Evapotranspiration Partitioning for Croplands Based on Eddy Covariance Measurements and Machine Learning Models

Agronomy, Journal Year: 2025, Volume and Issue: 15(3), P. 512 - 512

Published: Feb. 20, 2025

Accurately partitioning evapotranspiration (ET) of cropland into productive plant transpiration (T) and non-productive soil evaporation (E) is important for improving crop water use efficiency. Many methods, including machine learning have been developed ET partitioning. However, the applicability models in with diverse rotations not clear. In this study, are used to predict E, T obtained by calculating difference between leading derivation ratio (T/ET). We evaluated six (i.e., artificial neural networks (ANN), extremely randomized trees (ExtraTrees), gradient boosting decision tree (GBDT), light (LightGBM), random forest (RF), extreme (XGBoost)) on at 16 flux sites during period from 2000 2020. The evaluation results showed that XGBoost model had best performance (R = 0.88, RMSE 6.87 W/m2, NSE 0.77, MAE 3.41 W/m2) when considering meteorological data, ecosystem sensible heat flux, respiration, content, remote sensing vegetation indices as input variables. Due unavailability observed E or data sites, we three other widely methods indirectly validate accuracy our based XGBoost. estimation were highly consistent their 0.83–0.91). Moreover, estimated (T/ET) different crops. On average, maize highest T/ET 0.619 ± 0.119, followed soybean (0.618 0.085), winter wheat (0.614 0.08), sugar beet (0.611 0.065). Lower was found paddy rice (0.505 0.055), barley (0.590 0.058), potato (0.540 0.088), rapeseed (0.522 0.107). These suggest easy applicable reveal obvious differences among crops, which crucial sustainability resources improvements

Language: Английский

Effects of different surface water flow frequencies on water use characteristics of Tamarix ramosissima in the hinterland of the Taklimakan Desert

Journal of Hydrology, Journal Year: 2024, Volume and Issue: unknown, P. 132200 - 132200

Published: Oct. 1, 2024

Language: Английский