual. San Diego Municipal Code Ma y, 2003. http://www.sandiego.gov/planning/pdf/tripmanual.pdf The Duffey Company and Nelson/Nygaard C onsulting Associates, Transit Impact Fee Analysis: Technical Memorandum #2 Land Use and Trip Generation rates, prepared for the San Francisco Planning Department, September, 2000. Transit Route Location and Analysis, Understanding Transit: Basic course ma terial on Public Transportation prepared by Center for Urban Transportation Studies, University of W isconsinMilwaukee, July 2006. http://www4.uwm.edu/cuts/utp/ Transportation Department of North Central Texas C ouncil of Governments (NCTCOG), Dallas Fort Worth Regional Travel Model Description, June, 2006. Trip Generation Manual. San Diego Municipal Code. http://www.sandiego.gov/planning/pdf/tr ipmanual.pdf May, 2003.

PAGE 101

92 Usvyat, Len, L. Meckel and M. Dicarlantonio. Sketch Model to Forecast Heavy Rail Ridership. Proceedings of 12th TRB National Transportation Planning Applications Conference, Houston, Texas, May 2009. Wilbur Smith Associates, Lar edo Metropolitan Transportation Plan 20052030: Chapter 3 Demographics and Travel Demand Model, prepared for the Laredo Metropolitan Planning Organization, January, 2008. Xuehao C ., S Polzin, R Penyala, N. Siddiqui and I. Ubaka A Framework of Model ing and Forecasting Stop Level Transit Patronage Presented at 86th Transportation Research Board Annual Meeting Washington, D.C., 2007. Zhao, F., L. F. Chow, M. T. Li, I. Ubaka, and A. Gan. Forecasting Transit Walk Accessibility: Regression Model Alter native to the Buffer Method. Transportation Research Record No. 1835, TRB, National Research Council, Washington, D.C., 2003, pp. 3441.

PAGE 102

93 APPENDICE S

PAGE 103

94 Appendix A: Summary of Studies and Datasets Reviewed for Special Generator Enhancement A.1 Special Generat or: School For special generators like schools, colleges and universities, number of enrollments best describes the trip generation. This data can be obtained from the following datasets: 2000 U.S Census data, Current Population Survey (CPS) and American C ommunity Survey (ACS). 1) 2000 U.S Census Data : Data on school enrollment was obtained from answers to long form questionnaire filled by the sample of the population. People were classified as enrolled in school if they reported attending a "regular" public or private school or college at any time between February 1, 2000, and the time of enumeration. The Census 2000 Summary File 3 data are available from t he American Fact finder on the i nternet ( factfinder.census.gov ). This data file gives annual enrollments and is available by sex, age, type of school and type of college. The Census 2000 Summary File 3 (SF 3) sample d ata contains the following tables: P36 Sex by school enrollment by level of school by type of school for the population 3 years and over. P38 Armed forces status by school enrollment by educational attainment by employment status for the population 16 to 19. PCT23 Sex by school enrollment by age for the population 3 years and over. PCT24 Sex by college or graduate school enrollment by age for the population 15 years and over.

PAGE 104

95 Appendix A Continued 2) Current Population Survey (CPS) Data U.S Census Bureau conducts interviews for monthly Current Population Survey (CPS) and school enrollment data of households members 3 years old and over is obtaine d from CPS. This data gives annual enrollments for all the school and colleges in United States. Data is available by sex, age, race, type of school and type of college. The data can be used to study the trip attraction of schools and colleges based on the variable number of students enrollment. The dataset is easily available from the following link : http://www.census.gov/population/www/socdemo/school.html 3) American Community Survey (AC S) The American Community Survey (ACS) is a nationwide survey started in January 2006. ACS is started to replace decennial census long form by providing annual (or multiyear average) estimates of selected social, economic, and housing characteristics of t he population for many geographic areas and subpopulations. ACS gives school enrollment by age, sex, type of school and type of college for the population 3 years and over. 1 year and 3 year estimates of American Community Survey are easily available from the following link: http://factfinder.census.gov/servlet/DatasetMainPageServlet?_program=ACS&_submenu Id=datasets_1&_lang=en&_ts =

PAGE 105

96 App endix A Continued A.2 Special Generator: Airports FAA extracts passenger (enplanement) and cargo data from Air Carrier Activity Information System (ACAIS). This data is available only for Commercial Service Airports Commercial Service Airports are publicl y owned airports that have at least 2,500 passenger boardings each calendar year and receive scheduled passenger service. These airports are further classified into: 1) Primary Commercial Service Airports (that have more than 10000 passenger boardings per ye ar) and, 2) Non primary commercial service airports (that have at least 2,500 and no more than 10,000 passenger boardings each year). Passenger boarding and all cargo data is collected for a full calendar year and determines entitlements for the next full fiscal year (i.e., calendar year 2007 data determines Fiscal Year 2009 entitlement funds) The dataset is easily available from the following link : http://www.faa. gov/airports/planning_capacity/passenger_allcargo_stats/passenger/ A.3 Special Generator: Hospitals Amer ican Hospital Association (AHA) annual survey is an online survey taken by more than 6500 AHA registered hospitals throughout the United States. This database is used for market research and health care industry analysis on hospitals. The database captures information like facilities provided, hospital utilization, beds, admissions etc on each hospital. The number of beds information in this data can be used for the trip generation as no. of beds best describes the trip generation for hospitals. This dataset is available at state and regional geographic level.

PAGE 106

97 Appendix A Continued This dataset is not available online and can be ordered in the form of CD and book. More information can be obtained using the following link : http://www.ahadata.com/ahadata/html/AHASurvey.html A.4 Laredo Travel Demand Model Laredo Travel Demand Model serves as an important tool for developing comprehensive multimodal transportation plan for the Laredo Metropolitan area. In Laredo travel demand model uses 2000 U.S Census Bureau for socio economic data and Texas workforce commission 2003 for employment data. The special generators used in the Laredo travel demand model are Schools, College/University, Airports, Transit Center, Hospitals, Regional Shopping malls, Regional Sports facilities and Regional Parks. Trip generation for each special generator is modeled separately using linear regression analysis. The i ndependent variables used for each special generator are as follows: Table 26 List of Special Generators and Variables Used in Laredo Travel Demand Model Special generator List of variable s Schools, College /University Number of Students Number of Staff Airports Number of Boardings Number of Deplaning Passengers Transit Center Annual Bus System Transfers Hospitals Number of Employees Number of Beds Regional Shopping Malls Number of Emp loyees Regional Entertainment/ Sports Facilities Capacity of the Facility Regional Parks Acreage of the park Number of Visitors

PAGE 107

98 Appendix A Continued The data used for the linear regression analysis is obtained from the traffic counts and surveys conduc ted at the special generators. A.5 Lincoln MPO Travel Demand Model Lincoln travel demand model is used for the city of LincolnLancaster County (Lincoln MPO). In Lincoln travel demand model, special generators are considered as land uses that do not genera te or attract trips at the same rate as other land uses in the same land use category, hence they are assigned a unique trip rate. Nine special generators and the variables used to explain trip rates for these special generators used in Lincoln travel dema nd model are as follows: Table 27 List of Special Generators and Variables Used in Lincoln MPO Travel Demand Model Special generators List of variables Airports Number of Employees Prison Number of Employees Mall Area (1000 Sq.ft ) Medical Center Number of Employees University Main Campus Number of Students Heavy Industrial Area (acres) Low Retail Area (1000 Sq.ft) Low Office Area (1000 Sq.ft) Low Service Area (1000 Sq.ft) Trip attractions for the internal non residential la nd uses are estimated using a trip rate per unit (square feet, students, employees, etc.). These Non Residential trip rates are obtained using ITE Trip Generation, 7th Edition.

PAGE 108

99 Appendix A Continued A.6 Texas Travel Demand Model Package Special generat ors are modeled separately using trip production and trip attraction rates for that generator. Major regional amusement parks, Major sports facilities, Major regional airports, Military bases, Colleges, universities, communities colleges and High schools are considered as special generators in the Texas travel demand model. Special generator model requires more detailed information such as TAZ where it is located, number of hours in operation during a normal weekday, number of work shifts, and number of em ployees per work shift. All the data required for calculating trip attraction rates is obtained by conducting surveys at the special generators. Following variables are used by the linear regression models for each special generator: Table 28 List of Special Generators and Variables Used in Texas Travel Demand Model Special generators List of variables Military Base Number of Employees Schools, Colleges/Universities Number of Students Enrolled Hospitals Number of Beds Major Regi onal Airports Number of Flights/ Day Number of Deplaning Passengers /Day A.7 Dallas Fort W orth Regional Travel Model (DFWRTM) The modeling areas included in DFWRTM is the entire counties of Collin, Dallas, Denton, Rockwall and Tarrant, the western portion of Kaufman County, the northern portion of Ellis and Johnson Counties, and the eastern portion of Parker County. The employment types used in this model are Basic, Retail and Service. In DFWRTM, special generators and the variables used to explain trip rates are as follows:

PAGE 109

100 Appendix A Continued Table 29 List of Special Generators and Variables Used in DFWRTM Special generators List of variables Regional Shopping Malls Number of Employees University/Colleges Number of Employees H ospitals Number of Employees The trips attracted by special generators are computed by applying the trip attraction rates to the employment at respective sites and adding extra increment trips associated with each category of special generator. The number of incremental trips for each special generator type is obtained by taking the difference of cross classification model generated trip rates and trip rates obtained from regional travel survey.

PAGE 110

101 Appendix A Continued A.8 Trip Attraction Rates of Shopping C enters (SCs) in Northern New Castle County, Delaware Apart from ITE trip generation manual, this paper gives two possible approaches to calculate trip attraction for shopping centers. These two approaches are based on the survey of the movement patterns (N o. of people visiting and No. of vehicles). 1) Macroscopic Approach: In this approach, Trip attraction rate is a function of physical features of shopping centers like total parking space, total floor area, no. of stores and location of shopping center. ITE t rip generation manual does not consider these physical features i.e. Phenomenon of trip chaining is not taken into account in ITE. ITE uses gross leasable area (in 1000 Sq.ft) as independent variable and average number of vehicle trips ends per one day to shopping center as dependent variable. The relationship between total parking space, total floor area, number of stores and trip attraction rate is obtained by regression analysis. Shopping centers are classified based on number of stores, number of parkin g, availability of supermarket and discount retail store. Based on the composition of the stores in the SC, they are classified into following 4 groups: Type 1: This is a large SC with a large supermarket, a large discount retail store, one or two restaur ants, a bank, and many small stores are located. Type 2: This is a medium size SC where a medium sized supermarket, a medium sized discount retail store and many smaller stores are located. Type 3: This is a small SC where one supermarket and several s mall stores are located. Type 4: This is a collection of specialty stores, but does not include a supermarket or discount retail store.

PAGE 111

102 Appendix A Continued Data used in the linear regression analysis is obtained by conducting traffic counts survey for eve ry 15 min interval at various sites of SCs on weekday, Saturday and Sunday. In macroscopic approach, two models are used as the variables used (number of stores, floor area and number of parking spaces) are highly correlated and convey the same information Macroscopic models depend on the physical features and not the type of stores that is they are insensitive to the nature of stores. 2) Microscopic Approach: In this approach, importance is given to each store in SC. The main objective here is to determine weights for trip attraction rates (TAR) of each store. The weighted sum of TARs of individual stores gives TAR of Shopping Center. TAR for different stores in a SC is obtained by conducting survey for 15 min interval. Stores are classified into major and m inor based on the shares of each store and weight for each store is obtained using optimization technique. A major drawback of the microscopic model is the large volume of data that is required for calculation of the TAR of individual stores and the weight s. The number of people entering individual stores needs to be collected for different time periods.

PAGE 112

103 Appendix A Continued A.9 A Comprehensive Planning and Design Manual for Parkand Ride Facilities: Chapter 5 Suburban Parkand Ride Demand Estimation Te chniques A.9.1 Post Modeling Techniques This technique is used for individual park and ride facilities and follows the traditional transportation modeling methodology. The steps involved in this modeling technique are as follows: 1) Identify the production ends (home zones) and attraction ends (work zones) of the potential park and ride site. 2) Identify the various characteristics of attraction ends such as parking cost, availability, traffic congestion etc. 3) Determine total person trip interchange between the production zones and the attraction zones by using modal splits from the regional travel model or other data sources. 4) Determine the proportion of trip interchange for Park and carpool and Bus Park and ride users based on the characteristics of bus services and trip end density in attraction zones. 5) Estimate the number of parking spaces required at each site by developing trip interchange tabulations based on the park and ride demand share. A.9.2 Direct Regional Forecasting Techniques In this regional forecas ting approach, park and ride trip is actually modeled as a chained trip directly within the regional modeling process. Here, utility functions are used as they provide a measure of the attractiveness of one mode relative to another.

PAGE 113

104 Appendix A Continued M ultinomial logit modeling approach is used as the basic theory behind the approach is that travelers will choose the mode which is quickest and cheapest mode of travel. Probability of choosing mode i is given by: = w here, Along with decision to select park and ride versus the auto mode, commuter also decides on which park and ride lot to be used depending upon the traffic congestion conditions. The park and ride lots immediately upstream of traffic congestion tend to have high levels of demand. Logit coefficients for park and ride lots can be estimated using a trial and error approach, comparing estimates to observed occupancy and origin surveys until a level of accuracy is obtained. A.9.3 Site Level Forecasting based on Si te and Service Characteristics This model is based on the theory that site attributes and service characteristics define the attractiveness of the site to potential users. Therefore, park and ride demand is estimated based on the attributes of the park and ride location. This model assumes that attractiveness of one mode over another can be estimated by measuring the differences in site and service attributes between competing modes.

PAGE 114

105 Appendix A Continued Site specific demand is heavily influenced by number of characteristics such as location of lot, service characteristics and availability of competing lots and perceived convenience of the facility. A Park and Ride demand estimation study is done in the Greater Seattle metropolitan area for the King County Department of Metropolitan Services (Metro) on all bus transit network. The park and ride facilities were examined for their existing demand characteristics and the draw area associated with the patrons accessing the lot. A 1993 vehicle license plate survey was used as the basis for geocoding the residential location of vehicles observed in each of the 31 lots. Addresses for each observed parked vehicle were generated via a license plate search with the Washington Department of Motor Vehicles. The coordina tes of each vehicle accessing individual lots were compared to the coordinates of the lot being used and then plotted on a common scale. The resulting service area demand sheds for each lot were compared to generate a catchment area shape. In general, this methodology is all about defining a service area (catchment area) for the park and ride facilities and then developing equations based on the lot attributes using multivariate regression analysis. 1) Defining the market catchment area for park and ride It is defined based on the differences in parking costs, extent of transit network and perceived congestion in a region. Socioeconomic data can be collected for the defined catchment area and can be used to predict demand for the specific park and ride lot.

PAGE 115

106 Ap pendix A Continued The shapes of the catchment area having 50 and 85 % of the total observed users at each park and ride lot were considered. At the 85% user level, a parabolic shape nearly represents a catchment area of the lots. A circular pattern with a radial diameter of 2 to 2.5 miles, centered at the park and ride itself describes the average catchment area at the 50 % demand level. Individual market areas are smaller than standard market areas because of features such as lakes and mountains which red uces the likelihood of travel. Using this catchment area shapes, overlaps and gaps between the park and ride facility services can be determined. This will help us map coverage zones of each facility and locate areas of service duplication and poor service

PAGE 116

107 Appendix A Continued 2) Identifying the site level characteristics affecting park and ride demand The variables that can affect the demand for park and ride facilities at site level are as follows: No. of AM peak period express buses trips to CBD no. of AM peak period express buses trips to major employment centers other than CBD r atio of out of pocket auto cost to transit costs d istance between park and ride lot and destination (CBD) t otal population within the 50 % catchment area of lot % of lower m iddle and lower income households within the service area of lot t he average best schedule transit time between park and ride lot and destination, peak traffic on adjacent roadway facility n o. of home based work trips between market area and destination, e mployment demand measure at the destination r elative measure of congestion between lot and destination a ge of park and ride lot a vailability of priority treatments s afety characteristics of lot provision of passenger shelter and amenities t ransit I nformation parking costs at the destinatio n and park and ride lot access attributes. Service area population was determined by plotting the catchment area over a map of the 1991 Puget Sound Regional Council's TAZ system. Assuming that population is evenly spread throughout each TAZ, visual estimates were made of the percent of TAZ included within the catchment area. Transit costs are calculated by averaging weighted transit cost and auto cost is estimated by averaging weighted parking costs and driving cos ts to the major activity centers.

PAGE 117

108 Appendix A Continued 3) Site level demand estimation The variables mentioned above can be used to develop a planning tool to estimate the demand potential for park and ride facilities. The park and ride demand model is shown by following equation: Demand = N + aAa + bBb + cCc +.zZz w here, N = Constant, incorporating a measure of the minimum lot size. A, B, C, Z = independent variables. a, b, c, z = model coefficients to be estimated using least square method. a, b, c, z = variable exponents estimates using a least square method. The no. of variables to be used in the regression analysis are controlled by availability of data from past surveys and existing database and the potential ease of developing similar data for the ev aluation of future lots. All the variables used should be independent of each other and mutual independence can be evaluated by constructing a correlation coefficient matrix for all variables. Various PRD equations are developed based on the data items av ailable and R squared values are calculated to understand the percentage of variability in the data explained by the model. The demand obtained represents both transit oriented and nontransitoriented (e.g., carpool) demand for parkandride spaces. The proportion of trip interchange for Park and carpool (Nontransit) and Bus Park and ride (Transit) users is determined based on the characteristics of bus services and trip end density in attraction zones.

PAGE 118

109 Appendix A Continued This PRD model cannot be direc tly transferred to other regions. The two options available to transfer the PRD model to another location are as follows: Estimate a new PRD model, estimating the coefficients for each of the variables used in the several PRD equations, or validate the Seattle PRD equations, developing a correction factor that compensates for the inherent differences between the region being studied and the Seattle metropolitan area.

PAGE 119

110 Appendix A Continued A.10 Application of a Park and Ride Forecasting Procedure in the Greater Vancouver Transportation Model The regional transportation model being a nested logit choice model, Park and Ride was treated as a sub mode of transit in nested logit structure. The Park and Ride impedance was computed as the sum of auto impedanc e from origin to Parkand Ride and the transit impedance from Parkand Ride to the destination along with the weights and submodal biases applied to the park and ride trips. Also, the catchment areas at origin and destination ends were identified to avoid the creation of illogical trip chains. Modeling of several park and ride sites competing for same potential users is complex but can be effectively and efficiently achieved using the matrix convolutions. Assumptions on which the Park and Ride model is b ased: 1) Transit riders with abundant free parking will not use Park and Ride. 2) If transit impedance from origin to destination zone is lower than that from park and ride lot to their destination, trip makers will not use Park and Ride. 3) The generalized cost of park and ride will include parking charges and penalty representing the uncertainty of finding a parking place where demand exceeds the capacity. 4) Trip makers will be reluctant to use Park and ride if travel time and cost saved is less as compared to auto. 5) To match the observed distribution of origins of park and ride trips, it was necessary to apply weight to the auto leg of the trip. This weight is dependent on transit mode served by park and ride.

PAGE 120

111 Appendix A Continued 6) The effective transit impedance for trip makers using park and ride will be lower than for trip makers with no access to park and ride. The procedure to consider park and ride lots in travel model has following steps: 1) Compute Auto and Transit impedances for all dummy zones pairs representi ng park and ride sites. 2) Compute park and ride impedance using minimum path based on MIN PRI (ij) = Minimum (AI (ik) Wkm + TI (kj) + Pkm + SPk) w here, i = Origin j = Destination k = Park and Ride site. MIN PRI (ij) is the minimum park and ride impedance for all logical path i kj. AI (ik) is the Auto Impedance from i to k including any parking charge collected at k. Wkm is an Auto Impedance weight applied to all trips to k TI (kj) is the transit impedance from k to j. Pkm is a penalty or modal bias a pplied to all trips to k. It depends on transit mode (m) served by k. SPk is an additional penalty applied to all trips to k to ensure that parking demand does not exceed the available capacity.

PAGE 121

112 Appendix A Continued 3) Calculate enhanced transit impedance for origin destination with access to park and ride using: ETR (ij) = (ln (exp ((kj)) + exp ( (ij))))/ (Where, ETR (ij) is the enhanced transit impedance. model. 4) Run the distribution and auto/ transit mode split using ETR (ij) as the transit impedance. 5) Split forecast choice transit trips into walk and park and ride access modes using logit function 6) Compute park and ride impedance for each logical path based on: PRI (ij) = AI (ik) Wkm + TI (kj) + Pkm + SPk 7) Distribute forecast park and ride trips among competing park and ride lots based on following multinomial logit function: PRT (ij) = PRT (ij) exp ( (ij) (k) (exp ( (ij))) 8) After comparing the transit impedance with estimated peak hour capacity of parking lot, recalculate MIN PRI (ij) and PRT (ij) for all k. 9) Repeat split forecast choice transit trips and comparing transit impedance steps until demand at overloaded lots converge almost equal to capacity. 10) Separate forecast park and ride trips into auto and transit trips components. 11) Add auto leg of park and ride trips to the auto trip matrix and add the transit leg of park and ride trips to transit trip matrix. Assign auto and transit trips

PAGE 122

113 Appendix A Continued A.11 Intermodal Terminals This step of the freight carrier modeling process estimates the average total freight trips by mode that would be generated by the planned facil ity for a specific time period (daily, annual, etc.). The total trips generated by the facility include both production, originating from the facility, and attraction, destined to the facility, trips. The most common methods used for facility trip generati on include trip generation rates, regression equations, and surveys. Using trip generation rates is the simplest approach for trip generation, in which estimates of number of trips per employee are applied to the target facility to estimate the total trips generated. Trip generation rates also can vary based on truck types and the type of facility (land use). The trip generation rates used in this approach can be derived from previous surveys of freight flows associated with similar facilities or from stand ard sources providing average trip generation rates for facilities, based on facility and truck types. The use of regression equations for trip generation offers the ability to predict the total trips generated as a function of more than one facility varia ble, which makes this approach potentially more robust and reliable compared to the use of trip generation rates. For example, a regression equation predicting total daily freight trips as a function of land use category, number of employees, and building/ floor area. However, caution should be maintained when developing and using regression equations for trip generation, as equations with statistical inconsistencies will not result in reliable estimates.

PAGE 123

114 Appendix A Continued Conducting surveys is the mos t time and cost intensive approach for trip generation, but it can provide the most accurate results, compared to trip generation rates and regression equations. This approach is useful in the case of special trip generators such as intermodal terminals, in which trip generation estimates are derived through direct contacts with a limited number of firms (facility operators and users truck companies shippers, etc.). This approach is particularly effective if the planning agency has been building contact s with the freight community over a longer period of time.

PAGE 124

115 Appendix A Continued A.12 Discount Superstore Trip Generation This study aims at developing trip rates which truly represents the trip generation characteristics of discount superstores like Wal M art. To achieve this, a national discount superstore trip generation study was conducted by Texas Transportation Institute (TTI). An unbiased sample of 32 study sites was randomly selected from the 828 stores in original sample. The original sample was sel ected based on the following selection criteria: 1) Standard superstores (i.e., stores may or may not contain lube and tire centers and/or garden centers) 2) Located in a standard metropolitan statistical area (MSA). 3) At least two years old. 4) Free standing stores that could be isolated to perform an accurate count of inbound and outbound vehicles. 5) No construction, special promotions, or events at the store. Once the sites were selected, the study consisted of following steps: 1) Collecting site and trip generation da ta at 32 stores for the typical season during September to mid November and also the peak season (Thanksgiving and week prior to Christmas) with the help of trained supervisors. 2) Analyze data to determine trip generation rates or equations for both typical and peak seasons.

PAGE 125

116 Appendix A Continued Trip generation rates obtained using the survey data varied between individual superstores. Rates were developed using Gross Floor Area (GFA) as an independent variable. GFA data was obtained from an architecture fir m. The results show that the rates obtained from the national study are higher than the ITE trip rates, but the differences are not statistically significant except for the Sunday daily rate. The study concludes that high degree of variability and small numbers of observations in the ITE data are the reasons for this difference in the trip rates.

PAGE 126

117 Appendix A Continued A.13 Transit Impact Fee Analysis: Technical Memorandum # 2 Land Use and Trip Generation R ates This study is all about exploring land uses that might be incorporated into an expanded Transit Impact Development Fee (TDIF) and describing trip generation rates associated with these land uses for the San Francisco planning department. Based on the preliminary assessment of potential transit trip ge neration for each land use, following land uses categories were identified as potential candidate for generating high number of transit trips: 1) Office a. Professional/Business Office b. Professional Design Office 2) Lodging a. Hotel/Motel 3) Institutions a. Hospital, medical center b. Social/charitable service c. Child care facility d. Elementary/Secondary/Post secondary school e. Churches or other religious institution 4) Community Facilities a. Community Club House b. Community Cultural center

PAGE 127

118 Appendix A Continued 5) Assembly and Entertainment a. Th eatres b. Recreation Building c. Amusement Enterprise and parks/Citrus/Carnival d. Open air Stadium or arena 6) Commercial (Retail)/ Personal Services a. Local Oriented retail b. Regional retail c. Bar d. Full service restaurant e. Financial Services 7) Manufacturing and Processing a. Li ght Manufacturing assembly, packing, repair, processing b. Light Food Processing The trip rates for the above land uses are obtained from the following sources: 1) San Francisco Interim Transportation Impact Analysis Guidelines for Environmental Review, Interim Edition, January 2000, The Planning Department City and County of San Francisco. 2) Trip Generation, 6th Edition, Institute of Transportation Engineers 3) Citywide Travel Behavior Survey, Employees and Employers, may 1993, San Francisco Planning Department, San Francisco Public Utilities Commission, and San Francisco County Transportation Authority.

PAGE 128

119 Appendix A Continued A.14 Quantifying Special Generator Ridership in Transit Analyses The authors aim at developing an analysis process for analyzing the impact of s pecial generators on transit services in Denver area. The study area is Gold Line a freight rail corridor between downtown Denver and Golden, Colorado. According to the authors, there are three types of special generators: 1) Regular special generators are those special generators that produce trips on a regular, weekday basis. Examples: airports, regional shopping centers, hospitals and schools. 2) Periodic special generators are those generators that do not produce trips on a regular weekday basis. Exam ples: convention centers, stadia and arenas, fairs and festivals. 3) Special special generators include those sites or activities that cannot be easily classified as regular or periodic special generators. This paper focuses on impact evaluation of periodic special generators. In the analysis process, generators having 500,000 attendees annually or 8,000 average individual event attendees were only considered. If a generator did not meet the size criteria, it was merged with other events occurring at same pl ace.

PAGE 129

120 Appendix A Continued The trip generation model was developed for the periodic special generators: A.14.1 Trip Generatio n This model is based on the attendance projections, which can be estimated from the past history of the attendance for a particul ar type of generator. The trips are also allocated to trip purposes. For events on weekdays, the trips were split between home based non work and non home based trip purposes and for events on weekends, all trips are taken as home based non work trips. Tri p generation analysis is done on daily basis for all the periodic special generators. The annualization factor is total number of events days in a year. The trip generation results for the year 2015 for various special generators are obtained using the model. The annual growth factors are obtained based on the information from the operators of the special generators. A.14.2 Trip Distribution Home based nonwork and nonhome based trip attractions and nonhome based trip productions are estimated for the periodic special generators. Trip distribution of periodic special generators is simply a proportioning of trips from all parts of the region to a single site (for each periodic special generator). The basis for the proportioning mainly depends on the chara cteristics of each periodic special generator. The distribution of trips to and from the periodic special generators is made using a gravity model formulation. The gravity model is typically used to distribute trips from one origin to all destinations, not from one destination (i.e., the special generator) to all origins. However, the model can be applied in either direction.

PAGE 130

121 Appendix A Continued Zonal home based nonwork or nonhome based productions are used along with the periodic special generator att ractions for the generator. The distribution for each periodic special generator is independent of other generator. The results for both the distributions are obtainedone for home based nonwork trips and one for nonhome based trips. A.14.3 Mode Choice Mode Choice is dependent on numerous items including auto and transit travel times and costs. Mode choice model is used for home based nonwork and nonhome based trips. To simplify the mode choice procedure, trips obtained from trip distribution step are multiplied by the appropriate annualization factor. Several changes were made to the model in order to replicate base mode shares. The average event parking costs at the attraction zone was coded for each periodic special generator. The parking costs are a djusted to account for the average auto occupancy noted for each event. A.14.4 Transit Assignment Annual transit assignment is performed due to following reasons: 1) Each periodic special generator has a unique annualization factor. 2) This procedure eliminates the need to perform separate transit assignments for each special generator. This step gives the annual periodic special generator boardings.

PAGE 131

122 Appendix A Continued A.15 WACOG Connector Program Transit Feasi bility and Implementation Plan The main purpose of this report is to estimate potential transit demand between Bullhead City, Kingman, and Lake Havasu City in Arizona by examining transit dependent population and other potential riders. Arkansas Public Transportation Needs Assessment (APTNA) method and S urvey Research method were used to develop an intercity transit demand model. Transit demand is obtained by applying trip rates to the transit dependent population groups (elderly persons ages 60 and over, persons with disabilities under age 60, and persons living in poverty under age 60) in the APTNA method. The % of this population groups is obtained from the Arizona State Transit Needs Study. The frequency of the use of transit by different population groups mentioned above is obtained by the survey rese arch m ethod. A.15.1 Trip Production Based on the production and attraction between cities, gravity model was used to determine the intercity transit travel demand. The trip production for each city is calculated by multiplying the frequency of the use of transit by different population groups with the percentage of their respective population that uses transit. A.15.2 Trip Attraction Firstly, various trip purposes (medical, education, employment, recreation and county services) for which the trips are att racted are identified. Then, the proportion of trips attracted by these purposes (services) for the transit dependent population is obtained for each city.

PAGE 132

123 Appendix A Continued The annual trips by transit is obtained for medical services based on the num ber of beds in the hospital, for education services based on the number of enrollments in the college. The trips attracted by employment are obtained based on the labor force data for that year. Finally, the total intercity transit trips are calculated using the gravity model.

PAGE 133

124 Appendix A Continued A.16 Sketch Model to Forecast Heavy Rail Ridership The purpose of this paper is to study ridership potential for heavy rails by developing a model that considers variables related to area surrounding the station. A multivariate linear regression model was created only for nonCBD stations using current ridership data collected for all 474 U.S heavy rail transit stations for the years 2004 2006. The demographic information for both areas surrounding the stations a nd entire metropolitan area was obtained from the respective MPOs. Model was developed using data from following ten cities: Baltimore, Boston, Chicago, Cleveland, Los Angeles, Miami, New York (PATH train), Philadelphia, San Francisco, and Washington, DC Exclusive regions around each station were used so as to avoid double counting of population and employment around station areas. Various independent variables related to station area demographics, station specific transportation attributes, corridor demographic characteristics and metro area demographics were tested in the model. Along with the independent variables, the natural logarithm of the independent variables was also tested. The Person Product Moment was used to test possible linear correlations and Spearmans nonparametric coefficient was used to test possible nonlinear correlations between independent and dependent variables. The results show that best predictor of actual boarding is employment and transit service characteristics are the bes t predictors of natural logarithm of boarding.

PAGE 134

125 Appendix A Continued Following variables turned out to be statistically significant and important predictors: A.16.1 Binary V ariables 1) 1 if this is a terminal station,0 if not 2) 1 If this station is a second ary downtown, 0 if not 3) 1 if this is a special transit attractor station, 0 if not 4) 1 if there is parking available, 0 if not 5) 1 if there is connection to other rail modes, 0 if not A.16.2 Continuous V ariables 1) Distance to downtown, in miles 2) Midday headway in minutes 3) CBD density, in employees per square mile 4) Employment within 0.25 miles of the station 5) Employment within 0.25 to 0.5 miles of the station 6) Population within 0.25 to 0.5 miles of the station The results show a positive and strong relationship bet ween actual and predicted boarding for all 381 nonCBD stations. To evaluate the proposed lines or rail extensions, the results of the model are aggregated to route level as well as city level. The model performs better at the city level with an R squared value of 0.814 as compared to R squared value of 0.702 for the route level model. The model can be applied to other cities with similar characteristics based on the mean and standard deviations of the variables used in the model. Also model can be improved by considering the nonlinear models.

PAGE 135

126 Appendix A Continued A.17 Direct Ridership Forecasting Travel demand models do not consider changes in station level land use and transit service characteristics. So, direct ridership models are used to forecast transit patronage. Direct Ridership models have been used to evaluate and compare various variables influencing transit patronage. They are used for light rail [Sacremento regional transit (RT) & Salt Lake City (TRAX)], Commuter rail [Sonoma marin area rail transit (SMART)] and Heavy rail [Bay area rapid transit (BART). Direct ridership models use multivariate regression analysis based on the local land use data and data obtained from boarding & alighting counts at all stations. 30 Variables related to populat ion and income, employment, cost of travel, station characteristics, transit service characteristics and comparative auto and transit accessibility were used to discover combination of variables with stronger correlation with ridership. For BART, Ridership i s a function of variables like s um of population and employment within mile of station (POPEMP), population within station catchment area (POPCTCH), frequency of peak period feeder buses (BUS), number of station parking spaces (PARK), number of peak pe riod trains (TRAINS) and Train vehicle type;1 = BART & 0 = Caltrain (TECH). The two formulaes used are as follows: RIDERSHIP = 2.04 + 0.300 X POPEMP + 0.069 X POPCTCH + 0.560 X TRAINS + 1.787 X TECH RIDERSHIP = 2.400 + 0.233 X POPEMP + 0.021 X POPCTCH + 0.287 X BUS + 0.038 X PARK + 0.477 XTRAINS + 1.576 X TECH

PAGE 136

127 Appendix A Continued Also, log log regression analysis was performed to estimate elasticitys of the above mentioned variables. Similarly, models are used to obtain variables affecting commuter a nd light rail trains ridership.

PAGE 137

128 Appendix B: Results of the Exploratory Analysis Performed for the E nhancement of Trip Attraction Capability Table 30 Total Employment, Trips Generated by Employme nt, Trips Generated Using Trip Rates Area of Non Residential Land U ses and Special Gener ators for the Stops in all the F our Routes. Sr. no. Route Name Stop N ame Total Employment Trips Generated by Employment Trips Generated using Parcel land use based Trip Rates Absolute Percent Difference Total Weekday Boarding Area of the Non Residential land use Area of Special Generator Special Generator Dummy 1 F1 5000 Susquehanna St. 2,612 3,531.00 590.72 83.27 20.69 35917 0 0 2 F1 Edgewood Ave. & Shenandoah Ave. 2,345 3,211.60 576.06 82.06 15.07 323 57 0 0 3 F1 Edgewood Ave. & Chenago Blvd. 2,062 2,868.80 531.57 81.47 1.39 30473 0 0 4 F1 2626 W Edgewood Ave. 2,031 2,826.00 451.52 84.02 5.32 24343 0 0 5 F1 Edgewood Ave. & Edward St. 209 285.20 440.04 54.29 11.04 41160 0 0 6 F1 12th St. & West Palm Ave. 213 294.00 403.63 37.29 33.54 34506 0 0 7 F1 3120 12th St. 80 110.60 121.95 10.26 0.97 8311 2996 1 8 F1 12th St. & Melson Ave. 7 9.00 70.01 677.89 6.84 2996 2996 1 9 F1 12th St. & Prospect St. 9 11.80 70.01 493.31 4.89 2996 2996 1 10 F1 12th St. & Detroit St. 83 116.20 0.00 100.00 6.70 0 0 0 11 F1 Detroit St. & 10th St. 31 42.60 0.00 100.00 2.90 0 0 0 12 F1 Detroit St. & 8th St. 92 128.00 0.00 100.00 7.15 0 0 0 13 F1 Detroit St. & 6th St. 99 137.40 462.14 236.35 7.23 8024 7058 1 14 F1 Detroit S t. & 5th St. 96 133.20 761.52 471.71 8.89 10949 9983 1 15 F1 Detroit St. & 1st St. 17 23.80 39.61 66.43 6.16 4505 0 0 16 F1 Detroit St. & Detroit Cir. 84 113.40 39.61 65.07 6.26 4505 0 0 17 F1 Detroit St. & Commwealth Ave. 116 158.20 136.36 13.81 2.67 7 710 0 0 18 F1 Detroit St. & Lowell Ave. 121 165.20 38.39 76.76 1.58 4205 0 0 20 F1 Detroit St. & Broadway Ave. 162 224.80 758.60 237.46 4.41 31501 5220 1 21 F1 Broadway Ave. & St. Clair St. 132 182.80 1,755.71 860.45 4.59 193535 5220 1 22 F1 Broadway A ve. & Huron St. 105 143.80 635.68 342.06 0.95 114112 0 0 23 F1 Broadway Ave. & Superior St. 187 249.60 312.42 25.17 4.55 40893 0 0 24 F1 2606 Broadway Ave. 174 231.60 494.39 113.47 1.78 29116 0 0 25 F1 Broadway Ave. & McDuff Ave. 253 330.80 1,142.54 245 .39 6.41 67193 0 0 26 F1 McDuff Ave. & Beaver St. 299 392.00 1,185.59 202.45 9.65 79272 0 0 27 F1 McDuff Ave. & Strickland St. 361 477.40 1,187.25 148.69 0.34 85453 0 0 28 F1 McDuff Ave. & Warrington St. 275 370.40 141.82 61.71 5.29 12311 0 0 29 F1 McD uff Ave. & Fitzgerald St. 282 378.00 299.49 20.77 0.00 15979 3668 1 30 F1 2978 Fitzgerald St. 279 379.00 327.62 13.56 15.16 20010 3668 1 31 F1 Fitzgerald St. & Willow Branch Ave 296 398.60 388.32 2.58 5.45 25822 3668 1 32 F1 Fitzgerald St. & Cherokee St 128 167.60 388.69 131.92 4.17 25924 3668 1 33 F1 McCoy Creek Blvd. & Sunshine St. 26 31.60 173.43 448.83 0.43 19614 0 0 34 F1 McCoy Creek Blvd. & Leland St. 146 203.80 438.87 115.34 1.79 25896 0 0 35 F1 McCoy Creek Blvd. & King St. 182 254.40 438.87 7 2.51 7.75 25896 0 0 36 F1 McCoy Creek Blvd. & Nixon St. 189 260.40 515.61 98.01 1.06 31295 0 0 37 F1 Forest St. & Stockton St. 155 196.40 476.03 142.38 4.17 64917 0 0 38 F1 Forest St. & Woodlawn Ave. 277 367.20 719.62 95.97 0.92 134932 0 0 39 F1 Forest St. & Claude St. 328 437.60 1,051.14 140.21 1.56 178650 33368 1 40 F1 Forest St. & Copeland St. 563 763.00 935.74 22.64 4.38 162111 33368 1 41 F1 Forest St. & Goodwin St. 539 737.00 1,088.74 47.73 0.00 208788 33368 1

PAGE 138

129 Appendix B Continued Table 30 C ont inued Sr. no. Route Name Stop N ame Total Employment Trips Generated by Employment Trips Generated using Parcel land use based Trip Rates Absolute Percent Difference Total Weekday Boarding Area of the Non Residential land use Area of Special Generator Spe cial Generator Dummy 42 F1 Forest St. & Park St. 1,292 1,719.00 1,431.93 16.70 0.00 98933 0 0 43 F1 Park St. & Price St. 1,552 2,085.00 1,091.68 47.64 7.86 85497 0 0 44 F1 Park St. & Jackson St. 1,376 1,908.80 526.34 72.43 1.92 36306 0 0 45 F1 Water St & Jefferson St. 7,589 10,480.20 1,965.25 81.25 0.47 63289 0 0 46 F1 550 Water St. 9,692 13,316.20 7,647.36 42.57 0.18 235222 0 0 47 F1 Pearl St. & Water St. 12,182 16,730.40 8,032.69 51.99 2.09 251492 0 0 48 F1 Forsyth St. & Julia St. 23,225 32,174.00 16,677.95 48.16 0.55 577605 4874 1 49 F1 Forsyth St. & Laura St. 23,728 32,561.20 40,412.44 24.11 3.82 1524033 11685 1 50 F1 Forsyth St. & Ocean St. 22,410 30,714.80 30,244.94 1.53 0.48 1206460 11735 1 51 F1 Newnan St. & Adams St. 8,691 11,727.20 22,18 9.02 89.21 5.82 913534 11735 1 52 F1 Newnan St. & Duval St. 7,492 10,066.40 28,922.68 187.32 0.53 1185923 73981 1 53 F1 Ocean St. & Beaver St. 16,854 23,203.00 39,435.49 69.96 1.92 1607201 62246 1 54 F1 F.C.C.J. Station 14,456 20,197.20 30,239.71 49.72 101.53 1295519 85440 1 55 F1 Church St. & Newnan St. 6,226 8,313.00 27,222.83 227.47 0.05 1087644 52995 1 56 F1 Newnan St. & Ashley St. 5,878 7,836.80 29,862.56 281.06 1.43 1096236 69107 1 57 F1 Newnan St. & Brown St. 5,660 7,530.40 20,494.71 172.16 0.0 0 753088 69057 1 58 F1 Newnan St. & Ashley St. 5,467 7,280.60 7,643.59 4.99 0.46 327235 62246 1 60 F1 Beaver St. & Ocean St. 6,338 8,491.00 31,117.96 266.48 1.32 1329482 80566 1 61 F1 Beaver St. & Market St. 2,215 2,999.80 2,452.71 18.24 34.00 195657 62 246 1 62 F1 Union St. & Cemetery St. 643 839.00 455.33 45.73 0.41 46686 0 0 63 F1 Union St. & Palmetto St. 426 532.40 806.25 51.44 0.10 47529 4956 1 64 F1 Union St. & Spearing St. 428 535.80 1,266.67 136.41 0.15 63441 4956 1 65 F1 Union St. & A. Philip Randolph Bl 132 177.40 1,252.56 606.07 2.22 63571 4956 1 66 F1 Union St. & Van Buren St. 100 132.60 1,001.30 655.13 1.37 57897 4956 1 67 F1 Union St. & Franklin St. 109 144.00 1,036.95 620.10 0.88 64148 4956 1 68 F1 Franklin St. & Pippin St. 204 273.60 1,029.01 276.10 1.92 63278 4956 1 69 F1 Franklin St. & Odessa St. 249 326.20 2,928.91 797.89 2.10 128757 4956 1 70 F1 Franklin St. & Jessie St. 380 489.60 2,650.85 441.43 10.55 197328 0 0 71 F1 Franklin St. & Phelps St. 386 495.60 2,697.26 444.24 7.96 210349 0 0 72 F1 Franklin St. & E 1st St. 307 395.60 929.27 134.90 7.17 179032 1992 1 73 F1 1151 Franklin St. 484 635.80 1,088.93 71.27 1.71 212802 1992 1 74 F1 Franklin St. & E 3rd. St. 490 641.00 1,316.07 105.32 3.42 283968 1992 1 75 F1 Franklin St. & E 4th St. 381 516.60 980.18 89.74 2.92 189719 1992 1 76 F1 4th St. & Milnor St. 371 502.60 980.18 95.02 2.48 189719 1992 1 78 F1 Milnor St. & 5th ST 508 698.80 544.18 22.13 4.26 73599 2162 1 79 F1 1701 7th St. 514 715.40 880.26 23.04 2.63 59680 2162 1 80 F1 7th St. & Florida Ave. 454 625.60 986.57 57.70 0.00 67051 2162 1 81 P7 Herlong Rd & Fouracker Rd 51 69.80 148.28 112.44 26.42 15928 0 0 82 P7 Fouracker Rd & La Trec Dr 301 418.80 136.29 67.46 0.85 14929 0 0 83 P7 Fouracker Rd & Renoir Dr 629 830 .00 1,277.33 53.90 1.41 25732 0 0 84 P7 Fouracker Rd & Normandy Bvld 874 1,127.60 1,211.54 7.44 10.19 9755 0 0 85 P7 7952 Normandy Blvd 1,125 1,429.00 1,876.29 31.30 5.98 12749 6339 1 86 P7 Normandy Bl & Normandy Village Pwy 679 831.20 665.26 19.96 3.68 7759 6339 1

PAGE 139

130 Appendix B Continued Table 30 C ontinued Sr. no. Route Name Stop N ame Total Employment Trips Generated by Employment Trips Generated using Parcel land use based Trip Rates Absolute Percent Difference Total Weekday Boarding Area of the Non Resi dential land use Area of Special Generator Special Generator Dummy 87 P7 Normandy Blvd & Memorial Pkwy 572 743.60 3,759.36 405.56 8.06 54825 12871 1 88 P7 7016 Nomandy Blvd 400 501.80 4,026.77 702.47 2.46 70264 12871 1 89 P7 Normandy Blvd & La Marche Dr 629 764.80 185.53 75.74 13.52 8100 0 0 90 P7 Normandy Blvd & Granville Rd 489 606.80 231.11 61.91 2.98 10626 0 0 91 P7 Normandy Blvd & Lane Ave 863 1,079.60 2,594.19 140.29 13.60 79069 2776 1 92 P7 Normandy Blvd & Verna Blvd 733 942.80 2,413.66 156.01 7.31 73983 2776 1 93 P7 Normandy Blvd & Fountain Rd 675 884.60 2,997.47 238.85 1.31 79049 67734 1 94 P7 Normandy Blvd & Ellis Rd 473 628.20 3,020.04 380.74 2.74 81521 67734 1 95 P7 5476 Normandy Blvd 492 648.20 3,248.53 401.16 6.72 130831 67734 1 96 P7 5320 Lenox Ave 1,093 1,398.40 3,653.32 161.25 2.15 166109 86475 1 97 P7 Lenox Ave & Verna Rd 885 1,131.60 2,477.45 118.93 4.23 73353 51633 1 99 P7 Lenox Ave & Garth Ave 621 795.00 2,288.45 187.86 2.47 91689 19950 1 100 P7 Lenox Ave & Cassat Ave 649 818 .60 2,524.82 208.43 6.92 116428 19950 1 101 P7 Cassat Ave. & Lenox Ave 786 1,027.20 1,952.83 90.11 0.47 125458 2524 1 102 P7 4782 Lenox Ave 1,072 1,401.40 1,510.29 7.77 0.27 113957 5381 1 103 P7 Lenox Ave & Edgewood Ave 969 1,260.00 1,467.59 16.48 15.07 110948 5381 1 104 P7 Edgewood Ave & Roselyn St 703 942.80 1,390.42 47.48 3.42 105990 5381 1 105 P7 Edgewood Ave & College St 335 432.20 1,751.02 305.14 2.33 56848 33624 1 106 P7 Post St & Cypress St 409 526.40 1,562.85 196.89 16.12 50493 36121 1 107 P 7 Post St & Nelson St 189 240.80 860.10 257.18 4.72 19760 15067 1 108 P7 Post St & Brierfield Dr 119 158.40 41.33 73.91 2.64 7039 0 0 109 P7 Post St & Day Ave 136 179.40 437.43 143.83 7.22 20498 2745 1 110 P7 Post St & Shearer Ave 266 331.00 664.42 100. 73 0.95 28863 2745 1 111 P7 Post St & Plymouth St 315 393.80 1,930.53 390.23 13.08 45168 2745 1 112 P7 Post St & Willow Branch Ave 399 501.40 1,862.36 271.43 7.06 35607 3545 1 113 P7 Post St & Cherry St 178 232.40 1,301.16 459.88 7.96 18577 800 1 114 P 7 Post St & James St 224 288.80 186.15 35.54 0.59 8388 0 0 115 P7 Post St & King St 375 476.80 405.76 14.90 8.73 20916 0 0 116 P7 Post St & Acosta St 476 620.40 642.37 3.54 1.75 36505 5741 1 117 P7 Post St & Barrs St 705 929.20 605.82 34.80 0.91 45946 0 0 118 P7 Post St & Stockton St 702 928.40 949.12 2.23 5.87 50143 3438 1 119 P7 Post St & Osceola St 659 863.20 856.64 0.76 0.56 63845 658 1 120 P7 Post St & Copeland St 1,219 1,614.00 1,197.84 25.78 1.66 166761 0 0 121 P7 Post St & Goodwin St 1,179 1, 571.40 3,271.11 108.17 0.22 202950 13397 1

PAGE 140

131 Appendix B Continued Table 30 C ontinued Sr .no. Route Name Stop N ame Total Employment Trips Generated by Employment Trips Generated using Parcel land use based Trip Rates Absolute Percent Difference Total Weekda y Boarding Area of the Non Residential land use Area of Special Generator Special Generator Dummy 122 P7 Post St & Margaret St 1,579 2,111.00 3,776.85 78.91 5.60 206922 23136 1 123 P7 Park St & Riverside Pk 1,666 2,215.20 7,524.16 239.66 5.35 187129 23136 1 124 P7 Park St & Roselle St 5,913 8,161.20 5,887.74 27.86 2.44 186705 9739 1 125 P7 Park St & Edison Ave 5,577 7,710.60 1,938.55 74.86 0.83 102374 0 0 126 P7 Park St & Price St 1,552 2,085.00 1,091.68 47.64 5.01 85497 0 0 127 P7 Park St & Jackson St 1,376 1,908.80 526.34 72.43 2.61 36306 0 0 128 P7 Park St & Stonewall St 2,973 4,147.80 331.20 92.02 0.62 28391 0 0 129 P7 Water St & Jefferson St 7,601 10,492.80 1,965.25 81.27 0.58 63289 0 0 130 P7 Pearl St & Bay St 12,182 16,730.40 8,032.69 51.99 3.11 251492 0 0 131 P7 Forsyth St & Laura St 21,817 30,207.80 39,499.54 30.76 4.29 1424034 11685 1 132 P7 Forsyth St & Ocean St 22,763 31,202.00 30,342.72 2.75 0.43 1214902 11735 1 133 P7 Newnan St & Adams St 8,483 11,436.00 22,237.49 94.45 7.52 916345 11735 1 134 P7 Newnan St & Monroe St 7,859 10,565.80 28,809.63 172.67 0.32 1179367 73981 1 135 P7 Newnan St & Ashley St 5,878 7,836.80 29,862.56 281.06 2.26 1096236 69057 1 136 P7 Beaver St & Ocean St 6,318 8,463.00 31,117.96 267.69 4.78 1329482 80566 1 137 P7 Beaver St. & Laura St. 16,248 22,368.40 39,603.70 77.05 0.53 1636111 57869 1 138 P7 Beaver St. & Julia St. 15,128 21,111.80 25,083.70 18.81 0.86 1165508 52995 1 139 P7 F.C.C.J. Station 4,158 5,771.60 23,832.59 312.93 136.80 1066290 52995 1 140 P7 State St. & Julia St. 1,547 2,123.20 10,774.65 407.47 5.85 591368 52995 1 141 P7 Pearl St & 1st St 1,099 1,514.40 1,164.02 23.14 2.91 31574 0 0 142 P7 Pearl St & 4th St 1,260 1,757.80 967.65 44.95 0.65 64440 0 0 143 P7 Pearl St & 5th St 2,475 3,456.80 1,178.45 65.91 1.29 70875 0 0 144 P7 Pearl St & 6th St 1,335 1,858.60 2,170.70 16.79 2.95 58276 13424 1 145 P7 Pearl St & 8th St 1,671 2,329.20 2,297.75 1.35 0.13 78035 19349 1 146 P7 8th St Perry St 2,947 4,112.40 1,710.18 58.41 1.15 49330 14726 1 1 47 P7 8th St Boulevard St 4,372 6,078.00 1,218.70 79.95 14.52 45306 15986 1 148 P7 8th St. & Illinois St. 4,290 5,963.20 431.10 92.77 0.00 18296 8059 1 149 P7 8th St. & James Hall Dr. 3,056 4,240.60 94.86 97.76 34.56 5743 5743 1 150 P7 8th St. & Venus S t. 2,807 3,892.00 94.86 97.56 9.03 5743 5743 1 151 P7 8th St. & Francis St. 334 456.20 213.05 53.30 0.00 15855 13906 1 152 P7 Davis St. & 8th St. 2,904 4,025.40 243.17 93.96 0.02 17764 17764 1 153 P7 Davis St. & Reiman St. 1,745 2,404.00 31.14 98.70 0.3 3 1885 1885 1 154 P7 Davis St. & 11th St. W 107 136.00 79.47 41.57 0.41 4688 1885 1 155 P7 Davis St. & 13th St. 122 160.60 114.34 28.80 0.43 8507 1885 1 156 P7 Davis St. & Lincoln Ct. 127 167.60 156.04 6.90 4.50 13075 1885 1 157 P7 Davis St. & 17th St. W 88 113.60 86.60 23.77 0.00 9486 0 0

PAGE 141

132 Appendix B Continued Table 30 C ontinued Sr. no. Route Name Stop N ame Total Employment Trips Generated by Employment Trips Generated using Parcel land use based Trip Rates Absolute Percent Difference Total Weekday Boa rding Area of the Non Residential land use Area of Special Generator Special Generator Dummy 158 P7 18th St. & Venus St. 53 65.00 86.60 33.23 0.42 9486 0 0 159 P7 18th St. & Jupiter St. 72 91.40 505.01 452.53 1.11 14867 3826 1 160 P7 18th St. & Flanders St. 72 95.20 708.42 644.14 0.76 18593 7517 1 161 P7 Boulevard St. & 18th St. 140 185.00 1,282.80 593.41 4.18 37079 7517 1 162 P7 21st St & Boulevard St 46 61.40 1,450.13 2,261.78 2.47 47151 7517 1 163 P7 21st St & Saturn Ave 31 42.60 628.82 1,376.10 0. 43 14739 7517 1 164 P7 21st St & Brentwood Ave 36 49.20 505.01 926.44 4.73 14867 3826 1 165 P7 21st St & Davis St 31 42.20 565.40 1,239.81 0.05 18971 3826 1 166 P7 Davis St & 23rd St 52 69.60 138.93 99.61 1.23 11326 0 0 167 P7 Brick Rd & 25th St 43 57. 00 97.23 70.58 0.41 6758 0 0 168 P7 Brick Rd & 27th St 52 67.00 716.70 969.70 7.30 15992 11123 1 169 P7 Brick Rd. & 28th St. 123 156.00 649.69 316.47 1.04 12545 11123 1 170 P7 Brick Rd & 30th St 174 226.20 1,512.26 568.55 0.94 18764 16030 1 171 P7 30th st. & Brick Rd. 167 221.40 1,512.26 583.04 0.81 18764 16030 1 172 P7 Golfair Blvd & Brentwood Av 208 274.00 3,042.30 1,010.33 1.10 39320 24304 1 173 P7 Brentwood Ave & Woodbine St 219 289.40 3,012.97 941.11 4.94 38041 24304 1 174 P7 Brentwood Ave & Alder St 296 398.60 2,794.33 601.04 0.45 38490 21633 1 175 P7 4731 Norwood Ave 919 1,175.00 1,244.96 5.95 0.86 19072 0 0 176 P7 5030 Norwood Ave 969 1,239.00 1,225.96 1.05 1.01 17432 0 0 177 P7 Gateway Mall 826 1,044.20 101.83 90.25 61.53 7933 0 0 178 P7 5839 Norwood Ave 837 1,050.40 49.66 95.27 5.60 4418 0 0 179 P7 Norwood Ave & Crestwood St 235 308.00 612.91 99.00 2.57 17432 0 0 180 P7 Norwood Ave & Lynton St 279 368.20 654.19 77.67 0.73 20229 0 0 181 P7 Norwood Ave & Laurel St 219 284.80 657.88 131.0 0 1.42 20435 0 0 182 P7 Norwood Ave & Essex St 191 247.00 492.13 99.24 0.83 12693 0 0 183 P7 Norwood Ave & Carrollton Rd 176 227.20 217.59 4.23 3.38 11865 6121 1 184 P7 Dunn Ave & Regency Dr 1,008 1,256.20 7,978.22 535.11 16.67 73893 0 0 185 P7 1057 Dunn Ave 932 1,166.80 5,608.19 380.65 2.84 13264 0 0 186 P7 Dunn Ave & Bonnelly Dr 1,120 1,400.60 3,741.66 167.15 3.20 27636 19848 1 187 P7 1275 Dunn Ave 857 1,072.80 5,231.78 387.68 8.15 96098 77521 1 188 P7 Dunn Ave & Biscayne Blvd 589 754.20 5,323.79 605.89 4.25 104934 86513 1 189 P7 Highlands Library 395 528.60 153.76 70.91 0.99 7330 0 0 190 P7 Dunn Ave & Ray Greene Dr 360 479.60 246.75 48.55 3.36 15251 0 0 191 P7 Dunn Ave & Armsdale Rd 263 362.40 523.25 44.38 0.86 27554 6205 1 192 P7 Dunn Ave & E Pine Estates Rd 228 313.60 477.27 52.19 1.04 24911 6205 1 193 P7 2445 Dunn Ave 155 209.60 245.82 17.28 0.52 20752 0 0 194 P7 Dunn Ave & Irma Rd 104 137.40 125.33 8.78 0.07 9820 0 0 195 P7 Dunn Ave & Duval Dr 136 182.60 35.28 80.68 0.39 2046 0 0 196 P7 Dunn Ave & Lorence Ave 564 714.00 17.20 97.59 0.77 1485 0 0

PAGE 142

133 Appendix B Continued Table 30 C ontinued Sr. no. Route Name Stop N ame Total Employment Trips Generated by Employment Trips Generated using Parcel land use based Trip Rates Absolute Percent Differe nce Total Weekday Boarding Area of the Non Residential land use Area of Special Generator Special Generator Dummy 197 P7 Dunn Ave & Lucas St 469 582.40 32.27 94.46 1.18 2471 0 0 198 P7 Dunn Ave & Dobson Dr 23 32.20 0.00 100.00 0.10 0 0 0 199 P7 3737 D unn Ave 25 35.00 0.00 100.00 0.08 0 0 0 200 P7 Dunn Ave & N Wingate Rd 176 244.80 585.66 139.24 0.22 5722 5722 1 201 P7 Dunn Ave & N. Campus Blvd 176 244.80 585.66 139.24 4.93 5722 5722 1 202 P7 N. Campus Blvd & Key Adams Dr 176 244.80 585.66 139.24 0.06 5722 5722 1 203 P7 N. Campus Blvd. & Penny Camp Rd. 6 8.40 0.00 100.00 0.00 0 0 0 204 P7 N.Campus Blvd. & Capper Rd. 0 0.00 0.00 0.00 0.16 0 0 0 205 P7 F.C.C.J. Northside Campus 54 74.60 0.00 100.00 0.00 0 0 0 206 R5 F.C.C.J. Kent Campus 271 377.20 0.00 100.00 36.85 0 0 0 207 R5 Park St. & Glendale St. 236 327.80 0.00 100.00 1.94 0 0 0 208 R5 Park St. & Pinegrove Ave. 20 27.20 0.00 100.00 2.68 0 0 0 209 R5 Park St. & Van Wert Ave. 92 117.40 269.74 129.76 0.23 6275 6275 1 210 R5 Park St. & Ingleside Ave. 104 130.00 269.74 107.49 4.87 6275 6275 1 211 R5 Park St. & Talbot Ave. 138 170.80 569.79 233.60 0.04 39466 6275 1 212 R5 Park St. & Edgewood Ave. 119 144.60 300.04 107.50 0.14 33191 0 0 213 R5 Park St. & Valencia Rd. 58 69.60 300.04 331.09 0.00 33191 0 0 214 R5 Park St. & Seminole Rd. 55 76.80 226.20 194.53 0.88 16968 0 0 215 R5 Park St. & Aberdeen St. 28 38.60 0.00 100.00 0.00 0 0 0 216 R5 Park St. & McDuff Ave. 85 119.00 226.20 90.08 3.48 16968 0 0 217 R5 Park St. & Willow Branch Ave. 101 1 37.20 245.49 78.93 3.20 19081 0 0 218 R5 Park St. & Cherry St. 199 266.80 231.80 13.12 8.01 5395 1832 1 219 R5 Park St. & James St. 432 544.20 623.91 14.65 3.08 25798 7573 1 220 R5 Park St. & King St. 636 822.60 897.00 9.04 7.82 38446 7573 1 221 R5 Kin g St. & Oak St. 5,253 7,271.80 1,103.52 84.82 3.35 61940 7573 1 222 R5 Riverside Ave. & Barrs St. 5,358 7,442.40 2,438.84 67.23 16.90 172630 8521 1 223 R5 Riverside Ave. & Stockton St. 5,340 7,440.80 2,570.51 65.45 1.85 187041 2780 1 224 R5 Riverside Av e. & Osceola St. 5,379 7,493.60 3,094.89 58.70 0.89 223440 2780 1 225 R5 Riverside Ave. & Copeland St. 1,062 1,432.60 2,222.92 55.17 0.78 146483 2780 1 226 R5 Riverside Ave. & Goodwin St. 1,255 1,664.00 2,183.57 31.22 6.00 240851 0 0 227 R5 Riverside Av e. & Margaret St. 1,512 1,988.40 3,303.70 66.15 5.55 241595 0 0 228 R5 Riverside Ave. & Lomax St. 1,379 1,798.20 4,029.29 124.07 0.95 219050 23136 1 229 R5 Riverside Ave. & Post St. 1,682 2,224.80 6,966.01 213.11 1.43 182124 23136 1 230 R5 Riverside Ave & Riverside Park Pl 5,631 7,794.40 7,023.25 9.89 1.64 177938 23136 1 231 R5 Riverside Ave. & Roselle St. 5,538 7,667.20 5,421.03 29.30 0.81 145750 9739 1 232 R5 Riverside Ave. & Edison Ave. 5,526 7,685.00 1,649.66 78.53 3.78 98874 0 0

PAGE 143

134 Appendix B C ontinued Table 30 C ontinued Sr. no. Route Name Stop N ame Total Employment Trips Generated by Employment Trips Generated using Parcel land use based Trip Rates Absolute Percent Difference Total Weekday Boarding Area of the Non Residential land use Area of Special Generator Special Generator Dummy 233 R5 Riverside Ave. & Jackson St. 3,137 4,369.80 276.46 93.67 0.28 26941 0 0 234 R5 Riverside Ave. & Stonewall St. 2,940 4,101.60 264.42 93.55 1.18 21076 0 0 235 R5 Pearl St. & Bay St. 12,060 16,579.00 8,0 32.69 51.55 2.58 251492 0 0 236 R5 Forsyth St. & Julia St. 23,225 32,174.00 16,677.95 48.16 1.19 577605 4874 1 237 R5 Forsyth St. & Laura St. 23,559 32,324.00 40,412.44 25.02 3.20 1524033 11685 1 238 R5 Forsyth St. & Ocean St. 22,952 31,465.40 30,302.23 3.70 0.68 1211406 11735 1 239 R5 Newnan St. & Adams St. 8,380 11,291.80 22,189.02 96.51 4.63 913534 11735 1 240 R5 Newnan St. & Duval St. 7,978 10,756.40 29,130.87 170.82 0.78 1207881 73981 1 241 R5 Newnan St. & Ashley St. 5,878 7,836.80 29,862.56 281. 06 2.10 1096236 69057 1 242 R5 Newnan St. & Beaver St. 5,664 7,536.00 29,339.23 289.32 2.02 1069521 62246 1 243 R5 F.C.C.J. Station 14,486 20,229.00 30,397.42 50.27 143.15 1296758 52995 1 244 R5 Regency Square Hub 3,227 3,930.20 0.00 100.00 42.58 0 0 0 245 R5 9451 S Regency Square Blvd. 384 531.20 6,902.06 1,199.33 1.19 300806 186792 1 246 R5 9550 S. Regency Square Blvd. 1,874 2,483.60 2,623.56 5.64 1.06 114014 0 0 247 R5 S. Regency Square Blvd. & Monument Rd. 1,812 2,410.20 11.28 99.53 2.06 1248 0 0 248 R5 355 Monument Rd. 1,609 2,100.20 1,963.94 6.49 4.34 13413 10355 1 249 R5 445 Monument Rd. 855 1,033.40 1,963.94 90.05 2.60 13413 10355 1 250 R5 514 Monument Rd. 895 1,078.40 1,965.67 82.28 2.60 13605 10355 1 251 R5 544 Monument Rd. 42 54.60 1,965 .67 3,500.13 0.68 13605 10355 1 252 R5 989 Monument Rd. 281 349.40 0.00 100.00 0.57 0 0 0 253 R5 Monument Rd. & Treddick Pkwy. 104 131.00 0.00 100.00 3.20 0 0 0 254 R5 Monument Rd. & Lee Rd. 207 272.60 12.37 95.46 1.05 1368 0 0 255 R5 1431 Monument Rd. 118 149.60 12.37 91.73 1.56 1368 0 0 256 R5 1505 Monument Rd. 280 349.80 100.02 71.41 1.83 8635 0 0 257 R5 St. Johns Bluff Rd. & Monument Rd. 288 362.40 100.02 72.40 2.32 8635 0 0 258 R5 St. Johns Bluff Rd. & Causey Ln. 354 452.80 0.00 100.00 0.04 0 0 0 259 R5 St. Johns Bluff Rd. & S. Akers Dr. 146 197.60 285.73 44.60 0.13 112867 0 0 260 R5 St. John's Bluff Rd. & Lone Star Rd. 260 352.40 285.73 18.92 0.63 112867 0 0 261 R5 850 St. Johns Bluff Rd. 296 398.40 392.69 1.43 0.14 8388 4897 1 262 R5 St. Jo hns Bluff Rd. & Craig Industrial Dr. 96 129.60 210.51 62.43 0.04 4897 4897 1 263 R5 St. Johns Bluff Rd. & Airport Terrace Dr. 48 66.80 0.00 100.00 0.00 0 0 0 264 R5 St. Johns Bluff Rd. & Atlantic Blvd. 812 1,069.40 0.00 100.00 0.60 0 0 0 265 R5 St. John s Bluff Rd. & Theresa Dr. 865 1,134.00 137.38 87.89 3.82 25087 0 0 266 R5 St. Johns Bluff Rd. & Bradley Rd. 122 149.20 267.93 79.58 0.12 15428 0 0 267 R5 St. Johns Bluff Rd. & Lost Pine Dr. 299 405.20 156.79 61.31 1.29 10600 0 0 268 R5 St. Johns Bluff R d. & Fraser Rd. 350 471.40 0.00 100.00 0.58 0 0 0

PAGE 144

135 Appendix B Continued Table 30 C ontinued Sr. no. Route Name Stop N ame Total Employment Trips Generated by Employment Trips Generated using Parcel land use based Trip Rates Absolute Percent Difference Total Weekday Boarding Area of the Non Residential land use Area of Special Generator Special Generator Dummy 269 R5 St. Johns Bluff Rd. & Alden Rd. 210 276.80 0.00 100.00 0.81 0 0 0 270 R5 2656 St. Johns Bluff Rd. 244 335.80 27.63 91.77 3.15 7754 0 0 271 R 5 St. Johns Bluff Rd. & Judicial Dr. 415 567.80 75.16 86.76 0.00 21090 0 0 272 R5 St. Johns Bluff Rd. & Saints Rd. 525 712.40 132.47 81.41 0.33 37169 0 0 273 R5 St. Johns Bluff Rd. & Beach Blvd. 1,251 1,583.20 2,966.13 87.35 4.09 109166 1380 1 274 R5 Ce ntral Pkwy.& St. Johns Bluff Rd. 680 901.60 863.11 4.27 6.77 37682 0 0 275 R5 11655 Central Pkwy. 1,705 2,316.20 469.64 79.72 0.00 102634 0 0 276 R5 11710 Central Pkwy. 2,185 2,904.00 977.60 66.34 0.00 245164 0 0 277 R5 11818 Central Pkwy. 2,545 3,413.6 0 507.96 85.12 0.03 142530 0 0 278 R5 F.C.C.J. Southside Campus 3 4.20 57.71 1,274.05 8.23 6809 0 0 279 R5 Central Pkwy. & Beach Blvd. 899 1,208.20 0.00 100.00 0.38 0 0 0 280 R5 Beach Blvd. & Central Pkwy. 914 1,239.60 0.00 100.00 0.07 0 0 0 281 R5 120 00 Beach Blvd. 520 713.00 1,100.97 54.41 1.50 13806 13686 1 282 R5 Beach Blvd. & Sans Pareil St. 347 449.00 0.00 100.00 0.43 0 0 0 283 R5 3694 Kernan Blvd. 49 64.00 0.00 100.00 7.12 0 0 0 284 R5 Kernan Blvd. & Gehrig Dr. 10 14.00 0.00 100.00 0.04 0 0 0 285 R5 Kernan Blvd. & Mantle Dr. 4 4.20 0.00 100.00 0.45 0 0 0 286 R5 Kernan Blvd. & Hunter's Haven Ln. 6 7.00 0.00 100.00 0.11 0 0 0 287 R5 Kernan Blvd. & Blue Stream Dr. 12 15.40 68.50 344.81 0.09 7504 0 0 288 R5 Kernan Blvd. & First Coast Technology Pkwy. 44 57.40 68.50 19.34 0.14 7504 0 0 289 R5 UNF Dr. & Alumni Dr. 55 77.00 0.00 100.00 0.61 0 0 0 290 R5 U.N.F. Osprey Landing (U.N.F Dr.) 55 77.00 0.00 100.00 0.03 0 0 0 291 R5 U.N.F. Library (U.N.F. Dr.) 0 0.00 0.00 0.00 0.48 0 0 0 292 R5 U.N.F. Arena (U.N.F. Dr.) 0 0.00 0.00 0.00 1.28 0 0 0 293 R5 Town Center & Brightman Bl 90 109.60 0.00 100.00 0.05 0 0 0 294 R5 Town Crossing & Buckhead Branch 773 943.60 0.00 100.00 0.05 0 0 0 295 R5 Town Center Mall 1,735 2,109.00 0.00 100.00 0.20 0 0 0 296 U2 Regency Square Hub 2,934 3,568.00 0.00 100.00 89.82 0 0 0 297 U2 Arlington Expwy & Mill Creek Rd. 316 416.80 680.26 63.21 1.02 53020 0 0 298 U2 Arlington Expwy & Arlingtonwood Ave. 218 280.40 328.23 17.06 0.76 25137 0 0 299 U2 8109 Arlington Expwy. 529 723.60 51.45 92.89 0.00 3940 0 0 300 U2 Arlington Expwy & Townsend Blvd. 727 997.60 221.90 77.76 5.63 2168 2168 1 301 U2 7783 Arlington Expwy. 1,423 1,876.80 245.95 86.90 0.95 4350 2168 1 302 U2 Arlington Expwy & Alderman Rd. 1,040 1,351.40 136.80 89.88 0.44 14532 0 0 303 U2 7579 Arlington Expwy. 760 963.60 136.80 85.80 2.66 14532 0 0 304 U2 Arlington Expwy & Bert Rd. 239 307.60 1,242.53 303.94 2.98 22218 0 0

PAGE 145

136 Appendix B Continued Table 30 C ontinued Sr. no. Route Name Stop N ame Total Employment T rips Generated by Employment Trips Generated using Parcel land use based Trip Rates Absolute Percent Difference Total Weekday Boarding Area of the Non Residential land use Area of Special Generator Special Generator Dummy 305 U2 Arlington Expwy & Arling ton Rd. 399 524.80 1,126.46 114.65 7.20 9273 0 0 306 U2 6829 Arlington Expwy. 654 868.00 1,126.46 29.78 3.73 9273 0 0 307 U2 Arlington Expwy & Rogero Rd. 716 954.80 1,126.46 17.98 0.52 9273 0 0 308 U2 Arlington Expwy & Underhill Dr. 653 882.80 107.75 87 .79 2.34 18444 0 0 309 U2 Arlington Expwy & Cesery Blvd. 1,455 2,000.80 661.95 66.92 12.88 46517 2045 1 310 U2 Cesery Blvd. & Egret Point Ln. 1,239 1,710.80 521.59 69.51 3.16 20333 2045 1 311 U2 University Blvd & Saxony Woods Ln. 41 50.40 0.00 100.00 0. 46 0 0 0 312 U2 University Blvd. & Allen Pl. 205 265.80 647.24 143.51 0.16 18185 0 0 313 U2 University Blvd. & Atlantic Blvd. 728 945.40 947.71 0.24 22.00 36626 0 0 314 U2 University Blvd. & St. Cecilia Rd. 266 357.00 461.22 29.19 0.78 31337 0 0 315 U2 University Blvd. & Kellow Rd. 117 163.80 176.09 7.50 0.35 15458 0 0 316 U2 University Blvd. & Bartam Rd. 242 338.60 0.00 100.00 2.11 0 0 0 317 U2 University Blvd. & Coronet Ln. 1,438 1,893.20 3,368.82 77.94 4.93 41313 24135 1 318 U2 University Blvd. & Cruz Rd. 2,154 2,892.00 3,927.85 35.82 13.71 81020 63840 1 319 U2 University Blvd. & Booth Rd. 2,462 3,431.80 2,407.62 29.84 6.44 198637 39705 1 320 U2 University Blvd. & Harvin Rd. 2,522 3,516.40 3,844.79 9.34 4.46 230093 39705 1 321 U2 University Blvd & Kennerly Rd. 1,807 2,515.20 3,270.14 30.02 3.03 197755 0 0 322 U2 4140 University Blvd. 1,782 2,468.20 4,021.52 62.93 1.25 186811 2374 1 323 U2 University Blvd. & Bennett Rd. 793 1,065.40 3,000.60 181.64 2.41 95908 2374 1 324 U2 University Blvd. & B arnhill Dr. 861 1,150.40 4,183.23 263.63 4.09 96531 39512 1 325 U2 University Blvd. & Beney Rd. 902 1,161.00 2,979.25 156.61 2.44 84356 39512 1 326 U2 University Blvd & Mt. Carmel Terr 805 1,018.80 3,557.16 249.15 7.72 67332 37138 1 327 U2 University Bl vd. & Barnes Rd. 672 856.60 4,485.47 423.64 6.74 60180 3183 1 328 U2 University Blvd. & Spring Park Rd. 701 932.20 3,055.32 227.75 3.06 52650 8265 1 329 U2 University Blvd. & Cagle Rd. 1,045 1,367.80 1,253.63 8.35 1.83 121892 3430 1 330 U2 University Bl vd. & Richard St. 1,060 1,387.60 1,234.22 11.05 17.94 119555 3430 1 331 U2 University Blvd. & Philips Hwy. 1,138 1,464.00 805.01 45.01 2.06 59984 0 0 332 U2 University Blvd. & Powers Ave. 2,260 2,991.40 2,920.99 2.35 0.46 22623 2700 1 333 U2 University Blvd. & Chester Ave. 1,957 2,544.80 3,188.70 25.30 2.91 29679 6482 1 334 U2 6005 University Blvd. 2,112 2,757.00 3,030.23 9.91 6.11 39439 3782 1 335 U2 University Blvd. & St. Augustine Rd. 1,667 2,223.20 1,420.11 36.12 1.67 68117 16238 1 336 U2 Universi ty Blvd. & Minuteman Ln. 1,203 1,618.40 1,304.58 19.39 2.39 74113 12456 1 337 U2 University Blvd. & Graywood Rd. 520 705.00 866.79 22.95 1.24 49934 7026 1 338 U2 University Blvd. & Colgate Rd. 225 307.80 589.11 91.39 0.72 40276 0 0 339 U2 University Blv d. & Auburn Rd. 605 788.60 3,338.13 323.30 3.19 128462 67039 1 340 U2 San Jose Blvd. & Cornell Rd. 449 564.60 3,973.26 603.73 0.95 154487 67039 1

PAGE 146

137 Appendix B Continued Table 30 C ontinued Sr. no. Route Name Stop N ame Total Employment Trips Generated by Emp loyment Trips Generated using Parcel land use based Trip Rates Absolute Percent Difference Total Weekday Boarding Area of the Non Residential land use Area of Special Generator Special Generator Dummy 341 U2 San Jose Blvd. & Flanders Rd. 115 144.40 933. 32 546.34 0.21 45421 0 0 342 U2 San Jose Blvd. & Arcadia Dr. 97 123.80 822.02 563.99 1.95 40562 0 0 343 U2 San Jose Blvd. & San Amaro Dr. 60 78.40 119.61 52.56 0.43 5222 0 0 344 U2 San Jose Blvd. & E. Worth Dr. 210 271.40 200.58 26.09 0.80 5135 4493 1 345 U2 San Jose Blvd. & Gadsden Rd. 12 12.60 0.00 100.00 0.02 0 0 0 346 U2 San Jose Blvd. & Monterey St. 7 5.00 0.00 100.00 1.48 0 0 0 347 U2 San Jose Blvd. & Miramar Ave. 7 5.00 0.00 100.00 0.75 0 0 0 348 U2 San Jose Blvd. & Eutaw Pl. 11 10.60 0.00 100.00 0.05 0 0 0 349 U2 San Jose Blvd. & Morvenwood Rd. 13 13.60 0.00 100.00 0.11 0 0 0 350 U2 San Jose Blvd. & Mapleton Rd. 13 17.80 29.22 64.16 0.02 3201 0 0 351 U2 San Jose Blvd. & S Waterman Rd. 10 13.60 29.22 114.85 0.14 3201 0 0 352 U2 San Jose Blv d. & Saratoga Dr. 255 356.20 0.00 100.00 0.08 0 0 0 353 U2 San Jose Blvd. & Inwood Terrace 393 537.60 261.60 51.34 0.32 17916 0 0 354 U2 San Jose Bl & Oriental Gardens Rd 155 205.20 261.60 27.49 0.23 17916 0 0 355 U2 Hendricks Ave. & Lorimier Rd. 12 8.4 0 0.00 100.00 0.13 0 0 0 356 U2 Hendricks Ave. & Pineridge Rd. 15 12.20 0.00 100.00 0.03 0 0 0 357 U2 Hendricks Ave. & Marco Pl. 39 49.80 0.00 100.00 0.00 0 0 0 358 U2 San Marco Blvd. & Alford Pl. 1,010 1,297.00 6,732.15 419.06 0.01 119764 59975 1 359 U2 San Marco Blvd.& Balis Pl. 1,021 1,311.00 6,732.15 413.51 0.53 119764 59975 1 360 U2 San Marco Blvd. & Naldo Ave. 911 1,168.40 6,289.45 438.30 0.33 96930 58531 1 361 U2 San Marco Blvd. & Largo Rd. 553 709.80 6,509.19 817.05 0.09 97152 58531 1 362 U2 San Marco Blvd. & Landon Ave. 412 530.20 1,876.14 253.86 0.07 61526 19613 1 363 U2 San Marco Blvd. & LaSalle St. 477 615.00 1,500.29 143.95 0.48 59694 2224 1 364 U2 San Marco Blvd. & Phillips St. 878 1,159.80 1,034.45 10.81 0.72 61319 2224 1 365 U2 San Marco Blvd. & Nira St. 801 1,056.20 956.12 9.48 1.15 54977 2224 1 366 U2 Nira St. & Larue Ave. 833 1,104.60 922.39 16.50 0.47 52364 2224 1 367 U2 Nira St. & Flagler Ave. 953 1,286.60 1,257.68 2.25 0.03 47168 4885 1 368 U2 Nira St. & Hendricks Ave. 826 1,114.40 1,586.96 42.40 0.17 66399 4885 1 369 U2 Kings Ave. Station 3,884 5,346.20 5,944.75 11.20 0.10 177884 0 0


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 22 Ka 4500
controlfield tag 007 cr-bnu---uuuuu
008 s2010 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0004628
035
(OCoLC)
040
FHM
c FHM
049
FHMM
090
XX9999 (Online)
1 100
Rana, Tejsingh.
0 245
Enhancement of predictive capability of transit boardings estimation and simulation tool (tbest) using parcel data :
b an exploratory analysis
h [electronic resource] /
by Tejsingh Rana.
260
[Tampa, Fla] :
University of South Florida,
2010.
500
Title from PDF of title page.
Document formatted into pages; contains X pages.
502
Thesis (MSCE)--University of South Florida, 2010.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
538
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
3 520
ABSTRACT: TBEST is a comprehensive third generation transit demand forecasting model, developed by the FDOT Public Transit Office (PTO) to help transit agencies in completing their Transit Development Plans (TDPs). The on-going project funded by FDOT, related to TBEST, aims at further enhancing the capabilities of the TBEST model based on additional opportunities identified by the research team. The project focuses on enhancing TBEST's capabilities in following areas: 1) Improving the precision of socio- demographic data by using property appraisal data (parcel data) and, 2) Improving the quality of data regarding trip attraction. Based on the improvement areas, this study aims at performing an exploratory analysis to 1) Identify the differences in activity levels (population and employment) within transit stop buffers due to change in input data i.e. from aggregate census data to disaggregate parcel data. 2) Explore various strategies (development of employment based trip attraction and, parcel land use based trip attraction and exploring how special generators are dealt with in the past studies) to enhance the trip attraction capability of the TBEST model. The results obtained from this analysis provide insights on the strategies and helps define suggestions to further enhance the precision of TBEST model. The results show that use of parcel level data improves the accuracy in capturing the activity levels within the catchment area of each stop. The results also suggest use of parcel land use based trip attraction for stops with special generators or use of interaction variable (interaction between special generator dummy and size (square footage etc.) of the special generator) to enhance the trip attraction capability of the TBEST model.
590
Advisor: Abdul R. Pinjari, Ph.D.
653
Transit demand modeling
Trip rates
Special generator
Trip attraction
690
Dissertations, Academic
z USF
x Civil & Environmental Engineering
Masters.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.4628