Forest gradient response in Sierran landscapes: the physical template

Vegetation pattern on landscapes is the manifestation of physical gradients , biotic response to these gradients, and disturbances. Here we focus on th e physical template as it governs the distribution of mixed-conifer forests in California's Sierra Nevada. We extended a forest simulation model to ex amine montane environmental gradients, emphasizing factors affecting the wa ter balance in these summer-dry landscapes. The model simulates the soil mo isture regime in terms of the interaction of water supply and demand: suppl y depends on precipitation and water storage, while evapotranspirational de mand varies with solar radiation and temperature. The forest cover itself c an affect the water balance via canopy interception and evapotranspiration. We simulated Sierran forests as slope facets, defined as gridded stands of homogeneous topographic exposure, and verified simulated gradient response against sample quadrats distributed across Sequoia National Park. We then performed a modified sensitivity analysis of abiotic factors governing the physical gradient. Importantly, the model's sensitivity to temperature, pre cipitation, and soil depth varies considerably over the physical template, particularly relative to elevation. The physical drivers of the water balan ce have characteristic spatial scales that differ by orders of magnitude. A cross large spatial extents, temperature and precipitation as defined by el evation primarily govern the location of the mixed conifer zone. If the ana lysis is constrained to elevations within the mixed-conifer zone, local top ography comes into play as it influences drainage. Soil depth varies consid erably at all measured scales, and is especially dominant at fine (within-s tand) scales. Physical site variables can influence soil moisture deficit e ither by affecting water supply or water demand; these effects have qualita tively different implications for forest response. These results have clear implications about purely inferential approaches to gradient analysis, and bear strongly on our ability to use correlative approaches in assessing th e potential responses of montane forests to anthropogenic climatic change.