Two recent articles in the professional literature shed even more light on this interesting subject. The first is forthcoming in Water Resources Research and was produced by two scientists with Belgium’s Katholieke Universiteit Leuven and the research was funded by the Research Programme “Science for a Sustainable Development” of the Belgian Federal Science Policy Office. Ntegeka and Willems start their article noting “Due to the recent climate change stir, various researchers have sought for new techniques to probe the temporal variability of various observed time series.” “Stir” – that’s the first we’ve seen that one!

Anyway, the authors state “The analysis is based on the long-term high-frequency homogeneous rainfall series at the climatological station of the Royal Meteorological Institute of Belgium at Uccle that starts in 1898 and which is continued to date. The series is recorded by the same measuring instrument (a Helmann-Fuess rain gauge) at the same location since 1898 and processed with identical quality since that time. Trends or changes thus cannot be attributed to instrumental changes; the measuring accuracy is homogenous. The measuring frequency is unique as well: 10 minutes with more than 107 years of continuous data. The period 1898 - 2004 has been considered for the present study.”

The scientists present page after page of descriptions of their analytical procedures, but their figure below (Figure 1) largely summarizes the results. They quantify rainfall intensities with respect to a reference (like an average), and as seen in the figures, intensity levels are relatively high at times and low in other times. A case can certainly be made for increasing intensities over the past two decades, but when examined over the entire length of the record from Uccle, similar upward trends have occurred near the beginning of the record and in the 1940s. There is nothing unusual about the recent trends in precipitation intensities! Ntegeka and Willems conclude “In the winter and summer seasons, high extremes were clustered in the 1910s-1920s, the 1960s and recently in the 1990s. This temporal clustering highlights the difficulty of attributing ‘change’ in climate series to anthropogenically induced global warming.” We certainly agree.

Figure 1. Comparison of average quantile perturbations for 10 minutes (a), 1 day (b) and 1 month (c) rainfall extremes and 10-year blocks for summer and winter periods (from Ntegeka and Willems, 2008).

The second article of interest is by a pair of scientists with the School of Earth Sciences at Australia’s University of Melbourne. Davis and Walsh dare to title the piece “Southeast Australian thunderstorms: Are they increasing in frequency?” The pair collected thunderstorm data for stations in southeastern Australia, and indeed, they conclude that “There has been a significant increase in the number of thunderdays from 1941-2004” which must come as terrific news to the global warming crowd. However, the authors note “but much of this increase may have been a result of changes in observing practices in the mid 1950s.”

Figure 2 is the plot of thunderdays from 1941 to 2004 for the Laverton station, and it shows a highly statistically significant increase in thunderdays of nearly two days per decade. The authors present the same data in the second graph (Figure 3) in which they present the results of a break-point analysis. Suddenly, there seems to be no increase in thunderdays. They note that “Changes in observing practices can cause artificial trends to be introduced into climate data.” Australian weather observers generally follow the Australian Meteorological Observers Handbook which states “that the reporting of thunderdays in the phenomena section of the manual could now include all thunder being heard at the station.” Davis and Walsh note that “Previously, storms needed to be located within five miles of the station to be reported.” This change in reporting practices gives the upward trend that could be mistaken as a response to global warming, when in reality, the trend is a simple result of a change in observational practices.

Figure 2. Trends in warm season thunderdays from 1941-2004 for Laverton (from Davis and Walsh, 2008)

Figure 3. Break-point analysis for Laverton warm season thunderdays, showing discontinuity in the data record around 1968 (from Davis and Walsh, 2008)

In both studies, we can certainly see evidence that extreme precipitation events are on the rise. But in both cases, when we explore just a little deeper into the story, we find no evidence of any unusual upward trend in the frequency of extreme precipitation events. The more we learn, the more skeptical we become.

References:

Davis, S and K.J.E. Walsh. 2008. Southeast Australian thunderstorms: Are they increasing in frequency? Australian Meteorological Magazine, 57, 1-11.

Ntegeka, V. and P. Willems. 2008. Trends and multidecadal oscillations in rainfall extremes, based on a more than 100 years time series of 10 minutes rainfall intensities at Uccle, Belgium. Water Resources Research, [in press].  Source