Various biodiversity indicators have been proposed to capture habitat and taxonomic differences among coexisting species, as well as species richness and evenness. Biodiversity indicators are an essential tool for detecting and quantifying changes in ecological communities caused by environmental stress. However, biodiversity indicators may differ in their responses to different environmental stresses, making it critical to identify which indicators are most appropriate for specific monitoring purposes. In this study, we characterise 10 biodiversity indicators (species richness, total richness, taxonomic diversity, community tropicality, Simpson's diversity, Shannon's diversity, species dominance, Simpson's evenness, Shannon's evenness and Smith-Wilson evenness) based on their responsiveness to environmental change. The environmental responsiveness of biodiversity indicators is examined by 1) quantifying the dynamic similarity between different time windows within a single indicator time series using non-linear time series forecasting; 2) characterising and classifying biodiversity indicators based on the pattern of their dynamical similarity. We applied this analysis to long-term monitoring data of a fish community in Uchiura Bay, Fukui Prefecture, affected by temperature fluctuations due to thermal discharges from a nearby nuclear power plant (NPP). Our analysis identified three distinct groups of 10 biodiversity indicators. Group 1, including species richness and community tropicality, showed similar dynamics across all time windows, indicating high resilience to environmental change. Group 2, including species diversity indicators (e.g. Simpson's diversity and Shannon's diversity), responded to temperature changes, indicating their high sensitivity to temperature changes. Group 3, including species evenness indicators (e.g. Simpson's evenness and Shannon's evenness), showed frequently changing dynamics, presumably reflecting their sensitivity to various environmental changes. The resilient responses in species richness and community tropicality may reflect factors unrelated to temperature changes in the bay, such as the immigration of southern species via ocean currents. In contrast, the sensitivity in the dynamics of species richness and evenness indicators may be due to changes in the distribution of species abundance altered by the operation of the NPP and the associated decline in dominant fish species. Thus, the methodology developed in this study contributes to the understanding of how different biodiversity indicators respond to environmental change, thus facilitating more accurate and meaningful interpretations of biodiversity dynamics in natural ecosystems.