L'étude porte sur un magasin aquifère calcaréo-dolomitique très karstifié (karst à pitons ou tourelles) d'une région à climat tropical humide de la vallée de Li-Jiang.
L'échantillonnage hydrochimique concerne les zones saturée et non saturée de différents secteurs du karst.
La détermination des pressions de CO2 et des coefficients de saturation par rapport à la calcite ont fait l'objet d'une attention particulière. Les concentrations en calcium et bicarbonates (éléments prépondérants) varient respectivement dans les intervalles 3,1 à 4,7 et 2,5 à 4,8 mé.l-1.
La plus grande partie des échantillons de la zone saturée est en équilibre avec la calcite et la dolomite, sous des pressions de CO2 de l'ordre de 1 à 2.10-2 atm.
Les eaux de la zone non saturée sont celles de différents suintements recueillis au plafond d'une galerie naturelle. Elles se distinguent par une inégalité entre les teneurs en bicarbonates et en calcium, interprétée comme étant due à des échanges de bases avec les résidus de colmatage des fissures.
Les résultats hydrochimiques font supposer l'existence, dans le karsts de plusieurs domaines aquifères liés à la morphologie, interconnectés, avec drainage général par le Li-Jiang.
The main chemical features of groundwater in carbonate devono-carboniferous rocks in southern China
This paper deals with a research program aimed at studying the hydrogeologic properties of the water-rich carbonate rocks in Southern China. We focus our attention on the Guilin area (Guangxi province). Our results concern the hydrochemistry and we show how a study of time- and space-variations in the chemical composition of groundwater (saturated and unsaturated zone) may help in understanding the hydrodynamic conditions in a karstic aquifer.
The Guilin karstic area is mainly composed of a limestone and dolomitic limestone series extending in age from Mid-Devonian to early Triassic, reaching 3 000 m in thickness. In the Li-Jiang valley, the quaternary caver (alluvial deposits) is 10 to 20 meters thick. Geomorphology is named "tower karst", as it may be observed in various tropical regions. In fact, two different types of landform appear : the "fonglin" (peak forest) and the "fongchong" (peak cluster). In the former, isolated peaks (mean altitude : 250-300 m) rise in the plain (altitude around 150 m). In the latter, the peaks join together and delimit closed depressions (attitude up to 600 m in the Guilin area). Deep karstification is as important as surface karstification, which is related to the tectonic context (rocks are intensely fractured), to the subtropical wet climate (mean annual rainfall : 1 900 mm) and also to intense groundwater circulation.
Our hydrochemical study has been carried out on various types of water points (springs, sinkholes, large diameter wells, boreholes, an underground river, oozing points from the unsaturated zone). They are located in both the fonglin and the fongchong area, but the highest density is close to the contact fine between the two landforms. The chemical analyses are interpreted in three steps :
1) calculation (if possible) of the saturation coefficient relative to calcite and dolomite and of the CO2 equilibrium pressure,
2) study of time variations in the composition of water sampled at oozing points from the unsaturated zone and,
3) statistic study of space variations in the composition of water sampled at alt points.
Hydrochemistry is dominated by Ca2+ and HCO3- ions. Except for two aberrant points (probably resulting from a mixture with surface water), Ca2+ and HCO3- concentration varies in a fairly narrow range (respectively 3.12-4.72 meq.l-1 and 2.53-4.80 mea.l-1). Most of samples from the saturated zone are close to equilibrium with calcite and often with dolomite, under CO2 pressures about 0.01-0.02 atm. Small scale variations in lithology (presence of dolomite) and in CO2 pressure may therefore explain space variations in hydrochemistry and enable waters of the same origin to be identified.
The study of the unsaturated zone consists in sampling the water at three oozing points located in three fractures of the roof of a natural cave. Althrough the points are quite near to each other, water composition differs clearly between them. The most interesting result though is provided by the following experiment. One of the three points was initially dry. The flow occurred when breaking the concretion obstructing it. For six hours we studied the discharge and the water-composition variations at that point. The initial value of the discharge (6 ml.mn-1) rapidly decreases and reaches its stable value (4 ml.mn-1) after 1 h 30 min. Ca2+, Mg2+, Na+, K+ and HCO3- concentration show erratic variations, but the fact that Na+ and K+ concentration are linked suggests that these fluctuations are related to adsorption-desorption phenomena. It should be noticed that SO42- concentration is not erratic : it decreases from 0.44 meq.l-1 and reaches a stable value (0.08 meq.l-1) after four hours, i.e. much later than the discharge. This proves that the oozing point yields water from two different origins : a small reservoir and a network of connected fractures reaching a given potential boundary, both being initially static.
The statistic study of space variations identifies a mountain aquifer, in the Fongchong zone, and proves that relationships do exist between points which may be located quite far apart. Such relationships should be discussed within the tectonic setting : it has been shown that fracturation develops mainly in three directions (N000-020, N060-080, N120-130); the fracture opening is maximum in the N060-080 direction which corresponds to anormal and strike-slip faults. An open fracture in that direction, joining two points, may explain the relation between them. It also appears that the water composition of springs in the plain, far from the contact fine with the fongchong zone, may be similar to the fongchong one. This proves that water from the fongchong does not only yield springs on the contact fine between fonglin and fongchong. Thus, a groundwater flow occurs from fongchong to fonglin and the hydraulic boundary conditions between the two landforms way be very complex.
Finally, we propose the most probable flow pattern in the area studied. The various elements (unsaturated zone, mountain karst aquifer, fongchong aquifer, deep plain aquifer, surface plain aquifer) appear to be closely linked. Further work is underway, in order to estimate the depth of karstification, to identify the groundwater flow more precisely and to quantify if.