Mind the Cloud: Propagation of Cloud-Induced Bias in Lake Surface Water Temperature Remote Sensing and Modeling
Abstract
Satellite remote sensing is widely used to monitor lake surface water temperature (LSWT) due to its global coverage and relatively long-term record. However, most satellite-based LSWT products rely on optical sensors that cannot penetrate cloud cover, leading to systematic data gaps. Using synthetic datasets, we show that these gaps are not randomly distributed. Instead, cloud cover tends to coincide with specific thermal conditions, introducing geographically structured biases in annual mean LSWT estimates. We further conducted global-scale synthetic numerical experiments to understand whether cloud-induced bias propagates into model predictions when satellite-derived LSWTs are used for model development. We evaluated two modeling approaches: a physically-based model (air2water) and a data-driven model based on a Long Short-Term Memory (LSTM) neural network. While the LSTM model achieved higher predictive accuracy under complete data, it exhibited significantly amplified biases—particularly in warm-season temperature and ice duration—when trained on cloud-affected datasets. In contrast, air2water showed relatively stable bias patterns, reflecting its resistance to data gaps due to its physical constraints. Our findings underscore the importance of accounting for cloud-induced selection bias in optical satellite remote sensing and environmental modeling.
