The dynamic response of a single population to chemicals can be represented by a Weibull function. However, it is unclear whether the overall response can still be represented in this manner when scaled up to the community level. In this study, we investigated the responses of biological communities to polycyclic aromatic hydrocarbons by using an ecological model of Baiyangdian Lake in northern China. The community dynamics process was divided into the following three stages. In the first stage, toxicity, played a dominant role and strong, medium, and weak species responses were observed according to the toxicity sensitivity. In the second stage, the dynamic process was dominated by the interaction strength with three alternative dynamic pathways comprising of direct response, no response, or inverse response. In the third stage, the toxicity was again dominant, and the biomasses of all species decreased to extinction. The toxicological dynamics were far more complex at the community level than those at the single species level and they were also influenced by the interaction strength as well as toxicity. The toxicological dynamic process in the community was constantly driven by the competing effects of these two forces. In addition to the total biomass, the interaction strength was identified as a suitable community-level signal because it exhibited good indicator properties regarding ecosystem steady-state transitions. However, we found that food web stability indicators were not suitable for use as community-level signals because they were not sensitive to changes in the ecosystem state. Some ecological management suggestions have been proposed, including medium to long-term monitoring, and reduction of external pollution loads and bioindicators. The results obtained in this study increase our understanding of how chemicals interfere with community dynamics, and the interaction strength and total biomass were identified as useful holistic indicators.