Prosjekt Elg - trafikk i Stor-Elvdal 2000-2004. Hvordan unngå elgpåkjørsler på vei og jernbane?

Through the last ten years, two and a half times as many moose have been killed per kilometre on the railway line (0.70) than on the roads (0.29) in Stor-Elvdal, although the daily number of trains is only 12 compared to 2000 cars. Most of the collisions on the railway (95 ) and road (81 ) occur during 5 winter months. The radio collaring study showed that the moose generally migrate from higher elevated areas when snow gets too deep during winter, towards lower areas with less snow cover. Unfortunately, this is where the traffic arteries run. The length of the period with deep snow during winter explains all 85 % of the variation in number of accidents between years to only 54 % of the variation on the road. The moose are hit when they cross the road. During winter there are many moose and many crossings. During summer only few stationary moose cross the road and there are few accidents. In contrast, on the railway only few accidents happen while crossing, most are killed when they stay at the railway. The moose only stay at the railroad during periods with snow. In summer they cross the rails. Efforts against moose � train accidents should not reduce the moose crossings, but reduce the time the moose stay along the line. When we corrected for the length of the snow winters and analysed all the data for a long time period, we detected that the efforts Vegetation removal and Feeding each reduced the number of accidents to almost half (49 and 40 ) of the expected on the stretches where they were applied. The position where the accidents happened seemed more to be determined by coarse landscape qualities than by small scale qualities. The risk of accidents seemed to be large over longer distances. On the road there were some stretches with few accidents. We could not evaluate the effect of vegetation clearing along the road since the clearing started before we got accurate data. Also the experiments showed that both vegetation clearing within the railway fence and artificial feeding reduced the number of accidents to half of what was expected. In the experiment of clearing we found that increased amounts of winter forage outside the railway fence reduced the number of accidents when the forage was removed inside the fence. This again shows that the moose were killed when they stayed at the line, not when they crossed. Both the pellet group count and the radio marked moose showed that most moose keep close to the feeding sites. Also the new feeding sites attracted moose so that the GPS-marked moose were closer than 250 m from the feeding stations 40 percent of the time. Probable due to few GPS-marked moose the first year without feeding, the reduction in winter home range was not statistically significant, but the moose moved less daily when using feeding sites. When the feeding stopped for four days on feeding sites that had been used for years in Imsdalen, the moose became unrestful and the medium distance from the feeding stations increased from 279 to 818 m. One moose moved as far as 1343 m from the feeding site while the longest distance from a moose to a feeding site during feeding was 818 m. A feeding system connected with tracks from an all terrain vehicle down to 300 m from the main road kept the moose mainly away from crossing the road. The accidents on the road were reduced to the half of what was expected, but there was uncertainty in the estimate.