The High-resolution Urban Meteorology for Impacts Dataset for Atlanta Metropolitan Region (HUMID-Atlanta)

As urban areas host a large portion of the world's population, high-resolution gridded meteorological data within cities is required to answer impactful questions across fields including epidemiology and environmental sciences. Urban areas have complex land surface characteristics which significantly impact interactions between the urban surface and atmosphere. We develop a 1 km² grid spacing meteorological dataset for 2010-2023 using a coupled atmosphere-land-urban modeling system (WRF-Urban) over the Atlanta metropolitan area with the Local Climate Zone (LCZ) classification urban map. We then apply a multivariate machine learning bias correction technique to the WRF-urban temperature and moisture variables that learn shared dependencies, which is critical for heat exposure applications. The bias correction is well validated for air temperature but shows reduced spatial transferability for dew point. We also compute several common heat exposure indices such as National Weather Service heat index, Humidex index, wet bulb globe temperature index, universal thermal climate index, and apparent temperature. The dataset can be used to estimate population exposure to heat and quantify relationships between heat exposure and health endpoints in epidemiologic research, as input to quantitative risk assessments and economic evaluation of heat-health implications, and other microclimate applications.