(back to up to ScienceWorkRobin)
Orography/Bathymetry issues
Use of an envelope orography (e.g. Slingo and Pearson, QJRMS 1987) in low resolution atmosphere models has been shown to improve variability and blocking in a similar manner to parameterisations of gravity-wave drag. Preliminary attempts at in FAMOUS this also seem to improve the low latitude warm bias (esp. with the new ozone), and possibly help the high latitude cold one. Applying the envelope “algorithm” over Antarctica clearly makes the temperature worse and makes the orography look worse too so it’s not applied here - to a certain extent this is true of high northern latitudes too, although how to shade between areas where it is applied and those where it isn’t is not clear. The effect on atmospheric variability does not seem to be as expected - EKE diagnostics show a mixed bag of increases and decreases at different locations, but explicit N.Atlantic stormtrack/blocking plots show decreases in both, which is the opposite from what was expected. Storm track activity is too low anyway (common in GCMs, but especially so here), so this is bad news. As a result, we didn’t get any further with the use of envelope orography.
Original adtan FAMOUS used an extra-smoothed orography to try to reduce model instabilities that were eventually dealt with through the polar row homogenisation (I think). The original, less smoothed N24 orography has been reintroduced to FAMOUS now it’s slightly peakier, and helps both surface temperatures and midlatitude variablility a bit.
ancil: qparm.orog.4837 - release dump contains the new land already
Even with the removal of Iceland by Jones et al. the Atlantic MOC does not penetrate as far north in FAMOUS as in HadCM3. The bathymetry between Greenland and Scotland was
dug out in HadCM3 as part of tuning the AMOC and deep water formation, but this wasn’t done in FAMOUS. Trying this in FAMOUS does allow the MOC to push further north. However, it also initially strengthens the AMOC to unrealistic values whilst doing little to melt the excessive ice or raise SSTs except very locally to the deeper channel cut for it. We thus haven’t used this idea, since FAMOUS will be largely used for paleoclimate and other weirdnesses - deepening these channels is a pain that would have to be hand-done for every new landmask, and a bit too specific to modern climate.
The Southern Ocean in FAMOUS is generally too warm, there’s not very much Antarctic Bottom Water and the ACC’s too weak. It was suggested that drag at Drake Passage on the coarse FAMOUS grid might be responsible for the weak ACC, which might be correlated with the other two problems. Enlarging the DP didn’t help the ACC strength particularly, but one of the runs did produce some very odd overturning streamfunction plots - we don’t really understand the effects here, but they’re not immediately helpful. This idea has been shelved for the moment.
