Translating high-resolution climate change projections into erosion-vegetation feedbacks, sediment dynamics, and multi-century topographic evolution of dryland catchments

Changes in the properties of severe climatic events like rainstorms and droughts are expected to impact erosion rates significantly under modern global warming. Also over the recent geologic past and especially in drylands catchments, hydrological and vegetation transitions following changes in spatiotemporal properties of climatic phenomena have been suggested as triggers for periods of enhanced erosion that affected societies' sustainability and left pronounced topographic imprints. However, these potential drivers remain incompletely understood and quantified. This is, in part, because the discrete events that trigger erosion are hard to observe and the fine-scale processes needed to model erosion are computationally intensive to run over landscape evolution timescales.

To address this, we developed a new catchment-scale landscape evolution model based on the Landlab toolkit that explicitly represents episodic failures, sub-minute hydrology, overland-driven sediment transport, and erosion-vegetation links. We validated the model against event-based runoff and sediment records from the Lucky Hills site in the Walnut Gulch Experimental Watershed, Arizona, USA. After validation, we conducted a set of stochastic numerical experiments of landscape evolution in response to changes in sub-daily rainfall distribution, without considering changes in vegetation properties. We ran an additional set of simulations that integrated the landscape evolution model with historical and future climate records for the High Plains of Colorado, driven by a convection-permitting weather model (CPM). This experimental set allows us to explore changes in vegetation cover and its influence on sediment yield and topographic evolution under modern global warming.

We found that changes in the tail of the sub-daily rainfall distribution—changes similar to recent observations under modern global warming—could raise the total sediment yield by ~40% and alter the catchment morphology. Modeled sediment yield increased in response to the rising frequency of rare, high-magnitude storms, even when there was no significant change in the mean storm properties or annual rainfall. Further, we found that catchment erosion could increase even under a reduction in the mean conditions if the sub-daily rainfall distribution shifted toward a heavier tail. Numerical experiments driven by the CPM data confirm that under projected future conditions in the High Plains, erosion is expected to increase, even though the mean conditions become drier. Our simulations also reveal that the presence of vegetation impacts the morphology of the catchment, reducing channel density and preserving gullies' headcuts. Overall, this study contributes insight into the role of rainstorm properties and vegetation cover on landscape evolution, illuminates potential climatic triggers for past aggradation and degradation stages in low-order catchments, and provides valuable information for erosion risks under anthropogenic climatic and environmental changes.