Spatial distributed modelling of erosion and sediment transport in mountain catchments in the Pyrenees: challenges for calibration and validation

Abstract

Most erosion models applied at basin scale are based on spatially aggregated representation, for example, in a space subdivision of the surfaces to which behavior is assumed homogeneous (sub-basins, hydrological units, etc.). This approach leads to a desirable simplification of the calculations, but presents difficulties in addressing problems related to sediment transport and identification of sediment source areas, which should be treated by a continuous spatial representation. Spatially distributed models, however, allow a more accurate approach to this problem, at the cost of an increased computational complexity. The objective of this work is the calibration and validation of the model WATEM/SEDEM, an empirical-conceptual spatially distributed model, to predict erosion and sediment yield in two watersheds in the Central Spanish Pyrenees: i) the watershed of the Barasona Reservoir (1504 km2), which is drained by the Ésera and Isábena Rivers, the model calibration and validation was based on the depositional history of the Barasona Reservoir and suspended sediment records over 3 years (May 2005–May 2008) at the outlet of the Isábena River; ii) the experimental Arnás catchment (2.84 km2), the model calibration was performed based on a dataset of soil redistribution rates derived from point 137Cs inventories, allowing capture differences per land use in the main model parameters. The validation process was carried with the registration of six years of suspended sediment at the outlet of the Arnás catchment. The calibration process for watershed of the Barasona Reservoir showed the problem you have when trying to calibrate the parameters of transport capacity with a single variable (the export of sediment to the basin outlet), making impossible to find a single set of parameters that optimize the error function, making it necessary to adopt a compromise solution. For the experimental Arnás catchment the model calibration processes using spatially distributed sediment yield derived from 137Cs inventories allowed calibrating the empirical parameters of transport capacity in a satisfactory way, finding a single combination of values that optimizes the error function. These results show that the calibration parameters of transport capacity are a fundamental aspect of the model WATEM/SEDEM and other similar models. To obtain a reliable estimate of the spatial distribution of erosion and sediment transport requires a calibration and validation by means of spatially distributed data of soil loss, which in turn allows a calibration of spatially distributed parameters concerning transport capacity

Keywords

• erosion del suelo, transporte de sedimentos, inventarios de cs137, modelo watem/sedem, coeficiente de capacidad de transporte, cuenca experimental de arnás, pirineo central español,