A problem exists in Nolin River Lake and Rough River Lake in Kentucky, due to the increasing prevalence of cyanobacterial-based harmful algal blooms (CyanoHABs) and the threats they pose to local communities. These lakes were developed as artificial reservoirs from embankment. Further complicating the issue, the lakes are located within a heavily karstified region and there exists no plan or method currently for monitoring or managing CyanoHABs in a karst region with regard to groundwater inputs to the lake systems or their tributaries. A mixture of techniques and analysis methods was used to determine the best way to monitor and possibly detect the formation and occurrence of CyanoHABs in artificial lakes that are located in karst landscapes. The methods focused on determining the effect groundwater has on CyanoHAB occurrence and formation, how much nutrient pollution is entering the system, from where the pollution is originating and, ultimately, how best to monitor and develop management practices against CyanoHAB occurrence. Techniques used included dual nitrate isotope tracing, collecting hydrogeochemical data, lake discharge data, historical CyanoHAB data, and biological tracer monitoring in both lakes. The lakes under study showed varying degrees of the influence karst plays in their seasonal changes from summer to winter pools. Lake water temperatures never dropped below the temperatures needed for one of the dominant cyanobacteria, Cylindrospermopsis raciborskii, to grow. Calculations of nutrient loadings indicated that over 3.5 x 106 kg of nitrate moved through Nolin River Lake during the course of the study. The presence and concentrations of E. coli when paired with weather and geochemical data also revealed karst groundwater pulses exerting an influence through the system in response to precipitation events. The nitrogen and oxygen isotope data indicate that a wide variety of nitrate pollution sources are entering the system and that a variety of management techniques must be deployed to combat this complex issue. A holistic approach that focuses on management and education about karst processes and CyanoHABs is suggested, with an emphasis on broader community involvement beyond just the populations living adjacent to the lakes.

Ce travail de thèse s’inscrit dans une démarche d’approfondissement des connaissances du fonctionnement des aquifères karstiques. Pour cela, le site pilote des Sources du Toulon, siège d’une alimentation multiréservoirs, a été choisi et offre une fenêtre d’observation privilégiée sur les relations hydrogéologiques entre les aquifères de la marge Nord du Bassin aquitain. Elles sont utilisées depuis 1832 pour l’alimentation en eau potable de la ville de Périgueux, et constituent actuellement son unique ressource.L’objectif de ce travail a été de tester une approche multi-traceurs et un suivi haute-résolution pour caractériser le fonctionnement et la vulnérabilité d’un aquifère karstique multicouche.L’utilisation d’une approche couplée hydrodynamique et hydrochimie a mis en évidence l’alimentation des Sources du Toulon par une ressource profonde et captive en complément d’un aquifère libre plus superficiel. Ces informations ont été intégrées dans le contexte hydrogéologique global de la zone d’étude pour proposer une nouvelle délimitation du bassin d’alimentation des Sources du Toulon.L’utilisation d’un suivi haute-résolution a permis de préciser le fonctionnement hydrogéologique du système karstique étudié. Il en ressort des fonctionnements différents suivants les crues, avec des transferts de masse pouvant s’étendre de quelques jours à quelques semaines.L’apport des isotopes des nitrates (δ15N-NO3- ; δ18O-NO3-) a permis d’identifier deux origines des nitrates mesurés dans les eaux de surface et souterraines : les fertilisants minéraux de synthèse utilisés pour l’agriculture, et les rejets d’assainissement individuels.L’évaluation temporelle de la vulnérabilité spécifique des Sources du Toulon à certains marqueurs de contamination tels que les particules, les éléments bactériologiques et les nitrates a permis de distinguer plusieurs masses d’eau superficielles et subsuperficielles participant à l’alimentation pendant les crues, et jouant un rôle différent dans le transfert des contaminants. / This work is included in an approach for a better knowledge of karst aquifers. For this, the Toulon Springs pilot site was chosen, and provides the opportunity to study relationships between multilayered karst aquifers of the northern edge of the Aquitaine sedimentary basin. Toulon Springs are major regional springs and are located in Périgueux (Dordogne County, France). They have been supplying water to the metropolitan area of Périgueux since 1832.This work aim to test an innovative multitracer approach coupled with a physicochemical high-resolution auto-monitoring to characterize functioning and vulnerability of a multilayered karst aquifer.The use of a coupled hydrodynamic and hydrochemical approach has highlighted Toulon Springs supply by a deep and captive aquifer, in addition to a more subsuperficial aquifer. This information has been incorporated into the global hydrogeological situation of the study area to propose a new delineation of Toulon Springs hydrogeological cathment.The use of a high-resolution monitoring enables to specify the hydrogeological functioning of the studied karst system. The analysis performed on several flood events, identify that mass transfer can range from a few days to a few weeks.The use of nitrate isotope (δ15N-NO3- ; δ18O-NO3-) enables to identify two main nitrate sources in both surface water and groundwater: (i) inorganic fertilizer; (ii) sewage from individual house.The last part of this work was to analyse Toulon Springs temporal vulnerability during floods to marker of contamination such as particles, dissolved organic carbon, nitrate, and bacteria. Hence, several water type from surface runoff, unsaturated zone and saturated zone were identify as responsible of contaminant transfer: (i) water from saturated zone is responsible of vulnerability to mineral particles ; (ii) water from unsaturated zone is responsible of vulnerability to nitrare ; (iii) water from surface runoff is responsible of vulnerability to bacteria contamination.