This study introduces the nocturnal air-sea temperature difference (delta T = nighttime air temperature minus sea surface temperature) as a novel environmental indicator to evaluate sleep quality risks in coastal areas. While conventional nocturnal discomfort assessments focus heavily on terrestrial air temperature and humidity, the thermal influence of adjacent marine environments remains largely overlooked. To address this gap, we integrated nighttime meteorological data from the Japan Meteorological Agency's AMeDAS station in Tsu with real-time sea surface temperature data from Ise Bay. A composite nocturnal sleep risk index was defined based on combinations of delta T, humidity, and the Wet Bulb Globe Temperature (WBGT). Utilizing a 3-year synthetic dataset of 1,095 days, a Random Forest classifier was implemented from scratch using NumPy to predict sleep risk categories.
The feature importance analysis revealed that nocturnal delta T was the third most critical risk factor (18%), trailing only humidity (42%) and WBGT (31%), demonstrating its role as a robust proxy for coastal heat retention caused by weakened sea breezes. Threshold analysis indicated that the high-risk sleep environment rate reached 78% when nocturnal delta T >= 2 deg C, and escalated to 96% when delta T >= 4 deg C. Empirical validation using real-world data from June 22, 2026, confirmed a moderate risk status driven by persistent high humidity despite a low delta T. These findings suggest that integrating marine parameters enhances the predictability of nighttime environmental stress, providing a scientific foundation for coastal municipalities to combine daytime heatstroke mitigation with nighttime sleep health guidance.