Effect of the net radiation substitutes on maize and soybean evapotranspiration estimation using machine learning methods

Agriscientia

article

E. Walker¹  D. C. Fonnegra Mora¹  G. Fagioli²  V Venturini¹ 

[1] Universidad Nacional del Litoral ,(UNL), Facultad de Ingeniería y Ciencias Hídricas ,(FICH). Ciudad Universitaria. Ruta Nacional Nº 168 – Km 472;KILIMO S.A.

关键词: water stress;    net radiation;    crops;    machine learning;    Adaptive Boosting;   

DOI  :  10.31047/1668.298x.v39.n2.37104

学科分类：社会科学、人文和艺术（综合）

来源: Universidad Nacional de Cordoba * Facultad de Ciencias Agropecuarias

【 摘 要 】

Accurate evapotranspiration (ET) estimation is essential for water managementin crops, but it is not an easy task. Empirical ET methodologies require precisenet radiation (Rn) measurements to obtain accurate results. Nevertheless, Rnmeasurements are not easy to obtain from meteorological stations. Thus, thisstudy explored the use of machine learning algorithms with two Rn substitutes,to estimate daily ET: the extraterrestrial solar radiation (Ra) and a modelledRn (RnM). Support Vector Machine (SVM), Kernel Ridge (KR), Decision Tree(DT), Adaptive Boosting (AB), and Multilayer Perceptron (MLP) were applied tomodel FLUXNET Rn and ET observations. Adaptive Boosting produced the bestfield Rn measurements (RnO), yielding a Root Mean Square Error of about 16 %of the mean observed Rn. The resulting Rn (AB RnM) was used to model dailycrops ET employing the above-mentioned machine learning methods with RnO,AB RnM, and Ra, in conjunction with meteorological variables and the NDVIindex. The evaluated methods were suitable to estimate ET, yielding similarerrors to those obtained with RnO, when contrasted with ET observations. Theseresults demonstrate that AB and KR are applicable with rutinary meteorologicaland satellite data to estimate ET.

【 授权许可】

CC BY-SA