Cochise County and the City of Sierra Vista travel demand model development

              Cochise County & City of Sierra Vista
Travel Demand Model Development
June 18, 2010
Cochise County & City of Sierra Vista Travel Demand Model Development Page 2
Table of Figures
Figure 1: Historic HPMS Daily VMT Chart ............................................................................. 7
Figure 2: Number of Lanes .......................................................................................................... 8
Figure 3: Area Types .................................................................................................................... 9
Figure 4: Functional Class ......................................................................................................... 10
Figure 5: Input Network Speeds ................................................................................................ 12
Figure 6: Hourly LOS E Capacity ............................................................................................. 13
Figure 7: Other Speed & Capacity Stratification Variations ................................................. 14
Figure 8: Cochise County TAZ Structure ................................................................................ 15
Figure 9: Census Designated Places .......................................................................................... 17
Figure 10: 2007 Occupied Dwelling Units ............................................................................... 18
Figure 11: 2020 Occupied Dwelling Units ............................................................................... 19
Figure 12: 2040 Occupied Dwelling Units ............................................................................... 20
Figure 13: 2007 Total Employment .......................................................................................... 21
Figure 14: 2020 Total Employment ........................................................................................... 22
Figure 15: 2040 Total Employment .......................................................................................... 23
Figure 16: 2007 Counted Volume Flows .................................................................................. 28
Figure 17: 2007 Modeled Volume Flows .................................................................................. 29
Figure 18: Screenlines and Cutlines ......................................................................................... 31
Figure 19: Volume to Count Ratios .......................................................................................... 34
Figure 20: 2020 Modeled Volume Flows .................................................................................. 36
Figure 21: 2040 Modeled Volume Flows .................................................................................. 39
Cochise County & City of Sierra Vista Travel Demand Model Development Page 3
Table of Tables
Table 1: Data Fields for the Shapefile Network ......................................................................... 5
Table 2: Data Fields for the QRS-II Network ............................................................................ 6
Table 3: Historic HPMS Daily VMT .......................................................................................... 7
Table 4: Functional Class Codes ............................................................................................... 11
Table 5: Network Speed Stratification ...................................................................................... 11
Table 6: Network Capacity Stratification................................................................................. 13
Table 7: TAZ Structure by Place ............................................................................................. 16
Table 8: 2007 External Station Volumes .................................................................................. 25
Table 9: 2020 External Station Volumes .................................................................................. 26
Table 10: 2040 External Station Volumes ................................................................................ 26
Table 11: Estimated Accuracy of Model Parameters ............................................................ 27
Table 12: Model Validation Targets ........................................................................................ 27
Table 13: VMT Match by Area Type ........................................................................................ 30
Table 14: Screenline Results ...................................................................................................... 31
Table 15: Cutline Results .......................................................................................................... 32
Table 16: Cutline Results by VMT Range ............................................................................... 32
Table 17: Model Results by Functional Class ......................................................................... 33
Table 18: Model Results by Counted Volume Range ............................................................. 33
Table 19: 2020 VMT by Area Type .......................................................................................... 37
Table 20: 2020 VMT by Functional Class ............................................................................... 37
Table 21: 2020 VMT by Volume Range ................................................................................... 38
Table 22: 2040 VMT by Area Type .......................................................................................... 40
Table 23: 2040 VMT by Functional Class ............................................................................... 40
Table 24: 2040 VMT by Volume Range ................................................................................... 41
Cochise County & City of Sierra Vista Travel Demand Model Development Page 4
Context of the Project
Following the Arizona Department of Transportation’s Task Assignment MPD 12-09, Wilbur
Smith Associates has been charged with developing a combined travel demand model for
Cochise County and the City of Sierra Vista. The model was specified as being developed in the
QRS-II modeling platform. The route for developing the model included Network and Traffic
Analysis Zone (TAZ) development in TransCAD and ArcGIS shapefile formats for ease of
display and analysis, with all modeling work being performed in QRS-II. The final model files
were delivered in both QRS-II and in ArcGIS shapefile formats in June 2010. Following the
development of the 2007 base year model, forecast TAZ data was input to develop models for
the years 2020 and 2040.
The Wilbur Smith Associates model development team was lead by project manager, Dale
Miller, PE. Tom Cooney served as the modeling task leader and developed much of the model
input data. Charlie Sullivan performed the modeling work. Although not formally part of the
modeling team, Karen Lamberton of Cochise County and Jeff Pregler of the City of Sierra Vista
contributed to the process extensively, and verified the base year and developed the forecast
demographic data for the entire County and for the City of Sierra Vista, respectively. Training
and technical assistance was provided by Dr. Alan J. Horowitz of AJH Associates, the developer
of the QRS-II platform.
The QRS-II model file is a network with associated nodes. Network data is provided in the line
layer, while TAZ data for the model is contained in the nodes.
The network file for the model contains descriptive attributes and output modeled volumes.
Autos and trucks are modeled separately, and their output volumes are reported separately in the
network.
The TAZ data consists of the inputs which are used by the model to calculate the demand for
trips. The data is aggregated to the TAZ level, and includes average autos per household, total
occupied dwelling units, and employment in the three categories of basic, retail, and service. It
should be noted that employment data was originally provided in the two categories of retail and
non-retail; but the data was broken down into additional detail in order to use more precise
attraction rates per employment category.
Cochise County & City of Sierra Vista Travel Demand Model Development Page 5
Base Year 2007 Model Network
Model network development began by extracting Cochise County highway network information
from the ADOT statewide travel demand model and Cochise County and applicable city GIS
files. Industry-standard GIS layers, street maps, aerial photos, and other databases were all used
in network development. The complete network was designed to capture both regionally
significant and local level roads in order to support realistic model loading patterns.
A full range of network attributes were developed to support both the needs of the travel demand
model and the requirements of post-modeling display and analysis. The characteristics of the
QRS-II platform specified for the Cochise County and City of Sierra Vista model impose
limitations on the number and name of network data fields. Therefore, the model networks were
ported into the ArcGIS shapefile format in order to provide for additional network fields.
The more extensive number of network fields, populated in the ArcGIS shapefile format, are
shown in Table 1.
Table 1: Data Fields for the Shapefile Network
Cochise County & City of Sierra Vista Travel Demand Model Development Page 6
The QRS-II platform’s network fields, which are designed to be used only for direct modeling
needs, are noted in Table 2.
Table 2: Data Fields for the QRS-II Network
Two important considerations should be noted when working within the QRS-II modeling
platform and extracting or updating attributes to ArcGIS. The first point is that QRS-II does not
retain link ID’s when importing. Rather, links are renumbered with a Name attribute based on
their geographic order. All files used in QRS-II must list the links in the specified order.
Second, QRS-II orders link variables by defined link types. Three link types are defined in the
Cochise County and City of Sierra Vista model: 1-way street, 2-way street, and centroid
connector. As shown in Table 2, the reference code for extracting or updating a given attribute
varies depending on its link type. For example, the code to extract or update the link name and
assigned volume from the QRS-II network would be Bi for 1-way streets, Bst for 2-way streets,
and B89 for centroid connectors.
Ground traffic counts for area roadways were provided for the years 1998 through 2009. Most
of the years had just a small number of counts focused on area collectors. The data for 2007 was
both more comprehensive and more thorough, with a total of 1,239 counts provided on roads of
all functional class. Only the 2007 counts were coded into the network.
Truck percentages from the file SHSkdtFactors 2007-2008 ver2.xls were used to factor total
counts to develop counts for combination trucks. While these are factors rather than actual
counts, and while the factors were constant over relatively large distances of roadway, they
provided a useful and valuable measure to validate the truck model.
Cochise County & City of Sierra Vista Travel Demand Model Development Page 7
Complementing the ground count data, county-level HPMS VMT for the years 1998 – 2008 was
available from the ADOT website to provide another measure of traffic activity. Table 3 shows
the data for each year for Cochise County, along with the population trend data.
Table 3: Historic HPMS Daily VMT
The same data is presented in a chart in Figure 1. It is interesting to note that on-system VMT is
fairly stable throughout the period. Off-system VMT shows a slight but steady increasing trend.
Figure 1: Historic HPMS Daily VMT Chart
0
1,000
2,000
3,000
4,000
5,000
6,000
1998
1999
2000
2001
2002
2004
2005
2006
2007
2008
HPMS Daily VMT Trends
All Roads
On System
Off System
Cochise County & City of Sierra Vista Travel Demand Model Development Page 8
Color-coded views of several network data fields are shown in Figures 2 through 6, with
associated explanatory text and tables.
The number of lanes, shown in Figure 2, is based on the original street file and was verified
using aerial photography. Links on I-10 are detail coded with separate lines for each direction;
ramps and any frontage roads are included. It should be noted that there are some 2-way, 2-lane
ramps along I-10, and also in the Bisbee area.
Figure 2: Number of Lanes
Cochise County & City of Sierra Vista Travel Demand Model Development Page 9
Area Types were defined for the Cochise County and City of Sierra Vista model based on a
subjective evaluation of combined employment and dwelling unit density. Three defined area
types were used to stratify the assignment of network speeds and capacities to links, and to
stratify trip attraction rates. Area Type 1 represents the more dense urban areas such as portions
of Sierra Vista, Bisbee, and Douglas. Smaller urbanized areas such as Benson, Bowie, etc. were
classed as Area Type 2. Finally, the rural areas of the county were classified as Area Type 3.
The general effect of the area types is to define higher network speeds, lower network capacities,
and lower trip attraction rates for the more rural areas, while the more urbanized areas feature
lower speeds and capacities and higher attraction rates. Area Types for the Cochise County and
City of Sierra Vista model are shown in Figure 3.
Figure 3: Area Types
Cochise County & City of Sierra Vista Travel Demand Model Development Page 10
Functional classifications, shown in Figure 4, were defined and assigned based on the original
street file, standard network coding practice, and on the requirements of the QRS-II program.
Only a specific number of functional classifications are available within the QRS-II modeling
platform. In general, the response to this was to implement area types, with stratification of
functional class and other attributes replacing the original separate HPMS urban and rural coding
scheme.
Figure 4: Functional Class
It should be noted that the same functional classes are described by QRS-II with different
internal codes. The standard link functional classifications and the equivalent QRS-II codes are
shown in Table 4. The standard link functional classifications are used in the ArcGIS shapefile
format network, while the QRS-II codes are used in developing its Approach Codes to control
intersection capacity, throughput, and delay. The standard classifications are the numbers zero
through 8. The QRS-II codes are case-sensitive letters. Note that the Centroid Connector code is
Cochise County & City of Sierra Vista Travel Demand Model Development Page 11
the capital letter “O”, not the number zero, and that the code for a local street is a lower-case
letter “L”, rather than the number 1.
Table 4: Functional Class Codes
The scheme for estimating network speeds by Functional Class and Area Type is shown in Table
5. Two additional variables, the presence of continuous left turn lanes and the status of the road
as paved or unpaved, also have an effect on the designated input speed. It should be noted that
not all possible combinations are present for each stratification variable. For example, the
Interstate Functional Class is present only in defined rural area type, no links in the rural area
type have continuous left turn lanes, and there are no unpaved roads in the dense urban area type.
Table 5: Network Speed Stratification
Cochise County & City of Sierra Vista Travel Demand Model Development Page 12
The input network speeds are shown in Figure 5. These speeds represent the average operational
speed, and are different from the posted speed limit. The purpose of the stratification is to allow
the model to replicate the relative differences in observed driver behavior on different functional
classes of roads under different conditions.
Figure 5: Input Network Speeds
The stratified capacities of roadway links by Functional Class, Area Type, and other attributes
are shown in Table 6. To meet the requirements of the QRS-II model platform, the data is for
hourly capacity, one-way, per lane, at Level of Service (LOS) E. LOS E is considered the
saturation flow rate.
Cochise County & City of Sierra Vista Travel Demand Model Development Page 13
Table 6: Network Capacity Stratification
Stratified model total link capacities are shown in Figure 6.
Figure 6: Hourly LOS E Capacity
Cochise County & City of Sierra Vista Travel Demand Model Development Page 14
The two additional variations which affect link speeds and capacities are the presence of
continuous left turn lanes and the status of the road as paved or unpaved. These variables are
shown together in Figure 7. Note that continuous left turn lanes are present only in Sierra Vista
and Douglas. Unpaved roads are found throughout the county, and some roads have both paved
and unpaved segments.
Figure 7: Other Speed & Capacity Stratification Variations
Cochise County & City of Sierra Vista Travel Demand Model Development Page 15
Forecast Years 2020 and 2040 Model Networks
In the initial stage of development of the Cochise County and City of Sierra Vista model, no
future year network improvements have been defined. Planners have taken the eminently
reasonable approach of using the model to forecast and analyze network deficiencies, and to use
those results to generate potential network projects for further testing. It is expected that a
number of alternative projects will be developed using this process, and that they will be tested
in the model at a later time by County and City staff (or through a qualified consultant).
Cochise County TAZ Data
The TAZ structure developed for the Cochise County model features a total of 799 zones. Zone
size and density varies throughout the county, with the more dense urbanized areas having a finer
zone structure than the rural parts of the county. The TAZ structure is shown in Figure 8.
Figure 8: Cochise County TAZ Structure
Cochise County & City of Sierra Vista Travel Demand Model Development Page 16
Seven urbanized areas were identified as places in the original TAZ file; and these designations
have been retained. These seven places have sometimes included several places in one. For
example, the Sierra Vista place includes Sierra Vista, Huachuca City, Fort Huachuca, Sierra
Vista Southeast, Hereford, Palominas, and Miracle Valley. The Elfrida place includes Elfrida
and McNeal. In contrast, some smaller places, such as Tombstone, Gleeson, and Kansas
Settlement, have not been specifically marked. In general, the seven designated places have a
denser zone structure than the remainder of the county.
Table 7 shows the distribution of zones by place. Sierra Vista has the most zones with 304.
There are three zones on the western edge of I-10 which are within Pima County. Four zones on
the eastern side of the county are partially or totally in Hidalgo County, New Mexico, and
include the community of Rodeo. There are nine external stations, including two ports of entry
with Mexico at Naco and at Douglas.
Table 7: TAZ Structure by Place
Cochise County & City of Sierra Vista Travel Demand Model Development Page 17
Complementing the definition of places in the Cochise County model, Figure 9 shows the
Census Designated Places (CDP) for 2000. There are a total of 34 CDP’s in the county (plus one
for Rodeo, New Mexico).
Figure 9: Census Designated Places
Population
QRS-II trip modeling is based on households rather than population. As a result, occupied
dwelling units (DU’s) were the basis for the model and were calculated in the demographics
files, rather than population. DU’s for 2007, 2020, and 2040 are shown in Figures 10 to 12. For
display purposes, Fort Huachuca in TAZ 18 is shown with over 1,000 DU’s. However, in the
model, it was actually treated as a special generator.
Cochise County & City of Sierra Vista Travel Demand Model Development Page 18
Figure 10: 2007 Occupied Dwelling Units
Occupied DU’s for the 2007 model are shown in Figure 10. A total of 302 zones had no DU’s in
2007. There were 474 zones, or 60% of the total, with less than 500 DU’s. Only 14 zones had
500 or more DU’s in 2007; 13 of these were in Sierra Vista. The other higher-population zone is
TAZ 845 in Benson, south of 4th Street, between Ocotillo and Highway 80.
Cochise County & City of Sierra Vista Travel Demand Model Development Page 19
Figure 11: 2020 Occupied Dwelling Units
The forecast year 2020 DU’s are shown in Figure 11. The number of zones with no DU’s
dropped to 110, as the population is anticipated to spread throughout the county. The populated
zones with less than 500 DU’s increased to 653, now making up 83% of the total. The 27 zones
with 500 or more DU’s now include zones in Sierra Vista, Huachuca City, Whetstone, Benson,
Mescal, Pearce, and Bisbee.
Cochise County & City of Sierra Vista Travel Demand Model Development Page 20
Figure 12: 2040 Occupied Dwelling Units
Figure 12 shows the 2040 DU’s. Population is expected to continue to increase and to spread
throughout the county for the forecast year 2040. The number of zones with no DU’s dropped to
102. The populated zones with less than 500 DU’s decreased slightly to 649 as more zones
moved into higher population categories, but these are still 82% of the total. There are 39 zones
with 500 or more DU’s in 2040, which now include zones in Sierra Vista, Huachuca City,
Whetstone, Benson, Pomerene, Mescal, Pearce, Bisbee, and Douglas.
Cochise County & City of Sierra Vista Travel Demand Model Development Page 21
Employment
Employment is the main variable driving the trip attraction rates per zone. In contrast to
population (represented by dwelling units), which increased and showed more spreading through
the county, the increase in employment was more concentrated in the existing urbanized areas.
Much of the employment increase was concentrated in the Sierra Vista / Fort Huachuca area.
Figure 13: 2007 Total Employment
Total employment in 2007 was zero for 266 zones, as shown in Figure 13. A further 426 zones
have less than 100 employees. Another 82 zones have up to 500 employees, and 10 have
between 500 and 1,000. These 10 zones are in Benson, Sierra Vista, Bisbee, Douglas, and at the
Bisbee-Douglas International Airport. There are 6 zones in Benson and Sierra Vista with over
1,000 employees in 2007, including Fort Huachuca.
Cochise County & City of Sierra Vista Travel Demand Model Development Page 22
Figure 14: 2020 Total Employment
For the forecast year 2020, the number of zones with no employment dropped slightly to 260.
The next category of 0 to 100 zones also dropped slightly to 408, while 93 zones had up to 500
employees. The number of zones with between 500 and 1,000 employees increased to 18 and
now include TAZ’s in Willcox, Pomerene, Benson, Huachuca City, Sierra Vista, Bisbee, and
Douglas. The 11 zones with more than 1,000 total employees are in Benson and Sierra Vista.
Cochise County & City of Sierra Vista Travel Demand Model Development Page 23
Figure 15: 2040 Total Employment
For the forecast year 2040, the number of zones with no employment dropped slightly to 247.
The next category of 0 to 100 zones also dropped to 382. The zones with up to 500 employees
again showed an increase, growing to 118. The number of zones with between 500 and 1,000
employees increased to 26 and include TAZ’s in Willcox, Pomerene, Benson, Huachuca City,
Sierra Vista, Sierra Vista Southeast, Bisbee, and Douglas. There are 17 zones with more than
1,000 total employees; and again all of these zones are in Benson and Sierra Vista.
Cochise County & City of Sierra Vista Travel Demand Model Development Page 24
Model Design
All model functions for the Cochise County model are performed entirely within the QRS-II
platform. The platform design features extensive reliance on the state-of-the-art data,
algorithms, and procedures, which are embedded within the platform. This feature, according to
the QRS-II Reference Manual, “…permits elimination of unnecessary data and most calibration
exercises.” This feature makes QRS-II uniquely applicable for a range of studies, as it “…can be
used as a sketch planning tool when time for more careful planning is unavailable, or [it] can be
used for rigorous analysis….”
QRS-II was initially based on the parameters and procedures of NCHRP 187: Quick-Response
Urban Travel Estimation Techniques and Transferable Parameters. It has been updated with the
revision of that document to the 1998 NCHRP 365: Travel Estimation Techniques for Urban
Planning. Both reports feature compatibility with the procedures used in the Highway Capacity
Manual.
Trip Generation
Standard QRS-II procedures were used for trip generation. Production rates were modified
during the validation to enable a closer match to observed total VMT. Autos per household were
used as the chief production variable.
Attraction rates were varied by TAZ for the three defined area types. Employment was broken
into the categories of Retail, Non-Retail (basic), and Demographic 4 (service).
Three special generators were defined:
Ft. Huachuca / Sierra Vista Municipal Airport in TAZ 18
Cochise College in TAZ 582
Bisbee – Douglas International Airport in TAZ 596
The predominantly residential zones lying outside Cochise County were considered for
designation as special generators. However, after an examination of available counts in those
areas, and of the preliminary model results, it was determined that the standard trip generation
procedures in the model adequately captured their travel behavior. Therefore, these zones were
modeled using standard trip generation.
For automobile travel, the three standard trip types of Home-Based Work (HBW), Home-Based
Non-Work (HBNW), and Non-Home-Based (NHB) were used. Truck traffic was modeled as a
separate trip type, with trucks defined as multi-unit trucks only.
Cochise County & City of Sierra Vista Travel Demand Model Development Page 25
Trip Distribution
Trip distribution was based on the negative exponential function. The initial average trip lengths
for auto trips were set to the census journey-to-work figure of 21.45 minutes for HBW, with
HBNW and NHB correspondingly set to 7.20 and 6.78 minutes, respectively, using standard
relationships between the three trip types. However, during the validation and examination of
county-wide VMT and trip paths, and consideration of the geography of the county, this lead to
increasing the average trip length for HBNW and NHB to 15.28 and 14.36 minutes. For trucks,
the average trip length was set to 34.22 minutes, based primarily on the length of through trips.
External Stations
There are nine external stations defined for Cochise County, located as shown in Figure 8.
Tables of the total volumes, external-local volumes, and external-through volumes for autos and
trucks for the years 2007, 2020, and 2040 are shown in Tables 8 – 10.
Table 8: 2007 External Station Volumes
Cochise County & City of Sierra Vista Travel Demand Model Development Page 26
Table 9: 2020 External Station Volumes
Table 10: 2040 External Station Volumes
Growth rates for most of the external stations were based on ADOT traffic log AADT
projections. The two exceptions are the I-10 external stations. Their growth was expected to be
higher based on recent development, especially the western station at the Cochise County/ Pima
County border. For these stations, the higher growth rates from the I-10 reliever study of 2.46%
at the western station and 1.56% at the eastern station were used.
Cochise County & City of Sierra Vista Travel Demand Model Development Page 27
Base Year 2007 Traffic Assignment and Validation
Following the guidelines of the Model Validation and Reasonableness Checking Manual, it is
generally accepted that “…there are no absolute measures or thresholds that can be achieved to
declare a travel model or its components ‘validated’. The level of accuracy expected of a model
is somewhat subjective, and ultimately depends on the time and resources available… The
reliability of a model validation effort is always constrained by the quality and quantity of
validation data available.” The manual notes varying confidence levels for various model input
data and output parameters, as shown in Table 11. The overall point of the table is that there is
some level of uncertainty in model input parameters. As a result, the general practice is to be
wary of over-validating a model by too rigorous adjustments to match specified criteria. It is
generally preferable to have correct algorithms and behavioral trends that can be used with
confidence to predict trends in future behavior; instead of mangling the model to match
validation criteria which themselves have a range of error.
With that understanding, the target error ranges for
an urban model validation is noted in the manual
as shown in Table 12. It is generally accepted that
lower-volume facilities are validated to lower
standards. There are two reasons for this. First,
following the concepts of Table 11, there is usually
less confidence in the input data for lower-volume
facilities, so there should be correspondingly less
strict standards for their modeled outputs. Second,
higher-volume facilities in a model draw their trips
from a larger geographic range of zones and of
network paths. This makes them more responsive
to global model parameter adjustments, so that
their trips may be modified without violating the
validity of the modeling of overall traveler
behavior.
Table 11: Estimated Accuracy of Model Parameters
Table 12: Model Validation Targets
Cochise County & City of Sierra Vista Travel Demand Model Development Page 28
This brings up a further point in a model validation. That these accepted ranges are for urban
models with similar travel patterns and parameters within the geographic area. Cochise County
is a larger geographic area with rural zones and a mix of urbanized areas ranging from isolated
groups of houses like Paradise and Portal, to small clusters like Sumizona and St. David, to small
cities like Benson and Bisbee, to full-scale cities like Sierra Vista and its surrounding suburbs.
With the greater variability of the study area, there is a corresponding impact on the potential
accuracy of the model.
The overall flow of year 2007 counted traffic for Cochise County is shown in Figure 16.
Through traffic on I-10 is evident. The county shows its strongest volumes on I-10 from the
Cochise/Pima border to Benson and down Highway 90 to Sierra Vista. Traffic in the southern
part of the county in the vicinity of Douglas and Bisbee is also strong.
Figure 16: 2007 Counted Volume Flows
Cochise County & City of Sierra Vista Travel Demand Model Development Page 29
In comparison, the flows of traffic from the Cochise County and City of Sierra Vista model are
shown in Figure 17. The model has successfully generated the same overall patterns of traffic as
are recorded in the 2007 ground traffic counts. This is only one measure, but it shows that the
general operation of the model is correct.
Figure 17: 2007 Modeled Volume Flows
Cochise County & City of Sierra Vista Travel Demand Model Development Page 30
VMT by Area Type
Three area types were defined for Cochise County. Totaling the Vehicle Miles Travelled (VMT)
for all the links within each area type provides a measure of the model’s fit across various
geographic areas. Area types are shown in Figure 3. Statistics for the total counted VMT, total
modeled VMT, and their match for each Area Type is shown in Table 13.
Table 13: VMT Match by Area Type
Area Type 1 focuses on Sierra Vista, Bisbee, and Douglas. It covers slightly less than 1% of the
total land area and 4.3% of its roadway mileage. The mean value for its match for individual
links is 0.97, with a standard deviation of 0.57. The overall match is 0.87.
Area Type 2 has almost 2.1% of the county’s area and 13.2% of its roadway mileage. Its model
matches are 0.95 overall, with a mean match per link of 0.86 and a standard deviation of 0.82.
Area Type 3 constitutes 96.4% of the area of the county and 82.5% of its roadway mileage. The
match of modeled VMT to counted VMT for individual links averaged 1.22, with a standard
deviation of 1.30. Total modeled VMT to counted VMT is 1.06.
VMT by Screenlines and Cutlines
While the overall statistics for the model show a reasonably tight validation, the results of
screenline and cutline analyses examine the validation in more detail. Screenlines and cutlines
are collections of links used to aggregate the comparisons of modeled results to counts.
Screenlines are typically defined as large lines cutting across significant travel movements in a
study area. They present the big picture of the model. Cutlines are typically smaller, used to
check the validation for smaller and more specific areas.
For the Cochise County and City of Sierra Vista model, eight screenlines were defined, as shown
in Figure 18. Five screenlines capture east-west trips, one was defined to capture north-south
movements through the middle of the county, and two screenlines curve around the Sierra Vista
and the Bisbee-Douglas areas.
Cochise County & City of Sierra Vista Travel Demand Model Development Page 31
Figure 18: Screenlines and Cutlines
The higher-level view of the validation represented by the screenlines in Table 14 show the
reasonableness of the validation. The overall modeled-to-count ratio is an average of averages,
and so is not a rigorously valid statistic, but it is useful for comparison.
Table 14: Screenline Results
Cochise County & City of Sierra Vista Travel Demand Model Development Page 32
Complementing the screenlines, 42 smaller
cutlines were defined throughout the county,
with focus on Douglas and Sierra Vista. The
overall statistic for the 140 observations is
again an average of averages, but its value of
0.93 is generally reasonable and is
comparable to the 0.89 average for the
screenlines. However, a look at the individual
cutlines shows some with much less accurate
matches. Some of this is reasonable. For
example, cutline 23 has 7 observations, yet
the total counted VMT is less than 2,000.
These are low-volume roads with lower
expected accuracy. Table 16 aggregates
cutline matches by counted VMT ranges, and
again shows reasonable matches in the
aggregate values of the larger model.
Table 16: Cutline Results by VMT Range
The overall conclusion is clear that the
validation is less precise in detail than it is in
the larger picture.
Table 15: Cutline Results
Cochise County & City of Sierra Vista Travel Demand Model Development Page 33
VMT by Functional Class
Aggregating model results by functional class provides another view of the validation and a
diagnostic of the sources of the revealed imprecision. Table 17 shows the assigned VMT to
counted VMT ratios and average volumes for five functional classes. (Note: Figure 4 shows the
defined functional classes). It can be seen that the higher-volume facilities validated reasonably
well. The obvious outlier is the minor collector functional class, which includes unpaved roads
and very low volume roads. However, the statistics also show that the Minor Collector
functional class accounts for less than 2% of the total counted VMT.
Table 17: Model Results by Functional Class
VMT by Volume Range
A final aggregate view of the validation in Table 18 shows the relative mileage, VMT, and match
of the model by volume class. The lower match for the two categories of 20,000 – 25,000 and
25,000 – 30,000 are principally due to the functional class 2 roadways (primarily Highways 80,
90, and 92, Buffalo Soldier, and Fry).
Table 18: Model Results by Counted Volume Range
Cochise County & City of Sierra Vista Travel Demand Model Development Page 34
Assigned Volume to Count Ratio by Link
As a final graphic to show the model validation, Figure 19 shows the Volume to Count (V/C)
ratios in three categories to clearly display the geographic distribution of the model’s matches.
Overall, I-10 validated well. The model was somewhat high on the lower-volume roads in the
center of the county, between Elfrida and I-10. The relatively high volume functional class 2
roads in the south and west of the county were generally modeled low. Local-level links within
Sierra Vista and Douglas have mixed results.
Figure 19: Volume to Count Ratios
Cochise County & City of Sierra Vista Travel Demand Model Development Page 35
Overall, the 2007 model validation is not exemplary, but is generally reasonable for a full-county
mixed urban and rural model given the time and the levels of confidence in the input
demographics and counts. Any model could of course benefit from additional work, and this
model is no exception. If additional time and resources became available, the model validation
would benefit most from focusing on improving the following, in descending order of
importance and impact:
• Highways 80 and 90 south of I-10
• Highways 80 and 92 between Douglas and Sierra Vista
• Localized roads in Sierra Vista
• Localized roads in Douglas
• Low-volume roads between Douglas and I-10
• Low-volume minor collectors
Cochise County & City of Sierra Vista Travel Demand Model Development Page 36
2020 Model Results
The overall flow of year 2020 modeled traffic for Cochise County is shown in Figure 20.
Compared to the 2007 modeled flow shown in Figure 17, the same overall patterns of traffic as
are recorded in the 2007 ground counts. This general cross-check shows that the overall
operation of the 2020 forecast model is reasonable.
Figure 20: 2020 Modeled Volume Flows
Cochise County & City of Sierra Vista Travel Demand Model Development Page 37
2020 VMT by Area Type
The modeled VMT for 2020 is shown in Table 19, with a comparison made to the modeled 2007
data. The aggregate growth in VMT is greater in the more urbanized Area Types 1 and 2, as one
would expect. The compound annual growth in VMT is 2.7% for Area Type 1, 2.6% for Area
Type 2, and 1.8% for Area Type 3.
Table 19: 2020 VMT by Area Type
2020 VMT by Functional Class
The 2020 Modeled VMT results by Functional Class are shown in Table 20. The compound
annual growth rate in VMT ranges from 1.3% for Interstates to 3.0% for Minor Collectors. It is
interesting to compare this trend of percent increase (not volume increase) with Figure 1, which
shows HPMS daily VMT trends. In that figure, off-system roads are seen to have a stronger
trend of increase than on-system. The match of the model to the HPMS trends indicate that the
increase in input demographics for the model are reasonable.
Table 20: 2020 VMT by Functional Class
Cochise County & City of Sierra Vista Travel Demand Model Development Page 38
2020 VMT by Volume Range
Table 21 shows the 2020 modeled results by volume range, compared with the 2007 modeled
results. As with the previous tables, the 2020 data shows reasonable increases. Similar to the
higher off-system growth trend shown in Tables 19 and 20, Table 21 shows strong growth in the
lower volume ranges. However, unlike the previous tables, the growth across volume ranges is
more even across categories.
Table 21: 2020 VMT by Volume Range
Cochise County & City of Sierra Vista Travel Demand Model Development Page 39
2040 Model Results
The overall flow of year 2040 modeled traffic for Cochise County, shown in Figure 21, has
similar patterns to the 2007 and 2020 models. The link volumes, however, are noticeably
greater. Strong growth at the Cochise/Pima county external station on I-10 is a significant
contributor to the increased volumes. This general cross-check shows that the general operation
of the 2040 model is reasonable.
Figure 21: 2040 Modeled Volume Flows
Cochise County & City of Sierra Vista Travel Demand Model Development Page 40
2040 VMT by Area Type
Modeled results for 2040 are compared to the 2020 modeled results in Table 22. The trend of
higher growth rates in the less urbanized areas continues, although Area Type 2 exhibits a
relatively higher rate.
Table 22: 2040 VMT by Area Type
2040 VMT by Functional Class
Data for 2040 modeled VMT by functional class is shown in Table 23. Trends are reasonable,
with the compound annual growth rate for the 20-year period ranging from 1.9% for Major
Collectors to 3.2% for Minor Collectors. By the year 2040, VMT on Interstate 10 is modeled to
comprise almost 38% of the total VMT in Cochise County.
Table 23: 2040 VMT by Functional Class
Cochise County & City of Sierra Vista Travel Demand Model Development Page 41
VMT by Volume Range
Fairly even growth, with slightly stronger growth in the lower volume categories, is a feature of
the 2040 modeled VMT by volume ranges compared to the same categories for the 2020 model.
Again, the trends and relationships compared to 2020 seem reasonable.
Table 24: 2040 VMT by Volume Range
End of Model Documentation Report