Cover Image

Not for Sale

View/Hide Left Panel
Click for next page ( 32

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement

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 31
31 for and impacts of WVC countermeasures. This aspect of were developed for, and apply to, reported large-animal NCHRP 25-27 was conducted to investigate the hypothesis WVCs (as distinguished from data reported only as carcass that roadside carcass removal data not only indicate a different removal) and, with a view to the application of the models, magnitude for the WVC problem, but may also show differ- for use only with those variables for which data are readily ent spatial patterns than reported WVC data. The choice of available within the typical DOT safety databases. Because the database (collisions or carcasses) used to evaluate the animal exposure data (a measure of the numbers of animals WVC problem, therefore, may lead to the identification of involved in WVC that are near the road, and the amount of different hotspot locations and ultimately different counter- time they spend near the road over the course of a specific measure improvements. Patterns were examined visually by measured time unit) are not among these readily available GIS plots, and by the development of comparable negative variables, this approach will result in some unexplained binomial WVC and deer carcass removal models. WVC and variation in the dependent variable. The safety model inputs deer carcass removal data were obtained from the Iowa are limited to roadway (between shoulder edges) variables Department of Transportation (IaDOT). because few DOT databases include roadside information The creation and analysis of GIS-based data that include the (e.g., guardrail, roadside sight distance) or adjacent landscape attributes and location of roadway segment cross sections, (off right-of-way) characteristics. Even so, it is still necessary reported WVCs, and deer carcass removals can be used to an- to estimate models for lower levels of data availability that swer a number of questions: may exist in some jurisdictions. The result is three funda- mental levels of SPFs: Is the number of reported deer carcass removals different than the reported number of WVCs statewide and along Level 1: These SPFs include only the length and annual individual roadway segments? average daily traffic volume (AADT) of a segment. Are different "high collision" segments identified when Level 2: These SPFs require that segments be classified as reported WVCs and deer carcass removal data are used for flat, rolling, or mountainous terrain and also use the length the safety analysis of individual roadway segments? In and AADT of a segment. other words, do they have different occurrence patterns? Level 3: These SPFs include additional roadway variables Are there any apparent relationships between traffic flow, such as average lane width in addition to the Level 2 variables. roadway cross section characteristics, and reported WVCs? Are these relationships, if they exist, similar for deer carcass The SPFs can be used in a number of applications: removal data? Application A: SPFs can be used with caution to identify The activities completed as part of this aspect of NCHRP 25- roadway factors associated with a high propensity for 27 (e.g., plots, summary measures, and models) were used to wildlifevehicle collisions. These cautions pertain to possibly investigate and compare the patterns of two databases (i.e., counterintuitive inferences that may result from omitted, in- reported WVCs and deer carcass removals) that have been correctly specified, or correlated factors. This application can used to define and mitigate the WVC problem. be useful in roadway design and planning decisions that have implications for wildlifevehicle collisions. Application B: SPFs can be used in the identification of Research Approach: Methods and Data roadway segments that may be good candidates for wildlife The research approach emerged from a review of the exist- vehicle collision countermeasures. ing literature, specifically from a consideration of the gaps in Application C: SPFs can be used in estimating the effec- existing knowledge. tiveness of potential countermeasures that are considered for candidate segments. Application D: SPFs can be used in evaluating the effective- Methods ness of implemented countermeasures using state-of-the-art Aspect 1: Application of reported wildlifevehicle colli- methods for observational before-after studies.114 sion data. Predictive models for wildlifevehicle collisions (commonly called "safety performance functions" [SPFs]) For the last three applications, which are key elements in are crucial to state-of-the-art methods for filling safety analy- this project, collision history often is used as a predictor. sis gaps and developing the requisite guidelines for mitigat- However, it is now well recognized as a poor predictor be- ing these collisions. These models are derived from historical cause collision history tends to be short term (<3 years) rather data and relate collision frequency to physical roadway and than long term (3 years) and therefore subject to random roadside characteristics and to measures of exposure. They fluctuation and associated vagaries of regression to the mean.

OCR for page 31
32 The result is that for Application A, resources are often summary tables from these comparison activities are sum- wasted on safer sites that are wrongly identified and good marized in the "Findings and Results" section. Similar to As- candidates may be ignored. As a result, the countermeasure pect 1, negative binomial prediction models were also used. effectiveness estimates for Applications C and D can be exag- WVC and deer carcass removal prediction models (or SPFs) gerated. The regression to the mean problem cannot be that considered traffic flow and roadway cross section ele- overemphasized and so is illustrated in Appendix D. ments as potential input variables were created and com- While the SPF can provide less biased predictions than the pared. The results of these activities are also described in the collision count for Applications B, C, and D, estimates ob- "Findings and Results" section. tained from these models can have a high variance because of Several types of computer software were used to overlay, the inability to include potentially important explanatory present, and summarize the WVC and deer carcass removal variables in them. In recognition of this difficulty and the data within the GIS platform. Microsoft ExcelTM and True- problems with estimates from collision counts, an empirical BasicTM were used to manage the deer carcass removal data. Bayes (EB) procedure has been used.177 This procedure in The ArcGIS 9.1TM platform was used to present and analyze essence takes a weighted average of the two estimates, recog- the collision and carcass datasets spatially. ArcCatalogTM was nizing that both provide important clues as to a location's used as a file management program and applied specifically safety. In effect, by using the collision counts to refine the SPF for organizing spatial data. Most of the mapping activities prediction, the EB procedure accounts for factors, such as off took place in ArcMapTM. ArcGuideboxTM was used for some right-of-way characteristics and animal exposure, that affect of the more complicated spatial analysis, and the large size of wildlifevehicle collision frequency but are not in the model. the roadway inventory database files required the use of For example, a location that has more animal movements FileMakerTM. The modeling of the WVC and deer carcass than the "average" location, but that is similar in the charac- removal information was completed with SASTM statistical teristics of the prediction model, will tend to have more col- software. lisions than the "average" location. With EB refinement comes higher collision prediction accuracy. The EB proce- dure is illustrated by way of example applications, in the "In- Data terpretation, Appraisals, and Applications" section. Aspect 1: Application of reported wildlifevehicle collision The development of the SPFs involved determination of data. The models for predicting the frequency of reported which explanatory variables should be used, if and how vari- wildlifevehicle collisions were developed for rural two-lane ables should be grouped, and how variables should enter into and rural multilane roadways using Highway Safety Informa- the model (i.e., the best model form). Consistent with the com- tion System (HSIS) data from California, North Carolina, mon research practice in developing these models, generalized Utah, and Washington and for rural freeway roadways with linear modeling was used to estimate model coefficients, data from California, Utah, and Washington. These are the assuming a negative binomial error distribution. In specifying typical classifications used by DOTs in other aspects of safety a negative binomial error structure, the dispersion parameter management. Tables 6 through 9 summarize the data used. k, which relates the mean and variance of the regression esti- mate, is estimated from the model and the data. The value of Aspect 2: Comparison of wildlifevehicle collision and k is such that, the smaller its value, the better a model is for the carcass removal data. Three different databases were used set of data (See Appendix B). Conveniently, the dispersion to compare the magnitude and patterns of WVCs and deer parameter estimated in the SPF calibration is used to derive carcass removals in Iowa. First, 10 years of police-reported the weights for the two sets of information used in the EB WVC information in a GIS-acceptable format were acquired procedure. from the IaDOT. The data included the location of the WVCs and information provided on the police reports (e.g., sever- Aspect 2: Comparison of wildlifevehicle collision and ity, surface conditions, time of day, and age of driver). A large carcass removal data. The tasks completed for this research majority of the reported WVCs involved white-tailed deer were done to evaluate the value of collecting and plotting (Odocoileus virginianus). The reported WVCs in 2001, 2002, WVC and deer carcass removal data by location, and to test and 2003 were used in this analysis. The individual WVC the straw hypothesis that these two datasets may also identify locations were provided by the IaDOT and plotted by latitude different roadway locations for potential WVC countermea- and longitude coordinates. For example, the 2002 WVC sures. The magnitude and patterns of location-based WVC locations plotted on a roadway map of Iowa within a GIS reports and deer carcass removal datasets in Iowa were com- platform are shown in Figure 4. pared qualitatively through visual GIS plots and quantita- The two other datasets that were used included informa- tively (e.g., WVC frequency per mile). The GIS plots and tion about deer carcass removals and roadway cross sections.

OCR for page 31
33 Table 6. Data summary for rural two-lane roadways. Data Length (mi.) AADT Total Crashes/mile-year State period Total mean min max mean min max Crashes mean min max 1991- CA 8,349 0.644 0.001 26.137 4893 63 37041 5,378 0.068 0.000 16.670 2002 1990- NC 25,165 1.322 0.010 18.980 2466 2 80428 59,280 0.140 0.000 8.330 2001 1985- UT 9,260 2.503 0.010 40.380 1541 1 17424 15,334 0.186 0.000 6.250 2000 1993- WA 5,362 0.601 0.010 28.660 4334 87 23917 1,746 0.078 0.000 12.500 1996 Table 7. Data summary for rural multilane roadways. Data Length (mi.) AADT Total Crashes/mile-year State Period Total Mean min max mean min max Crashes mean min max 1991- CA 994 0.359 0.003 7.689 14312 304 78300 1,205 0.116 0.000 4.900 2002 1990- NC 1,185 0.803 0.010 9.440 11134 100 63332 5,406 0.347 0.000 8.330 2001 1985- UT 291 0.599 0.010 4.840 6162 186 61393 4,021 0.654 0.000 6.430 2000 1993- WA 322 0.423 0.010 63.440 12588 172 54274 251 0.218 0.000 12.500 1996 Table 8. Data summary for rural freeways. Years of Length (mi.) AADT Total Crashes/mile-year State Data Total Mean min max Mean Min max Crashes mean min max 1991- CA 1,659 0.536 0.001 14.917 22520 3275 86700 1,326 0.089 0.000 9.260 2002 1985- UT 700 1.928 0.010 13.730 10579 2776 64402 5,145 0.608 0.000 7.290 2000 1993- WA 400 0.685 0.010 8.320 18179 4124 49952 257 0.194 0.000 25.000 1996 Table 9. Variables available for modeling. State Roadway Variables State Roadway Variables AADT AADT Design speed in mph Average degree of curvature Divided/undivided Design speed in mph Lane width in feet Lane width in feet Shoulder width in feet Median type Median barrier type Median width in feet CA UT Median width in feet Number of lanes Number of lanes Paved roadway width in feet Surface type Percentage truck traffic Surface width in feet Shoulder type Terrain (level, rolling, mountainous) Speed limit in mph Terrain (level, rolling, mountainous) AADT AADT Shoulder type Average degree of curvature Shoulder width in feet Shoulder width in feet Median type Shoulder type Median width in feet Median barrier type NC Number of lanes WA Median width in feet Speed limit in mph Median type Surface width in feet Number of lanes Terrain (level, rolling, mountainous) Speed limit in mph Surface width in feet Terrain (level, rolling, mountainous)

OCR for page 31
34 (a) (b) Figure 4. Deer carcass removal (top) and individual WVC locations (bottom) in Iowa (2002). For this project it was only possible to plot the locations of the and shoulder width) for each roadway segment within Iowa deer carcass removals by IaDOT personnel to the nearest mile also were used. marker (Figure 4). The gender of the deer removed was also Figure 4 provides an example of the data from 2002. These noted if possible. Annual average daily volume estimates and data were compared visually and quantitatively on a cross section information (e.g., surface width, median type, statewide and sample corridor basis. The impact of the dif-