Figure 3. Relationship between environmental predictor variables and characteristics of small rodent population dynamics of Fennoscandian tundra, based on 385 sampling units from 22 locations (Table 1a). The panels for no. of above-zero days in winter and July mean temperature include data form all locations, while the panels for bioclimatic zone include data from the Norwegian locations only (n=17). Proportion of bioclimatic zone refers to the presumed optimal bioclimatic zone of each rodent genus (low alpine zone for voles, middle alpine zone for lemmings). The plots for community contribution include all available sampling units (n= 385 for winter climate and July temperature, 239 for bioclimatic zones). The plots for cycle amplitude include only the sampling units where a given species was present (n=334, 367, and 305 [lemmings, grey-sided voles andMicrotus ] for winter climate and July temperature, n=194, 225, and 181 [lemmings, grey-sided voles and Microtus ] for bioclimatic zones). Width of boxes is proportional to the number of observations. Environmental variables were calculated across a 25 km2 buffer zone, for figures with 9 km2 and 1 km2 buffers see Appendix 1 (Figures A6 and A7). Grey boxes indicate variables that were statistically significant in the best models in either model set (Table 4).

Tables

Table 1. Description of small rodent snap-trapping time-series included in this study. Time-series in part (a) are included in all analyses and time-series in part (b) only in location level analyses (see Appendix 2 for details). For additional information on the complete time-series see Appendix 2. For all time-series in (a) we only included units that were active ≥ 10 years. Time series numbers refers to Figure 1.
(a)