The South Pacific Convergence Zone (SPCZ) is a northwest to southeast oriented region of heavy rainfall and a location where intense Tropical-Extratropical interactions take place in the atmosphere. The sub-tropical to mid-latitude portions of the SPCZ (south of 20„aS and denoted by SM SPCZ) is known to be difficult to faithfully simulate in models, commonly appearing too zonal overall. An attempt has been made to get a clearer view of processes which may be important in maintaining the SPCZ using a combination of satellite based observations. The work is intended as a prelude for constructing a dynamical model, but the results seem interesting in themselves, especially when compared with NCEP/NCAR reanalysis results. The data sets used are QuikSCAT scatterometer estimated low-level (~10 m) winds, TRMM satellite microwave radiometer estimated SST's, and the Global Precipitation Climatology Project rain-rate estimates, which are based on a blend of microwave and infrared brightness temperatures. The period of analysis was Sept 1998 - Aug 2001, a period where La Nina or neutral conditions existed in the Pacific. Significant findings to date are:

(1) Low-level convergence in the tropics is spatially co-located with warm ( „d 25„aC) SSTs throughout the basin. It is also co-located with NCEP reanalysis-estimated rain rates. This suggests the algorithm for the reanalysis rainfall estimation may be too heavily biased toward SST's. Climate models using similar algorithms may have similar problems with rainfall representation.

(2) Averaged over the entire data period, near surface convergence is spatially co-located with the axis of maximum rainfall in the tropical portion of the SPCZ (and throughout the ITCZ), but not in the SM SPCZ. In this region, convergence occurred equatorward and eastward of the heaviest rainfall.

(3) Seasonal analysis revealed a similar near surface convergence pattern over the SM SPCZ during Austral Spring and Fall. However, the pattern changes with height such that convergence is co-located with the axis of maximum rainfall at the 850 and 700 HPa levels during these same seasons. This suggests a predominance of stratiform over deep convective rainfall in this region, with corresponding implications for the vertical distribution of diabatic heating.

(4) Lag correlation analyses of rainfall show only weak evidence of convective events propagating from the tropical to the subtropical and mid-latitude portions of the SPCZ. On the other hand, there is some evidence of propagation of convection from the sub-tropical to mid-latitude portions of the SPCZ. Composites of reanalysis height anomalies, at various lags, based on high rainfall events SM SPCZ show well developed wave-trains at 50 hPa and 20 hPa levels emanating from midlatitudes.

Taken together, the results suggest convection in the tropical SPCZ may be, like the Northern Hemisphere ITCZ, to a large extent dependent upon SST distribution. But the SM SPCZ appears to be largely due to the interaction between passing mid-latitude fronts with a moisture field "preconditioned" by the South Pacific High, an idea held by Kiladis et al (1992).