We encourage your correspondence, and although we cannot print all that we receive, we will include at least one short, perhaps edited, letter per issue.

Dear WeatherZine

In response to your editorial in WeatherZine #21 (Six Heretical Notions About Weather Policy:

When weather disasters happened in the past, the National Weather Service frequently benefited from an infusion of funds. After a 1974 tornado outbreak, for example, a research group was set up at the National Severe Storms Forecast Center. Lately, disasters come and go with little benefit to the weather community. The problem is that we have been ineffectual at convincing the public and policy-makers that (a) they are getting a bargain at the current cost, and (b) they could do even better with a larger investment.

If we ignore the role that human forecasters have in the process of generating forecasts, then we will get less from any level of investment than is possible. Although it seems we are on a juggernaut aimed at removing any role for human forecasters in the public weather service by 2050 (my forecast!), properly educated and trained humans can add considerable value to what any model can generate. We have chosen not to invest in meaningful training of our weather forecasters, thereby putting our hopes into improvement of the numerical model forecasts.

Much collected data goes unused because of a lack of access. Operational data has not been collected to be used in research. Once these data are used by the operational agencies, they go into some "black hole" where retrieval is difficult and expensive. Moreover, data quality issues tied to operationally-collected observations can be a serious barrier to their use.

Access to data collected in field programs typically is controlled and often limited to those directly involved in the data collection. This can seriously limit the general value of the data sets. We often are our own worst enemies at getting the full value from the data we collect at taxpayer expense.

As for forecast quality, the NWS continues to put forth only a token effort into a truly meaningful forecast verification program. Virtually nothing is happening with respect to "closing the loop" on any verification effort.

Another critical gap in our understanding is what it takes to be a high-quality human weather forecaster. We have no idea what characterizes the best human forecasters, by any measure of forecast quality. An obvious implication is that there is no role envisioned for them in the far future ...learning how humans do their part of the forecasting job well is apparently seen as a meaningless, dead-end task. The lack of any interest in this also tends to become a self-fulfilling prophecy.

We are very bad at applying what we know in the real world of operational forecasting. Decision-makers in research and operations are ignorant of the other's needs and interests so that whatever happens to address the gap is done by individuals with little or no support from either set of decision-makers. If only a small fraction of what is known from research is being applied in operations, we have only ourselves to blame!

The operational bureaucracy is very reluctant to change procedures to reflect new knowledge. We are still putting out superannuated "zone" forecasts when it is possible to provide gridded model fields of virtually any forecast quantity over the Internet. If obsolete products are allowed to continue indefinitely, there is little value to any research tied to product improvement.

The NWS is simply not equipped to provide good answers about the value of what they do, in spite of my belief that even their current forecast quality still manages to be extremely valuable, in spite of being significantly lower than what it could be if they made forecast quality a priority. I see little indication that anyone is doing much to address these critical issues. Defensive reactions to your "heresies" indicate a lack of understanding of what you are trying to accomplish with this dialog.

— Dr. Charles A. Doswell III
NOAA/National Severe Storms Laboratory
www.nssl.noaa.gov/~doswell/
doswell@nssl.noaa.gov
Standard disclaimer ... my views are my own, so don't blame my employer.

Comments? thunder@ucar.edu

[ Top of Page ]