Not long ago I criticized an AGU assessment of hurricane science for its demonstrably inaccurate treatment of seasonal climate forecasts. I hypothesized that the issue of seasonal hurricane forecasts had been caught up in the “hurricane-climate wars” between Bill Gray and Greg Holland. Holland and Peter Webster (who were both involved with preparing the AGU report) took serious issue with my even raising this hypothesis (how dare I!!), flatly denying any such relationship between the AGU report’s criticism of Gray’s seasonal forecasts and the global warming debate. However, In today’s hearing (that I participated in) Judy Curry’s testimony completely vindicates my raising this issue (Curry is a collaborator with Holland and Webster). Here is the relevant excerpt from her testimony (PDF):

It may take up to a decade for the observations to clarify the situation as to which explanation, natural variability or global warming, has better predictive ability. In the short term, evaluation of seasonal forecasts for the North Atlantic can provide some insights into the predictive capability of natural variability. Holland (2006) has conducted an assessment of statistical forecasts of North Atlantic tropical storm activity. Seasonal forecasts are based upon the statistics of North Atlantic tropical storms for the period since 1950. W. Gray commenced making seasonal forecasts in 1984. For the first decade (until 1994), Grays forecasts performed well (Figure 10), with a bias error of -0.2 storms per season for the June forecasts and a root mean square error of 1.8. In the period since 1998, Grays forecasts have performed much worse, with a notable low bias averaging -3.1 storms per season and a root mean square error of 5.2. NOAAs seasonal forecasts for the same period show little variation from Grays forecasts. It is argued here that the persistent low bias in the seasonal forecasts since 1995 indicates that the elevated activity in this period cannot be explained solely by natural variability seen in the historical data record since 1950.