The correctness of such claims can be evaluated rather simply by analyzing trends in extreme weather events over the historic past. If the observational data show no trend, or if they decline over time toward the present, the hypothesis that rising CO2 is increasing the frequency and/or magnitude of the events can be falsified. For under such circumstances, it cannot be concluded that rising CO2 is having any measurable effect on the extreme weather event under examination. Yet it is a bit more complicated than that.

Figure 3 presents a flow chart of the many questions that must be asked, and the steps that must be followed, before one is able to properly test the CO2-induced increase in extreme weather hypothesis. A critical step in this process centers on obtaining datasets of sufficient length to conduct proper statistical analyses. False signals can be obtained if a dataset is too short. Determining what constitutes a sufficiently long dataset begins with an understanding of how atmospheric CO2 and global temperature have changed over time.

Figure 3. Flow chart detailing questions that must be addressed and steps that must be taken to perform a proper analysis to test the model-based hypothesis that rising CO2 concentrations are increasing the frequency and severity of extreme weather events.

Atmospheric carbon dioxide concentrations have been rising since the dawn of the Industrial Revolution. Driven by gaseous emissions from the burning of fossil fuels such as coal, gas and oil, the air's CO2 content has risen from a mean concentration of about 280 parts per million (ppm) in 1800 to a value of approximately 400 ppm today. This historic rise in CO2, however, has not been uniform. Half of the increase has occurred since 1980 and three-fourths has occurred since the end of World War II (WWII). Therefore, if rising CO2 is having an effect on extreme weather, testing for such requires examination of extreme weather events that have occurred over a period of time in which a large fraction of the modern buildup of CO2 has occurred. Though it is perhaps somewhat subjective to designate at what point in time the rise in CO2 constitutes a "large fraction" of the modern increase, a good starting point would be since the end of WWII (~70 yrs), as three-fourths of its modern increase occurred since that time. However, because many extreme weather datasets do not extend back in time 70 years, a secondary starting option would be some interval of time between the end of WWII and 1979, as half of the modern increase in atmospheric CO2 has occurred during the past 35 years. If no trend or a declining trend in the data is observed over either of these two periods, the hypothesis that a given extreme weather event is affected by the rise in CO2 cannot be verified, and is more likely falsified.

But what if a rising trend were observed in the data, would that be proof of a CO2-induced influence? In a word, no. As shown in Figure 3, additional analyses must be performed.

It has already been established that, at a minimum, trends in extreme weather events must be evaluated over the period of majority buildup of CO2, which logically could be interpreted as the three-fourths increase that has occurred since the end of WWII or the one-half increase since 1979. If a rising trend is observed over this period, the parameter must further be examined over a much longer time period from which the full expression of its natural variability can be observed. And because extreme weather events are projected to increase in consequence of CO2-induced global warming, the only way to obtain an untainted view of their natural variability is to examine how these events responded to changes in climate over similar warm periods prior to the modern buildup of anthropogenic CO2. In most cases, this requires extending datasets back in time approximately 1,000 years to a climatic period known as the Medieval Warm Period, which was the last time global temperatures reached levels as warm as-or warmer than-they are today. Nevertheless, shorter datasets may still be used to falsify the CO2-induced extreme weather hypothesis, they just can't be used to prove it. Though a record may only extend back in time 200, 300, or 500 years, if it shows no trend or a declining trend, the hypothesis of a CO2-induced influence can be rejected.

The example below illustrates the importance of tempering claims of a CO2-induced influence on extreme weather events of the modern era and the need to study and evaluate their occurrence over a much longer period where the full expression of natural variability can be observed.

According to John Hooper's 14 August 2002 article in The Guardian, in the midst of that year's massive flooding in Europe, Gallus Cadonau (the managing director of the Swiss Greina Foundation) called for a punitive tariff on U.S. imports to force cooperation on greenhouse gas emissions, claiming that the flooding "definitely ha[d] to do with global warming" and stating that "we must change something now." He was joined in this sentiment by Germany's environment minister, Jurgen Trittin, who implied much the same thing when he said "if we don't want this development to get worse, then we must continue with the consistent reduction of environmentally harmful greenhouse gasses." A thorough analysis of historical flood accounts and more recent river-flow data by Mudelsee et al. (2003), however, revealed something very different.

What this team of German researchers did was to analyze historical documents stretching from the 11th century to 1850 and subsequent water stage and daily runoff records from then until 2002 pertaining to two of the largest rivers in central Europe, the Elbe and Oder rivers, seeking to determine trends in flood occurrence over the past thousand years. In so doing, the scientists report that "for the past 80 to 150 years"-which climate alarmists typically describe as a period of unprecedented CO2-induced global warming-"we find a decrease in winter flood occurrence in both rivers, while summer floods show no trend, consistent with trends in extreme precipitation occurrence." Thus, the strident claims of Cadonau and Trittin that global warming had caused the 2002 flooding failed to stand up to scrutiny when compared with historical observations. As the world recovered and warmed from the global chill of the Little Ice Age, flooding of the Elbe and Oder rivers did not materially change in summer and actually decreased in winter.

Blaming anthropogenic CO2 emissions for the European flooding of 2002 was obviously incorrect and not a reasoned deduction based on scientific evidence. If Cadonau and Trittin had properly followed the steps outlined in Figure 3, they would not have gotten things so wrong.

The dilemma and difficulty in most analyses of extreme weather events, however, is that modern records do not extend back that far in time. Indeed, most datasets only go back a few decades, rarely eclipsing a century in length. Thus, proxy records of extreme weather events must be collected and studied. And that is not easy to do. They take time and effort, and they are costly to produce. Nevertheless, they are necessary for those desiring to conduct proper scientifically-based analyses of extreme weather events.

At this point, as illustrated in Figure 3, it should also be noted that even if a location yields a positive trend in an extreme weather event across a time period of over 1,000 years or more, such a finding is still not sufficient to validate the model-based claims; for the mere existence of a positive trend does not prove it was caused by CO2-induced influences. And that brings up the third and final step required to properly establish a CO2 effect on extreme weather events: The influence of all other (non CO2-driven) variables that impact a given extreme weather event must be studied and factored out of observable trends.