Relative probability of Greater Sage-grouse winter occurrence within central region of Wyoming, USA at the combined landscape-patch scale (Filename: wysg_lp_central_winter_rsf.tif).
Dates
Start Date
1994
End Date
2010
Summary
Habitat selection studies can make important contributions to habitat prioritization efforts for species of conservation concern. We present a large-scale collaborative effort to develop habitat selection models for Greater Sage-grouse (Centrocercus urophasianus) across large landscapes (Wyoming, USA) and multiple seasons. Greater Sage-grouse are limited to western semi-arid landscapes in North America, range-wide population declines have been documented, and the species is currently listed a “warranted but precluded” from listing under the U.S. Endangered Species Act. Wyoming is predicted to remain a stronghold for Sage-grouse populations and contains approximately 37% of the remaining birds. We developed Resource Selection Function [...]
Summary
Habitat selection studies can make important contributions to habitat prioritization efforts for species of conservation concern. We present a large-scale collaborative effort to develop habitat selection models for Greater Sage-grouse (Centrocercus urophasianus) across large landscapes (Wyoming, USA) and multiple seasons. Greater Sage-grouse are limited to western semi-arid landscapes in North America, range-wide population declines have been documented, and the species is currently listed a “warranted but precluded” from listing under the U.S. Endangered Species Act. Wyoming is predicted to remain a stronghold for Sage-grouse populations and contains approximately 37% of the remaining birds. We developed Resource Selection Function (RSF) models to characterize and spatially predict seasonal habitat use. The models presented here can help identify important seasonal habitats for greater Sage-grouse across Wyoming. We compiled species data from multiple unique radiotelemetry studies in Wyoming (contributors from these local studies are described in the data lineage) and habitat data from high-quality, biologically relevant, Geographic Information System (GIS) layers across Wyoming. We developed habitat selection models for Greater Sage-grouse across Wyoming for 3 distinct life stages (hereafter referred to as seasons): 1) nesting, 2) summer/late brood-rearing, and 3) winter. We also developed patch and landscape models across 4 different extents, producing statewide models and 3 different regional models: 1) Southwest, 2) Central, and 3) Northeast Wyoming. Sage-grouse habitat selection in Wyoming varied spatially across regions and temporally across seasons. Preferred habitat components in each suite of seasonal models generally agreed with the extensive literature on Sage-grouse seasonal habitat requirements. We chose RSF thresholds for each model set that identified potentially important seasonal habitats for Sage-grouse. The collection of models presented here represents large scale and large extent resource management planning tools that are a significant advancement on previous large-scale tools in terms of both spatial and temporal resolution.We present multi-scaled models for each Sage-grouse season. Because Sage-grouse select biologically relevant landscape characteristics across multiple spatial scales, we therefore included characteristics such as vegetation, topography, anthropogenic, and hydrological variables at five different spatial extents. To create these different scale effects, we used neighborhood statistics (a.k.a. moving window) on the spatial data (e.g., a rectangular 3x3, or circular 564 m radius kernel, is passed over a dataset and a summary statistic is derived for each cell based on neighboring cells within the kernel). These different scales permit us to capture the effects of biotic, abiotic, and anthropogenic relationships that may influence the processes reflecting habitat selection. The patch scale included the two smallest window sizes that summarized predictor variables using radii of 0.045 km (0.006 km2) and 0.564 km (1 km2). The landscape scale included the three largest moving windows which summarized predictor variables at the following radii: 1.5 km (7.07 km2), 3.2 km (32.17 km2), and 6.44 km (i.e., 4 miles, 138.67 km2). Additional details about incorporating multi-scale predictors can be found in Fedy and others (2014).The RSF habitat surface represented by this dataset was developed from the combined landscape and patch RSF models. These data represents the Central Wyoming regional model for the winter life stage.Caution should be exercised in the interpretation of these data. We used resource selection functions (RSF) to produce the habitat selection probability surfaces. These probability values can range between 0 and 1. However, these values represent the relative importance of habitat selection and therefore RSF values cannot be compared across models and they should not be considered as a probability (percentage) for Sage-grouse selection. To make these continuous RSF surfaces more useful, we are also providing datasets that have been reclassified into several bins of habitat importance. However, using these data in novel areas (locations falling outside of the seasonal study site extents) requires careful scrutiny and knowledge of the species’ habitat for the region. We strongly recommend careful review of the manuscript that describes the generation of the data and their application.All references within this metadata record are provided in the supplemental and lineage section.
These data have been created in support of research and analysis of Wyoming Greater Sage-grouse (Centrocercus urophasianus). These data can be used to examine the relative probabilities of Sage-grouse habitat selection within Wyoming. Data users should be aware that models perform best within the study site extents where data were used for model development. The study site extents define the spatial extent of Sage-grouse use locations and available resources for a particular season, which were defined based on seasonal movements (Fedy and others 2012). Please refer to the manuscript (Fedy and others 2014) for extensive assessment and discussion of the application and performance of the RSF models in novel areas outside of the study site extents. The study site extents for each season are provided as different dataset products and therefore data users will require the study site extents, habitat models, and the regional boundary dataset to correctly use the models we developed. We are providing data for relative probabilities of Sage-grouse habitat selection (continuous surfaces), binned categories of relative use (categorical surfaces), and binary surfaces (habitat versus non-habitat) for each season (nesting, summer/late brood-rearing, and winter). These products are modeled at the statewide scale and for three different regional scales (southwest, central, and northeast). Each regional model is applied to the statewide extent, which will permit data users to investigate the use of one regional model applied to a different region. However, we caution data users about applying one regional model to assess habitat importance within a different region (Fedy and others 2014). The data user will select the season of interest and then evaluate the statewide and regional models before determining which model is best suited for their assessment. Auxiliary data should likely be used while selecting the appropriate regional or statewide model (Fedy and others 2014). For example, data users might use local Sage-grouse telemetry data (data not used to develop these models), lek data, or expert knowledge for model selection.
