GEOCHEMICAL STUDY OF A CRATER LAKE - LAKE PAVIN, PUY-DE-DOME, FRANCE - CONSTRAINTS AFFORDED BY THE PARTICULATE MATTER DISTRIBUTION IN THE ELEMENT CYCLING WITHIN THE LAKE

Suspended particles concentrations and settling particle fluxes collec ted by sediment traps were analyzed along a vertical profile in Lake P avin, France at the same time as dissolved element concentrations, whi ch are reported elsewhere (Viollier et al., 1995). Results for iron, m anganese, cobalt, vanadium, molybdenum, uranium, barium, rubidium and cesium are presented here, Particulate concentrations of all elements increase significantly just above the sediment-water interface. Mangan ese, iron and barium particulate concentrations exhibit a peak just ab ove the oxic-anoxic boundary. Molybdenum and uranium particulate conce ntrations increase drastically below the oxic-anoxic boundary. Mangane se and cobalt fluxes are at a maximum at 55 m depth, above the oxic-an oxic boundary. For all other elements, fluxes are at a maximum at 85 m depth, just above the bottom of the lake. In all cases settling parti cles fluxes are greater than buried fluxes, indicating that a dissolut ion reaction occurs at the bottom of the lake. Particulate iron is ess entially present as Fe(III) oxides in the mixolimnion, as Fe(II) phosp hates in the monimolimnion and as pyrite in the sediment. The change i n the Fe-bearing mineral leads to either important remobilisation or i mmobilisation of trace element into or from solution. Comparison of fl uxes in the sediment trap at 85 m depth with dissolved concentrations gradients above the bottom leads to the rejection of the previous hydr odynamic model of the lake. The fluxes and gradient data are consisten t with the tritium data only if the arrival of sublacustrine waters oc curs in the mixolimnion. Residence times of waters in monimolimnion in the model derived from the present study are about 20 times greater t han the previous values. The dissolved element concentrations profiles are explained only by eddy diffusion with a coefficient varying from about 20 m(2) a(-1) at the bottom of the lake to 0.2 m2 a(-1) at the m ixolimnion-monimolimnion boundary. Transport parameters obtained in th is way allow a quantitative description of the element cycles within t he lake and an estimation of the elemental ratio sedimentation versus exportation by the outlets. (C) 1997 Elsevier Science B.V.