Hydrogeologie

A distinctive feature of the Nhecolândia, a sub-region of the Pantanal wetland in Brazil, is the presence of both saline and freshwater lakes. Saline lakes used to be attributed to a past arid phase during the Pleistocene. However, recent studies have shown that saline and fresh water lakes are linked by a continuous water table, indicating that saline water could come from a contemporary concentration process. This concentration process could also be responsible for the large chemical variability of the waters observed in the area. A regional water sampling has been conducted in surface and sub-surface water and the water table, and the results of the geochemical and statistical analysis are presented. Based on sodium contents, the concentration shows a 1: 4443 ratio. All the samples belong to the same chemical family and evolve in a sodic alkaline manner. Calcite or magnesian calcite precipitates very early in the process of concentration, probably followed by the precipitation of magnesian silicates. The most concentrated solutions remain under-saturated with respect to the sodium carbonate salt, even if this equilibrium is likely reached around the saline lakes. Apparently, significant amounts of sulfate and chloride are lost simultaneously from the solutions, and this cannot be explained solely by evaporative concentration. This could be attributed to the sorption on reduced minerals in a green sub-surface horizon in the “cordilhieria” areas. In the saline lakes, low potassium, phosphate, magnesium, and sulfate are attributed to algal blooms. Under the influence of evaporation, the concentration of solutions and associated chemical precipitations are identified as the main factors responsible for the geochemical variability in this environment (about 92% of the variance). Therefore, the saline lakes of Nhecolândia have to be managed as landscape units in equilibrium with the present water flows and not inherited from a past arid phase. In order to elaborate hydrochemical tracers for a quantitative estimation of water flows, three points have to be investigated more precisely: (1) the quantification of magnesium involved in the Mg-calcite precipitation; (2) the identification of the precise stoichiometry of the Mg-silicate; and (3) the verification of the loss of chloride and sulfate by sorption onto labile iron minerals.

A model of anaerobiosis based on gas diffusion and microbial respiration is proposed for soil aggregates. Unlike previous models, it accounts for the dependency of O2 consumption on O2 and CO2 concentrations and, consequently, on chemical reactions of CO2 in soil solution. The model is tested with experimental data of respiration and O2 distribution within spherical remoulded aggregates of different sizes saturated with water. Most of the model parameters were estimated experimentally.Fick's law could describe O2 diffusion inside the saturated aggregates. The model agreed well with O2 profiles measured soon after saturation. Later, the model underestimated anaerobiosis, probably because of changes in radial distribution of microbial activity inside the aggregates. Respiration as a function of O2 concentration was satisfactorily described by the model. Nevertheless, each size of aggregate was analysed separately because of an apparent relation between size and maximal aerobic respiration. This seems to be associated with a non-homogeneous distribution of aerobic respiration. The model represents an improvement on those currently used to estimate anaerobiosis inside aggregates.

Under anaerobic conditions, microbes closely interact with geochemical reactions and can have an impact on the soil, the deep vadose zone, the underlying aquifer and the atmosphere. We have designed a model combining anaerobic microbial activities with geochemical reactions in the soil, and assessed it in batch experiments. The model describes the dynamics of six functional microbial communities, their decomposition after death, and the catabolism of carbohydrates through denitrification, dissimilatory NH 4 þ production, Fe(III) reduction, fermentation, acetogenesis, and SO 4 2reduction. It was combined with a model that thermodynamically describes acid-base, reduction-oxidation and complexation reactions in solution, and kinetic precipitation and dissolution. Batch incubations were done on a Calcic Cambisol, either without amendment, or after supplying (i) glucose or (ii) glucose and NO 3 -. Gases, mineral cations and anions, glucose, fatty acids and alcohols were measured during incubation. Net production of CO 2 was similar for both glucose treatments, about 40 times larger than in the control. For the glucose treatments, the main microbial activities were fermentation, acetogenic transformation of ethanol, and oxidation of H 2 . When the soil was enriched with NO 3 -, no H 2 was produced, and microbial activities were rapidly inhibited by NO 2 -. The model shows these trends as well as geochemical characteristics including pH and reduction-oxidation potential.

Soils of the terraces of the Niger River have locally undergone, prior to irrigation, a process of alkalinization. The use of the resulting nonsaline sodic soils [pHs 8.5–9.8 (s is “on saturated paste"), ECS = 2.2–3.2 dS m−1, SAR = 12–28 (mmol L−1)1/2, exchangeable sodium percentage (ESP) = 5–40] is greatly limited because of their alkalinity and sodicity. The mechanisms of degradation affecting the soil physicochemical properties, the water supply, and the mineral nutrition of crops were analyzed in the Lossa irrigation scheme in the Tillabery region in Niger. Increase in pH corresponds with an increase in the compactness and decrease in the permeability of the soils. In the subsoil a threshold effect is observed for an ESP ‐10 or a pH ∼8.5, and the hydraulic conductivity becomes very low, ∼0.05 mm h−1. Reducing conditions inhibit the mineralization of organic matter, favor denitrification, and cause a deficiency in nitrogen. The simultaneous increase in pH diminishes nutrient availability and causes deficiencies in phosphorus, then in potassium and zinc, in the most alkaline soils. While maize yield is greatly affected by soil degradation, an aquatic forage grass (Echinochloa stagnina) seems to be more adapted to these nonsaline sodic soils.

In some soils of the Niger valley, alkalinization and sodization are related to the concentration of alteration products of calco-alkaline gneiss to biotite, which releases significant amounts of cations and alkalinity. The result is an oversaturation of the soil solution with respect to calcite which precipitates with a positive calcite residual alkalinity (C-RA). Ca 2+ molality decreases while alkalinity and pH increase; Ca 2+ desorption is accompanied by adsorption of Na + on the exchange complex. K + , Mg 2+ and Na + concentrations in the soil solution are successively controlled by the formation of illites and smectites type silicates while kaolinite dissolves. These processes were successfully simulated according to the analytical data. They square with significant increase of the alkaline reserve in the soils according to chemical, physical and mineralogical alterations. These geochemical mechanisms were found to be involved in the formation of alkali soils at three sites of the region.

Evidence of soil degradation impeding soil tillage and irrigation in cultivated soils in Pakistan is identi®ed, described and represented in a general process of degradation. Based on a chemical analysis of soil characteristics, it is shown that a more general geochemical degradation process may occur in these soils. Two paths of salinization, i.e. neutral salinization and alkalinization inducing a process of sodi®cation, are identi®ed. The wide range of chemical properties of soils and corresponding geochemical processes can be explained by the great diversity of quality in irrigation water that is taken either from the canal or from the groundwater. The basic module of a geochemical model AQUA (ValleÁ s and DeCockborne, 1992) is calibrated with the help of a study of the soil geochemical properties (identi®cation of minerals, characterization of exchanges) and then used to assess the eect of four dierent water qualities on sandy and loamy soils. Based on these scenarios, the salinity, alkalinity and sodicity hazard of irrigation water is assessed by taking into account simultaneously the electrical conductivity and the residual alkalinity (calcite-residual alkalinity, residual sodium carbonates) or irrigation water and the soil cation exchange capacity: these three indicators appear the most relevant in the context of the study.

We investigated the effects of slurry application on water quality in wells, pasture-drainage conduits and rivers. The parameters determined were electrical conductivity, dry residue, pH, dissolved O 2 , NO À 3 , NO À 2 , Cl À , Na, PO 3À 4 organic matter content, COD, suspended solids, turbidity, colour, Fe, Mn, Cu and Zn. The data were analysed by principal components analysis and cluster analysis. The ®rst axis extracted by principal components analysis of the samples-by-variables matrix represented the degree of dilution of the water strongly related with saline content; the second axis represented redox conditions, affecting organometallic component. In general, the positions of the samples in this factor space re¯ected the major contamination processes affecting water resources of that type (wells, conduits or rivers). In the case of river-water samples, cluster analysis showed that month of sampling had a clear effect, re¯ecting the importance of rainfall. Finally, in all three water resource types, contamination was most severe in the area of highest livestock density and highest frequency of slurry application. #