Evaluation of Algorithms to estimate Daily Evapotranspiration from Instantaneous Measurements under All-sky Conditions

Zhaojie Ruan,Li Jia,Massimo Menenti

doi:10.1088/1755-1315/17/1/012133

Abstract

Instantaneous EvapoTranspiration (ET) can be estimated using a single set of instantaneous observations by polar orbiting satellites during the day. Daily, weekly or monthly total ET is required for hydrological studies and water resources assessment. This requires daily total ET regardless of cloud cover. The daily total ET is usually determined assuming that the evaporative fraction (EF) remains constant during at least the central hours of the day and that the product of EF times the daily total net radiation gives a satisfactory estimate of daily total ET. The impact of cloud cover is rarely discussed. In this paper, we used data collected at two experimental sites in the Heihe River basin in northwestern China: Arou with grassland in an alpine region and Yingke with agricultural crops in a semi-arid region. Two methods were evaluated to determine daily total ET with instantaneous observations: a) self-conservation of evaporative fraction (EF_const); b) assuming the diurnal course of ET is the same sinusoidal function of the time of the instantaneous observation as the solar irradiance (Rs) (ET-Rs). Daily ET calculated with the two methods were evaluated against observed daily values. The results showed that these assumptions did not hold and the accuracy of ET estimates obtained with either method was reduced by: a) diurnal variation of ET and b) the time lag between ET and net radiation under clear skies. Larger errors occurred when applying both methods under cloudy conditions during the growing season, while in the remaining part of the year the impact of cloud cover was lower.

Highlights

Evapotranspiration (ET) is an important component of the terrestrial water cycle, and accurate estimation of regional daily ET is essential in hydrology, agriculture and climatology
That under clear sky the evaporative fraction (EF) increases during the day in the growing season, while it has a bowl shape evolution with time in the other seasons
The assumption of the ET-Rs method is that the diurnal solar radiation and ET can be adequately described by a sine function for clear days

Summary

Introduction

Evapotranspiration (ET) is an important component of the terrestrial water cycle, and accurate estimation of regional daily ET is essential in hydrology, agriculture and climatology. ET cannot be directly measured over large areas and methods have been developed [1, 2] to use remote sensing data, e.g. land surface temperature (LST), to estimate ET over large areas. A variety of ET models [3, 4] have been developed, from simple empirical relationships Most of these models exploit polar orbiting satellite data with a higher spatial resolution, but lower observation frequency (one observation per day or less). Literature documents the use of methods based on the hypothesis of constant evaporative fraction, EF (see e.g. [6]) or on assuming that ET and solar irradiance Rs have a similar diurnal evolution, which on clear days can described by a sine

Methods

Results

Conclusion