Freight Trip Generation for the Roanoke Valley

Freight Trip Generation for the Roanoke Valley

Regional Freight Study Technical Report t
Roanoke Valley Area Metropolitan Planning
Organization (RVAMPO) and Roanoke Valley-
Alleghany Regional Commission (RVARC).
Freight Trip Generation for
the Roanoke Valley t
Technical Report
Fiscal Year t 2012 t Final 11-15-2012
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
1 | Page
ACKNOWLEDGEMENTS
This report was prepared by the RVAMPO in cooperation with the U.S. Department of Transportation
(USDOT), the Federal Highway Administration (FHWA), the Federal Transit Administration (FTA), the
Virginia Department of Transportation (VDOT) and the Virginia Department of Rail and Public
Transportation (VDRPT). The contents do not necessarily reflect the official views or policies of the
FHWA, FTA, VDOT, Department of Rail and Public Transportation (DRPT), RVAMPO or Roanoke Valley-
Alleghany Regional Commission (RVARC). This report does not constitute a standard, specification, or
regulation. FHWA, FTA or VDOT acceptance of this report as evidence of fulfillment of the objectives of
this planning study does not constitute endorsement /approval of the need for any recommended
improvements nor does it constitute approval of their location and design or a commitment to fund any
such improvements. Additional project level environmental impact assessments and /or studies of
alternatives may be necessary.
The Roanoke Valley Area Metropolitan Planning Organization (RVAMPO) fully complies with
Title VI of the Civil Rights Act of 1964 and related statutes and regulations in all programs and
activities. For more information, or to obtain a Discrimination Complaint Form, see
www.rvarc.org or call (540) 343-4417.
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
2 | Page
Table of Contents:
Purpose ………………………………………………………………………………….. 3
Motivation ……………………………………………………………………………… 3
Literature Review …………………………………………………………………….. 5
Methodology ………………………………………………………………………….. 9
Analysis ………………………………………………………………………………… 12
Results Summary …………………………………………………………………… 26
Application of Results to RVAMPO Study Area ……………………………. 27
Environmental Justice Discussion …………………………………………….. 31
Safety Discussion …………………………………………………………………… 35
Air Quality Discussion …………………………………………………………….. 37
Proposed Intermodal Center (Montgomery County) …………………… 38
2002-03 Wilbur Smith Regional Freight Study …………………………….. 39
Bibliography ………………………………………………………………………….. 42
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
3 | Page
Purpose:
The regional freight study has two separate, but interrelated purposes. The first purpose is to estimate
the relationship between number of employees in a business and the value, volume and weight of
outbound and inbound freight in the Roanoke Metropolitan Planning Organization (RVAMPO) Study
Area Boundary 2035. The second purpose is to apply the empirical results, found by pursuing the first
purpose, to the RVAMPO Transportation Analysis Zone (TAZ) structure, and to discuss the applied results
with regards to other regional transportation planning factors such as:
x Environmental Justice -which incorporates impacts on low-income and minority populations;
x Public Transportation;
x Non-motorized Transportation;
x Transportation Safety; and,
x Regional Air Quality
The reason that freight transportation is discussed in relation to these other planning factors, is that
freight transportation has an interrelation with passenger transportation, employment dynamics and
economic development implications. Freight vehicles that use the public right-of-way also intermingle
with passenger vehicles in the same transportation infrastructure. Finally, federal guidance encourages
metropolitan planning organizations and rural planning agencies to address transportation planning
with a multi-modal lens while incorporating larger community and economic dynamics.
Motivation:
According to federal law, every urban area with a population of 50,000 or greater in the United States is
required to form a Metropolitan Planning Organization (MPO) and develop, maintain and update a
regional Constrained Long-Range Transportation Plan (CLRTP). d]oo>ZdW[(}}v
estimating passenger travel demand for a base year and projecting passenger travel demand to a future
horizon year typically 20 years or more from the base year. Freight transportation is assessed indirectly
in this process through calibration and validation of the computerized 4-step travel demand model.
Essentially, traffic counts are taken which indicate the proportion of vehicles with 3 or more axels in a
traffic flow, and that proportion is reported as a truck percentage. This truck percentage is then
converted into passenger car equivalents using equivalents such as: a vehicle with a certain number of
axels is the equivalent of three passenger cars as far as traffic flow is concerned. The passenger car
equivalents vZv]vPZ^d((]]Pvuv^_~^•}(Z-step travel demand
model. This conventional indirect method of factoring in freight transportation is likely to be incomplete
given current realities of freight transportation demand such as:
x the increasing popularity of supply chain management and logistics management approaches in
manufacturing, light manufacturing, distribution and retail businesses;
x the increasing popularity of retail purchases from the internet which require shipment to the
purchaser; and,
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
4 | Page
x Z]v]vP}(Z](o(]oouvvo}P]]}]]v^}}_
o}P]]}ul}]Z]o(}]vP}vZ]^}]vX_
For the aforementioned reasons, researchers and planners desire to more completely assess and
estimate freight transportation demand and to incorporate that demand with the passenger travel
demand estimated by the conventional 4-step travel demand model. Fully incorporating freight travel
demand estimates into the transportation planning process is a complicated and multi-year research
endeavor. This study seeks to complete the first step of the process by investigating and assessing
regional specific (RVAMPO) relationships between the number of employees in a business and average
freight generated (outbound) and received (inbound) as measured by freight volume, value and average
shipment weight. These relationships will then be used with the socioeconomic data that is used in the
transportation planning process t}Pv^(]PZPv]}v_}(]o in the base year for the
various Transportation Analysis Zones (TAZs) in the RVAMPO Study Area Boundary (see map below).
The scope of this study will end at the regional freight generation profile. Subsequent research will be
needed over the years to incorporate the three other travel demand estimation steps (trip distribution,
mode choice and traffic assignment) specifically with regards to freight transportation.
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
5 | Page
The development of a regional freight generatiov}(]o]oo]oo(o^(]o_ or
^ov_}]vZv}]}vovv]vP}v]ooused as additional information to the
process. Additionally, the results of this study may be useful to the economic development community.
A relationship between employees and freight value could provide useful information to regional
economic developers.
Literature Review:
There is some past research on freight trip generation estimates within the United States and abroad.
Past research has generally focused on specific industry sectors of interest. There has generally not
been research into overall freight per employee or freight per square foot measures that can be applied
to an entire MPO at the TAZ level. This research will fill that gap by developing general freight
generation estimates and applying them to the RVAMPO TAZ structure where applicable. With this goal
in mind, following is a review of the applicable literature.
In 2003 Jones and Sharma presented a paper to the Transportation Research Board Annual Meeting in
tZ]vP}v]o^o}uv}(&]PZ^]d]&}]vPD}o(}El_dZ]
research estimated freight trips using and economic input-output model called IMPLAN which is
developed and maintained by the University of Minnesota. IMPLAN is tailored to model economic
transactions and multiplier effects over a specific geography, in this case Nebraska. An estimate of
freight trips were derived from the economic transactions modeled in IMPLAN. In this regard, freight
transportation was not surveyed directly; rather it was derived by the economic transactions within
Nebraska. Following is a table listing estimated freight tons per employee according to industry code in
Adams County:
(Jones and Sharma, 11)
Another report out of the Transportation and Economic Development Center at George Mason
hv]]]o^DP-regions and Freight: Evidence from Commodity Flow Survey and Freight Analysis
&u}l_X:}vZv’]((}oX]fically deals with the question of whether freight
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
6 | Page
generation measures such of annual tons per employee are different in so-named Mega Regions than in
other outside of Mega Regions. A geographic representation of Mega Regions using two different
definitions for North America follows:
(Gifford,
Chen, Kelekar, Zebrowski and Zhou, 10)
dZ]}]u}vuo}vo}v(}Pvo^(]PZo]v_}Z
inside and outside of Megaregions using the CFS area boundaries depicted above. The estimates derive
from the Commodity Flow Survey and the Freight Analysis Framework. Both are federal level
publications of aggregate freight data. A summary of relevant estimates to the RVAMPO study follow:
Mean Value Non-Mega-regions Mega-Regions t-value *
7RQVSHUHPSORHHLQ 3.38 1.45 5.3727
Value per employee (in millions) 1.76 1.43 3.2269
Ton-miles per employee (in millions) 1.41 0.34 5.4526
Value per ton (Outbound) 1.32 0.74 (5.4341)
Value per ton (Inbound) 1.09 0.80 (4.8618)
Outbound-to-Inbound Value ratio 0.90 1.05 (3.0154)
Outbound-to-Inbound Tonnage ratio 0.90 1.13 2.3024
Note: * – :HOFKVW-test yielded t-values indicating significant differences between the two groups.
(Gifford, Chen, Li, Kelekar, Zebrowski and Zhou, 15)
A third study, (}uZEZov]o^&]PZ]Pv]}v(]u_/]vPU
Meester and Tavasszy for the 42n European Congress of the Regional Science Association in 2002. Iding
et al. explicitly acknowledge that
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
7 | Page
^Z}}vP]}v]Poo}u]vP}UvZ]ouZv]}vZvP]v}Z}o}(
the private car in this context, the contribution of freight transport by trucks to traffic congestion has
received relatively little attention so farY. In order to predict traffic generation by industrial sites, the
relation between the different activities on an industrial site and the number of freight trips has to be
specified. Information about this relation is scarce however. In the Netherlands and elsewhere, it has
Zov]X_ (Iding, Meester and Tavasszy, 2)
Iding et al. summarized results from the previous freight trip generation studies in the United States in
the following table:
(Iding, Meester and Tavasszy, 5)
In order to establish up-to-date estimates of freight trip generation relationships, Iding et al. developed
a study methodology based on a large sample size and linear regression relationships. The initial sample
size of 10,000 firms yielded 1,529 responses or a response rate of 15%. Iding et al. separated industry
sectors by industry code and reported results of sectors with more than 10 respondents. They found
that
^&}u}}(Z}o]}vv]~ueither by area or by number of employees)
and number of freight trips can be proven. The strength of this relation varies considerably, however. In
some branches of industry (like wood products, chemicals, glass and pottery) R2
is rather high. For
wholeo]]]UZo]}v]lvZo]Pv](]vX_ (Iding, Meester and Tavasszy, 9)
Iding et al reported the regression coefficients for the various industry sectors in the study. Below is an
example of their results for outbound freight by area and number of employees in the Netherlands:
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
8 | Page
((Iding, Meester and Tavasszy, 10)
In summary, the literature review reinforced that freight trip generation rates have not been well
studied in the literature. There are three studies of note with partial applicability to the RVAMPO Study
Area summarized below:
Study Data Source / Approach Relevance to RVAMPO Study
Development of Freight
Statewide Trip Forecasting
Model for Nebraska IMPLAN , an economic
assessment package employing
an input-output model was used
to estimate freight trips based
on economic activity Estimated annual freight weight
generated per employee for a
variety of industries in Nebraska.
Mega-regions and Freight:
Evidence from Commodity Flow
Survey and Freight Analysis
Framework Commodity Flow Survey from
the Bureau of Transportation
Statistics and Freight Analysis
Framework from Federal
Highway Administration Estimated annual freight weight
generated per employee and
annual freight value generated
uo}Z^DP
Reg]}v_P}PZ]oX
Freight trip generation by firms
(Netherlands) 10,000 firms sampled with 1,529
responses. Data relationships
evaluated using linear regression Corroborates a linear regression
approach and provides R2
values
for comparison.
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
9 | Page
Literature review reinforces the concept that a freight generation study focused on the Roanoke Valley
is needed to uncover local freight generation relationships. Results from the existing studies are not
general enough to apply directly to the RVAMPO study area as they are.
A fourth study is mentioned separately because it had a direct influence on the design and execution of
this particular study. The Mobile Area Metropolitan Planning Organization Freight Plan t Final Report
June 21, 2010 prepared for the South Alabama Regional Planning Commission by the Center for
Management & Economic Research at the University of Alabama in Huntsville was used to develop the
survey instrument. Originally the plan was to partner with the Center for Management & Economic
Research at the University of Alabama in Huntsville and have them analyze RVAMPO data based on their
sophisticated technique using survey results and Freight Analysis Framework (FAF) national level data.
Unfortunately, funding did not come through for the Center for Management & Economic Research
analysis and RVAMPO staffs were required to develop an alternative in-house data analysis
methodology.
Methodology:
The methodology for this study centers around a two page survey instrument that was targeted at
businesses that ship or receive inputs or final product on a fairly routine basis. The target was not
necessarily transportation, freight or third-party logistics firms; rather it was firms in other areas of
business who generate freight in the normal operation of business. The two-page survey instrument,
depicted on the following pages, was derived from a similar survey developed for the South Alabama
Regional Planning Commission (Mobile Area) by the University of Alabama in Huntsville Center for
Management and Economic Research.
The sampling methodology for Phase 1 of data gathering process consisted of sampling businesses in the
top 10 freight producing industry classifications in the Roanoke Region on a geographic basis using
Geographic Information Systems software. The goal was to get a representative geographic coverage of
businesses, av}Zo^(]PZ]v]_]Z]Z]v}PZ]v(}u]}v
and discuss freight transportation issues. This involved a time consuming process of cold calling the
targeted businesses and setting up an appropriate interview time for a planner to visit the business at a
later date. In total 29 of the eventual 57 survey responses were obtained in this face-to-face manner.
After some time it became clear that the face-to-face method would not produce a sufficient quantity of
completed surveys in a reasonable amount of time. In Phase 2 of the data gathering process an
electronic version of the two-page survey was created and hosted on SurveyMonkey. The
SurveyMonkey link was posted to Facebook Pages and Linkedin Groups for the Roanoke Regional
Chamber of Commerce, the Roanoke Partnership, the Roanoke Blacksburg Technology Council and other
industry associations and groups. The SurveyMonkey link was distributed in both the City of Roanoke
vZ}v}l}v[}v}u]o}uv-newsletters. In addition, a planner would attend
regional conferences and events such as the Roanoke Career and Lifestyle Fair with small business card
sized survey invitations completed with the electronic survey link. The more mass market approach to
data gathering in Phase 2 garnered an additional 28 responses for a total of 57 responses from Phases 1
and 2. Since both a geographically diverse targeting method and a mass market method were employed
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
10 | Page
in data gathering, it is felt that the sample is representative of the general business community in the
Roanoke Valley. Due to the nature of the questions which ask about value and revenue, there was a
noticeable incidence of partially completed surveys where one or more responses were skipped due to
proprietary reasons of the businesses themselves. As such, some measures reported in the analysis
portion of this report will have between 20 and 40 data points rather than the full 57.
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
11 | Page
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
12 | Page
Analysis:
Ordinary Least Squares (OLS) linear regression analysis was used to model the relationship between
variables in the survey results. For single variable regressions there were some relationships that
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
13 | Page
produced statistically significant results based on F-test for the entire regression equation and t-test
and /or P-values for the individual parameters. The table below summarizes the status of the various
single variable relationships. The significant relationships will be described first and the relationships
that are not statistically significant will also be reported for informational purposes.
Statistical Significance of
Results /v}v K}v
vvo&]PZso
uo} z-^]Pv](]v z-^]Pv](]v
vvodlt]PZ
uo} z-^]Pv](]v 0 0 R
0 0 R0 0 R0 0 R
0 0 R0 0 R0 0 R0 0 R
vvos}ou~}(
^Z]uvuo} 0 0 R
0 0 R 0 0 R
0 0 R0 0 R
Annual Freight Value (Inbound) per Employee t Statistically Significant:
y = 87397x – 562993
R² = 0.7569
-$2,000,000$0$2,000,000$4,000,000$6,000,000$8,000,000$10,000,000$12,000,000$14,000,000
020406080100120
Annual Inbound Freight Value Number of Employees Annual Inbound Value Regressed on # of
Employees
Series1
Linear (Series1)
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
14 | Page
The relationship between the annual value of inbound freight value and the number of employees is
described by the following equation: Annual Value Inbound = ($87,397 x Number of employees) –
$562,993. The Adjusted R2
of this model is 0.74 (note the graph displays R2
instead of Adjusted R2
because of Microsoft Excel default functionality), which means that approximately 74% of the variation
in the data is explained by this equation. The regression equation as a whole is significant at the 1%
level according to the Significance F value, and the variable itself is significant at the 1% level according
to the P-value.
Essentially the equation provides an annual rate of $87,397 of inbound freight value generated per
employee. This compares with a ratio of the mean inbound freight value / mean number of employees
which equals $68,017. A summary of these two results follows:
Estimate for amount of Annual Inbound Freight
Value per Employee
Coefficient of Regression Analysis $87,397 per employee
Mean Value / Mean Number of Employees $68,017 average per employee
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
15 | Page
Annual Freight Value (Outbound) per Employee t Statistically Significant:
The relationship between the annual value of outbound freight and the number of employees is
described by the following equation: Annual Value Outbound = ($273,544 x Number of employees) –
$1,470,970. The Adjusted R2
of this model is 0.51 (note the graph displays R2
instead of Adjusted R2
because of Microsoft Excel default functionality), which means that approximately 51% of the variation
in the data is explained by this equation. The regression equation as a whole is significant at the 1%
level according to the Significance F value, and the variable itself is significant at the 1% level according
to the P-value. y = 273544x – 1E+06
R² = 0.5347
-$10,000,000$0$10,000,000$20,000,000$30,000,000$40,000,000$50,000,000
020406080100120
Annual Value of Outbound Freight Number of Employees Annual Value of Goods Shipped regressed on
Number of Employees
Value of goods shipped
Linear (Value of goods shipped)
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
16 | Page
Essentially the equation provides an annual rate of $273,544 of outbound freight value generated per
employee. This compares with a ratio of the mean outbound freight value / mean number of employees
which equals $223,717. A summary of these two results follows:
Estimate for amount of Annual Outbound Freight
Value per Employee
Coefficient of Regression Analysis $273,544 per employee
Mean Value / Mean Number of Employees $223,717average per employee
Annual Truck Weight (Inbound) per Employee t Statistically Significant:
y = 232375x + 1E+06
R² = 0.6352
020,000,00040,000,00060,000,00080,000,000100,000,000120,000,000140,000,000160,000,000180,000,000
0100200300400500600
Annual Inbound Truck Weight in lbs Number of Employees Annual Inbound Truck Weight Regressed on # of
Employees without 2 Outliers
Series1
Linear (Series1)
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
17 | Page
The relationship between the annual weight of inbound freight and the number of employees is
described by the following equation: Annual Freight Weight Inbound (in lbs) = (232,375 lbs x Number of
employees) + 1,333,590 lbs. The Adjusted R2
of this model is 0.61 (note the graph displays R2
instead of
Adjusted R2
because of Microsoft Excel default functionality), which means that approximately 61% of
the variation in the data is explained by this equation. The regression equation as a whole is significant
at the 1% level according to the Significance F value, and the variable itself is significant at the 1% level
according to the t-stat and P-value.
Essentially the equation provides an annual rate of 232,375 lbs of inbound freight weight generated per
employee. This compares with a ratio of the mean inbound freight weight / mean number of employees
which equals 249,100 lbs /employee. A summary of these two results follows:
Estimate for amount of Annual Inbound Freight
Weight per Employee
Coefficient of Regression Analysis 232,375 lbs per employee
Mean Inbound Weight / Mean Number of
Employees 249, 100 lbs average per employee
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
18 | Page
Annual Truck Weight (Outbound) per Employee t 0 0 RStatistically Significant:
The relationship between the annual weight of outbound freight is not statistically significant. The
Adjusted R2
of this model is negative. The regression equation as a whole is not significant at any
conventional significance level according to the Significance F value, and the variable itself is not
significant at any conventional significance according to the t-stat and P-value.
This lack of significance could be due to the fact that many businesses are shipping out less than
truckload (LTL) quantities due to logistics or just-in-time supply reasons. For instance a sophisticated
retailer such as Wal Mart may have an integrated ordering system whereby when a certain amount of
product is sold off the shelf an automatic replenishment order is sent to the supplier. This means that
the supplier may have shipping weight per shipment that fluctuates directly with retail patterns and
demand and does not correlate with number of employees at the supplier. It is noteworthy that the y = 120243x + 4E+07
R² = 0.0295
050,000,000100,000,000150,000,000200,000,000250,000,000300,000,000350,000,000400,000,000
0100200300400500600
Annual Truck Weight in Pounds Number of Employees Annual Truck Weight Regressed on # of
Employees
Series1
Linear (Series1)
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
19 | Page
difference between the regression coefficient rate and the mean is much larger than the previous
example. This is probably due to the fact that the regression equation is not statistically significant. A
summary of results follows (warning not statistically significant):
Estimate for amount for Annual Outbound
Freight Weight per Employee 0 0 R
Coefficient of Regression Analysis 120,243 lbs per employee
Mean Inbound Weight / Mean Number of
Employees 595,690 lbs average per employee
If the analysis is narrowed to the businesses that fall within the standard transportation classification
SCTG-33, which primarily includes metal work and metal fabricators, the results become statistically
significant. However, there are only 6 data points that correspond with SCTG 33 in our sample.
Following are the SCTG 33 results for comparison:
y = 4E+06x – 2E+07
R² = 0.8392
-50,000,000050,000,000100,000,000150,000,000200,000,000250,000,000300,000,000350,000,000400,000,000
020406080100
Annual Truck Weight in LBS Number of Employees Annual Truck Weight Regressed on # of
employees for SCTG 33 Only
Series1
Linear (Series1)
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
20 | Page
Interestingly, an exponential regression fits the date for SCTG 33 even better than a linear regression
(below):
Annual Volume (Inbound) # of shipments per Employee t 0 0 R
y = 500153e0.0832x
R² = 0.9403
050,000,000100,000,000150,000,000200,000,000250,000,000300,000,000350,000,000400,000,000450,000,000
020406080100
Annual Truck Weight in LBS Number of Employees Annual Truck Weight Regressed on # of
employees for SCTG 33 Only
Series1
Expon. (Series1)
Expon. (Series1)
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
21 | Page
The relationship between the annual (inbound) volume – # of shipments is not statistically significant. In
fact the regression output is very similar to the previous case of annual outbound truck weight. The
Adjusted R2
of this model is negative. The regression equation as a whole is not significant at any
conventional significance level according to the Significance F value, and the variable itself is not
significant at any conventional significance according to the t-stat and P-value.
This lack of significance could be due confusion over the word ^shipment._ The researchers meant that
Z}^Z]uv_o}Z]o~]XXlv]vP}o]vPZu]X/uo]v
the face-to-face interviews during Stage 1 of the data collection that some businesses use the term y = 3.83x + 1205.8
R² = 0.0211
02,0004,0006,0008,00010,00012,00014,00016,000
0100200300400500600
Number of Shipments as reported Number of employees Last Year – # of shipments vs. # of employees
Last Year – # of shipments
Linear (Last Year – # of
shipments)
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
22 | Page
^Z]uv_in a manner akin to an order or purchase order depending on which accounting or
enterprise resources planning software they were using. In this regard a particular vehicle could have
uv^Z]uv_}v](}uZ]]XdZ]}v(]}voed up in later versions of the
survey instrument; however, it could have influenced results for this question. Also, similar to the case
of annual outbound weight per employee, many businesses have switched to a supply chain
management, logistics management or just-in-time type of approach for inputs and /or finished product.
dZvu}(^Z]uv_}o]vu}}vuo}P]]uvPuv(}
rather than the number of employees. This may not have been an issue for the annual value analyses
vZ}(uoo^Z]uv_}o]ooPPP}ovvoo}((]PZXFor
this and other variables that do not show statistical significance in the regression analysis, the mean
(average) value per employee may be a better measure. Nevertheless, a summary of results follows
(warning not statistically significant):
Estimate for amount for Annual Volume
(Inbound) per employee 0 0 R
Coefficient of Regression Analysis 3.83 shipments per employee
Mean Inbound Weight / Mean Number of
Employees 23.80 shipments per employee
Once again there is a large discrepancy between the mean and the coefficient.
Annual Volume (Outbound) # of shipments per Employee t 0 0 R
y = 5.1282x + 2777.2
R² = 0.0405
05,00010,00015,00020,00025,000
020040060080010001200N
u
m
b
e
r
o
f
S
h
i
p
m
e
n
t
s
Number of Employees Annual Shipment vs. # of Employees
Last Year – # of shipments
Linear (Last Year – # of
shipments)
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
23 | Page
The relationship between the annual (outbound) volume – # of shipments is not statistically significant.
In fact the regression output is very similar to the previous cases of annual (inbound) volume and annual
outbound truck weight. The Adjusted R2
of this model is negative. The regression equation as a whole is
not significant at any conventional significance level according to the Significance F value, and the
variable itself is not significant at any conventional significance according to the t-stat and P-value.
]vZ]}UZ]ol}(]Pv](]v}o}v(]}v}Z}^Z]uvX_dZ
ZuvZZ}^Z]uv_o}Z]o~]XXlv]vP}o]vPZ
premises. It became clear in the face-to-face interviews during Stage 1 of the data collection that some
]vZu^Z]uv_]vuvvl]v}v}}Z}v]vP}vZ]Z
accounting or enterprise resources planning software they were using. In this regard a particular vehicle
}oZuv^Z]uv_}v](}uZ]]XdZ]}v(]}vo]vo
versions of the survey instrument; however, it could have influenced results for this question. Also,
similar to the case of annual outbound weight per employee, many businesses have switched to a
supply chain management, logistics management or just-in-time type of approach for inputs and /or
(]v]Z}XdZvu}(^Z]uv_}o]vu}}vuo}P]]
management factors rather than the number of employees. This may not have been an issue for the
vvoovovZ}(uoo^Z]uv_}o]ooPPP}ovvo
value of freight. For this and other variables that do not show statistical significance in the regression
analysis, the mean (average) value per employee may be a better measure. Nevertheless, a summary of
results follows (warning not statistically significant):
Estimate for amount for Annual Volume
(Outbound) per employee 0 0 R
Coefficient of Regression Analysis 5.13 shipments per employee
Mean Inbound Weight / Mean Number of
Employees 31.80 shipments per employee
As with all the cases in which the regression model is not statistically significant, there is a gap between
the coefficient of the regression and the mean per employee. In these cases it is probably advisable to
go with the mean, rather than the regression model.
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
24 | Page
Multiple Regressions for Three Cases of Statistical Insignificance:
For the three cases in which the OLS regression is not significant when regressed on employees (annual
truck weight outbound, annual volume inbound and outbound) a multivariate regression was performed
using both employees and square feet under roof as independent variables. Results for these three
cases follow:
Annual Truck Weight Outbound: – 0 0 R
The multivariate regression still does not produce an acceptable Adjuster R Square and the P-values for
the variables are not statistically significant at any of the conventional significance levels (10%, 5% or
1%). Also, the Significance F for the regression as a whole is not statistically significant at any of the
conventional significance levels (10%, 5% or 1%). In this case, the multivariate regression did not
substantially help in building a better model.
Annual Volume (number of shipments) Inbound: – 0 0 Rt.
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
25 | Page
The multivariate regression still does not produce an acceptable Adjuster R Square and the P-values for
the variables are not statistically significant at any of the conventional significance levels (10%, 5% or
1%). Also, the Significance F for the regression as a whole is not statistically significant at any of the
conventional significance levels (10%, 5% or 1%). In this case, the multivariate regression did not
substantially help in building a better model.
Annual Volume (number of shipments) Outbound: – Multivariate regression statistically significant at the
10% confidence level.
In this case the Adjusted R-square is not optimal; however the Significance F indicates that the
regression model as a whole is significant at the 10% level. Similarly, the square feet coefficient is
significant at the 5% level; however, the employees coefficient is still not significant and any of the
conventional significance levels (10%, 5% or 1%). In this case the multivariate regression did help to
achieve some statistical significance. Following is a single variable regression focusing on square feet
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
26 | Page
under roof:
The R-Square and p-values are similar for the coefficient of square feet under roof as in the multivariate
case indicating that the regression relationship between Annual Volume outbound and square feet
under roof is significant.
Results Summary:
The following is a summary of results from the regression analyses. The statistically significant cases are
depicted with green font and the non-significant cases are depicted in red font. The mean (average) is
also given to compare with the regression coefficient. The regression coefficient can be thought of as a
rate in the sense that when one more employee is added the annual outbound freight value increases
by $273,544. A similar relationship holds for the other coefficients. The average is merely the average
for the dataset.
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
27 | Page
The three cases in which the regression does not yield statistically significant results the mean may be
better suited to use for freight generation purposes. Please note that the gap between the regression
coefficient value and the mean is much greater with the statistically insignificant results than with the
statistically significant results.
The results can be useful to regional planning efforts. The statistically significant results for inbound and
outbound freight value per employee will allow for an estimation of freight value at the TAZ level for
transportation planning purposes. Freight value does not particularly help with traffic flow or
congestion measures; however, it could be of use in a joint transportation planning / economic
development planning endeavor. The statistically significant result for annual inbound truck weight per
employee could help by establishing freight attractions at the TAZ level for the region. The regression
coefficients for the three statistically insignificant cases cannot be used with confidence in long-range
planning. However, the means can be substituted as an alternative measure. There still may not be
much confidence in using the means; nevertheless, the means represent the average of the actual data
collected.
There are several possible reasons that contributed to the lack of statistical significance for the three
cases. First of all, there was initial confusion between the accounting and /or information technology
uv]vP}(^Z]uv_vZv}]}vuv]vP}(^Z]uvX_dZ]Pv]vv
Z}^Z]uv_}vZ]o]vP(]PZv]ng or leaving the premises. Some
]vZ}^Z]uv_}(}](]]uZ}u}Pv
their accounting /information technology system. Thus, from their point of view a particular vehicle
}ouv^Z]uvX_dZ]}v(]}vv}]o]vZPZ]vP}v
subsequent versions of the survey instrument clarified that shipment was intended to signify a vehicle
carrying freight. Additionally, the widespread adoption of supply chain management and logistics
management techniques by businesses and their suppliers and /or customers could contribute to a lack
of statistical relationship between shipments and truck weight versus employees. For instance, a
supplier to large retail establishments such as Wal-Mart or Target may have to conform to a delivery
(v]Z}u[o}P]]uUZ]Z]o]lo}]vov
inventories. Thus the driver in the relationship are retail market and logistics management dynamics,
not the number of employees at the supplier. The annual value results should be unaffected by these
issues because the many small shipments required by a logistics management system would aggregate
up to an annual value.
Application of Results to RVAMPO Study Area:
The freight generation rates per employee defined by the coefficients of the three statistically significant
regression equations were applied at the TAZ level for the RVAMPO Study Area 2035.
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
28 | Page
Annual inbound freight value per TAZ, depicted on the preceding page, and annual outbound freight
value per TAZ, depicted below, show similar geographic patterns as would be expected. The annual
value ranges per color are higher in the outbound annual freight value map as would be expected by
firms shipping in raw materials and then adding value and shipping out a more valuable finished
product.
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
29 | Page
The geographic pattern for the inbound annual truck weight per TAZ, depicted on the next page, shows
a somewhat similar pattern as the previous patterns pertaining to freight value with some slight
variations. This is due to different relationships between value per employee and truck weight per
employee estimated by the regression equations.
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
30 | Page
The map depicted on the next page relates the Outbound Annual Freight Value per TAZ to recent
Average Annually Daily Traffic (AADT) Truck Percentages estimated from published VDOT traffic counts.
In most cases higher AADT Truck Percentages are observed in close proximity to TAZs that are estimated
to generate higher annual freight value according to the regression equation. However, the area along
^o]Z}U_]]v]Z]oo}UZ}u]u}Z]PZvvo(]PZo
estimates and a relatively light AADT Truck Percentage. This is likely due to several factors:
x The employment along the corridor may currently be skewed toward office and retail uses.
x The regression equation likely over estimates freight generation for office and retail uses due to
averaging effects from light industrial and industrial uses.
However, there are two areas of US 419 that have potential for current and future freight transportation
impacts. The first area is from Starkey Road back to the interchange with US 220, and the second area is
in parts of the City of Salem and Roanoke County near the Salem border. Any future upgrades and
accommodations along US 419 should keep these two sections in mind.
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
31 | Page
Environmental Justice Discussion:
Shifting patterns of freight traffic provide both benefits and burdens for low-income and /or minority
populations. On the one hand employment that generates freight value is often needed in low-income
communities as it can pay better than other types of employment accessible to the population. On the
other hand, increased freight traffic can pose safety and other challenges to residents of a particular
neighborhood. It is beyond the scope of this document to evaluate whether there is a net benefit or
burden to low income neighborhoods with increased freight related employment. As of the writing of
this document, summer 2012, it is conceivable that community leaders would view freight related
employment as a net benefit due to its potential to decrease unemployment and increase average
wages. However, this cannot be demonstrated without a separate specific study on the matter.
The percentage of individuals under the national poverty line for the RVAMPO is depicted in the map on
the following page. It should be noted that the geographic patterns in the freight generation maps and
the poverty line map are not necessarily the same. At a planning level this would indicated that the
populations that fall under the national poverty line may not necessarily be overburdened by freight
producing businesses; however, they may not have greater access to employment and other benefits of 0 0 R
Corridor
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
32 | Page
Z]vX/v(UZu^]ou]uZ_vZ}v]ouo}uvvv
]v]]o[}v}]}vZ}PZo]transportation or other means.
/v}}]Z]}v}(^]ou]uZ_vi}o}(]PZvZ]v]o
locations of potential employees for those jobs this report will introduce some recent public
transportation measures that were developed in the RVAMPO to analyze activity (boarding and
alighting) at existing bus stops on the regions fixed route bus transit system. This is a novel approach to
combine freight and public transportation analysis and measures together. The hope is that existing
patterns in bus (public transit) usage will indicate whether the bus system can help to alleviate any
potential spatial mismatch between potential employment related to freight and those who are seeking
employment. In fact, several of the business stakeholders mentioned this issue during the freight
]v]XdZ]v}Zu^]ou]uZV_Z}Z]]]}vZvZ
needed to hire for several positions that the candidates for those positions lived in a different part of
the urban area and they could not use the bus to get to work because the bus lines did not extend out to
that particular business. These businesses were advocating for increased public transit so that they
could hire more seasonal employees.
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
33 | Page
The map above uses data that was collected as a joint effort by RVAMPO and Valley Metro (Greater
Roanoke Transit Company) staff. The data was collected between July 2010 and June 2011. Over 400
bus routes were randomly sampled yielding a 95% confidence level for the data collection period. An
index was then constructed from the data for each bus stop using the following formula Composite
Activity Index = average usage * stop frequency. The size and color of the dots above represent the
Composite Activit/v(}Z}(Z}}vsooD}[(]}uX
This data is overlaid on the outbound freight value per TAZ map that was developed using the regression
relationship documented earlier in this report. The concept is to identify areas that are predicted to
Pvo}}((]PZ}vZP]}vo]}vZ]Z}v[voZ]v
whose closest bus stops indicate a high activity index. It is anticipated that the combination of these
two factors could indicate that there is potential to extend bus service to address potential employment
spatial mismatch.
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
34 | Page
The two areas of note: Peters Creek t Hollins Corridor and West Salem /Roanoke County are depicted
with the red circles in the map above. Each of these areas has a potential spatial mismatch between
freight related employment and those who would benefit from such employment. Targeted rideshare
marketing should be considered as a way to match potential employees who live in other areas to
employment in these areas. In the case of the Peters Creek /Hollins area there is a corridor specific
planning processes at the local level underway as of the writing of this document. The Hollins Area Plan
can be found at the following link: 0 1 R
Much of the original feedback concerning the potential spatial mismatch between needed employees
and potential employment came from freight interviews with businesses located in the Roanoke Centre
for Industry and Technology in the City of Roanoke. A zoomed in section of the map depicts this area
with a red circle (next page). It is ironic that that industrial park is just outside of the Valley Metro
service area, and that the nearest bus stops show moderate to high activity. A separate discussion with
stakeholders concerning this situation is taking place during Fiscal Year 2013. 0 0 R
t Hollins
Corridor West
Salem /Roanoke
County
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
35 | Page
Safety Discussion:
There are many safety issues that could be discussed with regards to freight transportation. This report
will narrow its focus the potential safety issues that could result from the mixing of significant freight
volumes and bicycle transportation in the same corridor. The photo on the next page shows a bicyclist
beside a cement truck. The photo visually demonstrates the utility of bicycle lanes and wide shoulders
in corridors that are likely to carry both freight transportation and bicycle transportation volumes.
It is beyond the scope of this report to delve into the details and design of bicycle facilities that can
safely and harmoniously be used near corridors with substantial freight traffic volumes. Nonetheless,
this report will highlight three areas within the RVAMPO Study Area that pose the potential to carry
both significant freight and significant bicycle volumes. The three red circles depicted on the map on
page 36 show areas that already have various bicycle accommodations, and are areas predicted to
generate freight on a per TAZ basis. The three areas roughly correspond to the western half of the City
of Salem, the Route 419 corridor and Vinton /NE City of Roanoke area. It is noteworthy that the Peters
Creek /Hollins area that was discussed with reference to spatial mismatch between employees and
freight generating businesses is not highlighted on the map. This is due to the lack of current bicycle
accommodations in the area. Should more bicycle accommodations be constructed in the area they
should be designed with potential safety issues with regards to freight transportation in mind. 0 0 R
for Industry and
Technology
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
36 | Page
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
37 | Page
Air Quality Discussion:
The RVAMPO study area is covered under an Ozone Early Action Compact (EAC) and an Ozone Early
Action Plan (EAP), which were developed 2002-04. The EAC is essentially an agreement between local
governments, the Virginia Department of Environmental Quality (DEQ) and the Federal Environmental
Protection Agency (EPA) to pursue an Ozone EAP before an air quality plan would have been otherwise
required under traditional nonattainment designation. The EAP must incorporate the same scientific
rigor as the traditional approach and the EAP will be incorporated into the State Implementation Plan
(SIP).
In early March 2008 the Federal EPA revised the nationwide 8-hour Ozone Standard to 75 parts per
billion (ppb) based on a 3-PX/v]]}vZZZ}v}lZP]}v[-year average for
the 2006, 2007 and 2008 Ozone seasons are at 74 ppb, within the new nationwide standard. As such, it
is likely that the EAC /EAP will continue to be regarded as successful, and that the RVAMPO
transportation planning process will not have to include the traditional air quality conformity analyses
for the major planning products.
In Spring and Summer 2011, the Federal EPA postponed a new adjustment of the nationwide 8-hour
Ozone Standard until 2013. The Federal EPA has stated that the primary 8-hour Ozone Standard will be
revised to a final value somewhere within the range of 60 ppb to 70 ppb. The Federal EPA asserts that
the final standard will be set sometime in 2013.
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
38 | Page
The chart on the previous page, provided by Virginia DEQ on 04-17-2012 summarizes the ground level
ozone trends in the Roanoke Valley. The trends are downward which is positive for public health and
the prospects of complying with future national Ozone standards.
&]PZZ]o]o]vP}o}]Pv](]vZoovP}ZZsDWK^[v]o]
improvements. Recent inquiries by RVAMPO staff have not uncovered any local level anti-idling
ordinances in the Roanoke Valley. The localities involved in the original Ozone EAC /EAP process do have
anti idling policies for their own fleets of vehicles. It is beyond the scope of this report to discuss the
legal prospects of anti-idling ordinances in Virginia. It is hoped the large freight generators will
voluntarily develop anti-idling policies for their place of business.
Intermodal Center in Elliston (Montgomery County):
In 2008, the Virginia Department of Rail and Public Transportation (VDRPT) selected a site in Elliston,
Virginia for the regional Intermodal Freight Transfer facility for the multi-state Heartland Corridor
Project with Norfolk Southern (NS). The Elliston location is just outside the RVAMPO 2035 study area for
this plan. The following graphic illustrates the proximity of the selected site to the 2035 study area
(shown in purple).
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
39 | Page
Census 2010 results which were released in March 2012 indicates that the RVAMPO urbanized area
boundary (UZA) now extends into Montgomery County. As such Montgomery County will have voting
membership on the RVAMPO Policy Board by the Summer 2013. AS of the writing of this document, the
RVAMPO Policy Board has invited Montgomery County to appoint a liaison member to both the
RVAMPO Policy Board and the Transportation Technical Committee (TTC). The liaison member will be
present at meetings to advise the RVAMPO and TTC on issues pertaining to a new RVAMPO Study Area
Boundary 2040 and bylaws change to incorporate Montgomery County. The following map depicts the
RVAMPO UZA (defined by Census Bureau) in Red, and staff recommendations for the RVAMPO Study
Area Boundary 2040 in green. The proposed intermodal site is not included in the census defined UZA
boundary (Red), it is included in the staff recommended study are boundary (Green). This issue of
whether or not to include the proposed intermodal site in the new RVAMPO study area boundary will
have to be decided through the MPO process before Summer 2013.
2002-03 Wilbur Smith Freight Study:
In 2002-03 RVAMPO and the Roanoke Valley Alleghany Regional Commission (RVARC) contracted with
Wilber Smith Associates to conduct a regional freight study for the Roanoke Valley. That study used the
Reebie (now Transearch) freight database and developed and in-depth analysis of freight flows to and
from the Roanoke Valley. That study also included a freight stakeholder involvement process that
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
40 | Page
o}o]}(Z^d}&]PZ&]}vW}i_ZZ}o}v]]v(ovX
Below are the Ten Fast Action Projects from the original plan:
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
41 | Page
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
42 | Page
Bibliography:
Gifford, Jonathan, Zhenua Chen, Jing Li, Uma Kelekar, Nathaniel Zebrowski, and Xin Zhou. Mega-regions
and Freight: Evidence from Commodity Flow Survey and Freight Analysis Framework. Transportation and
Economic Development Center, George Mason University, 2007. Web. 2012.
RVAMPO and RVARC Regional Freight Study t Technical Report t Final 11-15-2012
43 | Page
Iding, Mirjam H.E., Wilhelm J. Meester, and Lori A. Tavasszy. Freight Trip Generation by Firms: Paper for
the 42nd European Congress of the Regional Science Association Dortmund, 2002. 2002. Web. 2012.
Jones, Elizabeth G., and Anshul Sharma. Proc. of Development of Statewide Freight Trip Forecasting
Model for Nebraska, Transportation Research Board Annual Meeting 2003, Washington DC. Jan. 2003.
Web. May 2012.
University of Alabama in Huntsville, College of Business Administration, Center for Management and
Economic Research. The Mobile Area Metropolitan Planning Organization Freight Plan Final Report.
Project No. SARPC 07-001, June 21, 2010

Leave a Reply

Your email address will not be published. Required fields are marked *