No abstract

Numerous models have been developed over the last decades to simulate the fate of pesticides at the watershed scale. Based on a literature review, we inventoried thirty-six models categorized as management, research, screening or multimedia models, each of them having specific strengths and weaknesses. Given this large number of models, it may be difficult for potential users (stakeholders or scientists) to find the most suited one with respect to their needs. To help in this process, this paper proposes a pragmatic approach based on a multi-criteria analysis. Selection criteria are defined following the user's needs and classified in five classes: modelling characteristics, output variables, model applicability, possibilities to simulate best management practices (BMPs) and ease of use. The relative importance of each criterion is quantified by a weight and the total score of a model is calculated by adding the resulting weights of satisfied criteria. This selection framework is illustrated with a case study that consists in selecting a model to develop water quality standards at the watershed scale with respect to the implementation of BMPs. This resulted in the selection of three models: BASINS, SWAT and GIBSI.

Many phosphorus (P) risk indicators have been developed to assess the risk of P loss from agricultural land to surface water. Most of these indicators are designed for land and climates where rainfall-induced erosion of particulate P from sloping landscapes is the main process of P transport. No indicators have been validated in the Canadian Prairies, where P losses are driven by snowmelt-driven runoff over nearly level landscapes and frozen soils. The objective of this project was to evaluate the relationship between water quality data for P from 14 watersheds and three existing P risk indicators for their potential use in the southern Manitoba prairie region of Canada. None of the indicators, including Birr and Mulla's P Index, a preliminary P risk indicator for Manitoba, and a preliminary version of Canada's National Indicator of Risk of Water Contamination by Phosphorus, was significantly correlated with mean concentrations of total P in water or P export per hectare. Although erosion risk was a significant factor influencing the value of these indexes, erosion risk was not correlated with either measure of P loss in these watersheds. Several other watershed characteristics, including average soil test P concentrations, livestock density, proportion of land in annual crops, and the land's inherent capability for agricultural production, were strongly correlated with P concentrations in water (r = 0.80***, r = 0.63**, 0.76***, and -0.70**, respectively). Therefore, these types of P risk indicators require modifications to estimate the risk of P loss under the soil, landscape, and climatic conditions of southern Manitoba.

Abstract. Watershed runoff is closely related to land use but this influence is difficult to quantify. This study focused on the Chaudière River watershed (Québec, Canada) and had two objectives: (i) to quantify the influence of historical agricultural land use evolution on watershed runoff; and (ii) to assess the effect of future land use evolution scenarios under climate change conditions (CC). To achieve this, we used the integrated modeling system GIBSI. Past land use evolution was constructed using satellite images that were integrated into GIBSI. The general trend was an increase of agricultural land in the 80's, a slight decrease in the beginning of the 90's and a steady state over the last ten years. Simulations showed strong correlations between land use evolution and water discharge at the watershed outlet. For the prospective approach, we first assessed the effect of CC and then defined two opposite land use evolution scenarios for the horizon 2025 based on two different trends: agriculture intensification and sustainable development. Simulations led to a wide range of results depending on the climatologic models and gas emission scenarios considered, varying from a decrease to an increase of annual and monthly water discharge. In this context, the two land use scenarios induced opposite effects on water discharge and low flow sequences, especially during the growing season. However, due to the large uncertainty linked to CC simulations, it is difficult to conclude that one land use scenario provides a better adaptation to CC than another. Nevertheless, this study shows that land use is a key factor that has to be taken into account when predicting potential future hydrological responses of a watershed.

Abstract:As part of a wider study on the adaptation of agricultural land use to climate change (CC), this paper presents an assessment of possible future hydrological regimes of the Chaudière River watershed, Québec, Canada. We first present a review of the various methods used to integrate outputs of General Circulation Models (GCMs) into hydrological models that are applied at a local scale. Following this review, the delta method, statistical downscaling, and a combination of both methods were selected for this investigation. Data from different GCMs (in the case of the delta method) corresponding to different gas emission scenarios and simulation members were also considered to provide a range of possible future conditions. We used the integrated modelling system GIBSI, which is based on the distributed hydrological model HYDROTEL, to simulate streamflows for a reference period and a short-term future period . For all three methods, results show a slight decrease in annual runoff (-5% on average). On a monthly scale, the effect is more heterogeneous depending on the method used, showing, in most cases, an increase in water discharge in the winter due to higher temperature and a decrease during the summer and fall. When using statistical downscaling, spring peak flow decreased slightly (-6.7% on average) while summer base flow remained unchanged. This study highlights the importance of using different methods and different sources of data in the assessment of potential CC effects on watershed hydrology. Résumé :Dans le cadre d'une étude sur l'adaptation du territoire agricole au changement climatique (CC), cet article présente une évaluation de l'effet potentiel des CC sur le régime hydrologique du bassin versant de la rivière Chaudière (Québec, Canada). Après avoir inventorié les méthodes existantes pour intégrer les sorties des Modèles de Circulation Générale (MCG) dans les modèles hydrologiques, trois d'entre elles ont été sélectionnées : la méthode des deltas, la mise à l'échelle statistique ainsi qu'une méthode combinant les deux précédentes. Afin d'obtenir une large gamme de conditions futures possibles, Le système de modélisation intégrée GIBSI, basé sur le modèle hydrologique distribué HYDROTEL, a été utilisé pour simuler les débits sur la période de référence et sur la période future . Les résultats montrent une légère diminution de la lame d'eau annuelle avec les trois méthodes (en moyenne -5 %). Sur une base mensuelle, l'effet est plus hétérogène selon la méthode utilisée, avec le plus souvent une augmentation de la lame d'eau en hiver du fait de l'augmentation de la température et une diminution en été et durant l'automne. Selon la méthode de mise à l'échelle statistique, le débit de pointe printanier diminuerait également (-6,7 % en moyenne) tandis que le débit d' étiage estival ne serait pas affecté. Cette étude met en évidence la nécessité d'utiliser plusieurs sources d'information et plusieurs méthodes pour évaluer les effets potentiels des CC sur l'hydrologie des bassins versants.

Abstract. Watershed runoff is closely related to land use, but this influence is difficult to quantify. This study focused on the Chaudière River watershed (Québec, Canada) and had two objectives: (i) to quantify the influence of historical agricultural land use evolution on watershed runoff; and (ii) to assess the effect of future land use evolution scenarios under climate change conditions (CC). To achieve this, we used the integrated modeling system GIBSI. Past land use evolution was constructed using satellite images that were integrated into GIBSI. The general trend was an increase of agricultural land in the 1980s, a slight decrease in the beginning of the 1990s and a steady state over the last ten years. Simulations based on thirty years of daily meteorological series showed strong correlations between land use evolution and water discharge at the watershed outlet, especially for summer and fall seasons. For the prospective approach, we first assessed the effect of CC and then defined two opposite land use evolution scenarios for the horizon 2025 based on two different trends: agriculture intensification or sustainable development. Simulation results showed that CC would induce an increase of water discharge during winter and a decrease the rest of the year, while land use scenarios would have a more drastic effect, agriculture intensification counterbalancing the effect of CC during summer and fall. Due to the large uncertainty linked to CC simulations, it is difficult to conclude that one land use scenario provides a better adaptation to CC than another, but this study shows that land use is a key factor that has to be taken into account when predicting potential future hydrological responses of a watershed.

Abstract. Wastewater sludges are used in agriculture as soil amendment and fertilizer, with regard to their organic matter and nutrient content. However, availability of nitrogen and phosphorus from sludge-amended soils and their transfer in runoff may lead to eutrophication of downstream surface water. The aim of this study is to establish and compare the effect of two different sludges on these transfers: an anaerobically digested and thermically stabilised sludge (Seine-Aval treatment plant, sludge no. 1), and a limed sludge (Saint-Quentin treatment plant, sludge no. 2). Experiments were performed on 12 sloping micro-plots (1 m × 1 m) submitted to sludge spreading and controlled rainfall simulation. Runoff water was sampled and analysed for concentrations in nitrogen species and phosphorus. Results show that spreading of sludge no. 1 increased both ammonium nitrogen (mean of 1.1 mg L −1 N-NH 4 vs. 0.2 mg L −1 N-NH 4 for control micro-plots) and particulate phosphorus concentrations (mean of 2 mg L −1 P vs. 1.1 mg L −1 P for control micro-plots) in runoff water. On the other hand, sludge no. 2 did not induce any significant effect on nutrient concentrations in runoff. These results are related to chemical composition and physical treatment of sludges. This study underlines the existence of a short-term risk of nutrient mobilisation by runoff after sludge spreading on soil, and the need to check precisely the impact of this practice on water quality.

The indicator of risk of water contamination by phosphorus (IROWC_P) is designed to estimate where the risk of water P contamination by agriculture is high, and how this risk is changing over time based on the five-year period of data Census frequency. Firstly developed for the province of Quebec (2000), this paper presents an improved version of IROWC_P (intended to be released in 2008), which will be extended to all watersheds and Soil Landscape of Canada (SLC) polygons (scale 1:1, 000, 000) with more than 5% of agriculture. There are three objectives: (i) create a soil phosphorus saturation database for dominant and subdominant soil series of SLC polygons – the soil P saturation values are estimated by the ratio of soil test P to soil P sorption capacity; (ii) calculate an annual P balance considering crop residue P, manure P, and inorganic fertilizer P – agricultural and manure management practices will also be considered; and (iii) develop a transport-hydrology component including P transport estimation by runoff mechanisms (water balance factor, topographic index) and soil erosion, and the area connectivity to water (artificial drainage, soil macropores, and surface water bodies).