Predicting the restoration effects by a structural dynamic approach in Lake Chaohu, China

In this study, an ecological model focusing on the phosphorus cycling was developed for the eutrophic Lake Chaohu, China. The parameters were calibration based on monthly observations from May, 1987 to April, 1988. The model results agreed with the observations. Subsequently, the long-term effects of six proposed restoration strategies for Lake Chaohu, such as phosphorus loading reduction and water level management, were predicted by a 25-year simulation. The results showed that Ecological Economy Water Level (EEWL), which makes the water depth lower in the spring than usual, had a significant impact on the phytoplankton (reduced by 90%) and submerged vegetation (recovered by a factor of two). In addition, an idealistic water level management method that decreased the water level by 1 m throughout the year (De1m) provided a better restoration effect than EEWL but with little practicability. Other methods, including phosphorus loading reduction, dredging of sediment and fish releasing, showed less significant effects than EEWL. Therefore, EEWL is recommended as the most practical and effective restoration method. The indicator of exergy generally increased during the restoration processes while De1m and EEWL had the first and second highest exergy values. Afterwards, structural dynamic models (SDMs) were introduced into a 5-year simulation conducted under EEWL or De1m to account for the rapid structure transition from phytoplankton domination to submerged vegetation domination. Two parameters (growth and phosphorus uptake rate of phytoplankton) are included in SDM1 and two more parameters (mortality rate of phytoplankton and submerged vegetation) are included in SDM2. In conclusion: (1) under both scenarios, the parameters in both SDM1 and SDM2 showed a decreasing trend and some fluctuations; (2) the exergy in SDMs was generally higher than that in non-SDM, indicating a better 'solution' for the ecosystem with a greater possibility to survive; (3) SDMs provided stronger changes in the model structure under De1m than EEWL, since the intensity of the forcing function variations were enhanced; (4) in SDM2 with more changing parameters, the first two parameters had similar variation trends as in SDM1 but more fluctuations and more realistic final values within theoretical ranges. The exergy was also higher than that in SDM1.