Air pollution is a major public health concern in Mongolia, particularly in Ulaanbaatar and other urban centers. During winter months, PM2.5 concentrations frequently rank among the highest globally. In the ger districts, informal settlements characterized by coal and wood-burning stoves, inadequate infrastructure, and limited access to clean heating, residents face disproportionately high exposures due to high household emissions and energy poverty. Chronic exposure to elevated PM2.5 has been associated with increased rates of cardiovascular and respiratory diseases, reduced life expectancy, and heightened susceptibility to seasonal illness.

This paper adapts the RF–CNN–LCN pipeline developed by Zheng et al. (2021) for Delhi and Beijing to capture fine-scale PM2.5 exposure patterns. It integrates satellite imagery from the 2024–2025 winter season with daily surface-based PM2.5 and meteorological variables (temperature, relative humidity, and wind speed/direction) to estimate high spatial resolution (~200m) ground-level PM2.5 concentrations. The goal is to identify local PM2.5 hotspots in Ulaanbaatar during winter.