IENE 2016

Avoidance behaviour of otherwise suitable habitats in close proximity to roads has been recorded for brown bears across the globe. In Greece, the current 55km operational part of the newly constructed KA45 highway (attached to the Egnatia Motorway  part of the TENT), cuts through bear habitat in  Kastoria region in an open valley framed by Gramos and Vitsi-Askio  mountains, where  a bear population of 150-220 ind. minimum is present. Despite the presence of 134 mitigation structures (mostly culvert underpasses) incremented at an average distance of 400m over the total length of the highway, since 2009  and up to 2014, 21 bear traffic fatalities occurred along this highway segment mainly due to poor fencing. We used telemetry and remote sensing technology in order to investigate and evaluate the oriented use of these mitigation structures by bears after the installation of a bear-proof fence in 2013-14 which dramatically decreased the number of traffic fatalities. The purpose of this double approach was also to investigate whether the bear-proof fence started functioning as an artificial barrier or as a triggering factor for more systematic use of the mitigation structures by bears. Nineteen adult and sub-adult bears (6 M and 13 F) were fitted with GPS/GSM radiocollars over three different periods: 2011-2012 (before bear proof fence installation) and 2014, 2015 (after bear proof fence installation) yielding 58,373 radiolocations. We analysed the spatial patterns of telemetry data to investigate the use of mitigation structures by radio-tagged bears, given the 30min radiolocation interval of the pre-programmed radio-collars. A simple model where straight lines connect consecutive radiolocations was used to describe the animals’ movement.  Mitigation structure expected attractiveness/suitability was classified according to the Openness Index (OI) and a correlated buffer zone was attributed to each structure, relative to the logarithmic ranked value of the OI. We examined the structure use with the assumption that all highway crossings and fixes (before or after a crossing) of tagged bears which fall in the buffer zone were successful use of the structure.

In order to cross validate the effective use of mitigation structures by tagged and non-tagged bears, a two folded methodology was developed. For tagged bears: their interactions with the mitigating structures were detected using appropriate RFID reader technology attached to mitigation infrastructure, while VHF collars functioned as tags. For non-tagged animals, that are currently the majority of the population, a sensor network based on the prototype EcoPrismaMultiSense and EcoTinySkylog nodes has been deployed on selected mitigating structures. Data from a multitude of analog and digital sensors such as (infrared, weight sensors, cameras) are fused together in order to detect bear presence and distinguish it from the presence of other species.

The overall approach aims at investigating whether this is currently a transitional adaptive phase for bears before they start using mitigation structures regularly or if this is a case of a highway starting functioning as a critical landscape barrier. The results of this study will essentially contribute in further adjustment of mitigation structures to ensure their effectiveness.