Testing a low-complexity spatially distributed model to simulate the intra-annual dynamics of soil erosion and sediment delivery

Erosion models simulating the intra-annual effects of hydrometeorological drivers and disturbances (e.g. vegetation clearcutting, tillage events, wildfires) need to represent temporal variability at time scales below the long-term annual average (e.g. the native timescale of the (Revised) Universal Soil Loss Equation). Here, we test a low-complexity, spatially distributed model (WaTEM/SEDEM: W/S), to simulate 15-day erosion and sediment dynamics. A standardised modelling routine was applied to four monitored and well-studied catchments in North-West Europe with open-access discharge (Q) and suspended sediment load (SSL) data, creating a model workflow implementable with predominantly pan-European Union data. Despite introducing temporally variable rainfall erosivity and crop cover into W/S, a temporally static calibration of transport capacity (TC) could not adequately replicate the SSL variability in most catchments. Instead, embedding seasonality into the TC through a multitemporal calibration routine improved the representation of the SSL variability by reducing and increasing the sediment transport efficiency in summer and winter, respectively. The optimal multitemporal TC revealed a negative relationship, or relative decoupling effect, between the gross erosion (i.e. the pixel-scale soil displacement) and the in-channel SSL. The net effect in most catchments was a reduction in the magnitudes of the simulated internal sediment fluxes at aggregated timescales compared to a temporally static TC calibration. While the multitemporal TC profiles give interesting insights into the potential seasonal biases within the model, care should be taken in its interpretation due to the confounding influence of potential error compensation in time and space, as well as unrepresented hydrological and erosion processes. Despite the complexities involved in the temporal downscaling of WaTEM/SEDEM, compared to single iteration long-term approaches, we show the utility of this approach to better understand the interdependencies between temporal scale and spatial redistribution rates within soil erosion models capable of large-scale applications.