An analysis of proxy measures of streamflow in the Mississippi River indicates the occurrence of large megafloods approximately 4700, 3500, 3000, 2500, 2000, 1200 and 300 years ago that were "almost certainly larger than historical floods in the Mississippi watershed" (Brown et al., 1999).  The authors of the study say these fluvial events were likely "episodes of multidecadal duration," spawned by the export of extremely moist air into mid-continental North America that was driven by natural oscillations in Gulf of Mexico ocean currents.

Think of what would happen if the United States were soon to experience another such multidecadal period of "megaflooding."  Before it was even in full-swing, the IPCC-inspired media would probably have driven the U.S. Senate to unanimous ratification of a dozen Kyoto protocols.  But would they have been right?

Such an occurrence of historically-unprecedented weather-related flooding sure sounds like what the climate alarmists are predicting to occur as a consequence of the ongoing rise in the air's CO2 content; but it is clear that the extreme hydrologic events documented by Brown et al. were in no way related to variations in the air's CO2 concentration, as they occurred over a period of near-constancy in this atmospheric property when its values were much lower than those of today. Hence, we can see how easy it would be to make an incorrect and dangerous rush to judgment about the effect of elevated levels of CO2 on streamflow magnitude, which appears to be totally unrelated to changes in this trace constituent of the atmosphere.

So what's really happening?  More recent streamflow data reveal that the conterminous United States is getting wetter in the mean, but less variable at the extremes, where floods and droughts occur.  This is the conclusion of Lins and Slack (1999), who computed secular trends in streamflow for 395 climate-sensitive stations from data obtained from more than 1500 individual streamgauges, some of which had records stretching all the way back to 1914.  These findings, coupled with those presented in our Subject Index summaries for Floods and Droughts, further rebut the climate alarmist claim that extreme hydrological events are increasing in frequency and/or magnitude, as well as the incorrect attribution of such events to the ongoing rise in the air's CO2 content.

In a paper published subsequent to our thus expressing our thoughts on this matter in our Summary of 11 October 2000, Hidalgo et al. (2000) reconstructed a history of streamflow in the Upper Colorado River Basin based on information obtained from tree-ring data.  In their study, they found evidence for times of recurring low streamflow, with "a near-centennial return period of extreme drought events" going all the way back to the early 1500s.

Both the virtual and real stories have been much the same in Europe.  Hisdal et al. (2001), for example, note that "the media often reflect the view that recent severe drought events are signs that the climate has in fact already changed owing to human impacts."  In studying more than 600 records of daily streamflow data obtained from the European Water Archive, however, they conclude that "despite several reports on recent droughts in Europe, there is no clear indication that streamflow drought conditions in Europe have generally become more severe or frequent."  In fact, they typically found just the opposite, i.e., that "overall, the number of negative significant trends pointing towards decreasing drought deficit volumes or fewer drought events exceeded the number of positive significant trends (increasing drought deficit volumes or more drought events)."

Tree-ring chronologies from northeastern Mongolia that were used by Pederson et al. (2001) to reconstruct annual precipitation and streamflow histories for the period 1651-1995 tell much the same story in that part of the world as well.  "Variations over the recent period of instrumental data are not unusual relative to the prior record," say the authors.  In fact, they too report that their reconstructions "appear to show more frequent extended wet periods in more recent decades."  In addition, they detected periodicities in the data that suggested "possible evidence for solar influences in these reconstructions."

Two hydrological model studies make some other important points.  In the first, Baron et al. (2000) evaluated the consequences of a doubling of the air's CO2 concentration and 2 to 4°C increases in air temperature on a high-elevation Rocky Mountain watershed.  Neither the extra CO2 nor the higher air temperatures produced much change in total runoff.  However, the 4°C increase in temperature caused seasonal snow melt to begin four to five weeks earlier than it does currently, allowing the melt water to infiltrate the soil more gradually and for a longer period of time than at present.  The authors indicated such a change would be especially beneficial, because the consequent gradual release of nitrates that are retained in the snowpack and otherwise released in a large pulse in the spring would relieve some of the ecological pressure caused by the high nitrates typically present in springtime flows.

Finally, De Walle et al. (2000) studied historical streamflow, climate and population data for 39 watersheds experiencing ongoing urbanization and 21 more-stable nearby rural watersheds in four regions of the United States to determine the effect of urbanization on mean annual streamflow in the face of potential global warming.  They found that "urbanization increased mean annual streamflow in rough proportion to average cumulative changes in population density," with an essential doubling of mean flow with complete watershed urbanization.  They also found that urbanization reduced the sensitivity of mean annual streamflow to changes in temperature, concluding that "estimates of the impact of future climate change on mean annual streamflow should include consideration of the concurrent effects of population growth."  Doing so, they found, greatly tempers the predicted adverse consequences of global warming and can sometimes even lead to predicted streamflow decreases being transformed into increases.

In conclusion, as ever more data are analyzed, there is ever more evidence that nothing unusual has been happening to streamflow rates the world over as the air's CO2 content has continued to rise.  In addition, there is little reason to believe that any changes that might possibly occur in the future would be deleterious.

References
Baron, J.S., Hartman, M.D., Band, L.E. and Lammers, R.B.  2000.  Sensitivity of a high-elevation Rocky Mountain watershed to altered climate and CO2.  Water Resources Research 36: 89-99.

Brown, P., Kennett, J.P. and Ingram B.L.  1999.  Marine evidence for episodic Holocene megafloods in North America and the northern Gulf of Mexico.  Paleoceanography 14: 498-510.

De Walle, D.R., Swistock, B.R., Johnson, T.E. and McGuire, K.J.  2000.  Potential effects of climate change and urbanization on mean annual streamflow in the United States.  Water Resources Research 36: 2655-2664.

Hidalgo, H.G., Piechota, T.C. and Dracup, J.A.  2000.  Alternative principal components regression procedures for dendrohydrologic reconstructions.  Water Resources Research 36: 3241-3249.

Hisdal, H., Stahl, K., Tallaksen, L.M. and Demuth, S.  2001.  Have streamflow droughts in Europe become more severe or frequent?  International Journal of Climatology 21: 317-333.

Lins, H.F. and Slack, J.R.  1999.  Streamflow trends in the United States.  Geophysical Research Letters 26: 227-230.

Pederson, N., Jacoby, G.C., D'Arrigo, R.D., Cook, E.R. and Buckley, B.M.  2001.  Hydrometeorological reconstructions for northeastern Mongolia derived from tree rings: 1651-1995.  Journal of Climate 14: 872-881.