This study evaluates simulations of the East Asian winter monsoon in e
ight GCMs that participated in the Atmospheric Model Intercomparison P
roject (AMIP). In addition to validating the mean state of the winter
monsoon, the cold surge and its transient properties, which includes t
he frequency, intensity, preferred propagation tracks, and the evoluti
on patterns of the surges, are examined. GCM simulated temporal distri
bution of the Siberian high and cold surges is also discussed.:Finally
, the forcing of the cold surges on the tropical surface wind and conv
ection, along with their interannual variation is analysed. The mean s
tate of the winter monsoon is generally portrayed well in most of the
models. These include the climatological position of the Siberian high
, the 200 hPa divergent center, and the large-scale wind patterns at t
he surface and the 200 hPa. Models display a wide range of skill in si
mulating the cold surge and its transient properties. In some of the m
odels, the simulated cold surge trajectory, intensity, frequency, prop
agation patterns and source regions are in general agreement with thos
e from the observed. While in others, the models cannot adequately cap
ture these observed characteristics. The temporal distribution of the
Siberian high and cold surges were realistically reproduced in most GC
Ms. Most models were able to simulate the effect of the cold surges on
the tropical surface wind, although a few models unrealistically gene
rated subtropical southerly wine! in the mid-winter. The relationship
between cold surges and the tropical convection was not satisfactorily
simulated in most models. The common discrepancies in the winter mons
oon simulation can be attributed to many factors. In some models: the
reason is directly related to the improper location of the large-scale
convective center near the western Pacific. The satisfactory simulati
ons of the monsoon circulation and the cold surges are partly due to t
he topographical characteristics of the East Asian continent, i.e., th
e Tibetan Plateau to the west and the oceans to the east. The correct
simulation of the interannual variation of the surface wind near the S
outh China Sea (SCS) and the maritime continent is a demanding task fo
r most of the models. This will require adequate simulations of many a
spects, including tropical convection, the Siberian cold dome, the ext
ratropical-tropical linkage, and the air-sea interaction. The discrepa
ncies noted here furnish a guide for the continuing improvement of the
winter monsoon simulations. Improved simulations will lead to an adeq
uate delineation of the surface wind and convection near the maritime
continent, which is essential for portraying the winter monsoon forcin
g in a coupled model.