In France, municipal solid waste incineration bottom ashes are widely used in road building. Leachates released from these materials may constitute a hazard for aquatic ecosystems. In order to assess this risk, a study of bottom ash leachates toxicity was carried out on 100 litre indoor microcosms. The 100 litre microcosms are glass tanks filled with a lake sediment and synthetic water, and inoculated with various organisms: microalgae, duckweeds, rooted macrophytes, cladocerans, pond snails, amphipods and chironomids. The microcosms were contaminated twice so as to reach a final leachate concentration of three percent. The contamination led to increased salt and metal concentrations. Among cladocerans, Daphnia magna was sensitive to contamination although recovery was observed, whereas Simocephalus vetulus and Ceriodaphnia dubia were not. Results regarding the amphipod Hyalella azteca depended on major exposure, via the water column (encaged organisms) or via the sediment (free organisms). Chironomids emerged similarly in control and contaminated systems. Among pond snails, leachates impaired survival and grazing activity of Lymnea stagnalis, whereas Physa sp. was not sensitive. Growth of all primary producers tested in the microcosms was not impaired. Copper may explain most of the observed effects. Due to the fact that concentrations studied here seem to be an overestimation of real field concentrations, and that bottom ash leachates characteristics decrease with time, it was concluded that the risks for lentic aquatic ecosystems would be minimal.
In France, most domestic wastes are eliminated by incineration. Incineration generates secondary wastes called municipal solid waste incineration (MSWI) bottom ashes. These materials are widely used in road building. Leaching of rainwater through MSWI bottom ashes produces leachates which have the following characteristics: a high pH, presence of organic matter (mainly cellulose and lignin + organic contaminants such as phthalates, polycyclic aromatic hydrocarbons (PAHs), polycyclic biphenyls, dioxin-like compounds), heavy metals (e.g., copper, chromium, zinc, lead), and salts (e.g., Cl−, K+, Mg2+, Ca2+, SO42−). The use of MSWI bottom ashes is regulated by a ministry circular (MATE, 1994). They must display minimal contamination levels and must be used in conditions where aquatic ecosystems will not be at risk from chronic discharges of released percolates. However, little is known about the ecological impacts of these effluents, especially on lentic ecosystems. Ecotoxicological assessment of liquid wastes is classically based on single-species tests. Due to the well-known limitations of this approach, efforts of some ecotoxicologists focused on more relevant approaches such as indoor microcosm testing, which allows study of the effects of pollutants in conditions which take into account the complexity of real aquatic ecosystems (Kimball and Levin, 1985; Forbes and Forbes, 1997), such as the simultaneous presence of organisms belonging to various trophic levels, inhabiting the sediment or the water column and interacting with each other and with their habitat. This approach was successfully applied in our laboratory on various issues using small (volume, 2 l) microcosms (Clément and Cadier, 1998); PAH bioavailability (Verrhiest et al., 2000); and ecological risk assessment of dredged sediment (Babut et al., 2002; Clément et al., 2004). This paper reports the results of a microcosm assay on MSWI bottom ashes percolates using 100 l aquaria. These microcosms consisted of glass tanks filled with lake sediment, synthetic water and various organisms (rooted macrophytes, microalgae, duckweeds, daphnids, amphipods, chironomids, pond snails). Systems were allowed to grow for 2 weeks, then percolates were added so as to reach a concentration of 1% (v:v). At 20 d, percolates were added to a concentration of 3%. The systems were monitored during 2 mo following contamination. Experiment duration was 95 d (Table 1). Survival, growth and reproduction of organisms were assessed.
Materials and methods
Leachates were produced using a laboratory lysimeter consisting in a polyethylene rectangular tank (1 m2) filled with 200 kg MSWI bottom ashes taken from an incinerator located in the Alps (France). Ashes were compacted in conditions similar to field road building conditions (Proctor optimum, Arquie and Morel, 1988). Five l of demineralized water were added daily to the ash material. Rain water was not used for the following reasons: presence of unknown pollutants, variability of its physico-chemical characteristics and difficulty of storage. The following day, leachates (ca. 5 l) were collected at the bottom of the lysimeter using a tap. Leachates were stored at 4°C and used as quickly as possible. At the end of a week, all 5-l sub-fractions were mixed together to constitute one fraction to be tested. Eight weekly fractions were produced (A, B, C, D, E, F, G, H). A composite leachate made of fractions A, B, C, D, E, F, G, H stored at 4°C was used in experiments. The physico-chemical composition of leachates are displayed in Table 2.
Six microcosms were set up as shown in Figure 1. The microcosms were glass rectangular tanks (0.70 m × 0.45 m × 0.42 m) that contained natural lake sediment (12 kg fresh weight) and a modified OECD medium (100 l) with salts, vitamins, micronutrients and macronutrients (Clément and Cadier, 1998) allowing development of invertebrates and primary producers. The systems contained compartments, cages and beakers that allowed isolation of some organisms (such as pond snails which thus could not eat plants), facilitated the monitoring and the sampling of other organisms (daphnids, benthic organisms and plants), or assessment of the effects in various exposure conditions (water/sediment, duration).
The systems were inoculated with the following organisms: cladocerans (3 species: Daphnia magna, Ceriodaphnia dubia, Simocephalus vetulus), microalgae (1 species: Pseudokirchneriella subcapitata), amphipods (1 species: Hyalella azteca), chironomid larvae (1 species: Chironomus riparius), pond snails (2 species: Lymnaea stagnalis, Physa sp.), rooted macrophytes (2 species: Elodea canadensis, Groenlandia densa), duckweeds (2 species: Lemna minor, Spirodela polyrhiza). Organisms were chosen as representative of the main trophic levels of aquatic lentic ecosystems: microalgae and rooted higher plants as primary producers, cladocerans as pelagic primary consumers, amphipods and chironomids as detritivore benthic invertebrates. Pond snails were chosen for their capacity to inhabit the water column as well as the sediment, and to control periphyton growth on tank walls. Cladocerans (except Ceriodaphnia dubia), microalgae, chironomids and duckweeds were bred in conditions described in Clément and Cadier (1998). Ceriodaphnia dubia was provided on request by another laboratory (Polden-Insavalor, Lyon, France) and used within 48 h. Ceriodaphnids were bred at 23± 2°C and a light regime of 16 h light:8 h dark, in a medium made with 50% commercial mineral water (Evian) and 50% used medium. They were fed 3 times per week with a mixture of fish food (Seramicron, 5 mg l−1), Chlorella vulgaris (24× 107 cells l−1) and Pseudokirchneriella subcapitata suspension (12× 107 cells l−1). The medium was renewed once a week.
Pond snails were collected at the river Rhône-Ain confluence (a pristine site upstream from Lyon) and bred several months in the lab, in a polyethylene tank filled with running ground water (10 l min−1) and fed with fresh lettuce. Macrophytes were collected in non-contaminated experimental ponds in Chambéry (ESIGEC, Savoie, France) and cultured in polyethylene 15-l aquaria filled with the same lacustrine sediment and the same medium as the ones used in microcosm assays. Characteristics and numbers of organisms initially introduced in the systems are displayed in Table 3. Chironomid larvae and amphipods were also introduced in small polyethylene cages filled with the same sediment as in the large compartment, positioned above the sediment of the big compartment and equipped with legs in order to minimize contact with the sediment (Figure 1). So as to monitor survival (maximum 10 d exposure) of organisms exposed to leachates via the water column, 250 ml glass beakers filled with overlying water and containing various organisms (10 individuals of each cladoceran species and 10 amphipods) were placed on the sediment. The beakers were capped with a synthetic tissue with a 200-μm mesh size which allowed water and microalgae transfer but prevented organism escape. In addition, water inside beakers was completely renewed on each daily measurement.
Overlying water in microcosms was not renewed but evaporation was compensated by weekly addition of ca. 5 l demineralized water per microcosm. Oxygen level was maintained at more than 80% using constant aeration by air bubbling with Pasteur pipets connected to an aquarium pump. The room temperature was kept constant so that the water temperature ranged between 20 and 22°C. The systems received 16 h d−1 2000 lux provided by two daylight fluorescent tubes (36 W) above each aquarium. Walls of aquaria were covered with a black plastic film so as to minimize periphyton growth. Ground fish food flakes (TetraMin) were deposited every day at the surface of sediment (30 mg d−1) to improve growth of benthic organisms (chironomids and amphipods). As a matter of fact, in the absence of added food, chironomids survive but do not emerge (Péry et al., 2002).
Three microcosms were used as controls and three were used as contaminated systems. After addition of water and sediment, and cross-mixing of waters (1 or 2 weeks) to ensure similar initial conditions in all systems (OECD, 1996; Barry and Logan, 1998), organisms were added. Microcosms were then monitored during a 1-mo period to allow development of populations and stabilization of physico-chemical conditions. Three microcosms were once contaminated with 1 l of the composite leachate and, 20 d later, with 2 l of the same leachate. The second addition was made due to the absence of clear effects following the first contamination. Survival, growth and reproduction of organisms were monitored during ca. 50 or 60 d after the first contamination. The measurements carried out on the systems are displayed in Table 4. Cladocerans in small compartments were monitored twice a week and neonates were discarded after counting, whereas populations were maintained in the large compartment. They were collected by siphoning and counted either totally or after sub-sampling when numbers exceeded 1000 individuals. Lymnea stagnalis snails were monitored by marking individuals with a number written on the shell with a permanent ink. Chironomid adults were trapped by covering the tanks with a net maintained in a wooden frame.
As a preliminary step to the microcosm assay, a D. magna (neonates) immobilization test on the composite leachate (mixture of leachates A to H, see leachate production) was launched following a standard protocol (ISO, 1989). In this 24 h assay, the pH of the leachate was not modified, and the following concentrations were tested: 1, 2, 4, 8 and 16%.
During the microcosm assay, another daphnid test was performed on the same leachate (concentrations: 1, 2 and 3%) and on the microcosms overlying waters sampled on day 58. The aim of this daphnid test was to see whether leachate toxicity was attenuated after 8 d following introduction in microcosms.
The amphipod H. azteca was also exposed 24 h and 5 d to the composite leachate (1–16%), following the principle of the daphnid standard test.
Data were analysed with Statview F-4.5. Significant differences (p < 0.05) between means were determined using the t-test.
Physico-chemical conditions in microcosm assays
The addition of leachates led to a significant increase of electrical conductivity. An increase of 35 μ S cm−1 per % of leachate was observed. This increase was mainly due to addition of chlorides, sulphates, sodium and potassium brought with the leachate (Figure 2). The pH of overlying waters varied between 7.9 and 8.4 during the course of both assays; leachate additions led to an increase of only 0.02 to 0.11 pH units, owing to the buffering capacity of the systems. Due to the composition of leachates, the dissolved organic carbon (DOC) content of the water column increased in treated microcosms, but only moderately (maximum 3–4 mg C l−1 in both assays). The contamination also induced an increase of overlying water copper concentration (Figure 3). The composition of the sediment was not modified by contamination. Pore waters of contaminated microcosms were only slightly more saline (Cl−, Na+, K+) and organic (DOC) at the end of assay, whereas no differences were noted for metals.
Effects on organisms
Daphnia magna was sensitive to the composite leachate at concentrations > 1% in the single-species test (Table 5). Note that, in this test, pH of the solutions varied between 7.6 and 8.5. Daphnid populations developed well during the pre-contamination period and reached 2400 individuals in some big compartments. However, a rapid decline and a high variability were observed after 22 d, hence daphnids were totally replaced before contamination (100 neonates in large compartments). Development in control microcosms was lower than in the pre-contamination period, and variability was still high, but significant effects on daphnid population were noted following the second contamination (nominal concentration, 3%; Figure 4). Survival and reproduction in control systems were on average a little higher in small compartments without sediment. Although variability was also high in small compartments, an effect of leachates following the first contamination was observed in both types of compartments. Effects of leachates were confirmed in daphnids exposed to the water column in submerged beakers (Table 6). Effects appeared only after the second contamination (day 50), but were attenuated with time after 7 d. This attenuation was confirmed by the single-species tests performed on day 58 on the composite leachate and on microcosm overlying waters; these were not toxic whereas the leachate was toxic at 2%. In addition, the leachate concentration of 3%, equivalent to the nominal concentration in treated microcosms, induced 55–70% mortality among daphnids.
Ceriodaphnia dubia populations developed well, reaching almost 10000 individuals in large compartments. Contrary to D. magna, Ceriodaphnia dubiawas not affected by leachate contamination. This was confirmed by the good survival of individuals exposed as long as 16 d in submerged beakers (Table 6).
Simocephalus vetulus was more difficult to collect by siphoning, due to the fact that it stands most often on the walls or any surface, and this may explain variability observed for results. However, populations reached high numbers in big compartments (3000 individuals). The leachates had no effect, and there was even a tendency to higher development in contaminated microcosms. The absence of toxicity was confirmed by results on Simocephalus vetulus exposed in submerged beakers (Table 6).
When exposed via the water column (Table 6), the amphipod H. azteca showed a higher sensitivity than D. magna, with high mortalities as soon as the first contamination (day 30). This higher sensitivity does not corroborate results of single-species tests on the composite leachate, which show a slightly lower sensitivity of H. azteca (Table 5). Hyalella azteca exposed in cages 10 to 17 d was not sensitive at all to leachates (Table 7). Amphipod populations in the big compartments developed moderately within the 80 days of experiments, the number of young individuals being low, suggesting that conditions were not optimal for these organisms. Results suggest an effect of contamination (44± 13 individuals in contaminated microcosms versus 125± 65 in control microcosms).
Emergence rates of Chironomus riparius reached 72–75%, and contaminations had no effect on this endpoint (Figure 5). Due probably to non optimal conditions, emergence in these microcosms was a bit lower than in single-species chironomid tests and in small laboratory microcosms, where emergence rate is most often > 80% (Clément and Cadier, 1998; Bonnet, 2000; Clément et al., 2004). Results of chironomid exposure in sediment cages are displayed in Table 8. Effects on survival were observed only for chironomid larvae introduced on day 45 and exposed 7 d to the second contamination on day 50, whereas growth of surviving larvae was not impaired.
Survival of snails was impaired following the second contamination, sublethal effects were also observed on grazing activity (Figure 6). Recovery was observed following reinoculation of snails after the second and last contamination. Physa sp. survived and reproduced well in all microcosms (141 to 318 individuals) There were no differences between control and contaminated microcosms.
The two duckweed species introduced in the systems showed linear moderate growth during the course of both assays (400 fronds maximum). Despite variability, there was no tendency to lower growth in the presence of leachates. Growth of rooted macrophytes was high, especially for Groenlandia densa, for which the number of leaves was multiplied by a factor of 3 to 4 during the 70 days of the test. No effect of leachates were observed on growth, morphology or colour. Due to cladoceran grazing and absence of effect on some cladoceran species, algal growth in the water column was always controlled, except when chlorophyll content increased on day 28 following algal reinoculation. Measurements and observations on periphytic algae at the end of assay showed similar diversity in control and contaminated microcosms, and no difference in final biomass which was however variable between microcosms (0.043 to 1.91 μg Chl a cm−2).
The microcosm protocol, used here with toxicants for the first time, needs to be improved; variability was high for some biological parameters and conditions for good development of organisms (e.g., D. magna, amphipods) were not always fulfilled. Indigenous organisms brought with the sediment (Oligochaetes at densities of 5000 individuals m−2 at the end of the assay, and Ephemeropterans) were observed during the assay. These organisms, which were not removed by sieving sediment through a 2 mm mesh, might have hindered the development of benthic organisms such as H. azteca and Chironomus riparius. In reported studies, a significant decrease of Chironomus riparius weight was observed when Tubifex tubifex was present in natural sediment at densities > 1460 individuals m−2, and H. azteca biomass was also decreased by 60% at T. tubifex density of 20000 individuals m−2 (Reynoldson et al., 1994).
In addition, our approach might be criticized on several points. First, it involves various organisms but ignores relationships among those organisms when considering results, failing thus to consider the ecosystem response as a whole. As a matter of fact, results are displayed for each population, and interactions between populations are only considered for grazers and microalgae. As a consequence, the ecotoxicological assessment of MSWI bottom ash percolates presented here appears as a multi-index assessment rather than an ecosystem impact assessment. The study of the ecosystem as a whole would require specific ecological approaches that should be developed in the future but are at the moment beyond the possibilities of our laboratory and the scope of our research. It should nevertheless been kept in mind that the ecosystemic dimension is de facto taken into account by our protocol, which enables the assessment of the effects of pollutants in conditions where populations interact between each other and with their physical environment. Another criticism would be that the protocol hesitates between the study of the response of populations in one complex environment (the main compartment of microcosms) and the study of some individuals encaged in particular conditions. For reasons explained in the Materials and Methods section, we actually propose both approaches because we think they are complementary rather than opposed.
In spite of these difficulties and limitations, this study brings new results on the toxicity of MSWI bottom ash leachates to aquatic organisms exposed in microcosms mimicking lentic ecosystems. The choice of one single contaminant concentration was dictated by the need to take into account variability, which generally increases with ecosystem complexity. An alternative would have been to study a range of 6 concentrations with no replication. Instead we chose to start the assay with the concentration 1% estimated in single-species tests as the lowest observed effect concentration (LOEC). Daphnia magna was only sensitive to the second contamination (composite leachate, nominal cumulated concentration, 3%), and finally recovered at the end of the assay, as observed in the big compartment and in the immerged beakers. The response of D. magna may be explained by Cu introduced with leachates. In the first assay, the Cu concentration varied for 60 d around 80 μg l−1. The 16 day-EC50 corresponding to a reduction of D. magna population in the lab is 16.1 μg l−1 (Enserink et al., 1991). Copper concentration following the first contamination (25 μ g l−1) was lower than the 48 h-EC50 values of 50 μ g l−1 reported by Toussaint et al. (1995) and 26 μ g l−1 reported by Assmuth and Penttilä (1995), but reached 80 μ g l−1 after the second contamination (Figure 3), and this graduation might explain the observed delayed effects. Moreover, decline of Cu content might corroborate recovery of daphnid populations. Other metals (Cr, Ni, Pb and Cd) in the leachates were generally at concentrations < 1 μg l−1, concentrations expected to have no effects on D. magna reproduction (Assmuth and Penttilä, 1995). It should however be kept in mind that, in a toxicant mixture such as this, possible effects due to additivity or synergy of single metals could be expected. Unlike D. magna, the other cladocerans Ceriodaphnia dubia and Simocephalus vetulus were not sensitive to leachate contaminations. Ferrari (2000) found, for some MSWI bottom ash leachates, a higher sensitivity for D. magna (48 h immobilization) than for Ceriodaphnia (7 d inhibition of reproduction), but this is not a general result since inverse sensitivities were observed with other leachates by Férard and Ferrari (1997). Our results do not support the general idea that Ceriodaphnia dubia and D. magna are equally sensitive to many organic and non organic substances (Elnabarawy et al., 1986), and that Ceriodaphnia dubia is more sensitive than D. magna to wastewater treatment plant effluents (Schroder et al., 1991). According to Versteeg et al. (1997), who analyzed literature data on acute and chronic toxicity tests, Ceriodaphnia dubia, Simocephalus vetulus, D. magna and D. pulex are equally sensitive to most xenobiotics and effluents. If Cu can explain most observed effects in our study, the tolerance of C. dubia was not expected since both cladocerans showed the same sensitivity to copper in reported studies (NOEC reproduction: 6.3 μg l−1 (BKH, 1995) or 12 μg l−1 (Carlson et al., 1986) for Ceriodaphnia dubia, 8.2 μg l−1on average for D. magna (Biesinger and Christensen, 1972; Van Leeuwen et al., 1988; BKH, 1995). It seems that in complex assays such as microcosm or mesocosm assays, sensitivities of cladocerans may significantly differ. In lake enclosures assays on a mixture of metals (As, Cd, Cr, Cu, Hg, Pb, Ni and Zn), Ceriodaphnia dubia was more affected than D. magna (Jak et al., 1996). In our study, differences observed in response of the three cladoceran species used in the assay might be linked to several factors (organism size, reproduction rate, trophic level of the microcosms) which could benefit more to small species such as Ceriodaphnia dubia than to large ones such as D. magna (Jak, 1997). In particular, the low level of food (algal density estimated through chlorophyll concentrations) might have been a handicap for D. magna neonates to reach maturity, as underlined by the low survival observed. Finally, our results suggest that MSWI bottom ash leachates might impair cladoceran abundance and diversity in the field, but that algal control ensured by this group through its grazing activity (Lampert et al., 1986) would not be affected, as shown in this study by low chlorophyll levels, due to tolerance of some species or recovery observed for the others.
The amphipod H. azteca exposed in various conditions (via the sediment and the water column in big compartments and cages, via the water column only in beakers and in single-species tests) showed contrasted and sometimes contradictory responses to leachate contamination. Exposure via water only led in all cases to high toxicity, even at high dilutions. As for D. magna, the observed effects might be explained by Cu since the LC50 of H. azteca ranges between 17 and 87 μg Cu l−1 for exposure durations of 96 h to 14 d (Ankley et al., 1993; Schubauer-Berigan and Dierkes, 1993; Kubitz et al., 1995). Simultaneous exposure via the sediment and the water column for organisms free in the large compartment or encaged led to contrasting results: absence of effects in cages, probable effects on free amphipods. The contamination via the water column and the absence of metal contamination of pore waters could explain the absence of effects for amphipods spending part of their life cycle inside the sediment. On the other hand, the observed effects might be explained by sediment avoidance due to other unfavourable conditions such as ammonia or lack of food, constraining amphipods to swim in the water column in order to get food. Unfortunately, such observations were not done. We observed amphipods in the water column in both tests, but we are not able to bring more detailed observations that could support these assumptions.
With regards to chironomid larvae, only one positive response was obtained for encaged chironomids exposed to the second contamination, a result that cannot be explained due to absence of detailed investigation on conditions prevailing in the sediment cages at that moment. However, most results suggest that Chironomus riparius was not sensitive to MSWI bottom ash leachates, as shown by emergence rates similar in control and contaminated microcosms. The fact that larvae spend most part of their life cycle inside the sediment and the absence of pore water contamination can explain this result.
The two gastropods species introduced in microcosms, Lymnaea stagnalis and Physa sp., showed different responses to MSWI bottom ash leachates. Lymnaea stagnalis was sensitive to the second contamination (effects on survival and grazing activity), but recovery was observed, and Physa sp. was not sensitive at all. As for cladocerans, variations of sensitivity among one group (the gastropods) suggest that, in the field, the functions of this group will be maintained due to the higher tolerance of one or several species.
The growth of all primary producers tested in the microcosms (suspended microalgae and periphyton, floating and rooted plants) was not impaired by MSWI bottom ash leachates, whatever the species. Growth inhibition of microalgae might have been expected for two reasons: i) microalgae, and Pseudokirchneriella subcapitata in particular, are sensitive to these effluents, with 72 h-EC50s ranging between 1 and 6% (Ferrari et al., 1999; Lapa et al., 2002; Triffault-Bouchet et al., 2004), and ii) Cu was present at concentrations toxic for microalgae (Ivorra et al., 1995; Clément and Zaid, 2004). Clément and Zaid (2004) found, for the same species, a 72 h EC50 of 25 ± 7.0 μg Cu l−1 and a 72 h-NOEC of 5 μg l−1. However, they used the same OECD medium but with 10 times less Fe-EDTA to minimize chelation of copper by EDTA (Walsh et al., 1987; Wong, 1989; Ivorra et al., 1995). Other NOECs (72 or 96 h) of 15 to 64 μg Cu l−1 were reported for the same species (BKH, 1995; RIVM, 1999). The same considerations apply to duckweed Lemna minor, for which Jenner et al. (1993) found a 14 d NOEC of 60 μg Cu l−1. In the microcosms, presence of EDTA and other natural chelating agents might have reduced Cu bioavailability to microalgae and perhaps to other primary producers.
Leaching of MSWI bottom ashes generally liberates copper at high concentrations (Chandler et al., 1997; Freyssinet et al., 2002). However, the concentration of 1% is probably a large overestimation of the expected environmental concentration, corresponding to infiltration of much rain water through the wastes and neglecting attenuation by soil or vegetation during the transport of leachates to the receiving aquatic ecosystem. The studied scenario is thus a ‘worst case’. With a more realistic expected environmental concentration of 0.01%, effects would be tiny or would probably remain at the sublethal level, and copper concentration would be around 1 μ g l−1, a value slightly less than the PNEC of 1.6 μ g l−1 proposed for this metal (INERIS, 2003). In addition, as time goes on, toxic load of MSWI bottom ash leachates rapidly decreases (Ferrari et al., 1999; Triffault-Bouchet et al., 2004) and a decrease of hazard is likely to happen. As a conclusion, it is most probable that, in the field, effects on lentic aquatic ecosystems due to MSWI bottom ash leachates would be very difficult to observe.
This work was funded by the French Ministry of Public Works and received a student grant (Gaëlle Triffault-Bouchet) from ADEME, the French Agency for the Environment and Energy Management.