TY - JOUR
T1 - Empirical bayes approach for estimating urban deer-vehicle crashes using police and maintenance records
AU - Gkritza, Konstantina
AU - Souleyrette, Reginald R.
AU - Baird, Michael J.
AU - Danielson, Brent J.
PY - 2014
Y1 - 2014
N2 - Deer-vehicle crashes are a growing problem in Iowa where deer-vehicle crashes represent 13% of all crashes reported. In 2009, these crashes resulted in nine fatalities and 451 injuries in the state. Deer-vehicle crashes are a problem even in urban areas of Iowa. It is known that deer-vehicle crashes are typically underreported. To address this underreporting, deer carcass salvage reports may be used to augment deer-vehicle crash reports. The objective of this paper is to exploit two sources of deer-vehicle crash data using a statistically reliable assessment methodology, empirical Bayes, to assess the potential for safety improvement of 150 urban highway sections in Iowa. Reconciliation of records to reduce double counting is discussed and a negative binomial regression model of deer-vehicle crash frequency as a function of roadway and environmental factors is estimated. The 25 most promising segments for deer-vehicle crash countermeasures are identified, mostly located on high-speed roadways, roadway segments with gravel right shoulders, and segments adjacent to grasslands. The methodology facilitates the identification of locations for countermeasure implementation as well as monitoring deer-vehicle crash trends.
AB - Deer-vehicle crashes are a growing problem in Iowa where deer-vehicle crashes represent 13% of all crashes reported. In 2009, these crashes resulted in nine fatalities and 451 injuries in the state. Deer-vehicle crashes are a problem even in urban areas of Iowa. It is known that deer-vehicle crashes are typically underreported. To address this underreporting, deer carcass salvage reports may be used to augment deer-vehicle crash reports. The objective of this paper is to exploit two sources of deer-vehicle crash data using a statistically reliable assessment methodology, empirical Bayes, to assess the potential for safety improvement of 150 urban highway sections in Iowa. Reconciliation of records to reduce double counting is discussed and a negative binomial regression model of deer-vehicle crash frequency as a function of roadway and environmental factors is estimated. The 25 most promising segments for deer-vehicle crash countermeasures are identified, mostly located on high-speed roadways, roadway segments with gravel right shoulders, and segments adjacent to grasslands. The methodology facilitates the identification of locations for countermeasure implementation as well as monitoring deer-vehicle crash trends.
KW - Deer carcass data
KW - Deer-vehicle crashes
KW - Empirical bayes
KW - Ranking
KW - Urban areas
UR - http://www.scopus.com/inward/record.url?scp=84892841363&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84892841363&partnerID=8YFLogxK
U2 - 10.1061/(ASCE)TE.1943-5436.0000629
DO - 10.1061/(ASCE)TE.1943-5436.0000629
M3 - Article
AN - SCOPUS:84892841363
SN - 0733-947X
VL - 140
JO - Journal of Transportation Engineering
JF - Journal of Transportation Engineering
IS - 2
M1 - 04013002
ER -