Ex-post evaluation of erosion control measures in southern Mali

As part of an impact study of a soil and water conservation (SWC) project in southern Mali, the effect of erosion control measures on soil erosion was evaluated. In one village, a baseline situation from 1988 was compared with the situation in 2003, after farmers had installed stone rows, live fences and grass strips, and had started cultivating across to the slope. This comparison showed a spectacular decrease in gully volume in cultivated fields of 87%, from 58 to 8 m3 ha−1. Estimated annual soil loss decreased with 77% from 42 to 10 t ha−1 year−1. However, baseline data on erosion gullies were not available for other villages. In the absence of baseline data, a simple ‘with–without’ comparison does not allow a correct evaluation because farmers install erosion control measures especially in fields with more erosion. Two alternative methods were used: a reconstructed baseline and a virtual time series. Using the reconstructed baseline, looking not only at active gullies but also at (partly) reclaimed gullies, we concluded that line interventions, gully interventions, and a combination of both, reduced the proportion of active gullies by 48%, 47% and 70%, respectively. Using a virtual time series, comparing erosion in fields with erosion control measures installed in different years, we concluded that erosion gradually decreased by 50% during the 3 years after installation of erosion control measures. In a separate study, we found a positive effect of gully interventions and sowing across to the slope in reducing sheet erosion. A reduction of the slope of the sowing direction by 1% reduced the cover of runoff deposit of coarse sand with 8%. Although a documented baseline is preferred, both a reconstructed baseline and a virtual time series are useful tools and make ex-post evaluations more relevant than a simple with–without comparison.