**3.3 Discussions**

242 Soil Erosion Studies

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

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

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

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

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

and more numerous daily rainfalls exceeding 10 mm.

and the repetition of daily rainfalls that exceed 10 mm are very low.

and a complete plant covering for the cultivated parcels.

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

decrease of hydrological surplus.

agricultural area in the test zone.

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 from monthly and seasonal representation of the hazard.

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 results.

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.

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

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

SCALES: An Original Model to Diagnose Soil Erosion Hazard

**4.2 SCALES data for the current period and for 2100** 

accumulation of rain is around 950 mm.

and winter crops (wheat).

**4.2.1 Data for the current period** 

units represented by agricultural parcels.

in 2100.

**4.2.2 Input data for a forecast of the soil erosion hazard to 2100** 

the GIEC that reproduce by default the current agricultural situation.

and Assess the Impact of Climate Change on Its Evolution 245

The climatic context is similar to the one described in the previous case study. Rainfall is quite constant all year long with an increase in autumn and winter. The annual

Agriculture combines cereal crops, fodder crops, and bovine breeding gathered in huge farms intensively exploited. Grassland cover half of the agricultural land of the catchment. These are almost always permanent grassland. Crops are covering an area equivalent to the one of the grassland. Arable lands are fairly divided between spring crops (maize silage)

The digital elevation model of Calvados at 20 m scale was used to estimate the slopes of the Lingèvres. The local climatic data comes from the same weather station than the one mentioned in the case study of the Branche catchment. This weather station is located 5 km from the Lingèvres. The identification of the agricultural parcel network and agricultural practices was based on field work. The data collected covers the period from September 2009 to august 2010. The soil data used for calculation of available water content and hydrous budget were extracted from the database related to the 1/50000 map of Calvados soils described in the first part of this document. Additional survey with soil boreholes also allowed refining the spatial resolution of pedological characteristics of the Lingèvres. Therefore the structural instability of the catchment's soils was deducted from already existing administrative data and from 18 additional analysis on representative of the soils and agricultural practices diversity. Input data of the model and their treatment at monthly and seasonal scale for an average climatic year had been integrated in elementary spatial

The agricultural and climatic data represent the two types of data that show a temporal variability. Nowadays, works relating to climate change allow us to have pieces of information that make consensus in the scientific community of climatologists for a distant future (GIEC, 2007a). Unfortunately, we cannot say the same for agricultural practices as the agricultural evolution depends at the same time on its interaction with climate, political choices and the socio-economic situation. Yet, it is not possible to precise with certainty what will be local agriculture in a distant future because we know little if nothing about the agricultural consequences of these interactions, political choices and socio-economic characteristics by 2100. So, our projection of the erosion hazard in 2100 leans on data from

To characterize the Norman climatic context by 2100, we used and adapted the GIEC's simulation data (GIEC, 2007a) concerning the A1B scenario (Cantat et al., 2009). These ones reveal a annual increase of temperatures in Normandy on the order of 2.8° C from now to 2100 (Fig. 12A) with a global warming being more intense during summer (+ 3.2° C). With regard to rainfalls, the annual accumulations would remain stable, which would nevertheless hide differentiated seasonal behaviors: +9% during winter and -21% during summer (Fig. 12B). All these data were used to determine the conditions of rainfall erosivity

conditions, it is not possible to carry out a monthly mapping of the hazard erosion for territories that exceed several hundred or thousand of km². This work must be limited to areas recognized as sensitive through the initial model's representation of the hazard or by the intermediary of the land managers' knowledge.
