Abstract
Although permafrost degradation contributes significantly to hydrological change in cold regions,gaps remain in our under-standing of streamflow variation induced by degrading permafrost in different river basins.We therefore used a deep learning model to simulate the long-term(≥ 30 years)monthly streamflow at 60 hydrological stations along the Lena River,the third longest circum-Arctic river.Analyzing the effects of precipitation,temperature,and thaw depth on streamflow variation throughout the Lena River Basin,we identified two feedback patterns relating streamflow to warming permafrost,observed in areas of continuous and discontinuous permafrost.In northern plain regions with continuous permafrost,94％of stations presented an increasing trend in annual streamflow from the 1900s to the 2010s due to permafrost degradation.The enhanced streamflow was mainly due to increased meltwater in the flood season.In southern regions covered by both continuous and discontinuous permafrost,approximately 38％of stations exhibited a declining trend in annual streamflow in response to permafrost degradation,with a high proportion(61％)located in mountain regions(elevation ≥ 500 m).The decline is attributed to the enhanced infiltration capacity of thawing frozen layers within discontinuous permafrost regions.Our study provides new insights into the mechanisms behind permafrost degradation-induced streamflow variation and highlights the importance of formulating tailored strategies for sustainable river management in cold regions experiencing climate change.
万方数据