24AI Deciphering Heat Wave Mechanisms in Bangladesh: Integrating Meteorological Insights and Machine Learning Techniques

In recent years, Bangladesh has experienced increasingly severe heat waves, particularly in Dhaka city, driven by a combination of meteorological factors. Climate change has been a primary contributor, significantly increasing the frequency, duration, and intensity of these heat waves. The rising global temperatures have exacerbated conditions, making Bangladesh particularly vulnerable. For example, during the heat wave of 2023, temperatures exceeded 41°C, marking the highest recorded levels in the last six decades. In 2024, the situation further deteriorated, with average temperatures ranging from 40°C to 42°C across various districts. A major factor in these extreme temperatures is the phenomenon known as a heat dome, where a high-pressure system traps hot air in the upper atmosphere, effectively creating a "lid" that prevents the escape of heat. This results in prolonged periods of extreme heat. Additionally, El Niño events, characterized by warmer surface waters in the equatorial Pacific, have led to hotter and drier conditions in Bangladesh, further amplifying heat wave events. The country's tropical location also heightens its susceptibility to these extreme weather events. South/southwesterly winds, carrying high moisture content, contribute to elevated temperatures due to the higher heat capacity of moist air compared to dry air. During these heat waves, relative humidity values have been observed to range from 40% to 80%, depending on the region. The mechanisms driving these heat waves involve complex interactions between atmospheric pressure systems, oceanic conditions, and local geographical features. The high-pressure systems associated with heat domes create a stable atmosphere, inhibiting cloud formation and leading to clear skies with intense solar radiation. The El Niño phenomenon disrupts normal weather patterns, causing shifts in temperature and precipitation that contribute to extreme heat. Moisture-laden winds from the south/southwest increase humidity levels, exacerbating the perceived heat. In this research, machine learning techniques have been employed alongside Numerical Weather Prediction (NWP) datasets and observational data to better understand the mechanisms behind these heat waves and to predict future events. Results indicate that machine learning models have shown promising accuracy when applied to NWP data. These models hold potential for providing reliable forecasts in future heat wave events in Bangladesh.