Edouard Patault

L’érosion hydrique est un processus majeur de dégradation des sols dans le monde avec des conséquences multiples : perte de terres agricoles, envasement des cours d’eau, coulées boueuses. En France, la région Hauts-de-France est la zone la plus touchée par ces pertes en terres qui peuvent dépasser les 10 t ha-1 an-1. Si les processus à l’origine de ces transferts sédimentaires ont été largement étudiés par la communauté scientifique ces dernières années, il reste néanmoins des verrous liés aux variabilités spatio-temporelles. De plus, l’efficacité des politiques récentes de lutte contre l’érosion n’est pas quantifiée. Cette étude propose une analyse de la variabilité spatio-temporelle des transferts sédimentaires selon plusieurs échelles (1-1000 km²), et une première évaluation de l’efficacité des politiques d’aménagement au sein d’un bassin du Nord de la France (la Canche ; 1274 km²). Une station de mesure haute-fréquence, a été implémentée à l’exutoire d’un sous-bassin versant de la Canche (la Pommeroye ; 0,54 km²) pour quantifier les transferts hydro-sédimentaires sur deux années hydrologiques contrastées. Selon nos résultats, les transferts varient de 29,4 à 70 t km-2 an-1. 40% du flux est exporté au cours de 3 épisodes érosifs majeurs (sur 48 enregistrés) et les paramètres forçants sont liés à la durée d’un épisode pluvieux et la quantité de précipitations. Sur ce même bassin, la prédiction spatiale à l’échelle centimétrique des transferts hydro-sédimentaires a pu être effectuée via le nouveau modèle d’érosion des sols WATERSED (BRGM) et l’efficacité d’un plan d’aménagement de lutte contre l’érosion a été quantifiée. Nos résultats permettent de valider l’opérabilité du modèle sur ce territoire. Les transferts sédimentaires à l’échelle du parcellaire agricole peuvent atteindre les 76 t km-2 pour un évènement donné et sont influencés par l’état de surface du parcellaire agricole. Une réduction significative (jusqu’à 84%) des transferts par les aménagements d’hydraulique douce est également observée. A l’échelle du bassin de la Canche, l’utilisation de traceurs chimiques et spectrocolorimétriques dans un modèle de mélange (Sed_Sat ; USGS) a permis d’évaluer d’une part les contributions des affluents de la Canche et d’autre part les contributions des sols et des berges (respectivement 30-70%). Des variations spatio-temporelles significatives ont pu être observées et les résultats tendent à montrer un potentiel impact positif des politiques récentes d’aménagement du territoire. Cette étude montre également que de nouveaux traceurs liés à la signature spécifique des particules magnétiques sont particulièrement prometteurs dans ce contexte pour tracer un signal d’érosion des sols. A terme, ces données pourraient être incluses dans des approches sediment fingerprinting. L’analyse selon plusieurs modalités spatio-temporelles et le couplage expérimentation/modélisation améliore donc notre compréhension de la dynamique des transferts sédimentaires sur le bassin versant de la Canche. Cela fournit des résultats essentiels pour orienter les futures politiques de lutte contre l’érosion des sols.

Erosion is recognized as a major threat worldwide and can be dramatically observed in Northwestern France as a consequence of water runoff. Recent regional studies in Normandy suggested that off-site erosion and runoff impacts led to significant economic costs over the last 25 years. Even if the regional planning strategy against erosion and runoff impacts could be seen as effective with a cost-benefice balance greater than 1, this strategy will no longer be as effective by 2050 due to climate change effects in the near future. To address this issue and conduct efficient land and water degradation neutrality strategies, local stakeholders now have to identify complementary strategies based on the deployments of nature-based solutions. However, there is a lack of references on the effectiveness of these complementary strategies.In this study, we conducted a modelling exercise with the WaterSed model at the regional scale (12,318 km²) aiming to: (i) quantify the hydro-sedimentary transfers reaching the karstic systems throughout the 15,000 sinkholes distributed across the territory, (ii) established the first regional cartography of vulnerability of sinkholes to runoff and erosion, and (iii) to evaluate the effectiveness of strategies considering nature-based solutions to prevent land and ground water degradation.The model was calibrated and validated using data of hydro-sedimentary transfer monitoring station on a local catchment. Multiples scenarios were explored (impacts of different nature-based solutions densities, localization of grasslands, ploughing of grass lands, soil and water conservation techniques, etc.) using semi-automatic positioning algorithm.The WaterSed model provided specific outputs like volume of runoff (m3) and volume of sediments (t) reaching the karstic system for different designed storms. The mean runoff per sinkhole was estimated between 7,700 and 23,200 m3 and the mean volume of sediment reached between 0.8 and 4.7 t per sinkhole.Our results suggest that increasing the density of nature-based solutions from 2 to 8 per km² can reduce from 0.5 to 1.3 % the runoff volume and from 5 to 15 % the sediment load reaching the sinkholes. Our results also suggest that a complement of 20 m to 250 m of grassland upstream a sinkhole can reduce the sediment load from 5 to 13 % and the runoff from 0.5 to 1.5 %. Our results suggest that the localization of ploughed-up grasslands can have a significant impact on the generation of hydro-sedimentary transfers (up to 10 % more sediment discharge).The results of this modeling exercise provided: (i) the first regional cartography of vulnerability of the 15,000 sinkholes to runoff and erosion, and (ii) local thresholds and valuable references to build and conduct efficient land and ground water degradation neutrality strategies with stakeholders.

Excessive sediment discharge in karstic regions can be highly disruptive to water treatment plants. It is essential for catchment stakeholders and drinking water suppliers to limit the impact of high sediment loads on potable water supply, but their strategic choices must be based on simulations integrating surface and groundwater transfers and taking into account possible changes in land use. Karstic environments are particularly challenging as they face a lack of accurate physical descriptions for the modelling process, and they can be particularly complex to predict due to the non-linearity of the processes generating sediment discharge. The aim of the study was to assess the sediment discharge variability at a water treatment plant according to multiple realistic land use scenarios. To reach that goal, we developed a new cascade modelling approach with an erosion-runoff geographic information system (GIS) model (WaterSed) and a deep neural network. The model was used in the Radicatel hydrogeological catchment (106 km² in Normandy, France), where karstic spring water is extracted to a water treatment plant. The sediment discharge was simulated for five design storms under current land use and compared to four land use scenarios (baseline, ploughing up of grassland, eco-engineering, best farming practices, and coupling of eco-engineering/best farming practices). Daily rainfall time series and WaterSed modelling outputs extracted at connected sinkholes (positive dye tracing) were used as input data for the deep neural network model. The model structure was found by a classical trial-and-error procedure, and the model was trained on 2 significant hydrologic years. Evaluation on a test set showed a good performance of the model (NSE = 0.82), and the application of a monthly backward-chaining nested cross-validation revealed that the model is able to generalize on new datasets. Simulations made for the four land use scenarios suggested that ploughing up 33 % of grasslands would increase sediment discharge at the water treatment plant by 5 % on average. By contrast, eco-engineering and best farming practices will significantly reduce sediment discharge at the water treatment plant (respectively in the ranges of 10 %–44 % and 24 %–61 %). The coupling of these two strategies is the most efficient since it affects the hydro-sedimentary production and transfer processes (decreasing sediment discharge from 40 % to 80 %). The cascade modelling approach developed in this study offers interesting opportunities for sediment discharge prediction at karstic springs or water treatment plants under multiple land use scenarios. It also provides robust decision-making tools for land use planning and drinking water suppliers.

RESUME Les départements de Seine-Maritime et de l’Eure sont très sensibles aux ruissellements. Des investissements ont été réalisés ces dernières années pour réduire les risques associés à ces processus. Pourtant l’efficacité de cette politique publique n’a pas été évaluée. À l’échelle d’un bassin pilote (la Lézarde, 212 km²), cette étude originale combine la modélisation en parallèle du ruissellement (MIKE11), des transferts sédimentaires (WaterSed), et une analyse coût-bénéfice selon différents scénarios d’aménagement du territoire et trois pluies de projet. Les simulations suggèrent que les ouvrages hydrauliques structurants sont économiquement efficients (bénéfice/coût~2). En conséquence du réchauffement planétaire il est très probable que les dommages par inondation augmentent. En parallèle, il est très probable que le montant des dommages évités augmente aussi du fait d’une sollicitation plus fréquente des ouvrages existants. L’efficience économique des ouvrages d’hydraulique douce, en complément des ouvrages d’hydrauliques structurants, n’est pas démontrée par nos simulations. La modification du paysage agricole aura une influence sur l’évolution des dommages: en modifiant les pratiques culturales, une réduction significative des dommages est attendue (-250 k€₂₀₀₀ an⁻¹), et en réduisant d’un tiers les surfaces en herbe, une augmentation des dommages sera significative (+400 k€₂₀₀₀ an⁻¹). La méthodologie développée dans cette étude pourrait être appliquée sur d’autres territoires afin d’évaluer les stratégies d’interventions publiques de lutte contre l’érosion et le ruissellement.

Excessive sediment discharge at karstic springs and thus, water treatment plants, can be highly disruptive. It is essential for catchment stakeholders and drinking water supplier to reduce the impact of sediment on potable water supply, but their strategic choices must be based on simulations, integrating surface and groundwater transfers, and taking into account possible changes in land use. Karstic environments are particularly challenging as they face a lack of accurate physical description for the modelling process, and they can be seen as a black-box due to the non-linearity of the processes generating sediment discharge. The aim of the study was to assess the sediment discharge variability at a water treatment plant according to multiple realistic land use scenarios. To reach that goal, we developed a new coupled modelling approach with an erosion-runoff GIS model (WaterSed) and a deep neural network. The model was used in the Radicatel catchment (106 km² in Normandy, France) where karstic spring water is extracted to a water treatment plant. The sediment discharge was simulated for five designed storm projects under current land use and compared to three land use scenarios (baseline, ploughing up of grassland, eco-engineering, best farming practices). Daily rainfall time series and WaterSed modelling outputs extracted at connected sinkholes were used as input data for the deep neural network model. The model structure was found by a classical trial and error procedure, and the model was trained on two significant hydrologic years. Evaluation on a test set showed a good performance of the model (NSE = 0.82), and the application of a monthly-backward chaining nested cross validation revealed that the model is able to generalize on new datasets. Simulations made for the three land use scenarios suggested that ploughing up 33 % of grasslands would not increase significantly sediment discharge at the water treatment plant (5 % in average). In the opposite, eco-engineering and best farming practices will significantly reduce sediment discharge at the water treatment plant (respectively in the range of 10–44 and 24–61 %). The coupling of these two strategies is the most efficient since it affects the hydro-sedimentary production and transfer processes (decreasing sediment discharge from 40 to 80 %). The coupled modelling approach developed in this study offers interesting opportunities for sediment discharge prediction at karstic springs or water treatment plant under multiple land use scenarios. It also provides robust decision-making tools for land use planning and drinking water suppliers.