WRF-Based Forecasted Weather Dataset from the Federal Institute of Espírito Santo, Serra, Brazil

This dataset contains short-term weather forecasts produced using the numerical Weather Research and Forecasting (WRF) model for the municipality of Serra, located in the state of Espírito Santo, Brazil. Different WRF versions were used as updates to the modeling system became available. The sequence of versions adopted was as follows: • 4.5.1, through January 2024; • 4.5.2, February through May 2024; • 4.6.0, June through November 2024; • 4.6.1, December 2024 through May 2025; • 4.7.0, June through July 2025, with LCZ/WUDAPT; • 4.7.1, August 2025 through July 2026, with LCZ; and • 4.8.0, from August 2026 onward, with LCZ. The initial and boundary conditions are obtained from the Global Forecast System (GFS), provided by the National Centers for Environmental Prediction (NCEP). Four nested computational domains were used, with a horizontal nesting ratio of 3:1 between successive domains and horizontal grid spacings of 27 km, 9 km, 3 km, and 1 km, corresponding to domains D01, D02, D03, and D04, respectively. The horizontal grid dimensions are 70 × 70 grid points for D01, 100 × 100 grid points for D02, 100 × 100 grid points for D03, and 100 × 82 grid points for D04. Domain D04 represents the region of interest. The cartographic projection used is Lambert, with a reference point at 20.1978° S and 40.2159° W. In the vertical, the WRF model uses 31 nonuniform eta levels, with higher resolution within the planetary boundary layer (PBL). The model top is set at 50 hPa. The atmospheric equations are solved in nonhydrostatic mode. Cloud microphysics is represented by the WRF Single-Moment 3-class (WSM3) scheme (mp_physics = 3). Longwave radiation is represented using the Rapid Radiative Transfer Model (RRTM) (ra_lw_physics = 1), while shortwave radiation is represented using the Rapid Radiative Transfer Model for General Circulation Models (RRTMG) (ra_sw_physics = 4). Turbulent processes are represented by the Yonsei University (YSU) scheme (bl_pbl_physics = 1) together with the Revised MM5 Monin–Obukhov surface-layer scheme (sf_sfclay_physics = 1). Land–atmosphere interactions are represented by the Noah Land Surface Model (sf_surface_physics = 2 and num_soil_layers = 4). Subgrid-scale convection is represented by the Betts–Miller–Janjic (BMJ) scheme (cu_physics = 2). WRF output files for D04 are generated at a temporal interval of 15 minutes (history_interval = 15). The TIME_SERIES_SOUNDING file contains data at different vertical levels, enabling characterization of the vertical structure of the atmosphere, and includes the following variables: TIME, X, Y, Z, U, V, W, PH, PHB, T, P, PB, QVAPOR, QCLOUD, and QRAIN. The TIME_SERIES_SURFACE file contains information on near-surface atmospheric conditions and includes the following variables: TIME, X, Y, Z, PH, PHB, P, PB, T2, U10, V10, HGT, RAINC, and SWDOWN. The variables included in the files are defined in the Variables_list.xls.本数据集包含针对巴西圣埃斯皮里图州塞拉市（Serra）制作的、基于数值天气预报研究与预报模型（Weather Research and Forecasting, WRF）的短期天气预报数据。 随着建模系统的更新迭代，研究采用了不同版本的WRF模型，所采用的版本序列如下： • 4.5.1：截至2024年1月； • 4.5.2：2024年2月至5月； • 4.6.0：2024年6月至11月； • 4.6.1：2024年12月至2025年5月； • 4.7.0：2025年6月至7月，搭配局地气候区（Local Climate Zone, LCZ）/城市土地利用数据库与工具包（WUDAPT）； • 4.7.1：2025年8月至2026年7月，搭配局地气候区（LCZ）； • 4.8.0：2026年8月起，搭配局地气候区（LCZ）。 初始场与边界条件取自美国国家环境预报中心（National Centers for Environmental Prediction, NCEP）提供的全球预报系统（Global Forecast System, GFS）数据。 研究采用四层嵌套计算域，相邻域间的水平嵌套比为3:1，各域的水平网格间距依次为27 km、9 km、3 km与1 km，对应D01、D02、D03与D04四个嵌套域。各域的水平网格点数分别为：D01为70×70，D02为100×100，D03为100×100，D04为100×82。其中D04为目标研究区域。 所采用的地图投影为兰伯特投影，参考点坐标为南纬20.1978°、西经40.2159°。 垂直方向上，WRF模型采用31层非均匀eta坐标层，在行星边界层（Planetary Boundary Layer, PBL）内具备更高的垂直分辨率，模型顶高设置为50 hPa。 大气动力方程采用非静力模式求解。云微物理过程采用WRF单矩3类（WSM3）参数化方案（mp_physics = 3）。长波辐射采用快速辐射传输模型（Rapid Radiative Transfer Model, RRTM）（ra_lw_physics = 1），短波辐射采用适用于通用环流模型的快速辐射传输模型（Rapid Radiative Transfer Model for General Circulation Models, RRTMG）（ra_sw_physics = 4）。 湍流过程采用延世大学（Yonsei University, YSU）参数化方案（bl_pbl_physics = 1），结合修正的MM5莫宁-奥布霍夫近地面层方案（sf_sfclay_physics = 1）。陆气相互作用采用Noah陆面过程模型（sf_surface_physics = 2，且num_soil_layers = 4）。次网格尺度对流过程采用贝茨-米勒-亚尼奇（Betts–Miller–Janjic, BMJ）参数化方案（cu_physics = 2）。 D04域的WRF输出文件的时间采样间隔为15分钟（history_interval = 15）。 TIME_SERIES_SOUNDING文件包含不同垂直层次的大气数据，可用于表征大气垂直结构，其包含的变量为：TIME、X、Y、Z、U、V、W、PH、PHB、T、P、PB、QVAPOR、QCLOUD与QRAIN。 TIME_SERIES_SURFACE文件包含近地面大气状况相关信息，其包含的变量为：TIME、X、Y、Z、PH、PHB、P、PB、T2、U10、V10、HGT、RAINC与SWDOWN。 所有文件中包含的变量定义详见Variables_list.xls。