Estimation of the yield and green water footprint of rainfed wheat based on remote sensing and machine learning

Abstract In this research, the relationship between remote sensing drought indices NDVI, EVI, SAVI, and LAI with the yield and green water footprint (WF) of rainfed wheat in 5 fields in Saqqez City (2001–2020) was investigated using multivariate regression (MR), random forest (RF), and support vecto...

Full description

Saved in:
Bibliographic Details
Main Authors: Mojgan Ahmadi, Hadi Ramezani Etedali, Abbass Kaviani, Alireza Tavakoli
Format: Article
Language:English
Published: SpringerOpen 2025-07-01
Series:Applied Water Science
Subjects:
Online Access:https://doi.org/10.1007/s13201-025-02542-x
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849334596999577600
author Mojgan Ahmadi
Hadi Ramezani Etedali
Abbass Kaviani
Alireza Tavakoli
author_facet Mojgan Ahmadi
Hadi Ramezani Etedali
Abbass Kaviani
Alireza Tavakoli
author_sort Mojgan Ahmadi
collection DOAJ
description Abstract In this research, the relationship between remote sensing drought indices NDVI, EVI, SAVI, and LAI with the yield and green water footprint (WF) of rainfed wheat in 5 fields in Saqqez City (2001–2020) was investigated using multivariate regression (MR), random forest (RF), and support vector regression (SVR) methods. Wheat yield of the fields)2001–2020) was simulated with the AquaCrop model. The results showed a high coefficient of determination (R 2) (R 2 = 0.97) between the yield simulated by the AquaCrop model and the observed yield of the fields. The high Nash–Sutcliffe efficiency (NSE) (0.86) and a small amount of underestimation in the calibration step showed the model has a suitable estimation. Results showed that in the simulation of the yield of rainfed wheat, the RF method had a high correlation, NSE was close to one, and root mean square error (RMSE) was less than 0.2 (ton/ha) and had good accuracy. The relationship between the remote sensing drought indices and the green WF of rainfed wheat, as shown by the results, is that the R2 varies between 0.87 and 0.73. The RMSE was between 0.13 and 0.1 (m3/ton) in different testing steps and the NSE was close to one. The relationship between WF climate variables and yield was examined. Results showed evapotranspiration (ET) and maximum temperature (Tmax) directly affected the green WF of rainfed wheat. The results showed that the RF method had a good estimate of the green WF of rainfed wheat. There is a significant relationship between the remote sensing drought indices and the green WF of rainfed wheat in the study area.
format Article
id doaj-art-2fbb3ff9a2aa49b6b57f1fd08be9a29d
institution Kabale University
issn 2190-5487
2190-5495
language English
publishDate 2025-07-01
publisher SpringerOpen
record_format Article
series Applied Water Science
spelling doaj-art-2fbb3ff9a2aa49b6b57f1fd08be9a29d2025-08-20T03:45:31ZengSpringerOpenApplied Water Science2190-54872190-54952025-07-0115811710.1007/s13201-025-02542-xEstimation of the yield and green water footprint of rainfed wheat based on remote sensing and machine learningMojgan Ahmadi0Hadi Ramezani Etedali1Abbass Kaviani2Alireza Tavakoli3Water Sciences and Engineering Department, Imam Khomeini International UniversityWater Sciences and Engineering Department, Imam Khomeini International UniversityWater Sciences and Engineering Department, Imam Khomeini International UniversityAgricultural Engineering Research Institute, Agricultural Research, Education and Extension OrganizationAbstract In this research, the relationship between remote sensing drought indices NDVI, EVI, SAVI, and LAI with the yield and green water footprint (WF) of rainfed wheat in 5 fields in Saqqez City (2001–2020) was investigated using multivariate regression (MR), random forest (RF), and support vector regression (SVR) methods. Wheat yield of the fields)2001–2020) was simulated with the AquaCrop model. The results showed a high coefficient of determination (R 2) (R 2 = 0.97) between the yield simulated by the AquaCrop model and the observed yield of the fields. The high Nash–Sutcliffe efficiency (NSE) (0.86) and a small amount of underestimation in the calibration step showed the model has a suitable estimation. Results showed that in the simulation of the yield of rainfed wheat, the RF method had a high correlation, NSE was close to one, and root mean square error (RMSE) was less than 0.2 (ton/ha) and had good accuracy. The relationship between the remote sensing drought indices and the green WF of rainfed wheat, as shown by the results, is that the R2 varies between 0.87 and 0.73. The RMSE was between 0.13 and 0.1 (m3/ton) in different testing steps and the NSE was close to one. The relationship between WF climate variables and yield was examined. Results showed evapotranspiration (ET) and maximum temperature (Tmax) directly affected the green WF of rainfed wheat. The results showed that the RF method had a good estimate of the green WF of rainfed wheat. There is a significant relationship between the remote sensing drought indices and the green WF of rainfed wheat in the study area.https://doi.org/10.1007/s13201-025-02542-xSupport vector regressionRandom forestMultivariate regressionGreen water footprint of rainfed wheat
spellingShingle Mojgan Ahmadi
Hadi Ramezani Etedali
Abbass Kaviani
Alireza Tavakoli
Estimation of the yield and green water footprint of rainfed wheat based on remote sensing and machine learning
Applied Water Science
Support vector regression
Random forest
Multivariate regression
Green water footprint of rainfed wheat
title Estimation of the yield and green water footprint of rainfed wheat based on remote sensing and machine learning
title_full Estimation of the yield and green water footprint of rainfed wheat based on remote sensing and machine learning
title_fullStr Estimation of the yield and green water footprint of rainfed wheat based on remote sensing and machine learning
title_full_unstemmed Estimation of the yield and green water footprint of rainfed wheat based on remote sensing and machine learning
title_short Estimation of the yield and green water footprint of rainfed wheat based on remote sensing and machine learning
title_sort estimation of the yield and green water footprint of rainfed wheat based on remote sensing and machine learning
topic Support vector regression
Random forest
Multivariate regression
Green water footprint of rainfed wheat
url https://doi.org/10.1007/s13201-025-02542-x
work_keys_str_mv AT mojganahmadi estimationoftheyieldandgreenwaterfootprintofrainfedwheatbasedonremotesensingandmachinelearning
AT hadiramezanietedali estimationoftheyieldandgreenwaterfootprintofrainfedwheatbasedonremotesensingandmachinelearning
AT abbasskaviani estimationoftheyieldandgreenwaterfootprintofrainfedwheatbasedonremotesensingandmachinelearning
AT alirezatavakoli estimationoftheyieldandgreenwaterfootprintofrainfedwheatbasedonremotesensingandmachinelearning