Chapitre 4:Rôle de la végétation sur les propriétés hydriques des sols

Chapitre 4 :ROLE DE LA VEGETATION SUR LES PROPRIETES HYDRIQUES DES SOLS

L’effet de la végétation sur le sol est suffisamment important pour qu'il entre parfois dans les critères de classification. Tel est le cas pour une végétation "naturelle", par exemple sous forêt. Cependant, dans de nombreux cas d'étude des propriétés hydriques des sols, le rôle de la végétation n'est ni pris en compte ni même explicité. Dans ce qui suit, les effets de la végétation et plus généralement de l'usage, incluant les pratiques agronomiques seront abordés. Deux sols de même dénomination ayant des caractéristiques pédologiques identiques situés respectivement sous culture céréalière et sous forêt de hêtres seront comparés. Ce travail a fait l’objet d’un article soumis à la Revue Plant and Soil.

91

Chapitre 4:Rôle de la végétation sur les propriétés hydriques des sols

4.1 SOIL WATER REGIMES UNDER FOREST AND CULTIVATION AND THEIR CONSEQUENCES ON OVERCONSOLIDATION AND WATER RETENTION PROPERTIES. F. BIGORRE & D. TESSIER* Laboratoire de Science du Sol, INRA Versailles, 78026 Versailles Cedex (France)

*Corresponding author. Phone:+33-1- 30833243; Fax: +33-1-30833259 e-mail: [email protected]

ABSTRACT The effects of old deforestations upon cultivated soils in temperate climates are not yet well known. Do cultivated and forest soils classified in identical soil units have the same physical properties? The answer to this question is important for all the disciplines which require a spatial management of environmental problems. The purpose of this article is to show how intensive cereal crops with high yields, cultivated on old forest soils, can greatly modify soil physical properties. Our study was carried out in the east of France on loamy soils that are representative of the leached soils (Luvisol, FAO classification) which represent large covers in Europe. Historically, part of them have remained beech-planted, whereas the other part were cleared in the middle of the 18th century and brought under cereal crop. The particle-size distribution and the pH are similar on both forest and cultivated soils. However, for the latter, there is a considerable decrease in water retention along with an increase in bulk density. The available water capacity calculated up to 1 m depth is 30 % lower than to that of the forest soil. There is a complete reorganization of the soil fabric whose principal feature is a collapse of the finest pores with the genesis of coarser pores associated with cracks. Water balance modeling based on the last 35 years made it possible to conclude that the soils under beech plantation have never undergone drought constraints as great as those undergone by the cultivated soils.

92

Chapitre 4:Rôle de la végétation sur les propriétés hydriques des sols

Cultivation has considerably increased the water uptake from soils. We suggest that the resulting high levels of desiccation caused the soil densification over the whole profile (~1 m) similar to an overconsolidation phenomenon. This work also showed the importance of determining soil physical properties on undisturbed soil samples, i.e. undried samples in order to avoid air drying which in turn erases a large part of the soil history and thus the water retention properties. Key words: cultivation, drought, overconsolidation, soil compaction, water balance, water properties.

INTRODUCTION In France, the intensification of agriculture has allowed a considerable increase in crop yield. In 1950 the average yield was of 2 t ha-1 and is nowadays of 7 or even 9 t ha-1 (Boulaine, 1992). This yield increase has had consequences on the water requirements of plants (Farahani, 1998). In many areas of France today, the water available for the plants has become the main limiting factor of agricultural production. The effects of such changes are not known very well, in particular the effects of drought intensity on soil structure, soil available water capacity and soil water regime. Many studies have addressed the effects of cultivation on soil properties. Most have been based on the comparison of plots with known histories, in geographical areas of relatively recent anthropogenic impacts (Canada, United States, Australia and Brazil). These studies have generally focused on the impacts of cultivation on soil organic matter stocks and quality, and on fertilizing elements (Arrouays et al., 1995; Bauer and Black, 1981; Besnard et al., 1996; Elliott, 1986; Blank and Fosberg, 1989; Zhang et al., 1988; Dormaar, 1979; Mann, 1986; Ihori et al., 1995; Voroney et al., 1981). A great deal of attention has also been put on soil structure stability. Although physical properties and especially water properties have been studied less often, several studies have shown soil compaction or densification following cultivation at depths of up to 0.5 m (Blank and Fosberg, 1989; Hammel, 1989; Bauer and Black, 1981; Low, 1972; Skidmore et al., 1975). Coote and Ramsey (1983) observed significant changes in the pore spectrum and water retention curve after 35 years of cultivation. Scott et al. (1989) showed that water retention was lower at a given water potential in a cultivated soil compared to a native prairie soil. Cattle et al. (1994) underlined a number of changes in the macropores of the soil profile, with thin section analysis.

93

Chapitre 4:Rôle de la végétation sur les propriétés hydriques des sols

Higher drought constraints, resulting from a yield increase of a grass crop, can cause an increase in the macropore size (Grevers and De Jong, 1990). The modifications in soil pore space have been explained by a reorganization of the structural elements following an increase in the desiccation levels reached by the soil after the introduction of more waterdemanding plants (Bouma and Hole, 1971). Some consequences on soil hydraulic conductivity were observed by Bouma et al. (1975). The plant water regime, depending on the intensity and rapidity of desiccation it allows through differential swelling and the creation of fracture planes, causes greater or lesser aggregate fragmentation (Dexter, 1991). The plants use these cracks for their rooting more or less intensively according to the nature of their root network (Matechera et al. 1994). Veenhof and Mc Bride (1996) and Ahuja et al. (1998) concluded that variations in soil water regime may originate an overconsolidation in the cultivated soils. The purpose of our study is to analyze the effects of a rotating crop wheat/barley/colza on the physical properties of a soil which was initially under forest cover (Beech). For this we compared the soil structure and water retention of the same soil either under beech cover or under wheat cover. Detailed attention has been given to the methodology used to establish the water retention curves as it seems essential to work on samples disturbed as little as possible and especially on samples that have not been air-dried.

MATERIALS AND METHODS

Study site of and sampling methods The studied soils were located in Lorraine in the East of France (Figure 4-1). They are representative of the leached soils in the large European geological basins. The area is in a contrasting semi-continental climate. In the last 40 years, the average rainfall has been of 755 mm for a Potential EvapoTranspiration (ETP Penman) of 670 mm. Two profiles representative of the luvisol (FAO, 1988) of this area were sampled on pits within a distance of fifty meters (Figure 3-3). The first one was under a beech forest, the second under a cereal crop with a rotation of wheat, barley and colza. The forest area has been managed for 250 years under regular beech grove (Fagus Sylvatica). The mean age of the settlement is more than 150 years and the oldest trees are 250 years old, which corresponds to

94

Chapitre 4:Rôle de la végétation sur les propriétés hydriques des sols

the period of clearing. Local documents show that the forest-cultivation limit has not changed since this period. At the end of winter, i.e. after full re-wetting, undisturbed decimeter-scale soil blocks were collected from the pits. We preferred to carry out the sampling respecting the soil horizons rather than constant incremental sampling. The depths used to compare the two soils were not always exactly identical but corresponded to the same horizons (Table 4-1). At the time of the sampling, the soil moisture was close to the field capacity. As shown by Hall et al. (1977), in such conditions, there is lower risk of volume disturbance during the experimental process in the high potential value range. In order to avoid desiccation and organic matter mineralization, the samples were kept in a cold room under hermetic packaging. Some of the samples were subsequently air-dried and sieved to 2 mm for physicochemical analyses. In order to assess spatial variability of soil properties, additional sampling was performed at three different depths (0-20, 40-60, and 60-80 cm) along transects including the pits in both forest and cultivated soils (Figure 3-3).

Physicochemical analyses The particle size distribution (AFNOR, 1996), organic carbon (AFNOR, 1996), pH in water (AFNOR, 1996) and cation exchange capacity (AFNOR, 1996) were measured on the pit samples. The following analyses were carried out on the auger samples for spatial variability determination : particle size distribution, total carbon and total nitrogen (C&N Element analyser, Fison Carlo Erba) and pH.

Physical analyses, water retention curves Small clods (5 to 10 cm3) were manually separated from the blocks. Water retention measurements were carried out at -1, -3.2, -10, -32, -100, -320 and -1600kPa on 5 to 15 replicates. A filtration device, (Tessier and Berrier, 1979) was used from -1 to -100kPa, whereas a pressure cell apparatus (Richards, 1948) was used at –320 and –1600 kPa. In order to establish pore water continuity, a kaolin paste was placed between the sample and the porous plate. The samples were brought to equilibrium for 7 days for each water potential (AFNOR, 1996). For low potential values (- 2.8 to -107MPa), the samples were placed under controlled hygrometry by using salt saturated solutions for 1 to 2 months.

95

Chapitre 4:Rôle de la végétation sur les propriétés hydriques des sols

Clod bulk density was measured by immersion in kerosene (Monnier et al., 1973). The dry weight was obtained after 48h in a drying oven at 105°C. Bulk density measurements on clods were compared with those on cylinders (360 cm3) in order to evaluate interclod porosity and thus a part of the structural porosity of the sample. The effects of intense desiccation on both soil structure and water retention curve were studied after applying experimental air-drying. After that, the samples were rehydrated under high vacuum conditions with a 6-cm capillary head of de-aerated water. Then, the water retention curves of the samples were established over dehydration. The water retention curves are presented by plotting the water ratio (water volume/solid volume) as a function of water potential. In order to calculate the solid volume, the solid density was determined using the pycnometer method (Blake, 1965).

Water balance modeling The soil water balance was modeled at daily time-step from 1961 to 1995. Two models using the same calculation methods were used. In these two models, plant transpiration was calculated by means of statistical relations depending on the potential ETP, the foliar index and on the soil water potential. For the forest, we used the BILJOUR model (Granier et al., 1999). For the crop system (wheat), we used the STICS model (Brisson et al., 1998). Cultivated Soil

Pit

Auger sampling for spatial variability

12m

Grass land

Beech Forest Studied zone

0

5

10

15

20

Scale in Meters Figure 4-1 : Studied zone location and sampling plan.

96

Chapitre 4:Rôle de la végétation sur les propriétés hydriques des sols

RESULTS

Physicochemical properties, spatial variability The soil analyses are presented in Table 4-1. In both profiles (under forest and under crop) the particle-size distributions were similar. A vertical clay gradient was observed indicating clay migration. The pH values were similar and relatively low (4.8-5.6). The cation exchange capacities (CEC) of the soils were similar and followed the pattern of clay distribution. The saturation rate by alkaline and alkaline-earth cations was below 0.6 in the first fifty centimeters and slightly higher in the lower horizons (from 0.6 to 0.9). Soil carbon and nitrogen contents were different in surface layers and identical below 45 cm. Another sampling carried out on 32 samples on transects around the profiles showed that the pH values were very homogeneous (Figure 3-3; Table 4-2). The soil pH was slightly higher in the 0-20 cm layer of the cultivated soil than that of the forest soil (5.61 and 5.26, respectively). Soil carbon and nitrogen contents were different in 0-20 cm layers and identical below 40 cm, whereas the C/N ratios were slightly higher under forest than in the cultivated soil in the first 60 centimeters. This difference can be related to the parent forest and crop vegetation. The particle-size distributions were similar in the 0-20 cm horizon, except for sands that represented less than 4% of the total. In the 40-60 cm depth, the fine fractions differed between the two soils and the coarse fractions were similar. In the 60-80 cm depth the particle-size distributions were identical except for the fine silts. The standard deviation percentages (std%) were generally low thus indicating great homogeneity between the results, confirming the representativeness of the two soil profiles. These results enabled us to conclude that the two soils were similar under forest and cultivation as far as their pedological development and geochemical environment were concerned.

97

Chapitre 4:Rôle de la végétation sur les propriétés hydriques des sols

Table 4-1: Pedological analysis and soil chemical properties of the two soil profiles. Particles size analysis

Exchangeable bases

-1

(cmolc.kg-1)

(g kg )

depth.

0-10 cm 15-20 cm 45-53 cm 62-70 cm 75-90 cm 2-15 cm 35-42 cm 45-50 cm 70-80 cm 80-95 cm

Hor. C.S F.S. C.L. F.L Clay C . . Forest soil A1 15 27 317 423 218 37 E 18 28 321 422 211 13 4 17 247 348 384 3 B1 3 Bt 18 16 213 327 426 3 B/C 11 17 240 359 373 Cultivated soil Ap 26 32 308 408 226 14.3 3.9 E/B 17 23 250 352 358 3.0 B1 23 22 217 328 410 27 29 199 326 419 3.2 Bt 29 29 201 321 420 2.8 B/C

pH Ca2+ Mg2+ K+ Na+ S/T CEC

5.6

7.31

1.28

0.65 0.08 0.87

4.8

1.11

0.28

0.20 0.02 0.28

5.8

5.1

6.16

2.74

0.23 0.06 0.62

14.9

5.1

8.03

3.41

0.23 0.09 0.68

17.4

5.0

5.46

2.44

0.23 0.06 0.58

14.2

5.0

3.82

0.37

0.35 0.01 0.72

6.3

4.9

5.65

0.75

0.25 0.02 0.57

11.8

5.1

6.71

1.85

0.10 0.04 0.58

14.9

5.1

7.77

2.40

0.22 0.06 0.72

14.6

5.2

7.98

2.58

0.11 0.06 0.70

15.3

(C.S.=Coarse Sand ; F.S.=Fine Sand ; C.L..=Coarse loam ; F.L.=Fine Loam)

Table 4-2: Spatial variability of soil data under crop and forest. Values in Italics are significantly different under crops and forest (ANOVA, p