This is an old revision of the document!
Note: the variables requested for PMIP2 are available on the PMIP2 variables' page
[ PMIP3 Wiki Home ] -
[ Help! ] -
[ Wiki syntax ]
I hope for these new sets of experiments that provision will be made to output some high frequency atmospheric data, i.e. 6 hourly, for selected
representative periods, 30 years would be Ok, but 100 years would be better. This would allow us to explore storm activity and variability
in detail and in particular extremes and to contrast with current and future storm climatologies. Only a selected number of variables would
be required, MSLP, precip., winds at several levels, ideally every 100hPa in the troposphere.
I believe this was a major failing in CMIP3 and to some extent PMIP. If PMIP models for these experiments are to be run at higher resolutions
than was previously the case then it makes storm studies more attractive and the neccessity for the higher frequency data.
Is there any information about how and where data will be archived?
Since PMIP3 is connected with CMIP5, the easiest way to proceed is to define the variable list from the CMIP5 list as soon as it circulates.
I totally agree that we should have these high frequency data. So please remain ready to interact on this point.
We should also have the data needed to force RCM, the RCM groups were interested when I presented CMIP3 at WGCM, and we already had this request for PMIP2.
The plan is for LSCE to take care of the data archiving of all PMIP3 data in France, though we have not chosen yet the actual storage/distribution site (i.e lab or computing center) yet.
PMIP3 data will be a subset of the CMIP5 database, and will be accessible with the same “tools” and the same CMIP5 accounts. We want to store all paleo data in the same location, even if the rest of the CMIP5 will be distributed (due to the expected huge size of CMIP5). This means that we will store only the data really relevant to PMIP3 (and requested by scientists), but it will make it possible to distribute PMIP3 data through alternative means (if necessary). If the software infrastructure chosen for CMIP5 allows it, we will maybe host a mirror of (some) present-day variables, if it seems useful.
the variable list from CMIP5 is circulating for comment and thankfully there are priority 1 request for some 6 hourly data such as
winds on several pressure levels as well as MSLP and a few other veriables. I hope this will provide motivation for outputing these
variables for the PMIP3 simulations. I guess the main question is for how long and what periods/experiments. Ideally I am interested in
all the proposed experiments but I guess because of archiving issues some selection is neccessary. Certainly the pre-industrial, last glacial
maximum and mid-holocene should be priorities. At a minimum representative 30 year periods should be chosen but the longer the better.