FAF2 Data Disaggregation Methodology and Results

presented to

Model Task Force

presented by

Vidya Mysore, Florida DOT

Krishnan Viswanathan, Cambridge Systematics, Inc.

November 28, 2007

Presentation Overview

Background

Florida FAF2 data

FAF2 Disaggregation method

Illustrative Example

Comparison with TRANSEARCH

Future Year Projections

Background

Growth in Florida

Increasing freight transportation

Capacity Constraints

FAF2 as potential data source

Audience of modelers and planners

Florida FAF2 Data

Florida FAF2 Data

  2002

Mode Within State From State To State

  Number Percent Number Percent Number Percent

Truck 487 85 50 68 85 42

Rail 60 11 17 23 37 18

Water (Domestic only) <0.1 <1 1 <1 37 18

Air, air & truck (Domestic only) <0.1 <1 0 <1 0 <1

Truck & rail <0.1 <1 0 <1 1 <1

Other intermodal 0 <1 1 1 5 3

Pipeline & unknown 27 5 5 7 36 18

Total 575 100 74 100 202 100

Florida FAF2 Data

Forecasts based on overall economic changes

Mode shares are assumed same in future

98 percent increase in commodity flows from 2002 to 2035

Truck increase is 108 percent

Domestic Water will decline by 51 percent

FAF2 Disaggregation Method

Data Sources

FAF2

County Business Pattern (CBP)

Public Use Microdata Samples (PUMS)

Census 2000

Three digit NAICS employment at county level

CBP most complete at MSA

Therefore use PUMS for employment allocation

Census 2000 for government and self-employed

FAF2 Disaggregation MethodSCTG Table NAICS Table

SCTG 2 to NAICS 3 Equivalency Table

CBP Data

NAICS 3 Employment Table

2002 Economic Census Data

2002 FAF2 Database

SCTG 2 to NAICS 3 Equivalency Table

County Level FAF2 Database by Value

SCTG 2 to NAICS 3 Equivalency Table

NAICS 3 Employment

Table

County Level FAF2 Database by Commodity

Census 2000 and QCEW Data

2002 FAF2 Database

DOMESTIC (kTon)

BORDER (kTon)

SEA (kTon)

Disaggregated to

Mode Split to

Truck (kTon)

Rail (kTon)

Water (kTon)

FAF2 Disaggregation Method

Mode Split to

Truck (kTon)

Rail (kTon)

Water (kTon)

SCTG 2 to NAICS 3 Equivalency Table

NAICS 3 Employment Table

County Level FAF2 Database by Commodity2002 FAF2 Database

County Level FAF2 Database by Commodity

2005 InfoUSA Data

TAZ Level FAF2 Database by FL Statewide Model Commodity Groupings

FAF2 Disaggregation Method

Develop relationships between commodity and employment and population data

Rationale is commodities end up in Zones that produce or consume them

Use relationships to develop factors for each commodity for freight flow disaggregation

Apply share of county tonnage to FAF2 regional tonnage to obtain disaggregated FAF2 O-D database

Illustrative Example

Florida Statewide Freight ModelFlorida Commodity Code Commodity Group Name STCC Codes SCTG Codes

1 Agricultural Products 1,7,8,9 1,2,3

2 Minerals 10,13,14,19 14,16,10-13

3 Coal 11 15

4 Food 20 4,5,6,7,8

5 Non-Durable Manufacturing 21,22,23,25,27 9,30,39,29

6 Lumber 24 25,26

7 Chemicals 28 20-23

8 Paper 26 27,28

9 Petroleum Products 29 17-19

10 Other Durable Manufacturing 30,31,33-39 24,32-40

11 Clay, Concrete, Glass & Stone 32 31

12 Waste* 40 41

13 Miscellaneous Freight 41-47,5020,5030 42

14 Warehousing 5010 42

Illustrative Example

Paper (SCTG 27, 28)

Pulp, Newsprint, Paper, and Paperboard

Paper or Paperboard Articles

Production Equation

0.362 (21.53) x Paper Manufacturing (NAICS 322)• R2 = 0.80

Attraction Equation

0.064 (4.56) x Paper Manufacturing (NAICS 322) + 0.043 (4.59) x Printing and Related (NAICS 323)

• R2 = 0.76

Illustrative Example

Estimate the annual tonnage of paper produced Pc(i) or attracted Ac(j) for each County

Aggregate the county productions Pc(i) and attractions Ac(j) to their associated Florida FAF2 regions to create PFAF2(i) and AFAF2(j)

Expand the FAF2 Regions matrix, FAF2(k,l), to Florida counties matrix, County (i,j)

Illustrative Example

If origin i and destination j are in Florida then County(i,j)=[FAF2(k,l)*Pc(i)/PFAF2(i)* Ac(j)/ AFAF2(j)]

If origin i is in Florida and destination l, is outside Florida then County(i,l)=[ FAF2(k,l)*Pc(i)/PFAF2(i)]

If origin k is outside Florida and destination j is in Florida then County(k,j)=[ FAF2(k,l)*Ac(j)/AFAF2(j)]

Illustrative Example    Paper 2002 (thousands of tons)

Origin Destination County FAF2 Zone

Disaggregation of Florida origins to Florida destinations

FAF2 Miami (20) FAF2 Jacksonville (19) #NA 16.27

Miami Dade County Baker County 0.16 #NA

Miami Dade County Clay County 0.11 #NA

Miami Dade County Duval County 5.51 #NA

Miami Dade County Nassau County 5.51 #NA

Miami Dade County St. Johns County 0.32 #NA

Palm Beach County Baker County 0.01 #NA

Palm Beach County Clay County 0.01 #NA

Palm Beach County Duval County 0.31 #NA

Palm Beach County Nassau County 0.31 #NA

Palm Beach County St. Johns County 0.02 #NA

Broward County Baker County 0.06 #NA

Broward County Clay County 0.04 #NA

Broward County Duval County 1.9 #NA

Broward County Nassau County 1.9 #NA

Broward County St. Johns County 0.11 #NA

Illustrative Example

    Paper 2002 (thousands of tons)

Origin Destination County FAF22 Zone

Disaggregation of Florida origins to other US destinations

FAF2 Miami (20) GA Rem (25) #NA 6.64

Miami Dade County GA Rem 0.27 #NA

Palm Beach County GA Rem 4.74 #NA

Broward County GA Rem 1.63 #NA

Disaggregation of other US origins to Florida destinations

GA Rem (25) FAF2 Miami (20) #NA 199.63

GA Rem Miami Dade County 113.05 #NA

GA Rem Palm Beach County 26.11 #NA

GA Rem Broward County 60.47 #NA

Comparison with TRANSEARCH

TRANSEARCH STCC 26 (Pulp, Paper, or Allied Products)

TRANSEARCH is unlinked trips and FAF2 is linked trips

  Origin (Production) Destination (Attraction)

 FAF2 (SCTG

27, 28)TRANSEARCH

(STCC 26)FAF2 (SCTG

27, 28)TRANSEARCH

(STCC 26)

Broward 25% 7% 30% 4%

Miami-Dade 71% 87% 57% 84%

Palm Beach 4% 5% 13% 12%

Total 100% 100% 100% 100%

Future Year Projections

Establish national control totals by commodity

Apply specific shipment growth by market and commodity

Apply specific purchasing and consumption growth by market and commodity

Future Year Projections

Summarize and compare results with national control totals

Adjust resulting freight flows so that volumes correspond with national level as follows:

For each market & commodity, adjust so shipments match purchases

For each commodity, adjust so that national control totals are satisfied

Future Year Projections

Use same methodology for Florida using CBP and Woodes & Poole (WP) data

WP data available at county level

Since we are focused on only tonnage, the data available to use can be used in a similar manner

Discussion

http://farm2.static.flickr.com/1191/875522713_c71061093b.jpg?v=0