A New Dataset on Meteorology and Observational Environments in a Typical High-Density City in South China

Dataset Description This dataset comprises a high-resolution urban meteorological observational dataset compiled from 48 automatic weather stations (AWS) across Shenzhen, a typical high-density megacity in subtropical China, for the year 2024. It was developed to address the critical gap in multi-element, high-density urban meteorological data for East Asian cities. Key Features: Spatial Coverage: 48 stations covering diverse urban land use types, including central business districts, high-density residential areas, industrial zones, coastal zones, and ecological reserves. Temporal Resolution: Hourly data from 00:00, January 1, 2024, to 23:00, December 31, 2024 (Beijing Time, UTC+8). Meteorological Variables: Air temperature (℃), station pressure (hPa), relative humidity (%), wind speed (m/s), wind direction (°), and hourly accumulated precipitation (mm). Quality Control: Raw data have undergone rigorous quality control procedures, including range checks, staleness tests, outlier detection (rolling 30-day window with ±4σ/±5σ thresholds), and visual inspection, referencing international urban flux data protocols. Gap-Filling: Missing data were filled using a 60-day rolling window bias-correction method against ERA5 reanalysis data, which was calibrated using local observations (the "UP" method). The gap-filled data show significant improvements (MAE reduced by 38%–73%) and align closely with observed diurnal and seasonal cycles. Ancillary Data: The dataset is accompanied by detailed site metadata, including geographic location, instrumentation height, and key urban morphological parameters (e.g., impervious surface fraction, tree cover, building height, and roughness length) derived from high-resolution remote sensing products (ESA WorldCover, 3D-GloBFP) within a 500m radius buffer. This dataset is the first of its kind for the Shenzhen region, offering a robust, continuous, and standardized dataset suitable for urban land surface model validation, local climate zone analysis, extreme event studies, and urban planning applications. It also serves as a critical benchmark for quantifying systematic biases in reanalysis products (e.g., ERA5) over urban canopies.