**3.2.2 Acquisition of input data**

To calculate slopes inclination, we used Digital elevation Model of Manche with a grid resolution of 50 X 50 m. The climate data start from the Torteval-Quesnay station, based 15 km East from our site. As for the general model, we collected the daily data for the 1991- 2004 period. These data were used for mapping the soil erosion hazard on monthly, seasonal and annual scales within the framework of an average climatic year. The data related to the agricultural practices result from a survey carried out among farmers who exploited catchment lands between 2005 and 2008. Those were used to map the annual hazard. For the other temporal scales, the hazard was estimated from agricultural data from the 2007-august 2008 period. Finally, the soil data (structural stability, available water content) were obtained from soil boreholes according to a density of 1 hole for 10 ha about agricultural area and from laboratory analysis concerning the structural instability of the superficial horizons. The spatial units of integration of the SCALES data correspond to the agricultural parcels.

#### **3.2.3 Results**

The monthly maps obtained for the September 2007-august 2008 period primarily reveal a significant intra-annual variability of the soil erosion hazard (Fig. 9).

One note a first sequence, between September and October, characterized by quite a low hazard on the majority of the cultivated parcels (approximately 300 hectares). The surfaces with medium level represent 2% of the catchment area. During this period, the erosivity is low because of nonexistent or very low hydrological surplus and insofar as the plant covering is relatively important because of the presence of fast-growing temporary crop (mustard) and wheat stubble.

SCALES: An Original Model to Diagnose Soil Erosion Hazard

**3.3 Discussions** 

results.

During summer, no cultivated parcel exceeds a low hazard level.

from monthly and seasonal representation of the hazard.

Fig. 10. Maps of soil erosion hazard at seasonal scale.

and Assess the Impact of Climate Change on Its Evolution 243

Autumn comes at the second rank for the soil erosion's most favorable seasons, before spring. The low hazard level is observed in the main parts of the cultivated parcels. But the medium hazard is still highly represented as it concerns 12% of the agricultural surfaces.

The initial conception of SCALES enables us to consider an adaptation of the model to succeed in the mapping the soil erosion hazard at intra-annual time scales. The evolutionary character of the model proves that we can go from a static temporal vision of the soil erosion hazard within the framework of an annual approach to a dynamic point of view starting

When we compare the monthly and seasonal data of the soil erosion hazard with the annual data (Fig. 11), we first notice that the initial version of the SCALES model doesn't enable to perceive the intra-annual variability of the hazard. In addition, the reading of the seasonal and annual hazards shows that the annual values are overestimated. Insofar as the input data of the initial model have changed, the results' comparison therefore appears to be delicate. However, we notice that the modularity of the SCALES model enables us to display the erosion hazard at different temporal scales and to get highly complementary

The adaptation of SCALES allows us to take into account the temporal as well as spatial variability of input data concerning climate and agricultural practices. Hence, for instance, the monthly erosive pressure of agricultural practices will move spatially from year to year according to the crop rotations decided by the farmer. Hazard levels will act in like manner. Thus, SCALES can be considered as a model which is spatially and temporally dynamic.

The monthly and seasonal approach of the soil erosion hazard needs to have local and precise information about agricultural practices. In addition, the characterization of the soil proprieties has to be based on field and laboratory data in a high spatial resolution. In these

A second sequence goes from November to January, which is different because of the levels of soil erosion hazard increase, with maximum levels in December. During this month, the parcels associated to a high soil erosion hazard represent 7% of the agricultural surfaces. December combines bare soils after the maize silage with abundant hydrological surplus and more numerous daily rainfalls exceeding 10 mm.

The third sequence goes from February to Match. Its marks a fall of the hazard levels on all the cultivated parcels because of an increase of the rate of plant covering and an important decrease of hydrological surplus.

The fourth period corresponds to April and May. We can notice higher hazard levels than in the previous period. The razing of temporary crops comes with the soil baring of the parcels used for maize production. In addition, this culture has a slow plant covering during the beginning of its growth, which leads to a prolonged exposition to intense rainfalls for the soil. Nevertheless, in spite of a large increase of bare soils, the erosivity is low because the hydrological surplus are nil and the repetition of daily rainfalls that exceed 10 mm are very low.

The last period, from June to August, is characterized by the lowest hazard levels that don't change during those three months. This can be explained by a nonexistent or low erosivity and a complete plant covering for the cultivated parcels.

At seasonal scale, the temporal variability of levels of soil erosion hazard, observed at monthly scale, is significantly smoothed (Fig. 10). However, the evolution in time of the soil erosion hazard intensity remains sensible. Winter appears as the period in which hazard levels are the highest. Yet they don't exceed the average level. This level concerns 32% of the agricultural area in the test zone.

Fig. 9. Maps of soil erosion hazard at monthly scale.

Autumn comes at the second rank for the soil erosion's most favorable seasons, before spring. The low hazard level is observed in the main parts of the cultivated parcels. But the medium hazard is still highly represented as it concerns 12% of the agricultural surfaces. During summer, no cultivated parcel exceeds a low hazard level.
