Forecasting Long-Series Daily Reference Evapotranspiration Based on Best Subset Regression and Machine Learning in Egypt DOI Open Access
Ahmed Elbeltagi, Aman Srivastava, Abdullah Hassan Al-Saeedi

и другие.

Water, Год журнала: 2023, Номер 15(6), С. 1149 - 1149

Опубликована: Март 15, 2023

The estimation of reference evapotranspiration (ETo), a crucial step in the hydrologic cycle, is essential for system design and management, including balancing, planning, scheduling agricultural water supply resources. When climates vary from arid to semi-arid, there are problems with lack meteorological data future information on ETo, as case Egypt, it more important estimate ETo precisely. To address this, current study aimed model Egypt’s most governorates (Al Buhayrah, Alexandria, Ismailiyah, Minufiyah) using four machine learning (ML) algorithms: linear regression (LR), random subspace (RSS), additive (AR), reduced error pruning tree (REPTree). Climate Forecast System Reanalysis (CFSR) National Centers Environmental Prediction (NCEP) was used gather daily climate variables 1979 2014. datasets were split into two sections: training phase, i.e., 1979–2006, testing 2007–2014. Maximum temperature (Tmax), minimum (Tmin), solar radiation (SR) found be three input that had influence outcome subset sensitivity analysis. A comparative analysis ML models revealed REPTree outperformed competitors by achieving best values various performance matrices during phases. study’s novelty lies use predict this algorithm has not been commonly purpose. Given sparse attempts such research, remarkable accuracy predicting highlighted rarity study. In order combat effects aridity through better resource also cautions authorities concentrate their policymaking adaptation.

Язык: Английский

Federated learning based reference evapotranspiration estimation for distributed crop fields DOI Creative Commons
Muhammad Tausif, Muhammad Waseem Iqbal, Rab Nawaz Bashir

и другие.

PLoS ONE, Год журнала: 2025, Номер 20(2), С. e0314921 - e0314921

Опубликована: Фев. 5, 2025

Water resource management and sustainable agriculture rely heavily on accurate Reference Evapotranspiration (ET o ). Efforts have been made to simplify the ) estimation using machine learning models. The existing approaches are limited a single specific area. There is need for ET estimations of multiple locations with diverse weather conditions. study intends propose distinct conditions federated approach. Traditional centralized require aggregating all data in one place, which can be problematic due privacy concerns transfer limitations. However, trains models locally combines knowledge, resulting more generalized estimates across different regions. three geographical Pakistan, each conditions, selected implement proposed model from 2012 2022 locations. At location, named Random Forest Regressor (RFR), Support Vector (SVR), Decision Tree (DTR), evaluated local (ET) global model. feature importance-based analysis also performed assess impacts parameters performance at location. evaluation reveals that (RFR) based outperformed other coefficient determination (R 2 = 0.97%, Root Mean Squared Error (RMSE) 0.44, Absolute (MAE) 0.33 mm day −1 , Percentage (MAPE) 8.18%. yields against site. results suggest maximum temperature wind speed most influential factors predictions.

Язык: Английский

Процитировано

2
Evaluation of Data-driven Hybrid Machine Learning Algorithms for Modelling Daily Reference Evapotranspiration DOI
Nand Lal Kushwaha, Jitendra Rajput, D.R. Sena

и другие.

ATMOSPHERE-OCEAN, Год журнала: 2022, Номер 60(5), С. 519 - 540

Опубликована: Июнь 20, 2022

Reference evapotranspiration (ET0) is one of the crucial variables used for irrigation scheduling, agricultural production, and water balance studies. This study compares six different models with sequential inclusion meteorological input such as minimum temperature (Tmin), maximum (Tmax), mean relative humidity (RH), wind speed (SW), sunshine hours (HSS), solar radiation (RS), which are necessarily in physical or empirical-based to estimate ET0. Each model utilized three variants machine learning algorithms, i.e. Additive Regression (AdR), Random Subspace (RSS), M5 Pruning tree (M5P) independently four novel permutated hybrid combinations these algorithms. To evaluate efficacy hybridizations stability models, a comprehensive evaluation independent was performed. With more variables, performances were found be superior terms prediction accuracies. The AdR6 that included all 6 selected outperformed other during testing period, exhibiting statistical performance MAPE (1.30), RMSE (0.07), RAE (2.41), RRSE (3.10), R2 (0.998). However, AdR algorithm, alone, capture about 86% variance observed data conforming 95% confidence band across irrespective number predict RSS comparison failed trends even variables. algorithms constituent better performers their accuracies but remained inferior an individual performer. All predictors higher values ET0 beyond 75% quartile.

Язык: Английский

Процитировано

30
Modeling Daily Reference Evapotranspiration from Climate Variables: Assessment of Bagging and Boosting Regression Approaches DOI Creative Commons

T. R. Jayashree,

N. V. Subba Reddy, U. Dinesh Acharya

и другие.

Water Resources Management, Год журнала: 2023, Номер 37(3), С. 1013 - 1032

Опубликована: Янв. 27, 2023

Abstract The increasing frequency of droughts and floods due to climate change has severely affected water resources across the globe in recent years. An optimal design for scheduling management irrigation is thus urgently needed adapt agricultural activities changing climate. accurate estimation reference crop evapotranspiration (ET0), a vital hydrological component balance need, tiresome task if all relevant climatic variables are unavailable. This study investigates potential four ensemble techniques estimating precise values daily ET0 at representative stations 10 agro-climatic zones state Karnataka, India, from 1979 2014. performance these models was evaluated by using several combinations as inputs tenfold cross-validation. outcomes indicated that predictions based on were most comparison with other input combinations. random forest regressor found deliver best among measures considered (Nash–Sutcliffe efficiency, 1.0, root-mean-squared error, 0.016 mm/day, mean absolute 0.011 mm/day). However, it incurred highest computational cost, whereas cost bagging model linear regression lowest. extreme gradient-boosting delivered stable modified training dataset. work here shows can be recommended ET 0 users’ interests.

Язык: Английский

Процитировано

18
Revolutionizing the Future of Hydrological Science: Impact of Machine Learning and Deep Learning amidst Emerging Explainable AI and Transfer Learning DOI Creative Commons
Rajib Maity, Aman Srivastava,

Subharthi Sarkar

и другие.

Applied Computing and Geosciences, Год журнала: 2024, Номер 24, С. 100206 - 100206

Опубликована: Ноя. 9, 2024

Язык: Английский

Процитировано

7
Forecasting Long-Series Daily Reference Evapotranspiration Based on Best Subset Regression and Machine Learning in Egypt DOI Open Access
Ahmed Elbeltagi, Aman Srivastava, Abdullah Hassan Al-Saeedi

и другие.

Water, Год журнала: 2023, Номер 15(6), С. 1149 - 1149

Опубликована: Март 15, 2023

The estimation of reference evapotranspiration (ETo), a crucial step in the hydrologic cycle, is essential for system design and management, including balancing, planning, scheduling agricultural water supply resources. When climates vary from arid to semi-arid, there are problems with lack meteorological data future information on ETo, as case Egypt, it more important estimate ETo precisely. To address this, current study aimed model Egypt’s most governorates (Al Buhayrah, Alexandria, Ismailiyah, Minufiyah) using four machine learning (ML) algorithms: linear regression (LR), random subspace (RSS), additive (AR), reduced error pruning tree (REPTree). Climate Forecast System Reanalysis (CFSR) National Centers Environmental Prediction (NCEP) was used gather daily climate variables 1979 2014. datasets were split into two sections: training phase, i.e., 1979–2006, testing 2007–2014. Maximum temperature (Tmax), minimum (Tmin), solar radiation (SR) found be three input that had influence outcome subset sensitivity analysis. A comparative analysis ML models revealed REPTree outperformed competitors by achieving best values various performance matrices during phases. study’s novelty lies use predict this algorithm has not been commonly purpose. Given sparse attempts such research, remarkable accuracy predicting highlighted rarity study. In order combat effects aridity through better resource also cautions authorities concentrate their policymaking adaptation.

Язык: Английский

Процитировано

17