Site occupancy analysis of the sagebrush lizard (Sceloporus graciosus) in the disappearing desert-scrub of the Columbia Basin

The shrub-steppe landscape of the Columbia Basin has been the target of agricultural and urban development, and the resulting fragmentation and degradation has led to the disappearance of unique arid mesohabitats. In central Washington, the sandy lowland habitats resembling desert-scrub, which under natural conditions are characterized by a shrub-and-sand mosaic, have become increasingly degraded by humans directly (e.g., flooding, agriculture) or indirectly (e.g. cheatgrass introduction). These habitats have unique community assemblages with species adapted to the sandy substrates and unobstructed matrix between shrubs. While much conservation literature focuses on the loss of shrubsteppe habitat, there has been little research on the effects of loss, fragmentation, and degradation of sandy desert-scrub in the Columbia Basin on small reptile inhabitants. The sagebrush lizard (Sceloporus graciosus), an obligate resident of these sandy habitats, serves as a potential model organism to document the effects of desert-scrub loss and degradation. Compared to historical sightings prior to 1980, scant observations of S. graciosus within the last decade may be a sign that this species might be disappearing from its Washington range. In this study, I aim to assess whether the sagebrush lizard is still present in much of its original range in Washington, and to determine which aspects of their habitat correlates with population density. Sandy desert-scrub habitats that are historically known to have or were capable of having sagebrush lizards in Washington and Oregon were chosen as field sites during the summer and early autumn of 2016. Landcover statistics, such as shrub cover and grass cover, and habitat characteristics including wildfire history were determined for use as potential correlates for the rate of sagebrush lizards encountered in each site. Landcover was quantified via supervised classification using aerial photographs, and model accuracy was verified by comparing transect samples from the field. Multidimensional scaling and analyses of similarity were performed to determine which sites formed groups that were dissimilar to others, and similarity percentages were used to quantify the degree to which each habitat factor contributed to site dissimilarity. In only 6 out of the 16 sites were sagebrush lizards detected during standard search surveys. In Washington, of the 7 surveyed sites with sagebrush lizard sightings within the past 30 years, only 2 (29%) had detectable individuals. There were no differences between sites where sagebrush lizards were present and absent unless sites within 100 m of a crop field were classified separately, whereupon all five of these sites then fell within one out of two groups with 60% similarity. Percent grass cover was the strongest factor contributing to group differences, with less grass and more open sand found in sagebrush lizard-present sites. Also, sagebrush lizard-absent sites had histories of wildfire. The need to separately classify otherwise suitable habitat near agriculture (and did not have sagebrush lizards) indicates additional negative influences affecting this species presence, since not incorporating this sole variable masks any differences among sites. In general, I infer that wildfire, cheatgrass invasion, and landscape development all have reduced suitable habitat for sagebrush lizards and likely explains their absence in their historic range. Given that the remaining pristine sandy desert-scrub habitats are now fragmented and few in number in the state of Washington, I infer from my results that there is very real cause for concern for the viability of this species in the state. If the challenge this species is facing is representative of other arid endemics, then the widespread changes to the shrub-steppe and desert-scrub landscapes in the Western USA provide many research opportunities in conservation ecology.