Habitat selection in moose and roe deer : a third order comparison

Sammanfattning: Ungulates are important animals in Swedish culture and economy, yet they are considered to cause considerable damage on forests to the disadvantage for the forest industry. At the same time, the forest industry is one of the reasons for the increased carrying capacity of moose and roe deer. Different factors affect the behaviours that cause the animals to inflict damage on the forest, but a main reason is access to browsing. The selection of habitat occurs on different spatial scales and this work looked closer on the selection at the 3ʳᵈ order, selection of habitat within the home range. The data that was used came from GPS collars from 7 female moose and 17 female roe deer that were tagged within the area of Nordmaling in northern Sweden as well as from camera traps within this area. The GPS data was analysed using Brownian bridge kernel method to receive the utility distribution and home range of the animals. Thereafter the data was connected to the ground cover map allowing a comparison of the usage of different types of habitat using a logistic regression. The GPS data was analysed for the full year, but also sectioned into the periods when the animals were expected to fawn, the rutting period, winter and the full year but excluding fawning, rutting and winter. The results showed for high usage of decidious trees for the full year in both moose and roe deer, while both groups seemed to avoid water. Moose also selected for clear-cuts throughout the year with an increase during winter. Urban areas showed for a higher usage than were expected for by roe deer. Rutting season increased the usage of arable land in moose while fawning season increased the usage of clear-cut areas by roe deer. The camera trapping data was also connected to the ground cover map and thereafter analysed with a Poisson regression. These results confirmed the high usage of decidious forest but contained high standard errors, likely as a result of low amounts of data. They also covered less habitat types as an effect of the sampling method. To benefit the forest industry, further studies would be needed on how the animals behave within the different habitats. However, the access to forage seemed to be a driving factor in the habitat selection with the exclusion of the rutting season in moose and the fawning season in roe deer.