River ecosystem health assessment in the Qinghai-Tibet Plateau: A novel hybrid method based on artificial intelligence and multi-source data fusion
Zhengxian Zhang,Xiaogang Wang,Yun Li,Yi Liu,Yuan Xu,Jingjuan Li,Wenhao Ding,Hongze Li,Hong Yang
DOI: https://doi.org/10.1016/j.eswa.2024.124078
IF: 8.5
2024-04-29
Expert Systems with Applications
Abstract:River ecosystem health assessment (REHA), an effective approach for identifying river ecosystem health, is crucial for achieving sustainable river management and ensuring water security. However, existing REHA methods still fail to consider the cumulated influences of uncertain inputs, stochastic environment and limited rationality of decision makers on REHA. Additionally, current REHA studies have mainly concentrated on plain areas, while the Qinghai-Tibet Plateau (QTP) remains largely unknown. Developing REHA techniques for plateau rivers is an urgent matter, due to the heightened fragility and complexity of river ecosystems in the QTP. To accurately assess river ecosystem health in the QTP, this study proposed Pythagorean fuzzy cloud (PFC) via coupling the Pythagorean fuzzy sets and cloud model. A novel PFC-TODIM model was developed by extending TODIM (the acronym in Portuguese for interactive and multicriteria decision making) to the Pythagorean fuzzy environment. The hybrid decision making framework was then created to handle REHA with uncertain inputs and stochastic environment, and the Senge Tsangpo River (STR) served as a case study in the QTP. We developed the indicator system based on multi-source data fusion, and employed Bayesian model averaging (BMA) method to reveal the potential risks and driving factors of river ecosystem health. Results showed that the developed models considered the limited rationality of decision makers, effectively handled REHA with uncertainties, and avoided overestimating river health levels due to ignoring the randomness and fuzziness of REHA. In STR, health statuses exhibited marked spatial differences. Sampling sites of 9.091%, 77.273 % and 13.636 % were excellent, healthy and subhealthy, respectively. Our findings highlight that dams, urban development, fish release, and grazing have adverse impacts on STR health, and effective protection measures are required to minimize human interferences on ecologically fragile areas. These findings can improve our understanding of the human disturbances and natural factors that interfere with river ecosystem health in the QTP.
computer science, artificial intelligence,engineering, electrical & electronic,operations research & management science