First Planning Workshop for the next US Global Reanalysis

June 5-6, 2000

University of Maryland, Department of Meteorology, Computer and Space Sciences Building, College Park, MD 20742
Room 2428
Map, Directions

Organizers: Eugenia Kalnay, Steve Lord, Siegfrid Schubert

Invitation Letter:
Dear colleagues: We would like to invite you to attend a Planning Workshop for the next US Global Reanalysis. It will take place on June 5 and 6 at the University of Maryland, in College Park. If you have not done so, please let Hazel McLean and Eugenia Kalnay know whether you are planning to come, or if you would like to suggest other invitees. If you would like to make brief presentation, please check the agenda below and let us know in which session and the topic. If you cannot come we would be grateful to receive your comments and suggestions and present them at the workshop for discussion. The idea of the Planning Workshop was triggered by an inquiry by Jim Hurrell on current reanalysis plans. He indicated that U.S. CLIVAR SSC is preparing a 3 to 5 years forward-looking set of recommendations to supply to NOAA, NSF, NASA, and DOE. At least some of these agencies have sought this input with the intent of using the information to help shape their future funding plans, and he and Michele Rienecker have the task of addressing the needs for the support of continuing, global atmospheric reanalysis. This requires estimating the cost (dollars, people, resources) of continuing global atmospheric reanalyses and flesh out the reasons for doing this. As you know, there have been several successful global reanalyses (NCEP/NCAR 50+ years, ERA-15 year, NASA/DAO 15 years, NCEP/DOE 20+). Now ECMWF is preparing to start their ERA-40 year reanalysis, and NCEP is preparing for a 20+ Regional Reanalysis over North America. As the methods and models used in data assimilation improve, it becomes desirable to create a more advanced global Reanalysis. For example, the NCEP/NCAR Reanalysis 1 was carried out with the system that was operational in 1995 and many improvements have been implemented at NCEP since then. The NCEP/DOE Reanalysis 2 has corrected several errors and incorporated better forcing of the soil moisture from observed precipitation and other improvements, but otherwise uses the same system as the first reanalysis. NASA/DAO has made very substantial improvements in their system as well. Since it takes about 4 years to prepare for a major reanalysis, we should start planning now. We apologize if you have not received the previous email announcing the workshop. Because this workshop has been informally organized, we have not secured any travel funds. As a result there is a stronger DC participation than would be desirable. Again, if you cannot come to the workshop we would appreciate very much if you would email us your thoughts and suggestions.

ISSUES IDENTIFIED On April 12 2000, S. Schubert (NASA), E. Kalnay (U.Md), S. Lord, M. Kanamitsu, R. Kistler, F. Mesinger and G. White (NCEP) met at NCEP and agreed that it was a good idea to have a workshop. In addition to the estimates of resources, etc, several issues were identified (notes courtesy of GW):

• Should we do a US (multi-agency) global reanalysis in the US? Can we take advantage and leapfrog the ECMWF 40 year Reanalysis (ERA-40), which in turn will leapfrog and take advantage of the previous NCEP reanalyses 1 and 2 and the NASA reanalysis? Such alternation would be good to provide the community with state-of-the-art reanalysis.
• What would be the role of the agencies? Some ideas discussed include performing a single (or an ensemble) US Global Reanalysis, by, for example, NCEP and DAO joining forces. Perhaps DoE can provide computational support, as they provided for the NCEP R2, and technical support. Would the Navy be interested in participating in this project?
• Should we use the "US global model infrastructure" being developed?
• The problem that has been identified by the WCRP as the most important in the reanalyses is the introduction of spurious trends due to changes in the observing system. Can this problem be ameliorated in future reanalyses so that the Reanalysis serves the inter-decadal variability community as well?
• How do we ensure that there is stable funding for Reanalysis? Many people think that it would be ideal if the National Weather Service adopted reanalysis as part of their operational services to climate. This would ensure continuous development of the global reanalysis system, rather than switching the concentration completely to the regional reanalysis as it is happening now at NCEP.
• Can we take advantage of the first reanalysis to better estimate the forecast error covariance (we originally assumed it was constant, whereas before 1979 this led to a suboptimal use of the fewer available observations). Can we use the first reanalysis to improve substantially the QC, and ensure that accurate isolated observations are given maximum weight?
• Should we use radiances? Probably yes. How do we ensure that we do not have a problem like the one identified by Mike Fiorino and Kevin Trenberth that introduced spurious jumps in the climate through bias corrections in satellite radiances in the ERA-15?
• Can we introduce new techniques such as cloud and precipitation initialization such as the one Arthur Hou has been developing at NASA and Russ Treadon at NCEP?
• Should we go for high resolution or for lower resolution and 4D-VAR data assimilation, or Kalman smoothing?

June 5, 2000
8:30- Coffee
9am-10:30 Individual presentations on past experience and lessons learned. (Schubert, Kistler, Kanamitsu, Fiorino, Jenne?, Woollen?, ... ) 10:30-10:45 Break
10:45-12:00 Requirements for a new system (Kalnay, Hurrell, Mehta?, Shukla?Ú)
12:00 -Lunch at Adele's (Maryland cuisine) or Seven Seas (chinese/japanese)
1:30-2:30 Discussion of the systems planned to be ready for the 3-5 year time frame (Lord, Atlas, Ú).
2:30-3:30 Position of the funding agencies (NOAA/NWS, NOAA/OGP, NASA, NSF, Navy, DOE)
3:30-3:45 Break
3:45-5:00 Brief presentations and discussion

June 6 2000
8:30- Coffee
9:00-10:30 Discussion of a consensus roadmap for the next Reanalysis (role of agencies, model, data assimilation, observations, schedule, necessary resources, funding, etc.)
10:30-12:00 Preliminary draft of the roadmap.

MEETING REPORT

Minutes provided by Jim Kinter and Siegfrid Schubert, with additions from David Legler and Jim Hurrell.

There was significant interest and motivation to develop a US Global Reanalysis Program. The first step in building interest among the community and the agencies is to develop a report that details the motivation and rationale for such a program. Additionally, the report will provide details on how such a program would be developed, the participants, the beneficiaries, and the projected costs.

"Volunteers" will lead the write-up of this report (see outline at end):
Section 1: Eugenia - ekalnay@metosrv2.umd.edu
Section 2: Siegfried - schubert@dao.gsfc.nasa.gov
Section 3: Chet - chet@iri.ldgo.columbia.edu
Section 4: Jim K. - kinter@cola.iges.org
Section 5: David L.- legler@usclivar.org
Section 6: Eugenia - ekalnay@metosrv2.umd.edu
Section 7: Siegfried- schubert@dao.gsfc.nasa.gov