SPARC DynVar Status & Plans Paul Kushner University of Toronto Background • The SPARC DynVar project was initiated at the 2006 SPARC SSG meeting. • The project’s broad aim is to understand the role of the stratosphere in tropospheric climate, using GCMs as the principal tools. • The project is described in a planning paper in the SPARC Newsletter, July 2007. • This is the DynVar project’s organizing group. I’ve highlighted the names of the project’s coordinators. Project Aims • We ask how GCMs’ stratospheric representation affects tropospheric climate, variability, and climate responses (high-top vs. low-top). • We focus on circulation & dynamics (complementing CCMVal and SOLARIS) • We center on comprehensive GCMs, but include a component on theory and simpliﬁed models. • We want to document the costs and beneﬁts of including a realistic stratosphere in an Earth Systems Model. Analysis Areas & their Coordinators The project is roughly divided into 4 analysis areas: A. DynVar Top (Giorgetta, Sassi): Effect of model stratospheric representation on the circulation and variability of the climate system. B. DynVar Intraseasonal (Perlwitz): Intraseasonal strat- trop coupling, with a view to seasonal prediction. C. DynVar Climate Change (Manzini): Stratospheric control of response to climate forcings. D. DynVar Ideal (Polvani): Cross-cutting theme, analysis of stratospheric inﬂuence using theoretical and simple-modelling approaches. Simulations • We want to set out coordinated experiments and analyses. Interested modeling groups include CCCma (Canada), NCAR, GFDL, UKMO, MPI. • We want to elucidate the role of the stratosphere for various degrees of coupling to the land-ocean surface: • Prescribed SST AGCM, Mixed-layer ocean/ AGCM, Dynamical ocean/AGCM, • Forced with standard climate forcings. • Theoretical component: explain the GCM results with simple models and theory. Overview of Activities • Our main activies for this year: • September 2007 Santorini Chapman Conference. • March 2008 planning workshop. • Participation in Ozone Forcing Dataset workshop in Bologna September 2008. • Main activities for coming year: • Setup of DynVar server • Publication of plan • Starting coordinated runs. Emerging Research Results Relevant to DynVar • There were several papers over the last year (Santorini, March DynVar Workshop, Bologna) that were connected in one way or another to DynVar themes. • Most directly relevant to DynVar were papers that emphasized a systematic comparison between high top and low top models. • This is an area where I think the project can make progress in the short term. Hi-Top/Low-Top Intercomparison: ECHAM Models Hi Top Minus Low Top: Temperature (OAGCM) (AMIP) Hi Top Minus Low Top: w* (OAGCM) Giorgetta et al. Circulation Response to Global Warming U SLP The Hi-Top Canadian GCM (CMAM) gets a tropospheric NAM response to CMAM 2XCO2. Hi-Top 10-3 hPa Lid The Low-Top Canadian GCM (Standard) does not. Standard Low-Top But a Low-Top version of CMAM also 3 hPa Lid gets the NAM response. Orographic GWD (OGWD) settings CMAM turn out to control the response. Low-Top 10 hPa Lid Standard Low-Top, CMAM OGWD Sigmond et al. 2008 Circulation Response to Global Warming U SLP If the CMAM GWD settings are put in the Standard GCM, it gets a NAM CMAM response too. Hi-Top 10-3 hPa Lid The OGWD controls lower stratospheric winds; stratospheric Standard resolution is secondary here. Low-Top 3 hPa Lid CMAM Low-Top 10 hPa Lid Standard Low-Top, CMAM OGWD Sigmond et al. 2008 Control of Tropospheric Response to Stratospheric Cooling Hi-Top and Low-Top CMAM get a U SLP tropospheric SAM+ response to Antarctic stratospheric cooling. CMAM Hi-Top But non-orographic GWD can control the response. CMAM The bottom row shows the impact Low-Top of using a non-conservative boundary condition on the GWD. CMAM Low-top models might represent Low-Top; SAM sensitivity accurately, but you Flux GWs have to be careful with to Space! implementation details. Shaw, Sigmond, Shepherd & Scinocca 2008 ENSO Teleconnections Hi Top Low Top Obs Cagnazzo & Manzini DynVar Planning Workshop • The workshop aimed to plan DynVar activities for the coming 2-3 years. • It consisted of a day of presentations and discussions on the project’s research themes, and a second day of working meetings on project plans. • The workshop was open to all DynVar organizers and participants, and the DynVar project itself is open to all interested researchers. • The workshop was held in conjunction with the GWD Initiative workshop. We met in plenary for part of the meeting, to coordinate on diagnostics. DynVar Planning Workshop Aims of the planning component of the workshop: • Summarize the science goals of DynVar and Analysis Area subprojects. (Completed) • Specify in detail which simulations will be run for the intercomparison (multiple high top and low top models). (Partially completed.) • Specify in detail the diagnostic requirements for the subprojects. (Partially completed.) • Outline a data access policy and distribution approach for model output. (Completed.) • Tentatively outline a plan for dissemination of the results. (Not started --- too soon.) DynVar Planning Workshop • To prepare for the planning meeting, we distributed a survey to gauge interest, projects, potential contributions from modelling centres. • The survey results provide input into the diagnostic requirements for the project. Lecture by Alan Plumb Thanks SPARC IPO! Science Content of the Workshop • We attracted some new people from outside the “usual suspects”. E.g. participants like David Straus from COLA/ IGES. • The idea to meet jointly with the GWD workshop was successful overall. • Appropriate diagnostics identiﬁed from global GCMs to be used by the GWD project. Planning Session • D. Waugh: overview of CCMVal plans & data policy. • F. Sassi: outline of DynVar simulations. The real question is who wants to commit to these runs? • J. Perlwitz: outline of survey results. Looks like about 20 or so research projects could use the data. • P. Kushner: Data policy. Data Policy and Server • DynVar will follow the CCMVal two-phase data policy and data formats (netcdf/CFMIP). • Data contributed to the projected will be served from the server dynvar that is being installed at U of T this month: • Server with 4TB of backed up storage. • Users will log in and will be able to preprocess the data before downloading (standard utilities installed) • Cost of about CDN$10K on NSERC funds. Simulations • Basic starting points for prescribed SST runs is CCMVal Ref 1 (prescribed radiative forcings). • CCMVal runs have occupied the groups fully this year. A clear commitment for any groups to do runs speciﬁcally for DynVar has not been actively solicited or provided. • An unresolved question is how to treat ozone forcing in such runs. Let’s focus on this issue. AM Signatures: Ozone Depletion Trend, Projection of GCM Response to 1968-1998 Trend onto SAM Ozone Depletion Z500 Surface winds and temperatures Thompson & Solomon 2002, Gillett and Thompson 2003 • Southern Hemisphere trends are strongly SAM related. • Photochemical ozone loss and recovery control a lot of this (see Son et al., Perlwitz et al.). Prescribed Ozone Forcing Runs • For DynVar models to test the effects of stratospheric resolution on the response to ozone forcing, we want to force a variety of AGCMs with a prescribed ozone forcing dataset. • At the time of the March workshop, we were pointed to the Bodeker et al. effort to establish an ozone forcing dataset. • PJK attended the ozone dataset workshop in Bologna and has proposed that DynVar become an early user of the dataset. • One idea would be to run tests on the new dataset that would look at the basic response to the 1970-2000 period of photochemical ozone loss, as in Gillett and Thompson. Status/Discussion • We have been busy, but PJK was not able to sustain the pace of DynVar past March 2008. PJK’s research group had major technical problems over the summer from which we are still recovering. • DynVar server should be up and running soon, and we can upload some data on it. • Revised project plan will be submitted to Newsletter (ok Diane?) • We have proposed to hold a follow on workshop for April 27-28 2009 in Hamburg. Marco Giorgetta will host. SPARC IPO support is now requested. Status/Discussion • We need to development a ﬁrm commitment from modelling groups to do some runs. But it seems unrealistic to get such a commitment right now with CCMVal being an overwhelming priority for many groups. • Need a clearer path forward for ozone forcing. • Should groups coordinate on coupled ocean/high-top atmosphere model testing?