Impacts of changing storm characteristics in a stochastic model for dryland vegetation pattern formation

In various drylands around the globe vegetation can be self-organized into striking stripe patterns, where vegetation alternates with bare soil over kilometer scales. These banded patterns are oriented transverse to a gentle slope of the terrain, of 0.5-2% grade. Models, as well as early observational studies, suggest these patterns exist because of positive feedbacks between the vegetation distribution and the soil water distribution that follows a rare rain storm. For example, there is enhanced infiltration of storm water in the vegetated regions compared to the bare ones, in part due to the increased surface roughness caused by the vegetation, which slows any overland water flow. We compare different models for the soil water re-distribution process on a gentle hillslope, following a rain storm, to investigate how that depends on precipitation characteristics such as typical storm depth and storm intensity, both of which are likely to be altered by climate change. Storm variability is captured as well through use of a simple stochastic rainfall model. Our modeling focus is on determining when an abrupt change in rainfall characteristics might lead to loss of vegetation bands, or even a complete collapse of this vulnerable dryland ecosystem.