Contribution of phytoecological data to spatialize soil erosion: Application of the RUSLE model in the Algerian atlas

Among the models used to assess water erosion, the RUSLE model is commonly used. Policy makers can act on cover (C-factor) and conservation practice (P-factor) to reduce erosion, with less costly action on soil surface characteristics. However, the widespread use of vegetation indices such as NDVI does not allow for a proper assessment of the C-factor in drylands where stones, crusted surfaces and litter strongly influence soil protection. Two sub-factors of C, canopy cover (CC) and soil cover (SC), can be assessed from phytoecological measurements that include gravel-pebbles cover, physical mulch, annual and perennial vegetation. This paper introduces a method to calculate the C-factor from phytoecological data and, in combination with remote sensing and a geographic information system (GIS), to map it over large areas. A supervised classification, based on field phytoecological data, is applied to radiometric data from Landsat-8/OLI satellite images. Then, a C-factor value, whose SC and CC subfactors are directly derived from the phytoecological measurements, is assigned to each land cover unit. This method and RUSLE are implemented on a pilot region of 3828 km(2) of the Saharan Atlas, composed of rangelands and steppe formations, and intended to become an observatory. The protective effect against erosion by gravel-pebbles (50%) is more than twice that of vegetation (23%). The C-factor derived from NDVI (0.67) is higher and more evenly distributed than that combining these two contributions (0.37 on average). Finally, priorities are proposed to decision-makers by crossing the synthetic map of erosion sensitivity and a decision matrix of management priorities. (C) 2021 International Research and Training Center on Erosion and Sedimentation, China Water & Power Press. Publishing services by Elsevier B.V. on behalf of KeAi Communications Co. Ltd.

Automated summary

The RUSLE model is commonly used to assess water erosion, with policymakers able to reduce erosion by controlling cover and conservation practices. However, using vegetation indices like NDVI may not accurately assess the C-factor in drylands. A method to calculate the C-factor from phytoecological data is introduced in this paper, aiming to map it over large areas using remote sensing and GIS.