date: Fri Oct 1 17:11:40 1999 from: Tim Osborn subject: COAPEC ideas to: m.kelly@uea Hi Mick As I said, I have two ideas for a NERC COAPEC proposal. One was to be a bigger project involving people in ENV, but it now looks as if they will be putting in a couple of different proposals and therefore won't be going along with my idea. I'm now keener to pursue the second idea, and to get you interested in it. So here's a brief outline. [Not so brief as it turned out!] Motivation One of the main questions to be answered in understanding seasonal to multiannual forecasting, the possibility of coupled atmosphere-ocean circulation on interannual to interdecadal timescales, and other problems is: IF, HOW and WHEN do mid-latitude SST anomalies generate an atmospheric response? Is the extra-tropical system a moderately-coupled system, or is it mainly atmospheric variations forcing the oceans with little return influence? This is one of the key motivations for both COAPEC and this proposed project. This question is well studied by using atmospheric models forced with varying SST fields, but is dependent upon the realism of the models used. As they gradually improve - especially at simulating weather systems and their response to external forcing/conditions - the answer should become clear. But at present, different models yield quite different results and the question cannot be answered. Other approaches involve looking at air-sea flux fields to identify which component is driving which, but the answer is uncertain due to large observational error margins in fluxes which get larger for longer records. The proposed project would supplement (NOT replace) these other approaches, by analysing observed and re-analysis SST and atmospheric fields in a novel way. Approach If we correlate SLP fields with SST fields, or composite them on the basis of the extremes of some SST index or SLP index (e.g., the NAO), we find significant correlations. But we cannot prove that the SST anomalies are causing the atmospheric response - the reverse could also be true. Lags or filtering might help, but cannot solve the problem. But what if we were to use daily atmospheric temperature and circulation observations to develop empirical relationships between circulation patterns (air flow characteristics) and near-surface temperature. In the mid-latitudes, the strength of short-timescale synoptic variability means that these relationships pick up temperature variations that are driven almost entirely by atmospheric changes rather than by SST variations. But the atmospheric circulation also varies on longer timescales (i.e, the statistics of the daily circulation patterns vary over months to decades), and these empirical relationships can be used to factor out the influence of such atmospheric changes. Having removed some temperature variability that we know (by the timescale of definition) to be driven by atmospheric variations, the residual temperatures may show much more clearly the influence of oceanic variability. If a composite of high NAO index winters shows a tripole of SST anomalies (cool in Labrador and in the trades, warm to the east of the US), it might turn out that the same composite in the *residual* temperature data set showed only a warm SST anomaly to the east of the US. The implication would be that the other two anomalies could be entirely explained by the anomalous atmospheric circulation, but that the remaining one couldn't and might therefore be driven by the oceans - and might itself have caused the NAO index to be high in the winters used for the composite. This approach has many spin-offs, related to explaining recent climate changes, validating climate models, contributing to downscaling methodologies etc., but these are not within the COAPEC programme. I envisage, therefore, a 9 to 12 month project specifically on the application outlined above, using NCEP daily data (a possible criticism could be raised here, but we have no "real" observations over the oceans on a daily basis). So the big question: how does it grab you? Let me know. Cheers Tim