Dataset of Performance of Daily Reference Evapotranspiration Estimation by Different Methods Across Brazilian Climates
Description
This database contains statistical performance metrics, per weather station, of 12 empirical methods for estimating daily reference evapotranspiration (ETo), evaluated against the standard FAO-56 Penman-Monteith method (Allen et al., 1998) at 584 automatic weather stations (AWS) of the National Institute of Meteorology (INMET), distributed throughout the Brazilian territory and classified by climate type according to the Köppen-Geiger classification performed by Alvares et al. (2013). Each row corresponds to a weather station and contains identification and location information for the station, the number of valid daily observations used in the analyses, the linear regression equations between each alternative method and the PM standard, and five statistical performance indicators calculated for each of the 12 evaluated methods: coefficient of determination (R², dimensionless), mean absolute error (MAE, mm day⁻¹), mean bias error (MBE, mm day⁻¹), percent bias (PBIAS, %), Nash-Sutcliffe efficiency index (NASH, dimensionless), and Kling-Gupta efficiency index (KGE, dimensionless). The 12 evaluated methods are: Hargreaves-Samani (HS), Hamon (Ham), Camargo (Cam), Benevides-Lopez (BL), Jobson (Job), Makkink (Mak), Turc (Tur), Jensen-Haise (JH), Tanner-Pelton (TP), Priestley-Taylor (PT), FAO-24 Solar Radiation (RS), and FAO-24 Penman (Pen). This database supports the results of the article " Performance of Daily Reference Evapotranspiration Estimation by Different Methods Across Brazilian Climates" (Moro et al., 2025, Brazilian Journal of Meteorology, v. 40, e40250027. DOI: 10.1590/0102-778640250027), whose main conclusions indicate that the Turc and FAO-24 Penman methods showed the best overall performance in most Brazilian climates, and that the Hargreaves-Samani method is the recommended alternative when only air temperature data are available. File: Brazilian_Performance_ETo.xls Format: Microsoft Excel (.xls), data spreadsheet and information spreadsheet Rows: 585 (Header plus one row per weather station) Columns: 97 Detailed description of the columns: Station identification and location: Code: INMET station code Name: weather station name Latitude: geographical latitude Longitude: geographical longitude Altitude: altitude (m) Situation: station status (active/inactive) Start_Date: station operation starts date End_Date: data period end date Koppen_Climate: Köppen-Geiger climate classification of the station State: Brazilian state where the station is located Region: geographical region of Brazil Bioma: Brazilian biome where the station is located n: number of valid daily observations used in the analyses Method performance indicators: Reg_Eq_[Method]: linear regression equation (string, format "y = a + b·x") R_squared_[Method]: coefficient of determination MAE_[Method]: mean absolute error MBE_[Method]: mean bias error PBIAS_[Method]: percent bias NASH_[Method]: Nash-Sutcliffe efficiency index KGE_[Method]: Kling-Gupta efficiency index.
Files
Steps to reproduce
Data acquisition and processing methods The data were obtained from the daily meteorological series of 584 automatic weather stations (EMA) of the Instituto Nacional de Meteorologia (INMET), distributed throughout the Brazilian territory. The period for each station was determined by data availability, with the start date being the operation start date of each EMA and the end date being December 31, 2023. The meteorological variables used were derived from the processing of hourly records — daily mean, maximum, minimum, or integral — and comprise: maximum (T_max, °C) and minimum (T_min, °C) air temperature; maximum (UR_max, %) and minimum (UR_min, %) relative humidity; global solar radiation (Rs, MJ m⁻² day⁻¹); and wind speed measured at ten meters (u₁₀), converted to two meters (u₂, m s⁻¹) according to Allen et al. (1998). For the analyses, only days with the simultaneous availability of all variables were used. Quality control The daily data were subjected to a quality control to ensure the integrity of the series and eliminate spurious data, according to the criteria described by Xavier et al. (2016; 2022): air temperature between −30 °C and 50 °C; relative humidity between 0% and 100%; solar radiation between 0.03·Ra and Ra (where Ra is the extraterrestrial radiation); and wind speed between 0 and 100 m s⁻¹. Calculation of ETo by the standard method and the alternative methods The daily reference evapotranspiration by the standard FAO-56 Penman-Monteith method (ETo_PM) was calculated according to Allen et al. (1998). Next, the ETo estimates were calculated by the 12 evaluated empirical alternative methods: Hargreaves-Samani (1985), Hamon (1961), Camargo (1971), Benevides-Lopez (1970), Jobson (Bowie et al., 1985), Makkink (1957), Turc (1961), Jensen-Haise (1963), Tanner-Pelton (1960), Priestley-Taylor (1972), FAO-24 Solar Radiation, and FAO-24 Penman (Doorenbos and Pruitt, 1977). Performance evaluation The performance of each method was evaluated per weather station using five statistical metrics calculated between the estimates of each alternative method and the ETo_PM values: coefficient of determination (R²), mean absolute error (MAE, mm day⁻¹), mean bias error (MBE, mm day⁻¹), percent bias (PBIAS, mm day⁻¹), Nash-Sutcliffe efficiency (NSE), and Kling-Gupta efficiency index (KGE; Gupta et al., 2009). For each station, simple linear regression equations were also fitted between the estimates of each method and the PM standard (ETo_method = a + b · ETo_PM). Computational tools All data processing — including quality control, calculation of ETo estimates by the 12 methods, computation of the performance metrics, and generation of the database — was performed in the Python programming language, with the aid of the NumPy, pandas, and Statsmodels libraries for data manipulation and statistical calculations.
Institutions
- Universidade Federal do Espírito SantoEspírito Santo, Vitória
Categories
Funders
- Coordenação de Aperfeicoamento de Pessoal de Nível SuperiorMinistry of EducationFederal District, BrasíliaGrant ID: 001