Precipitation and Soil Properties Determine Long-Term Consequences of Disturbance and Invasion in Drylands

Disturbance and invasive species have dramatic effects on desert plant communities, often resulting in degradation or shifts to alternative plant communities. Climate and soil properties determine water availability to plants, and have been thought to drive patterns of recovery following disturbance and potential for invasion. In chapter II we used a combination of natural gas pipelines and satellite imagery to understand how recovery from a uniform disturbance differs across precipitation and soil gradients. We used a recovery ratio (disturbed/undisturbed) of pipeline pixels and their undisturbed nearest neighbor pixel to quantify recovery in a comparable way across precipitation gradients. We found widespread evidence of incomplete recovery of NPP except in locations that receive the majority of their precipitation during winter. Recovery timing for shrub cover varied 50-120 years after disturbance, and was quickest in warm deserts where post-disturbance shrub species differed from the dominant shrub species in undisturbed locations. In chapter III we use natural gas pipeline data (same as chapter 1), but asked how disturbance affects production-precipitation relationships. We did this by comparing precipitation use efficiency (average NPP/average precipitation) and sensitivity to deviations in annual precipitation at disturbed and undisturbed sites across a precipitation gradient. We found that effects of disturbance varied according to shifts in vegetation type following disturbance, with general trends of lower PUE and higher sensitivity tied to losses of woody plants increases in grass species. In chapter IV we create a species distribution model for an exotic annual grass species Bromus tectorum using a novel approach that combines soil microclimate and germination models. The model was unable to explain abundance of B. tectorum across sites, but predicted presence well (73% accuracy). Our modeling approach allowed us to predict suitable B. tectorum habitat at an accuracy level similar to models using satellite imagery. Hindcasts of our model indicate that abiotic conditions that favor presence of B. tectorum are increasing with the biggest changes at mid-elevation sites (1765-2111m).