The aims of this study were to determine the potential use of biomarkers at multiple levels of biological organisation together with a fish community bioindicator to assess the estuarine health status of three temporarily open/closed estuaries. The estuaries investigated were the East Kleinemonde (EK), Old Woman's (OW) and Mtana (MTN), all of which are situated in the Eastern Cape Province. Three biomarkers, the acetylcholinesterase (AChE) assay, lipid peroxidation (LPx) assay, liver histopathology and a condition factor were used to determine sub-organism health and one bioindicator, the Estuarine Fish Community Index (EFCI), was used as a bioindicator of community health. The estuarine-dependent marine species Rhabdosargus holubi was selected as an indicator species for the sub-organism level analyses. The results from the community analyses indicated that the EK and OW estuaries were in 'good' condition, while the MTN was found to be in 'moderate' condition. Histological analyses revealed that R. holubi from all three estuaries showed signs of pathological changes to the liver, with the fish from the MTN eliciting the highest occurrence of these changes. The LPx assay found that R. holubi from both the OW and MTN showed signs of oxidative damage in the liver tissue, but those from the EK did not appear to be affected. The AChE assay showed that only the fish from the OW had been affected by anticholinesterase compounds. A laboratory study was undertaken using R. holubi as a positive control for the AChE and LPx assay. The fish were exposed to 3 μg/L chlorpyrifos, a known cholinesterase inhibitor, for six hours and their tissues were examined for changes to LPx levels and AChE activities. AChE activity was significantly inhibited (Mann Whitney U test, Z = 3.65, n = 38, P < 0.001) by the exposure, but LPx levels were not significantly affected. A composite index incorporating the biomarkers at different biological levels of organisation was developed. The index was designed to assist managers and scientists to determine whether the ichthyofauna of a system was being affected by environmental stressors and what management interventions could be undertaken to ameliorate the water quality in an estuary. The index was applied to the three estuaries investigated during the present study and both the OW and MTN were assessed to be in need of immediate management intervention. The fish in the OW were found to be stressed at all the sub-organism levels measured and the reason for this was hypothesised to be as a result of golf course activities in this adjacent estuary. A number of management actions are proposed to reduce the sub-organism stress observed in the fish from the OW. The livers of fish from the MTN were shown to be under stress; however the causative agent of this stress was unknown because there is no formal development in the MTN catchment. However, a possible contaminant source is proposed and management interventions to alleviate the stress on the biota of the MTN are suggested. The EK does not require immediate management intervention, however, continuous routine monitoring is recommended to ensure that conditions do not deteriorate. Shortcomings of the index were outlined and a number of suggestions were made in terms of other measures of biological health which could be incorporated into the index.

Dans un contexte de changements globaux, comprendre les interactions entre les parasites et l’environnement est essentiel afin de prévoir les futurs dynamiques d’infection ainsi que les changements dans le fonctionnement des écosystèmes. Cela dit, les patrons de distribution des parasites ainsi que les moteurs d’infection varient dans le temps et l’espace rendant la compréhension des mécanismes sous-jacents très complexes. Dès lors, les études comparatives se basant sur des données empiriques doivent prendre en considération les facteurs de variations entrant en jeu dans l’estimation des paramètres d’infection chez les populations naturelles. Dans une approche multiéchelles, nous avons exploré les sources de variation dans l’estimation de la prévalence d’infection en nous concentrant sur la maladie du point noir chez les communautés littorales de poissons dulcicoles. Nos résultats ont montré que la prévalence de l'infection est spatialement hétérogène dans le paysage, témoignant de l'existence de points chauds et de points froids de l'infection. Les biais d’échantillonnage lié aux méthodes ont mené à d’importantes variations dans l’estimation de la prévalence et dans les patrons spatiaux d’occurrence de la maladie. Nos résultats ont indiqué également qu’un faible échantillonnage a tendance à sures4mer la prévalence d’infection dans le paysage et que l’effort d’échantillonnage nécessaire pour estimer une prévalence fiable dépend de la méthode d’échantillonnage employée. Les caractéristiques physico-chimiques de l’eau et la structure locale des communautés de poissons se sont révélées les meilleurs prédicteurs d’infection à petite échelle. Nos résultats suggèrent notamment des effets de dilution par barrières d’obstruction et de compatibilité limitant la survie des cercaires. Plusieurs relations entre la prévalence d’infection et les prédicteurs environnementaux ont révélé de la nonlinéarité suggérant des interactions complexes. Notre étude contribue au développement de la compréhension des interactions entre les parasites et leur environnement, ainsi qu’aux biais potentiels dans l’étude des dynamiques d’infection. / In a context of global change, understanding the interactions between parasites and their environment is essential to predict future infection dynamics and changes in ecosystem functioning. That said, parasite distribution patterns and drivers of infection vary in time and space, making understanding the underlying mechanisms highly complex. Comparative studies based on empirical data must therefore take into account the factors of variation involved in estimating infection parameters in natural populations. Using a multi-scale approach, we explored the sources of variation in the estimation of infection prevalence, focusing on black spot disease in littoral freshwater fish communities. Our results showed that infection prevalence is spatially heterogeneous across the landscape with evidence of infection hotspots and coldspots. Method-related sampling biases led to significant variations in prevalence estimates and spatial patterns of disease occurrence. Our results also indicated that low sampling effort tend to overestimate the prevalence of infection in the landscape, and that the sampling effort required to estimate infection prevalence depends on the sampling method employed. Water physico-chemical characteristics and local fish community structure were found to be the best predictors of small-scale infection. Furthermore, our results suggest dilution effects due to obstruction and compatibility barriers limiting cercarial survival. Several relationships between infection prevalence and environmental predictors revealed non-linearity, suggesting complex interactions. Our study contributes to the development of our understanding of the interactions between parasites and their environment, as well as potential biases in the study of infection dynamics.