CMMAP LowCloudsTheme UpdateFrom Bjorn StevensDate September 21, 2007 Dear All, in preparation for the CMMAP advisory committee meeting on Monday I called a number of you to see what you have been doing, how it fits with our strategy and how to move things forward. If we are going to continue meeting like this, it is imperative that we begin gelling as a group by the next time we meet at UCLA in January, and so this round of calls has been beneficial in assessing where we are at and how we work more effectively together. Here are some thoughts that developed as a result of these exchanges. On Low Cloud Feedbacks:
addressed by both Virginia and Washington. Similar compositing strategies are being followed by both UCLA and Washington, where UCLA is teaming up with Durango to do in idealized contexts what is being explored by Washington in more realistic geometries. A lot of what is being done, or which people are committing to have done in the coming months fits well together and nicely within our strategy, but there is a lack of coordination. Suggestions: Cara Lyn and Brian try to set aside a week (in the next two or three) to work in a coordinated fashion to bring the question of the utility of the SCM framework for interpreting CAM to the next level. This will be followed by a visit of Cara Lyn to CSU (where Brian is sitting at the moment) in late October early November to try and arise at concrete results on these questions, so that this direction of our effort can begin looking more concretely at the question of parameterization choices in physically relevant contexts. To the extent we find that extending these forcings, and looking at them with Washington would be beneficial. Toward this end having Brian visit Matt in early December, or vice versa seems like it would be useful. Again to work together for a week. Finally, Washington and Virginia should try to find ways to coordinate their efforts (UCLA, by way of me has a minor interest in this as well) in looking at resolution effects of CAM-SP and their impacts on the representation of low clouds. A number of us continue to work on tool development. Here I would mention efforts at UCLA to develop the ability of the UCLA-LES to efficiently use thousands of processors, rather than tens. The development, evaluation and testing of microphsyical models for LES by the UW, UCLA and NCAR groups (how and should we work more closely together on these issues?) the development and evaluation of SGS models more appropriate for coarse resolution by UW and Virginia groups, as well as ongoing parameterization development for conventional GCMs. Finally Robert Pincus and I have been working to develop strategies for representing solar radiative transfer in ways appropriate for LES/CRM scales. This is being done in the context of the UCLA-LES and the FuLiou radiative transfer scheme (with McICA), but (if it works) can (and will) be extended to SAM and other radiative transfer models. Lastly Robert Pincus and Johh Helley have begun an initiative that will try to develop a infrastructure to support testing SAM variants on the GCSS cases. This work probably should be coordinated with the Virginia group who has been doing this with SAM for some time. Some things are not being actively pursued: (1) Talking to Matt I had the impression that UW was still inclined to pursue the big-brother approach with SAM that they suggested at the Kauai meeting, but this is a bit stalled. Is this something that is going to happen? (2) The physics-grid approach I was advocating is stalled in the context of CMMAP, but recently got a breath of fresh-air by DOE, and is only waiting on securing the personnel. I will keep people updated. Okay, those are my impressions, and suggestions. Comments are of course very welcome. Bjorn From Bruce Wielicki Date September 24, 2007 Chris and Bjorn: This sounds like a good overall direction. it is becoming more and more clear that the boundary layer problems are likely to dominate climate change issues for some time. I remain concerned that our "truth" fields for boundary layer t,q,wind vertical structure remains either field experiment (too few cases), or 4-D assimilation (poor accuracy). Sounds to me like we should be advocating for some group to take on a Climate 4-D Assimilation. The idea here would be for weather time scale assimilation to focus not on the traditional 500hPa height and other baroclinic storm metrics but on boundary layer structure metrics. This would like lead to some very different focus in the 4-D assimilation methods, data inputs, physics focus, model vertical resolution in the boundary layer, etc. As long as our 4-D assimilation methods focus on severe convection and baroclinic storms, we are crippled in testing cloud models in even a large ensemble weather prediction mode. CMMAP clearly does not have the resources to attack such a problem, but unless we clearly pose the shortcoming, the assimilation folks (EC, GMAO, or NCEP) clearly will not address it any time soon. Alan Betts was at the NEWS (NASA Energy and Water System) meeting last week and we were discussing his difficulties in constraining the combined land water/energy cycle because of related boundary layer issues that link cloud energy/evapotransporation in the land boundary layer. This may be one of that groups focii for diurnal to monthly to seasonal time scales (i.e. short time scales). Alan has worked quite a bit on this issue with ECMWF and is trying to get a better picture of combined land/boundary layer physics. He is finding large differences (factor of 2) in the modeling links of surface boundary layer moisture and energy fluxes. The polar regions are also an obvious boundary layer challenge for ice, cloud, and upper ocean. I know that MMF in its early stages has more capability to help convection, MJO, etc, but I'm glad to see a strong focus on attacking the boundary layer cloud issues. cheers bruce From Anning Cheng Date September 24, 2007 HI, Dear All: We have a few things to add about what we are doing now and will do in the next few months 1. Activities in Virginia We started to simulate other GCSS cases besides RICO using 3D UCLA-LES, 2D SAM and 1D IPHOC. We probably have all the major GCSS cases done before January. This part of work will focus on the resolution, domain size, dimensionality and other physical processes. It is surprising that the results depend highly on domain-size when grid-size is larger than 1 km for RICO, but not for DYCOM. We suspected that the interaction between unresolved and resolved scales may be an important issue. We also started to implement IPHOC into the 3D VV-CRM as a future tool to study the boundary layer and deep convective clouds in MMF context. The improvement to turbulence closure will primarily be conducted in the context of this new MMF although we have used the SAM for the preliminary tests, funded by the ARM project. 2. Relations to the group activities We have close cooperation with UCLA group about the sensitivity tests of the GCSS cases and will cooperate with Robert Pincus and John Helley on developing the infrastructure to support testing SAM on the GCSS cases. We are also glad to submit our results to document the SCM cases in the CMMAP wiki. Thanks for the good discussions. Anning and Kuan-Man |