Global temperatures are increasing at higher rates in the Arctic region, and these changes are affecting timing and availability of vegetation. Understanding the relationship between plants and their consumers is therefore necessary to gain a greater understanding into how climate change will affect herbivores in northern latitudes. We examined the relationship of plant productivity, measured by the normalized difference vegetation index (NDVI), to juvenile body mass in 57 populations of Rangifer tarandus managed by Sami people throughout Norway using generalized additive models. Optimal climate windows for NDVI were calculated at both the regional and national level. Since the populations cover a large area, spatial effects on juvenile body mass and NDVI were also considered using the related varying coefficient model structure. A model including covariates decomposed into spatial, temporal, and residual components was also considered.
Results/Conclusions
Models with NDVI averaged using regional climate windows were more effective at prediction and fit the data better than the national climate window. The decomposed model provided the best fit and lowest prediction errors, which indicates that the decomposed model is not just a useful effects analysis tool. Calculating the proportions of deviance explained for decomposed covariates adds to the usefulness of this model as a tool for analyzing spatial and temporal effects of covariates separately.