Proposition de protocole de traçage adapté au karst de la craie

Proposition of tracing test protocol in karstic chalk

Ginger Environnement, 16 allée Prométhée, 28000 Chartres

Reçu : 18 Décembre 2006
Accepté : 18 Mars 2007

Résumé

La craie est le plus vaste aquifère karstique français et de nombreux traçages y sont réalisés. Son comportement hydrogéologique spécifique résulte de la présence d’un épais recouvrement argilo-limoneux, d’une forte porosité matricielle et d’un potentiel de karstification plus faible que dans les calcaires. L’analyse de 144 traçages dans la craie normande permet de comparer la vitesse de transfert et la fréquence de restitution selon le « système-traçage » en question. Du fait de l’existence à la fois d’un karst d’introduction et de restitution, les systèmes perte-source sont les plus performants. Par contre les systèmes forage-forage ou ceux impliquant un point d’injection superficiel montrent des relations plus lentes et probablement soumises à des phénomènes d’adsorption à la traversée des formations superficielles et de la zone non saturée. Les principaux traceurs utilisés sont comparés entre eux sur le plan de leurs caractéristiques intrinsèques (pureté, adsorbabilité) et de leur potentiel de détection (bruit de fond, seuil analytique). On propose ainsi un protocole de traçage standardisé qui tient compte à la fois du type de système-traçage, de la nature du traceur et de la méthodologie analytique employée.

Abstract

Chalk of northern France is the main french karstic aquifer with an outcrop area of 70,000 km2. Its hydrogeological behaviour results from specific characteristics compared with limestone karsts : thick cover of silto-argileous terranes (10-40 m), high matricial porosity (20-40%), relatively low potential of karstification due to low altitudinal gradients between plateaus and valleys and to moderated tectonical background. Many tracing tests have been realized in chalk and 144 of them are analysed in the Normandie county, western part of French chalk. Restitution only occurs in 49% of tracer tests but it differs significatively depending on the « tracing-system » : high transfer speed (144 m/h) and positive results (62 %) are observed between sinkholes (locally called « bétoires ») and springs i.e. when exist an « introduction karst » on the plateaus and a « restitution karst » in the valleys. Although low transfer speed (12 m/h) and low positive results (22 %) are noted between surface injections (trench, ditch, creek) and borehole restitutions i.e. with lack of karstic features at both the extremities of the aquifer system. One of the main problems in tracing tests is due to the nature and the quantity of dye used. Principal dyes used are fluorescent ones (uranine, sulforhodamine, tinopal, naphtionate) and ionic ones (iodide, bromide, lithium, zinc). Standard dosage in Normandie is 1.4 kg/km of uranine but it must be adapted for other dyes according to: – at the injection stage, content of pure tracer in commercial product varies in the range 16-90 % while adsorption potential through unsaturated zone is between 0-35 % ; – at the restitution stage, background of classical dyes is between 0.01-0.10 µg/l (due to interferences for fluorescent dyes and to natural traces for ions) and analysis detection level is between 0.01-2.18 µg/l. An equivalence of dyes dosage is proposed on the basis of theses parameters : dosage might be related to uranine for other coloured fluorescent dyes, twice for uncoloured fluorescent ones, five times for bromide, ten times for iodide, more than hundred times for zinc and tenth of thousands for chloride. Use of chloride must be forbidden especially in Normandie because of its high background due to geochemistry of chalk and proximity of sea. We propose a standardised protocol notably based on the equivalence of dyes dosage and other points for the improvement of tracing tests: injection of a minimal quantity of water with the tracer, minimal duration of tracking with possibility of prolongation, minimal number of samp les collected and analysed and previou s measurement of background.