Quantification of soil losses and multivariate statistics of factors controlling water erosion (Allal Al Fassi watershed, Morocco)

Abstract

In order to preserve water resources and conserve agricultural land and biodiversity, a precise assessment of soil losses is necessary. It is for this reason that the empirical model of revised RUSLE soil loss equation was applied in order to quantify and describe the rate of erosion in Allal Al Fassi watershed, located in the Middle Atlas mountain chain (Morocco). To understand and define the relationships between water erosion and the factors aggravating its dynamics in Allal Al Fassi watershed, the use of multivariate statistics is in fact an integral part. Sub-models have been adopted to calculate factors of R erosivity, K erodibility, LS slope length and steepness, C soil conservation, and P anti-erosion practices using field data, remote sensing, and GIS (geographic information systems). The map of soil losses resulting from RUSLE model allows us to distinguish five classes. The average annual soil losses are estimated to be 624.07 t ha−1 year−1, with a wide range that goes from 3.57 to 3521.07 t ha−1 year−1 and a median equal to 335.48 t ha−1 y−1. They are greater upstream of watershed, mainly in the eastern part, where slopes are steep and long and soil are highly erodible with little or no development despite very variable conservation and very low to low erosion. However, they are low in areas with low to very low erosion, slope length, and steepness and erodibility, with low anti-erosion practices and strong soil conservation. Multivariate statistical analysis has shown that soil losses in Allal Al Fassi watershed are governed by LS slope length and steepness, K erodibility, and P anti-erosion practices and that the percentages of organic matter, sands, and silts have a significant impact on the variation of soil losses. The latter prove to be important for soils type C1 complex, entisols and inceptisols, soils on dolomitic/calcareous substrate, and soils occupied by matorrals and steppes. RUSLE model has proved useful in quantifying erosion and determining the factors controlling it in Allal Al Fassi watershed, highlighting priority areas that require urgent measures (preventive or remedial) and also where erosion risk reaches irreversible stages.