Focus Model Calibration 1.0 - DRCOG



Focus Model Calibration 1.0Denver Regional Council of Governments (DRCOG) Activity-Based Travel Model8/5/2010Suzanne ChildressContents TOC \o "1-3" \h \z \u Table List PAGEREF _Toc269106843 \h 3Introduction PAGEREF _Toc269106844 \h 5GISDK Modifications PAGEREF _Toc269106845 \h 7Erik and Shahida--- PAGEREF _Toc269106846 \h 7Population Synthesizer PAGEREF _Toc269106847 \h 8Regular Work Location PAGEREF _Toc269106853 \h 16Regular School Location PAGEREF _Toc269106854 \h 22Auto Availability Choice PAGEREF _Toc269106855 \h 24Daily Activity Pattern Choice PAGEREF _Toc269106856 \h 25Exact Number of Tours PAGEREF _Toc269106857 \h 26Work Tour Destination Type PAGEREF _Toc269106858 \h 29Work-Based Subtour Generation PAGEREF _Toc269106859 \h 29Tour Primary Destination Choice PAGEREF _Toc269106860 \h 29Tour Main Mode Choice PAGEREF _Toc269106861 \h 35Tour Time of Day Choice PAGEREF _Toc269106862 \h 37Intermediate Stop Generation Choice PAGEREF _Toc269106863 \h 38Intermediate Stop Location Choice PAGEREF _Toc269106864 \h 39Trip Mode Choice PAGEREF _Toc269106865 \h 41Trip Departure Time Choice PAGEREF _Toc269106866 \h 42Highway Assignment PAGEREF _Toc269106867 \h 43Transit Assignment PAGEREF _Toc269106868 \h 47Conclusions and Further Calibration Directions PAGEREF _Toc269106869 \h 48Appendix 1: Basic Aggregate Model Results PAGEREF _Toc269106870 \h 50Appendix 2: Assignment Model Summary File PAGEREF _Toc269106871 \h 54Table List TOC \h \z \c "Table" Table 1. Focus Model Flow PAGEREF _Toc269106938 \h 5Table 2. Observed and Modeled Vehicle Miles Traveled on Links with Counts PAGEREF _Toc269106939 \h 6Table 3. Observed and Modeled Transit Boardings PAGEREF _Toc269106940 \h 6Table 4. 2005 Regional Controls on Number of Households PAGEREF _Toc269106941 \h 9Table 5. Regional Percentages- Households By Age of Householder PAGEREF _Toc269106942 \h 10Table 6. Regional Percentages- Households by Presence of Children PAGEREF _Toc269106943 \h 10Table 7. Regional Controls Input and Output PAGEREF _Toc269106944 \h 11Table 8. 2005 PopSyn/ACS Number of Households By County PAGEREF _Toc269106945 \h 12Table 9. Difference in Percents Households by Income Group: ACS and PopSyn PAGEREF _Toc269106946 \h 12Table 10. Households Size By County Difference in Percents PAGEREF _Toc269106947 \h 13Table 11. Income Group Controls by Percents PAGEREF _Toc269106948 \h 13Table 12. Household Size Controls and Outputs by Percents PAGEREF _Toc269106949 \h 13Table 13. 2005 Employed People By County- PopSyn vs ACS PAGEREF _Toc269106950 \h 14Table 14. 2005 State Demographer Forecast Persons by Age Cohort By County PAGEREF _Toc269106951 \h 15Table 15. 2005 PopSyn Outputs Persons By Age Cohort By County PAGEREF _Toc269106952 \h 15Table 16. Percent Difference State Demographer - PopSyn Cohort By County PAGEREF _Toc269106953 \h 15Table 17. Average Distance To Work by Person Characteristics PAGEREF _Toc269106954 \h 16Table 18. Average Work Skimmed Distance and Logsums By Home District PAGEREF _Toc269106955 \h 20Table 19. Modeled and Observed Percent of Workers Regularly Working at Home PAGEREF _Toc269106956 \h 20Table 20. CTPP Target and Model Results- Workers by Home to Regular Workplace PAGEREF _Toc269106957 \h 21Table 21. Average Distance to School by Student Grade Level PAGEREF _Toc269106958 \h 22Table 22. Percent of Students attending a school located within school district boundaries PAGEREF _Toc269106959 \h 23Table 23. Percent of Modeled 2005 Households by Auto Ownership By County PAGEREF _Toc269106960 \h 24Table 24. Percent of Observed 2005 ACS Households by Auto Ownership By County PAGEREF _Toc269106961 \h 24Table 25. Percent of Modeled Households by Income Group By Auto Ownership PAGEREF _Toc269106962 \h 24Table 26. Percent of Observed TBI 1997 Households By Income Group By Auto Ownership PAGEREF _Toc269106963 \h 25Table 27. Modeled versus Observed Percent of Persons Making Tours by Purpose PAGEREF _Toc269106964 \h 25Table 28. Modeled versus Observed Percent of Persons Making Stops By Purpose PAGEREF _Toc269106965 \h 26Table 29. 2005 Modeled Persons by Type by Average Number of Home-Based Tours by Purpose PAGEREF _Toc269106966 \h 27Table 30. 1997 TBI Persons by Type by Average Number of Home-based Tours by Purpose PAGEREF _Toc269106967 \h 27Table 31. Average Tours per Person by Home District Observed and Modeled PAGEREF _Toc269106968 \h 28Table 32. Modeled and Observed % of Work Tours to the Regular Workplace PAGEREF _Toc269106969 \h 29Table 33. Percent of Work Tours by Number of Subtours PAGEREF _Toc269106970 \h 29Table 34. Average Modeled and Observed Tour Straight Line Distance PAGEREF _Toc269106971 \h 30Table 35. Target SuperDistrict to SuperDistrict Non-Mandatory Tours PAGEREF _Toc269106972 \h 33Table 36. Modeled SuperDistrict to SuperDistrict Non-Mandatory Tours PAGEREF _Toc269106973 \h 33Table 37. Modeled Number of Tours Destined to District Compared with Number of People and Jobs PAGEREF _Toc269106974 \h 34Table 38. Target and Modeled Work Mode Share All Destinations PAGEREF _Toc269106975 \h 35Table 39. Target and Modeled Non-Work Mode Share All Destinations PAGEREF _Toc269106976 \h 36Table 40. Target and Modeled Work Mode Share CBD Destinations PAGEREF _Toc269106977 \h 36Table 41. Target and Modeled Non-Work Mode Share CBD Destinations PAGEREF _Toc269106978 \h 36Table 42. Average Modeled Stops by Stop Purpose and Person Type PAGEREF _Toc269106979 \h 38Table 43. Average Observed TBI Stops by Stop Type and Person Type PAGEREF _Toc269106980 \h 39Table 44. Average Modeled Trip Distance (for trips with intermediate stops) by trip purpose PAGEREF _Toc269106981 \h 39Table 45. Number of Modeled Intermediate Stops Compared to Number of People and Jobs By District PAGEREF _Toc269106982 \h 40Table 46. Work Trip Mode Choice Shares Targets and Modeled PAGEREF _Toc269106983 \h 41Table 47. Observed and Modeled 2005 VMT on links with counts PAGEREF _Toc269106984 \h 43Table 48. Total Vehicle Miles Traveled By Facility Type on Links with Counts PAGEREF _Toc269106985 \h 44Table 49. Total Vehicle Miles Traveled By Area Type on Links with Counts PAGEREF _Toc269106986 \h 45Table 50. Modeled and Observed Volumes By Screenline PAGEREF _Toc269106987 \h 47Table 51. Modeled Transit Trips and Boardings PAGEREF _Toc269106988 \h 47Table 52. Average RTD Observed and Modeled Boardings By Sub-Mode PAGEREF _Toc269106989 \h 48IntroductionThe Focus activity-based travel model was recently developed and calibrated by Denver Regional Council of Governments and Cambridge Systematics. As activity-based travel model calibration is a new relatively new frontier, the calibration required several innovations for data comparison and calibration modification of variables, coefficients, and constants.Each Focus model component was calibrated individually and then the entire model was calibrated aggregately against roadway counts and RTD transit boardings. REF _Ref268868841 \h Table 1 shows the Focus model component flow. The network skimming, assignment and airport models were adapted from the earlier trip-based TransCAD model. All location, mode choice, and time of day model travel components were re-estimated for Focus using the 1997 Travel Behavior Inventory (TBI) survey of regional households. Table SEQ Table \* ARABIC 1. Focus Model FlowPopulation Synthesizer14. Work-Based Subtour GenerationTransCAD Initialization15. Tour Time of Day SimulationTransCAD Trip Generation16. Tour Primary Destination ChoiceTransCAD Skimming17. Tour Priority AssignmentPopulation Synthesizer18. Tour Main Mode ChoiceSize Sum Variable Calculator19. Tour Time of Day ChoiceRegular Workplace Location20. Intermediate Stop Generation ChoiceRegular School Location21. Trip Time of Day SimulationAuto Availability22. Intermediate Stop Location ChoiceAggregate Destination Choice Logsum Generation23.Trip Mode ChoiceDaily Activity Pattern24. Trip Time of Day Exact Number of Tours25. Write Trips To TransCADWork Tour Destination Type26. TransCAD Highway and Transit AssignmentIn this calibration, the model was calibrated up to 2005 data from the 1997 estimation. Data from 2005 was used wherever possible to ensure that the model was correctly capturing observed 2005 Denver travel behavior. The following 2005 datasets were used to calibrate against: 2005 American Community Survey (ACS), 2005 Colorado state demographer data, 2005 Colorado Department of Transportation (CDOT) highway counts, 2005 HPMS estimated regional VMT, 2005 Regional Transportation District (RTD) transit boardings and 2005 Compass trip-based model results. Unfortunately, because activity-based travel models are relatively new, large datasets were not available for some of the detailed travel behavior output by the model and many individual components had to be calibrated against the 1997 weighted expanded TBI survey data. The TBI and other older datasets like the 2000 Census were used in combination with growth factors to account for regional growth to 2005. One other dataset that was used for calibration in combination with growth factors was the 2000 Census Transportation Planning (CTPP) journey-to-work data. A new 2010 regional travel survey, Front Range Travel Counts and the 2010 Census should allow for a refreshed Focus calibration during 2011.Once comparisons were made of model results against the observed datasets, each model component was calibrated. The calibration involved changing utility function constants, coefficients, and adding variables. Then the model was re-run, results compared again, and modifications made again. This process was iterated as time allowed until satisfactory results were achieved.The major regional level model results of the calibration are shown in REF _Ref268870616 \h Table 2 and REF _Ref268870618 \h Table 3. These tables demonstrate that the aggregate model results match the observed counts and transit boardings well.Table SEQ Table \* ARABIC 2. Observed and Modeled Vehicle Miles Traveled on Links with CountsObserved VMTModeled VMT20,506,76820,906,583Table SEQ Table \* ARABIC 3. Observed and Modeled Transit BoardingsObserved Transit BoardingsModeledTransit Boardings269,741263,508The remainder of document first details how each individual model was calibrated. It finishes with the aggregate highway and transit assignment results, as well appendices showing overall model summaries.GISDK ModificationsThe first model component that was changed for calibration was the GISDK code based used for network skimming and assignment.Erik and Shahida---base code set came from Compass 2.0 / matching code used from estimation.changes made over Compass:use of trips passed from SQL Server focus, combined with old Compass DIA, I-E, E-E and commercial trips;auto operating cost based on VOC10 ten time period assignment changes4 transit periodsmany additional variables skimmed for focus model- ex generalized time; piecewise linear skim variablesValue of Time by area type changed to be more consistent with focus mode choice modelsPopulation SynthesizerAfter network skimming is completed, the Population Synthesizer (PopSyn) creates a recordset of individual regional households and persons. This section presents the results of validation tests performed on PopSyn developed for Focus. The Population Synthesizer creates a forecast of individual households and persons for chosen year. It operates by drawing household and person records from the 2000 Public Use Microsample (PUMS) with the goal of matching forecasted demographic controls. The 2005 synthesized population is validated by observing how well the synthesized data matches both the inputs and independent data sources. The inputs are also examined to verify their accuracy, and uncontrolled variables are compared to the American Community Survey (ACS) estimates and state demographer data. The controls are being adequately maintained in the model run. Overall the validation shows that PopSyn’s results are acceptable enough for use in the FOCUS.However, a few discrepancies between PopSyn’s data and external data sources were identified in this validation, as shown in the bullets below.The 2005 ACS estimated 3% fewer households in the 6 major counties than was given in the 2005 land use forecast given as input to PopSyn. The 2005 ACS estimated 5% fewer households in Denver County than PopSyn.Some of the variables that were uncontrolled like workers and person by age came out of PopSyn different from ACS and the state demographer data:PopSyn produced 6% more workers in the six counties than the ACS estimated. PopSyn produced 10% more persons age 25-44 and 9% fewer persons age 15-24 in the six counties than the state demographer’s forecast.Improvements to PopSyn require refinement to its inputs, which are controls on regional and zonal data. A better economic forecast will result in better regional demographics output by PopSyn. The land use model is used to create zonal controls; greater detail and accuracy with the land use model will improve PopSyn’s ability to project the types of households in each zone in future years. Better ability to forecast shifts in income groups by zone (gentrification, change in real income) would improve PopSyn ability to forecast households in the correct income groups in future years.The recommendations for PopSyn’s improvement are the following:Recommendation 1. Obtain a new economic forecast with greater detail and accuracy, including a regional forecast for households by number of adults, number of workers, number of children, and age of householder.Recommendation 2. Upgrade the land use model to make changes in income group distributions over time by zone. Develop the land use model’s ability to forecast more demographic characteristics (i.e. number of households by number of workers/number of children). Improvements to the land use model will allow PopSyn to output the types of households in each zone more accurately.2005 PopSynControl VariablesThe first validation tests described in this document concern the controlled household characteristics. The Population Synthesizer uses two sets of controlled variables for household characteristics: regional-level controls and zonal-level controls.? For the 2005 PopSyn run, the regional controls come from the 2035 economic forecast . The zonal controls are based on the land use model and 2000 Census data.? Two separate questions are posed about both the regional and zonal controls:First, are the controls valid?? Do they match other data sources?? Secondly, assuming the controls are valid, do the output control totals match the input control totals?PopSyn Regional ControlsThe regional controls provide targets for the Population Synthesizer to attempt to match demographics for the travel model region.? These controls were created based on the 2035 economic forecast. Because of some discrepancies between the shares of households by age of householder and households by type between the economic forecast for 2005 and the 2005 ACS, we chose to scale the economic forecast to the 2005 ACS shares for these categories. Also, the 2005 Economic Forecast was scaled to match the 2005 land use model for total regional households. The scaled regional controls are for total households by household type and age of householder as follows in REF _Ref268259729 \h Table 4.Table SEQ Table \* ARABIC 4. 2005 Regional Controls on Number of HouseholdsHousehold TypeAgeof HouseholderNUMBER OF HOUSEHOLDS FORECASTED AND SCALED TO LAND USE MODEL AND ACSYEAR2005201020152020202520302035One AdultNo Kids18-44127,464126,304129,259137,321150,649164,723175,435One AdultNo Kids45-6491,024103,164108,649115,192114,561116,427121,503One AdultNo Kids65 & over68,41880,059102,289132,151164,172194,074214,240One AdultWith Kids18-4448,01548,33850,31254,38260,59867,12972,354One AdultWith Kids45-6411,95014,54416,15818,02918,81519,87421,648One AdultWith Kids65 & over2,1483,1344,6176,7499,31811,94714,340Two or More AdultsNo Kids18-44136,051136,853140,596150,177165,863181,653194,438Two or MoreAdultsNo Kids45-64194,971221,177233,632248,607247,845252,207262,493Two or More AdultsNo Kids65 & over80,39997,097127,451167,025208,702246,262271,536Two or More AdultsWith Kids18-44224,469223,099230,728246,932272,744302,626326,056Two or More AdultsWith Kids45-6483,29295,479101,751109,268109,869112,571117,921Two or More AdultswithKids65 & over4,1775,1776,9189,25511,82714,25416,079 REF _Ref268259760 \h Table 5 compares the percentages of the households with householders in three age categories: 18-44, 45-64, and 65+ for the 2005 base year and for future years.? Note that the 2005 controls and the 2005 ACS are nearly identical because of the controls were scaled to match the ACS.Table SEQ Table \* ARABIC 5. Regional Percentages- Households By Age of Householder% of Households by Age of HouseholderForecast Year or Estimate Sources18-4445-6465+2005 ACS49%36%15%2005 Economic Forecast Control50%36%14%2010 Economic Forecast Control46%38%16%2015 Economic Forecast Control44%37%19%2020 Economic Forecast Control42%35%23%2025 Economic Forecast Control42%32%26%2030 Economic Forecast Control42%30%28%2035 Economic Forecast Control42%29%29% REF _Ref268259810 \h Table 6 shows the percentage of households with and without children in the PopSyn controls and from the ACS for the 2005 base year and future years.? The economic forecast controls in 2005 were scaled to the shares of households by presence of children from the ACS.? Note that share of households with and without children are nearly identical in 2005 for the ACS and the economic forecast because the controls were scaled to match the ACS.Table SEQ Table \* ARABIC 6. Regional Percentages- Households by Presence of ChildrenPercent of Households withChildrenPercent of HouseholdsWithout Children2005 ACS34%66%2005 Economic Forecast35%65%2010 Economic Forecast 34%66%2015 Economic Forecast33%67%2020 Economic Forecast32%68%2025 Economic Forecast32%68%2030 Economic Forecast31%69%2035 Economic Forecast31%69%The second question that needs to be answered is whether the regional controls are being adequately maintained during PopSyn’s operation.? REF _Ref268259867 \h Table 7 below compares the regional control totals input to the PopSyn’s outputs.? All the input controls match the output households with less than a 0.4% difference with several outputs substantially below this difference.? Therefore, PopSyn is effectively maintaining the regional controls.Table SEQ Table \* ARABIC 7. Regional Controls Input and OutputHousehold TypeAge of Householder% Difference Input and OutputOne Adult NoKids18-44-0.002%One Adult No Kids45-64-0.001%One Adult No Kids65 & over-0.012%One Adult with Kids18-440.042%One Adult with Kids45-640.203%One Adult with Kids65 & over0.383%Two or More Adults No Kids18-44-0.007%Two or More Adults No Kids45-64-0.006%Two or More Adults No Kids65 & over-0.010%Two or More Adults with Kids18-44-0.012%Two or More Adults with Kids45-64-0.011%Two or More Adults with Kids65 & over0.312%Zonal ControlsThe zonal controls provide targets for the Population Synthesizer to match on characteristics of synthesized households on a zonal level. PopSyn uses the 2812 zone system developed for the FOCUS project. The zonal controls are results of the land use model and 2000 Census data. An improvement to the controls would have to result from an improvement in the land use model. The controls for each zone are targets for the following nine statistics:ZONAL CONTROLS ( for each of the 2812 zones)(1) Total Households in the Zone(2) Percentage of Households in Zone with Income 0-30K(3) Percentage of Households in Zone with Income 30-60K(4) Percentage of Households in Zone with Income 60-100K(5) Percentage of Households in Zone 100K+(6) Percentage of Households of Size 1(7) Percentage of Households of Size 2(8) Percentage of Households of Size 3(9) Percentage of Households of Size 4+As with the regional controls, the first question we ask is:? Are the controls themselves reasonable? REF _Ref268259938 \h Table 8 compares the number of households from the 2005 land use model to the 2005 ACS. REF _Ref268259938 \h Table 8 shows that the 2005 land use model estimated 3% more households in the six counties overall than the 2005 ACS estimated. The land use modeling team asserted that differences in estimation methods between the land use model and ACS, as well as sampling error accounted for the differences seen in REF _Ref268259938 \h Table 8. Table SEQ Table \* ARABIC 8. 2005 PopSyn/ACS Number of Households By County?County2005 Land Use2005 ACSDifference%DifferenceACS Margin of ErrorAdams1452561413833873.53%+/- 1.8%Arapahoe2112242062504974.52%+/- 1.4%Boulder1181781134054773.34%+/- 1.7%Denver25376424157912185.25%+/- 1.4%Douglas8780787654153.30%+/- 1.1%Jefferson2133222113941928.01%+/- 1.0%Total1029553100166527887.83%? REF _Ref268260013 \h Table 9 shows the difference between the ACS and PopSyn income group distributions. The 2005 PopSyn estimates of percentages of households by income group came directly from the 2000 Census. The 2005 ACS estimates, in comparison, include inflationary effects from 2000 to 2005. As a result, the percent of households in the highest group of $100,000+ had more households in it in the ACS than in PopSyn. Furthermore, the $30-60K group had 4% fewer households in the ACS than in PopSyn. Improvements in the ability to forecast the number of households by income group changes over time would be required to cause PopSyn to more accurately predict income distributions. Table SEQ Table \* ARABIC 9. Difference in Percents Households by Income Group: ACS and PopSynDIFFERENCE IN PERCENTS: ACS %- POPSYN%INCOME GROUPAdams ArapahoeBoulder Denver Douglas Jefferson Total(in 1000s of dollars)0-30-2%0%1%-2%2%2%-1%30-60-5%-2%-3%-3%-3%-5%-4%60-1003%0%-4%0%-4%-1%0%100+4%2%6%4%5%4%4% REF _Ref268260060 \h Table 10 compares ACS and PopSyn percentage of households by household size. These differences look small and unproblematic.Table SEQ Table \* ARABIC 10. Households Size By County Difference in PercentsDIFFERENCE IN PERCENTS: ACS %- POPSYN%Household SizeAdams ArapahoeBoulder Denver Douglas Jefferson Total1-1%-3%-1%-1%-4%-2%-1%21%1%0%-2%0%-2%0%3-1%1%-2%1%1%0%0%40%0%3%2%4%4%2%The second question about the zonal controls that needs to be answered is if they are being maintained after PopSyn’s operation. REF _Ref268260119 \h Table 11 and REF _Ref268260127 \h Table 12 below reveal that the controls are being satisfactorily maintained. Finally, the number of households by zone is quite accurate as well with the majority of zones have the same number of households input and output.Table SEQ Table \* ARABIC 11. Income Group Controls by PercentsINCOME GROUPS ($)Controls (2000 Census)Outputs (2005 PopSyn Outputs)Difference0-30K25.6%24.8%0.9%30-60K32.1%31.6%0.5%60-100K25.2%25.5%-0.3%100K+17.1%18.1%-1.0%Table SEQ Table \* ARABIC 12. Household Size Controls and Outputs by PercentsHousehold SizeControls (2000 Census)Outputs (2005 PopSyn Outputs)Difference127.7%27.7%0.0%233.0%32.8%0.2%315.8%16.0%-0.2%4+23.5%23.5%0.0%PopSyn Uncontrolled VariablesThis section compares the uncontrolled variables output by PopSyn to ACS and state demographer estimates of these variables. As would be expected, the uncontrolled variables are farther from ACS estimates than are the controlled variables.The number of employed people and age are compared between PopSyn and ACS and state demographer data. Overall, the Population Synthesizer produced 6% more employed persons than in the ACS, as shown in REF _Ref268260202 \h \* MERGEFORMAT Table 13. It is possible that this difference was caused by the economic downturn that occurred after 2001, creating an overall loss of jobs in the region.? PopSyn’s forecast was based on the 2000 PUMS, prior to the economic downturn, which may have caused higher numbers of employed people to be output by PopSyn than the 2005 ACS sample. PopSyn estimates would improve if number of workers was added as a control.This discrepancy of employed people between PopSyn and ACS in 2005 is highest in the counties of Boulder (11% more in PopSyn) and Jefferson (8% more in PopSyn).? This may be a result of the employed people in these counties losing a larger portion of jobs during the economic turndown.Table SEQ Table \* ARABIC 13. 2005 Employed People By County- PopSyn vs ACSEmployed People by County??CountyPopSyn Output2005 ACSDifference%Adams 202,460196,1756,2853%Arapahoe291,102273,61417,4886%Boulder162,174146,10116,07311%Denver292,116277,42614,6905%Douglas136,101134,1811,9201%Jefferson297,055275,54921,5068%Total1,381,0081,303,04677,9626%PopSyn’s output for age by county was quite different from the state demographer estimates for age by county.? While the age of the householder is used as a control in PopSyn, the age of every individual is not.? REF _Ref268260240 \h Table 14 shows the state demographer age cohort estimates by county for 2005. REF _Ref268260275 \h Table 15, in comparison, shows the PopSyn outputs of age cohort estimates by county for 2005. Finally, REF _Ref268260323 \h Table 16 shows the percent difference between the state demographer and PopSyn estimates of age cohorts by county. Note that PopSyn produced 10% more persons age 25-44 and 9% fewer households age 15-24. These age groups are lumped together in the economic forecast, and were not controlled which caused the discrepancy in PopSyn to the economic forecast.Table SEQ Table \* ARABIC 14. 2005 State Demographer Forecast Persons by Age Cohort By CountyCounty?Age CohortAdamsArapahoeBoulderDenverDouglasJeffersonTotal0-1499,127112,09851,920121,80464,57999,295548,82315-2458,17976,05250,71570,86233,83877,214366,86025-44127,919154,88384,937194,35880,054140,852783,00345-6487,762140,67176,574129,42558,935157,534650,90165+28,34449,39323,95960,48411,68857,522231,390Total401,331533,097288,105576,933249,094532,4172,580,977Table SEQ Table \* ARABIC 15. 2005 PopSyn Outputs Persons By Age Cohort By County?CountyAge CohortAdamsArapahoeBoulderDenverDouglasJeffersonTotal0-1497,330119,57256,324112,22066,242114,668566,35615-2457,23969,81145,23576,19320,95663,661333,09525-44132,822174,11296,348202,91291,045166,851864,09045-6482,933129,70466,493117,70158,029140,740595,60065+33,09741,39422,06865,53710,36150,398222,855Total403,421534,593286,468574,563246,633536,3182,581,996Table SEQ Table \* ARABIC 16. Percent Difference State Demographer - PopSyn Cohort By CountyCountyAge CohortAdamsArapahoeBoulderDenverDouglasJeffersonTotal0-142%-7%-8%8%-3%-15%-3%15-242%8%11%-8%38%18%9%25-44-4%-12%-13%-4%-14%-18%-10%45-646%8%13%9%2%11%8%65+-17%16%8%-8%11%12%4%Total-1%0%1%0%1%-1%0%PopSyn ConclusionsThe most problematic statistics of PopSyn’s outputs involved age and number of workers, and especially at the county level. A few ways to improve PopSyn’s forecasted population would be to obtain a more detailed, accurate regional economic forecast and to enhance the land use model to forecast changes in income group by zone and any other demographic characteristics. For the most part, PopSyn’s output demographic characteristics are within a 5-10% range of the ACS estimates. The PopSyn synthesized population should be adequate to be used in all later model components to predict regional travel.Regular Work LocationThe population synthesizer specifies whether each created person is a worker or not. All workers are then run through the regular workplace location model to predict the locations of their regular workplaces based on their personal characteristics and home locations. The correct prediction the work location model is essential for the model overall to produce good results because a large amount of congested travel and transit occurs on the journey to work. The most important measures to capture are the distance to work and the distribution of home and work locations at a district level.Changes were made to the work location model to better match the observed distribution data. These changes included the additional of district-level dummy variables that either made certain districts more or less attractive by the application of different levels of positive or negative coefficients. For example, if we observed that the model was producing too few workplace choices in the CBD, we would add a term to the utility function of say 0.5* CBD District, where is a variable that CBD District =1 if a zone is in the CBD district and 0 otherwise. Over 30 calibration runs were conducted to make the observed and modeled distributions from home to work match better. REF _Ref268263780 \h Table 17 shows the average straight line distance from home to the regular workplace by person characteristics. Overall the model is matching the TBI average distance fairly well, with the average modeled distance to work being 8.3 miles and the average observed distance to work being 8.1 miles. The model performs less well in capturing specific subgroup behavior, such as the degree to which women and low income households travel shorter distances to work than average workers.Table SEQ Table \* ARABIC 17. Average Distance To Work by Person CharacteristicsPersonSubgroup2005 Focus Model1997 TBIDifferenceWomen7.77.10.6Men8.88.9-0.1Full Time Worker9.08.60.4Part Time Worker5.95.50.4Household Income<35 K6.86.40.4Household Income 35-100K8.48.40.0Household Income >100K9.19.2-0.1Age Under 205.14.90.2Age Over 208.58.20.2OVERALL8.38.10.2Figure 1 displays a histogram of the number of workers with a given distance to work in the TBI and the model. Overall the modeled distance distribution matches the TBI well.Figure SEQ Figure \* ARABIC 1. Distance Home to Work Observed vs. Modeled Histogram REF _Ref268264260 \h Figure 2 shows the districts which represent major regional travel sheds. This district structure will be used to aggregate regional travel model results so that trends can be determined. For example, REF _Ref268264371 \h Table 18 shows by these districts how the average modeled distance to work and work logsums vary. Note that the relative accessibility of each district, as represented in the logsums, impacts how far people are willing to travel to work.Figure SEQ Figure \* ARABIC 2. Districts MapTable SEQ Table \* ARABIC 18. Average Work Skimmed Distance and Logsums By Home DistrictHome DistrictAverage One Way Skimmed Distance to Work (Miles)Average LogsumBoulder Valley7-0.7CBD5-0.7DIA14-1.2East10-1.0East SE10-1.0Inner East6-0.8Inner North10-1.0Inner SE LRT7-0.9Inner SW LRT7-0.8Longmont10-0.8North13-1.1Outer NW11-1.0Outer SE LRT 10-1.0Outer SW LRT11-1.0Rural21-1.4Southeast13-1.2Southwest12-1.1W Northwest10-1.0West9-0.9 REF _Ref269107262 \h Table 19 shows the percent of workers whose regular workplace is at home as observed in the TBI at 8% and resulting from the model at 7%.Table SEQ Table \* ARABIC 19. Modeled and Observed Percent of Workers Regularly Working at HomeFocus Model Run Percent Work At Home7%TBI Percent Work At Home8%To find targets for travel from one home district to another, data from the 2000 Census Transportation Planning Package (CTPP) was used. Because large regional growth occurred between 2000 and 2005, the CTPP weighted expanded observed work trips were factored using a fratar method. The row sums were fratared to match the workers by home district, as output by the Population Synthesizer. The column sums were fratared to match the number of jobs by district, as output by the land use model. The resultant target work trips from home district to work district are shown in REF _Ref268264792 \h Table 20. This table also shows the modeled home district-regular work district pairs.The Denver International Airport (DIA) district has been omitted because all trips associated with DIA come from the Compass model.Table 20. CTPP Target and Model Results- Workers by Home to Regular Workplace?CBDInner SW LRTInner SE LRTOuter SW LRTOuter SE LRTInner EastEast SESoutheastEastInner NorthW NorthwestWestSouthwestNorthOuter NWBoulder ValleyLongmontRural??CTPP Target123456789101112131516171819Grand TotalCBD4,1817149373399271,0781877054631226461018319956213735610,944Inner SW LRT3,5165,5172,4672,8832,4372,3326241522,2721,3291,5593,2309442534001895958730,751Inner SE LRT13,8805,50219,9644,91918,7759,8365,7841,5784,4522,2132,0134,1881,3655581,2347882491,42298,720Outer SW LRT5,3213,2674,88717,61019,7962,5592,1781,1492,0181,7401,8323,9564,6203856521971552,30874,630Outer SE LRT9,4133,12610,13112,11455,1685,4585,4903,5383,1971,5201,6273,1803,2385031,0163351652,926122,144Inner East13,6503,5577,0291,7126,83515,2092,6975274,7682,4621,4833,11360973894356715965766,714East SE7,2453,86310,0185,13020,7557,88031,3234,18012,6875,2502,0942,8737381,1379632992702,620119,324Southeast4,5961,8405,8074,64326,6573,64613,68712,5304,5701,8331,0681,659901514760233863,04488,074East3,9952,1582,6591,1162,8693,8524,0364409,2042,7231,4811,25829257450117213668138,147Inner North5,1213,2332,4422,0532,9942,7652,2143918,00926,6197,8665,0709836,6357,1122,5479252,57289,551W Northwest7,5003,8222,7401,8553,5033,7801,4302975,3648,48226,19715,9012,0932,1987,5612,3637222,16797,975West11,5067,5225,1395,3447,8545,3312,1965885,3755,04312,28437,3536,7191,4062,5439803733,257120,813Southwest6,0423,7674,3568,71913,3182,7541,5335052,0232,2523,99711,97018,1615081,1803341492,22883,796DIA0000000000000000000North5,9181,8912,0451,3552,8712,5961,9082196,09918,1414,8353,86549017,28212,8244,3422,3083,82192,812Outer NW3,6959721,3306051,7901,6687311372,3654,9254,7203,4114033,35725,65717,6505,0671,36679,848Boulder Valley1,910245514237844885253453156846261,1271427267,19538,9544,9861,10860,794Longmont72322116917652323719838302879241396269795,18511,32126,5132,64250,771Rural5,0131,9652,9574,45810,9542,9473,7732,7094,6946,2494,3577,6622,5583,1234,0866,0514,47342,180120,210?Grand Total113,22553,18085,59175,270198,86974,81580,24429,09278,26092,65578,545110,82144,46541,07579,86887,53646,86775,6421,446,018Work Location Model?CBDInner SW LRTInner SE LRTOuter SW LRTOuter SE LRTInner EastEast SESoutheastEastInner NorthW NorthwestWestSouthwestNorthOuter NWBoulder ValleyLongmontRuralGrand TotalCBD407072567729554911992783968357348979111611715747125710926Inner SW LRT46316009286222392413226873413213351303142034447712603751453024430700Inner SE LRT13211564423504485415507890972981190449828522097474215255387222815866198555Outer SW LRT5068399655021970918587296718297981612156416724340455833446319049110374505Outer SE LRT8159374310137108865563850566089410333522060158337363072399536222532623121940Inner East12059366363561779525015428369942252563386228336546146439073047842766603East SE7101326411890303118464832635081421211148453522052949102082710723461141321119125Southeast3069195765833430283583424121641419140661766971186799139646815439290387927East404310652125522273629184387446101393012150614121895766732475549238083Inner North669229022885146326494177232023378652588979955331626583873672098608145289401W Northwest897333992792175527483973161915152691031026638130331428231782022898707115597811West1435482276443540768686529204034945615741108763733346511190292711602611372120611Southwest521742845066101491203529141668560169819803479106312044739996552789140683656DIA11513720154464823106591110284North616221992122117219383129191020156211387151893458430213141253245381809842497475Outer NW309511089905859891590853862424632253563393411427826692164503639123179799Boulder Valley12692462111072153661602348711201244844128852563843772343056660725Longmont4631511757516120010716366888679365519194267885130876196950692Rural45722498391543561566629174071321943664887467671882977238836075793277038625120528Grand Total1122195508594248718211907917630586351303777479492082803681085174400543590775798802444678660411449346Regular School LocationThe next component that is run after work location is school location. This model predicts for all students in the region where they will attend school. The model is segmented by four grade levels: pre-school, K-8, High School, and University to reflect that school selection varies by grade level.During school location calibration, increases and decreases on the distance utility function coefficients variables to make the modeled average better match the observed distances. REF _Ref268595530 \h Table 21 compares the average modeled and observed TBI distance to school by the four grade levels. All average distances are within 0.2 miles when comparing modeled to observed values. REF _Ref268595615 \h Figure 3 shows histograms of modeled and observed distances by student grade level for pre-school, K-8, and high school. One issue that can be seen is that the percent of K-8 students observed traveling less than 1 mile to school is about 60% in the observed data, but only around 45% in the model.Table SEQ Table \* ARABIC 21. Average Distance to School by Student Grade Level?Modeled2005TBI 1997Pre-School Student2.32.5K-8 Student1.81.6High School Student2.62.4University Student6.87.0Figure SEQ Figure \* ARABIC 3. School Location Distance To School: Modeled and ObservedFigure 4 shows a histogram of modeled and observed distances for university students. Note that the TBI has a different, almost linear distribution of distances to school increasing for increasing distances. In comparison, the modeled distribution is neither linear, exponential or normally distributed, but shows peaks at both 1-3 miles and 5-10 miles.Figure SEQ Figure \* ARABIC 4. Distance to School Histogram for University StudentsTable 22 compares the modeled and observed percent of students that attend a school within their school district boundaries. The school district they are assigned corresponds to their home zone school district. These percents do not refer to whether or not they go to a public school in the district, only if the school is located in the district. So they could attend a private school within the public school district boundaries and be included in the shares. The model includes variables for whether a school is in the student’s district values, and as a result the model matches the observed results fairly closely.Table SEQ Table \* ARABIC 22. Percent of Students attending a school located within school district boundariesStudent Grade LevelPercent in District: ModeledPercent in District: ObservedPre-School81%74%K-893%92%High School75%88%Auto Availability ChoiceAfter work and school locations are known, the model selects the number of autos available for each regional household. The available choices for each household are 0 autos, 1 auto, 2 autos, 3 autos, and 4+ autos.During calibration, the alternative specific constant for the 0 auto alternative was increased because the model was producing too few 0 auto households as compared to the 2005 American Community Survey (ACS). Also the coefficient on the transit accessibility was increased for the zero car alternative. The ACS showed that Denver County area had higher numbers of zero car households than was initially seen in the modeled data possibility because the transit accessibility in Denver allowing for greater numbers of 0 car households that had not been captured in the original eseimation.Table 23 and REF _Ref268684081 \h Table 24 shows the modeled and observed percent of households owning a given number of vehicles by the county in which they live. Table SEQ Table \* ARABIC 23. Percent of Modeled 2005 Households by Auto Ownership By County?% of Households in County?Vehicle Ownership LevelAdamsArapahoeBoulderDenverDouglasJeffersonTotal 6 CountiesNo Vehicle6%5%6%10%2%4%6%1 vehicle available28%30%29%38%18%29%27%2 vehicles available46%46%46%37%58%48%41%3 or more vehicles available:20%19%19%14%22%20%26%Table SEQ Table \* ARABIC 24. Percent of Observed 2005 ACS Households by Auto Ownership By County?% of Households in County?Vehicle Ownership LevelAdamsArapahoeBoulderDenverDouglasJeffersonTotal 6 CountiesNo Vehicle4%5%4%12%1%4%6%1 vehicle available32%34%29%43%20%29%33%2 vehicles available41%41%46%33%55%42%41%3 or more vehicles available:23%20%21%12%24%24%20%Table 25 and Table 26 show the percent of households by income group owning a given number of household autos in the model and the 1997 TBI data.Table SEQ Table \* ARABIC 25. Percent of Modeled Households by Income Group By Auto Ownership?Number of Household AutosIncome Group01234+<1534.5%49.4%12.4%3.3%0.4%15-309.2%51.5%29.1%7.5%2.7%30-751.7%29.9%45.1%16.5%6.8%75-1000.3%10.2%51.0%25.3%13.2%100+0.3%4.8%52.6%27.5%14.9%Total5.6%27.8%41.7%17.1%7.9%Table SEQ Table \* ARABIC 26. Percent of Observed TBI 1997 Households By Income Group By Auto Ownership?Number of Household AutosIncome Group01234+<1522.8%59.9%12.4%4.3%0.6%15-305.7%57.1%24.5%10.4%2.3%30-750.5%28.0%53.0%13.2%5.3%75-1000.0%6.9%59.4%26.2%7.6%100+0.0%2.0%62.5%24.2%11.2%Total4.4%32.2%44.3%14.1%5.0%Daily Activity Pattern ChoiceOnce basic information about where people work, go to school, and how many cars they have is predicted, the daily activity pattern choice model predicts the amount of travel they will undertake each day. The daily activity pattern choice model chooses for every person in the region whether or not they will make tours or intermediate stops for seven different purposes. The next model uses this information to determine the exact number of tours the people will take.The calibration aimed to predict a slightly higher number of people making tours of each type than the TBI due to the fact that people often forget some tours when responding to the travel survey. The percent of people by their person type making tours of a given type was compared to the TBI percent. The coefficients on the alternatives that involved tours of a given person type and tour type were decreased or increased to first match the TBI and then also to slightly increase the total number of tours. REF _Ref268698153 \h Table 27 compares the modeled and observed percent of people making tours of each of seven types. Similarly, REF _Ref268698455 \h Table 28 compares the modeled and observed percent of people making stops by each of seven types. All modeled tour purposes having a tour are within 2% of the observed tour purpose.Table SEQ Table \* ARABIC 27. Modeled versus Observed Percent of Persons Making Tours by Purpose TOURS% of Persons 2005 Focus Model1997 TBI ObservedDifference Focus %-TBI %Work Tour = 058%59%-1%Work Tour =1+42%41%1%School Tour=076%77%-1%School Tour =1+24%23%1%Escort Tour=087%87%0%Escort Tour=1+13%13%0%Personal Business Tour=087%88%-2%Personal Business Tour=1+13%12%2%Shop Tour=087%88%-1%Shop Tour=1+13%12%1%Meal Tour=095%95%-1%Meal Tour=1+5%5%1%Social Recreation Tour=083%83%0%Social Recreation Tour =1+17%17%0%Table SEQ Table \* ARABIC 28. Modeled versus Observed Percent of Persons Making Stops By Purpose STOPS% of Persons?2005 Focus Model1997 TBI ObservedDifference FocusTBI %Work Stop = 091%91%0%Work Stop =1+9%9%0%School Stop=097%98%-1%School Stop =1+3%2%1%Escort Stop=087%88%-1%Escort Stop=1+13%12%1%Personal Business Stop=079%81%-2%Personal Business Stop=1+21%19%2%Shop Stop=077%78%-2%Shop Stop=1+23%22%2%Meal Stop=088%88%0%Meal Stop=1+12%12%0%Social Recreation Stop=090%89%1%Social Recreation Stop =1+10%11%-1%Exact Number of ToursThe daily activity pattern predicts whether tours of a particular purpose will occur in each person’s day, and the exact number of tours model then predicts the exact number of tours of that purpose that will occur.As with the daily activity pattern, the coefficients on a given person type and tour purpose were increased or decreased so that the model results would first better match with the TBI. Then the coefficients were adjusted so that the number of tours was slightly higher than observed to account for respondent’s tendency to forget daily activities. REF _Ref268699443 \h Table 29 and REF _Ref268699494 \h Table 30 show the modeled and observed average number of home-based tours by tour purpose and person type. Note that the average number of tours was slightly higher in the model than observed to account for respondent forgetfulness. The total average number of tours by tour purpose looks very similar in the modeled and observed data. Table SEQ Table \* ARABIC 29. 2005 Modeled Persons by Type by Average Number of Daily Home-Based Tours by PurposePerson Type?Tour PurposeAge 0 to 4Pre Driving Age StudentDriving Age StudentFull Time workerUniversity StudentPart Time WorkRetiredNon Worker less than 65TotalWork0.000.000.240.870.270.630.040.050.44School0.210.940.920.050.870.000.000.000.24Escort0.350.120.170.130.140.340.320.250.20PersonalBusiness0.170.090.090.090.080.200.300.290.15Shop0.150.050.060.090.080.210.270.340.14Meal0.060.020.050.060.040.030.060.060.05SocialRecreation0.260.190.220.110.150.180.220.230.17All Purposes1.191.411.741.401.631.581.211.221.38 Table SEQ Table \* ARABIC 30. 1997 TBI Persons by Type by Average Number of Daily Home-based Tours by PurposePerson TypeTour PurposeAge 0 to 4Pre Driving Age StudentDriving Age StudentFull Time workerUniversity StudentPart Time WorkRetiredNon Worker less than 65TotalWork0.000.000.220.840.190.620.030.020.44School0.150.880.940.010.710.020.000.000.23Escort0.420.080.160.120.110.470.520.090.19Personal Business0.200.080.080.080.130.200.280.340.13Shop0.130.040.070.090.110.180.360.270.13Meal0.060.020.040.050.050.040.070.080.05Social Recreation0.200.160.150.110.250.150.250.220.15All Purposes1.161.651.651.291.531.681.521.031.33 REF _Ref268699717 \h Figure 5 displays the modeled and observed percent of people making a given number of tours in their day. A slight skew towards people making more tours in the model than observed can be seen in this figure, again reflecting that the model predicted more tours than observed because respondents tend to forget a small percentage of tours when responding to the travel survey.Figure 5. Percent of People by Number of Home Based Tours REF _Ref268699941 \h Table 31 shows the modeled and observed average number of daily tours by a person’s home district. The model appears to capture a fair amount of tour-making that is dependent on the accessibility of a person’s home district. For example, more accessible districts like the CBD tend to have a greater number of daily tours per person than less accessible districts like Rural.Table 31. Average Tours per Person by Home District Observed and ModeledDistrict1997 TBI2005 MODELBoulder Valley1.481.44CBD1.521.50East1.081.37East SE1.281.38Inner East1.311.45Inner North1.211.37Inner SE LRT1.221.41Inner SW LRT1.211.42Longmont1.451.36North1.231.36Outer NW1.341.39Outer SE LRT 1.431.41Outer SW LRT1.291.38Rural1.191.27Southeast1.321.38Southwest1.381.39W Northwest1.381.39West1.391.40Work Tour Destination TypeThe regular workplace location choice model selects, near the beginning of the model flow, each person’s regular workplace. Then the daily activity pattern and exact number of tours model select how many work tours a worker will make. This model component, work tour destination type, selects if each predicted work tour will go to the regular workplace. If the answer is that the work tour does go to regular workplace, the location is known for the regular location place location model, otherwise the work tour will be run through the tour primary destination choice model to predict its location.Because the model results were very similar to the observed TBI data, no calibration changes were required for the work tour destination type model. REF _Ref268700251 \h Table 32 shows the modeled and observed percent of work tours that are selected to go to the regular workplace.Table 32. Modeled and Observed % of Work Tours to the Regular WorkplaceModeled % of Work Tours to Regular Workplace86%Observed TBI % of Work Tours to Regular Workplace88%Work-Based Subtour GenerationThe work-based subtour generation model predicts for each home-based work tour the number of subtours that will be made. As with the work tour destination type model, the work-based subtour model performed quite well as compared to the observed TBI data, and so no calibration changes were required. REF _Ref268702183 \h Table 33 shows the percent of work tours that have a given number of subtours from 0 to 4 resulting from the model and the TBI.Table 33. Percent of Work Tours by Number of SubtoursNumber of Subtours on Work TourModeled Percent of Home-Based Work Tours TBI Observed Percent of Home-Based Work Tours081%81%118%17%21%1%30%0%40%0%Tour Primary Destination ChoiceThe tour primary destination choice model selects the zone to which a tour is destined. Tours destined to regular workplaces and school had their locations selected in the earlier models. However, work tours that do not go the regular workplace are run through this model. The tour primary destination choice model mainly focuses on destination zones for tours of non-mandatory purposes: escort, shop, meal, personal business, and social recreation.The first step during calibration was to make the modeled distance distribution better match the observed TBI distance distribution. REF _Ref268702495 \h Table 34 presents the 2005 modeled and TBI observed average straight line tour distances.Table 34. Average Modeled and Observed Tour Straight Line Distance?Average Straight Line Tour Distance FocusModeledAverage Straight Line Tour Distance TBI ObservedHome-Based5.25.1Work-Based3.33.2 REF _Ref269112508 \h Table 35 compares the modeled and observed average tour distances by purpose. REF _Ref268702509 \h Table 37 REF _Ref269112532 \h Figure 6 shows a histogram of modeled and observed tour distances for non-mandatory tours. Focus shows slightly fewer tours from 0 to 1 mile than the TBI, and slightly more tours from 1to 3 miles than the TBI.Table 35. Average Modeled and Observed Tour Distance By Tour PurposeTour PurposeFocus 2005 Average Modeled Tour Distance TBI 1997 Average Tour DistanceWork8.38.0School2.92.5Escort3.33.2PersonalBusiness5.04.8Shop2.43.0Meal3.93.5SocialRecreation5.65.8All Tours5.25.1Figure 6. Modeled and Observed Percent of Non-Mandatory Tours By DistanceAfter the distance distribution was calibrated, the district to district tours were calibrated to match observed data. Because the data was thin for the non-mandatory tour observations in the TBI, a different “super”district structure was used, as shown, in REF _Ref268702933 \h Figure 7 to ensure adequate that number of observations in most cells were statistically significant.Because the TBI was 8 years older than the model year and substantial regional growth occurred during this period, adjustments were made to the observed TBI district to district tours to account for the growth. First, the number of tours originating in a given district observed in the TBI was factored by the increase in persons living in each district from 1997 to 2005. Then the number of tours destined to a given district observed in the TBI was factored by the increase in jobs in each district from 1997 to 2005. Finally a fratar process was used to grow the tours so that row and column sums matched the new targeted originating and destined tours to reasonable level. The results of the fratar process used on the 1997 TBI non-mandatory superdistrict to superdistrict tours are shown in Table 35. In comparison, Table 36 shows modeled the superdistrict to superdistrict tours. Cells have been grayed out in these tables if there less than 15 observations in the TBI for the superdistrict pairs. Substantial differences are evident when comparing the modeled tours to the targets, for example in the SuperEast district. We are unsure whether this inconsistency is due to inadequacy in the target development or actual problems with the model. During calibration, dummy variables by district were added to push more or less tours to each destination district to better match the observed data. Particular attention as paid to ensure that the total tours destined to the CBD and SuperNorthWestBoulderLongmont were comparable. Figure 7. SuperDistrictsTable 36. Target SuperDistrict to SuperDistrict Non-Mandatory ToursDenver CBDRuralSuperEastSuperInnerLRTSuperNorthSuperNWBldrLngmtSuperSoutheastSuperSouthwestSuperWestGrand TotalDenver CBD25681843755948125773194790104741363Rural4085810395814184621212487632289006068112265East11220145615602121430171617620492356911836227914SuperInnerLRT2653217356999121358610341613152479351162131SuperNorth1023470559572832133528290871939133625223205629SuperNWBldrLngmt1179269965961952113923109865263387753343422SuperSoutheast23155314414492405315477243059311004494352015SuperSouthwest6677246324572424555703245815376826304248498SuperWest4707237011199787392718163306951444296701394797Grand Total55862774112727141872631631723621523241902564913887782063129Table 37. Modeled SuperDistrict to SuperDistrict Non-Mandatory ToursRow LabelsDenver CBDRuralSuperEastSuperInnerLRTSuperNorthSuperNWBldrLngmtSuperSoutheastSuperSouthwestSuperWestGrand TotalDenver CBD179662667673677283145181230309032365Rural10506745410575513860671543611752747112486137429SuperEast139301525259806389365328327819176327613302358557SuperInnerLRT74724893914010972115741654135391232021120207029SuperNorth335141061741354941568024619526822041?238084SuperNWBldrLngmt3501139127977876132360504845446118254355SuperSoutheast9775248255931513210831442210182232933656286606SuperSouthwest14991554755416716116821402188214156416982211059SuperWest9154132215328224111025523283330514431238156337645Grand Total55749828633834552180031901733296232831832051703149102063129Another way to determine if the tour primary destination choice model was accurately choosing destination zones was to compare the regional percent of tours destined to a district with the percent of people and jobs in that district. REF _Ref268763144 \h Table 38 shows this comparison of tours by district to people and jobs in the district. The column tours destined per persons+jobs shows that overall each person and job attracts approximately 1 tour. Districts with a higher number of tours are more accessible, like the CBD, and district with less tours are less accessible like the rural district. These results show that the model is performing reasonably well, even if the district to district comparison was somewhat questionable.Table SEQ Table \* ARABIC 38. Modeled Number of Tours Destined to District Compared with Number of People and Jobs?District NameNumber of Tours Destined% of Tours DestinedSum of People and Jobs% of People and Jobs% of Tours -% of People and Jobs Tours Destined Per Persons+JobBoulder Valley2185895%1791304%1%1.2CBD1829125%1173813%2%1.6East1544094%1566854%0%1.0East SE3188838%2982627%1%1.1Inner East2148505%1849605%1%1.2Inner North2017675%2734637%-2%0.7Inner SE LRT2542446%2416446%0%1.1Inner SW LRT1425944%1244643%0%1.1Longmont1356033%1367043%0%1.0North1848395%2102545%-1%0.9Outer NW2555566%2055215%1%1.2Outer SE LRT 41765910%39654810%1%1.1Outer SW LRT2054015%2045455%0%1.0Rural1737384%3138658%-3%0.6Southeast1562664%1836165%-1%0.9Southwest1698604%1841245%0%0.9W Northwest2543416%2566876%0%1.0West3385559%3327878%0%1.0Total3980066?4000640??1.0Tour Main Mode ChoiceGiven each tour’s destination and purpose, the tour main mode choice model predicts the main mode of tour. The trip mode choice model then predicts the trip modes given these tour modes.During calibration the alternative specific constants by mode and purpose were changed to make the modeled mode choices better match the observed mode choices. After one calibration run through transit assignment, the modeled transit boardings were too high as compared to observed boardings. Because substantial regional changes occurred from the 1997 TBI to the 2005 base model year, a new target was developed using the previous Compass model trip mode shares to account for regional shifts. The following describes the procedure used to develop the tour main mode choice target shares:The 1997 TBI mode shares for walk, bike, and school bus trips were taken directly since these modes are unavailable in Compass.The latest Compass 2005 mode shares for drive alone, shared ride 2, shared ride 3, walk to transit and drive to transit were factored to incorporate the TBI mode shares, so that the sum of all mode shares equaled 100%.The trip targets were bifurcated by trips to all destinations and trips to the CBD, to reflect the importance of capturing CBD behavior and its difference from other parts of the region. Also because trip mode choice behavior is significantly different for work and non-work trips, the targets were divided by work and non-work. This created four sets of targets: work-all destinations, work CBD, non-work all destinations, and non-work CBD.Next the four sets of trip mode shares had to be translated into tour mode shares. This was done applying the matrix of modeled trip mode share by tour mode shares and finding the proportion of tour mode shares that would produce the new target trip mode shares. REF _Ref268788214 \h Table 39 through REF _Ref268788219 \h Table 42 show the targeted mode shares from this process and the observed mode shares divided by work, non-work and CBD, all destinations. These tables reveal that the model mode shares and target mode shares are close.Table 39. Target and Modeled Work Mode Share All DestinationsTour ModeTarget ShareModel ShareBike0.6%0.6%Drive Alone66.6%67.2%Drive to Transit2.2%1.7%Shared Ride 218.5%19.0%Shared Ride 3+7.0%6.5%Walk2.2%2.3%Walk to Transit3.0%2.9%Table 40. Target and Modeled Non-Work Mode Share All DestinationsTour ModeTarget ShareModel ShareBike0.6%0.8%Drive Alone31.7%32.2%Drive to Transit0.30%0.15%SchoolBus5.0%3.5%Shared Ride 230.2%31.1%Shared Ride 3+22.6%22.3%Walk8.0%8.3%Walk to Transit1.7%1.7%Table 41. Target and Modeled Work Mode Share CBD DestinationsTour ModeTarget ShareModel ShareBike1.9%1.7%Drive Alone44.1%45.0%Drive to Transit13.6%11.1%Shared Ride 215.5%16.5%Shared Ride 3+5.4%5.6%Walk9.5%9.2%Walk to Transit10.1%10.9%Table 42. Target and Modeled Non-Work Mode Share CBD DestinationsTour ModeTarget ShareModelBike3%2%Drive Alone25%27%Drive to Transit3%2%SchoolBus0%0%Shared Ride 218%19%Shared Ride 3+14%12%Walk26%26%Walk to Transit11.1%11.2%Tour Time of Day ChoiceThe tour time of day choice uses all the previous modeled information about the tour and the person making said tour to predict the tour primary destination arrival time and tour primary destination departure time pair. Calibration changes were made to model alternative specific constants to push the model to better match the TBI shares by arrival time and departure time.The results of the calibration efforts are displayed in REF _Ref268790421 \h Figure 8 and REF _Ref268790423 \h Figure 9. These figures show the percent of tours by arrival and departure time hour resulting from the model and as observed in the TBI. Figure SEQ Figure \* ARABIC 8. Percent of Modeled and Observed Tours by Tour Primary Destination Arrival Time Figure SEQ Figure \* ARABIC 9. Percent of Observed and Modeled Tours By Tour Primary Destination Departure TimeIntermediate Stop Generation ChoiceAfter the time and mode of all tours is predicted, the model predicts the number of intermediate stops that will occur on each half-tour in the intermediate stop generation choice model. During calibration, coefficients by person type and tour purpose were changed to push the model to better match observed numbers of stops by person type and tour purpose. REF _Ref268790928 \h Table 43 shows the average number of daily by their purpose and person type resulting from the model. In comparison, REF _Ref268790986 \h Table 44 shows these averages as observed in the TBI. Table SEQ Table \* ARABIC 43. Average Modeled Stops by Stop Purpose and Person TypeStop PurposeFull Time WorkerPart Time WorkerUniversity StudentNon Worker Under 65RetiredDriving Age HS StudentChild Age 5 to 15Child Under 5Grand TotalEscort0.200.180.060.080.110.220.150.170.17Meal0.100.110.170.090.080.180.040.070.09Personal Business0.190.250.270.230.090.110.100.460.20School0.010.000.000.000.040.100.010.000.01Shop0.150.220.180.250.180.090.050.300.16Social Recreation0.070.070.130.070.050.130.120.160.09Work0.120.130.010.010.020.020.000.000.07Grand Total0.840.950.810.740.570.840.471.150.79 Table SEQ Table \* ARABIC 44. Average Observed TBI Stops by Stop Type and Person TypeStop PurposeFullTimeWorkerPartTimeWorkerUniversityNonWorkerUnder65RetiredDriving Age HS StudentChild Age 5 to 15Child Under 5Grand TotalEscort0.130.180.110.040.060.200.100.080.12Meal0.070.060.090.060.050.160.040.070.07PersonalBusiness0.170.220.260.250.140.110.100.380.18School0.000.020.000.000.030.110.030.000.01Shop0.200.300.380.370.150.090.070.300.21SocialRecreation0.070.070.100.050.050.170.130.190.09Work0.160.080.000.000.090.030.000.000.09Grand Total0.800.920.950.770.570.880.471.010.77Intermediate Stop Location ChoiceGiven the intermediate stops generated by the previous model component, the intermediate stop location choice model predicts in which zone the stops are located. During calibration, the coefficient on generalized time from origin to each alternative and alternative to origin was modified. The modifications were made so that the modeled average distance for trips that include at least one intermediate stop better match the observed TBI data. REF _Ref268846654 \h Table 45 shows the results of the calibration with the modeled and observed average trip distances by trip purpose. The modeled trip distance was forced to be intentionally slightly lower than the observed trip distance because total regional VMT was too high when the numbers were closer. That is, we made the trip lengths shorter for trips with intermediate stops to slightly decrease regional VMT to be in more in line with the counts.Table 45. Average Modeled Trip Distance (for trips with intermediate stops) by trip purposeTrip PurposeModeled Average Trip DistanceObserved Average Trip DistanceWork6.26.7School43.8Escort5.34.1PersonalBusiness4.44Shop4.43.2Meal4.54.6SocialRecreation4.35.1Total4.24.5 REF _Ref268847009 \h Figure 10 shows a histogram of the modeled and observed percent of trips with intermediate stops with given straight line distances. The TBI showed more 0-1 mile trips than the model, and less 1-10 mile trips than the model. The modeled distribution of trip straight line distances should be corrected to better match the observed distribution on the next calibration.Figure 10. Percent of Trips by Modeled and Observed Trip Straight Line Distancesleft10160Another way to examine if the intermediate stop location choice model is accurately selecting locations for stops is to see if the number of stops in each district predicted by the model is commensurate with the number of jobs and people in each district. REF _Ref268847275 \h Table 46 compares the share of intermediate stop locations in each district to the share of people and jobs in the district. The number of intermediate stops in each district appears approximately commensurate with the number of people and jobs. They are not exactly equal because the model incorporates many other variables including employment type and accessibility.Table 46. Number of Modeled Intermediate Stops Compared to Number of People and Jobs By DistrictDistrictNumberOfIntermediate Stopsin District% of Intermediate Stops In DistrictSum of People And Jobs% of People And JobsBoulder Valley2069227%1799544%CBD636442%1191733%DIA23240%194840%East232441%1608024%East SE33424311%3027008%Inner East1958036%1870405%Inner North2236637%2704087%Inner SE LRT520422%2422896%Inner SW LRT1426505%1256893%Longmont1004353%1384883%North1282314%2039025%Outer NW2312397%2111605%Outer SE LRT 37133112%40417310%Outer SW LRT2005466%079945%Rural686892%2907847%Southeast1447925%1853655%Southwest1448895%1899235%W Northwest2067917%2597676%West30147910%3361198%Sum3142957?4035214?Trip Mode ChoiceThe tour mode choice model selects the tour main mode. Given this information and the location of each intermediate stop, the trip mode choice model selects the mode on each individual trip.The alternative specific constants for each mode were calibrated in this model so that the modeled shares match the target shares better. The process to develop the trip mode choice targets was depicted in the section on tour mode choice. Please refer to this section for more information. Four sets of 2005 targets were developed using Compass data and TBI data in combination: work-all destinations, work-CBD, non-work-all destinations, and non-work-CBD. REF _Ref268848071 \h Table 47 compares the targeted and modeled work trip mode shares to all destinations and the CBD. REF _Ref269113606 \h Table 48 compares the targeted and modeled non-work trip mode shares to all destination and the CBD.Table 47. Work Trip Mode Choice Shares Targets and Modeled?Trip ModeTarget ModeShareAll DestinationsTarget Mode ShareCBDFocusModel Mode Share All DestinationsFocus Model Mode Share CBD DestinationBike1%2%1%2%DriveAlone79%53%80%52%DriveToTransit2%11%1%9%SchoolBus0%0%0%0%SR210%11%10%9%SR33%3%3%3%Walk3%12%3%15%WalkToTransit3%9%3%11%Table 48. Non- Work Trip Mode Shares Targets and Modeled?Trip ModeTarget Mode Share All DestinationsTarget Mode Share CDBFocusModel Mode Share All DestinationsFocus Model Mode Share CBD DestinationBike1%3%1%2%DriveAlone40%29%41%25%DriveToTransit0%2%0%2%SchoolBus3%0%3%1%SR226%17%27%14%SR319%13%19%10%Walk9%29%9%36%WalkToTransit1%7%1%10%Trip Departure Time ChoiceThe trip departure time choice model is last logit model in the Focus model stream. It selects the intermediate stop arrival time for trips on the outbound half-tour and intermediate stop departure time for trips on the inbound half tour. The times associated with the tour primary destination are known from the tour primary destination choice and are used to create windows of time availability when the intermediate stop can be scheduled. During calibration, the constants associated with each time choice were adjusted to better match the modeled time distribution to the observed TBI time distribution. REF _Ref268848929 \h Figure 11 contrasts the observed and modeled times associated with all trips, including trips without intermediate stops (so this figure includes trips whose times were predicted by tour primary destination choice). From this figure, it is evident that the TBI showed slightly more travel occurring from 11 am to 7 pm than the model. This problem should be addressed in the next calibration.Figure SEQ Figure \* ARABIC 11. Percent of Trips by Trip Time Choice TBI and ModeledHighway AssignmentThe trips with their predicted locations, modes, and times of day are passed to TransCAD for assignment.say something about changes here. REF _Ref268852783 \h Table 49 shows the observed and modeled VMT on links with 2005 counts. The modeled VMT is 1.9% higher than the observed VMT on links with counts.Table SEQ Table \* ARABIC 49. Observed and Modeled 2005 VMT on links with countsObserved VMTModeled VMT20,506,76820,906,583 REF _Ref268852659 \h Table 50 compares the observed VMT on links with counts to the modeled VMT on the same links divided by facility type of the link. The model is over-predicting larger facility types like expressway and freeway and under-predicting facility types like minor arterial and collector. This problem had been also seen in the latest Compass calibration, and should be addressed in a later calibration. Also, toll volumes are still being under-predicted by Focus, as they were in Compass 2.0. The problem should be remedied by adding multiple classes with varying values of time.Table SEQ Table \* ARABIC 50. Total Vehicle Miles Traveled By Facility Type on Links with CountsFacility TypeNumber of Counted Links Modeled VMT % Modeled VMT Observed VMT % Actual VMT % DifferenceFreeway210971189946.5%960520646.9%1.1%Expressway7117360818.3%15874757.8%9.4%Principal Arterial863807606038.7%745039036.4%8.4%Minor arterial31610009064.8%12783906.2%-21.7%Collector2183672211.8%5585562.7%-34.3%Toll (subset of Freeway) 1391514N/A145526N/A-37% REF _Ref268853119 \h Figure 12 displays the area types used in the model to summarize the VMT by area type on links with counts shown in REF _Ref268853195 \h Table 51. Most of the modeled VMT is close to the observed VMT by area type with the exception of CBD. A similar problem had been seen in the latest Compass 4.0 calibration. When this issue was investigated, it was found that many of the links were along the same road, and so if parallel links were selected by the model in path-finding the model may have under-predicted VMT on one road in favor of a parallel route. In other words, the under-prediction observed in the CBD may have been an artifact of the paucity of links observed in the CBD, rather than a major concern with the model trip distribution. The issue of whether CBD and CBD Fringe VMT are being under-predicted in favor of more suburban and rural areas should be addressed in a later calibration.Figure 12. Area TypesTable 51. Total Vehicle Miles Traveled By Area Type on Links with Counts Area TypeNumber of Links Modeled VMT %Modeled VMT Actual VMT %Actual VMT % DifferenceCBD59424000.2%742200.4%-42.9%CBD Fringe15212163905.8%14312157.0%-15.0%Urban585699031333.4%682592433.3%2.4%Suburban569757289736.2%744390136.3%1.7%Rural318508458324.3%473150823.1%7.5% Total 1,683 20,906,583 100% 20,506,768 100%1.9% REF _Ref268853732 \h Figure 13 shows the screenline locations used to determine how well the model was capturing flows into and out of important geographic areas. Observed and modeled volumes on screenlines with counts are shown in REF _Ref268854265 \h Table 52. Cells have been grayed out if less than 3 links were observed at the screenline, due to insufficient numbers of observations to make conclusions. The model has some problems with volumes to DIA as can be seen with the Tower Rd screenlines. This issue comes from the Compass part of the model (Focus does not forecast DIA trips) and will require revisiting in a later calibration.Figure 13. Screenline Locations352425-257175Table SEQ Table \* ARABIC 52. Modeled and Observed Volumes By ScreenlineScreenlineNameLinks with CountsObserved Volumes?Modeled VolumesPercent ErrorTotal All Screenlines157841,945,65542,865,4812%120Th10247,457278,22712%Boulder Circle110,97614,17829%Castle Rock259,52071,87021%Colfax16405,928420,0023%Colorado Blvd10419,719428,6192%DIA3100,86281,891-19%Downtown Circle19423,675395,146-7%Hampden10504,249487,426-3%Kipling9188,526203,9258%Tower Rd564,60338,788-40%Wadsworth20581,624600,0903%Total1053,007,1393,020,1610%Transit AssignmentTransit assignment takes the transit trips created by earlier model components and assigns them to the transit network. Focus has four periods for transit assignment: AM Peak, Midday (MD), PM Peak, and Early Late(EL). The trips for transit assignment come from Focus for all locations, except DIA trips which come from the old GISDK Compass part of the model. The assigned walk-to-transit trips come directly from the trip mode choice model. The assigned drive-to- transit trips, however, come from the tour mode choice model in production-attraction format. They are converted to OD format in the GISDK code. The reason for this simplification was that we could not find an easy way to track a drive to transit tour’s park and ride location, to ensure that the same park and ride was returned to, on the destination to origin part of the tour. The transit trips coming from the Focus model trip and tour mode choice models are summarized in REF _Ref268864249 \h Table 53. This table also shows that the 2005 Focus modeled boardings per linked trip were about 1.3 and the observed boardings per linked trip from RTD’s 2007 on-board survey were 1.4. Table SEQ Table \* ARABIC 53. Modeled Transit Trips and Boardings2005 Modeled Walk Transit Trips152,8662005 Modeled Drive Transit 2 X Tours49,218Modeled Total Transit Trips200,773Modeled Total Transit Boardings261,153Modeled Boardings Per Linked Trip1.3Observed On-Board Survey 2007 Boardings Per Linked Trip1.4 REF _Ref268864705 \h Table 54 compares the 2005 RTD Boardings by sub-mode to the modeled boardings by sub-mode. Overall the total boardings are 1% lower in the model than the observed data. A few issues that can be seen in REF _Ref268864705 \h Table 54 are that the express bus boardings resulting from the model are too high, as compared to the observed boardings, and the regional bus boardings resulting from the model are too high.Table SEQ Table \* ARABIC 54. Average RTD Observed and Modeled Boardings By Sub-Mode?Transit Sub-modeObservedModeledDifference% DifferenceMall Shuttle 46,729 39,413 -7,316-16%Denver Local Bus 123,821 124,329 5080%Limited Bus 17,497 15,884 -1,613-9%Express Bus 10,741 18,738 7,99774%Regional Bus 11,355 6,705 -4,650-41%Rail 34,578 34,644 660%Longmont Local 689 1,253 56482%Boulder Local 19,210 20,187 9775%Total 264,620 261,153 -3,467-1%Conclusions and Further Calibration DirectionsThe overall aggregate assignment calibration measures match observed datasets fairly well. Many minor outstanding issues exist that should be rectified in the next calibration. This next calibration should also use newer data from the Front Range Travel Counts survey and the 2010 Census.In terms of individual component calibration, the first model component with an obvious issue is intermediate stop generation. In this component, the TBI recorded an average of 0.3 stops per meal tours, whereas the model projected 0.7 stops per meal tour. Another model component with a problem evident is the trip time of day choice model. It is predicting too little travel from 11 am to 7 PM, as compared to the TBI.In terms of aggregate highway assignment measures, the Focus model is presenting several problems that the Compass model had in versions 2.0 and 3.0. Some of these problems should be fixed by making the same changes as in Compass 4.0. Others of the problems were also seen in Compass 4.0 and will require some innovation. For example, the Focus model was consistent with Compass in that it was predicting too much travel on larger facility types like expressways and too little travel on facility types like collectors. Modeled toll volumes are lower than observed and this should be improved by the inclusion of a multi-class assignment with varying values of time, as in Compass 4.0. Focus shows greater VMT in suburban and rural areas than the counts, as was seen in Compass 4.0, and this problem may need to be addressed in greater detail.For the transit assignment, Focus has a few problems with boardings by sub-mode. The express bus boardings resulting from the model are too high, as compared to the observed boardings, and the regional bus boardings resulting from the model are too high.Although many problems still exist with the Focus model 2005 calibration, none are a cause for major concern. All are on the scale of problems seen with existing versions of Compass. New person travel, commercial travel, and speed data from the Front Range Travel Counts should be used to rectify the existing problems with Focus. Our work never ends!The following appendices show aggregate model results for comparison purposes for later model runs.Appendix 1: Basic Aggregate Model ResultsThis appendix presents basic aggregate model results resulting from the activity generation, location choice, mode choice and time of day choice models.Total Persons2,715,548Total Person Trips11,128,553Total Vehicle Trips7,378,747Total Households 1,072,494 Average Person Trips per Person4.1Average Vehicle Trips Per Person2.7Average Household Size 2.53 ???????????ModeNumber of Person Trips Model Run% of Person Trips Model Run?PurposeNumber of Person Trips Model RunAverage Trips per Person Model RunAverage Straight Line Distance Model Run?Bike705400.6%Work 2,751,907 1.07.2Drive Alone561479250.5%School 1,382,855 0.52.9Drive to Transit454280.4%Escort 1,762,607 0.63.9SchoolBus2153051.9%Personal Business 1,646,718 0.64.2Shared Ride 2254916222.9%Shop 1,475,574 0.53.2Shared Ride 3+168834015.2%Meal 813,110 0.33.7Walk7921207.1%Social Recreation 1,295,796 0.54.8Walk to Transit1528661.4%Total 11,128,567 4.14.6Total11,128,553????????????Time of DayPerson Trips Model RunShare of Person Trips Model Run?Destination DistrictNumber of Trips Model RunShare of Regional Trips?4:00 AM26,3950%Boulder Valley 588,122 5%5:00 AM293,9323%CBD 294,015 3%6:00 AM366,1443%DIA181170%7:00 AM1,047,6229%East 313,651 3%8:00 AM1,119,72210%East SE 984,917 9%9:00 AM447,2314%Inner East 597,478 5%10:00 AM445,8404%Inner North 695,562 6%11:00 AM646,3426%Inner SE LRT 554,304 5%12:00 AM600,2735%Inner SW LRT 405,029 4%1:00 PM706,9106%Longmont 375,169 3%2:00 PM794,1657%North 546,768 5%3:00 PM949,4669%Outer NW 693,167 6%4:00 PM766,9217%Outer SE LRT 1,140,115 10%5:00 PM827,7057%Outer SW LRT 611,164 5%6:00 PM606,0755%Rural 535,178 5%7:00 PM452,1244%Southeast 525,375 5%8:00 PM415,1404%Southwest 528,189 5%9:00 PM236,3782%W Northwest 735,376 7%10:00 PM207,7512%West 986,865 9%11:00 PM68,8911%Total 11,128,561 ?12:00 AM30,8090%1:00 AM253270%2:00 AM230060%3:00 AM243980%?CountyPopSyn & Land Use Model Households% of Regional HouseholdsNumber of AutosNumber of Households% of HouseholdsAdams14526014%0598536%Arapahoe21143520%129787728%Boulder11795311%244712442%Broomfield167062%318296117%Clear Creek42810%4+846798%Denver25384224%Douglas877148%Tour PurposeTotal ToursAvg Tours Per PersonAverage DistanceGilpin22480%Work12354930.58.1Jefferson21332320%School6687590.22.9Elbert88011%Escort5494020.23.3Weld109311%Personal Business4316310.24.8Total Households1072494?Shop4130380.22.5Meal2190100.13.7Total Workers1444628Social/Recreation4754690.25.5Number Work At Home103671Total39928021.55.1% Work At Home7%?Average Straight Line Distance From Home To Regular Workplace (if works outside the home)8.3Tour ModeTotal Tours% of ToursBike280401%Average Straight Line Distance To School3.1Drive Alone171865143%Drive to Transit246091%SchoolBus957682%Shared Ride 2109212227%Shared Ride 3+69631817%Walk2561416%Walk to Transit811532%Total3992802?Tour Time of DayTour Destination Arrival HourShare of ToursTour Destination Departure HourShare of Tours?Destination DistrictNumber of Tours DestinedShare of Regional Tours4:00 AM19,8859%29131%Boulder Valley 218,589 5%5:00 AM33,91916%20%CBD 182,912 5%6:00 AM249,141117%3110915%DIA127360%7:00 AM685,163321%7735936%East 154,409 4%8:00 AM706,730331%11582954%East SE 318,883 8%9:00 AM231,311108%8479440%Inner East 214,850 5%10:00 AM167,84879%14388967%Inner North 201,767 5%11:00 AM210,29599%18128885%Inner SE LRT 254,244 6%12:00 AM207,30497%239507112%Inner SW LRT 142,594 4%1:00 PM187,47588%289885136%Longmont 135,603 3%2:00 PM151,64571%475500223%North 184,839 5%3:00 PM176,16083%594835279%Outer NW 255,556 6%4:00 PM182,71886%387246182%Outer SE LRT 417,659 10%5:00 PM183,97586%387804182%Outer SW LRT 205,401 5%6:00 PM216,484101%16031875%Rural 173,738 4%7:00 PM177,79483%19195490%Southeast 156,266 4%8:00 PM97,05045%259178121%Southwest 169,860 4%9:00 PM54,29725%12960661%W Northwest 254,341 6%10:00 PM45,74721%12203257%West 338,555 8%11:00 PM1,1021%4057219%Total 3,992,802 ?12:00 AM1,3631%2308411%1:00 AM18531%186829%2:00 AM15141%164308%3:00 AM20291%189869%Appendix 2: Assignment Model Summary File MODEL SUMMARY Model Label: 2005FocusCalibration Description: Final Calibration Run 2005 Focus Date: Tue Aug 03 13:34:04 2010COMPASS SOCIAL-ECONOMIC DATA~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ POPULATION HOUSEHOLDS EMPLOYMENT AVG HHSIZE POP/EMP RATIOTotal 2,699,986 1,068,817 1,323,179 2.53 2.04TMA 2,616,743 1,040,239 1,301,024 2.52 2.01CBD 15,162 10,638 91,187 1.43 0.17DIA 42 19 24,223 2.21 0.00The fraction of households in the TMA is: 0.973The ratio of HH in the 9-County area vs. TMA is: 1.014VEHICLE ASSIGNMENT~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 10 period Factor Hours VMT VHT Speed AM1 0.346 0.5 2,066,122 54,181 38.1 AM2 0.468 1.0 6,144,595 199,123 30.9 AM3 0.359 1.0 5,760,043 180,972 31.8 PM1 0.248 2.0 10,205,802 285,111 35.8 PM2 0.289 1.0 5,794,652 172,062 33.7 PM3 0.214 1.0 3,655,689 92,083 39.7 OP1 0.015 7.5 5,446,441 128,705 42.3 OP2 0.098 2.5 7,693,776 185,619 41.4 OP3 0.130 3.5 15,619,391 404,055 38.7 OP4 0.048 4.0 7,922,445 189,546 41.8Period Hours VMT VHT Speed AM 2.5 13,970,761 434,276 32.2 PM 4 19,656,143 549,257 35.8 Off-Peak 17.5 36,682,053 907,926 40.4 Peak Hours 2 11,939,248 371,185 32.2 Peak 6.5 33,626,904 983,533 34.2 All-Day 24 70,308,957 1,891,458 37.2 Facility Type VMT VHT Speed Freeway 25,003,018 408,794 61.2 Expressway 4,145,805 86,906 47.7 Principal 22,790,023 666,225 34.2 Minor 7,822,282 249,375 31.4 Other 10,547,829 480,158 22.0 Total 70,308,957 1,891,458 37.2 HOV 261,349 5,659 46.2 TOLL 800,483 11,197 71.5VMT in TMA is: 67,682,545VHT in TMA is: 1,841,847Average speed in TMA is: 36.7VMT in 9-County Region is: 68,630,101VHT in 9-County Region is: 1,867,633Average Speed in 9-County Region is: 36.7Total VHD is: 179,720Percent Delay is: 9.5% Total Highway Trips Assigned: 8,846,792COMPASS Interzonal trips are: 7,343,020COMPASS Intrazonal trips are: 444,017COMPASS Total vehicle trips: 7,787,037Total Vehicle Trips: 7,787,037Total Vehicle Trips in TMA: 7,578,828Total Vehicle Trips in 9-County Region: 7,684,931TRANSIT ASSIGNMENT~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ TRANSIT BOARDINGS Peak Off-Peak Walk Access Drive Access Total ------------------------------------------------------------Mall Shuttle: 25,411 14,002 26,351 13,063 39,413Local Bus: 73,436 50,893 107,471 16,858 124,329Limited Bus: 12,126 3,758 12,286 3,598 15,884Express Bus: 16,405 2,333 7,750 10,988 18,738Regional Bus: 5,239 1,466 2,707 3,998 6,705Rail: 21,231 13,413 14,839 19,805 34,644Skyride Bus: 823 430 811 442 1,253Longmont Local: 1,367 988 2,109 246 2,355Boulder Local: 12,068 8,119 17,696 2,491 20,187 -----------------------------------------------------------------------------Total: 168,106 95,402 192,020 71,488 263,508 TRANSIT PASSENGER MILES TRAVELED Peak Off-Peak Walk Access Drive Access Total ------------------------------------------------------------Mall Shuttle: 16,198 8,473 17,216 7,455 24,671Local Bus: 252,753 175,477 353,722 74,509 428,230Limited Bus: 64,262 20,308 61,693 22,877 84,570Express Bus: 157,703 22,820 67,474 113,049 180,523Regional Bus: 92,355 26,938 40,650 78,643 119,293Rail: 127,888 73,409 70,257 131,041 201,297Skyride Bus: 14,050 6,171 14,509 5,711 20,221Longmont Local: 2,702 1,991 4,139 554 4,692Boulder Local: 27,517 17,529 37,977 7,069 45,047 -----------------------------------------------------------------------------Total: 755,427 353,117 667,636 440,908 1,108,544 TRANSIT PASSENGER MILES TRAVELED PER BOARDING Peak Off-Peak Walk Access Drive Access Total ------------------------------------------------------------Mall Shuttle: 0.6 0.6 0.7 0.6 0.6Local Bus: 3.4 3.4 3.3 4.4 3.4Limited Bus: 5.3 5.4 5.0 6.4 5.3Express Bus: 9.6 9.8 8.7 10.3 9.6Regional Bus: 17.6 18.4 15.0 19.7 17.8Rail: 6.0 5.5 4.7 6.6 5.8Skyride Bus: 17.1 14.3 17.9 12.9 16.1Longmont Local: 2.0 2.0 2.0 2.3 2.0Boulder Local: 2.3 2.2 2.1 2.8 2.2 -----------------------------------------------------------------------------Total: 4.5 3.7 3.5 9.3 4.2 Total Transit Boardings: 263,508Total Transit Linked Trips: --Total Boardings per Trip: -- ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download