Volume 10, numéro 4, 1997

FR :

Sous climat soudanien caractérisé par une unique saison des pluies, les sols sont dénudés en fin de saison sèche suite au pâturage et aux travaux préparatoires au semis. Le ruissellement intense en début d'hivernage diminue progressivement avec la mise en place des couverts végétaux.

L'influence du développement de la végétation sur le ruissellement est étudié au moyen des données pluie-débit de 4 parcelles (50 m²) couvertes en mil, arachide, jachère ou maintenue dénudée du centre Sénégal au cours d'une saison des pluies (1994). Un modèle analogique de ruissellement ‡ stockage de surface (BADER, 1994), dans lequel l'infiltration est une fonction croissante de la lame d'eau en surface du sol est ajusté sur les données. Le modèle présente 3 paramètres: un paramètre de transfert n, un paramètre de ruissellement Hl et un paramètre d'infiltration S. Une analyse de sensibilité menée sur les données de la parcelle de sol nu montre que le paramètre n est le plus sensible des trois.

Le calage numérique des paramètres sur chaque crue au cours de l'hivernage permet d'étudier leur évolution temporelle. Cette évolution est cohérente avec l'occupation de chaque parcelle. Les paramètres n et S de la parcelle de sol nu sont invariants sur la saison tandis que ceux des parcelles en végétation s'écartent progressivement des valeurs obtenues sur sol nu. Pour les parcelles en végétation, les valeurs de S divergent de celles du sol nu lorsque l'indice radiométrique de végétation (N.D.V.I.) servant à l'estimation du couvert dépasse 0.30 - 0.35 environ.

L'évolution des paramètres n et S des parcelles en végétation peut être reliée au temps écoulé depuis le semis (mil, arachide) ou le sarclage initial (jachère) et à l'état d'humectation du sol (pour S). On montre également que le paramètre Hl peut être estimé linéairement à partir d'un indice de rugosité de surface descriptif de la microtopographie.

EN :

The Sudanese climate is characterized by a rainy season and a dry season (mean annual rainfall between 400 and 900 mm). At the end of the dryseason (June in the northern hemisphere), the landscape is completely bare under the effect of animal grazing or soil tillage. During the first rainfalls this leads to high runoff coefficients. These runoff coefficients decrease gradually as the amount of vegetation increases during the growing season (RODIER (1984-1985); ALBERGEL (1988)).

This is particularly true in the Groundnut basin of central Senegal where millet and groundnut are cultivated every other year. As the vegetative cover increases, a system of macropores develops in the soil and preferentially induces water infiltration through mesofauna burrows and along root systems. Hence, many authors have distinguished matrix infiltration governed by the generalized Darcy's law, from preferential infiltration through macropores characterized by a strongly heterogeneous spatial distribution (GERMAN, 1990). These macropores are thought to be responsible for the proportional increase in infiltration with increase in rainfall intensity observed on several experimental plots (BOUCHARDEAU and RODIER, 1960; VALENTIN, 1985; COLLINET,1985; ALBERGEL, 1988). A more complete surface ponding or a differential distribution of the macroporosity in relation with the microtopography can contribute to this phenomenon.

A conceptual runoff model accounting for surface storage, which views infiltration as a function of water depth on the ground surface, is proposed to describe the aforementioned phenomenon under three characteristic vegetative canopies of central Senegal (millet, groundnut and fallow). The model (BADER, 1994) is a distributed, three parameter model that accounts for transfer between spatial elements (parameter n), runoff (parameter Hl) and infiltration (parameter S). The model solves the equation of continuity according to an explicit scheme (forward time). The discharge exiting a spatial element is defined by a power function based on the water depth on the element. The value of the transfer parameter n (dimensionless) depends on the roughness and slope of the soil surface. Parameter Hl (meters) is equivalent to the water depth from which runoff occurs and is found in the discharge expression. Infiltration is defined as the product of the squareroot of the depth of ponded water of a plot and a S parameter (dimensionless) representing surface porosity.

The experimental work took place on 4 rectangular 50m² plots (10 m by 5 m) that were initially bare and weedy. At the beginning of the rainy season, two plots were cultivated in millet and groundnut, one left fallow and the fourth stripped by a powerful herbicide. The runoff was measured by a capacitive gauging system with each tank being equipped with a pressure transducer connected to a datalogger. A tipping bucket raingauge was also connected to the datalogger and rainfall and runoff were recorded simultaneously. The measurements were made to a precision of 4 mm in the tanks (0.16 mm uncertainty for surface runoff depth). With a total seasonal rainfall of 711 mm in 1994, the cumulative surface runoff varied between 40mm for the fallow plot to 150 mm for the bare soil plot. The cultivated groundnut and millet plots had cumulative runoff depths of 55 and 60 mm, respectively. The fallow plot would have had less runoff if it had been more than one-year old. The microtopography of each plot was evaluated using a profile meter. The surface roughness was estimated by the standard-error of measured relative elevations (GUILLOBEZ and ZOUGMORE, 1994). Measurements were taken after each significant rainfall and following tillage operations. The index of roughness varied following vigorous weeding of the groundnut plot to 5 mm for the fallow plot whose microtopography remained constant throughout the season. The development of the vegetative cover was indirectly followed by the calculation of a vegetation index (NDVI) derived from red and near infrared reflectances measured with a field radiometer. Although this index tends to saturate with full ground cover, it nevertheless remains a good indicator at the start of vegetative growth.

The proposed model was used to reproduce measured runoff during several storm events. Calculations were undertaken with a 10-s time step on a 1m-long spatial element with a uniform set of parameters for each plot. A sensitivity analysis was performed for all runoff events on the bare plot. Hydrograph characteristics (runoff volume, peak discharge and time-to-peak) were particularly sensitive to variations in the transfer parameter (n) and to a lesser extent to changes in the infiltration (S) and runoff (Hl) parameters. For the 42 measured runoff hydrographs for all fourplots, the results were excellent: 70% of the simulated hydrographs had a Nash's coefficient greater than or equal to 0.90.

For each plot, the seasonal chronicle of each parameter is coherent with the plot cover. The parameters for the bare plot were invariant throughout the rainy season. However, for the other plots, they varied with the vegetative cover. At the beginning of the growing season, they were similar to those obtained on bare soil and, as the vegetative cover increased, they varied until the NDVI exceeded 0.35 (approximately 20 days after seeding). The evolution of the n and S parameters for the cultivated plots was linearly extrapolated from past events (seeding for the cultivated plots and chemical weeding for the fallow plot) and for S to an antecedent precipitation index. Farming practices that modified surface roughness needed to be accounted for as well. For the transfer parameter (n) of the groundnut plot, an increase of approximately 0.4 was observed when a rainfall event followed weeding. No significant increase was seen for the millet plot. A linear relationship between the index of roughness and the roughness parameter (Hl) was also derived.

FR :

L'identification de pratiques agricoles qui minimisent les risques de contamination des eaux de surface nécessite d'évaluer l'importance des voies de transfert des herbicides vers les cours d'eau. Le but de cette étude est d'évaluer à l'échelle du champ agricole et pour une saison de culture l'exportation effective de l'atrazine et du métolachlore par ruissellement de surface et par drainage, ceci pour des conditions pédo-climatiques et agronomiques représentatives de la culture intensive du maïs-grain dans les Basses-Terres du St-Laurent (Québec). Pour les deux premiers événements pluviaux d'importance suivant l'application des herbicides, seulement deux des six champs étudiés ont présenté un ruissellement quittant le champ : les concentrations en herbicides ont atteint 1200 mg/L et 2400 mg/L. La charge exportée en herbicides semble inférieure dans le cas du non travail du sol (semis direct), comparativement au labour conventionnel. Les concentrations en herbicides dans l'eau de drainage sont inférieures à 6 mg/L (pour la majorité inférieures à 1-2mg/L) pour quatre champs, alors que deux champs ont présenté des concentrations atteignant 40-60 mg/L. La charge exportée par drainage apparaît être faible dans le cas de l'application d'herbicides en bandes, comparativement à l'application en surface totale. La masse en herbicides exportée par ruissellement (estimée à partir de coefficients de ruissellement probables) serait supérieure à celle par drainage. Une démarche destinée à diminuer les masses en herbicides exportées devrait ainsi viser la principale voie de cette exportation, c'est-à-dire le ruissellement de surface.

EN :

The use of pesticides in agriculture may result in the degradation of surface water quality. Since agricultural practices affect the transport of pesticides, there is a need to identify practices which minimize the contribution of the different transport paths to the streams, i.e. runoff and drainage. The aim of this study was to evaluate at the field scale and for one growing season the transport of the herbicides atrazine and metolachlor to surface water under soil, climatic and agricultural conditions representative of those encountered for intensive corn cropping in the St-Lawrence Lowlands (Quebec).

Six agricultural fields (Figure 1) were studied in 1995. Previous agricultural practices in 1994 and soil texture are summarized in Tables 1 and 2, respectively. Conventional practices (tillage with moldboard plow and application of herbicides over the entire area of the field) and conservation practices (no-till and banded application of herbicides over the seeded row) were studied. Each field was solely and entirely drained by one subsurface drain. The commercial formulation used in 1995 contained a mass of metolachlor two times higher than that for atrazine. Herbicide concentrations in runoff and drainage waters were monitored during the two first important rainfall events that occurred after herbicide application (Table 3). Sampled runoff corresponded to the water reaching a drainage channel or a stream. Drainage water was also collected for 3.5 - 4.5 months following the initial application. A total of 164 water samples was obtained. After sediment removal, metolachlor, atrazine and its dealkylated metabolite deethylatrazine (DEA) were extracted using a liquid-solid extraction procedure and analyzed by gas chromatography.

Only two fields produced runoff and the concentrations of parent-compounds (Figures 2 and 3) were high and varied during rainfall events between 60-500 mg/L (Field 2) or 130-2400 mg/L (Field 6). Concentrations found during the first rainfall event were higher than those encountered during the second event. The DEA/atrazine concentration ratio (DAR) was below or near 0.1, indicating runoff of recently applied atrazine (low degradation). These two fields present similar soil texture, pluviometry and sampling periods after herbicide application. Based on runoff coefficients observed for other agricultural fields (1-30%), it was estimated that the mass losses for herbicides (Table 4) would be higher under conventional tillage(Field 6) as compared to no-till (Field 2).

Significant transport of herbicides by drainage was observed during the two rainfall events. The losses of herbicides that occurred after these events and during a dry growing season (little or no drainage flow) were low. The drainage losses (concentration or masses) during the two rainfall events for Field 1 (clay) were very low. This was attributed to the low drainage capacity of the soil, to the low rainfall intensities as well as to the important delay between the initial application and the subsequent rainfalls. For silty clay loam to loam soils, the drainage flow increased in the 6-12 h period following the onset of rainfall, as did the herbicide concentrations. Metolachlor concentrations were slightly higher or close to those for atrazine: this was attributed to its possible more rapid decay and to its stronger tendency to adsorb to the soil.

During the rainfall events, four fields exhibited herbicide concentrations from drainage less than 6 mg/L (mostly < 1-2 mg/L). Fields 2 and 6 yielded parent-compound concentrations as high as 40-60 mg/L (Figures 4 and 5). The DAR values found for drainage water of Field 2 (0.1-0.5) were higher than those observed from runoff, indicating significant dealkylation of atrazine had occurred during its transport in the unsaturated zone. Field 6 allowed the monitoring of the DAR over the growing season and an inverse relationship was found between the DAR and atrazine concentration (Figure 6). This was attributed to the larger variation in atrazine concentration during a rainfall as compared to that of DEA. A DAR value near 1 was obtained at 1-2 months after application, indicating important degradation of atrazine.

The total mass losses of parent-compounds (two rainfall events) were evaluated (Table 5) except for Fields 2 and 4 which present frequent submerged drains. Banded herbicide application (Field 5) results in consistent lower losses of herbicide masses than those obtained for application over the entire surface (e.g. Field 5 compared to Fields3 and 6). It should be noted that the higher export observed for the entire surface application may be partly attributed to a shorter delay between application and rainfalls (Fields 3 and 6) or to a higher rainfall intensity (Field 6).

Although runoff reaching surface waters was limited, it was estimated that the total herbicide losses (Table 4) during the two rainfall events were higher than those from drainage (Table 5). In the perspective of reducing the herbicide loads reaching streams, it appears that remedial actions should focus on this main route of transport. Thus, complementary actions such as vegetated buffer strips to intercept crop land runoff may possibly be useful to limit herbicide transfer to streams in intensive agricultural zones.

FR :

Dans le domaine de la prévision des débits, une grande variété de méthodes sont disponibles: des modèles stochastiques et conceptuels mais aussi des approches plus novatrices telles que les réseaux de neurones artificiels, les modèles à base de règles floues, la méthode des k plus proches voisins, la régression floue et les splines de régression. Après avoir effectué une revue détaillée de ces méthodes et de leurs applications récentes, nous proposons une classification qui permet de mettre en lumière les différences mais aussi les ressemblances entre ces approches. Elles sont ensuite comparées pour les problèmes différents de la prévision à court, moyen et long terme. Les recommandations que nous effectuons varient aussi avec le niveau d'information a priori. Par exemple, lorsque l'on dispose de séries chronologiques stationnaires de longue durée, nous recommandons l'emploi de la méthode non paramétrique des k plus proches voisins pour les prévisions à court et moyen terme. Au contraire, pour la prévision à plus long terme à partir d'un nombre restreint d'observations, nous suggérons l'emploi d'un modèle conceptuel couplé à un modèle météorologique basé sur l'historique. Bien que l'emphase soit mise sur le problème de la prévision des débits, une grande partie de cette revue, principalement celle traitant des modèles empiriques, est aussi pertinente pour la prévision d'autres variables.

EN :

A large number of models are available for streamflow forecasting. In this paper we classify and compare nine types of models for short, medium and long-term flow forecasting, according to six criteria:

1. validity of underlying hypotheses,

2. difficulties encountered when building and calibrating the model,

3. difficulties in computing the forecasts,

4. uncertainty modeling,

5. information required by each type of model, and

6. parameter updating.

We first distinguish between empirical and conceptual models, the difference being that conceptual models correspond to simplified representations of the watershed, while empirical model only try to capture the structural relationships between inputs to the watershed and outputs, such as streamflow.

Amongst empirical models, we distinguish between stochastic models, i.e. models based on the theory of probability, and non-stochastic models. Three types of stochastic models are presented: statistical regression models, Box-Jenkins models, and the nonparametric k-nearest neighbor method. Statistical linear regression is only applicable for long term forecasting (monthly flows, for example), since it requires independent and identically distributed observations. It is a simple method of forecasting, and its hypotheses can be validated a posteriori if sufficient data are available. Box-Jenkins models include linear autoregressive models (AR), linear moving average models (MA), linear autoregressive - moving average models (ARMA), periodic ARMA models (PARMA) and ARMA models with auxiliary inputs (ARMAX). They are more adapted for weekly or daily flow forecasting, since the yallow for the explicit modeling of time dependence. Efficient methods are available for designing the model and updating the parameters as more data become available. For both statistical linear regression and Box-Jenkins models, the inputs must be uncorrelated and linearly related to the output. Furthermore, the process must be stationary. When it is suspected that the inputs are correlated or have a nonlinear effect on the output, the k-nearest neighbor method may be considered. This data-based nonparametric approach simply consists in looking, among past observations of the process, for the k events which are most similar to the present situation. A forecast is then built from the flows which were observed for these k events. Obviously, this approach requires a large database and a stationary process. Furthermore, the time required to calibrate the model and compute the forecasts increases rapidly with the size of the database. A clear advantage of stochastic models is that forecast uncertainty may be quantified by constructing a confidence interval.

Three types of non-stochastic empirical models are also discussed: artificial neural networks (ANN), fuzzy linear regression and multivariate adaptive regression splines (MARS). ANNs were originally designed as simple conceptual models of the brain. However, for forecasting purposes, these models can be thought of simply as a subset of non linear empirical models. In fact, the ANN model most commonly used in forecasting, a multi-layer feed-forward network, corresponds to a non linear autoregressive model (NAR). To capture the moving average components of a time series, it is necessary to use recurrent architectures. ANNs are difficult to design and calibrate, and the computation of forecasts is also complex. Fuzzy linear regression makes it possible to extract linear relationships from small data sets, with fewer hypotheses than statistical linear regression. It does not require the observations to be uncorrelated, nor does it ask for the error variance to be homogeneous. However, the model is very sensitive to outliers. Furthermore, a posteriori validation of the hypothesis of linearity is not possible for small data sets. MARS models are based on the hypothesis that time series are chaotic instead of stochastic. The main advantage of the method is its ability to model non-stationary processes. The approach is non-parametric, and therefore requires a large data set.

Amongst conceptual models, we distinguish between physical models, hydraulic machines, and fuzzy rule-based systems. Most conceptual hydrologic models are hydraulic machines, in which the watershed is considered to behave like a network of reservoirs. Physical modeling of a watershed would imply using fundamental physical equations at a small scale, such as the law of conservation of mass. Given the complexity of a watershed, this can be done in practice only for water routing. Consequently, only short term flow forecasts can be obtained from a physical model, since the effects of precipitation, infiltration and evaporation must be negligible. Fuzzy rule-based systems make it possible to model the water cycle using fuzzy IF-THEN rules, such as IF it rains a lot in a short period of time, THEN there will be a large flow increase following the concentration time. Each fuzzy quantifier is modeled using a fuzzy number to take into account the uncertainty surrounding it. When sufficient data are available, the fuzzy quantifiers can be constructed from the data. In general, conceptual models require more effort to develop than empirical models. However, for exceptional events, conceptual models can often provide more realistic forecasts, since empirical models are not well suited for extrapolation.

A fruitful approach is to combine conceptual and empirical models. One way of doing this, called extended streamflow prediction or ESP, is to combine a stochastic model for generating meteorological scenarios with a conceptual model of the watershed.

Based on this review of flow forecasting models, we recommend for short term forecasting (hourly and daily flows) the use of the k-nearest neighbor method, Box-Jenkins models, water routing models or hydraulic machines. For medium term forecasting (weekly flows, for example), we recommend the k-nearest neighbor method and Box-Jenkins models, as well as fuzzy-rule based and ESP models. For long term forecasting (monthly flows), we recommend statistical and fuzzy regression, Box-Jenkins, MARS and ESP models. It is important to choose a type of model which is appropriate for the problem at hand and for which the information available is sufficient. Each type of model having its advantages, it can be more efficient to combine different approaches when forecasting streamflow.

FR :

L'étude hydrochimique des eaux de la nappe libre de la Chaouia côtière montre une teneur excessive en sels dans ces eaux, spécialement des chlorures (jusqu'à 3 g.l-1).Cette salinité élevée peut, à terme, nuire gravement à l'économie de la région qui se consacre essentiellement à la culture maraîchère sous irrigation.

Des mesures isotopiques (18O/16O, 2H/1H) couplées aux éléments chimiques, notamment Br-/Cl-, excluent l'hypothèse d'une intrusion marine généralisée. L'existence d'une paléosalinité (SHIVANNA andal.,1993) n'est pas soutenue par les mesures des teneurs en isotopes radioactifs (3H, 14C). Il en est de même pour le lessivage de dépôts chlorurés préexistants dans les sédiments. L'apport essentiel en chlorures provient de l'altération de la roche mère, des fertilisants et des embruns marins lessivés par suite de l'irrigation et entraînés vers la nappe par percolation. Le recyclage de l'eau souterraine depuis au moins trente ans, par l'irrigation, n'a fait qu'augmenter cette salinité.

Toutefois en bordure de l'océan quelques puits semblent montrer la présence d'eau marine, en effet les débits d'exhaure des puits sont à la limite d'exploitation de la nappe et toute augmentation de débits dans ces puits engendrera fatalement une invasion marine généralisée.

EN :

In West of Morocco, the coastal plain Chaouia is located between Casablanca and Azemmour (south-west of Casablanca) over a distance of 65km (Figure 1). This plain aquifer spreads over 1100 km² in semi-arid climatic conditions with about 370 mm.year-1 of mean precipitation and 17°C in mean air temperature.

Irrigated agricultural farming is the main economic resource of the region. The only source of water is provided by highly mineralized groundwaters, which are harmful for the rural population and agricultural irrigation. The chloride content represents the main contribution of the salinity of ground waters. Most of the previous authors and authorities have attributed the high mineralization to seawater intrusion (YOUNSI, 1994). The objective of the present study is to determine the cause of salinity, considering three hypothesis: seawater intrusion, leaching of salts and evaporation of water.

The environmental isotopes (2H,18O, 3H, 14C) have been used with hydrogeology together with major and minor ion chemistry to identify the source of salinity and to estimate the residence time of ground water. Moreover an environmental isotope study was carried out in order to identify the origin of the salinization (MERLIVAT et al., 1970; COTECCHIA et al., 1974; GASPARINI, 1989; CABRAL et al., 1991; GOMEZ-MARTOS et al., 1993; HASHASH et al., 1995).

Both isotopic studies and Br-/Cl- ratios have been used to identify the salinity origin. The potential sources for this salinization are: seawater encroachment as a consequence of intense exploitation of the aquifer, dissolution of the halite from the several diapiric structures intruded in the aquifers. The isotopic waters sampling were carried out in August 1993 on several wells (Figure 2), in order to study the variations of the oxygen-18, deuterium, tritium and carbon-14 isotopes. Table 1 shows the results of the isotope and chemical analyses carried out. As oxygen-18 data for local precipitation are not available in Morocco, the oxygen-18 and deuterium mean contents in precipitation were taken from stations of the International Atomic Energy Agency (IAEA) network near the studied area (Gibraltar, Faro, Ponta Delgada) and/or deduced from published isotopic composition of shallow groundwaters in Morocco (KABBAJ et al., 1978; LOUVA and BICHARA, 1990; BOUCHAOU et al., 1995); the mean values for the precipitation are -4.0 ± 0.5‰ and -23.0 ± 2 ‰ (vs. V-SMOW) respectively. The oxygen-18 and deuterium mean contents in ground waters are -3.7 ± 0.3 ‰ and -20.1 ± 1.8 ‰ (vs.V-SMOW) respectively. The relationship between the oxygen-18 and deuterium contents of some groundwater samples indicates a good fitting along the world meteoric line (CRAIG, 1961) (Figure 3), suggesting that the ground waters are mainly of meteoric origin. There is no indication of possible mixing with seawater because there is no indication of a relation between the oxygen-18 content of the wells and their distance from the sea (Figure 4). On the other hand, no isotopic enrichment is noted during the infiltration of water from precipitation and/or irrigation.

The diagram d18O-Cl- (Figure 5) shows that all the data do not lie along a theoretical mixing line with seawater and/or an evaporation line. Every 18O data point is very close to the isotopic values of precipitation [-4.0± 0.5 ‰ (vs. V-SMOW)] but with a large scatter in Cl- content. In this zone, an increase in the salt concentration is not accompanied by an isotopic effect. The sources of the chloride must be found among the products of rock alteration, fertilizers and marine airborne salts. The recycling of salt water by irrigation increases the salinity of soil and ground waters.

Halite is characterized by a very low Br-/Cl- ratio(0.183 x 10-³). Water in some wells presents Br-/Cl- ratios similar to those of halite, but most of the groundwater samples have a greater ratio. These results argue against the hypothesis of dissolution of halite in the aquifer formation. But the Br-/Cl- ratio alone is not sufficient for detecting a possible seawater intrusion.

During August 1994, 20 drilled wells were sampled over the study area for determining tritium concentrations. Carbon-14 activities were measured only on some samples. The relatively high tritium content in the majority of the sampled waters [between 2 and 9 tritium units (TU)] indicates a recent groundwater recharge (Table 1). The use of natural radioactive isotopes (3H, 14C) in the ground waters gives an estimate of mean residence time of 50 ± 20 years (mixing model). Some waters show a mean residence time greater than 100 years because they are isolated from the general circulation and limited to local spots. The recent 14C activity of the groundwater (_100 pcm) allows us to reject the hypothesis of a mixing with paleoconnate waters as demonstrated elsewhere (SHIVANNA et al., 1993).

In the case of the coastal Chaouia (semi-arid environment), the use of both chemical tracers (Cl-, Br-) and isotopes (18O, 2H,3H, 14C) has shown that rainfall constitutes the essential origin of ground water in the area. The high salinity caused by seawater intrusion is probably limited in space to some coastal wells; neither evaporation of water before and during percolation, nor dissolution of evaporitic deposits (halite), has been found. The past marine transgression salt origin has also been discarded. The main source of the salinity is the washout of the rock alteration salts, fertilizers, marine airborne salts, and the unusable salts rejected by plants in the soil. Thus the increase of salinity is well explained by the recycling of salty ground waters by irrigation during at least thirty years.

FR :

La présence de métaux dans les boues municipales demeure l'obstacle majeure à leur valorisation comme fertilisant agricole ou forestier. L'utilisation des boues en milieu forestier diminue certes les risques de contamination de la chaîne alimentaire humaine. Cependant, nos connaissances sur le comportement à long terme des métaux ajoutés aux sols forestiers sont encore très limitées. Le risque de lessivage des métaux vers les écosystèmes aquatiques situés en aval des zones traitées doit être évalué. Cet article propose une comparaison des risques de contamination de l'eau selon que l'application des boues s'effectue en forêt ou en milieu agricole. Les concentrations de métaux dans les boues municipales, leurs formes et leur mobilité dans les sols amendés sont passées en revue. Et finalement, les normes québécoises qui ont été établies pour protéger le milieu récepteur contre les métaux présents dans les boues sont comparées à celles recommandées par l'U.S. EPA.

EN :

The presence of metals in municipal sewage sludge is still a major obstacle to their land application. Forest land application of sludge reduces the possibility of metals entering the human food chain. Present knowledge of their long-term behavior in forest soil is however limited. The risk of metals leaching to the aquatic ecosystem situated downstream from treated zone must be evaluated. This paper present a comparison of the risks of water contamination when sludges are applied to forests instead of cropland. The levels of metals in municipal sludge, their form and their mobility in sludge-treated soil are reviewed. Finally, the regulations of the Province of Quebec that have been established to protect forest land are compared to the recommendations of the U.S. Environmental Protection Agency.

Although sludge applications to forest land reduce the risk of human foodchain contamination, properties of forest soils (low pH, low cation exchange capacity (CEC), presence of large continuous pores) could represent an increased hazard of metals leaching to aquatic ecosystems. Metal concentrations in sludge vary considerably depending on the source of the sludge. Metals found in sludge are predominantly associated with the solid phase, while soluble and exchangeable forms generally represent less than 10% of the total metals.

Short term studies have revealed that metals tend to accumulate in the upper layers of soil, but few studies have evaluated long-term metal mobility or bioavailability. Many factors can contribute to an increase mobility (pH, oxidation-reduction potential, organic matter decomposition). The risk of aquatic ecosystem contamination by aluminum leached from sludge-treated soils hasn't yet been evaluated. It seems, however, that the risk exists if the soil become more acidic after sludge decomposition.

Limits on the amount of metals allowed in sludge have been fixed by many countries. Nevertheless, given the wide variation in the standards used to regulate the allowable content of metal in sludge or sludge-treated soils, disagreement presently exist among regulatory agencies. For example, the maximum cadmium limit allowed in sludge destined to be used as forest fertilizer in Quebec is 15 mg kg-1, whereas it is 85 mg kg-1 in the United States. Different philosophical views on environmental protection (zero tolerance or tolerance of a certain level of metals in soil), as well as the organism chosen when standards are set, are responsible for this inconsistency.

Before sewage sludge application to forests becomes a common practice, the associated risks of metals leaching to the aquatic ecosystem must be evaluated. Long-term studies should be undertaken in order to establish safe standards for applying sludge in a forest environment. Such additional efforts will enable good quality sludge to be considered as a usable and valuable resource in forestry.

FR :

L'objectif de ce travail est d'étudier dans un premier temps la dynamique du phosphore dans un système de traitement d 'eaux usées par lagunage facultatif, en relation avec les compartiments biologiques: phytoplancton, zooplancton et bactéries, dans un deuxième temps, d'évaluer, sous les conditions du climat aride, les performances épuratoires de ce système en ce qui concerne les formes du phosphore. Un intérêt particulier est porté au phénomène du relargage du phosphore par les sédiments.

Les résultats montrent qu'il y a une évolution saisonnière et cyclique du phosphore dans les bassins de lagunage, en étroite corrélation avec le phytoplancton et le zooplancton. Le phytoplancton représente la part importante des formes organiques du phosphore particulaire. Les meilleures performances d'élimination du phosphore sont notées au printemps (77% pour les PO4 et 64 % pour le phosphore total). La période la moins performante est la période automne-hiver où les rendements moyens ne dépassent pas 38%. Le phosphore éliminé est stocké dans les sédiments. En période estivale, des phénomènes de relargage du phosphore à partir des sédiments ont été mis en évidence, favorisés par une zonation verticale des eaux à cette période. Des essais aulaboratoire ont montré la grande aptitude des sédiments de la station de lagunage étudiée à relarguer le phosphore, avec intervention de l'activité biologique du sédiment.

EN :

Assimilation, sedimentation and release are the principal mechanisms controlling the behavior of phosphorus in aquatic ecosystems. There are numerous investigations of the phosphorus cycle in lakes and rivers in relation to the eutrophication process. There are, however, few studies about phosphorus cycling in waste stabilization ponds and most of these have discussed phosphorus removal.

The aims of this paper were: 1) to study phosphorus dynamics in relation to the biomass compartment (zooplancton, phytoplancton and bacteria); and 2) to evaluate the phosphorus removal efficiency of the treatment plant under arid climatic conditions. Of particular interest was the remobilization of phosphorus from the pond sediment, both under laboratory conditions and in the field.

The experimental installation consisted of two ponds receiving domestic wastewater (mean flow=120.4+5.8 m³.day-1 and specific organic load=56 kgBOD.ha-1. day-1; pond area=0.5 ha and depth=1.5 m). Samples were taken twice monthly from the surface, B1(S) and B2(S), and from the 1.5 m layers, B1(F) and B2(F). Composite (24 h) samples were taken from the inflow and the outflowof the plant to determine the removal efficiencies. In each sample, soluble reactive phosphorus (PO4), total phosphorus (PT) and total soluble phosphorus (PTD) were determined. Particulate phosphorus (PTP) was calculated as (PT-PTD) and soluble organic phosphorus (POD) as (PTD-PO4).

Phosphorus release was measured in a plexiglas flask containing sediment (collected by corer from the pond bottom) overlaid with distilled water. Over 15 days, daily measurements were made of PO4 and total phosphorus. Sediment dry weight (%), volatile matter, organic carbon (%) and total phosphorus were measured before and after each test. Phosphorus contents of the phytoplankton, zooplankton and bacteria were estimated using the methods of BOUGIS (1974).

The results show that cyclical and seasonal variations of phosphorus content were closely correlated with variations in the biological components. Significant correlation coefficients were noted between chlorophyll-a and particulate organic phosphorus. Organic phosphorus in algal cells was the more important form of particulate organic phosphorus.

Under the arid climate, the waste stabilization ponds were more efficient at phosphate removal during the spring-summer period (77% of PO4 and 64% of total phosphorus in spring). The first pond may play the major role in phosphate removal. Phosphorus is only eliminated from stabilization ponds through accumulation in the sediment. However, part of the phosphorus which accumulates in pond sediments remains potentially mobile. During the summer period, phosphorus release from the sediment, observed in situ, was favored by stratification of the overlying water. The same phenomenon was noted during the autumn, particularly the release of POD. Laboratory studies have shown that phosphorus release may amount to 55 mg P m-2d-1. Biological activity may play a significant role in this mobilization.

FR :

L'étude de la stationnarité de séries de débits annuels et saisonniers de 15 rivières roumaines et du Danube à Orsova met en évidence l'existence d'une "phase hydrologique" sèche importante depuis 1983. L'écoulement hivernal a diminué de 50% depuis cette date et semble principalement responsable de cette sécheresse hydrologique. L'extension géographique de cette sécheresse est précisée en Roumanie Centrale. L'analyse des débits du Danube montre qu'une partie importante de son bassin versant est touché par le phénomène.

EN :

A statistical study of annual and seasonal discharges series for 15 Romanian rivers and the Danube River at Orsova exhibits a very important dry hydrological period which began around 1983. The winter discharge decreased 50% during this period and seems to be the main cause of this hydrological drought. The spatial distribution of this drought is defined in Central Romania. The analysis of the Danube discharge series demonstrates that an important part of its watershed is in the same situation.