**3. Adaptation of the SCALES model at seasonal and monthly scales**

The SCALES model leads to the production of soil erosion hazard levels on a very precise spatial scale and for region-size territories. However, the results enable at best to compare median annual situations in a normal climatic context and during years with rainfalls higher than the normal climatic context. Yet, these are tendencies that hide an intra-annual variability of the erosion hazard, which shall be necessary to evaluate in order to take into account the quick change of climate conditions and of the surface state of cultivated soils. The next step is hence about an adaptation of the initial model in order to be able to evaluate the erosion hazard on seasonal and monthly scales.

#### **3.1 Data with intra-annual variability**

The data of the initial model characterized by an intra-annual variability correspond to agricultural practices and to climatic and pedoclimatic parameters (rainfalls and yearly positive hydrological balance). The modalities of the yearly repartition of the daily high intensity rainfalls and precipitations' volumes will affect the variability of the rain erosivity during months and seasons. As well, the agricultural practices associated to climatic conditions will affect the rate of plant covering and its evolution. Yet, the plant covering of a bare soil strongly intervenes on the probability of soil erosion by water. The adaptation of the SCALES model needs to get the monthly and seasonal data about 1° rates of plant covering and their evolution for cultivated parcels and 2° rain erosivity conditions.

the most represented since it covers 1400 km² (37%) of agricultural surface. It is located southwest of the department and locally in the East. The fourth level is the only one representing important hazard since the level 5 is absent. The affected parcels cover a surface of 100 km²

The assessment of the soil erosion hazard for a rainy year (2001, + 15% compared to the "normal" 1970-2000) has been achieved in order to study the potential impact of the rainfall conditions supporting high erosivity (Fig. 6b). Results are particularly interesting because they show that agricultural surfaces affected by a level 4 (strong hazard) extended more than 800 km² compared to a year with "normal" climatic conditions. The most affected areas are South-West, the central North/South axis and secondarly the eastern part of the territory. This brings us to considerate that Calvados is a department presenting a strong predisposition for the genesis of soil erosion by water, erosion which express itself as soon

Finer representation with zoom effect of levels of hazard at the scale of parcel unit reveals the occurrence of a mosaic of colors expressing very frequent spatial disjunction at local scale about erosion hazard (Fig. 7C). The rapid and brutal spatial variations of physic properties of the area associated with the interpenetration of grassland and crops parcels contribute to the strong heterogeneity of the results at local scale. The precision of input data of the model SCALES allows to obtain this type of conclusion. It also comes to the idea that the management of this issue assumes in priority an approach at the scale of a parcel or a

The aggregation of soil erosion hazard data at administrative and hydrologic scales (Fig. 7A-B) shows a significant loss of information when the basic scale is given up (Fig. 7C). SCALES model loses quickly its interest but can become a communication tool about the question of

The SCALES model leads to the production of soil erosion hazard levels on a very precise spatial scale and for region-size territories. However, the results enable at best to compare median annual situations in a normal climatic context and during years with rainfalls higher than the normal climatic context. Yet, these are tendencies that hide an intra-annual variability of the erosion hazard, which shall be necessary to evaluate in order to take into account the quick change of climate conditions and of the surface state of cultivated soils. The next step is hence about an adaptation of the initial model in order to be able to evaluate

The data of the initial model characterized by an intra-annual variability correspond to agricultural practices and to climatic and pedoclimatic parameters (rainfalls and yearly positive hydrological balance). The modalities of the yearly repartition of the daily high intensity rainfalls and precipitations' volumes will affect the variability of the rain erosivity during months and seasons. As well, the agricultural practices associated to climatic conditions will affect the rate of plant covering and its evolution. Yet, the plant covering of a bare soil strongly intervenes on the probability of soil erosion by water. The adaptation of the SCALES model needs to get the monthly and seasonal data about 1° rates of plant

covering and their evolution for cultivated parcels and 2° rain erosivity conditions.

**3. Adaptation of the SCALES model at seasonal and monthly scales** 

(2,6%). They are essentially gathered at the extreme south-west of Calvados.

as climatic conditions induce strong erosivity.

the erosion hazard on seasonal and monthly scales.

**3.1 Data with intra-annual variability** 

group of parcels.

soil erosion.

Fig. 7. Aggregation of the soil erosion hazard data: (A) at the administrative scale (municipality); (B) at the hydrologic scale; (C) significant deterioration of results from the parcel scale to the administrative scale due to aggregation procedure
