**2.3.2 Climate**

Rainfalls are regarded as the average factor of water erosion. It is allowed that the amount and the intensity of rains characterize the rainfall erosivity. Climate data comes from records of local meteorological stations of Météo France network. Insofar as the number of stations decreases with the lengthening of rains recording period, we decided to limit the reference period to 15 years (1991-2004) to be able to profit from a solid network of meteorological stations. This is composed of 22 stations located inside the Calvados and 19 stations located in periphery of this one. The main climatic data used are daily rainfalls.

Variables selected to appreciate the erosivity are (1) the yearly number of days with a Daily Intensity is above 10 mm (DI10) according to De Bruyn et al. (2001) and (2) the Yearly Positive Hydrological Balance (YPHB). This last variable at the same time allows us to take into account the amount of rains and the available water content (AWC) of soils. Let us recall that one of the major causes of erosion in Basse-Normandie is due to the saturation of soil. Also, we have to calculate beforehand the AWCs starting from the soil database presented further. Concerning the first variable, data of stations were interpolated by kriging method. In order to avoid interpolation errors related to the edge effect, we integrated data of the stations located just at outside of Calvados.

Classes are the following.

For DI10: < 20; [20-25[; [25-32[; [32-40[; >=40

For YPHB: <150 mm; [150-250 mm[; [250-350 mm[; [350-450 mm[; >=450 mm Those are respectively comparable with the levels of pressure 1, 2, 3, 4, 5.

#### **2.3.3 Agricultural practices**

In order to define the agricultural practices (Fig. 5), we had recourse to the data of the Agricultural census for 2000 realized at holding scale (source: Agricultural administration). 21 variables have been retained to evaluate the agricultural specialties of the 4844 holdings of the Calvados. They relate to socio-demographic characteristics of agricultural households, to juridical statute and economic dimension of the farms and to production systems (Dobremez and Bousset, 1996).

To calculate slopes inclination, we used Digital elevation Model of Calvados with a grid resolution of 20 X 20 m. The high resolution of DEM is essential because it allows us to carry out a topographical analysis of a very high degree of accuracy. Claessens et al. (2005) shows that DEM resolutions influence the slope inclination distribution: the coarser resolutions underline a larger contribution of lower slope angles (smoothing the effect on the landscape topographical representation). This report is in particular validated while comparing 50 m and 25 m DEMs. For Calvados, slope values result from the local cell-to-cell slope, rather than using a smoothing multiple cell windows, as done in major GIS procedures. Slopes were classified into six classes. Their limits were defined starting from values determined by statistics treatments of the cell slopes (classification according to the geometric progression method) and values coming from the literature (Le Bissonnais et al*.*, 2002). In contrast to the latter, the slopes larger than 15% have been regrouped in one class because of the absence of

The classes selected are as follows: [0-1%[; [1-2%[; [2-5%[; [5-10%[; [10-15%[; >=15%.

Rainfalls are regarded as the average factor of water erosion. It is allowed that the amount and the intensity of rains characterize the rainfall erosivity. Climate data comes from records of local meteorological stations of Météo France network. Insofar as the number of stations decreases with the lengthening of rains recording period, we decided to limit the reference period to 15 years (1991-2004) to be able to profit from a solid network of meteorological stations. This is composed of 22 stations located inside the Calvados and 19 stations located

Variables selected to appreciate the erosivity are (1) the yearly number of days with a Daily Intensity is above 10 mm (DI10) according to De Bruyn et al. (2001) and (2) the Yearly Positive Hydrological Balance (YPHB). This last variable at the same time allows us to take into account the amount of rains and the available water content (AWC) of soils. Let us recall that one of the major causes of erosion in Basse-Normandie is due to the saturation of soil. Also, we have to calculate beforehand the AWCs starting from the soil database presented further. Concerning the first variable, data of stations were interpolated by kriging method. In order to avoid interpolation errors related to the edge effect, we

In order to define the agricultural practices (Fig. 5), we had recourse to the data of the Agricultural census for 2000 realized at holding scale (source: Agricultural administration). 21 variables have been retained to evaluate the agricultural specialties of the 4844 holdings of the Calvados. They relate to socio-demographic characteristics of agricultural households, to juridical statute and economic dimension of the farms and to production systems

Those are respectively corresponding to the levels of pressure 0, 1, 2, 3, 4, 5.

in periphery of this one. The main climatic data used are daily rainfalls.

integrated data of the stations located just at outside of Calvados.

For YPHB: <150 mm; [150-250 mm[; [250-350 mm[; [350-450 mm[; >=450 mm Those are respectively comparable with the levels of pressure 1, 2, 3, 4, 5.

**2.3 Data processing 2.3.1 Topography** 

major dissected relief.

Classes are the following.

**2.3.3 Agricultural practices** 

(Dobremez and Bousset, 1996).

For DI10: < 20; [20-25[; [25-32[; [32-40[; >=40

**2.3.2 Climate** 

Adapted statistical treatments allowed us to draw up a typology of dairy farm (9 types) and non-dairy farm (13 types). A statistic aggregation (by summation) has been realized in order to assign them to the small administrative units (municipality scale). That led to the characterization of repartition profiles of different farm types for each municipality of Calvados which counts 706 of them. The following stage consisted in operating an Ascending Hierarchical Cluster followed by K-means method in order to reach a typology of farms (12 types) according to type repartition profiles (Bermond, 2004). Each type refers to modes of farm management, and to specific agricultural practices. A local farmer practice survey has been carried out in this direction, which enable us to produce our own data.

After selecting a sample of municipalities for each farm type, we interviewed the farmers about soil work methods used, plot localization and farm characteristics. This investigation showed that types of farm had notably evolved between 2000 and 2007. Thus, interviews have been used to update the 12 types of farm and to specify the current agricultural practices. This procedure has been applied to the sampled administrative units and, by extrapolation, generalized with all municipalities. Knowledge of crop rotations and management of intercrops allowed us to determine various modes concerning these agricultural input data of the SCALES model.

For crop rotations, the types are: winter crops, dominance of winter crop, balance winter crops/ spring crops, dominance of spring crops, spring crops. The passage of the first to the last type represents insofar the lengthening of the period during which soil is not protected. The duration of this period is in this way lower than 4 months for rotations based on winter crops and reaches a duration of 7 months in case of a succession of spring crops.

Those are respectively comparable with the levels of pressure 1, 2, 3, 4, 5.

Regarding the management of intercrops numerous publications underline the influence of different practices on soil erosion risks (Auzet, 1987; Martin, 1997; Martin 1999; Baumhardt and Jones, 2002; Le Bissonnais et al*.*, 2002; Lipiec et al., 2006; Strudley et al., 2008). Therefore the creation of a temporal plant cover like oilseedrape or mustard, in the period between two crops, will effectively protect the soil against run-off erosion. This measure will be less effective in case the crop partially covers the soil like for example rye (concept of scarce plant cover). One also considers that the wheat stubble correspond to this concept. With the absence of a temporal crop, soil tillage will permit to temporally reduce the erosion risk because of a better infiltration and a higher soil roughness. More the tillage operations are deeper, more effective are infiltration and soil roughness against soil erosion. The most unfavorable condition occurs when there is no tillage during the intercrop period remaining the soil bare. These different practices between crops or their absence (plant cover, scarce plant cover, deep ploughing, superficial ploughing, bare soil) are respectively comparable with the levels of pressure 1, 2, 3, 4, 5. The level 0 corresponds at the grassland and arboriculture areas.

## **2.3.4 Soils**

To achieve the aim of a fine diagnosis of soil erosion hazard, it was necessary to have a sufficiently precise soil database. It was not conceivable to exploit the Soil Geographical Database of France at scale 1 : 1 000 000. The regional BDSol-250 on a 1 : 250 000 scale (source INRA) does not exist. So we decided to create our own data.

SCALES: An Original Model to Diagnose Soil Erosion Hazard

spatial units in the north-west and center-east of the territory.

**2.4 Results** 

a rainy year

and Assess the Impact of Climate Change on Its Evolution 237

small adaptation of MWD (medium weight diameter of aggregates) classes found by Le Bissonnais and Le Souder (1995). This adaptation is based on the formation of two classes instead of one, which was initially provided for aggregates with a size larger than 2mm. For MWD of aggregate: > 3.5 mm; [3.5-2 mm[; [2-1.3 mm[; [1.3-0.8 mm[0.8-0.4 mm[; =<0.4 mm.

Combination of potential sensitivity of areas to erosion and rainfall erosivity in a "normal" climatic context leads to evaluation and mapping of mean erosion hazard (Fig. 6a). This document highlights of the existence of all levels of hazard. Level 0 shows parcels promoted with permanent grassland or orchards and/or with a slope lower than 1%. These parcels represent 1600 km² of agricultural surface (42 %). They are localized by the form of coherent

Fig. 6. Soil erosion hazard in Calvados at parcel scale (A) for a normal climatic contex, (B) for

Levels 1 and 2 of soil erosion hazard cover more than 500 km² of agricultural surface (13 %). It is located in majority in the central part of the Calvados along a North/South axis. Level 3 is

Those are respectively comparable with the structural instability levels 0, 1, 2, 3, 4, 5.

Fig. 5. Procedure for characterizing the agricultural practices in Calvados

During two years, we carried out nearly 8000 soil boreholes, which represent one borehole per 40 ha of agricultural land. Pedogenesis, soil thickness, coarse fragments, texture and hydromorphy have been registered. Data and Progressive knowledge of the soil landscape allowed us to produce a first soil map on which we selected 150 representative soil boreholes. Physical (granulometry, structural instability of surface horizon), hydrologic (AWC taking into account per cent of coarse fragments) and chemical properties have been determined. Structural instability has been evaluated starting from the INRA test of structural stability carried out on aggregates (Le Bissonnais and Le Souder, 1995).

Consequently, the soil features do not come from the application of the pedotransfer rules. This analytical step has been led us to finalize a global soil map on a 1 : 50 000 scale, to suggest at the same scale a map of soil structural instability, to propose another map in connection with spatial distribution of available water content of the soil, and, finally, combining the latter with interpolated rainfall data, to define the yearly positive hydrological balance. Classes of structural instability of the soil aggregates come from a small adaptation of MWD (medium weight diameter of aggregates) classes found by Le Bissonnais and Le Souder (1995). This adaptation is based on the formation of two classes instead of one, which was initially provided for aggregates with a size larger than 2mm. For MWD of aggregate: > 3.5 mm; [3.5-2 mm[; [2-1.3 mm[; [1.3-0.8 mm[0.8-0.4 mm[; =<0.4 mm. Those are respectively comparable with the structural instability levels 0, 1, 2, 3, 4, 5.
