deepET: Reference Evapotranspiration ($ET_o$) Estimation with Deep Learning

deepET is a web application built with Streamlit that estimates daily reference evapotranspiration ($ET_o$). It allows users to upload their own meteorological data and uses a suite of pre-trained deep learning models (ANN, LSTM, CNN) to generate predictions.

Features

• Upload Your Data: Bring your own meteorological data in a simple CSV format.
• Multiple Models: Choose from six different pre-trained deep learning models based on the data you provide.
• Baseline Comparison: Automatically calculates a baseline $ET_o$ using the conventional Hargreaves-Samani (HS) method for comparison.
• Interactive Visualization: View the results in an interactive plot showing your data (if provided), the HS model, and the deep learning model prediction.
• Data Export: Download the complete results, including your original data and all predictions, as a new CSV file.

How to Use the Web App

1. Prepare Your Data: Your data must be in a CSV file and contain the following columns with the exact names:

   • Date: The date for the measurement (e.g., YYYY-MM-DD).
   • Tmin: Daily minimum temperature (°C).
   • Tmax: Daily maximum temperature (°C).
   • Tav: Daily average temperature (°C).

2. (Optional) Add More Features: To use the more advanced models, you can include:

   • RH: Daily average relative humidity (%).
   • U: Daily average wind speed (m/s).
   • Eto: Your own measured or calculated reference $ET_o$ values for direct comparison in the plot.

3. Upload and Configure:

   • In the sidebar, upload your CSV file.
   • Enter the correct latitude and longitude (in decimal degrees) for the location where the data was collected. This is crucial for accurate calculations.

4. Run and Download:

   • Select one of the available deep learning models from the dropdown menu. The available models will be determined automatically based on the columns in your uploaded file.
   • The application will process the data and display the results.
   • Click the "Download Results" button to save the predictions.

Citation

If you use this application or the underlying models in your research, please cite the following publication:

Singh, A., Haghverdi, A., 2023. Development and evaluation of temperature-based deep learning models to estimate reference evapotranspiration. Artificial Intelligence in Agriculture. https://doi.org/10.1016/j.aiia.2023.08.003