Ecological impact of polycyclic aromatic hydrocarbons on Baiyangdian Lake based on an ecosystem model

Yong Zeng,Wei Yang,Yanwei Zhao
DOI: https://doi.org/10.1016/j.ecolmodel.2022.110103
IF: 3.1
2022-01-01
Ecological Modelling
Abstract:Conventional approaches to assessing the ecological impact of chemicals face uncertainty when extrapolating from single-species toxicity data to community-level effects on real ecosystems. Ecological models may be a useful tool to reduce that uncertainty because they enable a more mechanistic understanding of the effects of chemicals on species assemblages at the ecosystem level. Here, a case study was developed in AQUATOX 3.1 to simulate the effects of the polycyclic aromatic hydrocarbons (PAHs) on Baiyangdian (BYD) Lake in northern China. The model consisted of 12 groups of species: two phytoplankton, one macrophyte, three zooplankton, four benthic macroinvertebrates, one omnivorous fish, and one piscivorous fish. A simulation of a perturbed (or baseline) scenario over 1 year showed satisfactory calibration on annual average biomasses. Compared with a control scenario, the overall average biomass relative variation (V) decreases by 4.02%, the relative variation of eco-exergy (V-Ex) decreases by 5.64%, and the ratio of gross primary productivity to community respiration (P/R) increases by 5.58%. These results indicate that one-third of the species are remarkably affected, of which the biomass differences are larger than 20%. However, indirect effects due to ecological interactions simulated by the BYD model can either amplify or mitigate the effect of direct toxicity, leading to significant changes for some organisms' responses to chemicals in real ecosystems. The case study demonstrates that the BYD model has the potential to be an effective and useful tool to evaluate the toxicity effects of organic pollutants at an ecosystem level. Some suggestions are put forward to improve the model's accuracy, including (1) more monitoring and testing data for biological and toxicological parameters, (2) a step-by-step strategy for model calibration, and (3) sensitivity analysis before calibrating the parameters.
What problem does this paper attempt to address?