Complementary relationship with a convective boundary layer model to estimate regional evaporation

Citation
M. Sugita et al., Complementary relationship with a convective boundary layer model to estimate regional evaporation, WATER RES R, 37(2), 2001, pp. 353-365
Citations number
27
Language
INGLESE
art.tipo
Article
Categorie Soggetti
Environment/Ecology,"Civil Engineering
Journal title
WATER RESOURCES RESEARCH
ISSN journal
0043-1397 → ACNP
Volume
37
Issue
2
Year of publication
2001
Pages
353 - 365
Database
ISI
SICI code
0043-1397(200102)37:2<353:CRWACB>2.0.ZU;2-L
Abstract
An idea to use a simple convective boundary layer (CBL) model in the comple mentary relationship to estimate regional evaporation was explored. The CBL model simulated the potential specific humidity deficit D in CBL when the bulk stomatal resistance r(st) = 0. This value of D was then used in the Pe nman-type equation to derive evaporation E-po that would occur with ample s oil moisture under the prevailing weather condition. The same equation was also used to produce the potential evaporation E with the actual humidity d eficit, and these E-p and E-po values allowed the evaluation of the actual evaporation E by applying the complementary relationship E = etaE(po) - E-p ,, where eta is assumed as 2.0. This was tested with the data set obtained in Hexi Corn desert area in northwestern China with a modified version of a simple CBL model developed by Lhomme [1997]. It was found that the method produced better estimates of the daytime mean E values with smaller bias th an those obtained from a conventional application of the complementary rela tionship without the CBL model. Also, it was shown that the assumption of e ta = 2.0 in the complementary relationship was only approximate in most cas es. To take this into account, an additional procedure was explored in whic h eta was treated as a variable, and an iteration process with the CBL mode l determined the final 77 and E values. It was found that this process prod uced E values that have smaller systematic error and agree better with the measurements on average, but the unsystematic error got worse than that fou nd with 77 = 2.0, probably because of use of the CBL model with r(st) not e qual 0 in the iteration process.