Drinking water provision and quality at the Sahrawi refugee camps in Tindouf (Algeria) from 2006 to 2016.

Drinking water provision has been a constant challenge in the Sahrawi refugee camps, located in the desert near Tindouf (Algeria). The drinking water supply system is itself divided in three zones which pump groundwater from different deep aquifers. It is equipped with reverse osmosis plants and chlorination systems for treating water. The allocation of water supplied to the Saharawi refugees for human consumption in 2016 has been estimated at between 14 and 17 L/person/day on average. This supplied water volume is below recommended standards, and also below the strategic objective of the Sahrawi government (20 L/person/day). Yet the local groundwater resources are huge in comparison with estimated consumption, and hence there is great potential for increasing the supplied volume through effecting improvements in the supply system. The physico-chemical quality of the raw and supplied water between 2006 and 2016 has been assessed according to Algerian standards for human consumption. The raw water of two zones of the supply system presents a very high conductivity and high concentrations of chloride, nitrate, fluoride, sulfate, sodium, calcium, potassium and iodide concentrations of natural origin, which may entail health risks. The treatment of water in a reverse osmosis plant greatly improves its quality and osmosed water met the standards. However, the supply of osmosed and raw water needs to be combined in Zone 1, to avoid an excessive reduction in water volume, and the supplied raw water poses a risk to the health of the refugees. The present study provides an example of a drinking water supply system under extreme drought conditions and in the political and social conditions of a refugee camp. Furthermore, it establishes a reference for supplied water allocation and quality in the Sahrawi refugee camps.

Long-term non-isothermal reactive transport model of compacted bentonite, concrete and corrosion products in a HLW repository in clay.

Radioactive waste disposal in deep geological repositories envisages engineered barriers such as carbon-steel canisters, compacted bentonite and concrete liners. The stability and performance of the bentonite barrier could be affected by the corrosion products at the canister-bentonite interface and the hyper-alkaline conditions caused by the degradation of concrete at the bentonite-concrete interface. Additionally, the host clay formation could also be affected by the hyper-alkaline plume at the concrete-clay interface. Here we present a non-isothermal multicomponent reactive transport model of the long-term (1Ma) interactions of the compacted bentonite with the corrosion products of a carbon-steel canister and the concrete liner of the engineered barrier of a high-level radioactive waste repository in clay. Model results show that magnetite is the main corrosion product. Its precipitation reduces significantly the porosity of the bentonite near the canister. The degradation of the concrete liner leads to the precipitation of secondary minerals and the reduction of the porosity of the bentonite and the clay formation at their interfaces with the concrete liner. The reduction of the porosity becomes especially relevant at t=104years. The zones affected by pore clogging at the canister-bentonite and concrete-clay interfaces at 1Ma are approximately equal to 1 and 3.3cm thick, respectively. The hyper-alkaline front (pH>8.5) spreads 2.5cm into the clay formation after 1Ma. Our simulation results share the key features of the models reported by others for engineered barrier systems at similar chemical conditions, including: 1) Pore clogging at the canister-bentonite and concrete-clay interfaces; 2) Narrow alteration zones; and 3) Limited smectite dissolution after 1Ma.

Identifiability of diffusion and sorption parameters from in situ diffusion experiments by using simultaneously tracer dilution and claystone data.

In situ diffusion experiments are performed in geological formations at underground research laboratories to overcome the limitations of laboratory diffusion experiments and investigate scale effects. Tracer concentrations are monitored at the injection interval during the experiment (dilution data) and measured from host rock samples around the injection interval at the end of the experiment (overcoring data). Diffusion and sorption parameters are derived from the inverse numerical modeling of the measured tracer data. The identifiability and the uncertainties of tritium and (22)Na(+) diffusion and sorption parameters are studied here by synthetic experiments having the same characteristics as the in situ diffusion and retention (DR) experiment performed on Opalinus Clay. Contrary to previous identifiability analyses of in situ diffusion experiments, which used either dilution or overcoring data at approximate locations, our analysis of the parameter identifiability relies simultaneously on dilution and overcoring data, accounts for the actual position of the overcoring samples in the claystone, uses realistic values of the standard deviation of the measurement errors, relies on model identification criteria to select the most appropriate hypothesis about the existence of a borehole disturbed zone and addresses the effect of errors in the location of the sampling profiles. The simultaneous use of dilution and overcoring data provides accurate parameter estimates in the presence of measurement errors, allows the identification of the right hypothesis about the borehole disturbed zone and diminishes other model uncertainties such as those caused by errors in the volume of the circulation system and the effective diffusion coefficient of the filter. The proper interpretation of the experiment requires the right hypothesis about the borehole disturbed zone. A wrong assumption leads to large estimation errors. The use of model identification criteria helps in the selection of the best model. Small errors in the depth of the overcoring samples lead to large parameter estimation errors. Therefore, attention should be paid to minimize the errors in positioning the depth of the samples. The results of the identifiability analysis do not depend on the particular realization of random numbers.