干旱区生态环境知识资源中心(Arid): Comparative Assessment of Reference Evapotranspiration Estimation Using Conventional Method and Machine Learning Algorithms in Four Climatic Regions

Arid

DOI	10.1007/s00024-020-02473-5
	Comparative Assessment of Reference Evapotranspiration Estimation Using Conventional Method and Machine Learning Algorithms in Four Climatic Regions
	Raza, Ali 1; Shoaib, Muhammad 1; Faiz, Muhammad Abrar 2; Baig, Faisal 1; Khan, Mudasser Muneer 3; Ullah, Muhammad Kaleem 4; Zubair, Muhammad 5
通讯作者	Raza, Ali
来源期刊	PURE AND APPLIED GEOPHYSICS
ISSN	0033-4553
EISSN	1420-9136
出版年	2020
卷号	177 期号:9 页码:4479-4508
英文摘要	Reference evapotranspiration (ETo) is considered as an essential component in hydrological and agro meteorological processes. Its accurate estimation becomes an imperative in the planning and management of irrigation practices. ETo estimation also plays a vital role in improving the irrigation efficiency, water reuse and irrigation scheduling. The conventional physical model of Penmen Montieth (PM56) developed by Food and agriculture organization (FAO) has been recommended worldwide for ETo estimation. This model was firstly used in this study to determine ETo by using required meteorological data and obtained results used as referenced values. Afterward, five data machine learning algorithms/data driven models, support vector machine (SVM), multilayer perceptron (MLP), group method of data handling (GMDH), general regression neural network (GRNN) and cascade correlation neural network (CCNN), were applied to estimate ETo values. The climatic data of maximum and minimum temperatures, wind speed, average relative humidity and sunshine hours of six stations from Pakistan was used to train and test data driven model. Data driven models were also applied on other climatic stations without training data which lie in China, New Zealand and USA to further validate and investigate their performance. Comparison results indicated that model efficiency (ME) and correlation coefficient (r) of SVM were obtained (ranges: ME = 95-99%; r = 0.96-1) maximum for all the selected stations. Alternatively, model errors (RMSE = 0.016, MSE = 0.0001 & MAE = 0.08) for SVM were found minimum in comparison to GMDH, MLP, CCNN and GRNN. In addition, all data driven models show enough divergence from hyper arid to high humid climate except SVM which shows almost identical results for all the climatic zones in comparison to standard FAO-PM56 method. Finally, it can be concluded that SVM could be considered as a reliable alternative method for ETo estimation among data driven models.
英文关键词	Reference evapotranspiration machine learning algorithms cross validation climatic regions
类型	Article
语种	英语
国家	Pakistan ; Peoples R China
收录类别	SCI-E
WOS记录号	WOS:000523052300001
WOS关键词	ARTIFICIAL NEURAL-NETWORKS ; SUPPORT VECTOR MACHINES ; EQUATIONS ; MODELS ; PERFORMANCE ; EVAPORATION
WOS类目	Geochemistry & Geophysics
WOS研究方向	Geochemistry & Geophysics
资源类型	期刊论文
条目标识符	http://119.78.100.177/qdio/handle/2XILL650/315379
作者单位	1.Bahauddin Zakariya Univ, Dept Agr Engn, Multan, Pakistan; 2.Northeast Agr Univ, Coll Water Conservancy & Civil Engn, Harbin, Peoples R China; 3.Bahauddin Zakariya Univ, Dept Civil Engn, Multan, Pakistan; 4.Univ Lahore, Dept Civil Engn, Lahore, Pakistan; 5.NFC Inst Engn & Technol, Dept Civil Engn, Multan, Pakistan
推荐引用方式 GB/T 7714	Raza, Ali,Shoaib, Muhammad,Faiz, Muhammad Abrar,et al. Comparative Assessment of Reference Evapotranspiration Estimation Using Conventional Method and Machine Learning Algorithms in Four Climatic Regions[J],2020,177(9):4479-4508.
APA	Raza, Ali.,Shoaib, Muhammad.,Faiz, Muhammad Abrar.,Baig, Faisal.,Khan, Mudasser Muneer.,...&Zubair, Muhammad.(2020).Comparative Assessment of Reference Evapotranspiration Estimation Using Conventional Method and Machine Learning Algorithms in Four Climatic Regions.PURE AND APPLIED GEOPHYSICS,177(9),4479-4508.
MLA	Raza, Ali,et al."Comparative Assessment of Reference Evapotranspiration Estimation Using Conventional Method and Machine Learning Algorithms in Four Climatic Regions".PURE AND APPLIED GEOPHYSICS 177.9(2020):4479-4508.