The KDE+ software: a tool for effective identification and ranking of animal-vehicle collision hotspots along networks
- 698 Downloads
Objective identification of locations on transportation networks, where animal-vehicle collisions (AVC) occur more frequently than expected (hotspots), is an important step for the effective application of mitigation measures.
We introduce the KDE+ software which is a programmed version of the KDE+ method for effective identification of traffic accident hotspots. The software can be used in order to analyze animal-vehicle collision data.
The KDE+ method is based on principles of Kernel Density Estimation (KDE). The symbol ‘+’ indicates that the method allows for the objective selection of significant clusters and for the ranking of the hotspots. It is also simultaneously applicable to an unlimited number of road segments.
We applied the KDE+ method to the entire Czech road network. The hotspots were ranked according to their significance. The resulting hotspots represent a short overall road length which should require a more detailed assessment in the field. The 100 most important clusters of AVC represent, for example, only 19.7 km of the entire road network (37,469 km).
We present an objective method for hotspots identification which can be used for AVC data. This method is unique because it determines the significance level of hotspots in an objective way. The prioritization of hotspots allows a transportation manager to effectively allocate resources to a feasible number of identified hotspots. We describe the software, data preparation and present the KDE+ application to AVC data.
KeywordsKernel density estimation Spatial analysis Animal-vehicle collisions Roadkill Roads Network Hotspots GIS The Czech Republic
This work was financed by the Transport R&D Centre (OP R&D for Innovation No. CZ.1.05/2.1.00/03.0064). We would like to thank Andreas Seiler and the anonymous reviewer whose comments substantially increased the quality and lucidity of the text. We would also like to thank Jan Tecl for his help with information on accuracy of traffic-accidents data from around the world.
- Gunson KE, Teixeira FZ (2015) Identifying the patterns and processes of wildlife road interactions are important to inform road-wildlife mitigation planning. In: van der Ree R, Smith DJ, Grilo C (eds) Handbook of Road Ecology. John Wiley & Sons, Oxford, pp 101–110Google Scholar
- Hauer E (1997) Observational before-after studies in road safety. Pergamon Press, OxfordGoogle Scholar
- Niemi M, Nykänen A, Rita H, Vastaranta M, Väänänen VM (2014) Two spatial scales of moose-vehicle collisions. In: IENE 2014 International Conference on Ecology and Transportation, Programme and Abstracts. Seiler, A. (ed). 2014, Malmö, Sweden; publisher: IENE. p 74. Session ID: 3AGoogle Scholar
- Okabe A, Yamada I (2001) The K-function method on a network and its computational implementation. Geogr Anal 33(3):152–175Google Scholar
- Poledník L, Poledníková K, Roche M, Hájková P, Toman A, Culková M, Hlaváč V, Beran V, Nová P, Marhoul P (2005) Záchranný program–program péče pro vydru říční (Lutra lutra) v České republice v letech 2006–2015. AOPK PrahaGoogle Scholar
- Seiler A, Helldin J-O, Seiler C (2004) Road mortality in Swedish mammals: results of a drivers’ questionnaire. Wildl Biol 10:225–233Google Scholar
- Spellerberg IF (1988) Ecological effects of roads and traffic: a literature review. Global Ecol Biogeogr Lett 7:317–333Google Scholar
- Větrovcová J, Poledníková K, Poledník L, Beran V, Hlaváč V (2010) Databáze údajů o uhynulých jedincích vydry říční v ČR. Ochrana přírody 4:15–19Google Scholar