Determinants of the Asymmetric Parameter in the Generalized Complementary Principle of Evaporation

Abstract

The complementary principle, which was first proposed by Bouchet (1963), illustrates a complementary relationship among the actual evaporation, the potential evaporation, and the apparent potential evaporation. It has generated increasing attention for estimating evaporation by using only routinely observed meteorological variables (radiation, wind speed, air temperature, and humidity) without complex surface property parameters. However, this principle still poses great challenges because of the underlying uncertainties in estimating its critical parameter, namely, asymmetric parameter b. In this study, we adopted a sigmoid generalized complementary function and utilized the eddy covariance (EC) data from 217 sites around the world to determine b values in different ecosystems and their correlation with environmental factors. We found b has a mean value of 6.01 ± 0.08. The asymmetric parameter b is small in dry regions (i.e., the desert ecosystem, 0.42 ± 0.02) and increases as the land surface wetness improves. The ecosystem mean air temperature and vapor pressure deficit have negative correlations with b (Pearson correlation coefficients are −0.57 and −0.52, respectively), and the mean soil water content has a positive correlation with b (0.69). Besides, the sigmoid function has a favorable capability in estimating evaporation no matter based on the site-specific b values or the ecosystem mean b values. The ecosystem mean b values given in the current study also perform acceptably in the independent verifications, indicating these values can be applied extendedly for regional and global studies.

A review of the complementary principle of evaporation: from the original linear relationship to generalized nonlinear functions

“…Given that the third stage is uncommon, the polynomial B2015 performs well with calibrated parameters (Brutsaert et al, 2017;Han and Tian, 2018a;Liu et al, 2016;Zhang et al, 2017). However, observed points are located in the domain out of OMP at several flux sites, and the sigmoid H2017 shows the best performance in estimating evaporation as validated by using data from FLUXNET (Han and Tian, 2018a;Wang et al, 2019).…”

“…Thus, the constant α = 1.26 was suggested with acceptable weakening of the accuracy of E estimation (Han et al, 2012;Han and Tian, 2018a). In practice, α was also determined from the observed E values in wet condition when E was close to E Pen and/or E PT (Kahler and Brutsaert, 2006;Ma et al, 2015a;Wang et al, 2019). A novel method of using observed air temperature and humidity data in wet environments when measured E is lacking was proposed by Szilagyi et al (2017) and was successfully used for large-scale CR model applications .…”

Complementary principle of evaporation: From original linear relationship to generalized nonlinear functions

“…was suggested with acceptable weakening of the accuracy of E estimation (Han and Tian, 2018a;Han et al, 2012). In practice,  was also determined from the observed E values in wet condition when E is close to Pen E and/or PT E (Kahler and Brutsaert, 2006;Ma et al, 2015a;Wang et al, 2019). A novel method by using observed air temperature and humidity data under wet environment 375 was proposed by Szilagyi et al (2017) when measured E is lacking, and was successfully used for large-scale CR model applications (Ma and Szilagyi, 2019;Ma et al, 2019).…”

“…The b in the desert was 380 much smaller than those in the oases or irrigated farmlands (Han et al, , 2012. b was thought to be related to the characteristics of the atmosphere, i.e., the atmospheric humidity (Szilagyi, 2015), the Clausius-Clapeyron relationship between saturation-specific humidity and temperature (Lintner et al, 2015), or the characteristics of the land surface, i.e., the surface temperature (Szilagyi, 2007), the water availability of the evaporating surface (Han and Tian, 2018b;Lhomme and Guilioni, 2010), or the ecosystem types (Wang et al, 2019). Szilagyi (2015) applied a sigmoid function of relative humidity to 385 parameterize b -1 .…”

“…Szilagyi (2015) applied a sigmoid function of relative humidity to 385 parameterize b -1 . Wang et al (2019) used the ecosystem mean b values of 217 sites around the world in the B2017 with litter weakening of the evaporation estimation accuracy. However, the characteristics and determination methods of b need further studies toward a calibration-free evaporation estimation model.…”