Thesis: Cécile Grapeloup

Thesis Director: Lucilla BENEDETTI

Summary of the thesis: Chlorine is a hydrophilic element considered to be poorly retained in soils. It has several isotopes, including 36Cl. The latter has a natural cosmogenic origin, considered constant over time, and an anthropogenic origin, mainly linked to the nuclear tests of the 1950s to 1970s, whose fluxes were three orders of magnitude greater than the natural flux. A small fraction of the Cl brought to the soil is retained in the organic matter. However, the extent of this retention, the factors determining it and the duration of this retention were little known at the beginning of this thesis. By analysing 130 samples from the French Soil Quality Measurement Network (RMQS), we have shown that the inflow of Cl and 36Cl determines the formation of Cl and 36Cl stocks in soils in order 1. By normalizing these inputs by the fluxes, we show that the Corg stocks in soils were responsible for the formation of Cl and 36Cl stocks in order 2. At this scale, no impact of the soil occupation mode on the formation of Cl and 36Cl stocks has been demonstrated. We then sought to highlight this impact through the analysis of paired soils (forest-grassland or forest-crop) for three types of soil with contrasting functioning (cambisol, luvisol, podzol). This approach showed in agriculture: 1- non-systematic Cl contamination probably due to potassium fertilisation; 2- a loss of 36Cl probably due to erosion; 3- a decrease in 36Cl migration in podzol due to a reduction in podzolisation processes as a result of liming (increasing the soil pH). Using soil labelling with 36Cl flux from nuclear tests, we determined 36Cl retention rates and times. We show that the retention rate is low, below 5%. It decreases exponentially with depth (except for podzol) and varies with soil type (between 0.5% at the surface for podzol and 4.5% for cambisol). We show that this variability is related to organic carbon and that the 36Cl retention times correspond to the ages of organic carbon estimated by stable carbon isotopes (δ13C). These younger ages than those obtained by 14C dating are representative of the dynamic organic carbon fraction of soils. This method could be used for soils for which the stable carbon isotope dating approach (no dated C3/C4 vegetation change) is not possible. This retention time seems to vary according to soil type, with podzol showing much longer retention times for much lower retention rates. Application of the model to all RMQS forest soils would confirm this observation.

Thesis summary : Chlorine is a hydrophilic element considered poorly retained in soils. It has several isotopes including 36Cl. The latter has a natural cosmogenic origin considered constant over time and an anthropogenic origin mainly related to nuclear tests from the 50's to the 70's whose fluxes were three orders of magnitude higher than the natural flux. A small fraction of the Cl brought to the soil is retained in the organic matter. However, the importance of this retention, the factors determining it and the duration of this retention were little known at the beginning of this thesis. Through an analysis of 130 samples from the French Soil Quality Measurement Network (RMQS), we have shown that Cl and 36Cl influxes determine the formation of Cl and 36Cl stocks in soils to the order of 1. By normalizing the stocks by the fluxes, we show that the stocks of Corg in the soils were responsible for the formation of the Cl and 36Cl stocks at order 2. At this scale, no impact of the land use mode on the formation of Cl and 36Cl stocks has been demonstrated. We then sought to highlight this impact through the analysis of paired soils (forest-grassland or forest-crop) for three types of soil with contrasting functioning (cambisol, luvisol, podzol). This approach showed in agriculture: 1- non-systematic Cl contamination probably due to potassium fertilization; 2- a loss of 36Cl probably by erosion; 3- a decrease of 36Cl migration in podzol due to a reduction of podzolization processes because of liming (increasing the soil pH). Using soil labeling by 36Cl flux from nuclear tests, we determined 36Cl retention rates and times. We show that the retention rate is low, less than 5%. It decreases exponentially with depth (except for podzol) and varies with soil type (between 0.5% at the surface for podzol and 4.5% for cambisol). We show that this variability is related to organic carbon and that 36Cl retention times correspond to organic carbon ages estimated by stable carbon isotopes (δ13C). These younger ages than those obtained by 14C dating are representative of the dynamic organic carbon fraction of soils. This method could be used for soils for which the stable carbon isotope dating approach (no dated C3/C4 vegetation change) is not possible. This retention time seems to vary according to soil type, with podzol showing much longer retention times for much lower retention rates. Application of the model to all RMQS forest soils would confirm this observation.