Prediction of the daily spatial variation of stem water potential in cherry orchards using weather and Sentinel-2 data

Irrigation management is critical for orchard production and plays a key role when there is water scarcity. Climate change is changing global precipitation and temperature patterns, causing more frequent and intense drought periods in some parts of the world. Thus, there is a challenge to optimizing water use efficiency. The common practice for irrigation management is to apply the water lost by evapotranspiration. However, we could manage the irrigation by monitoring the plant's water status by measuring the stem water potential (Ψs), which is currently costly and time-consuming. Here, we predicted the daily spatial variation of Ψs. In the central part of Chile, during two seasons (2022-2023 and 2023-2024, between October and April), we measured Ψs in two orchards planted with sweet cherry tree variety Regina, and we monitored 30 trees weekly and biweekly. To predict the Ψs, we used the random forest (RF), extreme gradient boosting (XGBoost), and support vector machine (SVM) models. We selected vapor pressure deficit (VPD), reference evapotranspiration (ET0), relative humidity, and temperature as weather predictors. Spectral vegetation indices (VIs) and biophysical parameters were derived from Sentinel-2. We compared two split schemes for training and testing, one with complete randomness (rnd_split) and the other with daytime independence (tme_split). We found the best performance for the scheme rnd_split, reaching an R2 of 0.77 and a root mean square error (RMSE) of 0.234 MPa; with the tme_split, the performance decreased to an R2 of 0.62 and an RMSE of 0.36 MPa, both with XGBoost. The analysis of importance variables reveals that weather predictors, such as VPD, ET0, and temperature, have a higher weight in the model. These are followed by VIs that use short-wave infrared regions, which highlight the moisture stress index (MSI) and the disease and water stress index (DWSI). This model allows for precise monitoring of cherry water status, assists in optimizing the WUE, and can adapt to various climatic conditions.