Blame it on the Weatherman: How critical is rainfall to geomorphology?

The geomorphic activity of fluvial systems at any scale is ultimately driven by precipitation often in the form of rain. In numerical models, such as landscape evolution models, an input dataset of rainfall is commonly used to drive the model, often using a coarse spatial and/or temporal averaging. Beyond availability, characteristics of the rainfall data itself are frequently overlooked and the impacts of these on the results of the model not considered. However, landscape evolution models are sensitive to spatial and temporal variations in rainfall data and rainfall observations themselves contain spatial and temporal uncertainties, the nature of which varies between different observation methods.

This presentation synthesises the results of several linked studies highlighting the role rainfall can play in the modelling of geomorphology. First, we examine how the spatial and temporal resolution of the driving rainfall data is applied at influences the model outputs, with more than 100% difference in simulated sediment yields between the coarsest and finest resolutions used. Secondly, the role the source of the rainfall data plays, through comparison of observations from different methods, is explored showing that the uncertainty between the observations propagates non-linearly to simulated sediment yields.

To investigate these sensitivities the CAESAR-Lisflood model was used in combination with the STREAP weather generator to produce high-resolution estimates of rainfall, conditioned by observations, for the longer timescales required for landscape evolution studies. This pairing opened up the opportunity to investigate changes of geomorphic response to future predicted changes to rainfields due to climate change, showing that this is more complex than when considering changes to rainfall volumes alone.