Daily near-surface meteorological data for multiple WRF physics schemes over Xinjiang

Data description This dataset contains daily near-surface meteorological variables simulated using the Weather Research and Forecasting (WRF) model under multiple physical parameterization schemes over Xinjiang, northwestern China. The data are provided to support a systematic evaluation of WRF physics configurations and the identification of an optimal scheme for climate simulation in arid and topographically complex regions. Temporal coverage and experiment design The simulations were conducted for seven representative years (1960, 1970, 1980, 1990, 2000, 2010, and 2020) to sample multi-decadal climate variability. For each year, four representative months were selected—January, April, July, and October—representing winter, spring, summer, and autumn conditions. Within each selected month, simulations cover days 2–27 (26 days) to exclude model spin-up effects. Daily values were derived from the model output after removing spin-up periods. Physics schemes and numbering convention The dataset includes results from 34 WRF physics parameterization schemes. Variables are stored with a suffix indicating the scheme index (e.g. pre1, pre2, …, pre39). Although the scheme indices span 1–39, only 34 schemes are present. Several candidate configurations were removed because they were unsupported or incompatible, such as combinations that could not be compiled, failed at runtime, or could not be integrated stably. The retained scheme indices correspond one-to-one with the 34 physics-scheme groups analyzed in the accompanying manuscript, and the ordering is consistent between the dataset and the paper. Variables and units The dataset provides the following daily near-surface variables, stored separately for each physics scheme: Variable Description Unit tav 2 m mean air temperature °C pre Precipitation mm prs Surface pressure kPa win 10 m mean wind speed m s⁻¹ rhu Relative humidity % swdown Downward shortwave radiation MJ m⁻² Data format and structure All variables are stored in a single MATLAB file, var_all.mat. Each variable follows the naming convention variable + scheme index (e.g. pre1, pre10, prs23). Each variable is stored as a 102 × 728 numeric array, where the first dimension corresponds to 102 meteorological stations, and the second dimension corresponds to the daily time series aggregated across all simulated months and years (28 months × 26 days). Station location information is provided in the variable station, a 102 × 2 array containing the latitude and longitude of each station. The ordering of stations in station is consistent with the first dimension of all meteorological variables, enabling direct spatial association between station locations and time series data. Intended use This dataset is designed for physics-scheme evaluation, inter-scheme comparison, and regional climate analysis in arid regions. It enables users to quantify how different WRF physical parameterization combinations affect the simulation of near-surface climate variables and provides a reproducible basis for selecting robust configurations for climate research and data production over Xinjiang and similar dry, complex terrains.