Hydrogeochemical balance sheet of natural and anthropogenic impacts onto Orleans valley karstic network performed with major elements : the "dynamic confinement" model quantification.
Résumé
The Orleans valley aquifer comprises both the alluvia of the Loire river and its underlying calcareous stratum. This aquifer is fed by river recharge, thanks to a substantial karstic network in its calcareous part. The main outlets of the aquifer are the Loiret springs, including the famous "le Bouillon" spring. As a result, entries and exits of Orleans valley watertable make a natural observatory, allowing study of the transit of the chemical species inside the aquifer. Since 1997, this natural observatory has been improved with the installation of 52 piezometers (37 in the alluvial aquifer and 15 in the carbonate aquifer) within an alluvial quarry located in the middle of Orleans valley. Tracer experiments, carried out in this extended observatory, have shown that the porous calcareous and alluvial part of the aquifer constitute a "dynamically confined system". As a result, the hydrochemical input of the porous domain of the aquifer to the karstic flow must be negligible. The aim of this study is to confirm this theory with the use of major elements as large-scale temporal and spatial tracers of these exchanges. At "le Bouillon" karstic spring, the Na+, K+, Mg2+, Cl– and SO42– concentrations are closely correlated to those of the Loire river if a 3–4 day time lag is considered. This indicates a quasi-conservative transit of these elements in the karst. Conversely, calcite dissolution accompanying the organic matter biodegradation induces significant enrichments in Ca2+, HCO3– and NO3– (mean annual concentrations of which are, respectively, 27.0, 87.8 and 4.9 mg.L–1 in the Loire river and 37.3, 127 et 7.3 mg.L–1 at "le Bouillon" spring). After fertiliser spreading, the alluvial waters are highly enriched in NO3–, Cl–, SO42– (respectively 67.2, 24.0, 57.5 mg.L–1) compared to the Loire river (respectively 5.5, 12.7, 17.5 mg.L–1). The anthropogenic input is insignificant for Na+, of which the average concentration in the alluvial watershed (11.7 mg.L–1) remains close to the Loire river (12.9 mg.L–1). The alluvial watershed is depleted in K+ (1.3 mg.L–1) with respect to the Loire river (3.7 mg.L–1) and correlatively enriched in Mg2+ (17.0 mg.L–1 against 5.0 mg.L–1). High major element concentrations are measured in several calcareous piezometers confirming that vertical flows occur between the alluvial and calcareous parts of the aquifer. Furthermore, enrichment heterogeneity in those two strata is induced by a dynamic redistribution, with no significant leaching of anthropogenic inputs which were previously homogeneously spread. This redistribution is pulsed by ascents of the Loire river, impacts of which on the watershed are clearly identified on Mg/K-Na/K diagrams showing a main K {leftrightarrows} Mg exchange between Loire water and clays minerals. Taking into account average K and Mg concentrations in the different parts of Orleans valley's watershed, the volume of porous aquifer water brought to the karstic network flow mean estimated is 2.4 % of the total volume which transits between the Loire and the "le Bouillon" spring, showing the dynamic confining action of the aquifer porous domain. Taking into account more precisely seasonal river Loire and spring composition variation, these inputs can be more precisely established : 1.6% during winter and 1.2% during summer at "Le Bouillon" spring; 2.4% during winter and 3.9% during summer at "La Pie" spring. But such a weak global contribution of the porous domain accounts for 10% nitrate composition of the karstic springs. Seasonal spring nitrate composition balance is clearly explained by 60% river Loire, 30 % organic matter oxydation – carbonate dissolution and 10% porous domain inputs during winter, and 30% river Loire, 60% organic matter, – carbonate dissolution and 10% porous domain inputs. Same calcium mass balance calculations point out the necessity of CO2 winter complementary input by local rain fall penetrations.