Searching into the Cause of Southward Displacement of ITCZ in CFS
In a recent presentation, Pingping Xie et al. of NOAA/NWS/Climate
Prediction Center showed that the Climate Forecast System (CFS) free
runs differed from the Global Forecast System (GFS, the
atmospheric component of CFS) AMIP runs during winter and spring with
~10 degree southward displacement of ITCZ. Meanwhile, CFS SST
simulation was found significantly higher than the observation over the southeast Atlantic dry zone
and the southern ocean poleward throughout the year.
Further diagnoses revealed insufficient cloud amount especially over the
southeast Atlantic dry zone could be responsible, where the clouds were
mostly marine stratus with relatively low cloud top and distinct diurnal
cycle of maximum cloudiness observed in the early morning. Currently, CFS
does air-sea coupling once a day, not
resolving the diurnal cycle. Similar
problems were also found in the subtropical eastern Pacific dry zone.
Figure 1 DJF precipitation climatology.
From left to right are plots of
CMAP merged observation, GFS and CFS
simulations, respectively. The figures show distinct differences in
amplitude and southward displacement of the Atlantic ITCZ in
Figure 2 Mean annual cycle of Atlantic ITCZ (time-longitude
section of 30 -20W mean precipitation). Substantially strong amplitude
is found in
both GFS and CFS simulations. CFS simulation also shows a false double ITCZ
during spring and early summer time.
Figure 3 CFS simulation bias of DJF SST
climatology. It clearly shows the warm bias of the
tropical-subtropical eastern south Atlantic SST in association with
anomalous northwesterly winds above.
Figure 4 DJF total cloud amount observed by
(left) and simulated by GSF (middle) and CFS (right). Overall,
both simulations generate less clouds compared to the observation.
Over the tropical-subtropical southeast Atlantic dry zone, the
simulated cloud amount is particularly insufficient.
Figure 5 (a)Jan 1990 ISCCP cloudiness. The upper panel
shows the daily mean and the lower panel the time-longitude plot of
17-20N mean. A bi-polar structure seen over the dry
zone and the nearby continent could be caused by the sea breeze,
which presents a diurnal
Jan 1990 CDAS local circulation with daily mean removed (b) at 00Z
and (c) at
18Z. The upper panels show 700 hPa vertical velocity and the
lower panels the height-longitude section of 15-17S mean.
Strong rising motion over ocean and descending motion over vicinal land at 00Z
are reversed at 18Z.