Abstract:
Turbid water can be the source of important sanitary problems in karstic regions. It is the case of the Pays de Caux, in Haute Normandie, where the main resource in drinking water is provided by the chalk aquifer. In the case of the typical binary karst of the Pays de Caux, turbidity results from the input in sinkholes of turbid surface water induced by erosion on the plateaus. At some spring tappings, water may be very turbid in period of intense rainfall. The turbidity observed at a karstic spring is a complex signal which contains a part of direct transfer and a part of resuspension of the particles being transported. The aim of this study is turbidigraph separation, which would permit to distinguish the direct transfer and resuspension components of the turbidigraph. These two components are separated by comparing the elementary surface storm-derived water fluxes and elementary turbidity signals at the spring. The procedure takes place in three phases: (i) spring hydrograph separation by means of a two components mixing model (surface water and karstic groundwater) using specific electrical conductivity, (ii) decomposition of storm-derived water flux and turbidity thanks to the second-derivative method, (iii) comparison of the transfer times (~modal times) of the elementary turbidity and surface water flux signals, respectively. The mass corresponding to direct transfer, computed after signal decomposition, is then used to re-calculate a particle recovery rate, which passes so from 51+/-4 to 37+/-3%. Relations between particle flux and hydrodynamics show that resuspension can be either the fact of the dynamics of the introduction system, or that of the chalk karstic aquifer in general (case of resuspension not associated to surface water flux). In this sense, evolution of particle flux (and consequently of turbidity) can be also a marker of the karst structure.
