Volume 15, numéro 4, 2002

FR :

Ce travail explore les performances de deux types de membranes de nanofiltration (Desal DK et NF200) dans l'élimination dans les eaux de certains pesticides (l'atrazine et son métabolite la déséthylatrazine (DEA), la simazine, la cyanazine, l'isoproturon et le diuron) et évalue l'influence de la présence de matière organique ou inorganique dans la matrice d'eau sur l'efficacité de ce traitement.

Des eaux synthétiques, composées à partir d'eau distillée à laquelle a été ajoutée de la matière organique (acides humiques) ou inorganique (CaCl₂ ou CaSO₄), ont été traitées sur un pilote de nanofiltration durant 96 heures. Les taux rétention en pesticides et ceux de leur adsorption sur les membranes ont été calculés et comparés aux résultats obtenus sur une matrice d'eau distillée pure. Une influence du type de membrane et de la présence de la matière humique sur le taux d'abattement de certains pesticides a été constatée. L'influence de la matière inorganique est pratiquement insignifiante.

EN :

The intensive use of pesticides in agriculture has resulted in the contamination of groundwater and surface waters. The removal of these organic pollutants by the usual methods such as adsorption by activated carbon (in powdered or granular form) or oxidation by ozone have some disadvantages. Recently, the removal of organic pollutants by membrane retention (reverse osmosis, ultrafiltration and nanofiltration) has become increasingly popular and due to its low cost, nanofiltration has become an interesting option.

This study examined the efficiency of two different nanofiltration membranes (Desal DK and NF200) in the removal of some pesticides (atrazine and its metabolite desethylatrazine (DEA), simazine, cyanazine, isoproturon and diuron) from water and, in addition, investigated the influence that organic and inorganic matter may have on the efficiency of this removal. Synthetic waters were made from distilled water and organic matter (humic acids) or inorganic matter (CaCl₂ or CaSO₄) was added, as well as 1 µg/l of each pesticide. The samples were then filtered by a nanofiltration pilot for 96 hours. Samples of the feed, permeate and the retentate were taken after 4, 24, 48, 72 and 96 h. The samples were replaced with equivalent volumes of the original solution in the feed tank. The different samples were analysed by an on-line SPE / HPLC system. The different concentrations obtained allowed the determination of the proportion of the pesticides that adsorbed to the membrane.

The removal efficiency of pesticides from pure distilled water differed according to the membrane. For example, the Desal DK membrane eliminated more than 90% of all the pesticides (with the exception of diuron). In contrast, the NF200 membrane, eliminated more than 75% of all the pesticides (with the exception of diuron). The removal efficiency of Diuron was the lowest by both the membranes: 70 % by Desal DK and 45 % by NF200. The adsorption efficiency of the pesticides was similar for both membranes (between 30 and 40%). In pure water, pesticide removal is a function of both the specific properties of each pesticide (solubility, molecular mass, Stokes diameter, equivalent molar diameter, calculated equivalent diameter and polarity) and the physical characteristics of the membrane (molecular weight cut-off and current load).

The influence of inorganic matter on pesticide removal efficiency changed according to the type of membrane. For example, we noted an improvement in removal efficiency with the NF200 membrane from low removal with CaCl₂ to high removal with CaSO₄ for all pesticides examined including diuron. In contrast, for the Desal DK membrane, very little change was noted (a slight decrease in the removal efficiency of DEA and simazine with CaCl₂). Adsorption by the membranes remained stable for all the pesticides tested on the two types of membrane. The improvement in the removal of pesticides by the NF200 membrane was probably linked to pores being blocked by ions at high concentrations. It could be concluded from these results that elimination of pesticides is quantitatively linked to the physical characteristics of the membranes and that inorganic matter only has an effect with wide-pore membranes (NF200 membrane) and, furthermore, it has no influence on the adsorption of the pesticides on the membranes.

For water containing organic matter, we have noted an improvement in the removal of certain pesticides with the NF200 membrane when compared to distilled water (except diuron). With the Desal DK membrane, we observed a slight decrease in the removal of DEA, simazine and isoproturon, and a substantial drop for diuron (20 %) with no change for cyanazine and atrazine. Adsorption of the pesticides on the membranes remained unchanged with the NF200 membrane but increased by about 10% on the Desal DK membrane for all molecules. Pesticides, notably triazines, adsorb easily on organic matter by physiosorption (weak links) and by chemisorption (ionic links) to form macromolecules. The steric congestion and the density of these pseudo-complexes is high, which facilitated the elimination of certain pesticides with the (wide-pore) NF200 membrane by accentuating the effects of steric exclusion and electrostsatic repulsion and decreasing adsorption. For the Desal DK membrane, the increased adsorption of the pesticides on the membrane generated a more significant transition of these molecules in the direction of the permeate. This had a negative influence on the removal of some pesticides, depending on their size; the largest molecules underwent the least change. Diuron behaved differently from the other pesticides examined. This molecule did not bind to humic acids and its removal rate did not change with a wide-pore (NF200) membrane. However a greater adsorption of organic matter by the narrower-pore (Desal DK) membrane favored diuron adsorption and, consequently, its diffusion into the permeate. The effect of organic matter and, more specifically, of humic acids on the elimination of pesticides depends not only on the structure of the molecules, but also on the cut-off threshold of the membrane.

The two main mechanisms that govern the process of pesticide elimination by NF are repulsion (steric and electrostatic) by the membrane and adsorption on the membrane. This latter phenomenon must be minimized, to reduce the elimination of pesticides by fostering their transition in the direction of the permeate. In addition, removal of the pesticides by NF was favoured by the high-molecular weight organic matter fraction (i.e., humic acids). The influence of the inorganic matter (CaCl₂ and CaSO₄), for its part, is greater with the wide-pore membrane.

FR :

Alors que l'irrigation paraît être le recours essentiel face aux besoins croissants de la population mondiale en céréales, la dégradation des sols et des eaux annihile presque totalement les efforts d'aménagement. Aussi l'étude de la dégradation des sols dans la moyenne vallée du fleuve Sénégal constitue l'un des quatre objectifs du pôle de recherches sur les systèmes irrigués sahéliens qui regroupe quatre pays dont le Mali, la Mauritanie, le Niger et le Sénégal. Le suivi de l'évolution des eaux et des sols sous culture apparaît indispensable pour garantir une agriculture irriguée rentable et durable dans la zone. Dans cette région, les risques de dégradation des sols font suite à l'effet de deux processus de dégradation que sont l'alcalinisation et la salinisation des sols. Ces types de dégradation sont fortement corrélés à la remontée des nappes d'eaux souterraines. Dans cet article, l'étude des transferts d'eau dans les parcelles irriguées par le biais d'un suivi expérimental " in situ " met en évidence les relations existantes entre les eaux d'irrigation et les mouvements de la nappe. L'étude a permis une meilleure compréhension de la dynamique hydrique au sein de la parcelle : saturation du profil et processus de recharge de la nappe. Elle a également permis de montrer que les échanges avec la nappe pendant la période d'irrigation sont négligeables. Cette pratique de la riziculture inondée, entraîne une variation de stock importante qui représente plus de 40% des apports et qui se traduit par une remontée de nappe de près de 2 mètres.

EN :

Whereas irrigation appears to be the main approach to satisfy the growing worldwide demand for cereal, soil and water degradation continues to be an on-going problem in agriculture development. One of the main four concerns of the regional group for research on Sahelian irrigated systems (including Mali, Mauritania, Niger and Senegal) is soil degradation in the middle Senegal River valley. These soils are subject to various forms of degradation, mainly from salinisation and/or alkalinisation. These degradation processes are strongly correlated with water table dynamics, with water level fluctuations being significant. Therefore, to guarantee sustainable development of irrigated agriculture in the area, irrigation must be coupled with complete and permanent monitoring of soil and water quality.

In this paper, we present a complete study concerning water transfer in irrigated plots and its effects on the groundwater table. The experimental site is located in the Podor region, at 16°.37'N, 14°.52'W in the Donaye irrigated area. The surface area is about 50 ha. Water supply is assured by filling a main channel using a group of pumps on the Doué River. Irrigation of the parcels is performed with siphons from this channel. The experimental plot of 0.33 ha is used for an underwater rice crop. One or two rice crop production harvests are made every year, with the decision been taken by the farmers. There is no drainage system in the area.

The experimental plot was equipped with 8 piezometers located along a stream line. One is situated between the plot and the river in order to study the water movements caused by the exchange between the groundwater and the water in the river. Three are situated in the plot, one close to the Doué River, one in the middle of the plot and one near a dam, which is the opposite boundary compared to the river. Four other piezometers are located beyond the dam to estimate groundwater input and output at this boundary. Five tensiometers at 20, 40, 60, 95 and 135 cm depth are placed close to the piezometers located in the plot. Four water content profiles were measured during the irrigation period at depths 20, 40, 60 95 and 135 cm and the irrigation was performed over 84 days.

The water table level variations at 2 meters were recorded. The groundwater inflow decreased during the first part of the irrigation period due to infiltration below the irrigated experimental plot. This inflow increased during a second period due to water level variations in the river close to this site. At the output, the head gradient did not vary appreciably and the outflow was assumed to be constant during the observation period. Moreover, this gradient is quite low and the flow rate is very low. The water content and pressure profiles clearly show the infiltration of water in the soil during irrigation. The head gradients show the water movement in the unsaturated soil during infiltration and evaporation. Upward flow due to evaporation is observed at about 10 days after irrigation. The last measured water profile (76 days after the end of irrigation) shows that evaporation modifies the water content profile until at least a depth of 120 cm.

The water balance during the irrigation period showed that the input due to irrigation and precipitation was equal to 4150 m3. The evapotranspiration output was estimated to be 2370 m³. Groundwater exchange at the downstream boundary can be neglected during the duration of irrigation (84 days), since the average hydraulic gradient remained low (less than 0.8 %) and the hydraulic conductivity of the aquifer was not important (about 250 cm/day). It was assumed that the water storage quantified with the rise of the groundwater level was equal to the difference between surface input and evapotranspiration (more than 40 % of the contributions), leading to a 27 % change in water content, which is quite reasonable for this type of soil (clay).

The stored water was then recovered by evapotranspiration and groundwater outflow at the plot boundaries. We are away from a reasonable irrigation that would reduce the used water quantity and decrease the risk of soil degradation. This study allows a better understanding of the water dynamics in the experimental plot, which includes soil saturation, recharge processes, and exchanges between the aquifer and the river.

FR :

La présente étude s'inscrit dans le cadre d'un programme de recherche sur la gestion des lacs et barrages collinaires en zone méditerranéenne HYDROMED. Elle a pour objectif le suivi hydrologique et la sédimentation du barrage collinaire Saboun, situé dans la zone de Tanger au Nord-Ouest du Maroc. Un dispositif hydro-pluviométrique automatique a été installé à ce sujet depuis novembre1997. Il a permis le suivi du bilan en eau de la retenue et l'analyse de son fonctionnement hydrologique. Parallèlement, une étude de la bathymétrie par sondage ponctuel a montré que le taux de sédimentation est de l'ordre de 2% par an. Cette perte en eau de la capacité totale de la retenue est une conséquence directe d'une forte dégradation des sols du bassin versant estimée à 115 T/ha/an.

EN :

It is proposed that many Mediterranean countries could make better use of their water resources by constructing small hill reservoirs. Since 1980, Morocco has built more than 158 small dams in order to meet the requirements of rural population water use. Due to both environmental degradation and bad management, these small infrastructures are now in a critical situation. Furthermore, pollution and siltation, resulting in a deterioration of water quality and a reduction of water capacity, are problems at these sites.

The present study is part of the HYDROMED research program on the hydrological balance of small dams in the Mediterranean area. Its objective is the monitoring of hydrological balance and siltation of a small dam, Saboun, located in Tangier in Northwest Morocco. A remote hydrological station was installed in November 1997. This station allowed the examination of the dam water balance and analyses of its hydrological function.

The analyses of water balance demonstrated that the volume of water stored in the Saboun reservoir increased rapidly in November 1997, and was then followed by a progressive decrease. The load factor varied from 1.9 in 1997/98 to approximately 2 times the current capacity of reservoir, to only 0.3 in the following year 1998/99. These results show that even for a short hydrological period (2 years), the hydrological balance of the Saboun reservoir varied greatly as a result of the irregularity of its hydrological parameters. The essential reasons for this variability were the rainfall characteristics that control the runoff from the watershed and bad management practices.

In addition, a bathymetric survey was carried out by a punctual method to estimate the rate of dam siltation and sedimentation from the watershed. The results of this control demonstrated that the dam siltation rate was about 2 % per year. This loss of total water retention capacity of the dam is the consequence of high specific soil degradation of the watershed. The rate of sediment delivered from watershed and deposited in the reservoir was estimated at 37 T/ha/year. The total soil erosion for watershed was estimated at 115 T/ha/year by integrating both the sediment evacuated at the time of dam overflow and the sediment deposited in stream flow. This erosion, which comes essentially from cultivated soil (95 % of the watershed surface), may have serious effects on the hydrological balance of the dam, particularly by reducing the reservoir capacity of the dam, and on the reservoir water quality.

FR :

Le traitement des effluents urbains par réacteurs discontinus séquentiels (SBR : Sequencing Batch Reactor) constitue une solution alternative aux traitements par systèmes à boue activée en effectuant le traitement du carbone, la séparation liquide solide et l'élimination des nutriments au sein d'un bassin unique grâce à une gestion adéquate des cycles temporels de réaction. L'alternance de phases aérées et anoxiques suivie d'une période de décantation conduit en théorie à l'élimination quasi totale des ions nitrate formés lors de la phase de nitrification aérobie. Cependant, selon la charge appliquée, le carbone totalement dégradé lors de la phase préliminaire d'aération ne peut servir de source de carbone pour la dénitrification exogène.

Afin d'accélérer la dénitrification, plusieurs solutions sont possibles : l'une consiste à allonger la deuxième phase d'anoxie suffisamment longtemps pour traiter les ions nitrate résiduels au cours d'un processus de dénitrification endogène, l'autre à diminuer le temps de réaction aérobie tout en augmentant la fréquence des séquences aérobie/anoxie afin de conserver du carbone résiduel lors de la dénitrification. Une troisième solution réside dans l'ajout d'une source de carbone exogène suite à l'étape de nitrification de manière à permettre une assimilation plus rapide et plus efficace des ions nitrate formés (dénitrification exogène).

L'article compare les résultats d'abattement sur le carbone et l'azote d'une eau usée urbaine en utilisant les trois types de fonctionnement. Il en résulte la définition d'une stratégie globale de contrôle du procédé, chacun des scénarii pouvant être privilégié en fonction de la qualité de l'effluent de départ et des contraintes de traitement.

EN :

Wastewater treatment by a Sequencing Batch Reactor (SBR) provides an alternative solution to activated sludge treatment, by carrying out carbon treatment, liquid-solid separation and nutrient removal in a single tank, thanks to the appropriate management of the temporal reaction cycles. Alternating the aeration and anoxic phases, followed by a decantation period, leads, in theory, to the almost total removal of nitrate ions formed during the aerobic nitrification phase. However, depending on the applied load, the carbon that is totally degraded during the preliminary aeration phase, cannot be used as a source of carbon for exogenic denitrification.

Several solutions are possible in order to accelerate denitrification: one consists of lengthening sufficiently the second anoxic phase to treat the residual nitrate ions during the endogenous denitrification process; another strategy involves reducing the aerobic reaction time, while increasing the frequency of aerobic/anoxic sequences in order to preserve residual carbon during denitrification. A third solution lies in the addition of a source of exogenic carbon after the nitrification stage, to allow a quicker and more efficient assimilation of the nitrate ions that are formed (exogenic denitrification). This article compares the results of reducing carbon and nitrogen in wastewater, using three types of operation.

The cycle of reference has been established starting from previous bibliographical results (WUN JERN and DROSTE, 1989) and simulations using the model ASM1 (HENZE et al., 1986). It consists of an anoxic feeding, followed by an anoxic phase, then an aerobic phase and another anoxic phase. The cycle ends by the settling and decanting phases. The lengths of these different phases are: 1 h, 0.5 h, 4.5 h, 3.25 h, 1 h, 1 h. This reference cycle, carried out at the laboratory, leads to the elimination of 90% of the Chemical Oxygen Demand (COD), 95% of the Biological Oxygen Demand (BOD5) and more than 80% of total nitrogen, i.e. with residual concentrations of 60 mg×L^-1 for the Dissolved Organic Carbon (DOC), 5 mg×L^-1 for the Biological Oxygen Demand, and 10 mg×L^-1 for total nitrogen. These results are comparable with those in the literature (IRVINE et al., 1987, MELCER et al., 1987, YANG et al., 1999).

The results obtained during the reference cycle enabled us to reach a rate of total nitrogen removal of 85% and a global nitrogen concentration in the effluent of 11 mg×L^-1. Nitrification and denitrification rates yielded values of 0.8 mg N-NH4+×gMVS^-1 ×h^-1 and 0.8 mg N-NO₃ -×gMVS-1×h-1 respectively. Total nitrogen removal was not completely achieved because of the lack of available carbon. This lack of carbon favours endogenic denitrification, characterized by a slow denitrification rate 0.8 mg N-NO₃ -×mgMVS-1×h-1, compared to exogenic denitrification characterized by a higher nitrogen reduction rate (about 2 mg N-NO₃ -×mgMVS^-1 ×h^-1). This fact was experimentally confirmed with carbon addition in the form of acetate at the beginning of the second anoxic phase. In this case, carbon addition significantly improves the denitrification rate compared to the same experiment without exogenic carbon addition: 2 mg N-NO₃ -×mgMVS^-1 ×h^-1 versus 0.8 mg N-NO₃ -×mgMVS^-1 ×h^-1. However, this method raises operating costs for the process.

In order to reach complete nitrogen removal without the addition of synthetic carbon, it is possible to increase the anoxic phase time scale from 3 hours to 15 hours. Although the results in term of carbon and nitrogen removal are satisfactory when the anoxia phase is lengthened, the concentration in the discharged effluent is 0.33 mg total N×L^-1, and thus this technique decreases the productivity of the SBR. The feeding cycles of a biological reactor being variable, a regulation based on the use of the evolution of the pH, or the redox potential, can be considered (PAVELJ et al., 2001; ANDREOTTOLA et al., 2001). This regulation would make it possible to adapt the duration of the phases of anoxia to the necessary treatment.

To overcome this drawback, a possible approach consists in replacing the aerobic / anoxic phase in the reference cycle by five aerobic / anoxic phases during the same time. Unfortunately, this method leads to a decrease in nitrogen removal and in the nitrification rate, compared to classical cycle (65% versus 85%, and 0.4 mg N-NO₃ - ×mgMVS^-1 ×h^-1 versus 0.8 mg N-NO₃ -×mgMVS^-1 ×h^-1, respectively). The nitrification rate is, in this case, half that obtained in the reference cycle, probably due to delays related to the induction of nitrification and denitrification. This strategy, consisting of increasing the aeration / no aeration frequency, has to be optimized in term of nitrification and denitrification ratios.

A better solution from the economic and productivity points of view is the addition of wastewater at the beginning of anoxic phase. This strategy implies the modification of the cycle. First, after the anoxic feeding, an aerobic phase allows carbon and nitrogen oxidation. In order to supply an available carbon source for exogenic denitrifcation, a second feeding is introduced at the beginning of the second anoxic phase. This addition also contains ammonium ions and implies new nitrification and denitrification steps. This last denitrification phase is then endogenic.

Carbon addition in the form of wastewater leads to an improvement in nitrogen removal. The exogenic denitrification rate is twice the value for endogenous denitrification for the same cycle of operation (1.6 mg N-NO₃ -×gMVS^-1 ×h^-1 versus 0.9 mg N-NO₃ -×gMVS^-1 ×h^-1 respectively). This strategy yields a final concentration of 3 mg N×L^-1 and the nitrification and denitrification rates are similar to those of the traditional processes.

In conclusion, the addition of synthetic carbon in the form of acetate must be preserved as a means of acting quickly in the event of dysfunction (that can be detected by monitoring the redox potential or the pH), although this technique significantly increases the cost of operation. Although the best economic solution to improve denitrification is carbon addition in the form of wastewater, other strategies can be undertaken according to the goals of the treatment process. When the wastewater load is sufficiently weak (night period), the endogenous phase of denitrification can be lengthened. In the event of an important load, carbon addition (in synthetic form or as waste water) makes it possible to eliminate the nitrate ions exogenically.

FR :

Cette étude a pour objectif d'identifier les facteurs qui influencent la contamination des eaux de pluie par les produits phytosanitaires. Cinq sites contrastés ont été choisis de manière à être représentatifs des zones de productions légumières ou de plein champs et à couvrir les différents modes de contamination des précipitations. Il s'agit des sites de l'Ile de Ouessant, Landivisiau, Plouay, Ploufragan et Rennes. Les évènements pluvieux collectés sont choisis en fonction des caractéristiques de formation de la perturbation et du calendrier des épandages de pesticides.

Par ailleurs, les concentrations rencontrées pour le site de Rennes en 2000 (année très humide) ont pu être comparées à celles obtenues lors d'une étude conduite en 1996 sur un site proche mais pour des conditions climatiques plus habituelles (année humide à sèche).

Les analyses sont réalisées par extraction en phase solide suivie d'une analyse en chromatographie en phase gazeuse couplée à la spectrométrie de masse, ou par détection azote-phosphore spécifique (NPD) ou détection par capture d'électrons pour les composés halogénés (ECD). Dans ces deux derniers cas, la confirmation de l'identité des produits est réalisée par un système de double colonnes.

Sept évènements pluvieux distincts ont été collectés et analysés entre les 15 mars et 15 juillet de l'année 2000. Parmis les produits recherchés six molécules sont régulièrement retrouvées : l'atrazine et son métabolite la déethylatrazine (DEA), l'alachlore, le lindane (gamma HCH) et son isomère le béta HCH ainsi que la desméthryne. Nous avions déjà des observations analogues à l'issu de nos premières investigations de 1996.

Les analyses réalisées ont montré l'existence d'une contamination chronique du compartiment atmosphérique par l'atrazine et la DEA pour les zones sous influence agricole. Les niveaux de concentrations rencontrés sont faibles, de l'ordre de 10 ng/L. A ce bruit de fond s'ajoutent en période de traitement, des transferts depuis les parcelles traitées qui conduisent à des valeurs beaucoup plus élevées (de 0,1 à 0,7 µg.L^-1). La détection de l'alachlore et de la desmethryne est limitée aux périodes d'application de ces produits.

Du lindane, et dans un cas son isomère le béta-HCH, ont pu être mis en évidence sur quelques prélèvements, traduisant une contamination chronique du compartiment aérien due à la rémanence de ce type de composés.

Les concentrations dans les pluies, pour un site donné, sont très dépendantes de la pluviométrie. Alors qu'en année de pluviosité normale (1996), les concentrations en période d'application sont élevées, elles restent faibles pour une année humide (2000). Pour s'affranchir de l'effet de dilution, nous avons calculé des retombées massiques (mg.ha^-1) pour les deux périodes de collecte de 1996 et 2000. Les résultats de 2000 restent malgré cela très inférieurs à ceux de 1996 (d'un facteur quatre environ). Les surfaces emblavées et les différences d'usage entre les deux années sont trop faibles pour expliquer les écarts obtenus. Ceci nous permet de conclure que c'est l'intensité des transferts sol-air qui détermine les niveaux de contamination de l'atmosphère. L'humidité élevée des sols, pour une année humide, favorise la migration verticale des produits phytosanitaires dans les couches inférieures, réduisant ainsi les concentrations de pesticides susceptibles d'être transférés vers le compartiment aérien.

EN :

Contamination of rainwater by pesticides was investigated by analyzing samples from 5 locations in Brittany (France). These sites were chosen on the basis of various characteristics including agricultural practices around the site, the direction of prevailing winds (west to east) and the proximity to heavily populated areas. The sites chosen were: Ile de Ouessant (a natural reserve without pesticide application), Landivisiau (a semi-natural zone), Plouay (a rural site with intensive agriculture, mainly cereals and vegetables), Ploufragan (a suburban site) and Rennes (an urban area surrounded by intensive agriculture and orchards). The pesticides to be monitored included the compounds most commonly used in Brittany. New molecules were added to this list to take into account new homologues or uses. Some organochlorines and pesticides with specific application modes were thus included in the list. Analytical techniques involved solid-liquid extraction procedures followed by GC-NPD, GC-ECD or GC-MS after prior derivatisation. Derivatisation was performed with heptafluorobutyric acid to detect urea and sulfonylurea herbicides. Quantification limits were about 5 ng/L for GC-ECD or GC-NPD, and 50 ng/L for molecules detected by GC-MS after derivatisation.

Rain episodes were selected according to the prevailing meteorological conditions and the air mass trajectory. Only precipitation events greater than 6 mm were considered, in order to collect sufficient volume for analyses. Thirty-five samples (7 per location) were collected between 15 March and 15 July 2000, corresponding to various weather fronts.

Six pesticides were detected: atrazine and its metabolite de-ethylatrazine (DEA), alachlor, desmethryne, lindane and its isomer (beta HCH). Among those detected most concentrations were below 100 ng/L but higher concentrations were occasionally detected during and after spraying in agricultural areas. Lindane was detected at low concentration (<10 ng/L) in several samples collected during different periods and locations. These results demonstrate chronic atmospheric contamination due to the presence of this substance even though its use has been banned in the Economic European Community (EEC) (e.g. in 1999 in France).

The most frequently detected compound was atrazine, detected in 60 % of the samples. The presence of atrazine and its metabolite DEA in rainfall indicate chronic contamination of the atmosphere outside of spraying time (mid-March to mid-April). Detection of these compounds was generally limited to agricultural sites, and could also indicate transfer from agricultural areas during soil preparation for maize sowing. These hypotheses will be tested by monitoring atrazine in rainwater during a complete agricultural cycle. Important contamination by atrazine was detected during spraying time at agricultural locations, where concentrations ranged from 10 to 60 ng/L. Detection of alachlor was limited to application periods and rural sites, at levels ranging from 20 to 240 ng/L. However, amounts of alachlor were also detected in Ouessant, even though its use is banned on this isle. This presence was presumably due to the specific conditions of formation of the weather fronts that resulted in rainfall.

Concentrations of pesticides in rainwater obtained in 2000 were generally lower than those described in literature or in a previous study done in Brittany. Alachlor and atrazine concentrations were compared to those of a previous study conducted in Brittany in 1996. The two sites chosen were Rennes in 2000 and Le Rheu in 1996. These two locations, separated by 5 km, were not significantly different since Rennes is directly downwind of Le Rheu. Pluviometry was very important in the year 2000, showing a 30 % increase in rainfall compared to the average from 1961-1990. The concentrations of pesticides in both years were also very different. During a year of normal rainfall (e.g., 1996) concentrations of pesticides in application periods were high, but they remained low during wet years (e.g., 2000). These differences could not be explained by a dilution effect - when fall-out was calculated (mg/ha) to correct for the dilution effect, the differences remained important. Maize surfaces (132 000 ha in 1996 compared to 128 000 ha in 2000) and the quantities of pesticides applied (580 tonnes in 1994 in Brittany compared to 350 tonnes for atrazine, and 263 tonnes in 1998 compared to 119 tonnes in 1994 for alachlor) were too similar to explain the differences observed in 1998. Photodegradation processes, which are more important during normal or dry years, should also favour higher concentrations in 2000 than in 1996. Rain concentrations and fallout observed in 2000 remained lower than those obtained in 1996. Thus, it can be concluded that atmospheric contamination is lower for wet years (e.g., 2000) than for normal or dry years (e.g., 1996). Soil humidity, high for wet years, favours the vertical migration of these substances to lower soil layers, reducing pesticide concentrations available to be transferred to the atmosphere.

FR :

L'analyse de sensibilité constitue une étape importante dans le développement et la mise en place d'un modèle de simulation. Cette analyse de sensibilité du modèle CEQÉROSS effectuée sur un petit bassin versant agricole du Québec (Canada) montre que la dimension du maillage utilisé pour découper le bassin versant de Lennoxville en éléments de calcul exerce une influence directe, mais modérée, sur la charge sédimentaire transportée à l'exutoire. Les paramètres de calage associés à l'érosivité des précipitations et les variables d'entrée décrivant la topographie influencent considérablement les résultats de la simulation pour la période considérée et, de ce fait, nécessitent une plus grande attention lors des étapes de préparation des données et de calage du modèle. La forme linéaire de l'équation de production de sédiments en rivière exerce un rôle majeur sur les charges totales annuelles simulées. L'importance relative des paramètres associés à la granulométrie des sédiments transportés en rivière indique que la capacité du cours d'eau à transporter les particules grossières limite la charge totale annuelle évacuée à l'exutoire du bassin versant de Lennoxville pendant la période d'étude. Ces résultats doivent cependant être confrontés à d'autres analyses de sensibilité afin de décider des modifications éventuelles à apporter aux équations du modèle.

EN :

Sensitivity analysis is a significant step in the development of computer models used to simulate soil erosion and nonpoint source pollution because it allows to evaluate objectively the importance of changes which occur in the response of a model (output) following changes brought in the parameters and variables of the input data. The objective of this study consists of using sensitivity analysis in order to verify the influence of calibration parameters, input variables and the size of the simulation units (i.e. grid squares) on the annual suspended sediment load (net erosion) of a small agricultural watershed (78 ha) located in Lennoxville, province of Quebec (Canada) for the period October 1991 to September 1992. This sensitivity analysis uses the results of an hydro-sedimentological simulation carried out with the CEQÉROSS computer model (DUCHEMIN et al. 2001). The sensitivity analysis takes into account calibration parameters and input variables selected for their ability to influence soil and river erosion and sediment transport (table 2). The analytical procedure consists in using a sensitivity index to quantify the response of the model to extreme (minimum and maximum) values of parameters and variables. The principal results of the sensitivity analysis are presented in figure 3.

The sensitivity analysis of the annual load of suspended sediments indicate that the size of the grid squares used to segment the watershed (i.e. RESOLUTION) had a direct but rather weak influence on net erosion. The results indicate that the erosivity calibration parameters EXPRP and COEFAM, the transport capacity parameters FCTS and PSDS and the topographic input variable PENTD associated to soil erosion, have a considerable influence on the annual load of suspended sediments and, therefore require a greater attention to be brought during the steps of data preparation and model calibration. The results indicate also that calibration parameters FPUSS1, FPUSS2, FSSRIV1 and FSSRIV2 associated to river erosion are the most influential. The linear form of the river sediment production equation contribute to increase the influence of these parameters on the output of the simulation. The fraction of fine sediments (POSS1) transported in the river plays an insignificant role on the annual load of suspended sediments simulated by the CEQÉROSS model. The annual suspended sediment load varies conversely to the increase of coarse sediments (POSS4). The river capacity to transport the coarse grained particles hinders the annual suspended load evacuated from the Lennoxville watershed. The results of the sensitivity analysis must be confronted with additional sensitivity analysis (i.e. larger watershed, longer period of simulation) using the CEQÉROSS model in order to decide if modifications should be undertaken to the model equations.