Sheep breeds in contrasting environments - pasture utilization at different spatiotemporal scales

In Norway, approximately 2 million sheep are released onto outer fields for summer grazing. Indeed, during the summer months sheep are the most common large herbivores in most Norwegian mountain areas. The sheep, of which 80% are the heavy, docile Norwegian White sheep (NWS) and 13% short tailed, light-footed, agile and more gregarious Spælsau (SP), are released onto a range of summer grazing environments; from coastal to mountainous inland areas. However, little is known about phenotypic plasticity in ranging behavior of sheep, how different sheep breeds interact with different rangeland habitats and how they are able to adjust their foraging behavior at different spatiotemporal scales. Indeed, these complex behavioral processes are hierarchical, as large herbivores operate on several spatial and temporal scales to maximize nutritional acquisition and hence fitness. This is imperative for understanding the grazing dynamic of sheep and for a sustainable grazing management adapted to the available natural and ever-changing resources. In this study, 51 ewes of the two aforementioned breeds were fitted with GPS collars in two contrasting environments, during the summer grazing seasons 2013 and 2014. The two study areas were vegetation mapped and classified regarding their quality for sheep grazing (Spekedalen; poor, and Bratthøa; rich pasture). The collars logged position every hour and were equipped with sensors that recorded activity during time-to-fix (TTF). I defined several temporal scales from the whole grazing season to bounding individual time steps. I defined three spatial scales; 95, 50 and 20% utilization distribution areas (UDs). The spatial scales for all but the time-step temporal scales were extracted from the dynamic Brownian Bridge Movement Model (dBBMM). The time-step time-indexed UDs were based on Dynamic Brownian Motion variation extracted from every two consecutive timestamped relocations. I calculated the proportion of three sheep grazing quality classes: “Less Good”, “Good” and “Very Good”, within each UD and at all spatial scales. Using general mixed models I was thus able to infer area use across spatiotemporal scales, when both active and inactive, (papers I and II), habitat use across spatiotemporal scales and habitat selection, when active, (paper III) for both breeds in both environments, using a resource selection function (RSF) approach. I found that both breeds had non-significant larger seasonal UD area in poor Spekedalen as compared to rich Bratthøa. At finer temporal scales, 95% UD differences were found between the two study areas. However, these differences could be confounded by the higher density of salt blocks in Bratthøa, at least at large scale. SP had larger UD area than NWS at all but the finest temporal scale. The effect of breed and environment on sheep area use was stronger at the 50% UD as compared to the 95% UD scales, at all temporal scales. Although the two breeds may differ in their area use, I was not able to detect differences in habitat use and selection, in neither time nor space. However, effects of vegetation class and study area on habitat use and selection were found. My findings highlight the increased importance of the scarce “Very Good” patches with finer scales, in poor Spekedalen. Indeed, this high-quality class is probably more important for nutrient extraction and acquisition than the use indicates. The lack of environment by breed interactions in area and habitat use suggests that the two breeds respond equally to range quality, at all spatiotemporal scales. I appreciate that, in my study, the large individual variation may have overshadowed the effects of the intrinsic and extrinsic determinants. I conclude that scale has to be considered when comparing pasture utilization across spatial and temporal scales in contrasting environments and between sheep genotypes. My findings are thus important for management of grazing resources in multipurpose land use planning