Simulation of surface winds in the tropical Pacific during 1979-1988 by anAGCM at physical research laboratory (PRL)

Citation
B. Thomas et al., Simulation of surface winds in the tropical Pacific during 1979-1988 by anAGCM at physical research laboratory (PRL), J METEO JPN, 77(5), 1999, pp. 1083-1108
Citations number
40
Categorie Soggetti
Earth Sciences
Journal title
JOURNAL OF THE METEOROLOGICAL SOCIETY OF JAPAN
ISSN journal
00261165 → ACNP
Volume
77
Issue
5
Year of publication
1999
Pages
1083 - 1108
Database
ISI
SICI code
0026-1165(199910)77:5<1083:SOSWIT>2.0.ZU;2-H
Abstract
The annual cycle and the large-scale low-frequency variability of surface w inds over the tropical Pacific simulated by an atmospheric general circulat ion model (AGCM) are examined. The AGCM was integrated for 10 years with th e observed sea surface temperatures (SSTs) from 1979 to 1988 and the model outputs were compared with National Center for Environmental Prediction (NC EP) reanalysis. The AGCM successfully simulates the climatological features associated with the annual mean and annual cycle of the surface winds. The amplitudes of t he simulated annual harmonics are nearly comparable with observations over the south Pacific convergence zone (SPCZ). On the interannual time scale, t he AGCM realistically captures two warm events (1982-83 and 1986-87), and o ne cold event (1988). The first two empirical orthogonal functions (EOFs) a nd their principal components (PCs) of simulated wind anomalies reproduce t he important features, which are crucial for coupled air-sea interaction, l ike westerly anomalies associated with warm events, eastward propagation of westerly anomalies during 1982-83 and realistic convergence pattern in the first EOF mode over tropical eastern Pacific. Also, the simulated wind ano malies do not exhibit eastward migration during 1986-87 and the strength of anomalies is much larger during 1982-83 than 1986-87. Thus, the AGCM succe eds in reproducing the distinct characteristics of different warm events. I t is found that the PCs of wind anomalies are strongly correlated to the PC s of precipitation anomalies. Also, the correlation between PCs of observed and modeled precipitation anomalies is very high (0.93 for both PC1 and PC 2) at zero lag. Therefore, it appears that the AGCM's capability to represe nt the large-scale low-frequency variability of convective heating realisti cally contributes to its success in simulating the large-scale low-frequenc y part of surface winds. The most apparent deficiencies in the AGCM simulation include the incorrect location of westerly maximum, and less westward extension of westerly anom alies in EOF1. Also, the mean south-westerlies over Peruvian coast are weak er and not extended up to 170 degrees W, especially during January. To get more insight into the AGCM simulated winds, the response of a steady-state linear Matsuno-Gill type model to two type of forcing (AGCM simulated preci pitation anomalies, and AGCM simulated precipitation anomalies + SST anomal ies) is compared with AGCM simulated winds. The linear model, forced only b y the AGCM precipitation anomalies, can reproduce the AGCM simulated surfac e winds. The results of the linear model, when it is forced by both AGCM pr ecipitation anomalies and SST anomalies, suggest that the deficiencies of A GCM in the western Pacific are mainly related to the insufficient treatment of surface turbulent fluxes.