Abstract:Quantitative estimation of evapotranspiration (ET) plays a significant role in the mechanism study of water cycling and agriculture management. Remotely sensed satellites provide spatial continuous surface parameters, which make estimation of regional surface ET possible. However, surface parameters retrieved by remote sensing data is discontinuous temporally, which can only provide 1-2 instantaneous values at satellite over-passing time. Thus, time scale extrapolation process is of crucial necessity for daily ET estimation. Recently, many ET time scale extrapolation methods have been proposed, such as sine function method, constant evaporative fraction (EF) method, and crop coefficient method. It has been proved that different models have different applicability and accuracy over variant climate conditions and underlying surfaces. Many studies assemble these extrapolation methods into remote sensing models to estimate regional daily ET in China, however, little researches focus on the accuracy assessment of these methods. The purpose of this study is to compare and assess the accuracy of four commonly used ET extrapolation methods over variant climate conditions and vegetation types. The study was conducted at Gaoying flux site located in North China Plain and Tongyu flux long-term land surface processes observational station located in North-east China Plain to represent semi-humid and semi-arid climate respectively. The predominant crops of Gaoying site are winter wheat and summer maize planted in rotation. Tongyu long-term land surface processes observational station has two flux sites located on the sorghum and degraded grassland, respectively. Each site of Tongyu station as well as Gaoying site has an EC system to record half-hour latent heat flux and sensible heat flux. Four commonly used ET extrapolation methods selected from literatures were tested in this study. The selection was carried out considering their applicability to the Remote sensing data used. The first method is the constant evaporative fraction (EF) method, which assumes the evaporative fraction, defined as the ratio between latent heat flux and available energy is stable during daytime under fair weather. The second method is a variance to constant EF method (so called variant EF method), with a change that soil heat flux is neglected. One assumption implied in variant EF method is the average daily soil heat tends to be zero and could be excluded in estimating daily ET. The third method is sine function method, that proposed by the assumption that the ratio of instantaneous latent heat flux to daily ET to the diurnal trend of solar irradiance with a sine curve. The fourth method is the crop coefficient method, which implies the ratio of actual ET to reference ET (i.e. crop coefficient) stays stable during the day. Four wheat seasons and four maize seasons from 2005 to 2008 at Gaoying flux site, two seasons of sorghum, grass land and bare soil at Tongyu flux site are used in this study. By accumulating the half-hour LE measurements, daily ET is calculated and used as validation data. Results show that all the four methods have the similar patterns, but the variant EF method has the best accuracy and performs well over the typical croplands in north China. The assumption implying in the extrapolation methods that a certain parameter (such as EF, crop coefficient, etc.) remains stable during daytime does not work in the study areas. Therefore, estimation derived from morning satellite underestimates the daily evapotranspiration, but overvalues from afternoon satellite. The diurnal pattern of evaporative fraction will come next to get a better understand of extrapolation from instantaneous value to daily evapotranspiration, and a new extrapolation approach will be proposed.

Comparison of CLDAS and Machine Learning Models for Reference Evapotranspiration Estimation under Limited Meteorological Data

Optimal Methods for Estimating Shortwave and Longwave Radiation to Accurately Calculate Reference Crop Evapotranspiration in the High-Altitude of Central Tibet

Estimating daily reference evapotranspiration based on limited meteorological data using deep learning and classical machine learning methods

Improved Remote Sensing Reference Evapotranspiration Estimation Using Simple Satellite Data and Machine Learning.

Accessible Remote Sensing Data Based Reference Evapotranspiration Estimation Modelling

Nation-scale reference evapotranspiration estimation by using deep learning and classical machine learning models in China

Parameterization of the Ångström–Prescott formula based on machine learning benefit estimation of reference crop evapotranspiration with missing solar radiation data

Parameter regionalization based on machine learning optimizes the estimation of reference evapotranspiration in data deficient area

Research on methods for estimating reference crop evapotranspiration under incomplete meteorological indicators

Machine learning-based estimation of evapotranspiration under adaptation conditions: a case study in Heilongjiang Province, China

Utilizing Machine Learning Models with Limited Meteorological Data as Alternatives for the FAO-56PM Model in Estimating Reference Evapotranspiration

A Novel Integrated Method Based on a Machine Learning Model for Estimating Evapotranspiration in Dryland

Machine Learning-Based Estimation of Daily Cropland Evapotranspiration in Diverse Climate Zones

Improving solar radiation estimation in China based on regional optimal combination of meteorological factors with machine learning methods

Reference Evapotranspiration Modeling Using New Heuristic Methods

Optimal Alternative for Quantifying Reference Evapotranspiration in Northern Xinjiang

Modelling Seasonal Evapotranspiration of Arid Lands in China

Comparison of temporal extrapolation methods for evapotranspiration over variant underlying croplands

High performance machine learning approach for reference evapotranspiration estimation

Improving terrestrial evapotranspiration estimation across China during 2000–2018 with machine learning methods

Estimation of Evapotranspiration Based on METRIC and TTME Models from Landsat Data over Heterogeneous Land Surfaces in Arid Region of China