Woody Plant Dynamics in a Sonoran Desert Ecosystem across Scales: Remote Sensing and Field Perspectives

Abstract

Historic land uses impose discernable legacy effects that may influence ecosystem function, a concern of particular importance in actively managed landscapes. In recent history (ca. 150 years) tree and shrub abundance has increased at the expense of native grasses in savannas and grasslands. The magnitude and patterns of change are spatially heterogeneous, highlighting the need for analytical approaches spanning multiple spatial scales, from individual plants to patches to landscapes. The overarching goal of this dissertation was to explore long-term dynamics associated with woody plant encroachment with aerial photography and field studies to examine cover, density, soils and land use history at the Santa Rita Experimental Range.The first study characterized patterns in woody cover change on contrasting soils over 60 years using aerial photography. Woody patch dynamics revealed encroachment and stabilization phases in woody plant proliferation. Soil properties reflected the rate at which uplands reached a dynamic equilibrium, but not the endpoint (ca. 35% cover). Fluctuations around dynamic equilibrium reflected net change in patch growth and acquiescence combined with colonization and mortality. Efforts to characterize changes in land cover will require patch-based assessments beyond coarse estimates of percent cover.The second study capitalized on historic field measurements of shrub canopies to validate estimates of shrub cover derived from the earliest aerial photography, quantified detection limitations of 1936 aerial photographs for mapping shrub cover, assessed species-specific contributions to percent cover, and translated detection limitations to proportions of velvet mesquite (Prosopis velutina var Woot.) biomass missed with 1930s aerial photography.The third study was a field-based approach investigating how livestock grazing influenced mesquite cover, density, biomass, and stand structure over 74 years. The study supplemented traditional statistical analysis of grazing effects with methods quantifying spatial autocorrelation structure of mesquite density by grazing treatment. The outcome re-affirmed the supposition that mesquite cover may be dynamically stable at ca 30%, and revealed that livestock grazing slowed the shrub encroachment process from 1932 to 2006, counter to expectation. Results indicate that shrub growth trajectories persist long-term. Overall, this work affirms the importance of land use legacies and long-term perspectives in rangeland shrub dynamics.