Influence du pH sur les propriétés des sols : l'essai de longue durée des 42 parcelles à Versailles pH effect on soils properties: long time experiment of 42 plots in Versailles

Résumé

Les pratiques de fertilisation et d'amendement peuvent contribuer à modifier profondément les propriétés des sols. Ceci est mis en évidence par le dispositif des 42 parcelles de l'INRA à Versailles qui reçoit chaque année les mêmes traitements depuis 1929. Il présente l'originalité d'être en jachère nue. Les résultats reportés ici reposent sur 16 traitements que nous comparons à une parcelle témoin. Le sol est à caractère limoneux avec 14 à 22 % d'argile. Le pH, la capacité d'échange au pH du sol (CEC_sol), et les cations échangeables ont été mesurés. Des mesures de porosité, de rétention de l'eau à différents potentiels de l'eau ont été effectuées.

Des variations considérables de pH sont observées. Les engrais ammoniacaux conduisent à des parcelles très acides (pH 3,5 à 5,0), à faible CEC_sol, essentiellement saturée par des cations aluminium. Dans les parcelles chaulées ou recevant des amendements basiques, le pH est tamponné à 8,2 par les carbonates. La CEC_sol double du sol acide au sol chaulé où elle est majoritairement saturée par du calcium échangeable. Avec les sels neutres (par exemple KCl, NaNO₃) le pH n'a pas changé et pour le témoin il a légèrement baissé. Des observations de terrain montrent que ces différents états physico-chimiques sont en relation avec les états de surface du sol et influencent la porosité du sol en profondeur.

Ainsi dans les sols des régions tempérées la CEC_sol peut doubler dans un domaine très restreint de pH (6,0 à 7,5), orientant fortement les propriétés physiques des sols.

Mots-clés:

• Expérimentation longue durée,
• fertilisation,
• pH,
• acidification,
• propriétés des sols,
• capacité d'échange en cations

Abstract

In temperate regions, soil acidification is a real problem. The pH change causes transformations of both the chemical and physical properties of the soil. The 42 experimental plots at the Institut national de Recherche agronomique (INRA) in Versailles, France, were created in 1929. Since that time, each plot has received either fertiliser or amendments (Table 1). There are ten reference plots that have not been treated. Plots were dug up twice a year and left fallow. The results corresponding to 17 plots, 16 treated and one reference plot, are presented in this paper. The samples were collected in 1999 and compared to the 1929 soil reference. The pH was measured in water (AFNOR 1994), the cation exchange capacity (CEC) was determined at pH 7.0 (AFNOR 1994, CEC₇) and at the soil pH by the cobaltihexamine method (Ciesielski and Steckermann 1997, CEC_soil). Exchangeable cations (Ca²⁺, Mg²⁺, K⁺ and Na⁺) were measured after cobaltihexamine extraction and exchangeable aluminium by the KCl method (McLean, 1965). Water retention measurements were carried out with an apparatus developed by Tessier and Berrier (1979) and AFNOR (1996) at -10 kPa, with the Richards pressure plate cell at -1.6 kPa (Richards 1948) and with a controlled hygrometry dessicator at -107 MPa (Tessier 1984).

The average pH of the reference plot collected in 1929 was close to 6.3. In 1999, pH values ranged between 3.5 and 8.2. For plots treated with ammonium fertilisers the pH ranged from 3.5 to 5.5. In contrast, plots with basic treatment or liming had a soil pH that was close to 8.2 and was equilibrated with carbonates. Intermediate pH values (from 5.5 to 7.5) were obtained for plots treated with neutral salts such as KCl or NaNO₃, or with superphosphate.

The cation exchange capacity in soils at the reference pH (7.0) demonstrated variations in soil composition. For example, the CEC₇ varied with clay content as a function of depth or with organic matter loss or gain due to lying fallow or to manure application. When the CEC was measured at soil pH (CEC_soil), strong differences were observed. Two groups of plots can be differentiated in 1999. One ranged between 11 and 16 cmol+/kg; this range corresponded to plots with amendments or with fertilisers containing bases. A lower range of CEC_soil values (below 11 cmol⁺ /kg) was obtained with ammonium fertilisers. It is important to note that between pH=6.0 and pH=7.5 the CEC_soil doubled.

Six months after digging, the surface state of the soil was variable. Thick crusts dominated in acidic conditions, whereas the presence of dispersing cations (K⁺, Na⁺), and a smooth surface was present in soil treated with amendments. In the soil profile, bulk density increased with acidity and sodium or potassium cation concentrations. At -10 kPa, water retention measurements demonstrated that bulk density and sampling depth were linked with water retention: the higher the bulk density, the lower the water retention. At -1.6 MPa, water retention was correlated to the amount of clay and organic matter content. We also showed that the exchangeable calcium content and CEC_soil influence water retention. At -107 MPa, water retention depends on pH as well as the CEC_soil and exchangeable cation concentration.

The main purpose of this study was to show that after 70 years, plots subjected to intensive fertilisation or amendments and lying fallow (i.e., without organic matter restitution), have dramatically different soil properties. The first indicator of this evolution was pH. Ammonium fertilisers produced very acid plots (pH 3.5 to 5.0). In the presence of neutral salts (e.g., KCl, NaNO₃) and in the reference plot, the pH had decreased a little. In liming or basic treatment plots, the pH is controlled at 8.2 by carbonates. The second important factor to consider is the cation exchange capacity value. When measured at a reference pH (pH=7.0), the CEC₇ reflects the natural components present in a soil, but doesn't consider the influence of physico-chemical factors that operate in the native soil. After 70 years without organic restitution, the soil has lost approximately half of its original organic matter content, and as a result the exchangeable sites were approximately 2 cmol⁺ /kg.

The effective CEC at soil pH demonstrates the influence of pH on variable charges due to mineral constituents and organic matter. Plots with ammonium fertilisers have a low cation exchange capacity. Fertilisers with sodium and potassium cations have increased the exchangeable sodium (up to 11% of the CEC_soil) or potassium (18%) respectively. The cation exchange capacity doubled in going from acidic plots to liming plots, and in the latter the CEC is mainly saturated with exchangeable calcium.

In situ plot observations show that soil physico-chemical properties strongly influence not only soil surface state, but also soil profile porosity. Both acidification and dispersing cations are factors in soil degradation, whereas high pH values and calcium as the exchangeable cation produce strong structure stability and high porosity. It is also interesting to note that the CEC_soil can double between pH=6.0 and pH=7.5, thus influencing the physical properties of the soil. Comparing CEC₇ and CEC_soil facilitates the prediction of the effects of fertilisation and amendments on soil properties. The cation exchange capacity at soil pH can be used as an excellent indicator of soil quality.

Keywords:

• Long-term experiment,
• fertilisation,
• pH,
• acidification,
• exchangeable cations,
• cation exchange capacity