Rainfall spatial-heterogeneity accelerates landscape evolution processes

Nadav Peleg,Chris Skinner,Jorge Alberto Ramirez,Peter Molnar

doi:10.1016/j.geomorph.2021.107863

Abstract

Catchment hydro-morphological response is mainly conditioned on rainfall properties, such as rainfall intensity, storm duration and frequency, and the timing of these events. Rainfall spatial variability is likewise a major determinant affecting streamflow, erosion, and sediment transport, and is explored largely in the context of heavy rain triggering floods and fast morphological changes on hillslopes and in channels. In this study, we examine how the spatial structure of rainfall influences landscape evolution at the catchment scale over hundreds of years. To achieve this, multiple realizations of hourly rainfall fields, each differing only by their spatial distribution but identical in all other respects, were simulated using a weather generator. The impact of storm spatial-heterogeneity on the catchment morphology was then assessed with a landscape evolution model (CAESAR-Lisflood). A virtual “open-book” type catchment was used for this numerical experiment. The mean streamflow and low-flows remained the same while the magnitude of the annual peak streamflow increased by up to 12% in response to higher rainfall spatial heterogeneity. However, the erosion and deposition rates significantly increased (up to 50%) and the net erosion and deposition areas changed (increased by up to 9% and decreased by 13.5%, respectively) when the rain became less uniform in space. Furthermore, new gullies were found to be longer, deeper, and more branched in response to increased rainfall heterogeneity. The results suggest that heterogeneity in rainfall spatial patterns speeds up landscape development, even when rainfall volumes and temporal structures are the same. This implies that the spatial structure of rainfall may have more of an influence on catchment morphology at long time scales than previously thought.

Highlights

Rainfall is a crucial component in many hydrological processes, and hydro-morphological systems are highly sensitive to it (Fekete et al, 2004)
No change was detected when comparing the low flows emerging from the same coefficient of variation (CV) scenarios
Landscape models are sensitive to ‘geomorphological equifinality’ (Beven, 1996) issues because their evolution is heavily influenced by the landscape topography itself; these models have the potential to be robust to input uncertainties over long-term simulations (e.g. Hancock et al, 2016; Skinner et al, 2020). This would suggest that while our results indicate considerable changes to the gully development depending on the rainfall spatial structure, given enough time, each of the rainfall spatial scenarios will eventually result in similar changes in the stream pattern

Summary

Introduction

Rainfall is a crucial component in many hydrological processes, and hydro-morphological systems are highly sensitive to it (Fekete et al, 2004). Rainfall is known to be the main trigger for landslides (Iverson, 2000; Leonarduzzi and Molnar, 2020), debris flows (Guzzetti et al, 2008; Marra et al, 2016), and rapid erosion and transport of sediment (Valentin et al, 2005; Xu et al, 2015). Comparisons between storms with the same rainfall amounts reveal that longer-lasting storms can be more efficient in saturating the soil, enhancing streamflow and triggering more landslides (Montgomery et al, 1997; Struthers et al, 2007; Tarolli et al, 2011). The temporal properties of rainfall and their effect on hydromorphological systems have been investigated for many years, with most of the studies focusing on aspects of intensity, duration, and frequency of rainfall (Koutsoyiannis et al, 1998; Sivapalan and Bloschl, 1998; Willems, 2000; Borga et al, 2014; among many others)

Objectives

Methods

Results

Discussion

Conclusion