Diagnosing the causes of river deterioration using stressor-specific metrics

Abstract

More often than not, rivers are impacted by multiple stressors simultaneously affecting water quality, ecological flow, habitat diversity and ultimately lotic biodiversity. Identifying individual stressors as specific causes of deterioration can help inform water managers about stressor hierarchy and appropriate management options. Here, we investigate whether biological metrics from bioassessment schemes hold diagnostic capabilities to distinguish between the impact of individual stressors. We hypothesise that stressor-specific responses occur, when individual stressors show independent ‘modes of action’ (i.e. the specific stress-induced changes of environmental factors that modify the ecological niches of the species constituting the biological community). The stress receptors comprised three aquatic organism groups (macrophytes, benthic invertebrates, fish) represented by 437 biological metrics relevant in aquatic bioassessment. The stressor groups under investigation were physico-chemical, hydromorphological and hydrological stress. The data originated from official monitoring programmes with 769 sampling sites located at three broad river types in Western and Central Germany. Linear and non-linear variance partitioning was performed separately for each river type, with the non-linear analysis using a combination of boosted regression tree modeling and variance partitioning. We considered metrics to be potentially stressor-specific, if the corresponding models were explained predominantly by one stressor group. The linear analyses revealed 16 metrics that met our criteria. Subsequent non-linear modeling resulted in two genuinely stressor-specific metrics, both based on invertebrate data: The Index of Biocoenotic Region (specifically indicating hydromorphological stress) and the Relative abundance of alien invertebrate species (specifically indicating physico-chemical stress). We conclude that stressor-specific metrics can be empirically derived based on available monitoring data, and thus help support decision making in environmental management. However, their applicability is restricted to specific regions (e.g. river basin districts) reflecting the case-specific circumstance to which these metrics are conditioned.