Baumhefner, D., M. Blackmon, E. Kalnay, K. Mo, F. Molteni, J. Roads, P. Sardeshmukh, and S. Tracton, 1991: Summary of a workshop on numerical long-range prediction of 10-90 day time-mean flows. Bull. Amer. Meteor. Soc., 72, 1745-1752.


A workshop on numerical long-range prediction of 10-90-day time-mean atmospheric flow was held at the National Center for Atmospheric Research (NCAR), Mesa Laboratory in Boulder, Colorado, on 4-7 June 1990. It was cosponsored by the Climate and Global Dynamics Division of NCAR and the Climate Monitoring and Diagnostics Laboratory of NOAA. The National Science Foundation provided direct support for 20 graduate-level students. The coorganizers of the workshop were D. Baumhefner of NCAR and M. Blackmon of NOAA. The concept for this workshop developed from several very successful workshops that were held at the National Meteorological Center (NMC) in the past few years. Some of the basic ideas and early experiments on numerical extended-range prediction originated from these meetings, and some of those results are presented in this summary. The current workshop was attended by scientists from operational numerical weather prediction centers, research labs, and universities. The participants represented a broad cross section of interest in the subject, which led to a lively and productive program.

The main purpose of the workshop was to provide an overview of recent progress on the scientific problem of numerical long-range prediction, and it was specifically directed toward the time-mean flows strongly influenced by initial conditions. Workshop goals included assessment of the current status of skill at the extended range, both operationally and experimentally, discussion of the fundamental problems encountered as well as possible new approaches for future research, encouragement of the university research community to become involved in this area of research, and the coordination and planning of future experimental designs. The workshop was organized around the following topics: 1) current forecast skill, 2) prediction of skill, 3) forecast sensitivity, 4) analysis of forecast skill, 5) diagnosis of low-frequency phenomenon, and 6) prediction of low-frequency flow. The structure of each session began with keynote speakers, who reviewed the history, progress, and future of each topic, followed by general presentations. Extensive discussion was encouraged by the informality of the workshop and by providing ample time for each paper. Two formal discussion sessions were held, the first of which dealt with the possibility of a common forecast dataset experiment and archive for research purposes, followed by a debate on the outstanding problems currently facing the research community on this subject.