ODOT Freight Modeling Presented to the Ohio Conference on Freight Toledo, OH September 18, 2007 By Gregory Giaimo, PE Ohio Department of Transportation

Context of Freight Modeling at State DOT’s Part of a larger statewide passenger and freight model Developed primarily to study intercity highway corridors Prior to 1990’s, very few such models due to lack of computers and methods Since then about half the states have developed or are developing models

Statewide Model Status Source: NCHRP Synthesis 358, Statewide Travel Forecasting Models

Freight Model Types Freight modeling is handled in one of four ways by state DOT models 1.None 2.Traditional “4 Step” Techniques 3.Commodity Based 4.Integrated Land Use/ Economic/ Commodity Based

Source: TRB Electronic Circular E-C075: Statewide Travel Demand Modeling – A Peer Exchange

Statewide Models with No Freight Component Some small states have models exactly analogous to urban passenger travel demand models with no freight model

Commodity Based Models This approach models commodity flows These flows are then converted to trucks, trains etc., generally using static mode and payload factors by commodity This is the most common approach

Commodity Based Models (cont.) Obtain their commodity flow data from the Commodity Flow Survey or the Transearch database The Vehicle Inventory and Use survey is often used to develop payload factors

Integrated Land Use/ Economic/ Commodity Models This approach also models commodity flows The main difference from the previous type is the explicit econometric and land use models which feed commodity flows to the transport models Thus base year commodity flows are used to estimate these economic models instead of as direct model inputs

Ohio Model The five components directly related to the creation of freight flows will be hi-lighted: 1.Interregional Economic Model 2.Land Use Model 3.Activity Allocation Model 4.Aggregate Commercial Vehicle Model 5.Disaggregate Commercial Vehicle Model

Disaggregate Household Synthesis and Employment Spatial Disaggregation Models Ohio Integrated Land Use/Economic/Transport Model Aggregate Commercial Vehicle Model Disaggregate Commercial Vehicle Model Interregional Economic Model Aggregate Demographic Model Land Development Model Activity Allocation Model Visitor Model Long Distance Travel Model Short Distance Travel Model Assignment Model

Model Modules Interregional economic model of production & consumption by economic sector reflecting national forecasts Demographic model tied to economic activity reflecting migration and changes in population & household composition Activity allocation model to distribute model area economic and demographic forecasts to analysis zones with the related flows of goods & labor among zones from which travel demands are derived Land development model simulating developer behavior in response to demands & costs consistent with other development constraints Personal & household travel model reflecting person & household characteristics, zonal characteristics, inter-zonal economic flows & transport system supply characteristics, 2 components: short distance which looks like an activity/tour based urban area model and long distance, also tour based with purposes: business, recreation, other

Model Modules Aggregate model of goods and services transport arising from economic and demographic activity by zone very similar to the typical DOT commodity based transport model Disaggregate model of business-related person travel related to management functions, sales & support activities, provision of services and some short distance goods delivery. Model of visitor travel within and into the model area made by non- residents Transport system supply model incorporating air, intercity bus/rail, MPO transit & roadway networks with their corresponding level-of- service characteristics

Highway Network

Rail Network & Intermodals

Interregional Economic Model Establishes forecast flows of goods, services and labor (in $) between 14 regions of North America Uses exogenous national economic conditions and production composite utilities from the previous time step of the lower level models An inter-regional social accounting matrix based primarily upon IMPLAN data

Nation Economic Forecast Variables National Economic Growth Exogenous Economic Indicators

Industry Categories 1. Agriculture, Forestry, Fisheries 2. Primary Metals (Steel) 3. Light Industry 4. Heavy Industry 5. Transportation Equipment (Auto) 6. Wholesale 7. Retail 8. Hotel 9. Construction 10. Health 11. Transportation Handling 12. Utilities 13. Other Services 14. Grade School Education 15. Post-Secondary Education 16. Government

Land Use Model Creates developed floor space by category by zone Land develops (if possible) in response to increases in households and employment in the previous analysis year (see activity allocation model) but also due to: –Zoning –Flood plains –Slopes –School district quality

Water Service

Flood Plains

Severe Slope

Activity Allocation Model Subdivides activity to 5000 traffic analysis zones –Population –Employment –Labor Flow –Commodity Flow Uses transport utility equations relying on accessibilities and changes in developed land to redistribute activities from year to year

Activity Allocation Model Inputs include: –Regional flows from Interregional Economic Model –Floor space by category by TAZ from Land Use Model –Households from Aggregate Demographic Model –Transport costs from previous iteration of Transport Models

Aggregate Commercial Vehicle Model With the commodity flows established, the aggregate commercial vehicle model itself is very similar to the commodity representation in other statewide models Input is dollars of flow of goods and labor at the Traffic Analysis Zone (TAZ) level Output is flows of trucks between TAZ’s

Aggregate Commercial Vehicle Model Flow Chart Determine Mode Convert Goods Flows to Tons Total Dollars flows Determine Truck Type Determine Number of Truck Loads Determine Trucks by Hour of Day Based on traffic counts, conversion from annual to weekday assumes 300 equivalent week days per year. This value is obtained as follows: (52 * 5) weekdays plus (52 * 2 * 0.44) weekday equivalents for weekends minus 6 holidays. Trucks by type by hour by OD TAZ Fixed shares by commodity class, by distance for bulk commodities, mode choice model for inter- modal eligible commodities, based on CFS By commodity class and distance, from CFS By commodity class and distance, from VIUS By commodity class and truck type, from VIUS

Truck Mode Shares by Distance and Commodity

Dollar to Ton Conversion by Distance and Commodity

Truck Type by Distance and Commodity

Payload Factors by Distance and Commodity

Disaggregate Commercial Vehicle Model DCOM is designed to account for short distance commercial travel not related to the long distance shipping of freight (accounted for in ACOM) Long distance business travel is accounted for in the Long Distance Travel model of the personal transport model since these trips were obtained in the special long distance travel survey

Disaggregate Commercial Vehicle Model Employs a tour based microsimulation of employees Based on establishment surveys Analogous to HH based tour based model but based at the place of work Does not include route delivery vehicles

Disaggregate Commercial Vehicle Model

Employment categorized as: Industrial Wholesale Retail Transportation Handling Service

Disaggregate Commercial Vehicle Model Trip purposes: Service Meeting Goods (delivery) Other (includes such things as stopping for lunch or fuel)

Outputs

Forecast Volumes

Traffic Flow Maps District 2 Volumes Including Interstate District 2 Volumes Excluding Interstate

Toll Sensitivity Analysis Showing Volumes Changes

Congestion Management By ODOT Districts For Ohio Counties Congestion Comparisons

User Costs by Alternative

Predicted Changes by Economic Sector

Employment Forecasts ODJFS Emp Forecast (Ohio)ISAM Emp Forecast (Ohio+Halo) OSMP Industries %Change mid%Change Agriculture92,50085,100-8%184,588174,443-5% Heavy Industry493,100473,000-4%749,212705,277-6% Construction235,700269,10014%524,411596,69814% Light Industry180,900177,900-2%310,594303,279-2% Primary Metal Products61,50053,500-13%106,42491,873-14% Transportation Equipment159,900147,900-8%179,874166,250-8% Wholesale239,100261,4009%394,184425,5458% Retail1,020,8001,138,00011%1,504,4881,661,03610% Transportation Handling160,700184,70015%272,796310,02714% Utilities Services23,50020,000-15%155,928136,355-13% Other Services1,321,8001,520,10015%2,209,6742,519,59214% Education76,40084,10010%730,246829,95814% Health Care622,800757,30022%801,801964,78820% Hotel and Accommodation34,90037,6008%55,38359,5538% Government and Other756,500828,60010%334,501368,89910% Education + Government832,900912,70010%1,064,7471,198,85713% TOTAL5,813,8006,376,10010%8,514,1049,313,5739%

Interaction of Ohio’s Economy with the Rest of the US

Commodity Production

Changes in Commodity Production

Future Challenges Modeling community would like to move to more detailed freight modeling based upon econometrics and supply chains The main challenge that plagues this as well as current modeling efforts is the lack of geographically specific freight data

Top 10 Challenges Impeding Development of Freight Models by State DOT’s 1.Lack of DATA 2.Lack of DATA 3.Lack of DATA 4.Lack of DATA 5.Lack of DATA 6.Lack of DATA 7.Lack of DATA 8.Lack of DATA 9.Lack of Money 10.Lack of better theoretical formulations (but see 1-8 for why this is the case)

Questions?