The standard statistical atmospheric model that we use in the HCM generates winds based upon the current model's SST anomalies. This seems to work pretty well, but we are exploring the idea that the important wind anomalies may be more directly forced by strong tropical storms than by SST gradients. In this scenario, we would use outgoing longwave radiation (OLR) instead of SST to construct the statistical atmospheric model. However, because the model does not calculate OLR directly, we instead use a stand-in or proxy for OLR, namely a non-linear transfer function of SST. Several researchers have shown that such a function is a good estimate of OLR in the tropical region, which is what we care about.
The advantage of this technique comes in what particular model regions it emphasis. So, for example, consider the leading EOFs of SST anomaly and OLR proxy:
What you see is that they are basically similar, except that the OLR proxy is more concentrated in the central Pacific at the expense of the eastern Pacific. This might be a better match to the ocean model we are using, which does a better job of representing the SST anomalies in the central Pacific than the eastern Pacific. (It is interesting to note that many oceanic general circulation models have this characteristic; presumably it is related to the strong upwelling in the eastern part of the tropical Pacific).
Tests with the OLR proxy version of the HCM are currently underway.
Last modified: 24 July 1997