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

Published: 26 August 2026| Version 1 | DOI: 10.17632/44pxwb9h46.1
Contributors:
,
,
,
,
,
,

Description

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.

Files

Steps to reproduce

The dataset was generated using the Weather Research and Forecasting (WRF) model for Serra, Espírito Santo, Brazil, as part of a research collaboration between the Federal Institute of Education, Science and Technology of Espírito Santo (IFES) and the Federal University of Espírito Santo (UFES). The simulations were performed by the Laboratory of Alternative Energy Studies (LEAL) of UFES using meteorological data from an automatic weather station. The WRF output was processed at the grid point corresponding to the study location using the read_wrf_nc.f code, publicly available from the WRF project. The processed output was organized into two CSV files: TIME_SERIES_SOUNDING and TIME_SERIES_SURFACE. The LEAL-UFES team provides access to processed WRF output files through its data infrastructure. External access to these files may be limited to approximately two months or less after processing. All files corresponding to data processed through August 20, 2026, are permanently included in this dataset, ensuring access to the complete historical period covered by this publication. Original meteorological input data, personal authentication credentials, and restricted-access resources used during data generation and processing are not included. The dataset provides the processed WRF forecast data directly, allowing researchers to access, analyze, and validate the forecasts without requiring access to the original credentials or processing environment.

Institutions

Categories

Weather Forecasting

Funders

Licence