FR :

La bioaccumulation des métaux chez les mollusques peut être une conséquence de la désorption des éléments métalliques fixés sur les particules inertes ou vivantes en suspension dans l'eau qui leur servent de nourriture. Les meilleures expériences réalisées jusqu'ici semblent indiquer que les matières en suspension ont un rôle mineur dans la contamination des mollusques. Mais quelles que soient les précautions prises, les expériences ne simulent jamais parfaitement les phénomènes naturels. Aussi, nous avons choisi de reprendre cette question à l'aide d'une expérience en grand volume où le nombre de paramètres contrôlés est restreint mais où les organismes sont dans des conditions aussi proches que possible de la réalité.

Au cours d'un prégrossissement expérimental, les jeunes palourdes reçoivent des quantités variables contrôlées de nourriture phytoplanctonique. La fourniture d'une nourriture plus abondante a pour conséquence d'augmenter les quantités de cuivre et de zinc et au contraire de diminuer celles de plomb contenues dans chaque individu. Elle entraîne également une diminution des concentrations en cadmium et plomb traduisant une « dilution biologique » de ces métaux. L'augmentation de la densité de la population expérimentale a un effet négatif sur les quantités de cuivre et de zinc contenues dans les individus. A âge identique, les individus les plus grands présentent des concentrations plus faibles en cuivre et zinc et plus élevées en cadmium et plomb.

L'utilisation des eaux marines souterraines présente deux avantages: une production algale intense à un coût économique négligeable et la possibilité par échange thermique de réchauffer en hiver les eaux marines naturelles et ainsi de maintenir une croissance des mollusques toute l'année. L'emploi des eaux de forage n'entraîne aucun effet néfaste dans les phénomènes de bioaccumulation des métaux, du moins pour ceux étudiés ici : cadmium, cuivre, plomb et zinc.

EN :

Metal bioaccumulation in bivalves may occur as a consequence of the ingestion of inert or living particles with fixed trace elements. The best experiments like those carried out by BORCHARDT (1983, 1985) concerning Cd in mussels have shown that the role suspended matter plays in the contamination of molluscs is insignificant. But, however reliable the experimental methodologies, laboratory conditions never reproduce perfectly natural phenomena. We planned therefore to restudy the problem by using a largescale experiment where the procedure was characterized by a restricted number of controlled parameters and the organisms as close as possible to the real conditions.

During an experimental nursing of young carpet shells, we assessed the influence of various controlled quantifies of phytoplankton and of experimental population density on the transfer of metal from their environment to molluscs.

Carpet-shell brood (Ruditapes philippinarum) was distributed in several cylindrical containers the bottom of which consisted in a sieve. Food and seawater were renewed continuously by means of an ascending current (BAUD et BACHER, 1990). Nursing assays were carried out during summer over a period of 74 days. Eight groups of carpet shells were constituted according to food supplies (0, lx, 2x and 4x of Skeletonema costatum grown upon underground seawater plus natural phytoplankton) and population density (25 000 or 50 000 individuals per experimental container). Young molluscs were fed according to a cycle of 3 h-feeding periods and 2 h-periods with no food alternately. The average concentrations of algal cells in mollusc breeding seawater were 17.5, 35 and 70.10³ cells/L. This seawater was renewed at a flow rate of 3 m³/h.

At the end of the nursing period, molluscs exposed to different experimental conditions were separated by using sieves of different mesh-size (6, 8 and 10 mm). Young carpet shells were purged for 36 h in order to limit the overvaluation of bioaccumulated metal levels due to ingested matter (AMIARD-TRIQUET et al., 1984; KENNEDY, 1986). In each experimental and size-related categories, 90 individuals were sampled and divided into 3 groups of 30 specimens.

In these groups, soft tissues were separated from the shells and oven-dried at 80 °C for 48 h. The dry samples were powdered and three aliquot parts of about 100 mg each were digested with 1 ml of concentrated nitric acid (HNO₃, Suprapur) at 95 °C for 1 h. Then the trace element analyses were performed in this solution diluted with deionized water by dame (Zn) or by flameless (Cd, Cu, Pb) atomic absorption spectrophotometry using the Zeeman effect (AMIARD et al., 1987).

The influence of both food supplies and experimental population density on the dry weight of soil tissues of young carpet shells, their metal concentration and body burdens were examined by means of multi-linear regression analysis.

Increasing body burdens of Cu (4) and Zn (5) and decreasing body burden of Pb (3), corresponded to more abundant food supplies. The increase of phytoplankton supplies induced a decrease of Cd (6) and Pb (7) concentrations as a consequence of a « biological dilution » of these metals. Increasing density induced a depletion of Cu (4) and Zn (5) body burdens. Among individuals of the same age, the biggest ones exhibited the lowest concentrations of Cu and Zn (8 and 9) and the highest concentrations of Cd and Pb (6 and 7).

Increased food supplies induce a biological dilution of Cd and Pb in young carpet-shells. These results are in agreement with previous data concerning Cd, Cu, Pb and Zn in different species (MACKAY et al., 1975; BOYDEN, 1971; PHELPS et al., 1985; BERTHET, 1986). Thus front a sanitary point of view, the use of ground seawater for algal culture is not a risk since metal concentrations in molluscs are not enhanced.

FR :

L'utilisation hebdomadaire d'insecticides pour lutter contre les Simulies modifie l`abondance des populations benthiques. Les variations ne sont ni de la même ampleur ni du même signe, selon l'insecticide (Abate, chlorphoxime, B.t.) et les taxons.

Parmi les dizaines de rivières savanicoles traitées aux insecticides antisimulidiens, seules quelques unes font l'objet d'une surveillance de la faune non-cible. La problématique de ce travail est de savoir s'il est possible de généraliser à l'ensemble des rivières, les fluctuations d'abondance des populations observées sur une station de référence.

Sur la station de référence, l'évolution de l'abondance des populations pour les trois insecticides étudiés est mesurée à l'aide du rapport : R/P = log APTi/log AVT, où log APTi est le logarithme de l'effectif moyen pour la période de traitement à l'insecticide i et où log AVT est le logarithme de l'effectif moyen avant traitement. Pour chaque taxon ainsi que pour la faune totale on obtient donc un rapport Abate, un rapport chlorphoxime et un rapport B.t.

Les valeurs de ces rapports obtenus pour la station d'Entomokro, sont ensuite appliquées aux effectifs récoltés avant traitement sur trois autres stations. Les valeurs calculées sont comparées aux valeurs observées dans le cadre de la surveillance des rivières.

Pour la faune totale, l'écart entre les valeurs calculées et les valeurs observées est très faible, r = 0,98.

L'étude par taxon montre que la prédiction de ce modèle est excellente pour les Caenidae et les Hydropsychidae, bonne pour les Baetidae, les Chironomini, les Tricorythidae et les Orthocladiinae, médiocre pour les Tanypodinae et les Tanytarsini.

EN :

The weekly utilization of insecticides sprayed by the Onchocerciasis Control Programme to control Simulium damnosum larvae modifies the abundance of the non target benthic populations. These insecticides (B.t., Abate and chlorphoxim) have neither the same toxicity nor the same selectivity for the principal taxa. In consequence, the long-term variations of the abundance of the populations are neither of the same importance nor of the same sign (some of them remain the same or decrease white others increase), according to the product and the taxa. In long term, these variations being mainly the consequence of direct toxicity but also take into account the duration of the life cycles as well as spatial and trophic competitions.

Iin this study, only the saxicolous fauna is taken into account. If has been collected on rocks with the Surber sampler during the low water period (December to March).

Among the ten Savannah rivers treated with antiblackfly insecticides, only few are the object of a non-target fauna monitoring. The abject of this work is to land out if it is possible to generalize to all the rivers, the variations in population abundance observed for an insecticide on a control station.

In the reference station, the variation of population abundance for each of the insecticides is calculated by means of the ratio R/P = log APTi/log AVT, where log APTi is the logarithm of the average for the period treated with the i insecticide, and where log AVT is the logarithm of the average before treatment. For each taxa as well as for the total fauna, Abate, chlorphoxim and B.t. ratio were defined.

The values of these ratio obtained for the Entomokro station on the Marahoué river in the Ivory Coast are then applied to the data collected before treatment on three rivers or stations (Amou-Oblo on the Amou river in Togo, Asubende on the Pru river in Toge and Danangoro on the Marahoué river in the Ivory coast). The calculated values are then compared to the values collected after treatment in the river monitoring programme.

In the reference station the ratio of the populations collected before and after treatment are sometimes greater, sometimes lower than unit (when the ratio is greater than unit, it means that there is an increase of the abundance of populations during the processing period).

The ratios are very close to the one for the B.t. which is the less toxic insecticide against the non target fauna, the greatest for the chlorphoxim which is the more toxic product and intermediate for Abate which has an intermediate toxicity. It appears that the more insecticides were estimated to be toxic in gutter tests, the more population abundance differ tram the reference ones.

For the total fauna, differences between the values calculated and observed were very low (r 0,98). For the taxa, the prediction of this model is excellent for Caenidae and Hydropsychidae, good for Baetidae, Chironomini, Tricorythidae and Orthocladiinae, but mediocre for Tanypodinae and Tanylarsini. The authors submit some hypotheses to explain the bad adequation of the model for these two taxa; hypotheses based on taxonomic problems and niche releasing.

FR :

La végétation macrophytique des rivières acides d'Europe occidentale présente une grande unité, marquée par la constance d'un ensemble floristique de base, dont certaines espèces (Callitriche hamulata, Myriophyllum alterniflorum, Potamogeton polygonifolius, P. alpinus, etc) sont propres à ces eaux acides, d'autres (Callitriche platycarpa, Sparganium emersum, Elodea canadensis, Potamogeton crispus, etc) étant communes avec les eaux calcaires.

Cet ensemble subit des variations écologiques et chorologiques :

L'augmentation du niveau trophique des eaux conduit d'une végétation oligotrophe (différenciée par P. polygonifolius, Juncus bulbosus, etc) à un groupement mésoeutrophe (caractérisé par Callitriche obtusangula, Nasturtium officinale, etc). Ainsi, quatre types de phytocénoses ont pu étre distingués dans les Vosges du Nord en fonction de ce paramètre. D'autre part, une pollution organique localisée, par exemple par une pisciculture en Bretagne, provoque une augmentation des espèces eutrophes (C. obtusangula) et une régression des taxons plus oligotrophes (Scapania undulata).

L'alternance de biotopes courants et lents se traduit par un remplacement de phytocénoses : ainsi, en Bretagne, Oenanthe crocata et Ranunculus penicillafus sont associées sur les radiers, alors que Nuphar lutea et S. emersum caractérisent les mouilles. De plus, une diminution de l'éclairement permet la présence de groupements bryophytiques spécifiques.Les principales variations chorologiques entre les deux territoires correspondent à un enrichissement des zones atlantiques en macrophytes euatlantiques ou méditerranéoatiantiques, comme O. crocata ou Apium inundatum, qui manquent dans les régions médioeuropéennes. En outre, selon les territoires biogéographiques, l'écologie de certaines espèces peut varier (par exempte S. emersum, R. peltatus, P. polygonifolius).

Les macrophytes permettent donc, à l'instar des invertébrés benthiques, l'élaboration de biodiagnostics des eaux courantes.

EN :

For the study of the vegetation of acid rivers in France, two areas have been investigated by the authors : Armorican Massif and Northern Vosges (fig. 1).

This paper has two aims :

- assessing the general trends of macrophytic vegetation : floristic composition and plant communities, with special reference to the influence of physical and water quality parameters,

- pointing out and interpreting the variations observed between the two regions (chorological differences), and in the watercourses (ecological variations).

Floristic data and phytosociological surveys (from bryophytes to angiosperm macrophytes) were collected along 50 m. stretches (at least) by the stratified sampling method in the case of large watercourses, and by systematic surveys for smaller ones. Consecutive 50 m. long sketches were used to measure the effect of localized pollution. (The inventory is as precise as possible). A distinction was made between the aquatic communities and river-bankside ones. At the same time, physical parameters were measured or observed (depth, light conditions, width,...), certain water-quality parameters were measured (pH, conductivity,...) and water samples were collected for further analysis.

The data collected are gathered and compared in a table (table 1) showing differences between aquatic and subaquatic species, and between sunlit and shaded stretches. Trophic tendencies of most species are given according to either our observations or the literature.

The main ecological feature is the streaking longitudinal variation of aquatic species, which agrees with the zones described by ILLIES and BOTOSANEANU : crenon, rhitron and potamon. In sunlit areas the succession is dominated by the following species :

Potamogeton polygonifolius, Juncus bulbosus, Glyceria fluitans,...

-> Ranunculus spp., Callitriche hamulata, Myriophyllum alterniflorum,...

     ->Potamogeton crispus, Elodea canadensis,...

Areas in the shade are characterized by bryophyta; the succession is :

Scapania ondulata

->Rhynchostegim riparioïdes, Fontinalis antipyretica

As for the aquatic species, a longitudinal succession of bankside species exists :

Sparganium erectum, Carex paniculata, Carex rostrata

->Phalaris arundinacea

      ->Phalaris arundinacea, Typha latifolia

The macrophytic vegetation in acid rivers of Western Europe shows a great homogeneity, marked by the constancy of a floristic list, in which some species (Callitriche hamulata, Myriophyllum alterniflorum, Potamogeton polygonifolius, Potamogeton alpinus, Scapania undulata, etc.) only grow into acid waters, while others species (Callitriche platycarpa, Sparganium emersum, Elodea canadensis, Potamogeton crispus, etc.) grow also in calcareous waters.

This basic unit is submitted to ecological and chorological variations.

Ecological variations are due mainly to changes of the water quality. They are illustrated by two examples :

- two river systems in the Northern Vosges where a general survey shows longitudinal variations of communities without pollution or assessing different kinds of pollution,

- a pisciculture pollution with observations of species distribution on consecutive sketches.

An increase of water trophic level leads progressively from an oligotrophic vegetation (marked by Potamogeton polygonifolius, Juncus bulbosus, etc.) to a meso-eutrophic community (characterized by Callitfriche obtusangula, Nasturtium officinale, etc.). Four communities are distinguished in Northern Vosges (table 2) and characterize water quality (both pollution and trophic level) :

A : Potamogeton polygonifolius, Juncus bulbosus, Sphagnum spp. community, with the following main water-quality parameters : conductivity : 40-80 µS, phosphates : <20 ppb P, ammonia : <20 ppb N;

B : Potamogeton polygonifolius, Ranunculus peltatus, Callitriche hamulata community, (40-80 µS, 20-40 ppb P, ammonia : <40 ppb N)

C : Ranunculus peltatus, Callitriche hamulata, Potamogeton alpinus, Myriophyllum alteriflorum community, (60-120 µS, 40-120 ppb P, ammonia : <200 ppb N)

D : Callitriche obtusangula and Nasturtium officinale inside C community list, (80-130 µS, 40-300 ppb P, ammonia : -> 500 ppb N).

Two vegetation maps show the distribution of these communities (fig. 2). The succession of communities is related to water quality and fastened sequences are compared to normal ones. Without any pollution, A and B communities caver more than 5 km along the upper streams; under these conditions, the entire sequence A to D develops to 15 km, while with much anthropic pollution, it may occur within less than 3 km. A heavy pollution leads to the leads of communities as observed with the change tram A to C (without any B) due to a camping-site.

A localized organic pollution by a Breton pisciculture is responsible for water-quality changes (table 3): ammonia, nitrites, phosphates and conductivity increase. The succession of mineral nitrogen forms is assessed (fig. 3). In the N-NH₄+ and N-NO₂- peak areas, the eutrophic species increase (Callitriche spp., specially Callitriche obtusangula, Amblystegium riparium) and the oligotrophic ones decrease (Scapania undulata) (fig. 3). Tissue analysis of Ranunculus penicillatus and of Callitriche spp. does not show any enrichment in nitrogen in polluted areas: the ecological response is mainly an increase of nitrophilous macrophytes (Callitriche spp.).

The influence of water quality is more obvious in the Northern Vosges for most of the oligotrophic communities grow in forested sketches, white in Brittany these conditions are rare and eutrophication occurs in all the streams, even in the source areas. Floristically, neutrophilous or ubiquitous species are numerous. Some of them indicate either eutrophication or pollution : Amblystegium riparium and perhaps Octodiceras fontanum for bryophyta, Callitriche obtusangula, Potamageton perfoliatus and Sparganium. emersum (long leaves form), for aquatic species, Nasturtium officinale, Glyceria maxima for helophytes.

Physical parameters also lead to variations. The succession of swift and slow areas leads to changes in the communities. Thus, in Brittany, Oenanthe crocata and Ranunculus penicillatus are associated on the riffles, while Nuphar lutes and Sparganium emersum characterize the pools, even in the upper parts of the streams where there are watermill ponds.

The relationships between local hydrodynamic features and species distribution are discussed with regards to the stretches around the pisciculture area.

Main chorological variation is an enrichment of atlantic areas with numerous eu-atlantic or mediterraneo-atlantic species : Porella pinnata and Amblystegium fluviatile as bryophyla, Oenanthe crocata, Apium inundatum, Ranunculus omiophyllus as spermaphyta. As there are few or no species restricted to medio-european areas, Atlantic communities are often richer than eastern ones. But some invading species as Elodea nuttallii are spreading all over the country from North-Eastern France.

Ecological trends of certain species (Sparganium emersum, Ranunculus peltatus, Potamogeton polygonifolius) can change between the two biogeographical areas. So, Sparganium emersum grows in the upper oligotrophic zones in Northern Vosges while it characterizes lower meso-eutrophic zones in Brittany. In Brittany, Ranunculus peltatus un exist in eutrophic areas when the stretches are far from the sea. Potamogeton polygonifolius is very rare in swift streams in Brittany.

Therefore, macrophytes permit, as benthic invertebrates do, to get a river biodiagnosis. They are easily observed and mapped so they can be used for a general survey.

FR :

Cet article présente les résultats d'une étude, réalisée en usine de production d'eau potable, de la filtration biologique sur charbon actif en grains (CAG) utilisée en seconde filtration à la suite d'une étape d'ozonation. Les buts de cette étude étaient d'étudier la cinétique de colonisation bactérienne des filtres à CAG, d'analyser la qualité de l'eau en sortie de filtres durant la phase de colonisation biologique et de comparer le fonctionnement à l'équilibre biologique de deux filtres à temps de contact « eau-CAG » différents. Les résultats des mesures de biomasse bactérienne fixée, effectuées par une méthode de respiration potentielle de glucose, montrent qu'un volume d'eau filtrée de 30 - 50 10³ m³/m² de filtre est nécessaire pour coloniser le CAG. A ce moment, le biomasse fixée est d'environ 7 µgC/mf de CAG. La comparaison de la qualité de l'eau, en terme de carbone organique dissous biodégradable (CODB), entre l'entrée et la sortie du filtre étudié montre que durant la colonisation le processus de biodégradation, lié à l'installation d'une biomasse bactérienne suffisante, prend donc progressivement le pas sur le processus d'adsorption dominant au départ, mais qui diminue rapidement. La comparaison de filtres, fonctionnant dans des conditions différentes de hauteur de CAG et de vitesse de filtration, montre l'importance du paramètre temps de contact sur l'efficacité de la filtration biologique sur CAG. De ce point de vue, des temps de contact de 10 à 15 minutes offrent une bonne efficacité de la filtration avec un abattement de CODB pouvant dépasser 80 %.

EN :

During the two last decades, granular activated carbon (GAC) filtration has been used in drinking water treatment plant in order to remove by adsorption a part of the organic pollution. The adsorption capacity of GAC is however rapidly saturated, requiring frequent regeneration or replacement, which are both expensive and fastidious.

An interesting alternative, introduced in some French and Canadian treatment plants, consists in using GAC filtration without regeneration, taking advantage of the activity of microbial communities colonizing the GAC particles. In fact, this biological filtration, introduced in the treatment lines after land filtration and an ozonation stage, has the advantage of specifically removing biodegradable compounds which are the most undesirable fraction of dissolved organic-carbon (DOC).

This paper presents results of a study carried out at the Choisy-le-Roi treatment plant to understand better of the processes involved in biological filtration and to derive guidelines for a rational management of these filters. The aim to the study was to follow the bacterial colonization of GAC filters, to test the quality of the filtered water during the colonization phase and to compare the functioning of two colonized inters working with different empty bed contact times.

During the period of the study (from March to September 1989), GAC samples were collected at various depth in the filters in order to estimate the fixed bacterial biomass by a method of potential glucose respiration. Samples were also collected in the inlet and outlet waters for measuring the DOC and its biodegradable fraction (BDOC), and also the bacterial abundance estimated by epifluorescence microscopy.

Fixed bacterial biomass measurements on GAC were made during the colonization of the GAC filter nr 56 (table 2). The data showed that a filtered water volume of about 30 to 50 10³ cubic meters per square meter of GAC filter was required to reach a stable level of colonization (fig. 1). At this stage, the fixed bacterial biomass was around 7µgC/ml (fig. 1 and table 2).

The quality of the water, in terms of dissolved organic carbon (DOC) and biodegradable dissolved organic carbon (BDOC), was tested during this colonization. The DOC in the inlet water was in the range of 1.71 to 2.28 mgC/l with a biodegradable fraction between 0.4 and 0.7 mgC/l. A removal of DOC was observed in all the sampled situations, it ranged between 0.49 mgC/l and 1.48 mgC/l; the BDOC removal ranged between 0.31 mgC/l and 0.50 mgC/l (table 3). At the beginning of the colonization, the removal was due to adsorption, practically equal fractions of BDOC and NBDOC were removed. The adsorption process rapidly decreased as indicated by the decrease of non biodegradable dissolved organic carbon removal (fig. 2). Later, biodegradation increased with the bacterial colonization of the GAC. So during the colonization, biological processes took turn with adsorption so that a significant removal of biodegradable dissolved organic carbon was maintained from beginning to end.

Enumerations of bacteria were performed during colonization in the inlet and outlet waters of the nr 56 GAC filter by epifluorescence microscopy; they showed an increase of bacteria in the outlet water compared to the inlet water (table 4). This increase was nevertheless negligible in terms of carbon compared to the removed carbon; the average exportation was 0.0015 mgC/l white the removal was in the range 0.3 to 0.5 mgC/l.

Two filters (filters nr 56 and 58) with various functioning conditions were compared after their biological colonization. A vertical profile of BDOC in the nr 56 filter is presented in figure 3; it shows that the must significant part of biodegradation took place in the first half of the filter, where the bacterial biomass was greater. A better BDOC removal was observed for filter 58 than for filter 56 (respectively 0.59 and 0.44 mgC/l) with empty bed contact time of respectively 15 and 10 minutes. In figure 4, the percentage of BDOC removal is plotted against empty bed contact time for filter nr 56 and 58. These data confirm previous results obtained with GAC pilot filters at the Neuilly-sur-Marne treatment plant (data also presented in figure 4) showing that the percentage of BDOC removal increases with increasing empty bed contact times. From this point of view, the actual contact times of 10 to 15 minutes applied in the treatment plant have shown a good efficiency of the filtration with BDOC removed higher than 80 %.

GAC biological filtration is a very promising way of removing dissolved organic matter in drinking water treatment plants.

FR :

La coagulation par le Fe(III) de substances humiques aquatiques est optimale à pH = 5,5 pour un rapport massique Fe/COT de l'ordre de 2, quant à l'élimination du COT. Le but de cet article est d'étudier l'influence d'une préoxydation sur l'élimination de la matière organique par coagulation dans ces conditions.

La préoxydation (ozone, bioxyde de chlore ou chlore) à faibles doses induit une légère dégradation des rendements de coagulation des acides fulviques. Un taux de préoxydant fort (≈ 0,5 mg oxydant par mg de carbone organique initial), pratiqué avant la coagulation, implique de mettre en oeuvre une dose de coagulant plus élevée [≈ 3 mg Fe(III) par mg de COT]. Cette augmentation est moins marquée dans le cas d'une préchloration.

Les substances humiques ne représentant qu'une partie de la matière organique des eaux naturelles, des manipulations sur des eaux brutes ont donc été entreprises. L'effet de faibles doses d'ozone conduit à une légère amélioration de la coagulation. Cependant de fortes doses d'oxydant (0,5 mg d'ozone ou de bioxyde de chlore par mg COT) nécessitent l'emploi de plus fortes doses de coagulant. De plus, le rendement d'élimination est aussi affecté.

EN :

We present here some works which take place in the particular frame of the study of the drinking water treatment of reservoir waters containing organics at high concentration, mainly humic substances. Previous studies (LEFEBVRE and LEGUBE, 1990) on coagulation of fulvic acid solutions have proven that the optimal removal of organic matter was reached at pH = 5.5 with 2 mg of ferric iron per mg of organic carbon. The main question in this study is to know the impact of preoxidation (ozone, chlorine dioxide, chlorine) on the removal of organics by iron(III) coagulation.

The reconstituted waters obtained by dissolving fulvic acids (table 1) in a solution of salts in high-purity water (table 2), are defined by both a low inorganic content (as calcium, sulfate and bicarbonate ions) and a high organics concentration, main characteristics of real impounded waters (table 3). The experimental procedures of coagulation-flocculation-clarification followed the treatment lines shown in figure 1. Ozonation was carried out in a semi-continuous system (batch solution), by bubbling ozone (fig. 2). Chlorination and preoxidation with chlorine dioxide were performed in a batch reactor. The preoxidation steps were carried out at pH of reconstituted waters or of raw waters. A Dohrmann DC 80 was used for the determination of TOC. The accuracy of TOC measurement was found to be ± 0.05 mg/l C (for the range of studied concentration). Residual iron was analysed by atomic absorption on a Perkin Elmer 2380 apparatus (oxy-acetylene flame).

As shown in table 4, preozonation induced a slight decrease of Cebron fulvic acid removal by iron(III) coagulation at high ozone dose applied (= 0.5 mg O₃ per mg TOCi). In order to try to understand why ozone partially inhibits the efficiency of iron coagulation to remove fulvic acid at acidic pH, we determined the optimum dosage of coagulant required to obtain the best percent of removal on preozonated fulvic acid solutions. Figure 4 shows that preozonation at about 0.5 mg O₃ per mg TOCi appears to have shifted the region of the optimal TOC removal towards the higher Fe/TOCi mass ratios (≈ 3 mg Fe(III) per mg TOCi, for two fulvic acids).

Prechlorination (> 1 mg Cl₂/mg TOCi) induced a decrease of Cebron fulvic acid removal by iron(III) coagulation (table 5). Hence, we examined the optimum dosage of coagulant required to obtain the best percent of removal on prechlorinated fulvic acid solutions (fig. 6). The region of the optimal TOC removal was still obtained for a Fe/TOCi mass ratio of 2. However, the difference between the percent TOC remaining for inital fulvic acid of 10 and 15 mg/l was not significant (respectively 61.4 % and 61.9 %). A greater prechlorinatlon dosage (> 0.5 mg Cl2/mg TOCi) would probably give a shift of the optimal mass ratio.

Preoxidation with chlorine dioxide (even without ClO₂ residual) before a coagulation induced a decrease at TOC (table 4). So, we studied the shift of optimum dosage of coagulant required to remove the organic matter. Figure 5 shows that preoxidation (≈ 0.5 mg ClO₂/mg TOCi) shifted the region of optimal TOC removal towards higher Fe/TOCi mass ratio (≈ 3 mg per mg for Cebron fulvic acid, in figure 5).

The fulvic acids represent only an amount of TOC of a raw waters. So we carried out experiments on real impounded waters. For the study of the influence of the preozonation on the coagulation efficiency, we chose to work at the optimal coagulant dose determined from previous experiments. Several runs were conducted with different dosages of ozone between 0 and 2 mg/l O₃. Results reported in tables 6 and 7 indicate that for these two raw waters (Moulin Papon and Cebron), the preozonation slightly improved the efficiency of iron coagulation at acidic pH. These results can only be compared with Cebron fulvic acid when applied ozone dose was 0.2 mg O₃/mg TOCi.

Consequently, other experiments were carried out at a high ozone dose (- 0.5 mg O₃/mg TOCi). They showed a shift in the region of the optimal TOC removal towards the higher coagulant dose (fig. 7), as already observed with the fulvic acid solutions.

Only one experiment was made with chlorine dioxide in the case of Moulin Papou raw water. This preoxidation (≈ 0.5 mg ClO₂/mg TOCi) induced a shift to higher coagulant dosage but also a decrease of the efficiency of iron coagulation whatever the applied coagulant dosage (fig. 8).

As compared to humic chlorination literature, little information exists concerning ozonation of humics. Furthermore, in spite of small amounts of identified by-products in the literature, ozonation was usually found to increase smaller size materials in humic substances (GLOOR et al., 1981; VEENSTRA et al., 1983; FLOGSTAD and ODEGAARD, 1985; ANDERSON et al., 1986; AMY et al., 1987; LEGUBE et al., 1989). Moreover, analyses of carboxyl-group on the Cebron fuivic acid allowed us to show that the carboxyl content increased as ozone dosage increased (fig. 9).

Some works (VAN BREEMEN et al., 1979) have proven the rate of carboxyl groups in the stoichiometry of coagulation reaction of humics with iron(III) at slightly acidic pH. Consequently, it is not surprising that preozonation of fulvic acids appears to have shifted the region of TOC removal up into the higher iron dose range, according to the enhancement of the carboxyl content in the fulvic acids by ozone. This effect of preozonation was already reported by others (JEKEL, 1983; RECKHOW and SINGER, 1983).

Chlorination (GLAZE and PEYTON, 1978) or oxidation with chlorine dioxide (NORWOOD et al., 1983; COLCLOUGH Of al., 1983) led also to a decrease of the molecular weight of aquatic organic matter. The oxidative action of chlorine and chlorine dioxide gave many products like aromatic and aliphatic acids (chlorinated or not chlorinated) and probably increased the carboxyl content of humic substances. Hence, those oxidants inhibit the coagulation-flocculation of organic matter.

FR :

En France, les sécheresses consécutives des années 1985, 1986, 1989 et 1990 ont mis en lumière les problèmes relatifs à l'alimentation en eau potable, l'irrigation des terres agricoles et la préservation des écosystèmes aquatiques. Dans le cas des zones humides, continentales et littorales, caractérisées par une compartimentation hydraulique souvent complexe, le manque de connaissance se fait particulièrement sentir. Bien que de nombreux travaux aient permis d'évaluer l'évaporation des masses d'eau et l'évapotranspiration de certaines espèces d'hydrophytes et d'hélophytes, les études débouchant sur des bilans quantitatifs restent peu fréquentes. Le bilan hydrologique du marais de Moëze (2250 ha) a été calculé par décade entre le 11/06/89 et le 31/08/89. Il prend en compte le débit au droit de l'ouvrage d'alimentation, les volumes prélevés pour l'irrigation hors marais, les infiltrations et l'évapotransplration sur les 318 km de canaux. L'estimation de la consommation d'eau des parcelles est globalisée au niveau des mesures d'infiltration.

Les pertes par infiltration sont secondaires (9,4 %) au regard des volumes prélevés pour l'irrigation (38,0 %) et évapotransplrés par les canaux (43,7 %) dont 51,1 % uniquement par les 28,6 % des plans d'eau colonisés par Typha latifolia.

L'optimisation de la gestion estivale de l'eau d'un marais littoral agricole nécessite dans un premier temps de minimiser les pertes. C'est essentiellement sur la consommation d'eau des canaux colonisés par les hélophytes que l'on peut intervenir. Nous proposons un abaque qui permet d'évaluer l'importance des économies d'eau réalisées en fonction de plusieurs scénarios d'aménagement du réseau hydraulique.

EN :

In France, the drought that occurred during the years 1985, 1986, 1989 and 1990 have emphazised the problems of freshwater supplies for human consumption, for irrigation and for the conservation of aquatic systems. Today the water has to be economized through a rationalized management. Water balance must be evaluated in order to compare supply an demand. Hydrological functioning is particularly badly known as far as continental and coastal wetlands are concerned, probably because of a generally very complex hydrological partition. Many papers deal with the evaporation rate from a clear water surface or the evapotranspiration rate from several species of hydrophytes and halophytes. However studies of quantitative water budgets of wetlands remain few in number.

This paper reports an analysis of the budget summer water in that was a salt marsh now containing freshwater. The 25 km² marsh of Moëze is located on the French Atlantic coast; it has been progressively constituted by filling up a tidal bay since the flandrien period. The soils correspond to fluvio-marine silts, locally called « bri » accumulated over several tens of meters thick. The marsh is bounded on the North by the ancient limestone coast, on the South by the Arnoult River, and on the West by the coastline. Its drainage network includes permanently flowing main canals and also small silted-up ditches which sometimes dry up in summer and are largely colonized by aquatic plants, particularly Typha latifolia. The channels network is very dense (144 meters of ditches per hectare) and complex because of a close connection between all the canals and ditches. The regional oceanic climate is characterized by a surplus water balance from October to April (+ 315 mm) and quite a short one tram May to September (-338 mm).

The important terms of the water budget equation in this study were : the quantity of water pumped from the Arnoult River through the inflow sluicegate (Qa), the precipitations (P), the irrigation out of the marsh (lr), the evapotranspiration of water bodies (Epo), the seepage through canals and ditches (ls), and the change in water soil strorage (Vs). These terms are not equally susceptible to be measured. Groundwater seepage and evapotranspiration are difficult to measure and they are often determined by difference, but, they contain the residual error of ail the terms. The methods to evaluate each term of water budget were carried as follows :

1. As the inflow gate functions as a siphon, the flow rate (Qa) was calculated with the drowned orifice formula (LENCASTRE, 1984). The upstream and downstream water levels were permanently recorded by 2 limnigraphs. The upstream and downstream water velocities were measured every 2 days with a micro-currentmeter.

2. The precipitation values (P) used in the water budget equation correspond to the average of 4 rain gauges placed around the marsh.

3. The evapotranspiration of the channel network (Epo) was directly estimated through 4 experimental floating tanks (0.50 x 0.55 x 1.05 m size) previously used by GIRAUD (1985). One of the tanks was placed in clear water, the others was planted the typical aquatic vegetation of the marsh (Lemna sp., Ceratophyllum sp., and Typhia latifolia). The drop the water levels in the tanks corresponded to the loss of water due tou evapotranspiration. All the tanks were filled up to a fixed level, and the amount of water added, measured every 2 days.

4. The outputs for irrigation (Ir) concern 298 ha of maïze out of the marsh, and 23 farmers. The water amounts taken off were estimated form an inquiry of irrigation practices associated to a field control.

5. The water losses by seepage (ls) through canals and ditches were directly measured on the field by using the closed basin method. A length of canal was closed by 2 watertight bulkheads. The fall of water level was recorded and the amount of water added to maintain a fixed level was measured. This method is considered by KRAATZ (1977) as being the most accurate specially for low seepage. The fall of water level never exceeds 10 % of the water depth in the basin. 34 canals and ditches in the marsh were sampled. According to CHEVALLIER et al. (1984), 3 parameters influencing the soil permeability were measured : granulometry, CaC03 content and sodicity. After the sampling plan we have retained 4 experimental canals (average length =47 m, average water surface = 135 m2, average depth = 0.44 m).

6. The water strorage in the soil (Vs) was evaluated by analyzing the groundwater table fluctuations and moisture changes.

The water budget calculated for 10-day periods depending on the climatology calculations, from 10th June 1989 to 31st August 1989. The water losses due to seepage were secondary (9.4 %) compared to the amounts of water taken off by irrigation (38.0 %) and channel network evapotranspiration (43.7 %). The water consumption of helophytes such as Typhia latifolia was 2 to 3 times higher than the evaporation of a clear surface water body as shown in figure 3. In the marsh of Moëze, 51.1 % of channel evapotranspiration was due to the colonization by Typhia latifolia of the canals and ditches although they represent only 28.6 % of the channel network surface.

This study shows that is possible to quantify a water budget for a large scale wetland from field measurements associated to experimental approaches, with a satisfactory accuracy : less than 10 %. To reduce the water consumption of the marsh of Moëze, three essential recommendations may be given : the reduction of the global channel network surface, the cleaning of a part of ditches colonized by Typhia latifolia, or the combination of both techniques. According to the different management schemes, it is possible to predict the amounts of freshwater saved (fig. 5).

FR :

L'article que nous proposons s'inscrit dans le cadre des problèmes d'optimisation bimensionnelle (irrigation & salubrité) des ressources en eau durant la période d'étiage. Sur le cas du système NESTE, la résolution est effectuée selon deux approches :

- un modèle de programmation dynamique avec état de dimension deux (niveau des réserves, niveau dans la rivière) où, dans la solution numérique, les variables sont discrétisées;

- un modèle « synthétique » où l'on calcule une probabilité de non dépassement caractérisant l'état hydrique des ressources du système. Une règle empirique permet d'associer à cette grandeur une décision de consigne à effectuer.

Les résultats numériques sont comparés sur une série de chroniques historiques. Les avantages et les inconvénients de chacune des deux approches sont mis en lumière sur le cas réel du système NESTE.

EN :

This paper deals with bicreteria (irrigation & water quality) weekly operation of a water resource system during dry period. Two ways of handling the problem are assessed and compared on a real case study :

- a stochastic dynamic programming modal with a two dimensional state (reservoirs level, river level) that is numerically solved by discretization ;

- a more « synthetic » model where the state is expressed in term of a tail aera probability related to the consumption of all the present water resources in the future. A practical decision rule is based upon the associated critical value.

Numerical results are plotted on historical varies for both methods.

From the present application to the NESTE system, the conclusions are :

1) Both procedures allow the system manager to formulate operating strategies in a rational way :

- An operating rule can be derived to allocate water so as to meet a combination of the various objectives. It is expressed as a feedback law linking what we know from the state of the system to how we control its evolution.

- Both methods need a parameter to be set up by stochastic simulation.

- They give close results on the basis of the past data and can be conveniently proposed to system managers.

2) The system analysis approach is based on stochastic dynamic programming. If can be efficiently used to derive optimal feedback ruses of operation and can routinely deal with complex decisions such as limiting irrigation when a shortage is to occur or take the risk to keep going and decrease output targets for water quality management. At the same time, this procedure entails heavy computing time, uneasy interpretation of the weighting coefficient between irrigation and water quality objectives, and a rather artificial elicitation of the global compromise.

Such an approach is very well fit for simulation because it is composed of elementary blocks that are gathered in a transition relationship to describe the system's dynamic evolution. This approach also provides a means to get an optimal policy as long as the system manager accepts the necessity to formulate an objective function consistently with dynamic programming (i. e. stages are separable and additive). Of course this optimal allocation should be carefully examined because of modal uncertainties influencing both the system response and the hydrological behaviour.

3) The synthetic method may appear more attractive from the engineering point of view for the following reasons :

- the state is easily interpreted in terms of « dry year with a return period of 10, normal year, exceptionally wet year » and so on. The trade-off coefficient is the volume one wants to keep in the reservoirs at time T for a wet year. Consequently if the parameter is chosen with « good sense », no optimization scheme is needed.

- there is no computation except a mass balance equation and a normal probability law adjustment which is very easy because it deals with cumulative quantities.

4) Such models are designed to serve only as multicriterion decision making aids. In very dry days such as occurred in summer 1976 or 1989 in France they cannot create additional water resources... still, they can help the system manager by constant up dated multidimensional estimation of the risks that may be encountered when following different operation rotes. In the case of the NESTE system, a real-scale experiment began in 1989: in real time operation, both models worked on line as decision-making supports, and the system manager made a thorough study of the hydrological conditions when the two approaches did not agree on the same stragegy for the following week.

FR :

On connait l'extrême variabilité des ressources en eau superficielle et les problèmes que cela pose en aménagement du territoire. En conséquence, générer de longues séries de débits annuels (ou mensuels) est une nécessité pour l'étude de la disponibilité des ressources en eau (BURGES et LETTENMAIER, 1977).

Cette génération de longues séries peut se faire à partir de la connaissance des pluies si l'on dispose d'un modèle fiable de transformation pluie-débit. Un premier modèle linéaire visant à l'estimation du seul débit annuel, déjà utile pour un premier dégrossissage de la ressource, est bâti à partir de la pluie annuelle et de sa répartition entre les mois. Il constitue un outil de base que l'on a jugé néanmoins un peu trop rudimentaire. La recherche d'un modèle conceptuel simple nous a semblé une voie plus prometteuse. Exploitant une recherche déjà entreprise sur ce type de modèle au pas de temps journalier, on a construit un modèle mensuel à trois paramètres dont l'élaboration a été justifiée point par point par une meilleure efficacité des résultats fournis sur une soixantaine de bassins versants appartenant à la partie septentrionale de l'Algérie. Nous avons analysé les paramètres du modèle et obtenu une estimation approchée de la matrice des variances-covariances de ces paramètres afin de disposer d'un outil similaire à celui existant pour le modèle linéaire général.

EN :

Availability of long series of annual or monthly streamflows appears to be a necessity in water resources studies. In regions where such data do not exist we may resort to rainfall data to obtain the required long series of streamflows. In this case we need a reliable rainfall-runoff model. Such a model is often called a water balance model although this designation ignores the dynamics involved in the model. A first attempt was to establish a linear model connecting an annual streamflow to the concurrent rainfall and its allocation within the year.

This model was calibrated on 63 catchments throughout northern Algeria, an area of about 200,000 km² where mean annual streamflows range from 5 to 350 mm.

The standard error of the model, expressed as a percentage of the mean annual streamflow was about 41 %. This relatively poor result led us to explore the capabilities of lumped conceptual models. Hydrologic models may be of the following three types :

- the linear black box model,

- the lumped conceptual model,

- the physically based model.

The linear black box model may benefit from huge mathematical developments but is very poor from a hydrologic point of view. The physically based model is a permanent search to comply more and more to the whole catchment and all processes which are supposed to take place in it. It is tantamount to sheer utopia since the required fine description of the catchment will be definitely out of reach for economical reasons. The only workable solution lies in the design of a lumped conceptual model. But we have to bear in mind that such models cannot rely on a comprehensive physical basis and thus, all specifications have to be justified by the outputs of the model.

We benefited from a recent research upon a daily conceptual model that served as a starting sketch for the monthly model. After having tested numerous alternatives we worked out a simple model with only three parameters to be calibrated. The standard error of this model was about 27 % of the mean annual flow. According to recent analyses about the sensitivity of model parameters, we calculated the matrix of covariances of the parameters.

This was necessary because, due to the lack of a long series of data, we could not achieve an actual validation of the model against a streamflow sequence not used in the calibration process. However, the extreme parcimony of the model may alleviate the need of such a validation and the model compared favourably with well-known water balance models. As a conclusion, we may hope that the tool we have devised will be of valuable help in water resources management.