Chlorinated solvents source identification by nonlinear optimization method.

In this work, chloride ions were used as conservative tracers and supplemented with conservative amounts of chloroethenes (PCE, TCE, Cis-DCE, 1,1-DCE), chloroethanes (1,1,1-TCA, 1,1-DCA), and the carbon isotope ratios of certain compounds, the most representative on the sites studied, which is a novelty compared to the optimization methods developed in the scientific literature so far. A location of the potential missing sources is then proposed in view of the balances of the calculated mixing fractions. A test of the influence of measurement errors on the results shows that the uncertainties in the calculation of the mixture fractions are less than 11%, indicating that the source identification method developed is a robust tool for identifying sources of chlorinated solvents in groundwater.

Biomass partitioning of plants under soil pollution stress.

Polluted sites are ubiquitous worldwide but how plant partition their biomass between different organs in this context is unclear. Here, we identified three possible drivers of biomass partitioning in our controlled study along pollution gradients: plant size reduction (pollution effect) combined with allometric scaling between organs; early deficit in root surfaces (pollution effect) inducing a decreased water uptake; increased biomass allocation to roots to compensate for lower soil resource acquisition consistent with the optimal partitioning theory (plant response). A complementary meta-analysis showed variation in biomass partitioning across published studies, with grass and woody species having distinct modifications of their root: shoot ratio. However, the modelling of biomass partitioning drivers showed that single harvest experiments performed in previous studies prevent identifying the main drivers at stake. The proposed distinction between pollution effects and plant response will help to improve our knowledge of plant allocation strategies in the context of pollution.

Using 1,1,1-Trichloroethane degradation data to understand NAPL dissolution and solute transport at real sites.

Analytical and numerical models describing the evolution of contaminant concentrations in the plume associated with the dissolution of NAPL source and degradation processes were presented in the literature. At real sites and particularly in complex aquifers like chalk, it is difficult to understand how the sources of contaminants evolve with time. 1,1,1-Trichloroethane (1,1,1-TCA) is one of the few compounds with a well-known hydrolysis constant, that can help to improve knowledge of the contaminant sources and transport rates of dissolved contaminants in groundwater by dating the spill. In this work, different scenarios that could explain the evolution of the concentrations of 1,1,1-TCA and its degradation product 1,1-Dichloroethene (1,1-DCE) at a real contaminated site were investigated by analytical and numerical modelling. The results show that (1) the peaks of concentration time series do not correspond to a single contamination event even in the case of a complex medium, (2) the multiphasic behavior of the concentration time series is dictated by the dissolution in a heterogeneous medium, and (3) the persistence of the concentrations can arise from a small residual organic phase or transport in dual domain medium.

Status of Dieldrin in vegetable growing soils across a peri-urban agricultural area according to an adapted sampling strategy.

Since the fifties, organochlorine pesticides (OCPs) had been used in agriculture to protect vegetables. Two decades after their ban by the Stockholm convention in 2001, OCPs are still present in agricultural soils inducing vegetable contamination with concentrations above Maximum Residue Level (MRL). This is a major concern for a 5 km2 peri-urban vegetable growing valley located in the south west of France. In the present work, the sampling method was developed to clarify the spatial distribution of one OCP, Dieldrin, and its relationship with soil properties at the scale of study area. A total of 99 soil samples was collected for physicochemical analyses and Dieldrin concentrations. Results show Dieldrin concentrations in soils up to 204 µg kg-1. The horizontal distribution of this pesticide is heterogeneous at the study area scale but homogeneous in each reference plot studied. About 85% of the contamination was located in the top soil layers (0-40 cm depth), but Dieldrin may still be quantified at a depth of 80 cm. Among all soil physicochemical parameters analysed, SOM was the most significantly related (P < 10-4) with Dieldrin concentrations, once different grain size fractions were considered. Moreover, results indicate a 33 times higher Dieldrin concentration and/or extractability for coarse sand than for other grain size fractions. These results show that the developed sampling method is adapted for the study area scale as it helps understanding the factors influencing the spatial distribution of Dieldrin. Historical amendments are the predominant factor for the horizontal contamination and deep ploughing for the vertical contamination. Also, the variations of coarse sand repartition in soils prevents identification of relationships between SOM and Dieldrin contamination in bulk soil. Further investigation is required to explain these relationships but these results highlight why no clear relationship between OCPs and SOM was previously identified.

Effect of NAPL mixture and alteration on 222Rn partitioning coefficients: Implications for NAPL subsurface contamination quantification.

Soils and groundwater are often contaminated by complex organic mixtures also called Non Aqueous Phase Liquids (NAPLs). Several techniques such as drilling, monitoring of soil gas or injection of tracers are traditionally used to quantify NAPLs in aquifers but are complex to perform. The use of natural soil gas such as 222Rn could be an easy and cheap alternative. This method requires the knowledge of the radon NAPL-water partitioning coefficients (Kn-w.). Once spilled on soil, NAPL will undergo degradation (evaporation, effects of sun light among others) and this degradation could impact the Kn-w. This study aims at investigating the partitioning coefficients of complex NAPLs such as commercial diesel fuel and gasoline in relation to degradation such as evaporation and UV-degradation. For that purpose, batch experiments and GCMS investigations were carried out. The results show different Kn-w for the commercial diesel fuel (60.7 ± 6.1) and gasoline (37.4 ± 5.6). The results also show different Kn-w behaviors in relation with degradation. Degraded diesel fuel display opposite Kn-w values (74.8 ± 7.5 and 25.1 ± 2.5 for UV degraded and evaporated diesel fuel, respectively), compared to fresh one. Degraded gasoline shows no significant variations of the Kn-w compared to fresh one. The molecular investigation reveals the removal of the most volatile fraction for the evaporation treatment, whereas UV-degradation do not have pronounced effects on the chromatogram pattern. For the gasoline molecular investigation, no difference is observed between the treatments excepted a very slight removal of the lightest compounds under evaporation. These results show that NAPL degradation have effects on the Kn-w for diesel fuel and no significant effects for gasoline, at least with these degradation paths. This Kn-w variation will have in fine effects on 222Rn activity interpretation and NAPL subsurface quantification.

Laboratory-scale experimental and modelling investigations of 222Rn profiles in chemically heterogeneous LNAPL contaminated vadose zones.

The potential of LNAPL delineation by 222Rn soil-gas monitoring in a chemically heterogeneous vadose zone was investigated in this study based on laboratory (batch and columns) experiments and numerical modelling. An enhanced version of the MIN3P reactive transport code was used to simulate Rn transport in both uncontaminated and NAPL-contaminated vadose zones and results were validated against analytical solutions and laboratory experiments. Results show that 222Rn activity profiles are mainly controlled by porous media 222Rn production, vadose zone fluid saturations and especially the type and distribution of NAPL in contaminated areas. The results also show that decreases in 222Rn activity and variations in activity gradients provide evidence for the presence and saturation of NAPL. This study demonstrates that LNAPL delineation via 222Rn gas surveys at contaminated sites works best, if gas measurements extend as deep as possible and include regions where 222Rn activity decreases due to elevated NAPL content. In addition, collection and analysis of depth-discrete gas samples allows the characterization of vertical NAPL distribution based on the 222Rn activity gradient. The determination of 222Rn production in the unsaturated zone, as well as water capillary pressure curves are of key importance in enabling the discrimination of an uncontaminated from a NAPL-contaminated area.

LNAPL source zone delineation using soil gases in a heterogeneous silty-sand aquifer.

Source delineation of hydrocarbon contaminated sites is of high importance for remediation work. However, traditional methods like soil core extraction and analysis or recent Membrane Interface Probe methods are time consuming and costly. Therefore, the development of an in situ method based on soil gas analysis can be interesting. This includes the direct measurement of volatile organic compounds (VOCs) in soil gas taken from gas probes using a PID (Photo Ionization Detector) and the analysis of other soil gases related to VOC degradation distribution (CH4, O2, CO2) or related to presence of Light Non-Aqueous Phase Liquid (LNAPL) as (222)Rn. However, in widespread heterogeneous formations, delineation by gas measurements becomes more challenging. The objective of this study is twofold: (i) to analyse the potential of several in situ gas measurement techniques in comparison to soil coring for LNAPL source delineation at a heterogeneous contaminated site where the techniques might be limited by a low diffusion potential linked to the presence of fine sands and silts, and (ii) to analyse the effect of vertical sediment heterogeneities on the performance of these gas measurement methods. Thus, five types of gases were analysed: VOCs, their three related degradation products O2, CO2 and CH4 and (222)Rn. Gas measurements were compared to independent LNAPL analysis by coring. This work was conducted at an old industrial site frequently contaminated by a Diesel-Fuel mixture located in a heterogeneous fine-grained aquifer. Results show that in such heterogeneous media migration of reactive gases like VOCs occurs only across small distances and the VOC concentrations sampled with gas probes are mainly related to local conditions rather than the presence of LNAPL below the gas probe. (222)Rn is not well correlated with LNAPL because of sediment heterogeneity. Oxygen, CO2, and especially CH4, have larger lengths of diffusion and give the clearest picture for LNAPL presence at this site even when the gas probe is somewhat distant.