Impact of lake thermodynamics on atmospheric river activity

<title>Abstract</title> <p> Lakes, as sensitive indicators of climate change, play a crucial role in regional hydro-thermal cycles. However, the interaction mechanism between the thermal state of inland waters and Atmospheric River (AR) activity remains unclear at a global scale. This study investigates the spatio-temporal covariation between lake surface water temperature (LSWT) and AR frequency using data from 92,245 global lakes and a multi-algorithm AR database from 1981 to 2019. Spearman's rank correlation analysis reveals that LSWT in 6,228 lakes (6.75%) has a significant positive covariation with AR frequency ( <italic>p</italic>  &lt; 0.05), with these lakes being spatially aligned with global high-frequency AR corridors. A random forest attribution model identifies annual LSWT as the most dominant predictor of AR frequency (feature importance: 0.361), surpassing the influence of large-scale air temperature (feature importance: 0.315). A key non-linear mechanism is revealed: a step-like surge in AR frequency occurs as LSWT crosses the ~ 4.0°C thermodynamic threshold, suggesting that intensified evaporation from longer open-water seasons provides "moisture pre-conditioning" and a positive feedback that modulates AR systems. This study confirms that lakes are not only passive responders to climate warming and highlights the need to incorporate their dynamic thermal feedbacks into regional water cycle assessments. </p>