Using the ERA5 dataset of atmospheric variables to estimate daily reference evapotranspiration in Sicily, Italy.

Crop evapotranspiration (ET) plays a key role in many hydrological processes involving the soil-plant-atmosphere system. The concept of reference crop evapotranspiration (ET₀) was introduced to estimate the atmosphere evaporation demand independently of crop type, development stage and management practices. Among the available methods to estimate ET₀, the Penman-Monteith equation proposed by the Food and Agriculture Organization of the United Nations (FAO56-PM), is considered one of the most accurate, so that it is assumed as a reference to calibrate other simplified procedures. In several regions of the world, the limited availability of meteorological observations to estimate ET₀ can be overcome by using gridded reanalysis dataset created by data assimilation of weather observations. Different datasets with relatively high spatial resolution but different in terms of Spatio-temporal resolution have been generated and are freely downloadable at the global scale. The latest ERA5-Land product released in 2019 is characterized by a spatial grid to 0.1° latitude and 0.1° longitude. The database provides several land variables at hourly time-step including, among others, air temperature, dew point temperature and solar radiation at 2.0 m above the soil surface, as well as the wind speed components at 10 m height.

The objective of the research was to assess the suitability of ERA5-Land dataset of climate data to predict daily reference evapotranspiration in Sicily, Italy. For the period 2006-2015, the performance of the reanalysis data to capture the local climate variables was assessed based on the comparison with the corresponding ground data measured by a network of 39 climate stations in Sicily belonging to the Agrometeorological Information Service (SIAS). After evaluating the statistical errors associated with each climatic variables retrieved from the ERA5-Land, the comparison between daily ET₀ values obtained with the FAO56-PM and considering both the dataset was carried out.

The analysis showed that air temperature, solar radiation and wind speed retrieved by the ERA-5 dataset resulted in quite good agreement with the corresponding measured on the ground, with an average root mean square error (RMSE) equal respectively to 1.8°C, 2.9 MJm^-2d^-1, and 1.3 ms^-1 and corresponding mean bias errors (MBE) of -0.4°C, 1.0 MJm^-2d^-1  and -0.1 ms^-1. On the other hand, relative air humidity was characterized by average values of RMSE and MBE respectively equal to 10.3% and 5.6%. When considering all the examined climate stations, the RMSE and MBE values associated with ET₀ ranged from 0.4 to 1.3 mm d^-1, and -1.0 and 0.0 mm d^-1, supporting the possibility to consider the ERA-5 data to obtain suitable estimations of crop reference evapotranspiration even for other Mediterranean countries where measured climate data are not available.