Coastal fish indicators response to natural and anthropogenic drivers–variability at temporal and different spatial scales

Abstract

Ecological indicators are increasingly used in marine and freshwater management but only few are developed towards full operationalization with known patterns of variability and documented responses to natural and anthropogenic environmental drivers. Here, we evaluate potential sources of indicator variability at two different spatial scales in three coastal fish-based indicators of environmental status in the Baltic Sea; abundance of cyprinids, abundance of perch and the proportion of larger perch. The study was performed on a data set covering 41 monitoring areas subject to different levels of anthropogenic impact, at a latitudinal range of 56–66°N and a salinity range of 2–8. Interannual variation was clearly minor relative to spatial variation. Small-scale spatial variation was related to water depth, wave exposure and water temperature. The remaining variation was assessed in relation to differences in natural and anthropogenic drivers between monitoring areas. Cyprinids showed a clear inverse relationship to water transparency, which was used as a proxy for eutrophication, indicating increased abundances in nutrient enriched areas. None of the indicators showed an expected negative relationship to the level of coastal commercial fisheries catches. Rather, a positive relationship for Perch suggested that the coastal fisheries were concentrated to areas with strong perch populations in the studied areas. The effect of salinity and climate (temperature during the growth season) among monitoring areas were small. The results emphasize the importance of assigning area-specific boundary levels to define good environmental status in the coastal fish indicators, in order to account for natural sources of variability. Further, although long-term monitoring in reference areas is crucial for obtaining a historical baseline, our results suggest that the status assessment of coastal fish would generally gain precision by increasingly including spatially based assessments. We propose that similar analytical approaches could be applied to other ecosystem components, especially in naturally heterogenic environments, in order to separate indicator variability attributed to potential anthropogenic impact.