Home

TELUM (Transportation Economic and Land Use Model)

1. FORECAST YEAR 2005 OUTPUT DATA HOUSEHOLD TYPE LI LM MH HH TOTAL 1 98 835 41 505 222 2 108 1234 39 4 20 3 124 148 5T 25 348 4 515 53 4 Tes LES 5 L09 226 198 T6 610 6 58 81 LNG 66 320 7 12 22 105 161 300 8 34 100 107 54 295 9 554 124 T1325 Si 341 10 34 60 103 21 217 TOTAL 682 1021 876 514 3093 Page 13 BASE YEAR TO FORECAST YEAR PERCENT CHANGE IS PRINTED IF BASE YEAR WAS 0 0
2. SUMMARY OF CRITERION VALUES AT START OF EACH ITERATION Criterion values show how well the estimated value fits the observed value for this locator type Household Type 1 by Zone MANA 01 ho Ko 10 This shows the gradual movement of the raw unscaled goodness of fit measure towards zero over the twenty iterations of the parameter estimation NPRPRPRPRPRPRPE RB 1oU0 SUMMARY OF ITERATION COUNTS SUMMARY OF VALUES OF 10 PARAMETERS AT START OF EACH ITERATION ITER 1 ITER 2 ITER 3 ITER 4 ITER 5 For the PARAMETER 1 7 296600 7 667818 7 855647 7 863288 7 897191 8 021053 8 072397 8 156564 8 169782 8 189485 advanced 8 266637 8 337057 8 361495 8 362821 8 384958 user this tells 8 496555 8 630968 8 639703 8 657081 8 699234 PARAMETER 2 5 317700 4 323781 3 803317 3 740408 3 433186 more about 2 755781 2 555497 2 254523 2 277145 2 265309 the search 2 268482 2 300696 2 293999 2 295932 2 305396 procedure 2 313066 2 372806 2 370031 2 366431 2 360747 ae PARAMETER 3 2 464600 2 405665 2 293798 2 142106 6 934908 799049 529587 49065
3. 00 00 07 00 85 392 22 44 06 28 71 DVPL D 50 80 63 95 87 90 90 79 94 9 6 LPBL 100 EV D ER LAND USE VACANT 5 VACANT lng 50 00 1 20 00 2 36 93 Iy 5 00 Lz 14 29 0 12 08 Ts 9 48 2 9 56 3s 20 94 Lis 54 72 13 Page 16 8 32 OTH ETS O 65 6 65 COO OO 3 STR UNUSBLE OO 00 00 0 0 0 A H O D e Di 100 CQ No ES 000 100 pei TOTL Ck Ck ck KK KKK SUMMARY OF BASE YEAR TO FORECAST YEAR CHANGES TOTAL REGIONAL PERCENT CHANGE LI LM MH HH TOTAL 1 6 3 30 3 29 2 08 Es c N oe oe oe ZONAL MEAN ABSOLUTE PERCENT CHANGE EE LM MH HH TOTAL 16 88 10 46 8 17 352 97 F 8
4. Ey c diem i Always check this 110899 Your input use it to to be sure it is the specifically identify run you think it is your runs gt ck CK 0 RUN FOR HOUSEHOLD INCOME GROUP 1 TRAVEL FUNCTION TYPE Confirm that this is what ER OF ZONES 10 you wanted ER OF EMPLOYMENT S I C GROUPINGS ER HOUSEHOLD TYPE See notes on model formu ER AND USE ATTRACTIVENESS VARIABLI lation DITIVE LAG TERM INCLUDE DRAM MODEL STRUCTURE TOTAL LAGGED HOUSEHOLDS EMPLOYMENT TO HOUSEHOLD HEAD CONVERSION MATRIX R Based on regional data 1216 0322 2412 6050 converts employees by type at 0522 0521 3285 5672 place of work to households by 2747 3215 2901 1138 type at place of work 3000 3635 2311 1054 FIRST FOUR EMPLOYME ES IN FIRST T Is this data correct 145 153 a 50 Does it match 56 the target year I6 in the EMPAL 198 calibration
5. 1 53 89 155 243 125 120 320 401 216 280 2 89 71 108 157 177 200 392 1449 245 253 3 155 108 74 139 157 214 372 388 180 145 4 243 157 139 132 291 338 509 523 3177 0 5 125 IT i157 291 80 72 220 278 98 217 6 120 200 214 338 72 59 200 300 159 289 7 320 392 372 509 220 200 117 165 215 376 8 401 449 388 523 278 300 165 132 209 329 9 216 245 180 317 98 159 215 209 97 163 10 280 253 145 230 217 289 376 329 163 85 MEAN VALUE OF UNSCALED IMPEDANCE 229 THIS IS USUALLY IN TENTHS OF MINUTES NOTE K FACTORS FOR SCALING FORECASTS READ IN K FACTORS MULTIPLIED BY 90 Page 9 8 27 FORECAST YEAR 2005 INPUT DATA Does this match the output of the TELUM EMP ruin 880 3093
6. INPUT DATA SET VARIABLES gt ko ko ko Ck ce ck ck xj D FCST YR nd Is this the data you meant to use 1 4 2 5 3 7 4 6 5 10 6 9 7 1 8 8 9 9 0 7 nd Does it match the other runs HOUSEHOLDS BY TYPE Check these data too SZ 00 1 7 10 Ck
7. k ko KKK KKK IMPEDANCE DATA UPPER LEFT CORNER OF MATRIX FIRST 10 X 10 ZONES Ck ck ck ck ck ck kk k kc KKK 1 2 3 4 5 6 7 8 9 10 MEAN VALUE OF UNSCALED IMPEDANCE THIS IS USUALLY IN TENTHS OF MINUTES Do these impedance values BER RRB RRR RRB RRR RRR Ree eee make sense THE ELEMENTS OF THE IMPEDANCE MATRIX Look at your WILL BE DIVIDED BY 10 000 BEFORE CALIBRATION map p e Y x VALUE OF MAXIMUM LIKELIHOOD CRITERION FOR UNIFORM DISTRIBUTION What is the C W 921 0340 relative distance between zones VALUE OF MAXIMUM LIKELIHOOD CRITERION FOR PERFECT FIT Are these C B 751 8290 impedance values consistent WORST VALUE OF SEARCH CRITERION IS C W C B 169 2050 STARTING VALUES OF PARAMETERS Aisha E The statistical Does this Beta 2 2338 calculations number make Empl 3 6737 begin here sense Land 5 4202 Lambda 2 5000 IMPEDANCE RESCALED x 0 10
8. FINAL VALUES OF EMPAL PARAMETERS These are the values you will type into the BETA VALUE ADJUSTED TO REFLECT INTERNAL SCALING USE THIS ADJUSTED VALUE AS INPUT TO EMPAL EMPAL control card OR TO NEXT RUN OF CALIB for this employment ASYMPTOTIC STANDARD ASYMPTOTIC ERRORS T VALUES Alpha 1312 3 90 Beta 0237 20 59 Empl 3660 20 43 Land 6187 6 53 Lambda 1091 5 40 A goodness of fit measure Statistically R SQUARED VALUE FOR COMPARISON this is not as useful as the next one RSQ 8482 This is a better measure of goodness B W LR 8927 of fit of the EMPAL equation to this data EST WORST LIKELIHOOD RATIO THE RANGE OF THE LIKELIHOOD RATIO IS BETWEEN 0 0000 AND 1 0000 FOR A PERFECT FIT THE LIKELIHOOD RATIO WOULD BE EQUAL TO 1 0000 7 16 KKK KKK These are measures zone by zone of the sensitivity of this locator type Employment Type 1 to each attractiveness variable ZONAL LOCATION ELASTICITIES
9. FORECAST YEAR 2005 OUTPUT DATA ESDNTL 47 54 85 47 03 sok 78 85 eL 63 00 0 5 ESIDENTIAL 5 35 725 45 5 O ETAIL R 83 53 98 69 46 51 32 64 72 71 47 93 r9 95 96 97 16 97 21 8 Page 15 ED LAND USE ETAIL NOLO NN T1233 12 6 8 31 ELOP DEV EC BAS 00 STA 64 89 29 5 22 2 3 9 07 7 RESIDENT POPULATION 1 616 2 621 3 789 4 259 5 1379 6 724 7 679 8 667 9 TESZ 10 493 TOTAL 7001 TOT DVLPD BASIC 1 Le 0 2 4 0 3 4 0 4 194 5 5 6 1 6 34 0 7 14 0 8 15 0 9 9s 0 10 12 O TOTAL 9 14 FORECAST YEAR 2005 OUTPUT DATA
10. KK I 00 Do these LM 00 match the MH 00 9 5 00 spreadsheet EMPLOYMENT HSHOLD CONVERSION MATRIX EHOLD TYP LM LInd Hind Servi Retail JOBS PER EMPLOYEE REGIONAL 1 000 NET COMMUTING RATE REGIONAL 1 000 Page 3 8 24 LAGGED TOTAL HOUSEHOLD INPUT DATA
11. Ck ko kc ko ko KKK MyCity 12 06 00 Baseline 2000 2005 CK ck ck ck ck ko KKK ECASTS WILL BE MADE USING THE ADDITIVE LAG S FORMULATION OF TELUM RE FOR LOCATION SURPLUS WILL NOT BE CALCULATED ERATION NOT AVAILABLE WITH AN ADDITIVE LAG TERM Page 1 822 TRIP GEN Check all parameters against calibration runs ACkCkck ck ckckck ck ck ck ck kk VARIABLES LAND USE HOUSEHOLD ALLOCATION PARAMETERS TRIP FUNCTION HOUSEHOLD TO HOUSEHOLD ATTRACTION PARAMETERS LOCATED HOUSEHOLDS Check all parameters against calibration runs Check against land consumption regression results ERS ET HOUSEHOLD ATTRACTION PARAM 7801 8866 9988 1 0000 e w uHe LAGGED TOTAL REGRESSION PARAMETERS
12. 1 211 2 213 3 317 4 152 5 605 6 301 7 298 8 301 9 362 10 211 TOTAL 3030 MEAN 3035 Page 4 kk ck ck Ck ck KK KKK KKK KKK BASE YEAR 2000 INPUT DATA HOUSEHOLD TYPE Correct 1 2 3 4 5 6 7 8 9 0 1 TOTAL 675 1009 848 497 3030 Page 5 8 25 BASE YEAR 2000 INPUT DATA PULATION VARIABLES NONWORKING TOTAL POPULATION POPULATION 277 488 357 630 416 T33 199 350
13. EMPLOYMENT TYPE Light In Heavy In Servic Retail TOTAL 1 2 3 00 207 20 10 2 04 2 10 14 05 3 9 60 8 i 0 4 02 01 17 2SL3 02 5 20 4 39 11 SLS ld 6 00 36 09 02 03 7 16 00 08 S01 00 8 09 1 74 06 03 03 9 520 25 Sw 14 10 02 41 03 02 00 Page 7 8 20 SUMMARY OF BASE YEAR TO FORECAST YEAR CHANGES TABLE OF GINI COEFFICIENTS OF DISPERSION EMPLOYMENT BAS ECAST Y Light In 577 Is employment Heavy In 722 in your region Service 726 dispersing Or Retail 7996 concentrating TOTAL EMP 1 575 5 NOTE THAT 0 00 IS TOTALLY DISPERSED 1 00 TOTALLY CONCENTRATED KK KK KKK
14. ck CK REGIONAL LOCATION 8 AVERAGE STANDARD ELASTICITY DEVIATION IMP 0 4342 VACDEV 1246 0126 PERDEV 1864 0189 RESLND 9881 1000 LIHH 6 5805 6659 LMIHH 1 0907 1104 UMIHH 8482 0858 UIHH 2 4283 2457 LAGHH 1755 0613 THE INTERPRETATION OF A LOCATION ELASTICITY IS FOR A CHANGE IN THE LISTED VARIABLE OF 1 00 Just as it THE LOCATOR WOULD INCREASE OR DECREASE BY THE ELASTICITY AMOUNT IF ELASTICITY 0 4316 THEN FOR A 1 00 Says INCREASE IN THAT SPECIFIC VARIABLE FOR THAT ZONE THERE WOULD BE A 0 4316 DECREASE IN THAT LOCATOR IN THAT ZONE ALL OTHER THINGS BEING EQUAL I EGIONAL ELASTICITY PECIFIC LOCATOR TO TH TH GIVES THE AVERAGE SENSITIVITY OF THE SPECIFIC VARIABLE 2 9 7 32 This table compares the input data to
15. HOUSEHOLD TYPE LI LM MH HH TOTAL 1 52 30 34 22 29 2 zd 0 3 46 ls 3 I3 2 T3 75 10 4 31 28 34 154 24 5 Ll 2 s li 6 20 20 2 6 7 2 35 Fe 2 Ls 8 2 10 225 20 2 9 28 M 2 6 10 Avis T2 Gu 25 3 TOTAL 1 de 3s 3z Page 14 8 30 FORECAST YEAR 2005 OUTPUT DATA PULATION VARIABLES NONWORKING TOTAL POPULATION POPULATION 344 616 346 621 440 789 145 259 TO 1379 404 724 379 679 372 667 432 T1345 215 493 3908 7001 LATION 271 275 349 114 610 320 300 295 341 218 OY 3093 14 EMP POPUI Page ARY PO I RS AR TIO 0 0 0 0 0 SUPPLEM GROUP QU POPULA
16. IMPEDANCE MATRIX UPPER LEFT 13 13 ENTRIES 1 3 89 155 243 125 120 320 401 216 280 2 89 71 108 157 177 200 392 1449 245 253 3 155 108 74 139 157 214 372 388 180 145 4 243 157 139 132 291 338 509 523 317 230 5 125 177 157 291 80 72 220 278 98 217 6 120 200 214 38 72 59 200 300 159 289 7 320 392 372 509 220 200 117 165 215 376 8 401 449 388 523 278 300 165 132 209 329 9 216 245 180 317 98 159 215 209 97 163 10 280 253 145 230 217 289 376 329 163 85 MEAN VALUE OF UNSCALED IMPEDANCE 229 THIS IS USUALLY IN TENTHS OF MINUTES Page 3 BASE YEAR 2000 INPUT DATA EMPLOYMENT TYPE Light In Heavy In Serv
17. ck ck ck CK 1 2 3 4 5 6 7 8 9 0 1 MEAN VALUE OF UNSCALED IMPEDANCE 540 THIS IS USUALLY IN TENTHS OF MINUTES Does this match the THE ELEMENTS OF THE IMPEDANCE MATRIX EMPAL calibration WILL BE DIVIDED BY 100 000 BEFORE CALIBRATION input Check consistency again VALUE OF MAXIMUM LIKELIHOOD CRITERION FOR UNIFORM DISTRIBUTION C W 1957 1980 VALUE OF MAXIMUM LIKELIHOOD CRITERION FOR PERFECT FIT C B 1830 2000 The statistical WORST VALUE OF SEARCH CRITERION IS C W C B 126 9972 calculation STARTING VALUES OF PARAMETERS begin here ALPHA 7 2966 BETA 5 3177 VACDEV 2 4646 PERDEV 49344 RESLND 1 0938 LIHH 10 7542 LMIHH 7892 UMIHH 0529 UIHH 3 1115 LAGHH 5800 GRADIP SEARCH PROCEDURE PARAMETERS ALLOW UP TO 20 OUTER ITERATIONS EACH CONTAINING UP TO 3 LOOPS EACH OF WHICH MAY HAVE UP TO 5 STEPS 7 24 OUTER ITERATION 1 ko ck
18. HE REGIONAL ELASTICITY HE SPECIFIC LOCATOR TO GIVES TH T SUMMARY OF CALIBRATION RESIDUALS ko kc ck CK ZONE OBSERVED ESTIMATED RESIDUAL PCT DIF 1 6 00 12 82 6 82 113 6510 2 1 00 2 91 1 91 190 9278 3 110 00 83 53 26 47 24 0670 4 92 00 85 93 6 07 6 5946 5 35 00 25 91 9 09 25 9575 6 18 00 24 84 6 84 38 0065 7 44 00 39 91 4 09 9 2983 8 53 00 52 72 28 5364 9 40 00 71 02 31 02 71 5408 10 1 00 42 58 58 4391 EAN ABSOLUTE PERCENT ERROR 36 957 R ES WITH 0 OR 1 OBSERVED ARE OMITTED These are alternative measures of goodness UM OBSERVED LOCATOR VALUE Jm ORE UM OBSERVED LOCATOR VALUE 110 3 FOR 3 SMALLEST 25 OF ZONES 100 000 ZONES HAVE 2 00 OF THE
19. GRADIENT SEARCH BEGINNING OVER AGAIN AT ITERATION 1 GRADIP SEARCH PROCEDURE PARAMETERS ALLOW UP TO 20 OUTER ITERATIONS EACH CONTAINING UP TO 3 LOOPS EACH OF WHICH MAY HAVE UP TO 5 STEPS 7 11 ITERATION 1 Meaning decrease parameter Meaning increase parameter Since X is greater in importance than Y then lambda is a more important parameter ER OUTER ITERATION k ck CK LOOP GRADIENT FIRST STEP SIZE CRITERION OUT 1 395 96 18358830 72 lt 69 2 26458 1 13653300 1 3 21 13 1 08056700 29 32 PARAMETER 1 Alpha LOOP 1 LOOP 2 LOOP 3 PARAMETER VALUES 1 1325 91 1274 1 0338 NORMALIZED DERIVATIVES 0095 0 5 DERIVATIVE VALUES mA 0 PARAMETER 2 Beta LOOP 1 LOOP 2 LOOP 3 PARAMETER VALUES 2 2338 2 2492 2 6278 NORMALIZED DERIVATIVES 06 2 6 DERIVATIVE VALUES 3 95 0 x PARAMETER 3 Empl LOOP 1 LOOP 2 LOOP 3 PARAMETER VALUES 3 6737 3 6996 4 2891 NORMALIZED DERIVATIVES 0481 176
20. ck ck CK LOOP GRADIENT FIRST STEP SIZE CRITERION OUTER ITERATION 1 1 334563 2 01660100 67 82 2 18 49 33019340 63 88 3 16 62 1 08013100 48 28 ET PARAMETER VALU Check as NORMALIZED DERIVATIVI for EMPAL DERIVATIVE VA PARAMET BI PARAMETER VA NORMALIZED DERIVATIV DERIVATIVE VALU PARAMET 3 VAC PARAMETER VALU NORMALIZED 17217 DERIVATIVE VALU PARAMET R PARAMETER VALU NORMALIZED DERIVATIV DERIVATIVE VALU ET PA ETER VALU NORMALIZED DERIVATIV DERIVATIVE VALU PARAMET 8 UMI PARAMETER VALU NORMALIZED DERIVATIV DERIVATIVE VALU PARAMET 9 UI PARAMETER VALU NORMALIZED DERIVATIV DERIVATIVE VALU PARAMETER 10 LAG PARAMETER VALU NORMALIZED 17217 DERIVATIVE VALU APPARENT RIDGE BETWEEN PARAMETERS 7 25 OUTER ITERATION 20 k kk kk ck CK ck ck ck LOOP
21. ck ck ck ck Ck LMIHH UMIHH Are these correct ie 1 0 1 1 Do they match the EMPAL Se Bo d calibration output 100 22 116 154 1037 155 147 147 LOWS 67 131 T1545 10 107 C 000 10 BUNDNE 893 SDEV 45 314 53 49 7 22 ck ck CK TOTAL HOUSEHOLDS 1 1 KKK KKK ck ck Are these zonal total households correct for the lag year 1 2 3 4 5 6 7 8 9 0 MEAN 3233 SDEV 954 7 23 k ck ck ck IMPEDANCE DATA UPPER LEFT CORNER OF MATRIX FIRST 10 X 10 ZONES ko k kk ko
22. Households Worksheet Determine the number of households by income group previously defined by you that are located in each zone For the lag year you need only to specify the total number of households for each zone Employment 2000 IN 0 A e o e e e o v 1 Figure 3 TELUM DATAPREP Households Worksheet for Current Year Zonal Data Note The Household Percentages table is automatically calculated in TELUM DATAPREP Households 1995 Household Percentages 2000 0 0 7 7 97 0 0 00 oom om 0 0 0 0 Figure 4 Households Lag Year Figure 5 Household Percentages as Calculated by TELUM Note The Household Percentages table is automatically calculated in TELUM DATAPREP 4 11 Land Use Worksheet Each variable below defines the amount of area per zone occupied by each category Land Area per Land Use Figure 6 TELUM DATAPREP Land Use Worksheet Land Area per Land Use Percentages 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 Figure 7 Land Use Percentages as calculated by TELUM 4 12 Projec
23. Sk ko ko ko ko 7 14 SUMMARY OF RESULTS FROM GRADIENT SEARCH gt k ko kk ko ck ck ck ck CK SUMMARY OF CRITERION VALUES AT START OF EACH ITERATION Criterion values show how well the estimated value fits the observed value for this locator type Employment Type 1 by zone AANA 01 ho nr O o This shows the gradual movement of the raw unscaled goodness of fit measure towards zero over the twenty iterations of the parameter estimation calculations n hb nNerrrrrrr O 00 ENSS SUMMARY OF ITERATION COUNTS ES OF 5 PARAMETFE ART OF EACH ITERATION ITER 1 ITER 2 ITER 3 ITER 4 ITER 5 132500 033581 053970 002871 078082 For the 100167 229657 263458 289479 452845 advanced 483281 523777 616717 640374 779006 597689 623174 643204 721527 771924 user this tells 233800 628638 670289 891550 920331 more about 932977 940356 979645 985618 094076 ite 081080 096258 154366 168411 252357 721851 764874 770424 809712
24. Page 5 FORECAST YEAR 2005 OUTPUT DATA EMPLOYMENT TYPE Light In Heavy In Service Retail TOTAL 1 12 0 601 241 8554 2 79 28 83 103 294 3 CT Ss 9 9 40 4 177 3A T 3T 36 628 5 66 10 144 112 332 6 36 14 1535 89 291 7 T 2 25 354 63 8 Ba 3 45 68 119 9 9 Ss 1025 70 183 10 30 8 133 IX 288 TOTAL 430 450 1933395 880 3093 Page 6 8 19 BASE YEAR TO FORECAST YEAR PERCENT CHANGE 0 01 1 0
25. 4 10 LAND USE WORKSHEET ee i ee 4 11 PROJECTIONS WORKSHEET 1 1 1 rr ei ne 4 3 THE EMPLOYMENT TO HOUSEHOLDS CONVERSION MATRIX 4 3 5 DATA CHECK AND CONSISTENCY REPORT teas 4 6 6 APPENDIX FOR DATA PREPARATION 4 17 1 The TELUM Process The following flow chart outlines the general processes the user will follow in using TELUM This chapter tells how to prepare and compile the data required You will then calibrate the model using current and lag year data This process will provide the parameters statistically estimated equation coefficients which serve to fit the models to your data You will the use these parameters to assure accurate forecasting into future time periods Later you may develop policies that affect the final forecasts To evaluate these policies you may modify the data input variables or the parameters and you may impose constraints on household and employment locators Data Collection IDEU TEETE Travel Impedance TIPU Development of Regional Analysis of Parameters Forecasts of Employment and Goodness of Fit and Population Residual Errors DOPU Conversion Ratios Land Consumption Model Parameter Estimation Using Multiple Regression MCPU Development of Attractiveness Residual Files TELUM Res and TELUM Emp Parameter Estimation Using CALIBTEL Software MCPU Forecasting Process
26. 189 35 148 156 C 000 100 C THIS PRINTOUT OF A PORTION OF THE EMPLOYMENT INPUT IS TO VERIFY THAT THE CORRECT INPUT FILES WERE USED 7 21 KKK k ck ck RESIDENTIAL ATTRACTIVENESS VARIABLES CK ck ck ck ck kk kc KKK VACANT LAND ES LAND Are these correct O0 OPPOPPPPODO Q1 C0 r2 01 4S BNE NN MEAN 1 920 34 SDEV 0 048 1 k ck Ck ck HOUSEHOLDS BY TYPE kc k ko ko kk ck
27. ENT TYPE 1333 EMPLOYM 450 Page 10 430 1 2 3 4 5 6 7 8 9 0 1 TAL TO ko ko KKK Ck ck ck ck KKK ck ck ck ck CALCULATIONS CALCULATIONS I INDUSTRIAL FOR LAND USE COMMERCIAL FOR LAND USE Retail ED AS ED AS 2 Hind 4 ERE GROUP LInd ERE GROUP 3 Servi W TYPES ER RE ES W TYPES ER RE ES W ENT ES WE EGORI ENT ES WE EGORI PLOY E TYP E CAT PLOY E TYP E CAT ck Ck ck ck ck Ck ck ck KK ko KK EL OUTPUTS 828 ED ON MOD Page 11 NO CONSTRAINTS IMPOSI I ow I as a REGIONAL EMPLOYMENT HOUSEHOLD RECONCILIATION CHANGE IN HH DISTRIBUTION DUE TO CHANGE IN REG
28. 2005 Absolute Employee Constraint By Type and Zone your constraints in the table below COPY CONSTRAINTS FROM PRIOR YEAR 9 m Type 2 Total Constraints by Employment or Households The total constraint Type 2 directs the model to automatically populate the specified zone with your input constraint household or employment value For instance if a zone is constrained to a total value of 10 000 jobs TELUM forecasts 10 000 jobs to that zone If one employment or household type in a zone is constrained by a total value TELUM maintains the baseline proportion of employment types in that zone Enter total constraints in the same manner as shown for the absolute constraints in the worksheet screen shown above Type 3 Maximum Constraints A maximum constraint sets a maximum value of households or employment by type in a zone A maximum constraint operates only when the forecasted value for a specific zone and type exceeds your employment or household constraint input value for that forecast period An example of a Type III constraint occurs if high income households are constrained to a maximum of 2 000 in a given zone When and only when the maximum is reached during forecasting will this constraint affect the distribution and number of other households locating in this zone as well as the distribution of high income households in other zones in the region When the input value is exceeded the constraint is applied a
29. Integrated Urban Models 1983 Pion Limited London Chapter 6 It is important to note that in embedding the models into a streamlined user friendly interface a certain amount of user discretion had to be sacrificed TELUM MFCU is a linear modeling structure which for example a seasoned modeler may find limiting in that if a user decides it is necessary to rerun a forecast she must rerun all time periods In MFCU TELUM begins by running a set of Baseline BL forecasts This first forecast is performed using your DOPU data inputs TIPU impedance file and a set of statistical outputs the equation coefficients generated by TELUM during MCPU MFCU begins by telling you the name of your model forecast in subsequent forecasts you will assign your own name TELUM will then provide you with an opportunity to revise or add the following Employment and Population projections Add travel impedance files for future forecast time periods e Add constraints Change the number of time periods you wish to forecast for the region Next TELUM will begin the internal File Check in preparation for the model forecast Similar to MCPU begin Forecasting by clicking the GO button Once you click GO TELUM immediately begins forecasting your region s future employee and household locations for each forecast time period This process can take several minutes and should not be interrupted for any reason Once MFCU has completed the model for
30. Un RR G9 Maximum Absolute Deviation from Targets Industry 0 00E 00 Households 0 00E 00 Rescaling Factor 4DIV 0 Figure 9 TELUM DATAPREP Conversion Matrix Worksheet 1 5 Data Check and Consistency Report Once you have finished entering all of your region s zonal data inputs you must return to the first worksheet in your DOPU workbook labeled Data Check Data Check provides instruction on how to Run Data Check for your IDEU and DOPU zonal inputs Data Check completes a few statistical calculations that will tell you how strong the relation is between your data inputs TELUM reads these calculations and prepares a Data Consistency Report for you to review We strongly recommend you print this report and keep it on file for your agency 4 16 Appendix for Data Preparation Simple Numerical Example of the Conversion Procedure Consider the following numerical example First assume that we have a region where the reported employment in the region i e the number of jobs is 132 The number of households resident in the region is given as 70 Of the employed residents 90 work in the region Further there is net in commuting to work of 10 employees with 20 who live outside the region commuting in to work and 10 who live inside the region commuting out to work Finally there are five residents of the region who are unemployed but who would work if they had a job First we calculate the region
31. ck ck TOTAL REGIONAL PERCENT CHANGE Is this how you expected employment in your region to 027 028 023 071 grow or decline EMPLOYMENT TYPE Light In Heavy In Service Retail ZONAL MEAN ABSOLUTE PERCENT CHANGE EMPLOYMENT TYPE Light In Heavy In Service Retail TOTAL 134 388 122 108 088 NOTE THAT 0 01 IS 1 Page 7 8 21 TELUM RES Forecasting Output ELUM RES V3 10 HOUSEHOLD ALLOCATION MODEL VERSION OF 08 JUN 2003 COPYRIGHT S H PUTMAN ASSOCIATES 1989 2006 6 Dec 00 AT 4 17 34p This Copy Licensed for Use at Check H THIS RUN MADE ON Urban Simulation Laboratory Dept CRPlng Ck ck ck ck ck ck ck ck kk KKK KKK ko ck ck
32. 49 78 Figure 10 MAP IT Output 4 After you review Analysis of Forecast Spatial Patterns TELUM screen P7 8 1 prompts you to select one of the following tasks 1 Rerun a Model Forecast 2 View a prior Model Forecast 3 Runanew Model Forecast 4 Exit the TELUM system 3 Rerun a Model Forecast Upon completion of your baseline or policy forecast s you may elect to rerun the forecast with changes to the regional data inputs TELUM walks you through a series of information screens that are used to organize and store your next model forecast The following options are presented to screen P7 8 3 Change the original DOPU data set Change the regional control totals inside the DOPU Conversion Matrix Add or change a future year impedance file Change the regional employment and household projections NOTE When rerunning a policy forecast users may only select options C and D 3 15 Rerun the Rancho Carne Baseline Forecast The following charts provide the information and data you will need to provide TELUM for the Rerun of Rancho Carne s Baseline forecast Screen MFCU Rerun Variable Inputs Input Value P7 8 1 Forecasting Rerun Model Forecast P7 8 1A Rerun Model Forecast Baseline P7 8 1C Model Forecast File Storage OLD BL EMP Household P7 8 3 Do you wish to make changes Projections P7 6 Forecasting Time Periods Yes Enter new EMP Household P7 8 4 Projections See table below P7 13 Tr
33. 5 fe Jose Pi k Oc Bi Cij E Cij k jt where 7 7 E elasticity of type k employment to equal changes in all of the impedances for work to home trips originating in zone j calibrated TELUM EMP parameters for travel time and cij impedance between zones i and j The purpose of all this is to provide a means for assessing without the need for innumerable model runs the relative sensitivities of the various locators in the various zones to the different independent variables in the model structure This knowledge in turn provides a means for assessing the likely degree of impact of specific policy proposals on individual locator zone combinations 7 8 Sample Calibration Output Files TELUM EMP M MM EEEEE ITTT RRRRRR 0000 PPPPP 1 0 U SSSSS MMMM 3 1 0 OO P P I L U U S M M M EEEF RRRRRRR 0 OO PPPPPP I L U U SSSS M M E T R R 0 OO P I L U U S M M EEEEE T R RR 000 P I LLLLLL UUUUU 5 CCCCC A LL IIII BBBBBB CC 0 AA LL Lr BB B CC A A LL II BB B CC A A LL II BBBBBB CC AAAAAAA 11 BB B CC 6 A A LL II BB B CCCCC A A LLLLLLL IIII BBBBBB This Copy Licensed for Use at the Urban Simulation Laboratory Department of City and Regional Planning Check the type of University of Pennsylvania CALIB run CALIB V3 11 0 PROCEDURE VERSION OF 22 APR 1998 COPYRIGHT S H PUTMAN ASSOCIATES 1989 1998
34. EMP Location Elasticity Values pp 7 6 TELUM EMP Sample Calibration Outputs 7 8 8 Model Forecasting 1 TELUM Program Overview 4 Leer esee 8 1 2 Model Forecasting Unit Overview ecce ee ecce ee ee 8 2 3 The Baseline 5 8 3 Running a Baseline Model Forecast pp 8 3 Re run a Baseline Model Forecast nn 8 4 4 Running a New Model Forecast e eee eee eee eee eere eee ee ense eate setae sten ae enun 8 4 How to Re run a New Baseline Model Forecast en 8 5 Changing your Model Forecast Inputs 8 5 Making Use of Unobserved Factors in Forecasting 8 6 1 1 12 2 1 12 2 2 2 2 12 2 2 2 2 2 2 2 212 12 2 2 2 2 2 2 2 eos sost 8 7 The Employment and Household Location Model Formulations 8 7 The Residential Location Model TELUHMIRES 8 7 Review of the TELUM Model Forecasting Process pp 8 9 Model Forecasting Output Files pp 8 9 TELUM EMP Sample Forecasting Outputs sss 8 10 1 Introduction TELUM Welcome to TELUM This manual is intended to assist a new user of TELUM in learning to make efficient use of its many capabilities The manual contains a brief introduction to the history of the system along with detailed instructions for using TELUM In order to obtain useful and correct forecast results and in order to mini
35. Edit Session Optional 6 4 his step is optional Calculating the centroid value in an edit session allows you to undo your changes while doing so outside of an edit session is un doable To start an edit session from the editor toolbar click on the drop down menu Editor and select Start Editing 3 Calculate the X Values From the attribute table window in your shapefile right click on the heading for the field you just created XCoord In the menu revealed by the right click select Calculate Values this will bring up the Field Calculator window Click the advanced checkbox in the middle of the Field Calculator window In the first text box labeled Pre Logic VBA Script Code type Dim dblX As Double Dim pArea As IArea Set pArea Shape dblX pArea Centroid X Inthe second text box labeled Calc type dblX click OK 4 Calculate the Y Values The calculation of the Y values is the same as calculating the X values except that you use the field you made YCoord and in the Pre Logic VBA Script Code box you use the following script which merely replaces X with Y Dim dblY As Double Dim pArea As IArea Set pArea Shape dblY pArea Centroid Y Inthe second text box labeled Calc type dblY Click OK You will now have the X Y values for the centroid as fields in your shapefile These can be used along with the Pythagorean theorem
36. The purpose of multiplying by 1 0 divided by 1 0 minus the unemployment rate is to adjust back up the numbers of employees to account for the unemployed who it is assumed will have made location decision while employed and will therefore have located as others in their income when employed class l Methods for Calculation of the TELUM Res Conversion Ratios Estimates of the conversion ratios can be calculated from data collected by regional planning agencies and the U S Bureau of the Census The number of jobs per employee RJPE regional net commuting rate RNCR unemployment rates in each industry UNEMP and the number of employees per household by income group EMPHH can usually be estimated from regional planning agency surveys The employee to household conversion matrix CNV j is calculated by Computing the number of heads of household by income group employed in each industry Entering the results from step 1 into the conversion matrix spreadsheet in DATAPREP XLS along with estimates of the number of jobs per employee RJPE the regional net commuting rate RNCR unemployment rates in each industry UNEMP the number of employees per household by income group EMPHH the number of households by income group HHi and the number of employees by industry EMP 1 2 J Brugger Oct 25 C jeannette s intheworks 4edit_dataprep doc 4 18 5 MAP IT 1 What you can do with MAP IT With the MAP IT f
37. 0 010100 3 1 Install 7 17 3 1 Enter TELUM username and password pp 3 2 3 2 Review and Enter Regional Data pp 32 3 Data Organization and Preparation Unit and 1 4 1 7 ese 3 5 Work in the DOPU Workbook 3 5 Enter your Rancho Carne DOPU Data sse 3 5 Check your Data Consistency Results pp 3 7 MAP IT Check zones and Data check en 3 8 Launch MAPLE sa a eet tee n edere 3 8 4 Travel Impedance Preparation Unit eee ecce eee esee eere 3 9 TIPU Impedance Data Inputs em a em a Eiaeia 3 9 5 Model Calibration and Preparation Uniit sccsscssssscssssccsssscssssssssccssscees 3 10 Run Employment and Household Model Calibrations sse 3 10 Review the Calibration Analysis of Results and MAP IT sss 3 11 Run a Land Consumption Regression Model pp 3 12 Explanation of LANCON findings pp 3 14 6 Model Forecasting Unit 1 1 1 1 1 1 1 1 1 1 1 seen 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 esten aee 3 14 Prepare Files for a Forecast Model ee 3 14 Run a Baseline Model Forecast pp 3 14 Rerun a Model Forecast Ne 3 15 Run a Policy Model Forecast pp 3 16 MAP IT Map and Compare Forecast Results 3 17 4 Data Preparation 1 The TELUM ProGGSS td 4 2 2 Data Requirements for Agency Applications ecce esee eere eren ee enu 4 3 Region Lev
38. 0954 8156 4167 e 635 41 O hb TOTAL 29 0268 16 8208 13 MEAN 2 2348 1 7778 STD DEV 5320 4290 gt ko ko kk ko ck ck CK ELASTICITIES REGIONAL LOCATION ASTICITY IS FOR A CHANGE IN THE LISTED VARIABLE OF 1 00 THE LOCATOR WOULD INCREASE OR DECREASE BY TH THEN FOR A E FOR THAT ZONE THERE HAT LOCATOR IN THA Just as it n ELAS ICITY Says V 1 00 E AVERAGE SENSITIVITY OF HE SPECIFIC VARIABL me AVERAGE STANDARD ELASTICITY DEVIATION IMP 2 2348 5320 Empl 1 77178 4290 Land 8156 4167 I UAL 7 17 HE INTERPRETATION OF A LOCATIO AMOUNT IF ELASTICITY 0 4316 INCREASE IN THAT SPECIFIC VARIAB WOULD BE A 0 4316 DECREASE IN T ZONE ALL OTHER THINGS BEING EQ
39. 2 Model Forecasting Unit Overview This chapter discusses the process of model forecasting with TELUM The formulations of the models TELUM EMP and TELUM RES are discussed in detail in the chapter Appendix In this section you will learn how to run a Baseline model forecast and new Policy forecasts rerun your model forecasts and include aspects of local planning knowledge into your forecasts Analysis of phenomena as complex as the location of jobs and people in a large region requires complex tools TELUM contains a formal structure consisting of two modified versions of singly constrained spatial interaction models referred to here as TELUM EMP and TELUM RES These are followed by a multiple regression model which reconciles the competing employment and household demand for land The whole set of models run behind the TELUM interface The overall structure of the TELUM model forecasting process is rather straightforward Beginning with regional trends transportation facility descriptions and data on the current location of employment jobs population and households TELUM can forecast future location of jobs and households The forecasts are done in five year steps For example the 2000 data becomes the input for the next forecast 2005 and in turn 2005 becomes the input to 2010 The equation structure of the models is complex It is briefly described in the Appendix to this chapter The models are described in more detail in Putman S
40. 21 REGIONAL PERCENT CHANGE IN LAND USE TOTAL LAND AREA 00 UNUSABLE LAND 00 STREETS AND HIGHWAY 00 BASIC LAND T0 315 COMMERCIAL LAND 3 52 RESIDENTIAL LAND 54 32 VACANT DEVELOPABLE 23 14 TOTAL DEVELOPABLE 00 GINI COEFFICIENTS OF SPATIAL DISPERSION ACTIVITY BASE YEAR FORECAST YEAR LI 75235 571 LM 494 4923 MH 439 436 HH 546 2932 TOTAL 428 459 NOTE THAT 0 00 IS TOTALLY DISPERSED 1 00 TOTALLY CONCENTRATED Page 17 8 33 References Alonso W 1964 Location and Land Use Harvard University Press Cambridge Massachusetts Anas A 1982 Residential Location Markets and Urban Transportation Economic Theory Econometrics and Public Policy Analysis Academic Press New York Anas A and C Chu 1984 Discrete Choice
41. 4 Travel Impedance Preparation Unit In this section you will prepare a Travel Impedance File The term impedance refers to the travel time travel cost or a composite of both calculated by travel models not included in TELUM to describe differences in zone to zone difficulty of interaction An impedance file would typically be available from your agency s transportation department TELUM provides steps and examples for converting an impedance file for your TELUM project For the Rancho Carne project an impedance file is provided in your Tutorial folder labeled IMPD txt Place this IMPD txt file inside your CATELUMWDATA folder if it is not already there Instructions for creating an IMPD txt file with your regional data are provided in Chapter 4 Data Preparation TIPU Impedance Data Inputs The Travel Impedance Preparation Unit TIPU runs a data validation procedure on the contents of your IMPD txt file In order to run this validation you must enter the following data into TELUM as you work through the TIPU section 39 Screen TIPU Data Variable Input Value P5 30 3 Average Impedance 87 P5 30 4 Smallest Impedance 7 P5 30 5 Largest Impedance 197 P5 30 6 Top 4x4 10 45 1 87 45 7 30 42 31 30 8 67 87 42 67 23 P5 30 7 Bottom 4x4 14 61 39 28 61 14 34 33 39 34 28 21 28 33 21 29 When you have completed these steps correctly TELUM will search for the impedance file verify i
42. 6 2 0 al 14 0 8 16 0 9 11 0 10 12 0 TOTAL 90 BASE YEAR 2000 INPUT DATA OTHER LAND USE TOT AREA UNUSBLE STREETS VACANT 5 VACANT DEVELPBL DVPL DEV T 2 0 0 Tz 50 00 2 50 00 2 5s 0 0 T 20 00 5 80 00 3 6 0 0 2 30 0 6 69 60 4 20 0 0 Jis 3 19 20 96 81 5 7 0 0 1 14 29 7 85 71 6 3 0 0 T 33 33 3 66 67 7 15 0 0 Ts 6 67 LS 932793 8 17 0 0 Tt 5 88 17 94 12 9 T2 0 0 i 8 33 22 91 67 10 d 3 0 0 T 7 269 13s 92 31 TOTL 100 0 0 10 100 Page 8 IMPEDANCE MATRIX UPPER LEFT 13 13 ENTRIES
43. AT E ON Use at Dept Page 1 C COPYRIGHT S H P THIS RUN MADI This Copy Licensed for Urban Simulation Laboratory Always check the date and title to be sure this is the run you think it is 12 06 00 2000 2005 MyCity Baseline Do these match calibration values Are these the correct new regional forecast values EMPLOYMENT ALLOCATION PARAMETERS EMPL ATTR 213 AlI 271 038 TOTALS EL OUTPUTS Retail 880 Page 2 8 17 EGIONAL CONTROL TYPI Light In Heavy In Service Retail Light In Heavy In Service 430 450 1333 NO CONSTRAINTS IMPOSED ON MOD
44. Aki W Ci jt 1 AX EE 1 1 where Wie EL L 2 and 4 Atas Y grL stes 3 i where EF employment place of work of type k in zone j at time t 1 EF employment place of work of type k in zone j at time t L total area of zone j cijt impedance travel time or cost between z ones i and j at time t Pit1 total number of households in zone i at time t 1 A5 B k ak bx empirically derived parameters The Residential Location model TELUM RES TELUM RES is an aggregate form of a multinominal logit model of location choice When translated into computational form this yields a modified version of a singly constrained spatial interaction model There are two major modifications 1 a multivariate multiparametric attractiveness function is used 2 a consistent balanced constraint procedure is included in the model allowing zone and or sector specific constraints The multivariate zonal attractiveness term enables the inclusion of knowledgeable professionals input to the model structure in a consistent and replicable fashion The model is normally used for 3 5 the current maximum is 8 household categories whose parameters are individually estimated The model is described in more detail in Putman 1983 1991 as well as in numerous journal articles and reports For reference the equation structure is given here N n 2 QIB Wch exp B c 1 0 TIN 4 8 14 where 0 a E
45. Data Check 3 6 The Run Data Check Results button inside the Data Check worksheet will then be enabled when you click this button TELUM will then run a consistency check of your data In this report TELUM will highlight values that are not within the predefined ranges considered normal We strongly recommended that you review these reports and pay special attention to any highlighted values Check Your Data Consistency Results The following chart contains the values we expect you will see as you view your DOPU Data Consistency Check for Rancho Carne Value 0 786 0 924 0 186 0 997 0 830 0 946 0 921 0 967 0 926 2 5 1 8 1 11 0 11 5 2 6 0 Data Variable Correlations Total Current vs Lag Household Correlation Total Current vs Lag Employment Correlation Total Household vs Employment Correlation Total Household vs Population Correlation Employment Correlations AGR vs AGR LMFG vs LMFG PROF vs PROF RTL vs RTL FIRE vs FIRE Consistency Check of Regional Ratios Population per Household Population per Employment Percentage Change Households Employment Population Employment Current to Forecast Land Use Check P4 30 Screen P4 30 1 P4 30 3 P4 30 2 If consistency results in the tutorial or later with your regional data show red text or a warning this indicates low or high correlations or percentage change in parameters If this occurs you should d
46. FOR LAND CONSUMPTION EQUATIONS HI HH TOTLND 88820 0 ES LI HH 18870 BAS L 53870 COM L 41490 COMM L 0 BAS E 2 21350 EM COM 50120 BAS 2 05130 Page 2 8 23 DVL DV 18930 DVL DV 8 82340 DVL DV 2 86670 INTRCEPT 1 12010 RCEP 670 IN 1 INTRCEP 2 87890 gt ck Ce ck kk ENTIAL BASIC ERCIAL REGIONAL RATIOS EMPLOYMENT HOUSEHOLDS ETC EMPLOYMENT TYPE PERCENT UNEMPLOYED ck ck ck kk ck LInd 000 Hind 000 Servi 000 Retail 000 HOUSEHOLD TYPE EMPLOYEES PER HOUSEHOLD
47. MFCU 4 2 2 Data Requirements for TELUM Data for use in transportation location and land use models are required at different levels of spatial sectoral and temporal detail It is useful to distinguish between what information is needed about the overall modeling region and the zone specific information that 1s necessary for detailed spatial representation and analysis Region as it is used in TELUM means the geographic area you are modeling The region is defined by the Metropolitan Planning Organization MPO and may be a single county or an aggregate of multiple counties or parts of counties The following discussion is divided into region level data requirements and spatially disaggregated zonal data requirements Region Level Requirements There are three groups of data required at the region level First are the model parameters that are derived from a statistical analysis of regional patterns the process of calibration or of fitting the model to the data These parameters must be estimated prior to using the models for forecasting or policy analysis The second group of regional inputs is the regional ratios These include unemployment rates jobs per employee employees per household persons per household and other similar statistics The third group of inputs for the total region is the regional forecasts Here it is necessary to develop estimates or to obtain them from public or private sources of regional totals of popula
48. Models and the Housing Price and Travel to Work Elasticities of Location Demand Journal of Urban Economics 15 107 123 Ben Akiva M and S R Lerman 1985 Discrete Choice Analysis Theory and Application to Travel Demand MIT Press Cambridge Massachusetts Cochrane R A 1975 A Possible Economic Basis for the Gravity Model Journal of Transport Economics and Policy 9 34 49 Freeman A M 1993 The Measurement of Environmental and Resource Values Theory and Methods Resources for the Future Washington D C Martinez F 1992 The bid choice land use model an integrated economic framework Environment and Planning A 24 871 885 McFadden D 1974 Conditional Logit Analysis of Qualitative Choice Behavior in Frontiers in Econometrics P Zarembka ed Academic Press New York 105 142 McFadden _D 1978 Modelling the Choice of Residential Location in MU Interaction Theory and Planning Models A Karlqvist L Lundqvist F Snickars J Weibull eds North Holland Amsterdam 75 96 Neuburger H 1971 User Benefit in the Evaluation of Transport and Land Use Plans Journal of Transport Economics and Policy 5 52 75 Putman S 1983 ntegrated Urban Models Policy Analysis of Transportation and Land Use Pion London Putman S 1991 Integrated Urban Models 2 New Research and Applications of Optimization and Dynamics Pion London Shen P N 1995 Optimized Network Equilibrium Models of Combined Tra
49. Transportation and Land Use Modeling With Federal Highway Administration sponsorship in 1971 Professor Stephen H Putman began the development of what became known as the Integrated Transportation and Land Use Package ITLUP The specific intent of that package was to provide a means to properly represent the interrelationships of transportation and land use The original package developed in a university setting was able to demonstrate the general importance of these linkages which were previously overlooked in transportation policy analyses Another result of that work was the inclusion in the early 1980s of a portion of the package as part of the final release of the Federal Highway Administration s Urban Transportation Planning System UTPS software package In succeeding years extensive upgrades revisions and modifications were made to all portions of ITLUP All or portions of the integrated transportation and land use package have been applied in more than twenty different metropolitan areas in the United States as well as in cities overseas In the thirty years that have passed since the first work on ITLUP there has been a transformation in computer technology that was quite unimaginable in 1971 Perhaps nowhere is this more true than in the development and adoption for agency use of Geographic Information Systems GIS In the summer of 1997 METROPILUS a new land use modeling system was first applied in an operating agency Th
50. be accomplished by construction or modification of one sort or another After these links have been identified construction projects could be described and budgeted and the analysis would be completed in the form of a set of recommendations as to places where the network could be improved The defect in this procedure is that the congestion that results from the initial estimates of trip makers and thereby from the initial estimates of the locations of employment and households would in and of itself result over a long term span of years since forecasts traditionally are concerned with a re arrangement of the locations of employment and households Thus in order to properly estimate the congestion it is in effect necessary to know the congestion In order to properly know the congestion it is necessary to know the location of employment and population and the resulting demand for trip flow on the network and so on and so forth This is a classical example of a system that can only be properly analyzed by use of an interactive technique that includes both the direct and the indirect connections or as is sometimes described both the feed forward and the feed back connections amongst the elements of the system A complementary system to this one is traditional land use or urban design analysis In such a case descriptions of the transportation system which may include highway as well as transit are taken from exogenous sources That is to say so
51. land is defined as AU TAA DEV VAC Translating From Local Land Use Inventory to TELUM Land Use In many cases the land use inventories available to the planning agency will not have their land use categorized in the same terms as those described above This means that it will be necessary to use professional judgment to translate the data from the local land use inventory categories to your TELUM categories As an example the following definitions were used in work for the Houston region Total Area Land Only 1 6 no water Unusable Usable Unusable Land Parks Environmentally Constrained Usable Land Vacant Developable Developed Vacant Developable Total Vacant Land Environmentally Constrained Developed Commercial Residential Basic Streets and Highways Commercial Employment Land Retail and Office Employment Land Use Residential Land All housing types Basic Employment Land Industrial and Institutional Employment Land Use Streets and Highways Land Rights of Way 4 Data Organization and Preparation Unit DOPU Once you have completed the Initial Data Entry Unit IDEU you are ready to input your zonal data into the workbook provided in TELUM Data Organization and Preparation Unit DOPU DATAPREP Data may be transferred into DATAPREP from other worksheets Only values are permitted in DATAPREP Copy data into DATAPREP through Edit gt Paste Special Values or Text Note You must open the DATAPREP E
52. periods We recommend that you run your first baseline forecast with the data inputs you have provided in IDEU DOPU and TIPU If you don t agree with this first set of baseline forecasts you can always return to MFCU to rerun the baseline with any changes in data How to Run a Baseline Forecast Remember a Baseline forecast can only be made after you have successfully completed a model calibration run in MCPU This is because forecasting requires both the equation coefficients calculated by CALIB and the calibration residuals Use the following steps as a guide 1 When you enter MFCU the system will display BL as your first model forecast As you continue TELUM will ask you to change or add data as described above 2 Soon you will arrive at a screen asking you to Run a Model Forecast Click GO 3 TELUM will announce when it has completed your Baseline forecast Immediately following your first forecast your Spatial Analysis reports appear Carefully review your reports as well as your maps in MAP IT Here you will decide whether to keep your Baseline forecasts or rerun your baseline before proceeding to a new model forecast 8 3 Re Run a Baseline Model Forecast At times users tend not to agree with their Baseline forecasts Typically a user will find that they must make changes to the data In this case users may rerun their Baseline forecast A rerun is required when any of your starting data inputs change A u
53. question or concern in a timely manner After you have completed this tutorial you will be better prepared to build and run a TELUM project for your region 2 Initial Data Entry Unit IDEU In this section you will do the following 1 Install TELUM 2 Enter your TELUM username and password 3 Start your TELUM project 4 Review and enter regional data 1 Install TELUM Begin by performing a TELUM installation as described in Chapter Two of this manual 2 Enter TELUM Username and Password Once you have completed the installation process a TELUM icon will appear on your computer s desktop Double click on the TELUM icon Upon entering TELUM the program will prompt you to enter your username and password To obtain the username and password please contact TELUM user support or FHWA Resource Center please refer to Chapter 2 Section 4 User Support for contact information 3 Start your TELUM project After a series of welcome screens the opening screen appears with a number of colorful buttons These buttons later allow you to skip directly to different sections of the TELUM system You will only use these buttons later when you click EXIT in TELUM and subsequently wish to return to the section in which you stopped working However at this point your first time through you must begin by clicking on CONTINUE to enter the Initial Data Entry Unit IDEU 4 Review and Enter Regional Data IDEU is where you provid
54. the model estimate and calculates both absolute and percent differences Ck ck ck ck ck Ck ko ko k ko ko KKK SUMMARY OF CALIBRATION RESIDUALS ko k kk ko Ck ck ck ck CK ZONE OBSERVED ESTIMATED RESIDUAL PCT DIF 1 51 00 38 85 12 15 23 8267 2 31 00 33 33 3 55 3 127 00 165 83 3 30 5786 4 77 00 79 98 2 98 3 8729 5 107 00 113 87 6 87 6 4198 6 66 00 85 88 19 88 30 1164 1 152 00 123 48 28 52 18 7634 8 119 00 85 57 33 43 28 0933 9 109 00 107 18 1 82 1 6743 10 11 00 16 03 5 03 45 7355 EAN ABSOLUTE PERCENT ERROR 19 661 ZONES WITH 0 OR 1 OBSERVED ARE OMITTED These are alternative measures of goodness INIMUM OBSERVED LOCATOR VALUE 11 of fit AXIMUM OBSERVED LOCATOR VALUE 152 E FOR 3 SMALLEST 25 OF ZONES 26 2528 ESE ZONES HAVE 10 94 OF THE REGION TOTAL E FOR 3 LARGEST 25 OF ZONES 25 8578 ESE ZONES HAVE 46 82 OF THE REGION TOTAL THE SMALLEST ZONES ZONES W
55. to calculate the distances between the points 6 5 7 Model Calibration 1 Introduction to Model Calibration Each of your regional locator types will exhibit a different locating behavior in each region A particular locator type such as High Income Households may well exhibit different locating behavior in different regions It is therefore necessary to estimate the equation coefficients of the model equations for each locator type in your region The process of estimating these equation parameters is called model calibration TELUM performs a model calibration for each locator type in your region Due to the nonlinear structures of the TELUM forecasting model equations it is necessary to use specialized parameter estimation procedures for calibration The goal of calibration is to develop estimates of the parameters of a model s equations s which best fit the general model structure to your region s specific data set Most planners are familiar with this process in the context of multiple linear regression analysis TELUM s model calibration is analogous to regression analysis but uses different mathematics and a different computer program The computer program used here TELUM CALIB locates the optimum best fit parameter values by a method called gradient search This section provides an overview of the TELUM Model Calibration Preparation Unit MCPU is a small but critical unit in the TELUM system This unit provides three layers o
56. to calculate the next direction of ascent producing a new set of steps in a direction that 15 tangential to the previous zig zag step pattern This substantially increases the efficiency of the search process The top of the mountain is the point at which we will have found the values of the parameters to best fit the model to the eauations We will never get a perfect fit The values of best worst OUTER ITERATION LOOP GRADIENT FIRST STEP SIZE CRITERION OUTE ERATION 20 1 1 04110083 18 18 2 9 03014674 18 18 3 1 38 01877115 18 17 PARAMETER 1 Alpha PARAMETER VALUES NORMALIZED DERIVATIVES DERIVATIVE VALUES PARAMETER 2 Bet PARAMETER VALUE NORMALIZED DERIVATIVE DERIVATIVE VALUE PARAMETER 3 Empl PARAMETER VALUE NORMALIZED DERIVATIVE DERIVATIVE VALUE ER 4 Land PARAMETER VALUES NORMALIZED DERIVATIVES DERIVATIVE VALUES ER 5 Lambda PARAMETER VALUES NORMALIZED DERIVATIVE This 1s the final iteration ck ck ck kk
57. traffic congestion and transportation efficiency resulting from changes in urban design and land development patterns in combination with variations in socio economic factors Decades of transportation and land use studies of every sort have shown us clearly that there are relationships between transportation and land use or land development However if we look over all these many studies it is sometimes very difficult to understand how the varying results can be considered logically consistent One important result from the integrated transportation and land use package development was that its overall construct provided a clear way to see that often the apparently conflicting results from transportation and land use studies were in fact conflicting only because of the way in which they were initially viewed The most obvious example is in some of the traditional approaches to solving local congestion problems In such cases a study will be done of a physical transportation facility and a need will be defined for increased capacity of one sort or another on the network This capacity an additional highway lane a new road etc will be constructed and will result in the short term in an improvement of vehicle flow and a reduction in the observed congestion Unfortunately in the long term such strategies often have just the opposite result The increased network capacity is used by trip makers to make more trips and or longer trips Thus in the l
58. user documentation or appropriate technical support personnel in your organization 1 3 The following sources of information are designed to help you successfully navigate the TELUM system 1 Tutorial On your TELUM CD ROM you will find a folder labeled TUTORIAL containing all the files you will need to produce forecasts for a sample 20 zone region in TELUM We strongly recommend you use the Tutorial to become more familiar with the system and its data requirements A guide to the Tutorial is provided in this manual in chapter 3 2 Hyperlinks Hyperlinks are blue underline texts that provide additional ways to present more information about TELUM features Once you have clicked a hyperlink TELUM will automatically turn the underline text purple to indicate that you have already visited the text link Print buttons are often provided with the hyperlinks in case you need a printed copy of the information for use at a later date 3 MouseOvers MouseOvers are tips that pop up when the mouse pointer is paused over a text feature Currently this feature is only used in the TELUM MAIN screen and MAP IT Ea Tetum 1 5 xl TELUM Main You currently have a TELUM project started SA 283 Where would you like to go next Initial Data Entry Unit IDEU Gathers basic information the TELUM system will need to organize your region s data Data Organization and Preparation Unit DOPU Travel Impedance Preparation Unit TIPU Model Calibr
59. 10 Your new policy run name is FC2 Forecast 2 The following charts provide the information and data you will need to complete this New forecast 3 16 Screen MFCU New Run Variable Inputs Input Value P7 8 1B New Forecast Name FC2 P7 8 3 Do you wish to make changes Add Impedance file P7 15 Travel Impedance Data 1 2010 P7 16 Move your IMPD10 txt file into the TELUM DATA folder P7 19 Impedance File values see table below P7 6 Forecasting Time Periods Yes 7 8 44 Enter new EMP Household Projections No Page Future Year IMPD10 Input Value P7 19 Average Impedance 82 Smallest Impedance 7 Largest Impedance 197 Top 4x4 10 45 1 87 45 7 30 42 31 30 8 67 87 42 67 23 Bottom 4x4 98 427 273 28 427 98 238 33 273 238 196 21 28 33 21 29 5 MAP IT Map and Compare Forecast Results With a Baseline and FC2 forecast completed you can launch the Forecasting MAP IT if you have ArcView capabilities MAP IT provides a variety of Mapping Options to help you visually interpret the spatial changes in your region s employment and household location The Mapping Options in MAP IT Forecasting include Zonal Forecasts the zonal value for each employment and household category present Calculated Zonal Differences the zonal growth decline of each employment and household category between two time periods and or forecast runs Calculate
60. 3 4 DERIVATIVE VALUES 4 7 0 PARAMETER 4 Land LOOP 1 LOOP 2 LOOP 3 PARAMETER VALUES 5 4202 5 3928 4 8743 NORMALIZED DERIVATIVES 410507 0002 DERIVATIVE VALUES 20 1 4 1 0 ER 5 Lambda ER VALUES RIVATIVES VALUE PARAME NORMALIZED DE DERIVATIVE This is a record of the calculations from the first iteration of the statistical procedure The procedure used for these calculations is called gradient search It works by calculating the derivatives a term from calculus of the criterion goodness of fit with respect to each parameter in the model equation A positive derivative value means that the parameter should be increased in the next calculation a negative value means that the parameter should be decreased in the next calculation The normalized derivative values indicate the relative importance of each parameter at that step of the calculations In this case lambda dominates all three iterations loops 7 12 Note Iterations 2 through 18 have been omitted This 18 the next to last iteration Note The criterion is closer to zero than it was at the start and the derivative values are also closer 10 zero OUTER ITERATION 19 o
61. 3 469181 387740 380421 360463 358949 341801 289722 234142 221865 216898 196363 PARAMETER 4 531400 522979 506926 484966 391446 This is a 312310 295433 262035 259717 259604 f 260107 267642 267803 267786 2 SUM 273778 281328 281168 281120 281832 the changing parameter values 7 27 PARAMETER 5 1 093800 789592 685564 762948 1 083357 1 281465 1 399435 1 685125 1 715214 1 704325 1 683857 1 632733 1 627708 1 629245 1 630114 1 568497 1 512721 1 513424 1 510949 1 494942 PARAMETER 6 10 754200 10 952020 11 065630 11 094220 11 203920 11 239230 11 213190 11 136970 11 089880 11 052950 10 878330 10 712470 10 664970 10 660850 10 608870 10 385180 10 097560 10 085070 10 059680 9 998181 PARAMETER 7 789200 809661 804563 115293 652908 594173 598858 616039 650503 683216 829563 969021 1 009469 1 012404 1 054193 1 251672 1 507669 1 520983 1 548171 1 615023 PARAMETER 8 052900 051780 7 1 027421 171774 218433 328122 363993 390786 519930 646869 684789 688113 729788 911153 1 150730 1 163430 1 188834 1 250084 PARAMETER 9 3 111500 3 229865 3 295389 3 307671 3 361204 3 426301 3 433320 3 441918 3 445487 3 451411 3 475634 3 502177 3 510425 3 510799 2 3 562999 3 624335 3 629215 3 638989 3 663047 PARAMETER1O 580000 573710 510216 578353 999995 4853
62. 4 3 The installation package will then prompt you to select Full or Patch installation For a new TELUM installation you must select Full A Patch is only used for system upgrades TELUM xi Select Components O na plee mt 38644 k low sel e checkboxes for the Patch SUL options that you would like to have installed The disk space fields reflect the requirements of the options you have selected Disk Space Required 38644 k Disk Space Remaining 4082750 k Wise Installation Wizard Cancel 4 Installation will be completed a few seconds after your component selection is made 5 Your project filename will be by default TELUM EXE After installation a shortcut to TELUM will be added to your computer as a Desktop icon Your system will reboot automatically unless you are using a Windows XP operating system in which case a reboot is unnecessary 2 Configuring TELUM To open your project go to the Desktop and locate the TELUM icon for your project When you double click the TELUM icon you will see this screen 2 5 Username Settings When you open your TELUM project for the first time you will be prompted to create a username and password At this time this function is set with a default username and password When prompted enter you will need to enter username and password in order to start the application To obtain the username and password please contac
63. 5 k and Bi x Wei ewe 6 and y w amp 5 Dm D where E employment of type k place of work in zone j Ni households of type n residing in zone i total households residing in zone i at time 1 Li vacant developable land in zone i x 1 0 plus the percentage of developable land already developed in zone i Li residential land in zone i akn regional coefficient of type n households per type k employee impedance travel time or cost between zones i and j 7 B q r s b empirically derived parameters In the original formulation of TELUM RES all variables had the same time subscript Beginning in early 1994 with the more general availability amongst agencies of the necessary data several new formulations were examined in an attempt to include a lag term and thus increase forecast reliability which resulted in the current form of TELUM RES 8 15 Review of the TELUM Model Forecasting Process Each five year forecast step begins with the execution of TELUM EMP The model is normally used for 4 to 8 employment sectors whose parameters are individually estimated To forecast the location of employment of type k in zone j at time t 1 TELUM EMP uses the following input variables e employment of type k in all zones at time t e population of all types in all zones at time t e total area per zone for all zones e zone to zone travel cost or time between zone j and a
64. 6 400 ESTIMAT 63 24 VS 3 10 B 1 0710 4076 T FOR B 2 12 988 T FOR A 933 45 716 2499 4 283 066 849 4532 F PLOT OF OBSERVED 15 7 19 EST F 50 46 DF ERR B ERR A D ll 85 73 61 49 37 24 12 Sample Calibration Output Files TELUM RES M MM EEEEE ITTT RRRRRR 0000 PPPPP I L U U SSSSS E T R R 0 00 P P I L U U S M M M EEEE RRRRRRR 0 OO PPPPPP I L U U SSSS M M E T R R 0 OO P Ll U U S M M EEEEE T R RR 000 P I LLLLLL UUUUU SSSSS CCCCC A LL IIII BBBBBB CC C AA LL LI BB B CC A A LL LI BB B CC A A LL 11 BBBBBB CC AAAAAAA II BB B CC 0 A A LL Lr BB B CCCCE A A LLLLLLL 11111 BBBBBB This Copy Licensed for Use at the Urban Simulation Laboratory Department of City and Regional Planning University of Pennsylvania Check the type of CALIB run CALIB V3 11 DRAM DALISRATION PROCEDURE VERSION OF 22 APR 1998 COPYRIGHT S H PUTMAN ASSOCIATES 1989 1998 gt ko kk ko ck ck ck ck ck Ck Ck 41 THIS RUN MADI
65. 82 634541 729087 744735 760461 785842 793476 807624 801687 768293 746164 795101 818305 834210 808646 FINAL CRITERION VALUE 19 3167 FINAL VALUES PARAMETER DERIVATIVE ALPHA 8 716844 1 1 BETA 2 381420 2 1 VACDEV 188751 0 PERDEV 282303 al RESLND 1 496458 E LIHH 9 965671 LMIHH 1651311 Ll UMIHH 1 284517 ns UIHH 3 677471 6 LAGHH 842352 2 7 EXIT ON CONDITION 0 NORMAL TERMINATION If we know where people work and we know where people live then implicitly we know what length work to employment trips are being made 7 28 Time or distance IMPLIED TRIP FREQUENCY DISTRIBUTION STATISTICS barean dhe ie NOTE TIMES ARE IN MINUTES zones that are ck kk farthest apart KKK KKK KK KKK KKK KK KKK MAXIMUM OBSERVED IMPEDANCE VALUE 109 9000 AVERAGE ZONE TO ZONE IMPEDANCE INCL
66. 845478 procedure 673700 288951 665297 538830 534572 results 537230 583677 586141 600899 671438 204462 728042 784080 198553 884093 375297 374974 385520 424357 438279 420200 874246 719082 794995 968087 010081 201157 255280 279868 445621 m 454069 477771 533909 546471 617007 This is a 950011 967201 973132 001278 019510 summary of 990000 334169 630674 588068 538679 the changing 600212 591147 579781 596933 577096 599290 575865 595296 587496 585361 parameter 579696 591322 580765 574838 581872 values 5 FINAL CRITERION VALUE 18 1547 FINAL VALUES PARAMETER DERIVATIVE Alpha 2 854080 2 Beta 4 875161 2 Empl 7 478262 1 Land 4 042548 0 Lambda 588547 1 9 EXIT ON CONDITION 0 NORMAL TERMINATION ko kk Ck ck Ck ck ck Ck Ck SUMMARY OF COEFFICIENTS AND SIGNIFICANCE TESTS
67. Categories PROF RTL Household Data Available Current Year by Type O by Total O None Lag Year by Type O by Total O None Number of Household Household Categories LI Categories MI UMI Total Land Area of the Region Land Use Data Available for Your Project Street Vacant Usable Unusable Basic Commercial Residential Highway Developable Land Forecast Time Periods 6 Empl to HH Conversion PUMS Ratio O Default Ratio Empl per HH by Income Empl per HH Ratio 0 Default Ratio Unemployment Rate UR Ratio O Default Ratio Net Commutation Rate NCR Ratio O Default Ratio Regional Jobs Per Employee RJPE Ratio 1 Default Ratio PRINT P4 40 SCREEN Figure 1 IDEU Initial Data Entry Unit Report 3 4 3 Data Organization and Preparation Unit and MAP IT In this section you will do the following 1 Work in the Data Preparation Workbook 2 Check the consistency of the link between IDEU data and the GJS files in MAP IT 1 Work in the DOPU Workbook You will enter zonal data related to the employment household and land use activity taking place in your region Zonal data for Rancho Carne is available in the RC DOPU DATA xls file located in the Tutorial folder Enter Your Rancho Carne DOPU Data Upon entering the DOPU Workbook you will be prompted to answer if this is your first time or not entering DOPU module For the purpose of this tutorial click on the Yes button W
68. EQUIREMENTS 5 2 25 2 etre et petu iac Pb Eee a ad Petre each e eb na 4 3 SPATIALLY DISAGGREGATED ZONAL REQUIREMENTS enne nennen nennen 4 3 PREPARATION OF DATA 000101010101011 4 5 EMPLOYMENT DATA 0 0010101010101 4 6 HOUSEHOLD AND POPULATION DATA 00 0 0 0006 4 7 EANDUSE DATA n hareran dn a nn na 4 7 CHECKING THE CONSISTENCY OF THE TELUM RES LAND USE VARIABLES pp 4 8 Formulas for Unusable Useable and Vacant Land Variables en 4 8 Translating From Local Land Use Inventory Categories to TELUM Res Categories 4 9 ZONE TO ZONE TRAVEL TIME OR COST 2 otet dts oe esso ertet been Fes RONDE PNE Deoa cR 4 9 3 PREPARATION OF DATA INPUTS FOR IDEU e eene eee eee etn nannte ee tnan 4 3 REGION LEVEL REQUIREMENTS eeceeeeee 1 1 ne 4 3 SPATIALLY DISAGGREGATED ZONAL REQUIREMENTS 0 4 3 PREPARATION OF DATA INPUTS FOR IDEU uae dd a pd 4 5 EMPLOYMENT 1 a yi bre a 97 4 6 HOUSEHOLD AND POPULATION DATA pp 4 7 te a etie i Tub Nette dote een de dene 4 7 CHECKING THE CONSISTENCY OF THE LAND USE VARIABLES eene eene enne nnne 4 8 Formulas for Unusable Usable and Vacant Land Variables pp 4 8 Translating From Local Land Use Inventory Categories to TELUM Land Use Categories 4 9 4 DATA ORGANIZATION AND PREPARATION UNIT DOPU eee eere 4 9 EMPLOYMENT WORKSHEET 255 reete recited Gat ta ee rede dd ei 4 10 HOUSEHOLDS WORKSHEET
69. ERR A 835 434 034 4638 55 433 s 2 832 431 11 000 39 200 67 400 X AXIS PLOT OF OBSERVED The End 7 34 165 144 123 101 80 58 33 8 Model Forecasting 1 TELUM Program Overview The diagram below is an overview of your work thus far in preparation for your first model forecast This diagram shows the strong interconnections of each TELUM component It is critical as you begin model forecasting that you keep this linear connectivity in mind If you change your base year DOPU zonal data or TIPU impedance values you must also complete a rerun of calibration parameters before proceeding to a new round of model forecasting Data Editors and TELUM Model Parameter Forecasting Output Data Editors Model Files Components TELUM EMP Callbration TELUM EMP Data Prep Employment Forecast Year Location Employment Forecasting Data Flle TELUM EMP Controls Base Year Employment Data Base Year Population and Land Use Data Population and Land Use Data Edito Data ar TELUM RES Calibration TELUM RES Controls Forecast Year pe Employment Employment TELUM RES LLL Data Prep ssh ae Population and Location Forecast Land Use Data Base Year Population ng Population and and Land Use Land Use 3 Base Year Impedance Data Editor Data 8 1
70. GIS folder contains the following set of files after you complete DOPU GIS EMP PRN A space delimited text file containing the data you entered into your DOPU Employment worksheet GIS HH PRN A space delimited text file containing the data you entered into your DOPU Household worksheet GIS LU PRN A space delimited text file containing the data you entered into your DOPU Land Use worksheet The following files are added once you complete MCPU EMPRESEI TXT EMPRESE2 TXT EMPRESE3 TXT etc Comma delimited text files containing model calibration residuals for each of your corresponding regional employment categories DRMRESHI TXT DRMRESH2 TXT DRMRESH3 TXT etc Comma delimited text files containing model calibration residuals for each of your corresponding regional household categories The following files are added to your TELUM GIS folder after MCPU The xxxx in the file names indicate the year in each file name by forecast year EMPFCSTxxxx TXT Comma delimited text files containing model forecast outputs for each regional employment category by forecast year and forecast name DRMFCSTxxxx TXT Comma delimited text files containing model forecast outputs for each regional household category by forecast year and forecast name EMPLBLOS TXT Comma delimited text file containing baseline forecast outputs for Building a Shapefile Unfortunately we are unable to provide directions for building a shapefile We recommend
71. GRADIENT FIRST STEP SIZ RION PUTER ITERATION 20 5 88 02230959 2 21 05937040 4 43 02957990 ET ALP PARAMETER VALU IZED DERIVATIV DERIVATIVE VA RAMET Bl PARAMETER VA RMALIZED DERIVATIVI m DERIVATIVE VALUES This 1s the final Iteration PARAMET 3 VACDEV PARAMETER VALUES NORMALIZED DERIVATIV DERIVATIVE VALUES PARAMET PERDEV PARAMETER VALUES NORMALIZED DERIVATIV DERIVATIVE VALUES PARAMET 5 RES PARAMETER VALU NORMALIZED 17217 DERIVATIVE VALU PARAM 6 LI PARAMETER VALU NORMALIZED DERIVATIV DERIVATIVE VALU PARAMET 9 UT PARAMETER VALU NORMALIZED DERIVATIV DERIVATIVE VALU 7 26 SUMMARY OF RESULTS FROM GRADIENT SEARCH
72. IONAL EMPLOYMENT MIX BASE YEAR REGIONAL POPULATION TO HOUSEHOLD RATIO 2 4 FORECAST YEAR REGIONAL POPULATION TO HOUSEHOLD RATIO 2 264 REGIONAL SUMS OF INPUT VALUES OF ZONAL EMPLOYMENT FORECASTS LInd Hind Servi Retail 430 450 1333 880 ECASTS OF HOUSEHOLDS EGIONAL SUMS OF ZONAL FOR ADJUSTED FOR UNEMPLOYMENT AND EMP HH RATES LI LM MH HH 682 1023 876 514 FORECAST REGIONAL SUM OF EMPLOYED RESIDENTS 3093 INPUT VALUE OF REGIONAL POPULATION FORECAST 7001 COMPARISON OF OUTPUT YEAR INCOME GROUPS TO INPUT YEAR INCOME GROUPS HH 497 1642 HH 514 1661 GROUPS MH 848 8 I E GROUPS MH 876 2831 8 29 INPUT YEAR INCOME LI LM GROUP TOTALS 675 1009 GROUP SHARE 2229 3332 OUTPUT YEAR INCOM LI LM GROUP TOTALS 682 1021 GROUP SHARE 2206 3302 Page 12
73. ITH 0 OR 1 ERVED ARE OMITTED FROM MAPE CALCULATION This is the most RATIO OF ABSOLUTE ERROR SUM TO MEAN OF OBSERVED VARIABLE general of these measures 20 30 ARMO 17 865 represents a good fit THE MAPE AND MARMO STATISTICS ARE ALWAYS GREATER THAN OR EQUAL TO 0 000 FOR A PERFECT FIT ALL OF THESE STATISTICS WOULD BE EQUAL TO 0 000 7 33 k KKK KKK ck ck ESTIMATED ERVED VS REGRESSION OF OBS k ko ko ko kk ck ck ck ck CK Another way of comparing the model estimate to the actual data 15 by using a simple linear regression of one vs the other Here the estimated vs observed data are plotted 123 800 152 000 ED Y AXIS 95 600 ESTIMAT 52 67 VS DF 23 Bom OTIT F 30 84 3490 T FOR 8 2 154383 T FOR A 10 EST ERR B
74. Impedance Preparation 1 Introduction to the Travel Impedance Preparation Unit 6 1 2 How to Organize your Travel Impedance Data e 6 1 Organize your IMPD txt File nn 6 2 Create the IMPD txt 7 7 6 2 Enter your IMPD txt File into TELUM sss 6 2 Complete the Travel Impedance File Check ee 6 3 0 3 How to Add Future Year Travel Inpedance pp 6 3 7 Model Calibration 1 Introduction to Model Calibration 7 1 2 Starting Model Calibration File Check ee ee nenne 7 1 3 Begin Model Calibration Employment and Household Location 7 1 4 Analysis of Results for Model Calibration ecce ecce esee eee eene nenne 7 1 Analysis of MAPE and MARMO Results pp 72 Analysis of Location Elasticity Results pp 1 2 5 Land Consumption Calibration LANCON 6666666666666 7 3 How to Run a LANCON Calibration Regression ppp 7 3 Calibration Output Files 7 3 6 Model Calibration Appendix 1 1 1 1 1 1 1 1 eres 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 12 2 1 12 1 1 tae 4 Model Calibration and Goodness of Fit Theory pp 7 4 Asymptotic t Statistics in DRAM and EMPAL Calibrations 5 TELUM RES Location Elasticity Values 5 TELUM
75. KKK KKK ck ck KKK KK kk THIS RUN MADE 0 99 8 AT 9 17 59p mycity empal 1 110899 Your input use it to Always check this to specifically identify be sure it is the run your runs you think it is CK ck ck ck ck ck ko kc KKK IS RUN FOR EMPLOYMENT TYP EL FUNCTION TYPI Confirm that this 1s MBER OF ZONES 10 what you MBER OF EMPLOYMENT S I C GROUPINGS wanted MBER OF HOUSEHO YPES MBER OF LAND US TRACTIVENESS VARIABLI 7 9
76. M is an integrated interactive system that can be used to assist in evaluating the effects of a region s planned transportation improvement projects It may also be used to make long term forecasts of a region s spatial patterns as well as to produce forecasts which address the transportation and land use consistency that is required as input to the air quality estimates now required by the CAAA and TEA21 and ISTEA before that TELUM uses current and prior residential employment and land use data to forecast the future locations of each of those by employment sector household income group and land use type The interrelationships between transportation and land use can be just as important and in some cases more important than the individual direct consequences of either Having articulated a framework for examining analyzing or understanding the transportation and land use interactions it then becomes possible to consider the consequences of a wide assortment of different kinds of policies This includes policies that attempt to achieve their aims by changes on the demand side in terms of urban design and land use control as well as those that attempt to achieve their aims by acting on the supply side in terms of various kinds of transportation improvement These transportation improvements can be for highways transit or combinations thereof as well as in increases in utilization efficiency of existing facilities 3 Introduction to Integrated
77. N This would give the following Best Fit value of likelihood 1 gt NilnNi 2 1 7 4 The Worst Fit occurs when all values of the dependent variable are estimated by the mean of that variable For example if the region s total of Type 1 employment were divided by the number of zones to get the mean of Type 1 employment per zone and all zones were assigned an amount of Type 1 employment equal to the mean This is also known as the uniform distribution assumption where the Estimated Ni the Zonal Mean N and gives the following Worst Fit value of likelihood Ly gt NinN 3 From these two extreme values of likelihood we can construct a relative measure of goodness of fit which is analogous to the R measure but which is appropriate to the nonlinear equations of CALIBTEL and to the non normal distributions of the data This measure of Relative goodness of fit is called a Best Worst Likelihood Ratio and takes the following equation form oss Lob Lw 0 4 The computed value of this Best Worst Likelihood Ratio g has a range such that for a perfect fit o 1 00 and for the worst fit 0 00 Typical results obtained when fitting CALIBTEL give 0 70 0 95 The values taken by are independent of the magnitude of the dependent variables Asymptotic t Statistics in DRAM and EMPAL Calibrations In estimating nonlinear model parameters it is necessary to develop ways of assessing statistical signif
78. NCON uses your land use employment and household data to forecast the change in the amount of land by zone that will be used by each of these categories LANCON calibration is done by use of linear multiple regressions Much like the employment and household calibration analysis LANCON provides statistical measures for your region s land use data 3 12 At the startup of LANCON the calibration regression procedure will ask you to enter the following 1 Indicate for TELUM which employment categories in Rancho Carne are Basic Basic categories are industries that produce goods For Rancho Carne please check AGR and LMFG 2 Indicate for TELUM which household categories in Rancho Carne are Low and High Income Low and High incomes refer to households that fall in the bottom or top quartile quintile in the region respectively For Rancho Carne please check LI as your Low Income household category and UI as your High Income household category 3 Execute LANCON by clicking on the LANCON button This process takes time Please wait for the Continue button to become enabled before trying to proceed 4 Your LANCON findings will be displayed by Residential Commercial and Industrial Land Consumption category in the LANCON Statistical Report A sample report looks like this zz Tetum Lancon Statistical Report Commercial Land Summary Output Commercial Land Consumption Consumption by Zone Regre
79. POLS 1396 394 695 390 688 394 695 475 837 277 489 3971 7001 LATION 211 273 3 18 tola 604 301 298 301 362 212 OY 3030 6 EMP POPUI Page ENTARY PO ERS SUPPLEM GROUP QUART POPULATIO 0 OOTO O OO IG O BASE YEAR 2000 INPUT DATA ESDNTL 05 58 224 07 N 30 29 89 UI 453 09 ENUA R ESIDENTIAL 5 O 0 H 0 5 ETAIL R 99 94 98 32 61 42 27 56 72 42 99 72 95 03 96 09 97 39 97 52 ED LAND USE ETAIL How i T5 Dp Ls 70 j Page 8 26 ELOP DEV BASIC 01 10 sl 3 28 03 08 14 07 38 16 14 70 27 RESIDENT POPULATION 1 488 2 630 3 433 4 3504 5 1396 6 695 7 688 8 695 9 837 10 489 TOTAL 7001 TOT DVLPD BASIC 1 la 0 2 4 0 3 4 ie 4 19 4 5 6 2
80. PU Workbook and RC DOPU DATA xls by clicking on File gt Save After you have finished entering data into each of your DOPU worksheets click again on File lt Save There will be an X on the main worksheet next to each workbook name as shown below Microsoft Excel 1230 2 DATAPREP v 1 xls gt Ele Edt view Insert Format Tools Data Window Help E 2 18 xi 2 2 SRY 18 6 5 0 A MB 0m gt o sety ET Arial u BU EF Sz 09 9 3E 383 0 As Dig X A 8 5 D TE F 6 H I J K IL M N NN 1 d E Welcome to Data Check 5 6 7 When you have completed all of your worksheet data entries 8 1 up RUN Data Check save your Danprep workbook then click on me button 0 RESULTS and EXIT Run Data Check Results and Exit 11 Failure to run Data Check will stop your TELUM project Important 1 When you begin Data Check your Dataprep workbook will close 2 Review the Data Check Report generated by the TELUM system 3 Your Data Check Report is an important document Please print your report and keep it with your project 4 If you decide to change your DOPU data you must re enter the IDEU system M4 Data Check Employment Households Land Use Projections Conversion Matrix lal Ready FI NUM ji Figure 4 DOPU Opening Screen NOTE Please save and exit the RC_DOPU_DATA xIs file before running the DOPU Workbook
81. REGION TOTAL 3 LARGEST 25 OF ZONES 11 347 ES HAVE 63 75 OF THE REGION TOTAL TEST ZONES ZONES WITH 0 OR 1 RE OMITTED FROM MAPE CALCULATION This is the most RATIO OF ABSOLUTE ERROR SUM TO MEAN OF OBSERVED VARIABLI general of these ES 0 ARMO c 22 6695 measures 20 30 represents a good fit THE MAPE AND MARMO STATISTICS ARE ALWAYS GREATER THAN OR EQUAL TO 0 000 FOR A PERFECT FIT ALL OF THESE STATISTICS WOULD BE EQUAL TO 0 000 7 18 CK ck KKK KKK TIMATED ES REGRESSION OF OBSERVED VS Ck ck ck ck ck Ck ko kc ko gt kk ok Another way of comparing the model estimate to the actual data 15 by using 8 simple linear regression of one vs the other Here the estimated vs observed data are plotted 88 200 110 000 Y AXIS ED 6
82. T uses ESRI s ArcMap technology to produce maps that will 1 Display your model data inputs and outputs An example of a Rancho Carne MAP IT output follows Zones d Yew Insert Section Joos window hep x Zone With The Largest Total Area 260 05 3 2j SISIEN 1 35 4 72 Inches 46 457 5 1 x Total Households 2020 PT2 Density Gradients _ 050 gt 050 to 1 WB 337103291 Fn u Density Gradients 5 2 amies tt it 26 2 Measure regional sprawl MapIt Mapping options Eie Edt vew Insert Selection Tools Window Help FOBERSOU HDD Layers zone HouseholdsDensityGr 0 89to 0 50 E gt 0 50 to 0 22 C gt 4 22 to 0 60 E gt 0 60 to 3 37 Il 3 37 to 32 91 Layers s xe we 1 07 149 17 T 294 to 1006 5 7628 4 What if you do not have ArcView 8x mapping software While we strongly promote the use of MAP IT we recognize you may not have ArcView 8x available for your use Inside your TELUM project folder is a folder labeled GIS C TELUM GIS Your GIS folder contains a copy of all the text outputs TELUM generates and reads into the ArcMap software These text files may be formatted for use in your agency s mapping application GIS Folder Contents Your
83. UDING INTRAZONAL VALUES 54 0320 STANDARD DEVIATION OF IMPEDANCE 24 9670 MEDIAN TRIP LENGTH IS IN THE INTERVAL 40 2967 TO 42 1283 Y XR 9 184 KKK T8255 6 560 5 248 34936 AUGE 2 624 8 1312 8 X 7 109 900 88 286 66 673 45 059 23 445 1 032 PLOT OF TRIP FREQUENCY IN PERCENTS Y AXIS VS TRIPLENGTH X AXIS NOTE THAT DATA ARE GROUPED BECAUSE DATA SET HAS 10 ZONES 7 29 k ck SUMMARY OF COEFFICIENTS AND SIGNIFICANCE TESTS FINAL VALUES OF DRAM PARAMETERS These are the values you pum wil
84. UE Figure 7 MCPU Analysis 01 Results for Employment 3 11 MAP IT also appears in this section for you to view your Calibration Residuals Residual maps show where the model over and under estimates the zonal location of employment and households in the region An example of MAP IT outputs for Low Income Household Residuals follows LI Households Residuals Percent Error E gt 50 50 to 10 10 to 10 10 to 50 EH gt 60 Figure 8 MAP IT Output 3 To explain the role of calibration residuals let s assume that a zone in Rancho Carne contains the regional airport The presence of the airport shows a concentration of employment Under other circumstances a concentration of employment like this will cause an increase in household attraction which is not the case here because households are prohibited from locating near the airport While the model is unable to know zonal specifics a user may be able to identify the model s source of high over or under estimation A zone with a high over estimation of households as discussed in this example will appear in the darkest gray scale shade Zonal characteristics like this are important to document as they can later be added as local knowledge to the model to modify zone attractiveness and thus improve the accuracy of forecasts 3 Run a Land Consumption Regression Model The Land Consumption Model LANCON is the last section in the MCPU component LA
85. Users Manual TELUM Transportation Economic and Land Use Model Version 5 0 March 2005 Table of Contents 1 Introduction to TELUM 1 Overview of the TELUM Manwual etn 1 1 2 What can you do with TELUMP tii 1 2 3 Introduction to Integrated Transportation and Land Use Modeling 1 2 4 1 3 Operating Systeme 1 ett tsa ces ee pee eet 1 3 BLT 0 Ca omne nemen t EN 1 4 Hyperlinks uie ea 1 4 Job 1 4 Information 1105 eiecti oer tt hast iege eta ete ipee Lig t aru Eee res 1 5 Comment BOxes 3 i eee 2 12 212 i 1 5 eed 1 5 5 Flow of the TELUM System 1 5 6 Land Use Modeling An Overview 1 6 7 Forecasting with 1 8 2 Installation 1 UPD on 2 1 System Requirement oe nd nnne nennen nnne nnne 2 1 Overview of New TELUM Installation 2 2 Overview of an Upgrade to an Existing TELUM Installation 2 2 Running SETUP EXE de ita deet ie cete de rte le eral 2 3 2 Configuring TELUM 00 eere ee PEE Eee ree PEE siose 2 4 Username Setting Sin estet ree ere t E E 25 3 How to Change or Move your TELUM Project ee 2 5 Circumstances Requiring a TELUM Reinstallation 3 Tutorial 1 11901001101013 se M 3 1 How ol ne a i na 3 1 2 Initial Data Entry Umit
86. al ratio of jobs per employee RJPE We divide the total number of jobs in the region i e the total reported employment in the region 132 by the total persons working in the region This is the 90 residents who work in the region plus the 20 who live outside but commute into the region to work or a total of 110 This gives 132 110 or 1 2 jobs per employee Next in terms of persons employees we recall that 110 employees work in the region 20 of which commute in Ten other persons live in the region but commute out to work The regional net commuting rate RNCR equals 1 0 plus the ratio of the net commutation to the total employed persons at work in the region This ratio 1s therefore outbound commuters minus inbound commuters or 10 20 or 10 divided by 110 to yield 0 0909 which when added to 1 0 gives 0 9090 for RNCR Working back through the numbers we get 110 persons employed in the region if we divide the region s 132 jobs by the RJPE of 1 2 We may then multiply that 110 by the RNCR of 0 9090 to get the 100 employed residents of the region The unemployment rate UNEMP is calculated by dividing the number of unemployed workers by the total labor force or 5 divided by 100 5 yielding 0 0476 If we multiply the 100 employed residents by 1 0 divided by 1 0 minus the unemployment rate UNEMP we get 100 times 1 05 or 105 as the total of employed plus employable residents Finally we note that if we go back to the origina
87. an be useful to review this output to check the accuracy of input data and diagnose any calibration problems Two annotated sample calibration output files one for employment and one for households are included after the Appendix 7 3 Model Calibration Appendix Model Calibration and Goodness of Fit Theory The following section introduces the mathematical method used in calibration The calibration process involves fitting the TELUM equations to the data for a particular region The better the fit of the model to the data the more reliable the forecasts it produces In all socio economic data there is a systematic explainable component and a random unexplainable component The goal of model calibration is to adjust the model parameters so as to permit the model to explain as much as possible of the systematic component of the data To perform calibrations it is necessary to have one or more indicators of Goodness of Fit of the models to the data The equation structures of the TELUM models are intrinsically nonlinear and the data from which their parameters must be estimated are not normally distributed As such standard multiple regression techniques cannot do the job The parameters for the models are estimated by a computer program called CALIBTEL CALIBTEL contains procedures for each of the two models TELUM EMP and TELUM RES that are used to estimate model parameters for employment and household location respectively The procedure
88. ar time increments this causes several problems The principal problem is that too much takes place in ten years for the user to feel comfortable with this size increment In order to model the essential interactions between the various locating activities as well as between any of the activities and the transportation system a five year feedback or interaction time is much more appropriate than a ten year period Even so the exigencies of data availability may sometimes make it necessary to make do with less than the ideal data set 3 Sectoral Detail The sectoral detail situation is somewhat less nebulous Prior to the development of TELUM Emp and TELUM Res most earlier model applications used few or just one categories of locators While the final model outputs were produced in considerable detail the actual location procedures often involved only one or two locating categories A major goal in finding a means for modeling an increased number of locator types is to model each with individual and typically different locating behavior In addition the location behavior differences are determined from statistical analyses of the base year data for the specific region to which the models are being applied TELUM provides for the use of up to eight employment types and eight household types The precise number of different locators depends on both data availability and the intended use of the model outputs IDEU Zonal Input Checklist Use the
89. asticity measures the resulting percentage change in the number of households or employees in that zone For example suppose that for low income households in Zone 12 the location elasticity for residential land is equal to 0 2500 This means that a 196 increase in residential land in Zone 12 will result in 0 2596 increase in the attractiveness of Zone 12 to low income households The location elasticity values are static measures of model sensitivity This means that when a location elasticity value is calculated for a specific attractiveness variable in a zone all other attractiveness variables remain fixed In the example above the only variable that is allowed to change is the quantity of residential land in Zone 12 All other attractiveness variables in Zone 12 are assumed to be fixed as are the attractiveness variables including residential land in all other zones Because the location elasticity values are static measures of model sensitivity they will change as the values of the TELUM model attractiveness variables change e g the location elasticity values for forecast years will be different from the location elasticity values for the base year 7 2 The value of location elasticity for a specific attractiveness variable and zone is a function of 1 the value of the calibrated parameter for the attractiveness variable 2 the numbers of households or employees in the zone 3 the magnitude of the attractiveness variable and 4 the
90. at time 1 The residence and employment location forecasts produced by TELUM may then be used sometimes after a further step of spatial disaggregation as input to travel models that generate and distribute trips split trips by mode and then assign vehicle trips to the transportation network s and calculate congestion 1 8 So for example one could take the outputs of TELUM and use these as inputs to the trip generation and distribution components of some standard transportation planning model package Having completed the assignment of trips to the network using this package one could calculate the minimum paths through the network If multiple modes are being analyzed the minimum times through the networks via these different modes are calculated They would be combined in a composite cost calculation and then the composite cost estimates of zone to zone composite travel times or travel costs would be taken and used as inputs to the recalculation of employment and household location in TELUM Many different configurations of land use and transportation linkages have been tested using ITLUP and METROPILUS While the current implementation of the TELUM model system does not permit all of these configurations to be examined there is adequate scope in an agency setting for preparing an accurate baseline forecast which can be used as inputs to the agency s travel modeling system as well as for making forecasts of the consequences of variou
91. ation Preparation Unit MCPU SKIP TO MCPU Model Forecasting Unit MFCU P1 11 1 4 4 Information Tips Information Tips are small buttons labeled with 1 found inside the TELUM DOPU Data Preparation workbook By clicking on the information button TELUM pops up a small text box that describes the data requirements of the tables 5 Comment Boxes Comment Boxes are small red triangles that appear in the upper right corner of the TELUM DOPU Data Preparation workbook spreadsheet Comment boxes pop up when the mouse pointer is paused over a cell with a red triangle 6 HELP This feature is currently under construction In the future you will be able to access this manual and troubleshooting information from an internal TELUM file 5 Flow of the TELUM System To complete a full set of model forecasts for your region you must work through the TELUM system in the following order IDEU DOPU TIPU MCPU MFCU Once you have successfully completed each section you may move between modules to do additional work These modules are designed to gather organize calibrate and forecast your regional employment and households IDEU Initial Data Entry Unit Y TIPU Travel Impedance Processing Unit Y MCPU Model Calibration Processing Unit Y MFCU Model Forecast Calculation Unit TELUM Flowchart 1 5 In the following section we provide an overv
92. avel Impedance No When prompted by TELUM enter the following projections into the Employment and Household projections table Population Employment Year Total Pop AGR LMFG PROF RTL FIRE Total Emp 1995 57450 1977 3820 4123 11755 8 29083 2000 62819 1936 5434 6121 11121 6402 31003 2005 70016 851 9599 5670 10787 3 35010 2010 74750 834 10227 6324 11492 8346 37223 2015 77219 751 11084 6975 12299 8763 39872 2020 79664 669 11398 7248 12845 8938 41098 2025 81953 710 11611 7486 13141 9117 42065 2030 82776 684 11743 7573 13274 9482 42756 4 Run a Policy Model Forecast Agencies often run model forecasts in addition to their Baseline set referred to as a Policy forecast For a Policy forecast a user must intend to e Add or change a future year impedance file e Change the regional employment and household projections Run a New Forecast for Rancho Carne The following charts provide the information and data you will need to provide to run this New forecast for Rancho Carne In this new forecast you will add a Future Year Impedance file for the year 2010 Rancho Carne has adopted a new tax incentive policy in zones 6 7 16 17 and 18 that is expected to increase both employment and households to these areas With an increase in population and employment Rancho Carne s transportation department has issued new travel impedances for 20
93. cel must include both the Analysis ToolPak and the Analysis ToolPak VBA extensions Please load these extensions before installing TELUM Directions are provided below Please note that some computer installations especially in the case of networked systems have these extensions setup to be uninstalled when the user turns off their PC or even in some cases if they log off the network You will need to be sure that they are present prior to each use of TELUM Open an Excel workbook Go to Tools Add Ins click on Analysis ToolPak AND Analysis ToolPak VBA Note You might need to use your Microsoft Office CD ROM in order to install these extensions Please contact your systems administrator if you are experiencing problems loading these extensions Microsoft Excel must be set on Low Macro Security whenever you are running your TELUM project Open an Excel workbook Go to Tools Macro Security select Low ESRI ArcGIS must be installed on your computer in order to use TELUM mapping module called MAPIT Current version of TELUM is compatible with ArcGIS versions 8 3 and 9 x In addition to desktop installation of ArcGIS you also must install ArcGIS Developer Kit versions 8 3 9 0 and 9 1 or ArcGIS Desktop SDK for Visual Basic versions 9 2 This is necessary as some of the ArcGIS scripts and software libraries used by TELUM are not installed with the ArcGIS basic desktop installation If you do not have ArcGIS or you
94. cioeconomic class or group TELUM Emp specifies employment by industry type such as manufacturing or retail Thus the employment forecasts from TELUM Emp which yield spatial distributions of employment at place of work by employment type are converted to households by income group at place of work This conversion is accomplished by multiplying the matrix of employment forecasts by a set of conversion ratios that are derived from regional statistics This procedure provides the user with unique advantages Perhaps of greatest importance is that as the regional mix proportions of employment types varies so does the region s household income distribution For example a region experiencing a long term shift from manufacturing employment to service employment will have built into the model system the appropriate shift in the distribution of household incomes as a consequence of the different labor mixes of the different employment types This matrix of conversion ratios is constructed from Public Use Microdata Sample PUMS data provided by the U S Census Bureau Since there are thousands of observations in a PUMS data file a computer program must be written to compute this cross tabulation 4 13 While using PUMS data 15 suggested for building your conversion ratio matrix agencies unable to access PUMS data may use a default matrix The default matrix replaces the PUMS data with an even distribution of total persons employed by ind
95. d by definition but questions often arise in determining where to place certain categories that may have been defined differently for the original data files from which the data are being prepared For each zone it is necessary to know the following e Total zonal area e Residential area all types e Unusable area e g water or environmentally sensitive lands e Industrial area used for basic employment e Vacant developable area e Commercial area used for commercial employment Problems tend to arise in determining what constitutes vacant usable land area The models treat this category as developable land or land that can be used for residential industrial or commercial purposes How agricultural land parkland streets and highways and wetlands are to be treated is a matter for decision by the agency In forecast model runs it is necessary to have specific values for these categories and to develop a definition or set of definitions that are used for the baseline runs but may be tested as work progresses It is particularly important to consider issues of consistency in the land use category definitions The following are guidelines for consistency checking 4 7 Checking the Consistency of the Land Use Variables When a TELUM Res data set is constructed it is important that the land use data is internally consistent Land area must be in acres not square miles and should be consistent throughout all data sets or estimate
96. d Zonal Percent Change the zonal growth decline between two time periods and or forecast runs expressed as a percentage Simple Zonal Density total zonal developable land divided by the total number of zonal employment and or households observed 3 17 An example of Baseline and New Forecast MAP IT outputs follows 218 x Mapping Options Y File Edit View Insert Selection Tools Window Help zone TelumCalc 1 295 to 246 Bl gt 246 to 150 E gt 150 to 80 IH lt 80to 1 2005 Difference In Total Households Between Policies BL And 5 6 ki Difference 295 to 248 to 150 246 gt 5 0 150 gt D gt s0to 1 a s 2 13 5417 97 15 42 13 12 N 2 71 4 96 Inches i Figure 11 MAP IT Output 4 This concludes your TELUM Tutorial If you have questions please feel free to contact the TELUM staff J Brugger 2004 Oct 15 C Putman Putman Edits 22EDITTELUM 3 18 4 Data Preparation Data Preparation Index Because it 15 likely that users will make frequent reference to the sections of this chapter while preparing the data for their TELUM project we have included a separate chapter index here IC THB TEEUMPROCBESS 001 4 2 2 DATA REQUIREMENTS FOR AGENCY APPLICATIONS eee eee ee eene eee 4 3 REGION EEVEL R
97. d land consumption rates will be inaccurate The input data set for forecasting with the TELUM Res model has eight land use variables TAA Total Land Area AU Unusable Land AAAB Land Used for Basic Employment AAC Land Used for Commercial Employment GAAR Residential Land USBL Total Usable Land STS Land Used for Streets and Highways VAC Vacant Developable Land It is especially important for Unusable Usable and Vacant Land to be calculated consistently Finding consistent values for these three variables can be confusing since each variable s definition depends on the definition of the other two The best strategy is to fix the value for one variable and then determine the values for the other two variables Formulas for Unusable Usable and Vacant Land Variables 1 Vacant Land Fixed Usable and Unusable Land Calculated In the TELUM Res land accounting procedure developed land DEV is defined as DEV AAAB AAC GAAR STS If the vacant land values are known then usable land is defined as USBL DEV VAC Unusable land AU is defined as AU TAA DEV VAC 2 Unusable Land Fixed Usable and Vacant Land Calculated If the values of Unusable land are known then vacant land is defined as VAC TAA DEV AU Usable land is defined as USBL DEV VAC 4 8 3 Usable Land Fixed Unusable and Vacant Land Calculated If the values of usable land are known then vacant land 15 defined as VAC USBL DEV Unusable
98. do not have a version of ArcGIS compatible with TELUM installed on your computer you can not use MAPIT module Please select No when prompted by MAP IT The embedded GIS procedures of TELUM will only run with above versions of ArcGIS though you can export completed calibration and forecasting data from TELUM after completing the runs and use them in other GIS software Land use data values are to be in the unit of acres only When prompted by TELUM users are asked to place file s inside the TELUM project folder Files copied into TELUM must not have the read only property Check file properties after you copy them into TELUM folder especially if you are copying them from a backup CD many of which automatically set file property to read only To change the read only property please follow the instructions Select the file s then right click mouse on the file s go to Properties General and check if the property Read only is selected If it is then unselect it and click Apply Do the same with the rest of the files you copied in TELUM You should remove all former versions of TELUM prior to installing TELUM ver 5 BEFORE removing former versions of TELUM you have the option to save the data that you have already entered in TELUM s Dataprep xls spreadsheet In order to do that you should save your old dataprep xls with a new name and in a different folder of your choice AFTER installing the new version
99. e new system like its predecessor ITLUP contained several models for location analysis as well as software or links to external software for all the necessary data analysis statistical analysis and display including mapping of outputs and results In addition the components of the package could be applied as separate models and also could be connected to other models currently in use by a student or a planning agency for forecasting and analysis tasks The entire modeling system along with numerous utility programs as well as Graphical User Interfaces GUIs was embedded in ESRT s popular ArcView GIS operating environment Beginning in 1999 Professor Putman embarked on a five year U S Department of Transportation sponsored project to retool METROPILUS as a land use component for the Transportation Economic Land Use System TELUS The TELUS Land Use Model TELUM evolved from the earlier METROPILUS work to become a self contained novice friendly land use modeling system designed to project the location of new residential and nonresidential development based upon analysis of 1 prior and existing residential and nonresidential development 2 the location of transportation improvement s and 3 overall congestion in the system TELUM forecasts the location and amount of 1 2 household and employment growth for up to 30 years information needed by an MPO s external travel demand forecasting models to estimate network flows and sub
100. e region TELUM uses this rate to adjust the regional employment forecasts so they have the proper number of employees living in the region A rate less than 1 00 indicates net inbound and a rate greater than 1 00 indicates net outbound commuting In the absence of actual data a default value of 1 00 should be used which indicates no net in or outbound commuting is occurring 4 14 The regional rate of jobs per employee 1s a simple calculation of total reported jobs in the region divided by the total number of persons working in the region The number of total persons working in the region includes persons commuting into the region This value will typically fall below 1 50 A default value of 1 00 should be used 11 actual data is not available When the estimated conversion ratios are used in TELUM Res the conversion procedure defined by equation 1 must be consistent 1 6 the total number of households in each income group generated by the conversion procedure should be very close to the observed number of households in each income group In most cases the DATAPREP XLS conversion matrix spreadsheet computes an employee to household conversion matrix CNV j that guarantees consistency in the TELUM Res conversion procedure Industry HHI HH2 HH3 HH4 HH5 HH6 HH7 HH8 Total e 1 0 0 0 0 0 0 0 1 2 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 4 0 0 0 0 0 0 0 5 0 0 0 0 0 0 0 6 0 0 0 0 0 0 0 7 0 0 0 0 0 0 0 8 0 0 0 0 0 0 0 Ti o 589
101. e TELUM with the basic information the system will need to organize your region s data The purpose of IDEU is to help you prepare the data you will need in the next component of TELUM DOPU Listed below are the TELUM data requirements for IDEU The last column contains sample data you will use to run your tutorial session Enter Your Rancho Carne IDEU Data The sample data inputs below contain employment and household activity name abbreviations for the Rancho Carne region Employment and household activity name abbreviations will vary from one TELUM project to another The definition of each abbreviation follows the table below 32 Sample Data for Rancho Carne CA RC 20 62819 2000 1995 AGR LMFG PROF RTL FIRE 4 LI MI UMI UI 184 332 Available Available Available Available Available Available Available Available 6 PUMS Ratio EHIC Ratio UR Ratio RNCR Ratio Select Do Not Know TELUM IDEU Data Categories Name of your Region Number of Zones Total Regional Population Current Data Year Lag Data Year Number of Employment Categories Name of Employment Category 1 Name of Employment Category 2 Name of Employment Category 3 Name of Employment Category 5 Number of Household Categories Name of Household Category 1 Name of Household Category 2 Name of Household Category 3 Name of Household Category 4 Total Land Area of Region Total Land Usable Land Unusable Land Residential Land S
102. e first column contains an assigned identification number The second column contains your regional impedance zones starting with the impedance value of Zone 1 to Zone 1 followed by the impedance from Zone 1 to Zone 2 then Zone 1 to Zone 3 and so forth When you are finished you will have one long column of impedances from each regional zone to all other zones in the region A pictorial example is provided below 6 1 1 Organize your IMPD txt file In this example you see two tables of impedance numbers The first table shows your impedances in a matrix format where the impedances are read from left to right and top to bottom The second table shows how these impedance values are organized from the matrix table into your IMPD txt file Zone1 Zone2 Zone3 Zone4 2008 5 Zone 1 4 40 12 70 14 80 23 20 33 20 Zone 2 21 00 21 40 27 20 28 60 45 30 Zone 3 26 00 26 70 26 10 26 00 27 90 Zone4 31 50 33 30 33 30 33 40 33 40 Zone 5 50 50 40 50 40 50 42 30 42 30 Table2 oo00 1 0 0 ON 1 1 2 3 2 2 2 2 4 2 OA 0 4 Example 1 Organize your IMPD txt file 2 Create the IMPD txt Your IMPD txt will have two 10 space columns as pictured in Example 1 The first column contains an identification number of your choice The second column contains your impedance values If your impedance tab
103. e first set of model forecasts called Baseline Baseline forecasting is the spatial allocation of employment and households to zones based on an observed level of activity and calibrated attractiveness variables obtained from your current and lagged year data inputs During your Baseline forecast TELUM provides you with an opportunity to Change the number of forecast time periods e Add an additional impedance file for a future forecast time period e Change your region s total employment and household projections 2 Run a Baseline Model Forecast For Rancho Carne run your Baseline forecasts with changes to the forecast time periods impedance file and projections table as shown below Screen MFCU Variable Input Value P7 26 1 Change Regional Employment and Household Projections No P7 13 Add Future Impedance File No P7 6 Continue with Forecast Time Periods Yes P7 5 Impose Employment or Household Constraints No At the conclusion of your Baseline forecast TELUM will immediately provide an Analysis of Forecast Spatial Patterns to summarize the growth decline of the region s employment and household spatial allocations It is important to review this report carefully in conjunction with the MAP IT Forecasting results 3 14 BL Percent Change In Total Employment 2000 To 2005 Percent Change 428440 gp lt 38 39 to 11 96 BE gt 11 96 to 7 15 BH gt 7 15 to 43 87 gt 43 87
104. ecast you are immediately provided with an option to view a summary of your forecast outputs by policy and forecast year This report summarizes regional change in employee and household location and land use consumption For a more detailed graphical view of your region s zonal changes we recommend using MAP IT 82 3 The Baseline Forecast The Baseline Forecast is the initial set of regional forecasts made from your DOPU data TIPU impedance file and MCPU calibration parameters During future policy runs you will use your Baseline forecasts for a comparative analysis against your region s policy forecast s Running a Baseline Model Forecast In preparation for the baseline forecast TELUM provides you with an opportunity to do one or all of the following Reduce the number of forecast time periods you wish to use Add an impedance file for a future time period s Change your regional employment and population projects for your forecast time periods e Add zonal constraints Each of these options is described in more detail in the following sections For your first baseline forecast you may or may not wish to use the above options It is important to know that whatever you do in your baseline forecasts will be reflected in any of the policy forecasts you attempt later For instance if you decide to make a baseline forecast with only four forecast time periods all subsequent policy forecasts can only have four or fewer time
105. ecasts your agency believes accurately reflect the location of activity for the region 1 2 4 Running a New Model Forecast A new model forecast also referred to as a Policy Forecast can be run in the TELUM system after the baseline forecast Agencies typically run a new forecast when A change in policy is planned for one or more zones in the region A transit improvement or highway project changes impedance values in one or more zones Updates in the regional employment and population projections become available The agency wishes to analyze the What if we did this or What if the region doesn t grow as rapidly as we now expect that come along with policy changes 8 4 How to Run a New Model Forecast Begin the New Model Forecast by gathering your new data inputs New data typically includes updated household employment or impedance data policy constraints in the form of impedance value changes maximums and minimums on household and employment data or absolute parameter values Please see the proceeding section for directions to change regional projections DOPU and TIPU data and impose household and employment constraints by zone To run a new model forecast 1 Open TELUM and click Skip to MFCU 2 MFCU will ask you to select whether you wish to perform a New or Rerun model forecast Select New model forecast 3 Enter the name and a description of your new model forecast 4 TELUM wi
106. edance files to add for future forecast time periods 8 5 When running a new model forecast TELUM provides the options to change regional employment and population projections modify regional impedance files and impose location constraints on household and employment locators DOPU or TIPU changes must be made in the corresponding sections of TELUM Projection and impedance data can be updated by running a new model forecast as described below Changes to Employment and Population Projections and Impedance Files When new regional employment and population projections or impedance data becomes available the regional TELUM forecasts should be updated to ensure accuracy After viewing the baseline forecast and choosing to run a new model forecast you can update projections and impedance files By choosing to change regional employment and population projections you can manually alter the population and employment projections in the supplied worksheet in screen P7 8 4 as shown below zial Model Forecasting Unit Enter your new Employment Population Projection values for AGP Please manually update changes in your employment categories to the Total Employment column Forecast Time Employment Category Periods AGR LMFG PROF Employmen population 8 6 Employment and Household Constraints by Type and or Zone Any policy or infrastructure change expected to affect employee and household location and
107. el Requlirements ener nnns 4 3 Spatially Disaggregated Zonal Requirements pp 4 3 3 Preparation of Data Inputs for IDEU e eeeeeeee esee eene eene rnnt 4 5 Geographic Detail etn hte pie tetro en te qiie ei eerie 4 5 Temporal 4 6 Sectoral Detail x 1 a 4 6 Employment Data isee ee ct e ir dee a D TO e vele ben durs 4 6 Household and Population Data essen 4 7 Land Use Data 1 1 1 1 1 a en Vee 4 7 Checking the Consistency of the TELUM RES Land Use Variables 4 8 Formulas for Unusable Useable and Vacant Land Variables 4 8 Translating From Local Land Use Inventory Categories to TELUM RES 4 9 4 Data Organization and Preparation Unit 130 1207 4 9 Employment Worksheet i 4 10 Household Worksheet 4 11 Land Use Worksheet 1 a cte terere te gue 4 12 Projections Wotksheet 2 aede o ope en de 4 13 The Employment to Households Conversion Matrix 4 13 5 Data Check and Consistency Report ssa 4 16 4 7 Appendix for Data 6 5 Map It 1 What you can do with MAP TT 5 1 2 Howto run MAP IT aetas oae idee ee tete 5 1 3 What can MAP IT do for you ps 5 1 4 What if you do not have ArcView 8x mapping software RN 5 3 Building a Shape filenin eere e eg 5 3 6 Travel
108. ency from performing model forecast 6 How to Run a LANCON Calibration Regression 1 When TELUM opens to LANCON you will be asked to indicate which employment categories in your region are basic industrial Click all employment categories that apply as Basic employment for your region You must select at least one category 2 Next TELUM will ask you to check off all Household categories that are low income and high income within your region The terms low and high refer to your regions bottom and top quartile quintile category respectively You must select at least one category for each 3 Begin LANCON by clicking the LANCON button This procedure requires you to have the Analysis ToolPak installed in Microsoft Excel and your Macro security set to Low before LANCON can operate See the Installation Instructions for more details This process 15 highly sensitive and should not be interrupted Calibration Output Files In addition to the summary information which is provided for the user TELUM stores an detailed calibration output report in C TELUM Many new users of the models will find the summaries to be adequate for their purposes Our experience in Beta testing the system is that once users become familiar with the modeling process they sometimes want more information than the calibration summaries These files which have the suffix out contain the full report on each locator s calibration It c
109. ercial employment Residentially occupied land e Total usable land developed vacant developable e Land used for streets and highways e Vacant land developable e Employment by type e Land area occupied by basic employment industrial e Land area occupied by commercial employment e Zone to zone travel times and or costs These above data may not be available in convenient form in every metropolitan area This is often the case in urban areas outside the United States and Western Europe as well as in smaller areas in the U S In such cases it has been possible in past tests and applications to do useful planning and analyses with somewhat reduced data sets One example of using a reduced data set is to substitute for zonal employment data by type with zonal employment totals for the current or lagged year time periods A complete absence of employment data by place of work would make it impossible to run TELUM There are differences in the calibration vis a vis the forecasting and data requirements as well For the purpose of calibration a reduced employment data set of zonal totals is evenly divided between each of the region s employment types Once calibrated the model can produce forecasts using projected regional employment totals by type as an input for the model The model estimates where employment 4 4 and households are most likely to locate based on the initial attractiveness parameters found when the model was fitted
110. f consistency checking that assures the user that she is ready to proceed to forecasting A failure in TELUM s ability to perform a File Check Employment and Household Calibration and or Land Use Consumption Regression means there is a problem with your TELUM project and or associated project data and forecasting cannot be attempted 2 Start Model Calibration File Check File Check is a way for TELUM to check the consistency of your data files in preparation for model calibration and later forecasting If TELUM detects a problem with one of your files or finds a file is missing you will not be able to run model calibration An error in File Check is typically associated with your DOPU and TIPU data inputs In most cases users are asked to review and re run the DOPU Data Check and TIPU impedance program before attempting to re run File Check 3 Begin Model Calibration Employment and Household Location Once File Check is complete you can continue directly into model calibration Begin your model calibration by clicking on the GO button when prompted by TELUM Calibration may take several minutes to finish The length of time depends on the size of your region and the ability of the equations to reach an optimum parameter value as well as the computing power of your PC 4 Analysis of Results for Model Calibration Immediately following the conclusion of your model calibrations TELUM provides an Analysis of Results to summarize the mode
111. following sub sections as an aid in determining the data requirements for your region s model calibrations During data entry TELUM will provide you with hyperlinks that explain each requirement in more detail Employment Data For calibration of TELUM Emp it is necessary to have employment data by employment type and by zone for two time points roughly five years apart As stated before in most cases the current time is census year such as 2000 and then the lagged time is taken five years earlier For calibration of TELUM Res it is necessary to have employment data by employment type and zone for only one time point It is customary that the TELUM Emp current year matches the TELUM Res time point For example if the main TELUM Res time point is 2000 the TELUM Emp current year will ideally be 2000 as well The TELUM Emp employment data are the only data that require two time points in the calibration of the models the two time points are necessary for the calibrations only The employment sectors are usually taken as aggregations of the one digit NAICS North American Industry Classification System employment types In most of the recent TELUM Emp applications we used eight employment sectors that closely match the one digit NAICS or SIC codes 4 6 Household and Population Data For calibration of TELUM Emp it is necessary to have the population data usually by household type for one year or time point to match t
112. g the value of time not only for the different locator types such as income levels but for future time periods as well Finally if there are several modes involved all of the above issues are important as well as the question of how to calculate a composite multi modal cost If your TELUM forecasts are intended to be used in a model configuration linked to a traffic assignment package there is also the question of whether to aggregate the networks to match the TELUM zone system or take aggregated skim tree outputs as inputs to subsequent TELUM model forecasts Note that the preparation of the zone to zone travel times costs or composite impedances is not an integral part of TELUM We expect these data to have been developed by your agency during normal activities In this chapter emphasis is placed only on the organization of your impedance file TELUM must be able to read your IMPD TXT file so the values can be converted into a form suitable for use in the model calibration and forecasting units Because there are numerous travel model software packages and because many MPO s have customized the software they have it has not been possible to develop direct software links for the travel models to TELUM What we have done is develop a simple procedure for making this connection that we believe can be used with any travel model software package 2 How to Organize Your Travel Impedance Data Your travel impedance file must have two columns Th
113. gh Income HI LI is type 1 MI is type 2 HMI is type 3 and HI is type 4 The same concept applies to employment In a student project in which there were eight employment types the table below was used to organize a large number of constraints The corresponding employment sequence numbers are noted in the Type category The constraint type value and year were included to best facilitate accurate data entry into TELUM E3 Microsoft Excel ENVConstraints dbf H File Edit View Insert Format Tools Data Window Help Adobe PDF 5 2 1 YEAR Constraint Type Emp HH Zone Type Value 2 2005 3 Employment 153 1 4 3 2005 3 Employment 153 2 13 4 5 3 Employment 153 3 15 5 2005 3 Employment 153 4 25 6 2005 3 Employment 153 5 38 7 2005 3 Employment 153 6 18 8 5 3 Employment 153 7 24 9 5 3 Employment 153 8 16 10 2005 3 Employment 154 1 3 11 2005 3 Employment 154 2 3 12 2005 3 Employment 154 3 9 13 2005 3 Employment 154 4 0 14 5 3 Employment 154 5 4 15 2005 3 Employment 154 6 2 16 2005 3 Employment 154 7 9 17 5 3 Employment 154 8 3 18 2005 3 Employment 155 1 0 19 2005 3 Employment 166 2 0 20 2005 3 Employment 6 3 0 n ADE ENVConstraints Se rt Ready NUM 8 7 Type I Absolute Constraint by Specific Parameter An absolute constraint is a total zonal value that can be imposed for any employment or household variable Specified zones will be forecasted
114. he employment current year For calibration of TELUM Res it is also necessary to have the population data by household type for one point and again it is usually the same as the TELUM Emp current year The population data are usually derived from the decennial population census TELUM Res allows the use of lagged household variables which require household data for a prior time period normally five years earlier We strongly recommend that a lagged household total variable be used in calibration The household data by zone are divided into household types These are usually households by income category such as low income low middle income etc Most previous applications of DRAM the predecessor of TELUM Res have used four or five household types roughly corresponding to income quartiles or quintiles TELUM Res can handle as many as eight household types allowing for a greater number of income groups or for example a cross tabulation of income and life cycle In an application for Chicago income groups were divided into eight household categories In a Detroit application households were divided into four income groups and further subdivided into households with or without children for a total of eight household types Land Use Data For calibration of TELUM Res and the LANCON land consumption model it is necessary to have a current year data set of land use by category in each zone The categories are relatively straightforwar
115. here er elasticity of type n households to changes in residential land in zone i ai a matrix of conversion coefficients of type n households per type k employees Ef employment of type k place of work in zone j the calibrated TELUM RES parameter for residential land Li residential land in zone 1 pi the probability of a type n household with an employed head of household in zone j residing in zone i and households of type n residing in zone i For TELUM RES the location elasticity values for travel time are defined for a increase in the travel time for trips from all employment zones to the specified residential zone The equation for the location elasticity for travel time is as follows puc D Caney gt Je pr d p 6 66 Ni J where c elasticity of type n households to changes in travel times from all employment zones to residential zone 1 cj travel time between zones i and j and a B the calibrated DRAM parameters for travel time TELUM EMP Location Elasticity Values The TELUM EMP location elasticity values are exactly analogous to the TELUM RES location elasticity values However because TELUM EMP has an additive lag term the elasticity values must be multiplied by the potential term weighting parameter Because the lagged employment variable appears in both the potential term and the lag term of 7 6 TELUM EMP the location elasticity for lagged employ
116. icance as a substitute for the measures more readily calculated in the estimation of parameters of linear models with normally distributed variables The maximum likelihood estimator when correctly calculated is asymptotically normally distributed with a mean equal to the true parameter value and with a covariance matrix that can be calculated by use of second order partial derivatives These derivatives are calculated as part of the parameter estimation procedure and allow the computation of asymptotic t statistics that yield an indication of the significance of the individual parameters in the models equation structures Since the TELUM application is applied mostly for regions with more than 100 zones a good rule of thumb is that an asymptotic t value greater than 2 00 is an indication of a statistically significant parameter value TELUM RES Location Elasticity Values Each of the TELUM RES location elasticity values have the same mathematical definition except for travel time For the percentage of developable land developed and the household percentage variables the location elasticity values are defined for changes in one plus the value of the variable For example if the percentage of developable land 75 developed equals 66 the TELUM RES attractiveness variable is equal to 1 66 A 1 increase in this variable is equal to 0 0166 Location elasticity for any attractiveness variable shown for residential land a EN L ERER w
117. ice Retail 3 48 70 values 11 gt 5 82 j correct 3d 1154 1 16 0 560 201 2 82 32 76 90 3 25 12 15 12 4 180 32 32 5 55 16 130 98 6 36 10 169 87 7 1 27 35 Are these 8 9 e TOTAL 442 463 1303 822 3030 Page 4 8 18 ck Ck ck kc KK KKK BASE YEAR 2000 INPUT DATA HOUSEHOLD TYPE 1 2 3 4 5 6 j Are these 8 values 9 correct p e TOTAL 675 1009 848 497 3030 CK ck ck Ck KK KK KKK KK NOTE ATTRACTIVENESS K FACTORS INCLUDED K FACTORS MULTIPLIED BY 60
118. ield that contains the zone number For your Rancho Carne project select the ID field NOTE If you do not select the ID field when prompted as you enter MAP IT for the first time your data will not map correctly You can change the appearance settings of your maps as desired The following are examples of the Highest Total Employment and Upper Middle Income Household Location maps from the Rancho Carne data Check Zones gt File Edit View Insert Selection Tools Window Help E f Layers M zone D Zone With The Highest Total Employment 2 600 i it P 2 9 Tri acm p 05 7 06 Inches Figure 5 MAP IT Output 1 3 8 x Data Check File Edit View Insert Selection Tools Window Help UMI 2000 gt 16 to 157 E gt 157 to 314 Bl gt 314 to 1 256 gt 1 256 to 1 413 5 5 55 5 38 Zonal Totals 310 16 16to 157 BEL gt 157 314 B gt 31410 1 256 WB gt 1 25610 1 413 iss 0 09 8 42 Inches Figure 6 MAP IT Output 2 When you have finished working in MAP IT please exit ArcView to return to TELUM After you have viewed Check Zones in MAP IT TELUM will prompt you to answer three questions regarding your maps For the purpose of this tutorial please answer Yes to all of the Check Zone questions
119. iew of land use modeling and how TELUM can be used in your region While this background information is not required for running the TELUM system you may find these sections informative 6 Land Use Modeling Overview Phenomena as complex as the location of jobs and people in a large region require complex analysis tools In recent years a number of regional planning agencies in the US have carried out the process of implementing forecasting models of employment and household location and land use both for the purpose of doing forecasts and policy analyses as well as for the purpose of providing inputs to their transportation and air quality modeling efforts The overall approach as embodied in a package of computer programs and procedures involves several major components These are 1 first procedures for forecasting the spatial location of employment and households in a metropolitan region 2 second a procedure for using these location forecasts to produce a set of origin destination trip matrices 3 third a procedure when appropriate for doing mode split analysis 4 fourth a procedure for assigning in most cases only highway trips to a capacity constrained highway network and 5 fifth a set of procedures for linking the congested travel times back to the employment and household forecasting procedures Such an approach overall is the only one that allows for explicit representation analysis and evaluation of the effects on
120. k kokck k ko k ko k ok LOOP GRADIENT FIRST STEP SIZE CRITERION OUTER ITERATION 19 1 27 03241611 m 2 1 26 01169534 Criterion values 3 53 04695823 PARAMETER 1 Alpha LOOP 1 LOOP 2 PARAMETER VALUES 2 7275 2 7587 NORMALIZED DERIVATIVES 5655 2564 DERIVATIVE VALUE I N w N 63 0 50 WwW CO W PARAMETER 2 Beta p o o tu w LOOP 1 LOOP 2 PARAMETER VALUES 4 8097 4 8514 NORMALIZED DERIVATIVES 7561 1348 DERIVATIVE VALU Co Ds Co xj ca N N 4 65 O w BON BOW PARAMETER 3 Empl LOOP 1 LOOP 2 PARAMETER VALUES 7 4244 7 4271 NORMALIZED DERIVATIVES 0494 2485 DERIVATIVE VALUES 0 PARAMETER 4 Land Derivative Values LOOP 1 LOOP 2 LOOP 3 4 0013 4 0191 4 0187 23230 PARAMETER VALUE NORMALIZED DERIVATIV DERIVATIVE VALUE N ce C oo 2 e PARAMETER 5 Lambda LOOP 1 LOOP 2 LOOP 3 PARAMETER VALUES 5748 5726 5909 NORMALIZED DERIVATIVES 0414 9241 6429 DERIVATIVE VALUES 0 1 2 Co APPARENT RIDGE BETWEEN PARAMETERS The gradient search technique may be thought of as a mathe
121. l numbers we had 100 employed residents 5 unemployed but employable residents and 70 households This gave 105 employable employees per 70 households for a ratio of 1 5 employees per household EMPHH If we divide our 105 employed plus employable from the previous step we close the loop by getting our original 70 households Formulation of the TELUM Res Conversion Procedure uses the following equation to convert employees at place of work to households by income group at place of work HH Eun s 1 0 CNV x LEO i RJPE 1 0 UNEMP EMPHH where HH the number of households in income group j EMPi the number of employees in industry type 1 RNCR the regional net commuting rate RJPE the regional ratio of jobs per employee UNEMP the unemployment rate for industry type i 4 17 CNV j the percentage of employees in industry type i who belong to household income group j Note CNVj is an element of the employee to household conversion matrix EMPHH the ratio of employees per household for income group j It should be noted that The purpose of multiplying by the regional net commuting rate 15 to insure that when the final conversion procedure is applied to the employment forecasts from TELUM Emp the proper number of employees living in the region 1s used The purpose of dividing by the ratio of jobs per employee is to adjust the employment figures to account for persons holding more than one job
122. l s fit to your data inputs One of the first measures used for best fit analysis is the Best Worst Likelihood Ratio which 15 a normalized maximum likelihood criterion see the Chapter Appendix for more details In linear multiple regression analysis the best fit is measured by the R criterion In addition to the best fit criterion the Analysis of Results shows the statistical significance of the parameters obtained The statistical 7 1 significance here is measured with asymptotic t tests If the absolute values of the t statistics are too low they indicate that an equation coefficient is not statistically significant If the t values are sufficiently large generally greater than 2 00 for most data sets they indicate that the equation coefficient is likely to be statistically significant The formulas associated with these measures are provided in the Appendix to this Chapter Analysis of MAPE and MARMO Results Another set of goodness of fit measures examines the distribution of residuals or errors between the observed data and the models current best fit estimates A commonly seen form is the Mean Absolute Percent Error or MAPE This is the average mean of the absolute values of percent error between an observed set of say household data and the values that would be estimated by DRAM Unfortunately the value of MAPE can be easily distorted by large percentage errors in small zones For example if a zone with an observed ten ho
123. l selected for assembling data and performing analyses the level of detail will not satisfy every user s needs The greater the degree of geographic detail in the data i e the smaller the individual zones or analysis areas the greater the cost of obtaining the data the greater the required complexity of the model and inevitably the lower the statistical reliability of the forecasts Prior models have been executed at several levels of detail but the majority of applications have been in roughly the same zone size range The Houston data set used by the Houston Galveston Area Council some years ago contained five counties that were divided into 199 analysis zones Some of these zones were aggregates of just a few census tracts while others were somewhat larger At the rural edges of the region some zones were large in area but relatively low in population and employment A similar scale of analysis was used for the San Diego region which contained only one very large county but was divided into 161 zones Here too some zones contained few census tracts while others were aggregates of quite a few The comments regarding zone area vis vis zonal activity levels apply here Analyses of the Washington D C region were also done by an aggregation of several counties divided into 182 analysis zones Successful results have been obtained at the census tract level of detail for regions such as Atlanta Colorado Springs Kansas City and Sacrament
124. l type into the uation P DRAM control card for Parameters Trip ASYMPTOTIC Function STANDARD SYMPTOTIC forecasting ERRORS T VALUES ALPHA 4158 20 97 BETA 1020 23 36 3 VACDEV 1807 1 04 PERDEV 1 7797 16 RESLND 1105 13 55 HH LIHH 5301 18 80 io LMIHH 9808 1 68 UMIHH 4955 2 59 Ai UIHH 1 0772 3 41 Att LAGHH 0559 15 06 A goodness of fit measure Statistically R SQUARE UE FOR COMPARISON this 1s not as useful as the next one RSQ 7792 This is a better measure of goodness of fit of the EMPAL equation to B W LR 8479 this data EST WORST LIKELIHOOD RATIO THE RANGE OF THE LIKELIHOOD RATIO IS BETWEEN 0 0000 AND 1 0000 FOR A PERFECT FIT THE LIKELIHOOD RATIO WOULD BE EQUAL TO 1 0000 7 30 ck ELASTICITIES 1 0 1 0 ZONAL LOCATION ko kc ko ko Ck ck ck CK 00 41 00 TD OY 00 4 These are measures zone by z
125. le is developed in a Microsoft Excel spreadsheet you can format the columns and export the table into a text file following this procedure 1 In your Excel spreadsheet select the two columns with data usually the first two columns A and B and set the column width to 10 2 While columns are still selected go to Format gt Cell gt Number tab and select Number from the Category list Click OK 3 Click on File gt Save As Choose the file name and in the drop down menu in the Save as type box select Formatted Text Space delimited 4 Click Save and confirm on the following two screens 6 2 5 Locate your saved text file it should have an extension prn Open it in the Notepad 6 In the Notepad click on File gt Save as Click Save If you wish change the file name The file should be saved with the txt extension 7 Close the Notepad 8 Open the saved file and check if your data is saved as described in the TELUM Manual page 6 2 two 10 space columns If yes you are good to go further Save the file in the Data folder inside TELUM directory and rename the file to IMPD TXT 3 Enter Your IMPD txt File into TELUM 1 Place your completed IMPD txt file inside the TELUM project folder labeled DATA C TELUM DATA 2 TIPU performs an impedance file check to evaluate and confirm the format of your zonal impedance values In order to perform this check you must gather the fol
126. le practice and will reduce the danger of errors in your forecast Based on extensive testing it is extremely unlikely that you will be able to successfully complete a forecast without the operating skills gained through running the TELUM Tutorial Instructions for installing the TELUM Tutorial are covered in the next chapter of the manual 4 User Support For user support please visit TELUS website at www telus national org If you are an MPO State DOT or other public transportation or planning agency in the United States you can also contact TELUS development team Transportation Economic and Land Use System TELUS New Jersey Institute of Technology Tiernan Hall Suite 287 University Heights Newark NJ 07102 E mail telus njit edu Phone 973 596 5700 Fax 973 596 6454 2 6 If you are not an MPO State DOT or other public transportation or planning agency in the United States please contact Federal Highway Administration FHWA Resource Center for more information about user support for TELUM FHWA Resource Center Lisa Randall Planning Technical Service Team Leader 12300 West Dakota Avenue Suite 340 Lakewood CO 80228 Phone 720 963 3209 Fax 720 963 3232 lisa randall dot gov 2 7 3 TELUM Tutorial Tutorial Contents 1 Introduction 2 Initial Data Entry Unit DEU 3 Data Organization and Preparation Unit MAP IT DOPU 4 Travel Impedance Preparation Unit TIPU 5 Model Calibration a
127. ll ask you to indicate the change s and or data you wish to add for your new model forecast Select all options that apply You can choose to change employment and household projections add future year impedance files or impose constraints on regional employment and household data by zone You must make a change otherwise there is no reason for you to be running a new model forecast 5 TELUM will automatically transport you to where you have indicated changes are necessary Make your data changes and proceed back through the system to MFCU During this time we suggest you do not exit TELUM until you have completed your new model forecast 6 Once you return to MFCU continue through the TELUM screens to Run Model Forecast Select GO 7 You have completed a New model forecast Now you can review your Spatial Analysis and Map It Forecasting maps Changing Your Model Forecast Inputs Your model forecast inputs as you will recall are the DOPU zonal data TIPU impedance file and MCPU model parameter values Often agencies must change and or add data inputs to reflect a policy or updated data that becomes available within their modeling region This is a typical modeling function A change is necessary when Achange is made to the zonal and or regional DOPU data inputs e Anew or modified base year impedance file is added to TIPU Achange is made in the region s Employment and or Population projections e The agency has imp
128. ll other zones at time t Following the employment location forecasts produced by TELUM EMP a set of residence location forecasts is produced by TELUM RES The model is normally used for 4 to 5 household types usually income groups whose parameters are individually estimated A separate submodel within TELUM RES called LANCON calculates land consumption using a multiple regression based procedure for making a simple reconciliation of the demand for location by employers and households with the supply of land in each zone To forecast the location of residents of type h in zone i at time t 1 TELUM RES uses the following input variables e residents of all h types in zone i at timet e Jand used for residential purposes in zone i at timet e the percentage of the developable land in zone i that has already been developed at time t e the vacant developable land in zone i at time t zone to zone travel cost or time between zone i and all other zones at time t 1 e employment of all k types in all zones at time 1 The residence location forecasts produced by TELUM RES may then be used sometimes after a further step of spatial disaggregation as input to procedures exogenous to TELUM that generate and distribute trips split trips by mode and then assign vehicle trips to the transportation network s Many different configurations of land use and transportation linkages have been tested by the use of ITLUP and more recently TELUM While
129. low data inputs for TELUM a Average Zone to Zone Impedance The average is a simple algebraic mean of your region s zone to zone travel impedances b Smallest Zone to Zone Impedance Find your region s smallest zone to zone impedance value c Largest Zone to Zone Impedance Find your region s largest zone to zone impedance value d Top 4x4 Impedance Values Collect the 16 impedance values from the top left corner of your travel impedance matrix as pictured above in Table 1 e Bottom 4x4 Impedance Values Collect the 16 impedance values from the bottom right corner of your travel impedance matrix 4 Complete the Travel Impedance File Check With all of your impedance inputs entered TELUM will ask you to run the Impedance File Check File Check begins by looking for your IMPD txt inside the DATA folder If your IMPD txt is not inside the DATA folder or TELUM is unable to read your IMPD txt then TIPU will stop File Check and provide you with troubleshooting instructions You must complete TIPU before you can continue into model calibration 3 Future Year Travel Impedance Forecast Travel Impedances provide you with the option to include your Transportation Department s impedance files for future time periods Adding impedance file s in addition to your current TIPU impedance file is strictly optional Forecasts can be made based on your current year zonal impedances 6 3 However if you use the current year impedance
130. matical procedure for finding the highest point on a virtual mountaintop in a multidimensional space It progresses by calculating the direction of steepest ascent up the mountain from a particular point All the parameters are changed simultaneously in proportion to their normalized derivatives in order to move up the mountain When a step is taken having changed all the parameters the criterion is re evalulated in order to determine whether we are further up the mountain If so we take another 7 13 number of steps in the same direction we may find by recalculating the criterion that we have gone too far and are over the crest of the mountain and are moving back down In that case we calculate a new direction of steepest ascent and move off in that direction Even if the criterion does not decrease after a set number of steps in one direction a new direction of steepest ascent is calculated It is possible to encounter a long curved steep sided ridge on our way up the mountain The gradient search procedure would slow down appreciably in such a case as its search path would amount to numerous single steps back and forth over the ridge in a zig zag pattern CALIB can detect such a situation and will do the following a print a message apparent ridge and b use a different procedure
131. ment must be defined in terms of the values of the TELUM EMP potential term and scaled lagged employment Location elasticity for any attractiveness variable other than lagged employment shown for total land area in TELUM EMP is defined as k Bh Lj AE AE i t m 1 os 7 EL L En 2 Ni E p p 7 where el elasticity of type k employment to changes in land area in zone J the calibrated TELUM EMP potential weighting term b the calibrated TELUM EMP parameter for total land area L total area of zone j the probability of type k employment locating in zone j with work to home trips terminating in zone 1 and Exe employment of type k in zone j at time t un weighted potential term The equation for location elasticity for lagged employment is as follows 4 MB Nu ae 01 ph G 3 Bfa i E 0 ER 2 1 E 8 eux m 5 E 0 Ef Ek E 1 M 9 E where 1 c elasticity of type k employment to changes lagged employment jit 1 the calibrated TELUM EMP parameter for lagged employment and employment of type k in zone j at time 1 1 scaled lag term For TELUM EMP the location elasticity for travel time is defined to a 1 increase in the travel time for trips from all residential zones to the specified employment zone The equation for the location elasticity for travel time is 9 0 x j k y k gt 55
132. mewhere someone will provide an estimate of the zone to zone travel time and travel cost on various modes that a user might experience let us say in the year 2010 or 2020 Based on these estimates as well as on base year data regarding the initial locations of employment and households and on a set of regional forecasts of total employment and total households a calculation can be made that will estimate their location in the zones of the region Often a whole series of such forecasts will be made at five or ten year intervals from some base year out to some long term planning horizon The defect in this approach which is analogous to the defect in the traditional transportation planning approach is that no cognizance is given to the fact that the locations of employees and households will by virtue of the trips necessary to interconnect them congest the network The congested network times will in most cases be somewhat if not significantly different from the initial estimates of the network times What is needed is an interactive procedure that includes both the effects of the location of employees and households on the transportation system as well as the variation of the location of 1 7 employees and households caused by congestion induced changes in the transportation system characteristics It is precisely this integrated interactive process that the original ITLUP model system was designed to properly represent It is this transpo
133. mize disruptions to the health of your computer operating system it is imperative that users this means you follow this manual closely This system links to Microsoft s Excel to Microsoft s Access to other system utilities such as Wordpad and to ESRI s ArcView as well as to system components for Visual Basic and for FORTRAN computing languages For your ease of use there are over 500 User Interface Screens along with all the computer code to connect them Successful installation and use of a system this complex REQUIRES that you read and follow the instructions carefully 1 Overview of the TELUM Manual This manual is organized in the following manner You will find it useful to read each chapter thoroughly before starting the corresponding section of TELUM 1 Introduction History and Basics of TELUM 2 Installation Setting Up and Configuring TELUM on Your Computer 3 TELUM Tutorial Running a Sample Forecasting Project to Provide Familiarity and Experience in Using TELUM 4 Data Preparation Gathering Formatting and Entering Regional and Zonal Data 5 MAP IT Using GIS to Analyze and Forecast Results 6 Travel Impedance Preparation Unit Format and Input of Travel Impedance Files 7 Model Calibration Checking Data Estimating Equation Parameters and Explaining Calibration Results 8 Forecasting Explaining the Forecasting Process Implementing Baseline and Policy Forecasts 2 What can you do with TELUM TELU
134. nd Preparation Unit MCPU 6 Model Forecasting Unit MFCU 1 Introduction Welcome to the TELUM Tutorial This tutorial is a 20 zone student TELUM project for Rancho Carne a fictitious California coastal region Please use this document and the accompanying data inside the Tutorial folder to walk through the Rancho Carne project The Tutorial folder is located on your TELUM CD Rom This data set was originally developed by Daniel Schack and Leah Wright as one of the requirements for a course on Urban Simulation Modeling taught by Dr Putman It has since been modified to improve its teaching function How to use this Tutorial This tutorial is intended to provide an overview of TELUM as well as practice in data entry and output comprehension The Tutorial folder contains all the files and data you will need to complete a calibration and a set of model forecasts for the Rancho Carne region As you work through this project hyperlinks are available to help explain the various functions and data inputs more thoroughly than you will see here Throughout this tutorial you will see screen numbers used to refer to various sections of the TELUM software These screen numbers are located in the lower left corner of each TELUM screen e g P3 15 1 If you have a question or concern while you are working with TELUM you must record the screen number where the problem occurred Referencing a screen number improves the TELUM staff s ability to respond to your
135. nd surplus households or jobs are distributed to the next more desirable zones Enter maximum constraints in the same manner as prior type in the worksheet screen below 8 9 11x Model Forecasting Unit 2005 Absolute Maximum Household Constraint By Type and Zone Enter your constraints in the table below COPY CONSTRAINTS FROM PRIOR YEAR a 9 um Type 4 Minimum Constraints A minimum constraint sets a minimum value to the amount of employment or households forecasted in a zone by type The amount of households or employment in that zone is affected by the constraint only when the value is less than the input minimum The distribution of employees and or households is thus limited and redistributed to other zones only when the household or employment value does not meet minimum constraints Enter minimum constraints as before in the worksheet screen shown below 8 10 BET Model Forecasting Unit 2005 Absolute Minimum Employee Constraint By Type and Zone your constraints in the table below COPY CONSTRAINTS FROM PRIOR YEAR 9 mm m After all desired constraint types are entered TELUM will provide a Constraint Summary Report as shown below If there was an error in entering or applying a constraint using the BACK button will enable you to re entered or modify constraints for all time periods and types 8 11 Ea TELUM Model Forecasti
136. ng Unit 2005 Constraint Summary Report Maximum Employee Household by Employees Households by Type and Zone Type and Zone 400 emphh zone e value Employee 153 4 Employee 1153 1 Employee 153 Employee 153 Employee 153 Employee 153 Fmnloiee 1153 Total Employees Households by Employees Households by Type and Zone Type and Zone EE TELUM will then ask the you to verify a File Check in preparation for forecasting with the new constraint values Click GO if constraint entry is accurate and complete 5 Making Use of Unobserved Factors in Forecasting Residuals represent unobserved factors that influence employment and household location but are not captured in the TELUM EMP and TELUM RES model formulations Residuals are a means to capture the information that is contained in errors made by TELUM EMP and TELUM RES when predicting the location of employment and households for the base year In fact the mathematical formulations of TELUM EMP and TELUM RES allow us to make perfect estimates of base year employment and household location when residuals are not attenuated 1 e the attenuation parameter is set to 1 00 for a verification run However this does not mean that it is possible to make perfect forecasts of employment and household location in future time periods Since residuals are determined from base year model calibrations these unobserved factors have an implicit temporal specification As an example
137. o The location of regional boundaries also must be addressed Again there is not a rigidly set definition of a regional boundary In general it is best to try to set the boundary so that the amount of economic and social interaction crossing it is as small as possible It will not always be possible to achieve this goal and the models have provisions for including constrained external zones to address this problem Perhaps the most extreme case was illustrated in Orange County CA in a project done in the late 1980 s where the greatest interest was in the analysis zones within and immediately adjacent to the County Yet the County is closely tied to the rest of the Southern California region In this application a zone system was developed with small zones in and near Orange County and progressively larger zones outside The 4 5 external zones were constrained in the model runs The calibration results from this hybrid data set turned out quite well 2 Temporal Detail TELUM calculates zonal forecasts in five year increments starting five years beyond the current data year e g 2000 yielding forecasts for 2005 2010 2015 etc Virtually all tests of the predecessors of TELUM Emp and TELUM Res used five year or approximately five year increments There is little chance that sufficient data will become available in the near future to enable the use of shorter time periods While some attempts have been made to try to work with ten ye
138. o G9 Oo o Ool ol Employees per Household by Income Category 1 HHI HH2 HH3 HH4 HH5 HH6 HH7 HH8 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 Number of Households by Income Category HHI HH2 HH3 HH4 HHS HH6 HH7 HH8 Total Number of Employees by Income Category HHI HH2 HH3 HH4 5 HH6 HH7 HH8 Total Scaled Number of Employees by Income Category Household Target HHI HH2 HH3 HH4 HH5 HH6 HH7 HH8 Total 0 0 0 0 0 0 0 0 0 Figure 9 TELUM DATAPREP Conversion Matrix Worksheet 1 Extreme values of the other conversion ratios RNCR RJPE UNEMP and EMPHHj may make it impossible to compute an employee to household conversion matrix CNVi which ensures consistency In these cases it is likely that errors were made in the estimation of one or more of the conversion ratios 4 15 Adjusted Industry Emp UNEMP Emp Industry Target 1 00 1 U 5 5 5 5 5 5 OO Total Industry HHI HH2 HH3 HH4 HHS HH6 HH7 HH8 Total e 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 1 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 0 0000 00 1
139. of TELUM and when you have gotten to the Data Organization and Preparation Unit DOPU with your new dataprep xls spreadsheet open you will be able to transfer your data using Copy and Paste Special Values Open the old renamed dataprep xls directly outside of TELUM At the prompt click No The spreadsheet will open and you will be able to access your data NEVER copy an old TELUM spreadsheet into a newer version of the program The spreadsheets are not compatible and will cause malfunctions in your project FYI a safe way to uninstall former versions of TELUM is via the standard procedure that Microsoft Windows offers using the Add Remove Programs option 2 Go to Start Settings Control Panel Add Remove Programs Scroll down until you find TELUM Click TELUM and then click on the Change Remove button In addition go to TELUM folder and erase the whole folder How to Install TELUM Once you have checked off ALL of the above TELUM project requirements begin your installation by double clicking on the SETUP exe 2 Ifyou are installing TELUM from TELUS CD during installation TELUM will ask you to select version of ArcGIS you are using If you are using any of the 9 x versions please select ArcGIS 9 0 If you do not have ArcGIS installed on your computer please select I do not have ArcGIS 8 3 or higher Click Install button after making the above selections OR If you are downloading in
140. one of the sensitivity of this locator type Household Type 1 to each attractiveness variable 410900 o o 1 ELASTICITIES 1 0 1 0 ZONAL LOCATION LAGHH 2643 53997 0836 1840 1596 1937 1824 1922 1844 3006 2 1047 1755 0613 7 31 UIHH 8154 249175 1758 4788 3496 6977 1524 3860 6119 0725 6575 4283 2457 2 2 2 2 2 2 2 2 2 3 25 2 UMIHH 9834 0191 7600 8658 8207 9423 7518 8334 9123 0732 9620 2 0858 ZONE 300 0 41 00 O m TOTAL MEAN STD DEV KKK
141. ong run it has often been the case that an improvement in a transportation system most frequently in terms of highway construction while having a short term effect of improving the situation for travelers has a long term effect of doing just the opposite Indeed one of the consequences of highway construction in the absence of demand management or urban design in an attempt to in some way regulate land use and land development has been to spread greater network congestion over a larger number of links in the network The traditional transportation planning approach makes it very difficult to anticipate these kinds of system responses to particular policies In this analysis process a series of externally produced estimates of trip demands usually in the form of origin destination trip matrices is calculated using exogenously estimated sets of socio economic data So for example let us consider an agency preparing in 2004 long term transportation plans for the year 2030 or 2040 or beyond In such a case typically a series of socio economic forecasts in terms of employment locations and household locations spatially distributed over a 1 6 large region would be prepared first These would be based upon information about the highway system that the region was expected to have though in fact there would be even at this early stage in the process an inconsistency because the system that the region would be expected to have would show differen
142. or travel time can be forecasted for comparison with other policies and the baseline These changes are reflected by TELUM when entered as new population or employment projections updated impedance files or household employment and location constraints Constraints and changes in impedance values reflect regional and zonal policies that can alter the pattern of employment and household location Constraints serve two important roles in the forecasting process They limit growth in zones that are either increasing too fast due to model errors or increasing at a rate that deviates grossly from local knowledge Examples of local knowledge include the protection of environmentally sensitive or farm lands future increases or limits on households or employment values due to planned development or limits on infrastructure expansions There are four types of Employment and Household Constraints Absolute Constraints Total Constraints and Maximum and Minimum Constraints Constraint Data Entry Preparation Before entering constraints it is helpful to organize a sample input table that resembles the one below with Zone Type and Value columns as well as year constraint type and employment or household designation The Type category reflects not the constraint type as Type 1 Absolute Constraint but the employment or household category number For instance with four household categories Low Income LD Middle Income MD High Middle Income HMI and Hi
143. orking with your DOPU Workbook You may enter the Rancho Carne data into the TELUM DOPU Employment Household Land Use and Projection Worksheets manually or by using the copy option NOTE The data will paste correctly only if you use the paste special options in Excel Import data from your worksheet using paste special as values You need to import data into the blue cells only in your DOPU worksheets Once your data is correctly imported your zonal household population and employment numbers will appear in the worksheet An example of these worksheets as you will first see them before data entry into the DOPU workbook follows Households 2000 e 1 Toal 2 ol Pics Jone HHI HH2 HH3 HH4 HHS HH6 7 H P it Household 7 oa one 2000 2000 2000 2000 2000 2000 2000 Ouse OuP Population pe 2000 Quarters 2000 Household 2000 o ow o ow o ow o ow Figure 2 TELUM DATAPREP Households Worksheet for Current Year Zonal Data 3 5 ER Figure 3 TELUM DATAPREP Projections Worksheet NOTE The data will paste correctly as values only if you open both the DOPU Workbook and the RC DOPU DATA xls in the same instance of Excel Do this by opening the DOPU workbook through TELUM as instructed Then open your regional data workbook from which you will paste your data by choosing File gt Open and browsing to locate the other file After pasting save your DO
144. ouble check your data entries and agency sources for accuracy In the Racho Carne example there is a low correlation for total households vs employment as place of residence in this region do not correlate with 3 7 place of work 2 MAP IT Check Zones and Data Check If you have ArcView 8 3 installed on your computer system you have access to TELUM MAP IT a mapping tool for visually displaying your data and your calibration and forecasting results In this section TELUM will ask you if you wish to use MAP IT If you select No this will be the last time TELUM presents MAP IT as an option If you should decide to use MAP IT later you will need to revisit the TELUM DOPU section and select the Yes option If you select Yes MAP IT will require a set of regional shapefiles inside the main TELUM folder Inside your Tutorial folder is a set of shapefiles for Rancho Carne labeled ZONE SHP ZONE DBF ZONE SHP XML ZONE SHX Place these four files inside the main TELUM folder before enabling the MAP IT function Launch MAP IT TELUM begins MAP IT by instructing users to complete a mapping check called Check Zones Check Zones is used to ensure mapping consistency If they find their check zone maps to be inaccurate the user must correct the problem before this feature can be enabled for future use MAP IT will prompt you to select the shapefile f
145. pedance In some instances especially for educational projects it will be necessary to use geometric distances between zone centroids to approximate travel impedances Here we provide notes on calculating the location of zone centroids in ArcView Once the centroids are calculated a matrix may be created by simply calculating the centroid to centroid distances The diagonal elements of the matrix may be calculated by assuming that each geographic zone can be approximated by a circle Then Arc View can give you the zone s area and you can calculate backwards using the formula for the area of a circle to get an implicit zone radius to use as the intrazonal distance which is an acceptable approximation for the distance matrix diagonal values The centroid for a polygon is the geometric center listed in terms of two points the x coordinate and the y coordinate The calculation of the centroid in ArcGIS requires the addition of two fields in shapefile form containing the zones of your region and the use of a Visual Basic A script to calculate the two coordinates of the point 1 Add the fields that will hold the coordinates of the centroid for each zone Open the attribute table in your regions shapefile Click on the Options button at the bottom of the attributes table window and select Add Field Name the field XCoord set the type to Double click OK Repeat naming the second field YCoord 2 Start an
146. ployment data are required by place of work For the purpose of calibration of the employment model TELUM Emp employment data are required for two different time periods The ideal household and land use data are for a census year such as 1990 while the second lag time point of employment data as well as the lagged total households are for five years earlier The following is a list of spatially disaggregated input requirements for TELUM Res and TELUM Emp These requirements are for data for each zone or district in the region The design of these zone systems is not a trivial matter but it is often a matter over which the analyst has little or no control These data are required only for the base or starting year for a forecast and also for a lag year where noted for calibration In subsequent forecasts the outputs of each simulated time period become the inputs to the next TELUM EMP e Households by type e Employment by type current and lagged values e Total land area e Land area occupied by basic employment e Land area occupied by commercial employment e Zone to zone travel times and or costs TELUM RES e Households by type current and lagged lagged can be total households by zone e Total population e Total employed residents e Group quarters population e Total households e Total land area e Unusable land undevelopable restricted or reserved e Land area occupied by basic employment industrial e Land area occupied by comm
147. relative attractiveness of other zones in the region Location elasticity values will be larger when the calibrated parameter for the attractiveness variable is large in absolute value the number of households or employees is small relative to other zones in the region or the value of the attractiveness variable is small relative to other zones in the region 5 Land Consumption Calibration LANCON In TELUM land use by locating activities is calculated after the completion of the location demand calculation TELUM EMP calculates location demand by employers followed by the TELUM RES calculation of location demand by households LANCON takes both these calculated demands and estimates the actual change in the amount of land by zone that will be used by each of the demand categories If there has been a decrease in demand by a particular category then land currently in use by that category is released into a pool of land available for any use If there has been an increase in demand by a particular demand category then the addition of land to use by that category is calculated After the calculations are done for each demand category the sum of land used is adjusted by an increase in density to match the land available for such uses TELUM produces a LANCON Statistical Analysis section to highlight your region s land consumption reliability Many regions find their land use reliability to be low This is common and should not deter an ag
148. rtation and land use consistency that is required as input to the air quality estimates now required by the CAA and TEA21 and by ISTEA before that Even the earliest tests of ITLUP done nearly 20 years ago showed that the interrelationships between transportation and land use can be just as important and in some cases more important than the individual direct consequences of either set of phenomena Having articulated a framework for examining or analyzing or understanding the transportation and land use interactions it then becomes possible to consider the consequences of a wide assortment of different kinds of policies For the first time this included policies that attempt to achieve their aims by changes on the demand side in terms of urban design policies land use control policies and the like as well as policies that attempt to achieve their aims by acting on the supply side in terms of various kinds of transportation improvements either in highways or transit or combinations thereof as well as in access and increases in utilization efficiency of existing facilities 7 Forecasting with TELUM Land use forecasting is best done in time increments usually of five years length as this acknowledges the difficulties of obtaining data for any more detailed set of intervals while at the same time allowing for some amount of adjustment of employment residence land use and transportation forecasts in response to each other within these in
149. s for long range forecasting you are assuming that there will be no change in your region s transportation infrastructure How to Add Future Year Travel Impedance Future Year Impedance files are entered into your model forecast s from within MFCU Each time you prepare to re run or run a new model forecast MFCU will ask if you wish to enter IMPD files for future time periods If you indicate that you wish to add future year impedances you may enter as few as one or a number of impedance files equivalent to your forecast time periods 1 Prepare your Future Year Travel Impedance FY TIPU files in the same format as your TIPU IMPD txt file except this time you will include the forecast year in the impedance label e g An impedance file for year 2010 is saved as IMPD10 TXT 2 Place your FY TIPU file s inside your TELUM DATA folder 3 When you enter MFCU for your Baseline model forecast and or for new or re run model forecasts select Yes when TELUM asks if you would like to add Future Year Impedance file s 4 For your FY TIPU impedance files you must collect the same five data inputs for each additional impedance file as described for TIPU File Check earlier in this chapter TELUM will prompt you to enter your File Check inputs Upon successfully creating and File Checking your future impedance files TELUM will move all of your FY TIPU files in preparation for your model forecast 4 Substituting Geometric Distance for Im
150. s possible policy alternatives Throughout this manual we present a mix of theory discussion with instructions for specific model operation Our intent is to enable thoughtful users to make effective use of this powerful analysis tool In the next chapter we provide information on installing TELUM and follow this with a description of data requirements for use of these models We then provide detailed instructions for the preparation of a small test data set to be used for the education of students and new users 1 9 2 Installation of TELUM 1 Setup System Requirements It is imperative that the user follows these setup instructions closely We cannot over emphasize the importance of this preliminary check and setup of your computer Deviation from these instructions will cause your computer and TELUM to malfunction To aid your successful installation of the TELUM system use the following checklist Please review the list and confirm that everything is in order on your computer before you start the TELUM installation We cannot overemphasize the importance of this preliminary check and setup of your computer A successful TELUM installation depends on the following Your computer must be equipped with a registered Windows 2000 with the upgrade to Service Pack 5 or Windows XP with Service Pack 2 and you must have the Microsoft Office software package including full installation of Microsoft Access and Microsoft Excel Microsoft Ex
151. sequent congestion induced changes in travel times It is important to understand that within TELUM the DRAM and EMPAL models borrowed from METROPILUS and now known as TELUM RES and TELUM EMP constitute only a portion of a complete regional transportation location and land use model system Such a system would involve location and land use allocation models as well as a set of transportation analysis models including the steps of trip generation trip distribution mode split and trip assignment All of the agencies making use of TELUM are expected to have their own transportation analysis software already installed and operational The outputs of TELUM then become the inputs to the agency s own travel demand models and trip assignment package e g EMME2 MINUTP TRANPLAN TRANSCAD UTPS etc The congested network travel times and or costs produced as output from these packages may then be used as inputs to subsequent time period forecasts from TELUM Most planning professionals have a good intuitive sense of how employment and household location patterns develop over time and how those patterns are affected by changes in transportation systems Human intuition cannot however encompass all the thousands of data items and interactions that describe transportation location and land use in a metropolitan region Computer models such as TELUM can both process this data in a consistent fashion and by making explicit much of the intuitive unders
152. ser must rerun when one or all of the following apply A change is made to the zonal and or regional DOPU data inputs A new or modified impedance file is added to TIPU A change is needed in the region s Employment and or Population projects The agency wants to add future forecast time period impedance files How to Re run a Baseline Forecast Rerun your Baseline forecast by doing the following Open TELUM and click Skip to MFCU Upon re entering MFCU you will receive a screen asking What you would like to do next in model forecasting Select Rerun Model Forecast TELUM will ask you which forecast you would like to rerun Select Baseline Next TELUM will ask you to indicate the change s you wish to make for your new forecast Select the data component s you wish to change If you do not make a change there is no reason to rerun the Baseline forecast TELUM will automatically transport you to the section where you have indicated that changes are necessary Make your data changes and proceed back through the system to MFCU During this time we suggest you do not exit TELUM until you have completed your new set of Baseline Forecasts Once you re enter MFCU continue through the TELUM screens until you see Run Model Forecast Select GO Once you have completed the Baseline rerun you can review your Spatial Analysis and MAP IT Forecasting maps You may continue to rerun your Baseline forecast until you find a set of for
153. ssion Statistics Multiple R R Square Adjusted R Square Standard Error 227942 Observations 1 Significance 210183 4 MS F ANOVA if F 1 82445 Regression 5 0 958 0 192 4214 o 015122 1 74696 Residual 14 0 637 0 045 Total 19 1 595 cet 1 39198 1 2145 1 03701 0 859523 0 682036 0 504549 0 39318 1 24722 2 10125 2 95529 3 80932 4 66336 Observed Commercial Land Per Employee Lower Upper 95 0 184553 5 318 Coeff St Err fStat P value Intercept LnPerDev LnPerBas LnPerCom InPerLi InPerHI Predicted Commercial Land Per Emple PRINT Page 2 of 3 SCREEN P6 14 Figure 9 LANCON Statistical Report 3 13 Explanation of LANCON findings LANCON calibration is done with multiple regression analysis The results of LANCON calibration are given in terms of goodness of fit of the model to the data In an actual agency project the user might wish to examine land use types and zones for which there were large errors to see whether data corrections would improve results 6 Model Forecasting Unit In this section you will do the following Prepare Files for a Forecast Model Run a Baseline Model Forecast Rerun a Model Forecast Run a Policy Model Forecast MAP IT Map and Compare Forecast Results 1 Prepare Files for a Forecast Model When you enter the Model Forecasting Unit TELUM prepares your files for th
154. stallation files from TELUS TELUM website please select installation package that matches the version of ArcGIS that you currently have on your computer 3 TELUM installation will begin automatically Please follow the instructions on the screen 4 Atthe conclusion of installation your system will reboot automatically unless you are using a Windows XP operating system With Windows XP a reboot is not necessary 23 Running SETUP EXE 1 The program will prompt you to choose an installation directory The default directory is CATELUM Click the Browse button to install the TELUM system on a different directory NOTE Please do not use any spaces in the name of the directory Correct C TELUMNEW Incorrect C TELUM NEW Ei TELUM Destination Location O Setup will install TELUM in the following folder To install into a different folder click Browse and select another folder You can choose not to install TELUM by clicking Cancel to exit Setup Destination Folder C NTELUM Browse Wise Installation Wizard 2 If you click the Browse button you will see a dialog box similar to the one below which allows the user to install TELUM into an alternate directory x Destination Location ucd restore You 01 14telum C3 3dabm amp C3 3dbapi Set Toii addlag C3 adobeapp De arcais ls Copy of TELUM Documents and Settings 2
155. sts of residential location are to be made The Conversion Matrix After TELUM forecasts the location of employees at their place of work it calculates the resulting numbers and types of households which are then allocated to their places of residence This is the forecast of households at their place of residence Use the sample spreadsheet shown below which is in the Workbook DATAPREP to input the number of heads of household by income group that are employed in each industry Also fill in the number of employees per household You will find these numbers using guidelines in the US Census as a basis 11 you choose This number will typically fall within the range of 0 99 to 2 31 A default value of 1 00 may be substituted 1f the data based values are not available for your TELUM project The spreadsheet will calculate the percentage of households in each income group whose employed person works in each industry When converting from employees to households TELUM must account for households associated with persons currently unemployed An unemployment rate is used under the assumption that unemployed individuals made their location decisions while employed and therefore located like others in their income class If you do not have estimates of unemployment rates by employment type then use a default value of 0 0 for the percentage of unemployment The regional ratio of net commutation measures the extent of work trip commuting into or out of th
156. suppose that TELUM RES is calibrated for the base year 1990 Housing prices which are not explicitly included in the TELUM RES model formulation are unobserved factors which influence household location The TELUM RES residuals which are calculated for the TELUM RES base year 2000 calibration contain information about the influence of 2000 housing prices on household location Over time the influence of 2000 housing prices on household location will diminish The location of households in 2005 may be strongly influenced by housing prices in 2000 but by the year 8 12 2015 the influence of 2000 housing prices on household location will be negligible Therefore it is logical to assume that the effect of residuals on household location will also diminish over time For this reason residuals are usually reduced in TELUM EMP and TELUM RES forecast runs 8 13 APPENDIX The Employment and Household Location Model Formulations There are three special features of the TELUM EMP equations 1 a multivariate multi parametric attractiveness function is used 2 a separate weighted lagged variable is included outside the spatial interaction formulation 3 a constraint procedure is included in the model allowing zone and or sector specific constraints The model is normally used for 4 8 employment sectors with individually estimated parameters The equation structure is as follows k k EF AKE Pit
157. t TELUM user support or FHWA Resource Center please refer to Section 4 User Support below for contact information 3 How to Change or Move Your TELUM Project With an installed version of TELUM on your computer you can begin land use modeling activities for your agency If at any time you should decide to change or move your agency project you must re install TELUM onto another computer If you decide to reinstall TELUM on a machine where you are currently running a TELUM project it is imperative that you rename your old TELUM folder or modify the name of the new folder Otherwise TELUM will reinstall itself over the existing C TELUM folder and the resulting file mismatches will render both the original and new versions of TELUM inoperable Circumstances that require a TELUM Reinstallation The following types of project change cannot be accomplished within an existing TELUM project To change any of the following you must reinstall TELUM 1 The number of zones increases or decreases The number of employment and or household categories increase decrease and or change names 3 Your agency builds a new regional GIS shapefile for your TELUM project 4 Your TELUM project is being relocated to a different computer Once installation has been successfully completed we strongly recommend the you run through the Tutorial section in Chapter 3 to familiarize yourself with the data entry and program requirements This tutorial offers valuab
158. t characteristics to users as a function of what the users were doing about using the system i e the traffic flows and congestion In any case a set of forecasts would be developed and then based on the forecasts of the location of employment and households a set of estimates would be made of the number of trips originating from each zone and terminating in each zone Then a trip distribution procedure would be implemented that would calculate the number of trips going from each particular origin zone to each particular destination zone in the region These trips would then be assigned to the links of the proposed highway network Any of a variety of trip assignment algorithms might be used The intention of any of them would be to calculate how many trips would travel across each of the individual links in the highway network Then based on the number of trips using each link an estimate would be made of the congestion the increased time or cost that would be experienced by each of the users of that particular link in the network Once these congestion levels have been calculated for all of the links in the network it becomes possible to trace the minimum cost paths from each zone to each other zone over the congested network Looking at these minimum cost paths as well as the congestion levels on individual links of the system the conventional analysis procedure would then identify links that should have capacity increases which normally would
159. tanding of these phenomena effectively describe these important interactions In addition both TELUM EMP and TELUM RES contain provisions for user augmentation of forecasts This can be done by use of constraints on activity location which will be described in this manual Furthermore and of particular importance here by altering inputs to the models in order to represent policy assumptions policies can be evaluated by the same data rich replicable and behaviorally consistent process 4 TELUM Basics In this document we describe the components of TELUM how to organize a data set for your region how to use the data with the software and how to interpret the results obtained In the manual font styles are used to indicate whether directions refer to menu items directory names file names Excel spreadsheets or command buttons In the box to the right are samples of how these font styles are used Naming instructions for the files you create will be covered in each relevant section of the manual Before you begin using the TELUM application you should know some of the program features available to you These features are built in aids that help you work through the model preparation and forecasting without having to always refer to this manual Operating System TELUM HELP and the User Manual assume that you are proficient in the use of the Windows operating system If you need help with the operating system please consult its
160. tervals Each increment would begin with the execution of TELUM EMP The model is normally used for 4 to 8 employment sectors with individually estimated parameters To forecast the location of employment of type k in zone j at time t 1 TELUM EMP uses the following input variables e Employment of type k in all zones at timet e Population of all types in all zones at time t e Total area per zone for all zones e Zone to zone travel cost or time between zone j and all other zones at time t Following the employment location forecasts produced by TELUM EMP TELUM RES automatically produces a set of residence location forecasts This model is normally used for 4 to 5 household types usually income groups with individually estimated parameters Then the land use submodel LANCON calculates land consumption making a simple reconciliation of the demand for location by employers and households with the supply of land in each zone To forecast the location of type h residents in a zone at time t 1 TELUM RES uses the following input variables e Residents of all h types in zone i at time t the previous time period e Land used for residential purposes in zone i at time t e The percentage of the developable land in zone i that has already been developed at time t e The vacant developable land in zone i at time t e The zone to zone travel cost or time between zone i and all other zones at time t 1 e Employment of all types in all zones
161. the current implementation of the model system does not permit all of these configurations to be examined there is adequate scope for an agency to prepare accurate forecast inputs to the agency travel modeling system Throughout this manual we present a mix of theory discussion with instructions for specific model operation Our intent is to train thoughtful users to utilize this powerful analysis tool Model Forecasting Output Files TELUM stores a large model report that contains a copy of data inputs used and outputs generated for both employment and households during a forecast These reports are lengthy and TELUM provides a summary of each report s contents in the Spatial Analysis section As you become familiar with TELUM you may wish to review these detailed outputs for further insight into the forecasting process Further if you should later decide to rerun a model forecast TELUM will store the original detailed output files in a folder inside your main project folder in case you decide to review them again at a later date The next section contains a sample copy of a forecasting output file Several of the more important sections are highlighted for you 8 16 Always check the date and title to be sure this is the run you think it is EL 9 2006 6 18p CRPling EMPLOYMENT ALLOCATION MOD TELUM EMP V3 00 TELUM EMP Forecasting Output VERSION OF 20 NOV 2004 UTMAN ASSOCIATES 198 4 1 6 Dec 00
162. tion and employment for the forecast years i e for all future time points The following is a list of required region wide inputs for each model for each forecast time period TELUM Emp Employment e Regional ratios None e Regional forecasts Total employment for each employment type TELUM Res Households e Regional ratios Percent unemployment by employment type if available Employees per household by household type Matrix of households by income per employee Conversion Matrix by employment type Jobs per employee Net regional rate of employee commutation e Regional forecasts Total population Spatially Disaggregated Zonal Requirements A zone is a unit of spatial analysis that can be defined in terms of census tracts voting districts traffic analysis zones or an aggregation of these categories In previous land use model applications zones with an average population between 3 000 to 10 000 persons have worked best At this level of geographic detail the classes of data fall into substantive groups including population households by income and place of residence and employment by type and place of work These data are usually available from census publications The household and land use data are required for one time point by place of residence Prior work with TELUM Res calibrations has shown the benefits of adding a five year lagged total household variable by zone to the data set 4 3 Zonal em
163. tions Worksheet Regional projections are used in TELUM to control the sum of forecasted zonal employment and population Determine and enter regional forecasts for employment and population numbers in each forecast period Household size as it appears in the yellow box below is the average size for all household types This number is calculated automatically by TELUM from the data in your current year Households worksheet 0 0 0 0 0 0 0 0 Figure 8 TELUM DATAPREP Projections Worksheet The Employment to Households Conversion Matrix In the following section we describe the purpose of the Conversion Matrix after which you will see a sample of the worksheet template You can learn more about the conversion procedure by reviewing the numerical example at the end of this chapter in the Appendix section The TELUM system uses two models in a recursive sequence to forecast the location of activity The employment model TELUM Emp begins the sequence by producing a forecast of the spatial distribution of employment This is followed by the residential model TELUM Res which produces a forecast of the spatial distribution of households given the forecast location of employment Thus the output of TELUM Emp the forecast of the spatial distribution of employment is used as an input to TELUM Res The core of location forecasting in TELUM Res is done of households by household type In general the household types are specified by so
164. to the zonal data This procedure is also applicable for smaller reduced household data sets as well The ability to collect and use data during calibration that most accurately reflects the type and level of activity taking place in the region even with reduced variables is more important than having an artificially complete data set with low reliability 3 Preparation of Data Inputs For IDEU The success of every forecasting or analysis project is critically dependent upon the quality of its data inputs In general the more comprehensive and complex the proposed forecasting method the more extensive and expensive the required input data As most planning agencies have relatively limited resources available it is not possible for them to collect and process all the data It is necessary to develop a specific list of data requirements which can be specified along three dimensions 1 Geographic Detail 2 Temporal Detail 3 Activity Sectoral Detail In the following pages each of these three dimensions will be discussed In application default data values may be used to temporarily close data gaps Directions for preparing default values are provided in the subsequent data sections and in the TELUM help system 1 Geographic Detail The level of geographic detail employed depends on the requirements and limitations of models and data and competing interests amongst users of the analysis results No matter the final level of geographic detai
165. treets Vacant Developable Number of Forecast Time Periods Employment to Household Conversion Ratio Employment per Household by Income Unemployment Net Commutation Rate Regional Jobs per Employee 3 3 Screen P2 3 P2 4 P2 4 P2B 7 P2B 7 P2B 8 1 Name of Employment Category 4 P2B 9 1 P2B 12 P2B 12 1 Land Used for Basic Employment Land Used for Commercial Employment P3 13 1 P3 15 P3 16 1 P3 17 P3 18 P3 19 RC Rancho Carne AGR Agriculture LMFG Light Manufacturing PROF Professional Scientific and Management FIRE Finance Insurance and Real Estate LI Low income Households MI Middle income Households UMI Upper Middle income Households UI Upper income Households PUMS Public Use Microdata Sample EHIC Employees per Household by Income UR Unemployment Rates RNCR Regional Net Commuting Ratio After entering your regional data TELUM provides a summary report of the data you entered in IDEU as pictured below Please compare this report to your data to ensure accuracy Ei TELUM i e x Initial Data Entry Unit Report Name of the Region RG Number of Zones Estimated Total Population Current Year Lag Year Employment Data Available Current Year by Type ObyTotal O None Lag Year by Type O by Total O None Number of Employment seen Employment Categories FIRE
166. ts contents and display the frequency distribution of impedences by zone You will then automatically be escorted into the next section MCPU 5 Model Calibration and Preparation Unit In this section you will do the following 1 Run a Model Calibration 2 Review the Calibration Analysis of Results and MAP IT 3 Run a Land Consumption Regression Model 1 Run Employment and Household Model Calibrations Model calibration is a process of estimating the model parameters equation coefficients to obtain a match between observed and model estimate distributions of a region s employment and household location The Employment and Household Model Calibrations are calibrated by use of non linear regression TELUM will use your IDEU DOPU and TIPU inputs to run the model calibration The calibration is computationally intensive and may take up to 30 minutes to complete Upon completion of the employment and household calibration TELUM immediately provides a calibration Analysis of Results section 3 10 2 Review Calibration Analysis of Results The Analysis of Results contains statistical measures used to describe the model s fit to your data The better the fit of the model to the data the more reliable the forecasts it can produce An example of calibration results for Observed vs Estimated Employment Location for the Retail category of employment follows PRINT SCREEN EXIT OAK CONTIN
167. unction you can view the location of your regional data and model forecast outputs You can keep track of where the TELUM models over and under estimate zonal employment and household location decide 11 your region is more or less sprawled in future time periods and track how regional employment and households change their location over time and or with policy influence In this chapter you will learn about the various mapping tools provided by MAP IT how to use TELUM data outputs in mapping software other than MAP IT 2 How to Run MAP IT To use the MAP IT function you must have an installed version of ArcView 8x on your operating system If ArcView 8 3 is not available on your computer you can not use this function and must indicate NO in the first MAP IT screen that reads Will you be using the MAP IT function If ArcView 8 3 is available on your PC then you must have a GIS compatible shapefile of your region 1 Begin by confirming that you have a shapefile column with a numbering scheme that matches your regional zone inputs from JDEU and DOPU 2 Name your region shapefile s ZONE 3 Place your ZONE labeled shapefiles inside your TELUM folder C TELUM 4 Answer Yes when TELUM asks if you will be using the MAP IT function 3 What can MAP IT do for you GIS 15 a powerful computer mapping system and much more It is a tool for managing information of various types according to where it is located MAP I
168. used for the estimation of parameters for these models in their current aggregate form is gradient search In effect the partial derivatives of a goodness of fit criterion with respect to each specific parameter are calculated The values of these derivatives determine the direction of parameter search Putman 1983 The appropriate goodness of fit measure for the calibration of TELUM CALIBTEL is the likelihood function a measure derived from the notion of maximum likelihood as developed in econometrics This measure has the general form L XNlhN 1 where L is the computed likelihood measure Ni is the observed value and Ni is the estimated value of the dependent variable In TELUM RES the dependent variable would be households of a particular type located in a particular zone and in TELUM EMP the dependent variable would be employment of a particular type located in a particular zone It is important to note that in this equation form the magnitude of L is conditional on the magnitudes of the data being used In a region with millions of households L will be larger than it will be in a region with hundreds of thousands of households The Best Fit is when the difference between the models estimate of the dependent variable and the observed values in the calibration data set is as small as possible A perfect fit would be obtained if for each independent variable observation i e locator type and zone the Estimated Ni the Observed
169. useholds is estimated to have fifteen it is a 5096 error If a second zone with 1000 observed households is estimated to have 1050 it is a 596 error The value of MAPE for these two zones taken together is 27 595 a value that exaggerates the forecast error of the model One way to deal with this bias is to state the value of MAPE for just the smallest and largest observations zones in the data set along with the MAPE for all the zones taken together In that comparison we might see a MAPE of 500 in the smallest zones which account for 2 the region s low income households and a MAPE of 12 for the largest zones which account for 87 of the region s low income households The MARMO measure is another way to give error measures that are weighted by the size of the observation and is a good summary measure of likely forecast error levels The best and unachievable value for MARMO is 0 0 which would indicate a perfect fit of model to data Normal values of MARMO vary from 10 0 to 40 0 for each locator type Analysis of Location Elasticity Results Location elasticity measures the sensitivity of household and employment location to changes in the specific attractiveness variables of the TELUM models The location elasticity values are calculated both for individual employment or residential zones and for the regional average values for each variable for each locator For a 1 increase in an attractiveness variable in a zone the location el
170. ustry between each household type Default values have produced reasonable results when measured against matrices produced using PUMS data Calculate your default employment to household matrix by dividing the industry employment totals Employment worksheet cells B503 to 1503 by the total number of household income sectors in your region Enter the numbers in the corresponding cells of Input Table 1 in the DATAPREP Conversion Matrix worksheet x Note The unique feature of capturing regional employment and household shifts is lost when industry employment is evenly distributed between all household income sectors There are five different types of regional conversion ratio used in the TELUM Emp to TELUM Res connection These regional ratios are e Regional estimates of the number of jobs per employee e Regional estimates of the net into or out of the region commutation rate e Regional unemployment rates for each employment type e Regional employee to household conversion matrices e Estimates of the number of employees per household for each household type Regional control totals of employment by type are inputs to TELUM Emyp and no regional totals of households are input to TELUM Res The regional totals of households by income group are completely determined by the conversion of employees to heads of households This conversion procedure must be consistent with the actual numbers of households observed within a region if accurate foreca
171. vel and Residential Location Choices Ph D Dissertation Department of City and Regional Planning University of Pennsylvania Philadelphia PA Williams H C W L 1976 Travel Demand Models Duality Relations and User Benefit Analysis Journal of Regional Science 16 147 166 Wilson A G et al 1981 Optimization in Location and Transport Analysis Wiley Chichester Sussex texas2 doc wp60dos 09 feb 95
172. with your total inputted number for that locator type An example of an absolute employment or household constraint by type and zone is setting an employment value to a single zone If the manufacturing employment category is set to a total of 100 jobs in a specific zone TELUM will automatically forecast 100 manufacturing jobs to that zone The location of other employment types within this zone will likely be effected by this constraint To impose zonal constraints choose the year and type of constraint as indicated in the screen shot below Model Forecasting Unit Select Constraints for Constraint Type I H Year2005 Employment Constraint Household Constraint Year 2010 C Employment Constraint T Household Constraint Year 2015 C Employment Constraint T Household Constraint Year 2020 C Employment Constraint r Household Constraint Year 2025 C Employment Constraint T Household Constraint Year 2030 C Employment Constraint T Household Constraint 2 9 m mu Using the constraint data preparation table the constraints can be entered as shown in the table below If the policy run requires constraints for more than one year the constraint entry process can be facilitated by using the constraint copy button shown below on the right of the screen This copies the exact zones values and types of constraint and must be modified if the project constraints change over time periods 8 8 Model Forecasting Unit
173. xcel file and your data file in the same run of Excel for Paste Special Values or Text to be enabled e g open one file and then go to File gt Open to locate and open the second file Files in the same Data input cells are marked in a blue font Cells absent of blue indicate that DATAPREP automatically calculates the values Please note that the numbers shown in these spreadsheets are only for reference to help you understand how they will be used 4 9 Employment Worksheet For this section you need to have already determined the four to eight employment types how many workers employed by each type and how many employees work in each zone This data is necessary for the current year as well as the lag year T 1 which is defined in TELUM as approximately five years prior Once your data 15 correctly imported your zonal household population and employment numbers will appear in the worksheet An example of these worksheets as you will first see them before data entry into the DOPU workbook follows Employment 2000 gt gt gt gt gt 0 0 0 0 0 0 0 0 0 0 Figure 2 TELUM DATAPREP Employment Worksheet for Lag Year 4 10
174. you consult your mapping application and or ESRI s ArcMap documentation for assistance 5 3 6 TELUM Travel Impedance Preparation 1 Introduction to the Travel Impedance Preparation Unit TIPU Impedance is a term referring to the travel time travel cost or composite of both calculated by travel models to describe differences in zone to zone difficulty of interaction An agency transportation department typically builds impedance files and should be consulted for this section of TELUM The TELUM model requires an input data file containing zone to zone travel times and or costs for the region Several recent applications have made use of composite travel times costs developed from multimodal travel models and are often described in terms of impedos or some such other unit name The time point of these travel times should be the common current year for both models The issue of peak versus off peak travel times is important here The best solution when it is available is to use the afternoon peak travel times as these are the best inputs for location modeling In the issue of time versus cost for impedance values the most common choice is time To use cost would be perfectly acceptable except for the issues of inflation for the future year costs In addition it is sometimes possible to consider a generalized cost which can be calculated by combining travel time and travel cost In that case there is the problem of estimatin

TELUM (Transportation Economic and Land Use Model)

Contents

Download Pdf Manuals

Related Search

Related Contents