This chapter describes the specification, the creation of calibration targets and the calibration summaries for trip generation models for model area resident trips, truck trips, internal-external trips, external-internal trips, and external-external trips.
Throughout the documentation, residents will be used to designate people who are residing in the four counties for which detailed dissaggregate synthesized population is used as the starting point. These four counties are: Riverside, San Bernardino, Orange and San Diego. The production model follows a standard cross-classification approach where average trip rates are applied to households stratified by socio-economic variables. The attraction model is a simple rate model, where a separate attraction rate is applied to zonal variables, including employment by employment type.
For the residents, trips are modeled for the following six categories.
The RIVCOM model uses trip production rates calculated from the SCAG Activity Based Model (ABM) 2016 base year validated model outputs. Please note that RIVCOM model trip production rates were calculated from model output data from only two counties (Riverside and San Bernardino) out of the six counties in the SCAG ABM region, to better represent RIVCOM model travel patterns.
Work-purpose trip production rates were calculated for six household worker (0-workers, 1, 2, 3, 4, 5+ workers) by five household auto (0-cars, 1, 2, 3, 4+) categories. This parameter file is here.
Non work trip purpose production rates were calculated for six household size (size 1, 2, 3, 4, 5, 6+) by five household auto categories (same as above). Non-work purposes include HBSH, HBO and NHB. The non work trip production parameter file is here.
The HBSC trip production rate is applied to households with 5 to 17 year old students. The HBU trip production rate is applied for households with 18 to 25 year old students. These parameters are available here.
Trip generation models apply the parameters described above to the actual number of households in different categories and to the number of students by age categories mentioned above. The number of households in different categories for each TAZ is stored in the file.
generation/HHDisaggregation. The number of students in the two age categories (5 to 17 year old students and 18 to 25 year old students) in each TAZ is stored in the file.
generation/Students_Summary Both these files are created during the population synthesizer step by summarizing the dissaggregate households and population output.
Trip production rates do not vary by household income group. The reason behind this was that, the auto ownership segmentation is a good proxy for household income, and also cuts down on model runtime to have fewer segmentation. In addition, for assessing tolling, auto-segmentation is as useful as income segments. Even though the trip production is not segmented by household income, the resulting trip productions are aggregated based on the household income level. The reason this is possible is because the household disaggregation file, is segmented by household income level (low income and high income).The cut-off between low income household and high income household is $40k, in 2018 dollars. The trip productions are aggregated into five market segments by household income and auto sufficiency categories.
HBW purpose travel market segments are defined as (based on household workers):
|Segment Name||Segment Definition|
|Zero autos||households with zero autos|
|Low income insufficient||low income (<=$40k) households, where autos = 1 & workers > 1|
|Low income sufficient||low income (<=$40k) households, where autos = 1 & workers = 1, or autos greater than 1|
|High income insufficient||high income (>$40k) households, where autos = 1 & workers > 1|
|High income sufficient||high income (>$40k) households, where autos = 1 & workers = 1, or autos greater than 1|
HBO and HBSH purpose travel market segments are defined as (based on household size):
|Segment Name||Segment Definition|
|Zero autos||households with zero autos|
|Low income insufficient||low income (<=$40k) households, where autos = 1 & hh size > 1|
|Low income sufficient||low income (<=$40k) households, where autos = 1 & hh size = 1, or autos greater than 1|
|High income insufficient||high income (>$40k) households, where autos = 1 & hh size > 1|
|High income sufficient||high income (>$40k) households, where autos = 1 & hh size = 1, or autos greater than 1|
HBSC, HBU and NHB purposes were treated as a single segment and not classified into five market segments.
The trip attraction model uses the following attraction variables including employment and enrollment:
The RIVCOM attraction model parameters were asserted from several other models including the SCAG trip-based regional model. The current version of trip attraction model parameters by purpose are here.
The trip attractions were balanced back to trip productions at the end of the trip generation procedures, except for the NHB purpose. For the NHB purpose trips, trip productions was set equal to trip attractions.
The table below summarizes the trip generation model. The overall household-level target trips rate for each purpose, computed from SCAG 2020 RTP base year (2016) model output, is compared to the model-predicted trip rates by purpose. The results shows that the model predictions closely align with the targets. The minor differences are calibration updates to get the assigned network volumes to match better the observed counts.
|Trip Purpose||Target Trip Rate||Model Trip Rate|
Truck trips that are internal to the four-county model region are covered by this model. Truck trips are segmented into three categories in the model framework: commercial vehicles (CV), single unit trucks (SUT); and multi-unit trucks (MUT). For calibrating the truck trips, SCAG model output was used, where the truck trips segments were slightly different. The segments in the SCAG model were Light-Heavy Duty Truck (LHDT), Medium-Heavy Duty Truck (MHDT) and Heavy-Heavy Duty Truck (HHDT). LHDT was target for CV, MHDT was target for SUT and HHDT was target for MUT.
The truck trip generation and attraction parameters are a function of employment in the TAZ by different categories. The model parameters were adopted from the Hickory, North Carolina Model and then calibrated to match the number of truck trips in the SCAG 2020 RTP base year (2016) model output for: Riverside, San Bernardino and Orange counties. This parameter file is here.
truck generation. It has both the trip generation parameters and trip attraction parameters for the three truck segments. The calibration targets and the results of calibration are shown below.
|Truck Segment||Target Trips||Model Trips|
Trip generation for external trips are done separately for passenger cars and trucks. There are 30 external stations in the model and are shown in the map below. The external station ID can be seen by hovering on the map markers in the map.
Passenger cars will be discussed first followed by trucks. The passenger car external models are divided into three segments: External-External, Internal-External and External-Internal. Trip generation for each of these three segments are based on the volume of traffic at each of the external stations.
In the base year, observed count data at the external stations can be used and for future year, estimates of external station data is used. In RIVCOM, count data at all external stations was not available; therefore the model uses the SCAG 2020 RTP base year (2016) model output is used to specify the traffic volume at all the external stations for both the base year and future years. These numbers are specified in the file
external_awdt. The observed or estimated traffic volume at each of the external stations are specified in the field AWDT2018, AWDT2019, and so on. This volume is the sum of the three constituent segments: EE, IE and EI.
External-External trips for passenger cars are obtained from the total volume specified in the external_awdt.csv multiplied by the EERatio field in the same file. Therefore, EERatio is the fraction of station volume that are EE trips. The model outputs from SCAG Model outputs are used to obtain the fraction of external station volume that are EE trips.
The table above shows the external stations that had EE passenger car trips. Each EE trip gets counted twice, once at the entry station and once at the exit station; therefore, based on the above summary there are total of 5,619 EE passenger car trips in the base year. Due to the formulation of the EE trip generation model, the model predictions for the number of EE trips match exactly with the inputs shown in the table above. It may be noted that only a subset of external stations have EE trips. The reason for this is that, the EE trip estimate (as a function of the total external trips) for a station was obtained from the SCAG base year model outputs and in those outputs only few of these interchanges had EE trips.
Within the RIVCOM model framework IE and EI trips are treated together in a combined way as IE/EI trips in all the sub-models. At each external station, the EE passenger car volume is subtracted from the total passenger car volume at the station to obtain the total IE/EI volume. This total volume is further segmented by occupancy: Drive alone, Share Ride 2 and Share Ride 3. The volume by these occupancy categories are computed using external station level factor stored in the input file
eiie_occup. The factors vary by external station and was derived based on the predicted model volume share of DA, SR2 and SR3 in the SCAG 2020 RTP base year (2016) model output. The table below shows the total IE/EI trips generated at each external station, as well as the resulting vehicle trips for the three vehicle occupancy categories.
Truck EE and IE/EI trip generation models are similar to the passenger car models except that there is no occupancy level segmentation for trucks. External truck trip generations are based on the volume of truck traffic at each of the external stations. SCAG 2016 RTP ABM model outputs are used to specify the truck traffic volume at all external stations. These numbers are specified in the file
external_truck_awdt. The observed or estimated truck traffic volume at each of the external stations are specified in the field AWDT2018, AWDT2019, and so on. This volume is made up of the three constituent segments: EE, IE and EI. Truck EE trips are obtained from the total truck volume specified in the external_truck_awdt.csv multiplied by the EERatio field in the same file. Total IE/EI truck volumes at each external stations are obtained by subtracting the truck EE volume from the total truck volume at the external stations. The table below shows the truck model volumes at external stations for EE truck trips and IE/EI truck trips.
Riverside County Model, 2020