Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 161
APPENDIX K
Existing and New HMCFS Data
Analysis Examples
K.1 Existing Data from Freight Analysis
Framework Database
Description
The spatial data from FHWA's Freight Analysis Framework (FAF) are available at county and
state levels in terms of estimated tons and values for commodity groups. The commodity classi-
fication system in the FAF uses the Standard Classification of Transported Goods (SCTG) codes
at the two-digit level.
Limitations
Because the data are modeled based on a stratified national sample of economic activity and
not actual traffic flows, they are only generally applicable for a local HMCFS and should only
be interpreted in terms of commodity groups that can be expected to be present in a region or
state. Data can only be approximately associated with hazmat class level for the vast majority of
commodities.
Supported Objectives
Increasing awareness about hazmat transport and minimum scenarios definition.
How to Use the Data
1. Develop a listing of commodity flows for your state using Geographic Information Systems
(GIS).
2. Identify commodity groups associated with hazmat transport and use the listing to indicate
what may be transported in your region.
K.2 Existing Data from BTS/Census Bureau
Commodity Flow Survey
Description
The Bureau of Transportation Statistics/Census Bureau 2007 Commodity Flow Survey (CFS)
data are applicable at a state or national level. If an LEPC is interested in using national esti-
mates of hazmat shipments by different modes (including trucks) for local estimates, this is a
good source.
K-1
OCR for page 162
K-2 Guidebook for Conducting Local Hazardous Materials Commodity Flow Studies
Limitations
Data should only be considered generally applicable for a local HMCFS in terms of commodi-
ties that may be expected to be present in a region or state. Estimates have a very high degree of
variability for local networks since they are drawn from a national sample of shipments. They may
be off by a large degree, and additional survey data are necessary to provide further information
about the validity of the data.
Supported Objectives
Increasing awareness about hazmat transport and minimum scenarios definition.
How to Use the Data
1. Access the report at the Internet address listed for the report in Appendix G.
2. Select the desired table, and review the information for hazmat shipments by mode, class, or
characteristic for your state.
3. Develop corresponding listings and tables as an indication of what may be transported in your
region.
Application Example
A local entity is interested in information about transportation of all hazardous materials and
Hazard Class 3 materials. First, they need to access the 2007 Commodity Flow Survey informa-
tion. There they might identify Table Sector 00: CF0700H04: Hazardous Materials Series: Haz-
Mat Shipment Characteristics by Mode by Hazardous vs. Nonhazardous Status for the United
States: 2007. This table shows that a total of 3,344,658 million ton-miles of all commodities (haz-
ardous and non-hazardous) were shipped in the United States, including 1,342,104 million ton-
miles by truck. A total of 103,997 million ton-miles of truck transport were of hazardous materials.
Around 7.7 percent of truck ton-miles shipped were associated with transport of hazardous
materials (103,997/1,342,104).
Another table of interest might be Table Sector 00: CF0700H07: Hazardous Materials Series:
HazMat Shipment Characteristics by Mode by Hazardous Class or Division for the United States:
2007 & 2002. This table shows that a total of 181,615 million ton-miles shipped for all modes are
associated with Hazard Class 3, Flammable or Combustible Liquids, and 55,934 million ton-miles
by truck.
· Hazard Class 3, Flammable or Combustible Liquids, corresponds to 5.4 percent of all ton-
miles shipped for all commodities by all modes (181,615/3,344,658), and 4.2 percent of all
truck ton-miles shipped (55,934/1,342,104).
· Of the hazardous materials shipped by truck in the United States, 53.8 percent were Hazard
Class 3, Flammable or Combustible Liquids (55,934/103,997).
A third table of interest might be Table Sector 00: CF0700H08: Hazardous Materials Series:
HazMat Shipment Characteristics by Mode by UN Number for the United States: 2007. This table
shows that a total of 23,665 million ton-miles shipped by truck are associated with UN/NA Num-
ber 1203 (Gasoline), 16,408 million ton-miles with UN/NA Number 1993 (Various Petroleum Dis-
tillates, including diesel fuel), and 5,729 million ton-miles with UN/NA Number 1202 (Diesel Fuel),
for a total of 45,802 million ton-miles for these UN/NA numbers by truck. (Note that there also are
other UN/NA IDs that are for Class 3 hazardous materials. These IDs are used as examples).
· Of the Hazard Class 3 Flammable or Combustible Liquids shipped in the United States by
truck, 82 percent were associated with UN/NA Numbers 1203, 1993, or 1202 (45,802/55,934).
OCR for page 163
Existing and New HMCFS Data Analysis Examples K-3
· Of the hazardous materials shipped by truck in the United States, 44 percent were associated
with UN/NA Numbers 1203, 1993, or 1202 (45,802/103,997).
K.3 Existing Data from HPMS Combined with
Existing Data from VIUS or CFS
Description
The FHWA's Highway Performance Monitoring System (HPMS) contains information for
annual average daily traffic (AADT) levels for major roadway segments including the state and
national highway systems. The U.S. Census Bureau's 2002 Vehicle Inventory and Use Survey
(VIUS) data are summarized in Appendix H. CFS data are described in the previous section.
Limitations
Commodity flows estimated using these sources should only be considered generally applica-
ble for a local HMCFS. They have a very high degree of variability since they mix a local estimate,
a local annual sample, and a national annual sample; they may differ from the true value by a large
degree. Additional survey data are necessary to provide further information about the validity of
the data.
Supported Objectives
Increasing awareness about hazmat transport and minimum scenarios definition.
How to Use the Data
1. Obtain AADT estimates for major roadway segments in your jurisdiction.
2. Determine the percentage of truck traffic in the local area that makes up total traffic (estimate
or other information source).
3. Apply the percentage of total traffic that is trucks to the AADT values to estimate the truck
traffic levels.
4. Apply the overall percentages of hazmat truck traffic from the bottom row of the 2002 VIUS
data table to the estimated truck traffic levels, or apply percentages of hazardous materials by
truck versus all commodities by truck from the 2007 CFS, for a crude estimate of numbers of
hazmat trucks on applicable segments
5. Present the information in lists and tables, as applicable.
Application Example
A local entity is interested in estimating the number of trucks per day transporting Hazard
Class 3 materials over a particular Interstate highway segment. According to the HMPS traffic vol-
ume map, the AADT of an Interstate section is over 100,000 (all vehicles). The local entity assumes
that truck traffic is 15 percent of the overall traffic volume (caution: this value is used as an example
only and has no applicability to roadways in your jurisdiction). This corresponds to over 15,000
trucks per day, on average.
Based on the 2002 VIUS data, a total of 2.3 percent of U.S. miles are driven by trucks while
requiring a Class 3 placard or "Combustible" placard. According to the 2007 CFS, Hazard Class 3,
Flammable or Combustible Liquids, corresponds to 4.2 percent of all truck ton-miles shipped for
all commodities. Using these estimates and assuming that all trucks on the roadway section are
driven the same distance through the jurisdiction, the local entity might expect to have between
OCR for page 164
K-4 Guidebook for Conducting Local Hazardous Materials Commodity Flow Studies
350 and 630 trucks per day carrying Class 3 liquids on the Interstate segment (15,000 * 0.023 = 345;
15,000 * 0.042 = 630).
K.4 Total Truck Counts Combined with Existing Data
from VIUS or CFS
Description
This method uses counts of the number of trucks on different roadway segments to identify
truck traffic volumes, rather than HPMS traffic level estimates. However, this method still neces-
sitates application of national percentages of hazmat truck traffic from the bottom row of the
2002 VIUS data table (found in Appendix H) or 2007 CFS data.
Limitations
By eliminating some of the measurement error from the previous method, this method is
probably slightly more relevant at the local level than estimates generated entirely from existing
data sources. However, commodity flows estimated using these sources should still only be con-
sidered generally applicable for a local HMCFS. They have a very high degree of variability since
they mix a local annual sample with a national annual sample; they may differ from the true value
by a large degree. Additional survey data are necessary to provide further information about the
validity of the data.
Supported Objectives
Conducted with convenience or representative sampling, supported objectives may include
increasing awareness about hazmat transport and minimum definition of training scenarios
(depending on the quantity and quality of data).
How to Use the Data
1. Determine truck traffic levels and patterns. This may range from a general estimate of truck
traffic in the entire jurisdiction to levels of truck traffic by time for represented locations.
2. Apply the overall percentages of hazmat truck traffic from the bottom row of the VIUS data
table to the estimated truck traffic levels, or apply percentages of hazardous materials by truck
versus all commodities by truck from the 2007 CFS, for a crude estimate of numbers of hazmat
trucks for represented locations.
3. Present the information in lists, tables, or charts, as applicable.
Application Example
A local entity is interested in estimating the number of trucks per day transporting Hazard
Class 3 materials over a particular Interstate highway segment. A state DOT performs counts of
trucks on a section of Interstate highway and provides the data to the local entity. The 2007 aver-
age annual daily truck traffic (AADTT) for the Interstate section was 9,210.
Based on the 2002 VIUS data, a total of 2.3 percent of U.S. miles are driven by trucks while
requiring a Class 3 placard or "Combustible" placard. According to the 2007 CFS, Hazard Class 3,
Flammable or Combustible Liquids, corresponds to 4.2 percent of all truck ton-miles shipped
for all commodities. Using these estimates and assuming that all trucks on the roadway section
are driven the same distance through the jurisdiction, the local entity might expect to have
OCR for page 165
Existing and New HMCFS Data Analysis Examples K-5
between 210 and 390 trucks per day with a Hazard Class 3 Flammable Liquids placard on the
Interstate segment (9,210 * 0.023 = 212; 9,210 * 0.042 = 387).
K.5 Truck Type/Configuration Counts Combined with
Existing Data from VIUS
Description
This method uses counts of trucks by type and configuration on different roadway segments,
rather than generic truck counts. This allows for application of national percentages of hazmat
traffic for each truck type and configuration from respective rows of the 2002 VIUS data table
(see Appendix H).
Limitations
By further specifying the nature of truck traffic over the generic truck counts, it is probably
slightly more relevant at the local level than estimates generated using only generic truck counts.
However, these estimates still have a high degree of variability since they mix a local annual sam-
ple with a national annual sample; they may be off by a large degree. Additional survey data are
necessary to provide further information about the validity of the data. These estimates should be
considered as only having low-to-medium applicability for a local HMCFS in terms of level of
hazmat traffic that may be expected to be present in a community.
Supported Objectives
Conducted with convenience or representative sampling, supported objectives may include
increasing awareness about hazmat transport, minimum scenarios definition, and maximum
scenarios definition (depending on the quantity, quality, and validity of data).
How to Use the Data
1. Determine truck traffic levels and patterns by type and configuration. This may range from
estimates of truck traffic in the entire jurisdiction to levels of truck traffic by time for specific
locations.
2. Apply the percentages of hazmat truck traffic from the corresponding rows of the VIUS data
table to the observed truck traffic levels by type and configuration for a crude estimate of
numbers of hazmat trucks for represented locations.
3. Present the information in lists, tables, or charts, as applicable.
Application Example
A local entity is interested in estimating the number of trucks per day transporting Hazard Class 3
materials over a particular Interstate highway segment. They have information that shows that the
2009 truck traffic on the Interstate segment was 500 tank trucks per day, 2,500 flatbed trucks per
day, 3,000 refrigerated van trucks per day, and 3,500 standard van trucks per day (the LEPC only
counted trucks by type, not configuration).
Based on the 2002 VIUS data, 23.3 percent of U.S. tank truck miles, 0.4 percent of flatbed
miles, 0.4 percent of refrigerated van miles, and 1.3 percent of standard van miles are driven
while requiring a Class 3 placard or "Combustible" placard. Using these estimates and assuming
that all trucks on the roadway section are driven the same distance through the jurisdiction, the
OCR for page 166
K-6 Guidebook for Conducting Local Hazardous Materials Commodity Flow Studies
local entity might expect to see around 190 Hazard Class 3 Flammable or Combustible Liquids
trucks per day on the Interstate segment ((500 * 0.233) + (2,500 * 0.004) + (3,000 * 0.004) +
(3,500 * 0.013) = 184).
K.6 Placard Counts Combined with Total Truck Counts
Description
By counting the total number of trucks observed on a roadway segment and observing whether
or not each truck has a hazmat placard, a locally relevant estimate of the total percentage of truck
traffic that has a hazmat placard can be made. This may be particularly useful for locations for
which specifically identifying a placard (e.g., by class/division or number) are challenging, such as
locations that are some distance from the observed traffic, or where traffic is travelling at high rates
of speed with limited time for truck observations.
Limitations
For purposes of locally relevant identification of presence or absence of hazardous materials, this
method is sufficient. However, it does not inform about the types of hazardous materials being
transported without application of national estimates such as the 2002 VIUS data. When this is
done, estimates mix locally relevant survey data with local and national samples. They may be off
by a moderate-to-high degree. Follow-on survey data may provide further information about the
validity of the information.
Supported Objectives
Conducted with convenience or representative sampling, supported objectives may include
increasing awareness about hazmat transport and minimum scenarios definition (depending on
the quantity and quality of data).
How to Use the Data
1. Determine truck traffic levels and patterns. This may range from a general estimate of truck
traffic in the entire jurisdiction to levels of truck traffic by time for represented locations.
2. Determine placarded truck traffic levels and patterns. This may range from a general estimate
of placarded truck traffic in the entire jurisdiction to levels of truck traffic by time for repre-
sented locations.
3. Create a ratio of placarded trucks to overall trucks that can be estimated for applicable loca-
tions and times.
4. Present the information in lists, tables, or charts, as applicable.
Application Example
A local entity is interested in estimating the number of trucks per day transporting Hazard Class 3
materials over a particular Interstate highway segment. They conduct a 24-hour placard count dur-
ing a weekday on the Interstate segment. Four-hundred trucks were observed to have a hazmat
placard during the count. The 2007 AADTT for this section of roadway was 9,250, according to the
state DOT. The entity assumes this represents the daytime, weekday traffic level during their plac-
ard count. Using the observed placarded truck count, over 4.3 percent (400/9,210) of trucks on the
Interstate might display a hazmat placard if current truck traffic levels are similar to 2007 traffic
levels. After applying some statistics (see Section K.10) and assuming the placard counts follow a
OCR for page 167
Existing and New HMCFS Data Analysis Examples K-7
Poisson distribution, the local entity is 90 percent confident that the true placard count falls some-
where between 368 and 434 observations, or between 4.0 and 4.7 percent of AADTT.
Based on the 2002 VIUS data, a total of 2.3 percent of U.S. miles are driven by trucks while
requiring a Class 3 placard or "Combustible" placard. Based on the 2007 CFS, 53.8 percent of
hazardous materials shipped by truck in the United States were Hazard Class 3, Flammable or
Combustible Liquids. Using the state DOT AADTT numbers with VIUS data and assuming that
all trucks on the roadway section are driven the same distance through the jurisdiction, around
210 Hazard Class 3, Flammable and Combustible Liquids trucks per day (9,210 * 0.023 = 212)
could be expected on the Interstate. Using the placard count with 2007 CFS data, around 220
Hazard Class 3, Flammable and Combustible Liquids trucks per day (400 * 0.538 = 215) could
be expected.
K.7 UN/NA Placard ID Counts
Description
Observing and identifying specific placards enables recognition of particular truck/roadway
transport hazmat hazards, including the relative proportion of different types of hazardous
materials carried by trucks. Since it does not include a count of trucks, this method may be
appropriate for use by a single data collector at busy traffic locations where both counting of
trucks and identifying UN/NA placard IDs is too difficult.
Limitations
When conditions permit, it is more advantageous to count the number of trucks or num-
ber of trucks by type/configuration in addition to counts of specific UN/NA placard IDs. Reli-
ability of information will depend on the sampling framework applied and the accuracy of data
collection.
Supported Objectives
Conducted with convenience, representative, or cluster sampling, supported objectives may
include increasing awareness about hazmat transport, minimum scenarios definition, maximum
scenarios definition, emergency planning, and identifying equipment needs (depending on the
quantity and quality of data).
How to Use the Data
1. Group UN/NA placard ID information according to class/division, specific ID, TIH classifi-
cation, and associated initial response actions, or other categories.
2. Determine levels and patterns of observed placards (by hazmat grouping). This may range
from a general estimate of observed placards for the entire jurisdiction to levels of observed
placards by time for specific locations.
3. Calculate proportions of hazmat placards observed for each grouping.
4. Present the information in lists, tables, or charts, as applicable.
Application Example
A local entity is interested in estimating the percentage of placarded trucks transporting Haz-
ard Class 3 materials over a particular Interstate highway segment during the daytime. They
OCR for page 168
K-8 Guidebook for Conducting Local Hazardous Materials Commodity Flow Studies
determine the following information for a daytime, weekday 8-hour placard ID count on the
Interstate segment:
· 50 placards with UN/NA placard ID 1203 (Gasoline),
· 25 placards with UN/NA placard ID 1993 (Various Petroleum Distillates),
· 12 placards with UN/NA placard ID 1863 (Aviation Fuel),
· 5 placards labeled "Combustible" or "Fuel Oil,"
· Total number of placards counted: 200, and
· Peak hourly placard count rate from 11 A.M. to 12 P.M. is 35 placards per hour.
Approximately 46 percent of the trucks observed with placards on the Interstate had a Haz-
ard Class 3, Flammable or Combustible Liquids placard ((50 + 25 + 12 + 5) / 200). After apply-
ing some statistics (see Section K.10) using a binomial distribution and assuming that daytime,
weekday hazmat traffic patterns are consistent with the observed time period, the entity identi-
fies that this percentage can be expected to range between 39 and 53 percent with 90 percent con-
fidence. These data provide some estimates, and since the sample was over a limited time period,
follow-on data are necessary to validate the information. The same type of estimates can be
repeated for individual placard IDs that are included in the sample.
K.8 UN/NA Placard ID Counts Combined with
Total Truck Counts
Description
This method includes a count of total trucks and counts of specific UN/NA placard IDs. Not
only can it be used to identify the presence of commodities associated with specific UN/NA plac-
ard IDs, it also can be used to estimate the proportion of observed truck traffic that is placarded.
More locally relevant information is obtained about hazmat transportation using these counts,
but the complexity of the survey is limited because only the total number of trucks is counted,
not the different types of trucks.
Limitations
Reliability of information will depend on the sampling framework applied and the accuracy
of data collection.
Supported Objectives
Conducted with convenience, representative, cluster, stratified/proportional, or random sam-
pling, supported objectives may include increasing awareness about hazmat transport, mini-
mum definition of training scenarios, maximum definition of training scenarios, emergency
planning, identifying equipment needs, comprehensive planning, and route analysis (depend-
ing on the quantity and quality of data).
How to Use the Data
1. Group UN/NA placard ID information according to class/division, specific ID, TIH classifi-
cation and associated initial response actions, or other categories.
2. Determine levels and patterns of observed placards (by hazmat grouping). This may range
from a general estimate of placarded truck traffic in the entire jurisdiction to levels of plac-
arded truck traffic by time for specific locations.
OCR for page 169
Existing and New HMCFS Data Analysis Examples K-9
3. Determine truck traffic levels and patterns. This may range from a general estimate of truck
traffic in the entire jurisdiction to levels of truck traffic by time for specific locations.
4. Calculate proportions of hazmat placards observed (by grouping) to total truck traffic.
5. Present the information in lists, tables, charts, or maps, as applicable.
Application Example
A local entity is interested in estimating the percentage of all trucks transporting Hazard
Class 3 materials over a particular Interstate highway segment during the daytime. The entity
collects the following information for a daytime, weekday 8-hour placard count on the Inter-
state segment:
· 50 placards with UN/NA placard ID 1203 (Gasoline),
· 25 placards with UN/NA placard ID 1993 (Various Petroleum Distillates),
· 12 placards with UN/NA placard ID 1863 (Aviation Fuel),
· 5 placards labeled "Combustible" or "Fuel Oil,"
· Total number of placards counted: 200,
· Total number of trucks counted: 5,000,
· Peak hourly placard count rate for 11 A.M. to 12 P.M. is 35 placards per hour, and
· Peak hourly truck count rate for 1 P.M. to 2 P.M. is 600.
In addition to estimates discussed for the previous example, the following are identified.
Approximately 1.8 percent of all trucks observed on the Interstate had a Hazard Class 3, Flamma-
ble or Combustible Liquids placard ((50 + 25 + 12 + 5) / 5,000). Approximately 4 percent of all
trucks observed on the roadway had a hazmat placard (200/5,000), assuming that daytime, week-
day hazardous materials and overall traffic patterns are consistent with the observed time period.
Hazardous materials truck traffic appears to peak during the late morning. These data provide
some estimates, and since the sample was over a limited time period, follow-on data are necessary
to validate the information. The same type of estimates can be repeated for individual placard IDs
that are included in the sample.
K.9 Placard ID Counts Combined with
Truck Type Counts
Description
This method includes a count of trucks by size/configuration in addition to counts of specific
UN/NA placard IDs. It can be used to identify the presence of commodities associated with specific
UN/NA placard IDs, estimate the proportion of observed total truck traffic that is placarded, as
well as proportions of different types/configurations of trucks that are placarded. Although more
complex observational truck traffic sampling can be performed without conducting interviews or
examining shipping manifests, this method is probably the most complex that can be accom-
plished using HMCFS volunteers.
Limitations
Reliability of information will depend on the sampling framework applied and accuracy of
data collection.
Supported Objectives
Conducted with convenience, representative, cluster, stratified/proportional, or random sam-
pling, supported objectives may include increasing awareness about hazmat transport, minimum
OCR for page 170
K-10 Guidebook for Conducting Local Hazardous Materials Commodity Flow Studies
scenarios definition, maximum scenarios definition, emergency planning, identifying equipment
needs, comprehensive planning, and route designation (depending on the quantity and quality
of data).
How to Use the Data
1. Group UN/NA placard ID information according to class/division, specific ID, TIH classifi-
cation and associated initial response actions, or other categories.
2. Determine levels and patterns of observed placards (by hazmat grouping) for each truck type.
This may range from a general estimate of placarded truck traffic in the entire jurisdiction to
levels of placarded truck traffic by time for specific locations.
3. Determine truck traffic levels and patterns by type and configuration. This may range from
estimates of truck traffic in the entire jurisdiction to levels of truck traffic by time for specific
locations.
4. Proportions of hazmat placards observed (by grouping) to truck traffic (by type and config-
uration) may be calculated.
5. Present the information in lists, tables, charts, or maps, as applicable.
Application Example
A local entity is interested in estimating the percentage of all trucks transporting Hazard Class 3
materials over a particular Interstate highway segment during the daytime, as well as variations
in traffic levels. The local entity collects information for truck type, configuration and UN/NA
placards for a daytime, weekday 8-hour count on an Interstate segment. The LEPC assumes that
daytime, weekday traffic patterns are consistent with the observed time-period, and summarizes
the information as listed in Table K.1.
Table K.1. Example summary of truck type, configuration, and
UN/NA placard information.
Location: Roadway Segment Description
Date: June 20, 2011
Time: 8:00 A.M. to 4:00 P.M.
Trucks Observed with
Truck All Trucks
Truck Type Class 3 Other Total
Configuration Observed
Placards Placards Placards
Straight 10 10 20 50
Tank Tractor-Trailer 74 56 130 250
Subtotal 84 66 150 300
Straight 0 1 1 400
Box Van Tractor-Trailer 7 22 29 1,600
Subtotal 7 23 30 2,000
Straight 0 0 0 200
Refrigerated Van Tractor-Trailer 0 1 1 1,000
Subtotal 0 1 1 1,200
Straight 1 8 9 200
Flatbed Tractor-Trailer 0 6 6 500
Subtotal 1 14 15 700
Straight 0 1 1 200
Other Tractor-Trailer 0 3 3 600
Subtotal 0 4 4 800
Total 92 108 200 5,000
OCR for page 171
Existing and New HMCFS Data Analysis Examples K-11
Table K.2. Example summary of percentage trucks with UN/NA placards,
by truck type and configuration.
Location: Roadway Segment Description
Date: June 20, 2011
Time: 8:00 A.M. to 4:00 P.M.
Truck % Placarded Trucks with % All Trucks with
Truck Type Class 3 Class 3
Configuration
Placard Other Placard Placard Any Placard
Straight 50% 50% 20% 40%
Tank Tractor-Trailer 57% 43% 30% 52.0%
Subtotal 56% 44% 28% 50%
Straight 0% 100% 0.0% 0.3%
Box Van Tractor-Trailer 24% 76% 0.4% 1.8%
Subtotal 23% 77% 0.4% 1.5%
Straight -- -- 0% 0%
Refrigerated Van Tractor-Trailer 0% 100% 0.0% 0.1%
Subtotal 0% 100% 0.0% 0.1%
Straight 11% 89% 0.5% 4.5%
Flatbed Tractor-Trailer 0% 100% 0.0% 1.2%
Subtotal 7% 93% 0.1% 2.1%
Straight 0% 100% 0.0% 0.5%
Other Tractor-Trailer 0% 100% 0.0% 0.5%
Subtotal 0% 100% 0.0% 0.5%
Total 46.0% 54.0% 1.8% 4.0%
Table K.2 summarizes the proportions of hazmat trucks and proportions of all trucks with a
Hazard Class 3 placard and other placards.
Table K.3 summarizes the hourly 90 percent confidence intervals using statistics (see Section
K.10) for proportions of placarded trucks versus all trucks.
As shown, these estimates have a moderate degree of variability. They are based on locally rel-
evant survey data, but the sample was over a limited time period. They may be off by a moder-
ate degree, but appear to suggest that some differences in hazmat traffic patterns exist, if they
follow the same pattern. They also provide information about the type of hazmat transport haz-
ards that may be expected, and when risk is greatest. Follow-on survey data may provide further
information about the validity of the information.
Table K.3. Example summary of percentage trucks with UN/NA
placards, including confidence intervals.
No. Trucks Observed % Trucks with Hazmat Placard
with Placards 90% Confidence Intervals
Hour of Day
With Hazmat Mean
Total Lower Upper
Placard
8 A.M. 25 500 5.0% 3.62% 6.86%
9 A.M. 25 650 3.8% 2.78% 5.29%
10 A.M. 20 550 3.6% 2.53% 5.19%
11 A.M. 40 700 5.7% 4.43% 7.34%
12 P.M. 25 550 4.5% 3.29% 6.24%
1 P.M. 25 800 3.1% 2.26% 4.31%
2 P.M. 20 650 3.1% 2.14% 4.40%
3 P.M. 20 600 3.3% 2.32% 4.76%
Total 200 5,000 4.0% 3.6% 4.5%
OCR for page 172
K-12 Guidebook for Conducting Local Hazardous Materials Commodity Flow Studies
K.10 Comments on Statistical Analyses
The 1995 Guidance (1) includes a discussion of statistical considerations for traffic count data,
including flows that vary randomly, or in daily, weekly, or seasonal patterns. A table is provided
in the Guidance for confidence intervals based on a Poisson distribution, which can be used for
calculating probabilities of discreet event data such as truck counts. This is not the only distri-
bution that is applicable for count information. For example, the data in Table K.3 were evalu-
ated using a binomial distribution modified for extreme proportions (below 0.1 and above 0.9).
Other analyses might include regression models.
The research team is not minimizing the technical expertise of local entities in their primary
fields, but the fact is that most (e.g., LEPCs) do not have actively involved personnel who are
well versed in transportation statistical methods. The team suggests that LEPCs and other local
entities conducting an HMCFS at objective levels where statistical considerations are impor-
tant (see Appendix D) seek the advice of transportation professionals who are trained in these
analyses. Individuals with this sort of expertise can often be found at universities, local (e.g.,
MPO), state, or federal agencies, or consulting firms. A number of potential statistical methods
may be applied and these can be found in statistics and transportation engineering textbooks
or other sources.
K.11 Interviews with Hazmat Shippers, Receivers,
and Carriers
Description
Interviews with hazmat shippers, receivers, and carriers, as well as with emergency responders
and managers, and other key informants, are discussed in Chapter 5. Limited information from
interviews can be used to confirm hazmat presence and help define priority sampling locations
and frameworks.
Limitations
Unless many interviews are conducted, it is unlikely that sufficient information will be obtained
using this method to develop reliable estimates of hazmat transportation over roadway network
segments.
How to Use the Data
1. For each interview, list the date, time, and identity of the individual, along with a description
of information relevant to the HMCFS project.
2. Compile the interview results in lists or paragraphs.
K.12 Shipping Manifests (Origin/Destination)
Description
This is the most resource-intensive new data collection method described in this guidebook for
an HMCFS. An examination of shipping manifests can be used to confirm hazmat presence, help
define priority sampling locations and frameworks, and provide information about the percent-
age of non-placarded shipments that are carrying hazmat.
OCR for page 173
Existing and New HMCFS Data Analysis Examples K-13
Limitations
As with interviews with shippers, receivers, and carriers, a great deal of shipping manifest
information is needed to develop reliable estimates of hazmat transportation over local roadway
networks, and full use of information obtained from shipping manifests requires advanced trans-
portation modeling techniques.
How to Use the Data
1. For each manifest examined, list the date, time, carrier, hazmat commodity and quantities,
along with a description of information relevant to the HMCFS project provided by the carrier,
including their origin and destination, routes taken, and the ultimate origin and destination of
the shipment, if known.
2. Compile the results in tables, and summarize data accordingly.
K.13 Hypothetical Application of HMCFS Data
To illustrate these applications, consider the hypothetical case study of "Center County LEPC."
Sometown, Texas, is the main city in Center County. Sometown is approximately 30 miles from
Megacity and has an Interstate highway that runs through it. The county has a history of agricul-
tural production and is the location for an industrial facility that uses and ships hazardous
materials, and it has a small crude petroleum processing facility. Sometown is a demographically
young and growing community with a small paid fire department and a mostly unincorporated
surrounding area that is served by volunteer fire departments (VFDs). It has been several years or
longer since most of the VFDs have conducted any hazmat training or reviewed their standard
operating guidelines for hazmat response. The last time mutual aid agreements or emergency
response service incident command procedures were reviewed for any department in the county
was in 2003, and the county population has grown by 50 percent since then.
Although the Center County LEPC is interested in hazmat transport throughout the county,
they are particularly interested in a stretch of Interstate highway east of Sometown that has the
industrial facility on one side and subdivisions on the other side, including a large elementary
school. Center County LEPC decides to conduct a hazmat CFS mostly to help them define train-
ing needs but possibly other applications as well. The LEPC wants to better understand the vari-
ability underlying the collected data and understand whether hazmat transportation patterns may
vary by time of day. One of the LEPC members knows a faculty member from Megacity Univer-
sity who lives in Center County and agrees to assist with statistically evaluating the data, where
needed, as part of a class project.
Assume that the analysis examples given in this appendix apply to the Interstate segment of inter-
est to Center County LEPC, and that the LEPC might have obtained information about hazmat
transport over the segment by any one of those methods. Using information from Sections K.1,
K.2, K.3, or K.4, the LEPC might be able to raise awareness of local officials about the potential
magnitude of the problem, or identify that a large number of Class 3 hazmat trucks may be going
through their community and plan for training accordingly. Beyond that, however, few conclu-
sions can be drawn. Using the information from Sections K.5 or K.6, the LEPC has better informa-
tion about the types of incidents that can be expected, and although some estimates of the
magnitude of potential exposure improve, the reliability of conclusions is still lower.
Using information from Sections K.7 or K.8, the LEPC can start to get a better handle on the
type, magnitude, and source of potential exposures, although additional data would be advised.
Using information from Section K.9 improves on this even further by providing information
OCR for page 174
K-14 Guidebook for Conducting Local Hazardous Materials Commodity Flow Studies
about when potential exposures might occur. Not only does the LEPC have better information
about hazmat transport over the segment, but the locally relevant evidence provides justification
if the LEPC needs to request modifications to practices or allocation of additional resources from
other local, state, or federal agencies.
For example, by examining the statistical variability of the data (confidence intervals), it
appears that the proportion of truck traffic carrying hazardous materials during the late morn-
ing period (11 A.M.12 P.M.) over the segment may be significantly higher than the early after-
noon period (1 P.M.3 P.M.). This information is not conclusive since the intervals identify the
likely range of hourly hazmat truck traffic averages at a 90 percent level of confidence. But it
appears to make sense since the shipping manager of the industrial facility near the segment was
interviewed (Section K.11) and indicated they do most of their shipments in the late morning.
Occasionally, some of those shipments are Class 2.3 gases by large flatbed truck. Although traf-
fic during the 8 A.M. to 9 A.M. period has a high average as well, it is not statistically different from
any other time period. Say, for example, the elementary school sends half-day students home at
11:30 A.M., and those buses use the roadway segment of concern (it is the shortest, most direct
route). The LEPC, the school district, and the industrial facility may want to consider whether
there are alternate routing options for buses or tractor-trailers, even if these routes are less direct.
The community and school may also wish to review building air infiltration rates and shelter-
in-place, evacuation, and emergency notification systems to ensure that protocols and proce-
dures reflect potential hazards.
