I\'ve always been a bit chagrined that the land mass on which I currently reside, Tasmania, happens to fall across three different cells in the global model, and consequently isn\'t represented in any of them (I think it covers about 45% of one of them). There may be more than merely parochial interest in this, as recent work (http://www.csiro.au/news/OceanSupergyre.html) indicates that this area may be critical in controlling ocean circulation and therefore heat transport. In this sort of situation, is it possible that the imprecision introduced in approximating the geography could influence the global climate significantly? Would it not be worth a sub-suite of runs in which one of the Tasmanian-vicinity cells (the 45%-covered one) is flipped from \'ocean\' to \'land\', to find out?

---

If you run the SAP model (http://attribution.cpdn.org), you might have more luck with the Tasmanian cells (the SAP model is higher resolution, although shorter duration).

The long duration (and therefore coarse grid) models have code in to try to handle the features which are too small for the grid (i.e., mountain ranges and so forth).

---

I'm a volunteer and my views are my own.
News and Announcements and FAQ

There\'s also a nice article on RealClimate about appropriate scales here.

Thanks very much for both these links, Iain and Mike. Particularly interesting was that one of the RealClimate discussion correspondents asked an almost identical question to mine: \"Do climate scientists expect any surprises as resolution increases? Were there “surprises” between the 1980s GISS models and the latest models? I note that in our region (Australia) your minimum scale map leaves out Bass Strait (between Australia and Tasmania) and Torres Strait (between Australia and PNG). These are significant water ways for local\" (indeed, I would say global) \"climate and ocean currents. They are also about 150 km wide - close to 200 km - so why would they be omitted?\"

to which the response was \"One Japanese model does have a very high spatial resolution, but I don’t think there are any particular surprises.\"

So it would appear that the answer to my original question (which was not so much about replicating Tasmanian climate in a GCM, rather whether the effective omission of Tasmania would significantly affect the GCM outcome) is \'no\'. However, I\'d still like to see some more detailed investigation of this issue (essentially an elaboration of the response above), which I don\'t think the SAP model really addresses.

---

The reason that the resolution isn\'t much higher is that doubling the resolution of the model increases the amount of computer time required by 16 (since all four dimensions have to double). So a 160 year model with 3 seconds per timestep at the current resolution would take 48 seconds per timestep at double resolution (taking 55,000 hours = 6.3 years to run, compared to 5 months).

To combat this, \'regional\' models can be used. These are high resolution in one small part of the world, but coarse resolution elsewhere (mentioned in the link above. For more information, search for \'PRECIS\' and \'hadley\').

For more links related to climate science, see the \'climate science\' README in my signature.

---

I'm a volunteer and my views are my own.
News and Announcements and FAQ