Linear Regression Machine Learning Algorithms for Estimating Reference Evapotranspiration Using Limited Climate Data

Abstract

A linear regression machine learning model to estimate the baseline evapotranspiration based on temperature data for South Korea is developed in this study. FAO56 Penman–Monteith (FAO56 P–M) reference evapotranspiration calculated with meteorological data (1981–2021) obtained from sixty-two meteorological stations nationwide is used as the label. All study datasets provide daily, monthly, or annual values based on the average temperature, daily temperature difference, and extraterrestrial radiation. Multiple linear regression (MLR) and polynomial regression (PR) are applied as machine learning algorithms, and twelve models are tested using the training data. The results of the performance evaluation of the period from 2017 to 2021 show that the polynomial regression algorithm that learns the amount of extraterrestrial radiation achieves the best performance (the minimum root-mean-square errors of 0.72 mm/day, 11.3 mm/month, and 40.5 mm/year for daily, monthly, and annual scale, respectively). Compared to temperature-based empirical equations, such as Hargreaves, Blaney–Criddle, and Thornthwaite, the model trained using the polynomial regression algorithm achieves the highest coefficient of determination and lowest error with the reference evapotranspiration of the FAO56 Penman–Monteith equation when using all meteorological data. Thus, the proposed method is more effective than the empirical equations under the condition of insufficient meteorological data when estimating reference evapotranspiration.

Modeling Daily Reference Evapotranspiration from Climate Variables: Assessment of Bagging and Boosting Regression Approaches

“…The techniques such as Extreme Learning Machine (ELM), Generalized Regression Neural Network (GRNN), Artificial Neural Network (ANN), and Linear Regression models demonstrated their effectiveness in the evaluation of ET 0 using only temperature data (Feng et al 2017;Kim et al 2022) under different time scales in a variety of climatic zones. The modeling of ET0 by using input combinations composed of various climatic variables, such as the temperature, relative humidity, solar radiation, and wind speed, has been explored by using the ANN (Gocić and Amiri 2021;Elbeltagi et al 2022a, b), support vector machine (SVM) and tree-based models (Fan et al 2018), adaptive neuro-fuzzy inference system (ANFIS)-firefly algorithm (Tao et al 2018), hybrid ELM with optimization (Zhu et al 2020), and fuzzy logic-based hierarchical fuzzy system (HFS)-particle swarm optimization (PSO) algorithm (Roy et al 2021).…”

The mathematical catalyst deactivation models: a mini review

“…Similarly, the optimization methods applied an articial neural network (ANN), Genetic Algorithm (GA) and Particle Swarm Optimization (PSO). 124 These optimization methods were designed iteratively to estimate the best kinetic and deactivation parameters which were expected to yield the minimum error between the experimental and predicted Table 6 The error evaluation expressions 125…”

Reference evapotranspiration prediction using machine learning models: An empirical study from minimal climate data

“…In recent years, several research scholars have utilized ML models to estimate ET 0 (Bellido-Jiménez et al, 2022;Chen et al, 2022;Farooque et al, 2022;Nagappan et al, 2020;Rajput et al, 2023c;Wu et al, 2019). Accuracy analysis of regression and ML methods advocates that these techniques estimate ET 0 more accurately than empirical techniques (Chia et al, 2020;Kim et al, 2022), which utilized minimal climatic parameters. Artificial neural network (ANN), which has become popular, is extensively employed in ET computations (Dimitriadou & Nikolakopoulos, 2022;Tikhamarine et al, 2019).…”