Appendix B: Technical Memorandum #1
NEXT STEPS
Below is a list of the next information to be provided to TDOT and GDOT for review and approval upon completion:
• 2007 and 2035 socioeconomic data (currently expected in November 2008)
• List and corresponding map of capacity-adding transportation projects added to the 2000 model to update to 2007
• Populated validation worksheet once validation complete
As noted earlier, once the modeling efforts and draft LRTP document have been completed, full model documentation, including a user’s guide and all model files, will be provided to TDOT and GDOT.
As discussed during the modeling workshop held on August 20, 2008, please provide any comments back to the RPA within 10 days, if not sooner. We thank you in advance for understanding our time constraints and look forward to coordinating with you throughout the development of our model and LRTP. If you have any questions about the model, please do not hesitate to contact Melissa Taylor at the RPA via email at [email protected] or phone at (423) 757-0077.
Attachments:
Table of TDOT Validation Requirements Empty Validation Worksheet
Attachment #1
Model Validation Documentation Requirements Specified in TDOT Division Procedure: MPO Model Approval
and
“Minimum Travel Demand Model Calibration and Validation Guidelines for State of Tennessee,”
University of Tennessee, Center for Transportation Research
Socioeconomic Data
TAZ Maps with corresponding existing and future
SE data, including external stations PDF Maps TDOT*
Person trips/household by purpose
Validation
Worksheet U Tenn**
Travel Survey Data
Source, year, and type of travel survey information
Text (overview provided in
Submittal #1) TDOT*
Household travel diary survey (2002) used to determine trip generation rates, average trip lengths, auto occupancy factors, and other characteristics used in 2000 model development and validation.
Percent IE/EE splits were calculated from the Chattanooga Urban Area Origin-Destination Study conducted in 2002. However, it was determined that these percentages were not reliable as the interstate
surveys were conducted on ramps at interchanges and rest areas.
Vehicles were not pulled off to the side of the road on the mainline interstate segments where most EE trips occur. The vehicles that were surveyed were already stopping at the interchange or rest area.
Therefore, the results provide a significant bias towards IE trips and EE trips are underestimated. The percent IE/EE splits from the origin
and destination survey were subsequently compared to the old 2000 MINUTP model and it was determined that the old 2000 MINUTP
model percent splits appeared more reasonable.
Will need to
If no survey, provide source of trip data
Text (overview provided in
Submittal #1) TDOT* QRFM used for truck trip production rates; NERPM for attraction rates Network Data
Provide files (networks, turn
penalties/prohibitors, speed/cap, FT codes) Files TDOT*
Table of significant projects added to network by
model year Excel Table TDOT*
Explanation of how MPO dealing with
cross-border traffic Text TDOT* K-factor and screenline at state line
Determine during validation Trip Generation
Description of trip generation method (P/A, O/D, Special Generators; Cross-Class, Regression, or Discrete Choice)
Text (overview provided in
Submittal #1) TDOT*
Cross-Classification method based on auto availability and children per household; Trip production rates for HBW, HBSCHOOL, HBSHOP, HBSR, HBO, and NHB calculated using 2002 HH Travel
Diary Survey; Truck production rates from QRFM; Attraction rates
from NERPM; Special generators included same
Description of trip purposes
Text (overview provided in
Submittal #1) TDOT*
HBW, HBSCHOOL, HBSHOP, HBSR, HBO, NHB, LTTK, MDTK,
HDTK same
Description of how EE and EI trips addressed Text TDOT*
The MINUTP splits were used as a starting point and were later modified during the validation process.
The percent IE trips were applied to the 2000 AADT to calculate the total number of IE trips at each external zone for the TAZ geography file. The EETRIPS input matrix file is generally the residual left after estimating IE trips in the TAZ geography file. The percentage of EE trips was applied to the AADT per external zone and then distributed
from each origin zone to each destination zone using distribution patterns from the old MINUTP 2000 model EETRIPS file. For external
zones added to the TransCAD model, and not in the MINUTP model, it was assumed that 100 percent of these trips were EE trips since
these were low volume roadways.
Whether trip rates developed for vehicle and/or
person trips Text TDOT* Person trips same
Special generators (how and where) Text TDOT* Yes; Airport, UTC, Chatt State College, Chester Frost National Park
Determine during validation Trip Distribution
Files (friction factors, terminal time, and
K-factors) Files TDOT*
Trip length by trip purpose for base and future year
Validation
Worksheet TDOT*
Comparison of HBW mean trip length compared to Census JTW mean trip length
Validation
Worksheet TDOT*
Trip Length Frequency Distribution (TLFD) by Trip Purpose Compared to Census JTW and Household Travel Survey (Charts)
Validation Worksheet (add
charts) U Tenn**
Area-to-Area Flow of Trips Map U Tenn**
Were friction factors in gravity model kept constant between base and future years?
Text (overview provided in
Submittal #1) TDOT*
Yes. The friction factor file used in the new CHCNGA 2000 TransCAD model is similar to the
file used in the CHCNGA 2000 MINUTP model. The key difference is the number of trip
purposes. Since separate friction factors were not available for the expanded home-based
other purposes, the same home-based other friction factors are used for home-based
school, home-based shop, and home-based social recreation trips.
Also, as discussed,
gamma functions from the Quick Response Freight Manual were
used for truck purposes. same
Were terminal times used?
Text (overview provided in
Submittal #1) TDOT* Yes same
Were K-factors used and if so, what was the basis?
Text (overview provided in
Submittal #1) TDOT*
Yes; To improve the Georgia/Tennessee State Line screenline in the CHCNGA 2000 TransCAD model, a 0.25 K-Factor was added from
zones in Georgia to
zones in Tennessee. To improve the Tennessee River Crossing screenline, a 0.25 K-Factor was added to zones from the north side of
the bridges to the rest of the model region.
Determine during validation
Mode Split TDOT*
Description of mode split and choice analysis
method (i.e. nested logit?) Text (n/a) TDOT* n/a n/a
Were vehicle occupancy factors by trip purpose used to convert person trips to vehicle trips?
Were they kept constant between years?
Source?
Text (overview provided in
Submittal #1) TDOT* Yes; Yes; Source is HH Travel Survey same
What network model coding logic was used to
code access to transit in the network? Text (n/a) TDOT* n/a n/a
Trip Assignment
Files (loaded network, summary table including estimated link vols & costs, text file with user
inputs and model outputs) Files TDOT*
Comparison of screenline vols with counts (2 N/S, 2 EW)
Validation
Worksheet TDOT*
Comparison of cutline, screenline, and cordon line vols with counts
Validation
Worksheet TDOT*
Was time-of-day assignment performed? If so, details.
Text (overview provided in
Submittal #1) TDOT* No same
If no time-of-day assignment, how were 24-hour vols converted to peak hour vols? What conversion factors used? Relationship between peak hour & 24-hour vol?
Text (overview provided in
Submittal #1) TDOT*
Peak hour factors applied to achieve temporal distribution based on
2002 Household Travel Diary Survey same
Fraction of total VMT assigned to IZ and centroid connector trips? Is it reasonable?
Validation
Worksheet (add) TDOT*
Traffic counts on 10% of regionwide highway segments by functional calss (freeways and
principal arterials at a minimum & screenlines) Table U Tenn**
Percent Differences in Daily Volume by FT and screenline
Validation
Worksheet U Tenn**
Percent Difference in Peak Hour Volume by FT
Validation
Worksheet (add) U Tenn**
Percent RMSE by Link Volume Group
Validation
Worksheet U Tenn**
Scatter Plot of Assigned vs. Observed Link Traffic Volume
VMT, VHT, and average speed for entire system and by FT
Final Files (complete datasets) Files TDOT*
Final Documentation (complete model validation report, including user's guide)
PDF of Final
Report TDOT*
Attachment #2
Model Validation Worksheets
(Template – Not Populated with 2007 Statistics)
ga 2007 MODEL VALIDATION SUMMARY TEMPLATE 9/18/2008= Statistics in green-colored cells will be updated when 2007 validated m eration of Total Trips by Purpose Unbalanced200420042000New FDOT Productions% by ProductionsAttractionsProductions% by ProductionsNashvilleMemphisMontgomeryGuidelines*2 Work0#DIV/0!0212,11312.72%16.79%25.15%16.00%18-27%12-24% School0#DIV/0!085,0685.10%11.02%5-8% Shop0#DIV/0!0159,4209.56%7.17%10-20% Socrec.0#DIV/0!092,9195.57%7.67%9-12% Other0#DIV/0!0289,01317.33%28.10%14-28% Based0#DIV/0!0434,51326.06%26.42%18.00%22-31%20-33% *4 Vehicles0#DIV/0!0163,3309.80%8.79%11.00% ht-Truck#DIV/0!120,28673.65% ium-Truck#DIV/0!31,58019.34% vy-Truck#DIV/0!11,4647.02% rnal0#DIV/0!0230,90613.85%8.82%13.00% ternal030,0253.00% SOV#DIV/0!0.00% HOV#DIV/0!0.00% ht-Truck#DIV/0!0.00% vy-Truck#DIV/0!0.00% 0#DIV/0!1,667,282100.00%100.00%100.00%100.00% rip Rates & Socioeconomic Summaries 2007200020022004200020002001TDOTNew FDOTFHWA sureNashvilleMemphisMontgomery*5 KnoxvilleSavannahTarget*1Guidelines*2Target*3 ousehold#DIV/0!2.502.562.652.682.452.53n/a2.0-2.7n/a s per Household0.009.448.598.208.648.407.668.5-10.58.0-10.08.0-14.0 s per Person0.003.783.363.093.233.433.03n/a3.3-4.03.5-4.0 s per Employee0.005.184.866.414.915.705.54n/an/an/a er Person#DIV/0!0.730.690.510.660.600.55n/an/an/a tion0395,0611,206,665761,3461,181,701299,180232,011 holds0158,055471,298310,412416,830111,79391,834 yment0287,918833,862457,796604,578196,799127,000 s / Household YearPerson Trip/HH (TransCAD)20070.00 (TransCAD)20009.44 (MINUTP)20007.44 get8.50 to 10.50 T Guidelines8.00 to 10.00 argetn/a
20072000 Chattanooga (TransCAD)
Chattanooga 2007 (TransCAD)Chattanooga (TransCAD)TDOT Target *1 47-54% 20.89%
39.00%39.18%
tribution rip Length (in Minutes) 2007200020002000200020022000200420002001TDOTNew FDOT urpose Chattanooga (TransCAD)Chattanooga (TransCAD)
CTPP (Hamilton)CTPP (Catoosa)CTPP (Walker)NashvilleKnoxville *6 MemphisMontgomerySavannah *7Target*1 Guidelines*2 d Work0.0016.3621.2023.7026.5019.2612.0519.0019.9019.8012-35 d School0.0014.48n/an/an/an/a6.82n/an/an/an/a d Shop0.0013.78n/an/an/an/an/a12.30n/a15.109-19 d Socrec.0.0011.38n/an/an/an/an/a13.90n/an/a11-19 d Other0.0014.41n/an/an/a11.487.9813.4017.4015.408-20 Based0.0015.44n/an/an/a12.367.1413.2016.7013.406-19 l Vehicles0.0019.64n/an/an/a17.2015.80n/a ht-Truck0.0018.79 ium-Truck0.0017.95 vy-Truck0.0022.17 ternal0.0041.04n/an/an/a27.8019.9526-58 0.0018.68n/an/an/a14.378.5014.3619.8016.58n/a Travel 20072000 Two-DigitTDOTNew FDOTFHWA urpose Total Trips
Intrazonal Trips% IntrazonalTotal Trips
Intrazonal Trips% IntrazonalTarget*1Guidelines*2Target*3 d Work00#DIV/0!212,11314,9317.04%1-4% d School00#DIV/0!85,0689,45911.12%10-12% d Shop00#DIV/0!159,42015,8289.93%3-9% d Socrec.00#DIV/0!92,91919,58521.08%4-10% d Other00#DIV/0!289,01332,26611.16%3-7% Based00#DIV/0!434,51334,8538.02%5-9% l Vehicles00#DIV/0!163,3304,9953.06% ht-Truck00#DIV/0!120,2863,5502.95% ium-Truck00#DIV/0!31,5801,3094.15% vy-Truck00#DIV/0!11,4641361.19% 00#DIV/0!1,436,376131,9179.18%3-5%
Chattanooga 2007 (TransCAD)Chattanooga 2000 (TransCAD) Two-Digit
ignment (1st of 3 pages) Screenlines TDOTNew Chattanooga 20072000AccuracyFDOT Total VolumeTotal CountVol/Count RatioVol/Count RatioLevel *1 Guidelines *2 * 000.000.971 Central City Partial Cordon+/- 10%+/- 5% 000.000.942 Tennessee River Crossing+/- 10%+/- 5% 000.000.973 North-South Downtown+/- 10%+/- 5% 000.000.914 East-West Downtown+/- 10%+/- 5% 000.000.905 Missionary Ridge+/- 10%+/- 5% 000.000.996 South Chickamauga Creek+/- 10%+/- 5% 000.001.097 North-South North Hamilton Co.+/- 10%+/- 5% 000.001.068 Ringgold+/- 10%+/- 5% 000.001.029 TN/GA State Line+/- 10%+/- 5% 000.001.0010 External Cordon Line+/- 10%+/- 0% TDOTNew FDOTFHWA Volumes Over Counts% DifferenceVolumes Over Counts% Difference Accuracy Level *1 Guidelines *2 Target *3 1Interstate1.0000.98-1.6+/- 7%+/- 5% (+/- 6%)+/- 7% 2Expressway1.0000.94-5.6+/- 7%+/- 5% (+/- 6%)+/- 7% 3Principal Arterial Divided1.000.00.98-2.1+/- 10%+/- 7% (+/- 10%)+/- 10% 4Principal Arterial Undivided1.000.00.94-6.1+/- 10%+/- 7% (+/- 10%)+/- 10% 5Minor Arterial1.0000.88-12+/- 15%n/a+/- 15% 6Collector1.0000.98-2.2+/- 25%+/- 10% (+/- 15%)+/- 25% 7Rampn/an/an/a 8One-Way1.0000.71-29.1n/a+/- 15% (+/- 20%)n/a 1CBD1.0000.88-11.8 2CBD Fringe1.0000.92-8.2 3Residential 1.0000.96-4.0 4OBD1.0000.92-8.4 5Rural 1.0001.109.7 6Urban Undeveloped1.0001.021.9 7Mountainous1.0000.95-5.0 11.000.00.93-6.8 21.000.00.98-1.6 31.000.00.95-5.3 41.000.00.92-7.8 5 1.0000.95-4.7 nce for Daily Volumes (calculated for Links with Counts) Chattanooga 2007TDOTNew FDOTFHWA TransCADTarget *1 *9Guidelines *2 Target *3 0.00%60%60% 0.00%47%47% 0.00%36%36% 0.00%29%29% 0.00%25%20-30% *1025% 0.00%22%15-25% *1022% 0.00%21%5-20% *1121% etermination R2 (Correlation Coefficient of Actual Counts and Model Volumes) *1 TDOT Target: R2 more than 0.88 *2 New FDOT Target: N/A *3 FHWA Target: R2 more than 0.88
25-50%
2000Chattanooga 2007ounts
Assignment (2nd of 3 pages) n Square Error TDOTFHWA ngeTransCADTransCADTarget*1 Target*3 0.00%73.10%115.7645-100 990.00%33.60%43.1435-45 ,9990.00%18.10%28.2727-35 ,99925-30 ,99921.70%15-27 ,99916.90% ,9994.40% ,9997.20%10-20 ,999 ,999 ,999 0.00%23.40%30.0032-39 iles Traveled YearVMTHPMS*12 ga (TransCAD)200700 (TransCAD)200431,796,87533,316,412 (TransCAD)200426,881,55026,980,700 (TransCAD)200024,159,507n/a ery (Tranplan)20008,296,866n/a (TP+)20015,697,4235,743,828 rget *1 arget *3 (same as FDOT)
31.7 28.0
VMT / Person 0 26.4 24.4
0VMT / HH
Chattanooga 2007Chattanooga 2000 n/a
0.00% 0.00%15-25 10-19
New FDOT Guidelines & Standards*2 64.567.5 74.277.8 62.024.6 10-16 for small urban area / 17-24 for large urban areaT Guildelines *2 30-40 for small urban area / 40-60 for large urban area
+/- 5% difference b/w model and estimate
25.38 30.25 19.200.00%
Assignment (Last of 3 pages) unctional Classification ypeVMTHPMS% Difference 1Interstate00.0% 2Expressway00.0% 3Principal Arterial Divided00.0% 4Principal Arterial Undivided00.0% 5Minor Arterial00.0% 6Collector00.0% 7Ramp0 8One-Way00.0% omments/Conclusions: R, Minimum Travel Demand Model Calibration and Validation Guidelines For State of Tennessee Model Validation and Reasonableness Checking Manual ncludes combined purposes for NHB Work and NHB Nonwork, where appropriate. ge Trip Length for NHB for Knoxvile was obtained by averaging 7.88(NHB Work) and 6.40 (NHB Others). ge Trip Length for Internal-External for Savannah was obtained by averaging 18.4 (IE Truck) and 21.5 (IE Passenger Cars). guidelines (*1 ) suggest FHWA target should be selected where both of FHWA and Michigan DOT criteria are available (TDOT guidelines, page15). and MDOT targets were switched in Table 3 of TDOT guidelines (page 12), thus the original source of FHWA Checking Manual (*3 ) was referred for con S stands for Highway Performance Monitoring System. ets for medium urban are size (200k - 1M) was chosen for Chattanooga Region.
n/a
New FDOT Guildelines *2TDOT Target *1 27-33% 18-22% 8-12% n/a
(all the same) *13 33-38% 33-38% 27-33%
FHWA Target *3 Chattanooga 2007