A novel hybrid WOA-XGB model for estimating daily reference evapotranspiration using local and external meteorological data: Applications in arid and humid regions of China

Shicheng Yan, Lifeng Wu, Junliang Fan, Fucang Zhang, Yufeng Zou and You Wu

Agricultural Water Management, 2021, vol. 244, issue C

Abstract: The information of reference evapotranspiration (ET0) is vital for optimizing irrigation scheduling, planning water resources and assessing hydrological drought. However, accurate estimation of ET0 is difficult if long-term or complete climatic variables are unavailable, especially in developing countries like China. This study proposed a novel hybrid extreme gradient boosting (XGB) model with the whale optimization algorithm (WOA) to estimate daily ET0 at four stations in the arid region and four stations in the humid region of China. Particularly, its performances were evaluated under the local and three external scenarios with seven incomplete combinations of maximum and minimum temperatures (Tmax and Tmin), relative humidity (RH), wind speed (U2), relative sunshine duration (n/N) and extra-terrestrial radiation (Ra) for the period 1966–2015. The results showed that U2 was the most influencing variable for daily ET0 estimation in the arid region, followed by n/N and RH, while n/N was more important than RH and U2 in the humid region. Locally trained and tested WOA-XGB models greatly outperformed their corresponding simplified FAO-56 PM models, with the average decrease in root mean square error (RMSE) by 40.1% and 38.9% in the arid and humid regions, respectively. Compared with local WOA-XGB models, the prediction accuracy of externally trained WOA-XGB models with local or external testing data decreased by 18.1% or 69.9% in the arid region, and 16.8% or 67.9% in the humid region, respectively. However, external WOA-XGB models with synthetic testing data from the target and adjacent stations overall improved the prediction accuracy of local WOA-XGB models by 5.7% and 9.6% in the arid and humid regions, respectively. The results indicated that external WOA-XGB models with local testing data produced acceptable daily ET0 estimates. However, when synthetic data were employed during testing, external WOA-XGB models gave excellent daily ET0 estimates, which were comparable to or even better than local WOA-XGB models. This is a promising strategy that allows more accurate estimation of daily ET0 when lack of long-term historical or complete recent data.

Keywords: Reference evapotranspiration; Extreme gradient boosting; Whale optimization algorithm; Cross station; FAO-56 Penman–Monteith (search for similar items in EconPapers)
Date: 2021
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (13)

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0378377420321417
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:agiwat:v:244:y:2021:i:c:s0378377420321417

DOI: 10.1016/j.agwat.2020.106594

Access Statistics for this article

Agricultural Water Management is currently edited by B.E. Clothier, W. Dierickx, J. Oster and D. Wichelns

More articles in Agricultural Water Management from Elsevier
Bibliographic data for series maintained by Catherine Liu ().