It was Dr. Gene Willeke's idea to create a National Drought Atlas. Gene worked at IWR during the first year of the National Drought Study while on sabbatical leave from the University of Miami (Ohio). Looking back now, the Atlas seems an obvious idea, but it was not then. Had it not been for a series of fortuitous coincidences, the project probably would have been too ambitious even for a national study.
We knew that assessing the risk that drought posed to any region required a statistical analysis of regional precipitation or streamflow, or both. Because droughts are relatively long events in relation to the time we have been recording streamflows and precipitation, this statistical analysis was not a trivial exercise; an analyst would have to find pertinent precipitation or streamflow stations that had been recording for at least a few decades, assure that the data were relatively continuous and reliable, and then fit a distribution to the data. Even if all went well, one would have little confidence in the calculated return intervals of long droughts. A USGS report published around 1990 typically characterized the recurrence intervals of historic droughts in terms like "> 10 years".
Gene led the development of the Atlas from his original inspiration to a widely reviewed draft report. That report was large (593 pages, each 11" X 17", mostly in Courier 7 font). The web version lacks only the Palmer Drought information, which we saw as much less valuable than precipitation or streamflow, but it also contains much more than the 600 page paper Atlas did. The web version automates much of the cross referencing and calculation required when using the paper version, and it is linked to many other sources of information. And of course, it is more widely available in this form.
Gene had the right idea, but even he didn't realize at first how good his timing was. The first instance of good fortune was the National Drought Study itself; it gave the Corps funding to carry out unusual projects like the Atlas. But we were lucky on other fronts, too. Because of growing concerns about our ability to detect incipient climate change, NOAA and USGS had just developed 48 state databases of high quality, long record length precipitation and streamflow data (the HCN datasets, respectively), which assured us unimpeachable data at minimum cost and time. John Vogel, then with the National Weather Service, informed us of our final bit of luck; Jon Hosking and Jim Wallis, IBM researchers, had just demonstrated a new statistical method known as "l-moment analysis". This method allowed us to produce much robust estimates of non-exceedance frequencies of long duration events than we would have otherwise. This was especially true for the precipitation analysis, because data from nearby and similar stations were considered together, a process, known as regionalization. The Atlas was the first practical application of the l-moment methodology. Since that time, it has been widely recognized as a landmark improvement in statistics, and has been adopted for many purposes. Dr. Vogel used it in later work on precipitation, and Great Britain uses l-moment analysis as part of its official method of estimating flood frequencies.
Special thanks are owed to the Thomas J. Watson Research Laboratory of IBM. The development of Wallis and Hosking would have been impossible without IBM's commitment to basic research. IBM also allowed them to contribute a substantial amount of time (and computer time) to the Atlas project simply because of its scientific value. They led the review and further refinement of the HCDN and HCN data and provided the mathematical leadership for the project. Dr. Ned Guttman of the National Climate Data Center worked with Dr. Wallis and Dr. Hosking, including an intensive year of number crunching, to develop the precipitation regions (clusters) and quantiles. He has since written many papers on the Atlas. He and Jim Wallis have provided me constant encouragement and help in producing this web version of the Atlas.
A number of people were involved in planning the Atlas, especially Will Thomas (U.S. Geological Survey), Ken Kunkel (Director, Midwest Climate Center, Champaign, IL), and Clive Walker (U.S. Department of Agriculture, Soil Conservation Service). Zoltan Montvai was the National Drought Study point person at Corps Headquarters. He not only represented us and made sure the budget process worked, he took an active interest in the technical details of the study. Richard Punnett (Corps of Engineers) made significant contributions in review and in developing specific applications of the Atlas material. It was he who told us we needed to estimate the frequencies of long duration wet periods, so that reservoir operators could better estimate the odds of re-filling reservoirs. The National Drought Study held several workshops in which Corps of Engineers personnel from districts and divisions and headquarters participated, including an intensive review at IBM's Thomas J. Watson Research Laboratory. Advice from our workshops and comments made by attenders at several meetings of the American Association of State Climatologists and the American Society of Civil Engineers helped us make the Atlas more user friendly. Substantial financial support was also provided by Miami University. Mark Kochan and Beth Leamer, who were graduate assistants at Miami University, did valuable literature and geographic analysis in the early stages of Atlas preparation. Finally, I’m grateful for the continued support of Dr. Gene Stakhiv at IWR, who has championed the effective application of good science to practical water resources issues throughout his career.
revised 1 Aug 06