Predicting Drivers of Collective Soil Function With Woody Plant Encroachment in Complex Landscapes

Dryland (arid and semiarid) ecosystems are extensive, home to a third of the human population, and a major contributor to terrestrial net primary productivity and associated biogeochemical cycles. Many dryland systems are undergoing woody plant encroachment, which can substantially alter landscape-scale soil nutrient dynamics via long-recognized islands of fertility mechanisms. To effectively constrain soil biogeochemistry responses to woody plant encroachment, predictions are needed for microbial biomass and especially microbial activity in addition to existing predictions for soil nutrients-referred to collectively hereafter as collective soil functioning. Here we evaluated whether collective soil functioning could be predicted from a suite of metrics including plant cover, precipitation, soil physiochemical characteristics, and topographic variables across complex landscapes undergoing woody plant encroachment by mesquite (Prosopis velutina). Plant cover alone predicted nearly half of the variability (R-2 = 48.5%) in collective soil functioning and had a significant effect on each component of this index (soil nutrients, microbial biomass, and microbial activity). Prediction strength for collective soil functioning increased to 55.4%, and the error term decreased by 13.4% when precipitation, soil physiochemical characteristics, and topographic metrics were also included in models (plant and environment model). Besides the expected effects of plant cover, other significant predictors of collective soil functioning included state factors such as topography, precipitation, and parent material along with soil age and bulk density. These results illustrate that mesquites influence many components of soil functionality but the strength of this effect depends on which component is analyzed and which environmental variables are considered.