Analyse de la distribution des sols en fonction de l'orientation de la pente : application à un secteur du Vexin français

Abstract

Many authors have evoked the major influence that slope orientation has on soil type and distribution. For example, very thin soils develop in sedimentary rock residuum on southwest-facing slopes, whereas thick soils develop in loess deposits on northeast-facing slopes (Fig. 1). Soils are thus closely related to the underlying geological materials. The present study, which aims to quantify the relationship between soil distribution and geomorphology was carried out in the part of the Paris Basin known as the 'Vexin français' region, to the northwest of Paris near Pontoise. Methodology. The standard Digital Terrain Model (DTM) enables the development of a rigorous statistical methodology applicable to large areas that consists in comparing soil data with topographic data (Fig. 2). Two types of data stored in two databases were used: 1. point data from boreholes with the results of sample analysis stored in the semantical DataBase (DB). 2. polygon data representing the soil Cartographic Units (CU), stored in the graphical DB of a Geographic Information System (GIS). The topographic data, from the standard DTM provided by the Institut Géographique National (IGN), were computed from the topographic map. This DTM has a pixel size representing 50 x 50 m. Results. The phenomena studied range in scale from a few metres to several kilometres. A rule governing soit spatial organisation was established, after acquiring a global perspective of the whole region. Figure 6a. in which 18 boreholes are plotted in 18 pixels of CU 76 (calcic brown soils developed in hard limestone), reveals that only one borehole falls outside the N140° to N320° field; Figure 6b shows the same phenomenon with all the 976 pixels of CU 76. Samples are more commonly located along slope orientations near the N140°-N320°-trending line (Fig. 6a). The pixels of CU 76 are shown as percentages of the total pixel population for each direction (Fig. 6c); N300° seems to be predominant. CU 45 (orthic luvisol in loess) covers a very large area (13871 pixels). The 99 auger boreholes (Fig. 7a) proved sufficient to characterise this soil unit. In contrast to Figure 6a, most of the 99 auger boreholes fall to the northeast of the N140°-N320°-trending line, representing an oversampling along the N320° and N120°-N140° directions. The pixel distribution ranges between N0° and N80° (Fig. 7b) and the relative importance of the N0° direction is revealed by the percentage of the total pixels (Fig. 7c). The results, obtained on low-angle slopes (average 2 to 3%), are comparable to those of Bourennane (1992, 1997) for the Beauce region to the south of Paris. CU 10 and CU 17 show a similar soil-distribution trend, also corresponding to slopes facing opposite directions, but steeper (averaging respectively 11 and 9%). CU 10 consists of podzolic sandy soils in Stampian sands and CU 17 of sandy loam soils contaminated by loess on the same Stampian sands. CU 10 soils, represented by 1512 pixels, occur on southwest-facing slopes plotting within the N120° to N300° field whereas CU 17 soils, represented by 1417 pixels, occur on northeast-facing slopes plotting within the N300° to N120° (particutarly N0° to N80°) field (Fig. 8). The rules of soil distribution were then tested using a few CU pairs, the results of which are generatised for the whole map (Pontoise 1-2) using all the 591 auger boreholes (Fig. 9); three soil sets are identified: - soils in loam formations (L): 17% of the auger boreholes, - soils in other sediments (S): 60% of the auger boreholes, - soits in heterogeneous formations. reworked on the slopes (P): 23% of the auger boreholes. A total of 82% of the loamy soils (L) are located ta t/w northeast of a N315°-N135°-trending line, and 79% of the soils developed in other sediments (S) are located to the southwest of the same line. The P soils show a regular distribution in all directions. These results show excellent precision for use in soil mapping and illustrate the spatial organisation of soils in the Vexin français region. Interpretation and discussion. Soil distribution patterns are necessarily related to depositional processes and pedogenetic evolution; two main factors seem to influence soil distribution, namely wind and topography. The wind effect. Many Quaternary specialists believe that the loess of the Vexin plateau is eolian and derived from the north or northwest (the Channel was above sea level during the Weichsetian). After the last glacial event, the dominant winds may also have changed to westerly. The results of the soil distribution model were then compared with present dominant wind directions (Fig. 10). During the autumn, the mean direction (N240°) of strong to medium (> 5 m/s) winds is opposite to that of the loamy soils distribution (N60°), whereas the mean direction of low-velocity winds (< 4 m/s) is perpendicular. In summer however, mean wind frequency is the highest (N300°), with marked N320° lowvelocity winds. Wind efficiency has an effect on transport capacity: loess particles are detached from the slopes facing the dominant wind direction and deposited on the leeward slopes. Wind efficiency can be evaluated by its kinetic energy (Ec). The Ec of médium-and high-velocity (> 5 m/s) winds is presented on Figure 11 and is seen to be most efficient on bare soils during summer, autumn and winter. Ec is maximum in the N40° to N12O° field, which is slightly different from the maximum distribution of silty soil (Fig. 9) in the N0° ta N9O°field. Lower velocity (< 4 m/s,) winds,from N17O° in autumn and winter and N320° in summer, are very efficient for carrying small particles such as loess, thus extending the range of efficient wind energy to match the pattern of silty soil distribution. In winter, daily freeze-thaw alternations probably also influenced, and perhaps still do, the differential erosion of south-southeast- to northwest-facing slopes; such alternations do not occur on north-northwest- to south-east-facing slopes. The topographic effect. Tectonics can play a major role in shaping landforms and, indirectly. can influence the distribution of many soils and superficial deposits. Although the sedimentary layers in the Vexin region dip shallowly to the northeast or southwest and in the Beauce region they are horizontal, the same Soil Organisation Model (SOM) pattern is observed. This is also the case for the Yonne department (where the sedimentary layers dip perpendicularly to those of the Vexin region) and in the west of the Paris Basin (accentuated by an eastern to northeastern dip). The SOM has also enabled a better knowledge of the tectonic features along the Vigny anticline, many of which are not recorded on the 1:5O,000-scale geological map. Conclusion. The presented Soil Organisation Modeb is shown to be applicable to the 'Vexin français' region and can probably be extended to a larger area of the Paris Basin. It is thus proposed to use it is for predicting soil type and superficial deposit distribution, and inasmuch could provide a convenient tool for soil and geological surveys. It can be used to focus on areas of uncertainty (Fig. 13) where the soil type is not known. These areas are of limited extent and correspond to slope orientations probably related to major tectonic features.