Identification of eroded areas using remote sensing in a badlands landscape on marls in the central Spanish Pyrenees

Identification of erosion areas on a regional scale can be very useful for environmental planning, and can help reduce land degradation and sediment yield to streams and reservoirs. Remote sensing techniques were used to determine erosion and erosion risk areas in a badlands landscape in the Esera River catchment (Spanish Pyrenees). The size, sparse vegetation cover, and high erosion level in the badlands, enabled good visual and digital discrimination relative to other land covers and surfaces. The maximum likelihood algorithm was used for obtaining a spectral distance map to the bare soil signature characteristic of badlands on marls. The ROC (receiver operating characteristic) curve analysis was applied to this map for obtaining an optimum classification of the badlands. Two alternative classification thresholds were set to determine erosion areas and areas at risk of erosion. Two classification performance statistics, the model's sensitivity and specificity, were calculated as a means of expressing the uncertainty-omission and commission errors-associated to both maps. Most erosion risk areas coincided with low vegetation cover surrounding the badland areas. (c) 2008 Elsevier B.V. All rights reserved.