With the rapid urbanization process, the problem of sudden water pollution in China is becoming increasingly prominent. After pollution occurs, the primary task is to identify the source of pollution for emergency disposal. Nowadays, there are many studies on the sudden water pollution source identification, but they are generally not applicable to the looped river network cases. The reason for this issue is that the problem of identifying pollution source in a looped river network is severely ill-posed, involving complex calculations and being rather difficult to solve. Therefore, this research proposes a creative method for identifying the source of sudden water pollution in a looped river network, which includes: (1) the task of identifying pollution source is divided into two steps: the first step involves identifying the river with a pollution source, and the second step involves identifying detailed information about the pollution; (2) at the first step, the BP neural network algorithm is used to calculate the specific river where a pollution source is located, and the flow and water quality models are further used to calculate the concentration of pollutants at the outlet of the identified river when observations are not available in this river; (3) at the second step, after the river identified, the detailed information of pollution source including the release position, the release time, and the release mass, can be further estimated using the backward location probability density method; (4) the proposed method has been applied to a typical looped river network case, and the results show that this method can effectively identify the source of sudden water pollution from a looped river network. This research achievement can serve for the emergency response, responsibility management, and pre-emergency planning of sudden water pollution in the river network areas, with great practical significance and broad application prospects. |