Article
Engineering, Civil
Dilip Kumar Roy, Kowshik Kumar Saha, Mohammad Kamruzzaman, Sujit Kumar Biswas, Mohammad Anower Hossain
Summary: The study evaluated the potential of the PSO-HFS model in predicting ET0, demonstrating its superior performance compared to benchmark models. By ranking the models using Shannon's Entropy concept, the PSO-HFS model showed excellent performance on both training and testing datasets.
WATER RESOURCES MANAGEMENT
(2021)
Article
Environmental Sciences
M. Rastgou, H. Bayat, M. Mansoorizadeh, A. S. Gregory
Summary: This study compared the performance of different neural network methods and the M5 tree method in predicting the soil water retention curve in Iran, finding that RBF-based PTFs performed better in terms of the IRMSE criterion, and the developed PTFs and methods were more reliable than those by other researchers.
WATER RESOURCES RESEARCH
(2022)
Article
Environmental Sciences
Ozgur Kisi, Salim Heddam, Behrooz Keshtegar, Jamshid Piri, Rana Muhammad Adnan
Summary: In this study, the viability of RM5Tree in prediction of daily streamflow in a cold climate was investigated. Benchmark results showed that RM5Tree offered better accuracy compared to other models, both in predicting daily streamflow and estimating downstream station's flow using only upstream data.
Article
Engineering, Civil
Saeid Janizadeh, Mehdi Vafakhah, Zoran Kapelan, Naghmeh Mobarghaee Dinan
Summary: The study developed a novel flood susceptibility model based on BART methodology and tested its predictive performance in the Kan watershed in Iran, outperforming the previously published NB and RF models. The results highlighted the importance of altitude and distance from the river in assessing flooding susceptibility.
WATER RESOURCES MANAGEMENT
(2021)
Article
Biodiversity Conservation
Mumtaz Ali, Mehdi Jamei, Ramendra Prasad, Masoud Karbasi, Yong Xiang, Borui Cai, Shahab Abdulla, Aitazaz Ahsan Farooque, Abdulhaleem H. Labban
Summary: Reference evapotranspiration (ETo) is an important climate parameter affecting plant water use. A novel MVMD-BRT model was developed to accurately forecast daily ETo, and it outperformed benchmark models in terms of accuracy.
ECOLOGICAL INDICATORS
(2023)
Article
Environmental Sciences
Qiong Su, Vijay P. Singh
Summary: The Priestley-Taylor (PT) method is commonly used to calculate reference evapotranspiration (ETo) in hydrologic and crop models, but its default coefficient may not be reliable across different climatic regions. This study derived an analytical expression of PT coefficient (PTa) using the Penman-Monteith method, which improved the accuracy of ETo estimation. The global monthly PTa dataset is open-source and can be incorporated into models. The study also found that radiative component was the main driver of global ETo changes, and the impact of available energy and wind speed on ETo variations intensified in a warming climate.
WATER RESOURCES RESEARCH
(2023)
Article
Green & Sustainable Science & Technology
Jitendra Rajput, Man Singh, K. Lal, Manoj Khanna, A. Sarangi, J. Mukherjee, Shrawan Singh
Summary: This study aimed to compare the performance of regression techniques and machine learning models in estimating daily reference evapotranspiration (ET0). The M5Tree model outperformed other models in both training and testing phases, while the POR model showed the worst performance. The developed ET0 model can be utilized for efficient irrigation planning in semi-arid regions, especially in the absence of a weighing type field lysimeter.
ENVIRONMENT DEVELOPMENT AND SUSTAINABILITY
(2023)
Article
Plant Sciences
Eduardo Salgado, Nieggiorba Livellara, Esteban Chaigneau, Fernando Varas, Italo F. Cuneo
Summary: The objective of this study was to assess the relationship between hourly daily shrinkage (HDS) and hourly reference evapotranspiration (EToh) in walnut trees and pomegranate plants under different irrigation regimes. The data shows that the relationship between EToh and HDS is better than previous models.
Article
Environmental Sciences
Stavroula Dimitriadou, Konstantinos G. Nikolakopoulos
Summary: The aim of this study was to investigate the usability of artificial neural networks (ANNs) for estimating reference evapotranspiration (ETo) in the Peloponnese Peninsula, Greece. The results showed that ANNs can be a good predictive model for ETo, even with a simple architecture. The most influential factors for ETo were found to be the mean temperature (Tmean) and wind speed (u(2)).
Article
Agronomy
Hadeed Ashraf, Saliha Qamar, Nadia Riaz, Redmond R. Shamshiri, Muhammad Sultan, Bareerah Khalid, Sobhy M. Ibrahim, Muhammad Imran, Muhammad Usman Khan
Summary: The estimation of reference evapotranspiration (ETo) in Punjab, Pakistan was conducted to assess the spatiotemporal variation of climatic parameters on ETo. The results indicated that ETo was mainly influenced by minimum temperature, maximum temperature, and windspeed, with higher values in southern Punjab compared to northern Punjab. Accurate estimation of ETo can contribute to improved irrigation scheduling for different crops in Punjab.
Article
Engineering, Mechanical
Bo Wang, Xionggang Ke, Kaifan Du, Xiangjun Bi, Peng Hao, Caihua Zhou
Summary: A novel experiment strain field reconstruction method (EXP-SFRM) is proposed to improve the accuracy and monitoring effect of conventional strain field reconstruction based on finite element analysis. The method introduces the coordinates of strain gauges to establish connections between discrete strain gauges and finite element analysis. Experimental tests are conducted to validate the high accuracy and good monitoring effect of EXP-SFRM. Compared with conventional FE-SFRM, EXP-SFRM exhibits lower average errors and predicts the buckling regions more accurately.
INTERNATIONAL JOURNAL OF MECHANICAL SCIENCES
(2023)
Article
Mathematics
Zhaohao Wu, Deyun Zhong, Zhaopeng Li, Liguan Wang, Lin Bi
Summary: This paper presents a method for orebody implicit modeling based on normal estimation of cross-contour polylines. The method can automatically estimate normals and reorient them, showing advantages of low calculation, high efficiency, and strong reliability.
Article
Environmental Sciences
Ali Raza, Muhammad Shoaib, Muhammad Azhar Inam Baig, Shakil Ahmad, Mudasser Muneer Khan, Muhammad Kaleem Ullah, Sarfraz Hashim
Summary: The study suggests that in estimating crop water requirements and irrigation scheduling, MLP stands out as a superior alternative to traditional methods, enhancing the accuracy of ETo estimation.
FRESENIUS ENVIRONMENTAL BULLETIN
(2021)
Article
Green & Sustainable Science & Technology
Soo-Jin Kim, Seung-Jong Bae, Min-Won Jang
Summary: A linear regression machine learning model based on temperature data was developed to estimate reference evapotranspiration in South Korea. Compared to temperature-based empirical equations, the proposed model achieved higher accuracy and lower error when using all meteorological data.
Article
Water Resources
Ahmed Elbeltagi, Ali Raza, Yongguang Hu, Nadhir Al-Ansari, N. L. Kushwaha, Aman Srivastava, Dinesh Kumar Vishwakarma, Muhammad Zubair
Summary: This study addresses the challenge of limited climatic data in developing countries and applies five machine learning algorithms to predict daily reference evapotranspiration. It identifies the most influential climatic parameters and finds that the AR-M5P model performs better in predicting ET0 values compared to other algorithms.
APPLIED WATER SCIENCE
(2022)
Article
Soil Science
Khabat Khosravi, Ali Golkarian, Rahim Barzegar, Mohammad T. Aalami, Salim Heddam, Ebrahim Omidvar, Saskia D. Keesstra, Manuel Lopez-Vicente
Summary: In this study, machine learning models coupled with resampling algorithms were developed and tested for soil temperature forecasting. The BA-KStar model performed the best for the 5 cm soil depth, while the DA-KStar model outperformed the others for the 50 cm soil depth. All hybrid models showed higher prediction capabilities compared to the linear regression model.
Article
Computer Science, Artificial Intelligence
Rana Muhammad Adnan Ikram, Barenya Bikash Hazarika, Deepak Gupta, Salim Heddam, Ozgur Kisi
Summary: Accurate streamflow estimation is crucial for water management. This study aims to improve prediction accuracy by adopting new data preprocessing techniques and machine learning methods for streamflow estimation in high-altitude basins. The results show that the epsilon-AHELM method outperforms other methods, and the EWT technique is more effective in reducing prediction errors compared to EMD and EEMD techniques. Overall, it is recommended to use EWT-epsilon-AHELM for streamflow estimation.
NEURAL COMPUTING & APPLICATIONS
(2023)
Article
Engineering, Multidisciplinary
Rana Muhammad Adnan, Reham R. Mostafa, Hong-Liang Dai, Salim Heddam, Alban Kuriqi, Ozgur Kisi
Summary: This study investigates the feasibility of using limited climatic input data to predict monthly pan evaporation. The proposed RVM-IMRFO algorithm significantly improves the accuracy of prediction compared to other models, by an average improvement ranging from 27.65% to 8.63% in various evaluation metrics.
ENGINEERING APPLICATIONS OF COMPUTATIONAL FLUID MECHANICS
(2023)
Article
Multidisciplinary Sciences
Bijay Halder, Iman Ahmadianfar, Salim Heddam, Zainab Haider Mussa, Leonardo Goliatt, Mou Leong Tan, Zulfaqar Sa'adi, Zainab Al-Khafaji, Nadhir Al-Ansari, Ali H. Jawad, Zaher Mundher Yaseen
Summary: Climatic conditions are causing health emergencies and changes to the earth's surface. Human activities, particularly those related to industry and transportation, contribute to climate change and air pollution. Monitoring systems like Sentinel-5P and Google Earth Engine (GEE) are used to track air pollutants and their effects. Fluctuations in air pollution levels across different regions emphasize the need for further investigation and management to protect the environment and human health.
SCIENTIFIC REPORTS
(2023)
Article
Geochemistry & Geophysics
Solmaz Khazaei Moughani, Abdolbaset Osmani, Ebrahim Nohani, Saeed Khoshtinat, Tahere Jalilian, Zahra Askari, Salim Heddam, John P. Tiefenbacher, Javad Hatamiafkoueieh
Summary: A study in Iran used a deep-learning algorithm called convolutional neural network (CNN) to map surface spring potential and compared the results to predictions made by other advanced data-mining models. The study found that CNN provided the best prediction of spring locations, followed by the AB-LMT model.
Article
Biodiversity Conservation
Sungwon Kim, Youngmin Seo, Anurag Malik, Seunghyun Kim, Salim Heddam, Zaher Mundher Yaseen, Ozgur Kisi, Vijay P. Singh
Summary: Total phosphorus (T-P) is the concentration of phosphorus in water and is an important parameter for eutrophication in lakes and rivers. The current research uses neuroscience-dependent approaches to predict river T-P concentration. Singular techniques, such as machine learning and deep learning models, were developed, and double-platform synthetic techniques were integrated with prior data-processing algorithms. Evaluation was done using statistical standards and visual references, and results showed that the double-platform synthetic techniques did not always lead to more accurate predictions. The best accuracy was achieved by the singular techniques for both stations using specific models, while the double-platform synthetic techniques also achieved good predictive accuracy.
ECOLOGICAL INDICATORS
(2023)
Article
Automation & Control Systems
Meysam Alizamir, Jalal Shiri, Ahmad Fakheri Fard, Sungwon Kim, AliReza Docheshmeh Gorgij, Salim Heddam, Vijay P. Singh
Summary: This study evaluated the prediction accuracy of new machine learning methods, including WLSTM, WMLPANN, LSTM, MLPANN, and MARS, for modeling daily solar radiation. Different combinations of climatic data were used as input, and the results showed that the WLSTM method outperformed the other methods in estimating solar radiation values at two stations in Illinois, USA. The average RMSE values of the other methods decreased when using the WLSTM method, indicating the successful application of the hybridization of LSTM with the wavelet transform technique in improving the prediction accuracy of solar radiation based on climatic parameters.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2023)
Article
Meteorology & Atmospheric Sciences
Rana Muhammad Adnan, Amin Mirboluki, Mojtaba Mehraein, Anurag Malik, Salim Heddam, Ozgur Kisi
Summary: This study compares the performance of Long Short-Term Memory (LSTM) tuned with Grey Wolf Optimization (GWO) and several other machine learning models in the prediction of monthly streamflow. The results show that LSTM-GWO is the best model in terms of accuracy and is particularly effective in predicting peak streamflow. Additionally, the study demonstrates that LSTM-GWO outperforms other models, such as ANN-GWO, SVR-GWO, and ANFIS-GWO, in predicting streamflow using data from different stations.
THEORETICAL AND APPLIED CLIMATOLOGY
(2023)
Article
Engineering, Civil
Ahmed M. Al-Areeq, S. I. Abba, Bijay Halder, Iman Ahmadianfar, Salim Heddam, Vahdettin Demir, Huseyin Cagan Kilinc, Aitazaz Ahsan Farooque, Mou Leong Tan, Zaher Mundher Yaseen
Summary: In this study, the authors improve flood susceptibility mapping using machine learning models and compare the performance of ensemble algorithms (Light GBM) and Elastic-net Classifier. They create a flood inventory map using satellite images and field observations, and evaluate the accuracy of the models using receiver operating characteristic (ROC) curve and area under the curve (AUC). The results indicate that the traditional Elastic-net Classifier model performs better than the ensemble algorithm in terms of accuracy. These algorithms have the potential to provide a practical and affordable method for geospatial modeling of flood vulnerability.
WATER RESOURCES MANAGEMENT
(2023)
Article
Environmental Sciences
Meysam Alizamir, Zahra Kazemi, Zohre Kazemi, Majid Kermani, Sungwon Kim, Salim Heddam, Ozgur Kisi, Il-Moon Chung
Summary: The likelihood of water contamination near landfills is high due to leachate percolation. Therefore, it is crucial to create a reliable framework for monitoring leachate and groundwater parameters to control water quality. An efficient hybrid artificial intelligence model, ELM-GWO, has been used to predict landfill leachate and groundwater quality at Saravan landfill in Iran.
Article
Green & Sustainable Science & Technology
Meysam Alizamir, Kaywan Othman Ahmed, Jalal Shiri, Ahmad Fakheri Fard, Sungwon Kim, Salim Heddam, Ozgur Kisi
Summary: Reliable and precise estimation of solar energy is crucial in energy management, especially in developing countries. This study developed different artificial intelligence models for solar radiation estimation in Kurdistan region, Iraq, and found that the proposed VMD-DELM algorithm significantly improved the accuracy of daily solar radiation prediction.
Article
Energy & Fuels
Abdelhamid Ouladmansour, Ouafi Ameur-Zaimeche, Rabah Kechiched, Salim Heddam, David A. Wood
Summary: This study investigates the relationship between rock porosity and drilling variables using machine learning models, specifically the K-nearest neighbor, random forest, support vector regression, and eXtreme Gradient Boosting models. The XGBoost model provides the most accurate porosity predictions for the individual and combined reservoirs.
GEOENERGY SCIENCE AND ENGINEERING
(2023)
Article
Environmental Sciences
Ala Bouchehed, Fares Laouacheria, Salim Heddam, Lakhdar Djemili
Summary: In this study, three machine learning methods (SVR, RVM, and GPR) were used to predict seepage flow through embankment dams. The models were developed using measured seepage flow and piezometer level data obtained from the dam. The models were calibrated and validated using various performance metrics, and the results showed that the proposed models are a good alternative to the in situ measured data. The RVM model had the best performance, followed by the GPR model, while the SVR model performed the poorest.
ENVIRONMENTAL SCIENCE AND POLLUTION RESEARCH
(2023)
Article
Ecology
Salim Heddam, Khaled Merabet, Salah Difi, Sungwon Kim, Mariusz Ptak, Mariusz Sojka, Mohammad Zounemat-Kermani, Ozgur Kisi
Summary: In this study, a new hybrid machine learning model is proposed for predicting water temperature based on signal decomposition algorithms. The experimental results show that the new model achieves high predictive accuracies and has the potential to improve water resources management and assessment of water temperature variability over time and space.
ECOLOGICAL INFORMATICS
(2023)
Article
Computer Science, Interdisciplinary Applications
Javad Hatamiafkoueieh, Salim Heddam, Saeed Khoshtinat, Solmaz Khazaei, Abdol-Baset Osmani, Ebrahim Nohani, Mohammad Kiomarzi, Ehsan Sharafi, John Tiefenbacher
Summary: In this study, a vote algorithm was developed to improve the performances of three machine-learning models for soil temperature forecasting. The results showed that the V-RF model performed the best for 3-day ahead forecasting at a depth of 5 cm, while the V-RT model was the least effective.
JOURNAL OF HYDROINFORMATICS
(2023)