|
|
THE ITCZ AND TROPIC-WIDE CLIMATE VARIABILITY
A. Axisymmetric experiments.
In an attempt to understand the role of SST in determining the nature of the Inter-Tropical Convergence Zone (ITCZ) a series of GCM integrations have been performed using an ocean covered version of the Unified Model (UM).
The ITCZ appears as a narrow cloudy band on or near the equator in satellite images and reflects the regions of strong low-level convergence and intense precipitation within the tropics.
The strength and position of the ITCZ is influenced by the underlying SST but the exact nature of the relationship is not clear even to the extent that there may be a single ITCZ or two ITCZs either side of the equator.
Figure 1 shows the climatological precipitation for January and it's clear that the rainfall maximum varies in both intensity and position throughout the tropics.
Figure 1: Precipitation climatology for January (mm/day) (taken from the Xie and Arkin climatology)
Investigations of this phenomenon have been performed with forcing from fixed SST distributions which are symmetric about the equator and vary in the latitudinal direction only (figure 2 (a)).
Figure 2: Axisymmetric experiments (a) Sea Surface Temperature (SST) used to force the model (in K) and (b) Zonal mean time mean precipitation for each experiment (mm/day).
Clearly the strength of near-equator SST gradients (red and blue cases) has a large influence on the intensity of tropical precipitation (figure 2(b)). These intense precipitation maxima coincide with the upward branch of a strong Hadley circulation. This has the effect of smoothing out these large temperature gradients in the tropics such that in the mid and upper troposphere there is no sharp peak in temperature that is so prominent in the imposed SST.
In addition, for weak enough SST gradients (green case), a transition from a single strong equatorial ITCZ to a weaker twin ITCZ regime occurs. Although two distinct precipitation peaks appear at 17 degrees N and S it is not strictly speaking an ITCZ feature but instead results from convection being triggered in preferential regions on the equatorward side of mid-latitude systems that extend over warm waters in the sub-tropics. There is still precipitation present near the equator but this is only 4 mm/day and results due to a local radiative convective equilibrium and is influenced little by the large scale circulation.
B. Sea Surface Temperature Anomaly Experiments
A further motivation for using aqua-planet integrations is to identify possible mechanisms for coherent modes of variability within the tropics. An observational study by Ward (1993) has revealed that such a coherent mode exists in northern hemisphere summer on yearly to decadel timescales involving in phase variability of Sahelian and Indian rainfall.
One interpretation of this mode, summarised in figure 3, is that the distribution of western pacific SST leads to changes in the convective activity causing an intensification (suppression) of the local Hadley circulation and through some teleconnection mechanism results in a suppression (intensification) of the Hadley circulation elsewhere in the tropics. This is apparent in the distribution of low-level convergence and wind anomalies seen in the Sahel wet (Sahel dry) phase of the mode shown in (figure 3).
Figure 3: The Tropic Wide Mode (TWM) shown for the Sahel wet/ La Nina phase of the mode (the signs of the anomalies are reversed for the Sahel dry/El Nino phase of the mode).
Experiments to identify the mechanisms involved in the local and remote response to SST have been performed with the addition of an imposed SST anomaly. The anomalies vary in position, size, strength and shape in order to determine the sensitivity in the response of the model. The changes in precipitation due to forcing from an SST anomaly are quite marked.
Figure 4 shows the precipitation difference between an experiment with a 3K confined SST anomaly and the control experiment (blue line in figure 1). Near to the SST anomaly maximum at 95 degrees E there is a substantial increase in precipitation (blue areas indicate increases in excess of 2 mm/day above the control experiment) of some 200% compared to the control experiment. Of note is the position of the precipitation maximum which is not exactly over the SST anomaly maximum but some 5 degrees to the west. The red regions in figure 4 indicate areas of rainfall suppression in excess of 2 mm/day so it is immediately apparent that the SST anomaly is able to influences most of the tropics through the suppression of rainfall.
Figure 4: Precipitation difference (in mm/day) between an experiment with a 3K SST anomaly (centred on the equator at the dashed line) and the control experiment (blue line in figure 1). Blue areas indicate regions with precipitation increases greater than 2 mm/day and red areas indicate regions with precipitation decreases greater than 2 mm/day (contour interval 5 mm/day positive values (blue regions) and 1 mm/day negative values (red regions)).
Analysis of the circulation in the presence of an SST anomaly shows a large intensification of the local Hadley circulation near the longitude of the increased precipitation and an associated decrease elsewhere giving capping of convection and suppressed precipitation.
Other work using the aqua-planet has looked into the transient tropical wave activity seen in the model fields and how these vary in the presence of a SST anomaly of different shapes and sizes. Figure 5 shows a hovmoller plot of precipitation again from an experiment with a 3K confined anomaly centred at 95 degrees E.
Figure 5: Hovmoller plot of precipitation averaged between 5N to 5S from an aqua-planet experiment with a 3K confined SST anomaly centred on the equator at 95E. The more intense precipitation is given by red and orange colours and the weaker precipitation is given by blue colours. The scale runs from 0 to 80 mm/day. Time runs from bottom to top and left to right.
The influence of the SST anomaly can clearly be seen as a region of higher precipitation which is quasi steady over the anomaly. Transient wave signals are evident as successive propogating strong and weak precipitation bands, the strongest of which propagates west to east and can be classified as a Kelvin wave type disturbance with a phase speed of about 18 m/s. A weaker Rossby wave signal can be seen propagating east to west away from the anomaly, but with a slower phase speed.
References
Ward, M.N., 1993: Tropical north African rainfall and worldwide monthly to multi-decadel climate variations. Ph.D. Thesis, University of Reading, U.K.
Return to DYNAMICAL PROCESSES home page.
This page is maintained by R.B.Neale@reading.ac.uk
|
|
