Land cover change (1950-2011) impacts on soil erosion in a Mediterranean catchment

Hari G. Roy, Dennis M. Fox±, Karine Emsellem

UMR 7300 ESPACE CNRS, University of Nice Sophia Antipolis, Nice, France 2
± Dennis Fox: (


Vineyard is reported to be one of the land covers most vulnerable to soil loss in the Euro-Mediterranean region. Mediterranean areas are particularly vulnerable to soil erosion due to high rainfall intensities, agricultural activities on steep slopes, low organic matter, low nutrient contents, and rapid land-use changes indicated several reasons for accelerated soil erosion in vineyards: soil is bare much of the cultivation period and vineyards are relatively common on steeper slopes. During spring and winter, soil surfaces of many Mediterranean vineyards remain almost bare and highly vulnerable to loss due to high moisture content, loose upper layer, and high intensity rainfall events; moreover, abandonment of farmland, expansion of vineyard in the upland forests and cereal fields have also accelerated soil erosion risk. Vineyards in the Mediterranean area have the highest soil erosion rates – greater than rain fed cereals, olives, grassland, and forest cover.

In Southeast France, Roy et al. (2014) showed that between 1950 and 2011, vineyard lost more than a third of its initial cover. Urban and suburban encroachment explain much of the loss, but conversion to green space was also important. Vineyard losses in the plain were compensated in part by new fields on steeper slopes. The catchment river (La Giscle) flows into a tourist port located in the Gulf of St Tropez and port authorities have had to dredge the channels of sediments to permit navigation.

The objective of this study was to determine the impacts of land cover changes in the Giscle catchment over the 1950-2011 time interval. The study is based on Revised Universal Soil Loss Equation (RUSLE) model predictions for each period. RUSLE has been used extensively in predicting soil erosion, including in the Mediterranean basin, and land cover trends cited here are comparable to many other Mediterranean locations where near-coastal zones are under intense development pressure.

Materials and Methods

Grape production is the main agricultural activity in the 235 km² catchment that flows into the Gulf of St Tropez, SE France. Roy et al. (2014) mapped land cover from orthorectified aerial photographs and vineyard cover went from about 2,426 ha in 1950 to 1,561 ha in 2011 (Figure 2). Much of the loss in vineyard cover is due to urban/suburban pressure and land prices. Vineyard fields are located mainly in the sandy floodplain and have spread under urban pressure onto steeper slopes in recent years.

The RUSLE module in IDRISI estimates average annual soil erosion rates and determines spatial patterns of soil loss (Eastman 2012). The model was run using a 25 m DEM for 1950, 1982, 2003, and 2011 to estimate soil erosion based on land cover maps produced by Roy et al. (2014). RUSLE parameters for erosion estimation are described below and the equation is the following (Renard et al., 1997):


A = average annual soil loss (T acre-1 yr-1) (converted to T ha-1 yr-1); R = rainfall-runoff erosivity factor; K = soil erodibility factor; LS = slope length and gradient factor; C = cover management factor; P = support practice factor.

Changes in farming techniques were impossible to account for historically and were ignored.

Results and Discussion

As steeper slopes were cultivated, mean erosion rate increased from about 11.8 T ha-1 yr-1 in 1950 to a peak of 14.7 T ha-1 yr-1 in 2003 (Figure 1). Since the RUSLE model does not account for sediment deposition, actual erosion rates in this study are probably slightly overestimated.

 Figure 1  Figure 2
Figure 1: Mean and median erosion rates. Figure 2: Total erosion.

(click on images to enlarge)

Total erosion reflects changes in mean erosion rate and vineyard area, where vineyard area dominates total erosion trends (Figure 2). The small loss in vineyard area between 1950 and 1982 is compensated by an increase in mean erosion rate so the net result is a negligible increase in total erosion (Figure 2). Despite a further increase in mean erosion between 1982 and 2003, total erosion drops by about 21%. Total erosion decreases again slightly in 2003-2011 as both vineyard area and mean erosion decrease. The changes in total erosion rate correspond to about 291 T ha-1 yr-1 and 224 T ha-1 yr-1 for the 1982-2003 and 2003-2011 time intervals, respectively. Total soil loss represents the amount of sediment that can potentially be injected into the stream network, potentially causing problems for aquatic biodiversity (pesticides and nutrients adsorbed to sediments) and channel navigation in the ports downstream.


The study quantified the impact of land cover change on soil erosion in vineyards in the Giscle catchment in south-east France where RUSLE was used in the IDRISI GIS environment to create and compare soil erosion rates and maps of 1950, 1982, 2003, and 2011. Public subsidies and urban/suburban pressure decreased vineyard area by about 30% over the study period. Although mean erosion rates generally increased, the net effect is a decrease in the total sediment loss in the catchment due to the major change in vineyard land cover. The shift of vineyards to steeper slopes is unusual as most studies observe the abandonment of steeper slopes in favor of more intensive agriculture in alluvial plains. This contradiction is probably only true for vineyards (and not cereal crops) and may be present elsewhere in the Mediterranean. Total erosion in 2011 represents about 75% of erosion in 1950-1982, and this is probably good news for water quality and port authorities in charge of maintaining navigation channels in the Gulf of St Tropez ports.



Renard, K. G., Foster G.R., Weesies G.A., McCool D.K., and Yoder D.C., 1997. Predicting soil erosion by water: a guide to conservation planning with the revised universal soil loss equation (RUSLE). Agriculture Handbook (Washington), (703).

Roy H.G., Fox D.M., Emsellem K., 2014. Spatial dynamics of land cover change in a Euro-Mediterranean catchment (1950-2008). Journal of Land Use Science 10:277-297, doi: 10.1080/1747423X.2014.898105