“In preparing for battle I have always found that plans are useless, but planning is indispensable.” – Dwight D. Eisenhower
The National Research Council (NAS/NRC) is developing a research agenda to advance subseasonal to seasonal (S2S) forecasting. Here’s their statement of task:
An ad hoc committee will conduct a study that will identify opportunities to increase forecasting skill on subseasonal to seasonal (S2S) timescales based on the 2010 NRC report Assessment of Intraseasonal to Interannual Climate Prediction and Predictability and progress since. The report will describe a strategy to increase the nation’s scientific capacity for research on S2S forecasting. The committee will develop a 10-year scientific research agenda to accelerate progress on extending prediction skill for weather and ocean forecasts at spatial and temporal resolutions to aide in decision making. The committee’s report will cover: * Identification of potential sources of predictability and assessment of their relative value for advancing predictive skill; * Identification of process studies for incorporating new sources of predictability into models; * Application and advancement of ocean-atmosphere-ice-land coupled models; * Key observations needed for model initialization and verification of S2S forecasts; * Uncertainty quantification and verification of probabilistic products; * Approaches to communicating this type of prediction in a way that is useful to and understandable by decision makers; and * Computational and data storage and visualization infrastructure requirements.
The committee is chaired by Ray Ban; membership of the full committee, agendas for the meetings, and other information is available on the link above.
In the public session of today’s meeting, the committee will be discussing the value of improved S2S forecasts. They’ll hear first from Fern Gibbons and Sara Gonzalez-Rothi, of Senate staff. That should be quite interesting and informative. A second panel for the morning was supposed to feature Jeff Lazo, an NCAR economist and lead author of the seminal paper (in an otherwise very skimpy literature) entitled U.S. Economic sensitivity to weather variability, published in the AMS Bulletin in 2011; and Lawrence Friedl, who directs NASA’s Applied Sciences Program, arguably one of the best-structured and most strategic approaches to applying scientific advances in our field to beneficial use. These two individuals are what former NOAA Administrator D. James Baker used to refer to as “PWAKS” – People Who Actually Know Something. Paul Higgins (who directs the Policy Program here at AMS ) and I were invited to fill in around the edges. However, through one of those accidents of scheduling, he and I may be the only folks to show up.
In my remarks, I’ll be largely repeating two simple points covered in posts to this LOTRW blog over the years and in the corresponding book.
First, costs of S2S forecasts, or any similar services, for that matter, are difficult enough to estimate, but the benefits or value are harder to measure still. Many factors contribute to this, but one of the biggest is that value or benefit, and the allocation of that value across sectors of society and individuals, are determined as much by public policy as they are by science or engineering. To repeat the (perhaps over-used) example from the blog and from the book, deregulation of electricity and the construction of regional grids have combined with the rise of solar and wind power to dramatically increase the value of weather forecasts. By contrast, regulation of dam operations on watersheds requiring that management decisions be based on reservoir levels without reference to forecasts changes in those levels effectively reduces the value of (in this case, S2S) weather forecasts to zero. This latter has been documented by Rayner, Lash, and Ingram: WEATHER FORECASTS ARE FOR WIMPS: WHY WATER RESOURCE MANAGERS DO NOT USE CLIMATE FORECASTS; Climatic Change 69: 197–227, Springer (2005).
Second, the societal benefit from advances in forecast skill (or advances in any kind of science or technology for that matter) depends on the application of the innovation (in effect, SB=IA). Application of science isn’t something that just happens; it’s a scientific discipline or set of disciplines, and is therefore a fit subject for research and analysis in its own right. This has been argued very elegantly, for example, by former AMS President Bill Gail in a number of settings (e.g., here).
Both these realities suggest that valuation ought to be viewed much as Eisenhower viewed planning. Specific valuations ought to be taken with a grain of salt. But the thought process of characterizing value is priceless.
In this light, it would seem that (1) valuation of S2S forecasts and (2) applications of S2S forecasts each deserves a research agenda in its own right – agendas that ought to be developed and pursued with vigor in parallel with efforts to improve the forecasts themselves.