My interest in this topic dates from when I joined a regional water authority at the start of 1976. I could not fathom why, in some resource analyses, one studied reservoir inflows over set periods beginning in a given month, whereas in others one studied the minimum inflows over set durations without regard to starting month.
The severe drought of that year tested resources to the limit. Were the supply difficulties attributable to the severity of the event, or might mismanagement or misdesign be partly to blame? It was uncomfortable to find that there was no accepted way of verifying drought severity through an analysis of rainfall records - something that even an unloved water authority could not be accused of falsifying.
Soon after, authoritative analyses of long-duration rainfall depths appeared (Tabony, 1977), from which the ubiquitous Tabony Tables emerged. No longer working in water resources, it was some years before I realised that the problem had not been solved. The tables are correct in themselves but their misapplication to assess drought severity presumes, without foundation, that the resource system under stress is only sensitive to low rainfalls of the given duration and starting month.
The opportunity to look at the problem afresh came in an abstract study of dependent time-series (Reed, 1995b). While I was grateful to NERC for supporting this excursion, I was annoyed that my favourite industry maintained a casual approach to drought severity assessment. Shamefully, it suits water companies that (mis)application of Tabony Tables exaggerates drought rarities: it makes supply restrictions a little easier to justify. In 1995, Yorkshire Water's reputation was undone, not by the nonsense assertion that they were dealing with a 1000-year drought but because an executive dared to shower outside the region.
In 1996, I tried a new tactic at a seminar designed to boost water industry funding for ecological and hydrological research: doing a straight presentation but handing out a flyer that parodied the drought severity assessment problem. Sadly, its only success was in its uncanny prescience of English football fortunes (Tel and Glenda - why they parted).
The paper to the CIWEM Centenary Conference (Reed, 1995b) provides the bones of a solution method. It is built around a simple simulation that interprets the significance of the rainfall drought to the water-resource system. It remains to apply the technique to different types of resource system, and to generalise results across the UK. A spin-off application allows appraisal of the benefits - in terms of increased drought resistance - of cross-connecting resource systems. The gains are less than one might think: major droughts tend to have a large footprint.
While such a study might exploit long-term 5 or 7-day rainfall totals, more readily available monthly rainfall series suffice to assess all but the most quickly responding resource systems.
If you are interested in any of the above, please do get in touch. It was pleasing to discover the need for the research, and the relevance of the method, remembered in a CEH review for the Environment Agency in 2003.