Multiscale drivers of spatially variable grass production and loss in the Chihuahuan Desert

TitleMultiscale drivers of spatially variable grass production and loss in the Chihuahuan Desert
Publication TypeConference Paper
Year of Publication2011
AuthorsPillsbury FC, Peters DC, Yao J, Okin GS
Conference Name96th ESA Annual Meeting, Earth Stewardship: Preserving and enchancing earth's life-support systems
Date Published08/2011
PublisherESA
Conference LocationAustin, TX
ARIS Log Number265788
AbstractHistoric regime shifts from grass- to shrub-dominated states have been widespread in the Chihuahuan Desert and other arid and semiarid regions globally. These patterns of grass production and shifts to shrub dominance are spatially variable, and show a weak correlation with precipitation, suggesting that drivers and processes at different spatial and temporal scales are important determinants of the production and recruitment of grasses. Using long-term data from the Jornada Basin USDA-LTER site, we tested the hypothesis that spatially contagious processes of soil redistribution and shrub invasion can overwhelm the importance of fine-scale drivers (e.g., precipitation, soil depth, vegetation composition) to explain variation in grass production and loss. We found that both fine-scale drivers and contagious spatial processes explain variation in grass production in different parts of the landscape. Grass production on locations surrounded by stable soils is primarily a function of precipitation, soil characteristics, and vegetation composition. However, locations in close proximity to invasive shrubs and erosive soils have poor relationships with local drivers. In these locations, spatial processes of sand deposition and dune formation associated with invading shrubs limit the ability of grasses to exploit otherwise favorable conditions for establishment and growth. These results suggest that focusing on fine-scale processes, such as competition, is of limited utility when spatial processes interact with broad-scale drivers to produce observed patterns. A more explicit focus on multiple interacting spatial and temporal scales is often necessary to more fully understand significant ecological regime shifts.