Transfer and Deep learning models for daily reference evapotranspiration estimation and forecasting in Spain from local to national scale (Data and Code)

Yu Ye,Aurora González-Vidal,Miguel A. Zamora-Izquierdo,Antonio F. Skarmeta

Department of Information and Communication Engineering, University of Murcia, Murcia, 30100, Spain

The structure of this repository is organized as follows:

• all data murcia: Real meteorological and ET0 data in the Region of Murcia, obtained from http://siam.imida.es/.
• all data siar: Real meteorological and ET0 data in Spain, obtained from https://servicio.mapa.gob.es/websiar/AltaUsuario.aspx?dst=3.
• forecastMurcia: Forecast meterological data in the Region of Murcia, obtained from https://www.weatherbit.io/.
• forecastSiar: Forecast meterological data in Spain, obtained from https://open-meteo.com/.
• createStandarMLmodels.ipynb: The code to create the standard ML models (RF, SVR and MLP).
• createLSTMmodelsForTL.ipynb: The code to create the TL models (LSTM).
• transferLearningCompareTLNoTL.py: The code to get results from models that use TL or not.
• tlScriptCompareTLNoTL.bat: Script to call the transferLearningCompareTLNoTL.py.
• plots.ipynb: Notebook to create the figures.
• getSIARData.ipynb: Notebook to download data from SIAR.
• <stationcode>M*.keras: LSTM models trained from <stationcode>.
• finalModels<stationcode>NO-TL.pkl: Standard models trained from <stationcode>.
• all_medidas<stationcode>.xlsx: Complete local results of the station <stationcode>.
• all_scalers<stationcode>.pkl: Scales for the data from <stationcode>.