Changes of Banded Vegetation Patterns in Drylands and their Drivers

Banded vegetation patterns on hillslopes in arid and semi-arid landscapes are theoretically expected to change—expand and contract in their size and/or change in their biomass density—as local meteorological and climatic conditions change. Some areas show stronger effects than others. In this study, using Google Earth Engine (GEE), we collected 32 years of data on patch bandwidth and normalized difference vegetation index (NDVI) from seven sites of banded vegetation patterns throughout the world. Results show that changes in annual precipitation affect either patch bandwidth or biomass density. To investigate the drivers of banded vegetation patch dynamics further, we constructed a Bayesian hierarchical model using GEE data and a set of predictor variables. Results from the statistical model suggest that slope, soil texture, soil pH, and NDVI are effective predictors for variation in patch wavelength and that the influences of these predictors vary from site to site. These findings may suggest that in some regions, climatic conditions are changing at a rate that vegetation cannot match. We are investigating further conditions that lead some sites to respond to changes in precipitation with adjustment in biomass density, while other sites in the vegetation bandwidth.