Grasslands in California's inland Mediterranean climate zone vary greatly over time and space, largely due to fluctuating rainfall and heterogeneous geology, topography, and soils. In light of a dramatic invasion of exotic species into these grasslands, conservation management goals typically include the preservation and enhancement of native vegetation. Developing specific management targets to achieve these goals, however, is complicated by uncertainty about pre-invasion conditions and the spatial and temporal complexity of the system.
The Tejon Ranch, the largest, contiguous, privately-owned property in the state, supports 44,000 ha of California's inland Mediterranean grassland. The mission of the Tejon Ranch Conservancy (Conservancy) is to preserve, enhance, and restore the native biodiversity of Tejon Ranch. In 2009, the Conservancy partnered with the University of California Range Ecology Lab led by James Bartolome to describe the grasslands on the ranch, in order to build the understanding required for management planning. As a doctoral researcher and co-leader of the study from 2010 to 2014, my objective was to develop a scientific framework to inform reliable predictions about the distribution of plant species over space and time in the ranch's grasslands. I investigated three basic questions:
1. Does species composition correlate with geologic, topographic, and edaphic landscape composition, at differing spatial scales?
2. What are the drivers of inter-annual community change at the ecological site scale?
3. What are the alternative states of the ecological sites?
Chapter 1 is a description of how I investigated the first question in the western Mojave Desert landscape of the ranch. I collected topographic, edaphic, and ground cover data at 35 small (0.25 m2) plots across a 64 km2 (6.4 x 107 m2) extent in the spring of 2010. Fortuitously, the timing and amount of rainfall in 2009-2010 resulted in high diversity and abundance of native annual forbs and grasses across the landscape. I encountered 46 species; all were annuals except for three perennial bunchgrass species in 9 plots. I found that elevation, soil cation exchange capacity, soil silt percentage, and soil total nitrogen explained 40% of the spatial variation of the 25 species encountered at more than one plot. I identified nine distinct species assemblages, four with plant cover dominated by native annual species, and five with plant cover dominated by exotic annual species. The five exotic assemblages were constrained to two geographic areas of the landscape. Both areas contained sediments from degraded dolomitic roof pendants of the Tehachapi Mountains and featured soils high in clay and nitrogen. I hypothesized that atmospheric nitrogen deposition is preferentially increasing soil nitrogen in these zones, promoting persistent exotic species dominance there. Overall, results suggested that species composition did in fact correlate with landscape composition, as perceived at the scales of geology type and landform, and measured at the smaller scale of the plot. I conclude by recommending that restoration management planning incorporate considerations of biogeochemical nitrogen cycling.
In Chapter 2, I explain how the Range Ecology Lab and Tejon Ranch Conservancy investigated my first and second research questions in the San Joaquin Valley landscape of Tejon Ranch. Using thirty-five 3024 m2 study plots across a 294 km2 extent, we discovered that species distribution was more strongly correlated with geology, slope, and elevation than with USDA Major Land Resource Area or soil map unit. Accordingly, we used geology, slope class, and elevation class to divide the grasslands into 52 ecological sites (with 14 ecological sites representing physical conditions of 85% of the grassland area). With a focus on four geographically extensive ecological sites, I demonstrate how we verified the accuracy of our ecological site model, identified community types (i.e., community phases) at the ecological sites, and determined the drivers of community phase shifts between years at the ecological sites (i.e., community pathways). The Pleistocene terraces and Mafic bedrock slopes ecological sites each supported a single community phase in all three years. In contrast, at both the Holocene flats and the Miocene hills ecological sites, phase shifts were observed from 2010 to 2011, and again from 2011 to 2012. These inter-annual shifts in community phases were driven more strongly by variation in rainfall than by rodent bioturbation or livestock grazing. At both ecological sites, October and November rainfall exceeding 2 cm was a prerequisite for community phases dominated by exotic annual grasses, whereas less precipitation in those months promoted community phases with a higher relative abundance of native annual forbs. A concurrent wildlife study on the Holocene flats ecological site revealed that community phases with dense exotic annual grasses are unsuitable for a suite of special-status vertebrates. The Conservancy is using fall rainfall exceeding 2 cm as a cue to suppress exotic annual grass biomass using cattle grazing, in order to enhance conditions for native annual forbs and wildlife. This management prescription would not have been possible without organizing the landscape into ecological sites and tracking the grassland community on those ecological sites across multiple years.
In Chapter 3 I describe how we explored my third research question. We used state-and-transition models to catalogue our understanding about historical and contemporary states at the four focal ecological sites in the San Joaquin Valley landscape of Tejon Ranch. Landscape reconnaissance revealed that all but the Pleistocene terraces currently support an alternative state in addition to the herbaceous state measured in 2010-2012 - one with greater relative cover of perennial grasses, shrubs, and/or oaks. Using soil phytoliths and historical accounts, we developed hypotheses about historical states (circa 1772) for the four ecological sites. We propose that the Holocene flats ecological site likely supported a native forbland with an open canopy of Atriplex sp., the Mafic bedrock slopes ecological site likely supported a forbland-oak savanna matrix, and the Pleistocene terraces and Miocene hills ecological sites likely supported grasslands dominated by perennial grasses. Using the state-and-transition models as guides, I describe potential restoration management for each ecological site. The Conservancy is currently spending limited restoration funds to enhance conditions for native annual forbs and wildlife in the current annual grassland state of the Holocene flats. However, significant recovery of native biodiversity appears possible on the Pleistocene terraces and Mafic bedrock slopes ecological sites, if the Conservancy elects to spend the time and money required to cross the thresholds from the current annual grassland states to alternative states on those ecological sites.