Habitat Assessment of the Rare and At-Risk Species Umtanum Desert Buckwheat (Eriogonum codium) for Increased Successes with Assisted Migration

Abstract

“Edaphic islands” are comprised of unique or challenging soil habitats and host 15-20% of global floral taxa including numerous rare and endangered species. These communities in general, and small populations of edaphic endemic species in particular, are at increasing risk from climate change related stress, threats from invasion by generalist non-native species, and exposure to novel anthropogenic disturbances, including altered fire intervals. The rare and endangered Umtanum Desert Buckwheat (Eriogonum codium) is a slow growing, long-lived perennial plant highly restricted to a 1.9 acre discontinuous band in the shrub-steppe of Hanford Reach Nuclear Facility (Department of Defense). Little is known about the habitat characteristics of its single population or whether its extremely limited distribution is due to the presence of unique soil properties at this location. Previous work has theorized E. codium may require specific edaphic conditions and/or low inter-species competition. The single population is in decline and outplanting efforts undertaken at nearby locations with similar site and soil conditions have largely failed. Assisted migration could enable otherwise impossible translocations of this species outside its native site, greatly reducing its risk of decline, genetic loss, and extinction. In 2020, the Washington Natural Heritage Program (WNHP) developed a habitat suitability model for E. codium which identified 115 acres of likely habitat, and seedlings were planted at five sites across two locations in 2020 and 2021. The intent of this study is to inform management strategies for the assisted migration of E. codium and determine the effects of outplanting locations and site selection on outplanted seedling success. To do so, this study characterizes E. codium’s current habitat and compares these habitat characteristics to past, current, and potential future assisted migration locations at two different spatial scales. In Chapter 1, site, soil, and vegetation habitat characteristics were measured at the single location where E. codium occurs (Native Site). Differences in E. codium site occupancy, abundance, and between the five remaining subpopulations and sampling transects was used to characterize E. codium’s associated habitat type and responses to a range of biotic and abiotic habitat variables. Inter-species competition was found to be an important factor shaping E. codium habitat. Site and soil characteristics, such as heat load, rock and bareground cover, and the abundance of fine and very fine gravel grades, were found to interact and form spatially heterogenous microhabitats and vegetation community dynamics directly impacting E. codium’s growth and survival. Cluster analysis of vegetation functional group composition identified six distinct vegetation habitat types delineating three perennial graminoid community types, each associated with either high, medium, or low abundance of E. codium. Site, soil and vegetation characteristics were compared between these “optimal” E. codium habitat groups to elucidate E. codium’s ability to occupy a range of microhabitat conditions. These results are explored as a means to theorize on E. codium’s evolutionary ecology and the role of microhabitat differences on its current confinement to the edge of Umtanum Ridge. In Chapter 2, site, soil and vegetation habitat characteristics were measured at past, present, and potential assisted migration locations, and compared to characteristics at the Native Site. Badger Mountain is considered a highly suitable past assisted migration location where seedlings were outplanted in 2020. Snow Mountain is considered a moderately suitable assisted migration location where seedlings were outplanted at three sites in 2020 and 2021, with additional trails ongoing at the time of this study. Gingko Petrified Forest State Park is considered a highly suitable assisted migration location, although no seedlings have been outplanted there to date. Site-level habitat characteristics at Gingko State Park and Snow Mountain were compared to those at the Native Site to evaluate the accuracy of the WNHP habitat model at the spatial scale utilized by managers when selecting assisted migration sites. Gingko State Park shared more habitat similarities with the Native Site than Snow Mountain did, consistent with the output from the WNHP habitat suitability model. However, similarity to the Native Site was quite different between the three Snow Mountain sites which was not reflected in the model. These results suggest that the habitat model may not be useful at selecting assisted migration sites at spatial scales finer than landscape level, or distinguishing highly suitable and moderately suitable habitats when they co-occur within a given location. Overlaps between the range of values at the Native Site and at assisted migration sites indicate the existence of potentially suitable microhabitats at fine spatial scales. Targeting these microhabitats for seedling establishment may be important for assisted migration. To assess this, differences in site, soil and vegetation characteristics around outplanted seedlings (transect-scale) were related to seedling survival rates and to habitat characteristics at the Native Site (high, medium, and low optimal habitat groups from Chapter 1). Overall seedling survival was very low, possibly due to differences in soil characteristics from the Native Site, and/or increased inter-species competition with forbs, annual graminoids, and especially other woody shrubs including more widely adapted Eriogonum species. However, heterogeneity in microhabitat niches at very fine spatial scales may have allowed for very limited seedling establishment (14% at Badger Mountain, 3% at Snow Mountain) and should continue to be the focus of future studies.