Forecasting of Reference Crop Evapotranspiration in Huang-Huai-Hai Plain Based on XGBoost Machine Learning Model and Numerical Weather Prediction Data

doi:10.11924/j.issn.1000-6850.casb2023-0793

Abstract

Abstract:

This study seeks to enhance the prediction accuracy of reference crop evapotranspiration (ET₀) in the Huang-Huai-Hai Plain. Three models, namely BP, XGBoost, and CatBoost, were developed utilizing historical daily meteorological data from 2020-2021. These data included maximum air temperature (T_max), minimum air temperature (T_min), total solar radiation (R_a), sunshine hours (S), relative humidity (RH), and wind speed at a height of 2 m (U₂) for training purposes. The models were tested and forecasted using daily numerical weather prediction data from Xinxiang city, located in northern Henan Province, for the year 2022. The forecasted results were compared with the ET₀ data calculated using FAO-56 Penman-Monteith model. The results showed that R_a, T_max, and T_min were most correlated with ET₀ among all factors, and therefore were considered as input factors for model running. The three models generated acceptable ET₀ accuracy at 1-16 d forecast scale. In model testing, XGBoost model had R=0.875, RMSE=0.230 mm/d, MAE=0.181 mm/d, and MAPE=8.45%, respectively. On average, the R value of XGBoost model was increased by 10.2%, and values of RMSE, MAE, and MAPE were decreased by 39.9%-62.4%, compared to BP and CatBoost models. In view of the accuracy and stability of the XGBoost model, it can be a recommended model for ET₀ forecasting in the Huang-Huai-Hai Plain.

Key words: P-M model, BP neural network model, reference crop evapotranspiration, machine learning, forecasting models, Huang-Huai-Hai Plain, XGBoost, CATBoost

ZHU Chunxia, QIN Anzhen. Forecasting of Reference Crop Evapotranspiration in Huang-Huai-Hai Plain Based on XGBoost Machine Learning Model and Numerical Weather Prediction Data[J]. Chinese Agricultural Science Bulletin, 2024, 40(28): 126-133.

Figures/Tables 6

References 47

[1]	强小嫚, 张凯, 米兆荣, 等. 黄淮海平原地区深松和灌水次数对冬小麦-夏玉米节水增产的影响[J]. 中国农业科学, 2019, 52(3):491-502. doi: 10.3864/j.issn.0578-1752.2019.03.009
[2]	张东霞, 秦安振. 冬小麦―夏玉米作物蒸散量及其水热关系研究[J]. 作物杂志, 2022(6):145-151.
[3]	李金琴, 王宇飞, 侯旭光, 等. 不同节水种植模式对玉米产量的影响[J]. 北方农业学报, 2017, 45(4):30-34.
[4]	PERERA K, WESTERN A, NAWARATHNA B, et al. Forecasting daily reference evapotranspiration for Australia using numerical weather prediction outputs[J]. Agricultural and forest meteorology, 2014, 194:50-63.
[5]	CAI J, LIU Y, LEI T, et al. Estimating reference evapotranspiration with the FAO Penman-Monteith equation using daily weather forecast messages[J]. Agricultural and forest meteorology, 2007, 145(1-2):22-35.
[6]	付浩龙, 罗玉峰, 李亚龙. 基于气温预报的HS公式短期逐日参考作物腾发量预报评价与分析[J]. 长江科学院院报, 2018, 35(2):51-56. doi: 10.11988/ckyyb.20160868
[7]	刘梦, 罗玉峰, 汪文超, 等. 基于天气预报的漳河灌区参考作物腾发量预报方法比较[J]. 农业工程学报, 2017, 33(19):156-162.
[8]	CHEN Y, BAI M, ZHANG Y, et al. Multivariable space-time correction for wind speed in numerical weather prediction (NWP) based on ConvLSTM and the prediction of probability interval[J]. Earth science informatics, 2023, 16(3):1953-1974.
[9]	LIU B, LIU M, CUI Y, et al. Assessing forecasting performance of daily reference evapotranspiration using public weather forecast and numerical weather prediction[J]. Journal of hydrology, 2020, 590:125547.
[10]	SARMAS E, SPILIOTIS E, STAMATOPOULOS E, et al. Short-term photovoltaic power forecasting using meta-learning and numerical weather prediction independent long short-term memory models[J]. Renewable energy, 2023, 216(11):118997
[11]	ZHANG W, CHANG C, XUAN Y. The impacts of climate change on bank performance: What’s the mediating role of natural disasters[J]. Economic change and restructuring, 2022, 55:1913-1952.
[12]	FAN J, WU L, ZHANG F, et al. Evaluation and development of empirical models for estimating daily and monthly mean daily diffuse horizontal solar radiation for different climatic regions of China[J]. Renewable & sustainable energy reviews, 2019, 105:168-186.
[13]	QIN A, NING D, LIU Z, et al. Insentek sensor: An alternative to estimate daily crop evapotranspiration for maize plants[J]. Water, 2019, 11:25.
[14]	LIU Q, YANG Z, CUI B, et al. The temporal trends of reference evapotranspiration and its sensitivity to key meteorological variables in the Yellow River Basin, China[J]. Hydrological processes, 2010, 24:2171-2181.
[15]	FAN J, WU L, ZHANG F, et al. Climate change effects on reference crop evapotranspiration across different climatic zones of China during 1956-2015[J]. Journal of hydrology, 2016, 542:923-937.
[16]	YANG Y, CUI Y, BAI K, et al. Short-term forecasting of daily reference evapotranspiration using the reduced-set Penman-Monteith model and public weather forecasts[J]. Agricultural water management, 2019, 211:70-80.
[17]	杨会娟, 粟晓玲, 郭静. 基于支持向量机的干旱区月潜在蒸散发的模拟[J]. 中国农村水利水电, 2016(7):6-10.
[18]	张乐坚, 程明虎, 田付友. 人工神经网络及支持向量机在降雨量预报中的应用[J]. 高原气象, 2010, 29(4):982-991.
[19]	冯丽丽, 张琨, 韩拓, 等. 基于人工神经网络的不同植被类型蒸散量时空尺度扩展[J]. 兰州大学学报(自然科学版), 2017, 53(2):186-193.
[20]	谭君位, 崔远来, 张培青. 实时灌溉预报与灌溉用水决策支持系统研究与应用[J]. 中国农村水利水电, 2015(7):1-4,9.
[21]	任传栋, 王志真, 马钊, 等. 基于深度学习及传统机器学习模型估算山东省参考作物蒸散量[J]. 节水灌溉, 2022(3):67-74.
[22]	GRANATA F. Evapotranspiration evaluation models based on machine learning algorithms- A comparative study[J]. Agricultural water management, 2019, 217:303-315.
[23]	PATIL A, DEKA P. An extreme learning machine approach for modeling evapotranspiration using extrinsic inputs[J]. Computers and electronics in agriculture, 2016, 121:385-392.
[24]	李俊. 基于改进的BP神经网络的水面蒸发预测[J]. 节水灌溉, 2017(11):42-45.
[25]	杜云皓, 仇锦先, 冯绍元. 改进GA-BP模型在地下水位埋深预测中的应用[J]. 节水灌溉, 2017(9):81-84.
[26]	QIN A, FAN Z, ZHANG L. Hybrid genetic algorithm-based BP neural network models optimize estimation performance of reference crop evapotranspiration in China[J]. Applied sciences, 2022, 12(20):10689.
[27]	FAN J, YUE W, WU L, et al. Evaluation of SVM, ELM and four tree-based ensemble models for predicting daily reference evapotranspiration using limited meteorological data in different climates of China[J]. Agricultural and forest meteorology, 2018, 263:225-241.
[28]	张薇, 霍树义, 贾悦. 机器学习模型在河北省参考作物蒸散量计算中的比较[J]. 节水灌溉, 2018(4):50-53.
[29]	郝林如, 郭向红, 雷涛, 等. 气象要素缺失条件下不同机器学习模型计算参考作物蒸散量比较[J]. 节水灌溉, 2022(7):102-108.
[30]	牛曼丽, 李红岺, 李新旭. 基于CatBoost的温室日参考作物蒸发蒸腾量估算模型研究[J]. 节水灌溉, 2022(1):14-19.
[31]	HUANG G, WU L, MA X, et al. Evaluation of CatBoost method for prediction of reference evapotranspiration in humid regions[J]. Journal of hydrology, 2019, 574:1029-1041.
[32]	刘战东, 肖俊夫, 牛豪震, 等. 地下水埋深对冬小麦和春玉米产量及水分生产效率的影响[J]. 干旱地区农业研究, 2011, 29(1):29-33.
[33]	陈志月, 吴立峰, 刘小强, 等. 基于GPR、CatBoost、XGBoost三种模型预测江西地区水面蒸发量[J]. 水资源与水工程学报, 2020, 31(6):116-125,131.
[34]	李可利, 张鑫. 基于ANFIS的陕西省参考作物蒸散量计算[J]. 自然资源学报, 2020, 35(6):1472-1483. doi: 10.31497/zrzyxb.20200618
[35]	黄滟淳, 崔宁博, 陈宣全, 等. 基于不同机器学习模型的川中丘陵区参考作物蒸散量模拟[J]. 中国农村水利水电, 2020(5):13-20,27.
[36]	冯培存, 魏正英, 张育斌, 等. 基于思维进化算法优化BP神经网络的温室甜瓜ET₀预测研究[J]. 节水灌溉, 2019(9):36-39.
[37]	CHEN T, GUESTRIN C. XGBoost: A scalable tree boosting system[A]. //Proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining (KDD)[C]. New York: ACM, 2016:785-794.
[38]	ZHANG Y, ZHAO Z, ZHENG J. CatBoost: A new approach for estimating daily reference crop evapotranspiration in arid and semi-arid regions of Northern China[J]. Journal of hydrology, 2020,588.
[39]	KISI O, GENC O, DINC S, et al. Daily pan evaporation modeling using Chi-squared automatic interaction detector, neural networks, classification and regression tree[J]. Computers and electronics in agriculture, 2016, 122:112-117.
[40]	LU L, ZHANG D, ZHANG J, et al. Ecosystem evapotranspiration partitioning and its spatial-temporal variation based on eddy covariance observation and machine learning method[J]. Remote sensing, 2023, 15(19):4831.
[41]	FENG Y, JIA Y, CUI N, et al. Calibration of Hargreaves model for reference evapotranspiration estimation in Sichuan basin of southwest China[J]. Agricultural water management, 2017, 181:1-9.
[42]	FENG Y, PENG Y, CUI N, et al. Modeling reference evapotranspiration using extreme learning machine and generalized regression neural network only with temperature data[J]. Computers and electronics in agriculture, 2017, 136:71-78.
[43]	谭鑫, 罗童元, 谢亨旺, 等. 基于气温预报的江西省参考作物腾发量预报方法比较与分析[J]. 中国农村水利水电, 2022(4):150-155.
[44]	白依文, 鲁梦格, 程浩楠, 等. 基于天气预报信息的参考作物需水量预报研究[J]. 中国农村水利水电, 2021(10):175-179.
[45]	SHARMA G, SINGH A, JAIN S. Hybrid deep learning techniques for estimation of daily crop evapotranspiration using limited climate data[J]. Computers and electronics in agriculture, 2022, 202:107338.
[46]	ABDULLAH S, MALEK M, ABDULLAH N, et al. Extreme learning machines: A new approach for prediction of reference evapotranspiration[J]. Journal of hydrology, 2015, 527:184-195.
[47]	JHAVERI S, KHEDKAR I, KANTHARIA Y, et al. Success prediction using Random Forest, CatBoost, XGBoost and AdaBoost for Kickstarter Campaigns[A].// 2019 3rd International conference on computing methodologies and communication (ICCMC)[C]. IEEE, 2019(2):1170-1173.

	T_max	T_min	T_mean	R_a	S	RH	U₂
T_min	0.912^**	1.000
T_mean	0.943^**	0.867^**	1.000
R_a	0.865^**	0.835^**	0.876^**	1.000
S	0.635^*	0.685^*	0.638^*	0.785^*	1.000
RH	0.762^*	0.635^*	0.586	0.672^*	0.324	1.000
U₂	0.124	0.213	0.215	0.314	0.121	0.168	1.000
ET₀	0.884^**	0.715^*	0.735^*	0.727^*	0.477	0.458	0.125

	T_max	T_min	T_mean	R_a	S	RH	U₂
T_min	0.912^**	1.000
T_mean	0.943^**	0.867^**	1.000
R_a	0.865^**	0.835^**	0.876^**	1.000
S	0.635^*	0.685^*	0.638^*	0.785^*	1.000
RH	0.762^*	0.635^*	0.586	0.672^*	0.324	1.000
U₂	0.124	0.213	0.215	0.314	0.121	0.168	1.000
ET₀	0.884^**	0.715^*	0.735^*	0.727^*	0.477	0.458	0.125

模型	训练阶段				验证阶段
模型	R	RMSE/(mm/d)	MAE/(mm/d)	MAPE/%	R	RMSE/(mm/d)	MAE/(mm/d)	MAPE/%
BP	0.817b	0.496a	0.864a	23.5a	0.757b	0.485a	0.452a	33.6a
CatBoost	0.857b	0.335b	0.268b	15.2b	0.831a	0.281b	0.226b	11.4b
XGBoost	0.923a	0.183c	0.122c	7.32c	0.875a	0.230b	0.181c	8.45b

模型	训练阶段				验证阶段
模型	R	RMSE/(mm/d)	MAE/(mm/d)	MAPE/%	R	RMSE/(mm/d)	MAE/(mm/d)	MAPE/%
BP	0.817b	0.496a	0.864a	23.5a	0.757b	0.485a	0.452a	33.6a
CatBoost	0.857b	0.335b	0.268b	15.2b	0.831a	0.281b	0.226b	11.4b
XGBoost	0.923a	0.183c	0.122c	7.32c	0.875a	0.230b	0.181c	8.45b

模型	预报天数/d	R	RMSE/(mm/d)	MAE/(mm/d)	MAPE/%
BP	1~8	0.65b	0.185c	0.240b	21.6b
	1~16	0.54c	0.230b	0.281b	27.6ab
	1~32	0.33e	0.315a	0.311a	30.7a
	1~56	0.31e	0.326a	0.325a	32.6a
CatBoost	1~8	0.62b	0.190c	0.183c	9.63c
	1~16	0.55c	0.176c	0.153cd	12.6c
	1~32	0.30e	0.275b	0.221b	22.6b
	1~56	0.28e	0.281b	0.226b	24.8b
XGBoost	1~8	0.72a	0.118c	0.114d	4.65d
	1~16	0.74a	0.126c	0.113d	10.7c
	1~32	0.48d	0.226b	0.193c	16.7b
	1~56	0.43d	0.230b	0.181c	15.6b