Compounding effects of human activities and climatic changes on the river water temperature evolution mechanism for ecological effects

Understanding the drivers and ecological effects of river water temperature (Tw) change remains a complex and challenging task in the field of ecohydrology. We developed the extreme‐point symmetric mode decomposition to quantitatively identify the trend, period and mutation of water temperature (Tw). The sparrow search algorithm‐long short‐term memory model was constructed to reconstruct the historical Tw to quantitatively identify the effects of human activities (HA) and climate change (CC) at different time scales. The random forest method was used to analyse the contribution rates of different meteorological factors and to clarify the main factors. Our results demonstrate that the Tw had a cooling trend in the 1980s, and the other decades have a warming trend, and the annual Tw distribution shows an obvious ‘unimodal’ distribution. The interannual Tw has a short period of 2–3a, 6–12a and a long period of 10–15a. At the interannual scale, as the distance from estuaries decreases, the influence of HA gradually diminishes, and only the great Tw is governed by CC. Contributions from HA and CC to Tw vary at the seasonal and monthly scales due to geographical differences. In Cuntan, the contribution of HA to Tw is greater than that of climatic factors. The contribution rates of climatic and HA in the three regions of Yichang, Hankou and Datong also exhibit their respective characteristics and fluctuation patterns. The maximum temperature is the most important factor affecting the change of Tw. Abnormal Tw changes result in delayed spawning and characteristic reproduction Tw attainment dates for the Four Major Chinese Carps and Acipenser sinensis.