December 7, 2007

Sign of the Times

Filed under: Droughts, Floods, Precipitation

In the Wednesday December 5th, 2007 issue of the New York Times appeared a story by Felicity Barringer titled “Precipitation Across U.S. Intensifies Over 50 Years.” In it, Ms. Barringer reports on a new study released by an organization called Environment America that she described as “a national group that advocates new laws and policies to mitigate the effects of climate change.” The focus of the Environment America report was on the results of an analysis they performed examining trends in “extreme” precipitation frequency across the United States (apparently Environment America has spread these results to their arms in individual states, Environment Colorado, Environment California… you get the idea). The research was funded by the Pew Charitable Trusts. As Ms. Barringer reported, Environment America concluded that “Across the United States, the number of severe rainfalls and heavy snows has grown significantly in the last half-century, with the greatest increases in New England and the Middle Atlantic region” and, of course, that this was just as predicted to occur from global warming. Lest you think that more precipitation is a good thing, Environment America is quick to warn that “An increase in the frequency of storms delivering large amounts of rain or snow does not necessarily mean more water will be available” and that “[w]hile it may seem like a paradox, scientists expect that extreme downpours will be punctuated by longer periods of relative dryness, increasing the risk of drought.”

That the New York Times chose to highlight the Environment America report is interesting for a number of reasons, among them: 1) the report did not appear in the scientific literature (rather it was released by an environmental organization), 2) a day prior to the Times article, a paper was published in the peer-reviewed scientific literature which concluded that “extreme” precipitation was increasing across the United States (but no mention of it was made by Ms. Barringer), and 3) a paper published in the peer-reviewed scientific literature back in 2004, also described increases in “extreme” precipitation events across the United States, but also cautioned that based upon the nature of precipitation, increases in precipitation amounts are virtually always accompanied by apparent increases in “extreme” precipitation events and thus singling out “extreme” precipitation events without discussing overall precipitation trends is alarmist by its very nature. Oh, we forgot to mention, that the two papers in the scientific literature that found increases in extreme precipitation across the U.S. included among their co-authors some folks often deemed industry-funded global warming skeptics (Brommer, Cerveny, Balling, 2007; Michaels, Knappenberger, Frauenfeld, Davis, 2004). Perhaps our papers didn’t warrant coverage in the Times because we failed to include breathless prose about how global warming was to blame, or perhaps because we failed to outline steps that should be taken to reduce carbon dioxide emissions, or maybe we didn’t time the release of our papers to coincide with a major international meeting on climate change (although one of them was published by chance this week). Or maybe, it was that pointing out that the basic findings by some global warming “skeptics” closely matched that from “concerned” environmental organizations would take all the fun out of vilifying the skeptics, so it was best to simply ignore them to avoid the awkwardness altogether.

A closer look at the Environment America paper clearly shows it to be the inferior of the three papers—at least if scientific significance is the standard by which it is judged. If global warming alarmism is deemed the most important attribute, well, then there is no contest.

Environment America basically makes some alarmist claims, tests the one that it knows it can demonstrate to be true, does not discuss those findings in the proper context, chooses not to investigate claims that likely won’t be supported by the data, and then throws in some factual errors for good measure.

For instance, Environment America makes the following assertions in their report “When It Rains, It Pours”:

In summary, scientists expect global warming to alter general precipitation patterns over the contiguous United States in four key ways:

• Storms with extreme rates and amounts of rain or snowfall will become more frequent.

• Summers will tend to be drier while winters will be wetter. Total precipitation will increase over most of the country but not in the Southwest.

• The frequency of extreme events will increase much more than total precipitation.

• Precipitation will become increasingly likely to fall as rain rather than snow — a simple consequence of increased temperatures. Paradoxically, the number of dry days will also increase, because intense downpours will punctuate longer intervals of relatively dry weather.

These assertions are all testable with the precipitation dataset compiled by Environment America, but the only one they examined was the first one. And, given that annual total average precipitation has shown a general increase in the United States over the past century or so, it was a pretty safe bet that “extreme” precipitation events were increasing as well.

Why a safe bet? Because, based on the nature of the distribution of precipitation events, there is a strong association between total precipitation and precipitation from heavy or extreme events. If you think about it you realize that this is obvious. You can’t very well get a lot more precipitation from a lot of light precipitation events—it takes 20 additional 0.10-inch events per year to contribute as much additional precipitation as one 2.00-inch event, or four 0.5-inch events. Twenty additional raindays in a year is a large change while a few extra days with thunderstorms are hardly noticeable. By and large, the total annual precipitation that a location in the U.S. receives is highly dependent on the number of heavy or “extreme” raindays. This is true now, it was true 100 years ago, and it will be true in the future.

And you can see this by comparing the findings from Environment America for changes in extreme precipitation frequency with a plot of the annual total precipitation averaged across the U.S. (note this comparison was not made in the Environment America report). The top half of our Figure 1 is the annual precipitation history for the United States (as compiled by the National Climatic Data Center) from 1948 to 2006 (the period of record in the Environment America analysis). The lower half of Figure 1 is the annual frequency of “extreme” precipitation events across the U.S. as calculated by Environment America.

Figure 1. Top: The annual precipitation history of the United States, 1948-2006 (source: National Climatic Data Center). Bottom: the average annual frequency of extreme precipitation events across the United States, 1948-2006 (source: Environment America).

Notice how closely Environment America’s “average annual frequency of extreme rainstorms” tracks the observed average total annual precipitation across the U.S. Low values in both datasets in the 1950s (when the U.S. was in a major drought), an increase from the 1950 to the early 1980s, low values in the late 1980s, high values in the mid-1990s, low values in the late 1990s-early 2000s and near average values in 2006, the most recent year. Overall there is an upwards trend from the drought conditions of the 1950s to the general wetness of the past few decades. This close correspondence between the frequency of extreme precipitation events and total average precipitation is what we wrote about back in 2004 (Michaels et al., 2004), and what is by and large the way things have to be—that is, the more precipitation you get in a year, the more of it comes via heavier events and vice versa. So if the climate is changing so that we get more precipitation, it virtually has to be the case the more of it will come in heavy or extreme events. C’est la vie.

So this claim had to be true, it was a no-brainer.

The second claim, that summers will tend to be drier while winters will be wetter, could have easily been analyzed to see if it was indeed occurring—after all since Environment America claims that there first global warming prediction is on-going, shouldn’t all of them be?

But, alas, no actual data concerning seasonal precipitation trends is presented by Environment America. Wonder why? Well, it is because they likely wouldn’t have found any evidence for it.

A quick and dirty analysis can be had by using the U.S. climate data visualization tool available at the website of the National Climatic Data Center ( Figure 2 shows what we found when we plotted the U.S. averaged precipitation data from 1948-2006 by season. Virtually the opposite of the Environment America assertion—there was no trend in winter precipitation, while spring, summer and fall seasons all showed a tendency for increasing precipitation.

Oh well, so much for that one.

Figure 2. U.S. precipitation by season, 1948-2006 (source: National Climatic Data Center).

What about the third claim, that the frequency of extreme events will increase much more than total precipitation? Another easily tested hypothesis given their precipitation dataset, but another in which no results are presented. Why not? Well, you can probably imagine. But just in case you can’t, Figure 3 depicts what we found when we performed a related analysis in our 2004 paper (Michaels et al., 2004). The top half of Figure 3 depicts the amount of precipitation that falls on the wettest day each year (i.e “extreme” precipitation) averaged across the U.S. from 1910 to 2001. Notice, as we mentioned previously, that we indeed found that the wettest days of the year were getting wetter, or to put it another way, precipitation during extreme events was increasing. However, contrary to the expectation of Environment America, the amount of precipitation falling on the wettest day as a percentage of the annual precipitation did not change at all over the same period (Figure 3 bottom). In other words, precipitation amounts in extreme events were not increasing more than total precipitation, or, to put it another way, the increase in extreme precipitation was not “disproportionate” when compared to the overall precipitation.

Another assertion down the drain.

Figure 3. (top) Average amount of precipitation that fell on the wettest day of each year across the United States, 1910-2001. (bottom) Percentage of total annual precipitation that fell on the wettest day of each year across the United States, 1910-2001 (image source: Michaels et al., 2004).

Which bring us to their fourth claim. The first part of it, that “precipitation will become increasingly likely to fall as rain rather than snow — a simple consequence of increased temperatures” is likely true (given the climate of the U.S.—not so in Antarctica) as we demonstrated in one of our other peer-reviewed papers (Davis et al., 1999). And as to the second part of the claim, that “paradoxically, the number of dry days will also increase, because intense downpours will punctuate longer intervals of relatively dry weather” there is no indication of this going on at all, a fact that Environment America gets completely backwards (emphasizing how the lack of peer review harms the veracity of their report).

Environment America could have tested this claim themselves, by simply tallying the number of dry days in the precipitation dataset each year and seeing if the number was declining. But, apparently they chose not to (or did so and didn’t like the results), because they present no analysis of their own to support the claim. Instead, they attempt to use the scientific literature to do so. But, they fail miserably, in that they mischaracterize the results they cite and fail to cite other results which show that both the number of raindays and the amount of soil moisture has been increasing across the U.S.

On page 23 of their “When It Rains, It Pours” report, Environment America writes “Since the 1970s in the contiguous United States, an apparently unusual increase in precipitation intensity has occurred. At the same time, the annual number of days with rain or snowfall has decreased.” Huh? We know total precipitation, in general, across the United States has been increasing, but have never heard that the number of precipitation days was decreasing. No reference was given to this statement. One reference that we know of that did examine the trend in precipitation days across the United States was Karl and Knight (1998), so we checked that source and found that Karl and Knight concluded “Clearly, the total annual increase in precipitation frequency [across the United States] of 6.3 days per century significantly contributes to the increase in precipitation.” In other words, precipitation across the U.S. has been increasing, in part due to increases in the number of days with precipitation. Just the opposite of what Environment America reported. Another recently published scientific journal article, this one by Andreadis and Lettenmaier (2006), investigated trends in 20th century drought characteristics across the U.S. They found “Droughts have, for the most part, become shorter, less frequent, and cover a smaller portion of the country over the last century.” More findings that stand in opposition to Environment America’s claims.

So just where did Environment America get the idea that the number of days with rain was declining across the U.S.? Well, we kept reading and came upon the following (p. 23 of their report):

In 2002, Vladimir Semenov and Lennart Bengtsson at the Max Planck Institute for Meteorology in Germany compared actual observations of precipitation intensity with the results of two climate models over the contiguous United States during the 20th century. They found general agreement between the model and reality in terms of the trend toward more frequent extreme precipitation. They also observed that for the northeastern quadrant of the United States, the annual number of days with precipitation has been declining since the 1970s (simultaneously with an increase in the frequency of extreme downpours) – and the model generally reproduces the trend, albeit overestimating the absolute number of days with precipitation. [ref. 61]

Their reference 61 is to:

V.A. Semenov and L. Bengtsson, “Secular Trends in Daily Precipitation Characteristics: Greenhouse Gas Simulation with a Coupled AOGCM,” Climate Dynamics 19: 123-140, 2002.

So we pulled the Semenov and Bengstsson paper (as any peer reviewer would have done) and checked out whether indeed they had reported that “They also observed that for the northeastern quadrant of the United States, the annual number of days with precipitation has been declining since the 1970s.” It turns out that Semenov and Bengtsson did no such thing. In fact, what Environment America interpreted and reported as observed changes across the northeastern United States were actually climate model simulations of the precipitation characteristics there. So Environment America’s lone piece of support for their contention that there has been a decline in raindays across the United States are not based upon observations (which show an increase in raindays), but instead based upon a climate model simulation. Bad, bad, bad. This would not have survived peer-review.

Basically, all of the claims that Environment America made about precipitation across the United States were testable. Environment America chose (or only reported on) the few that they knew had to be correct based simply upon the general characteristic of the weather and the weather trends over the U.S. during the past 50 years (that is, the more rain you get, the more comes from “extreme” events, and the warmer it gets, the less precipitation falls as snow). The other more contentious claims, like, even though we are getting more rain it should be coming in fewer events, and even though we are getting more rain droughts should be getting worse, are in fact wrong.

So, what we are left with is nothing but a climatology lesson from Environment America, trumped up by a bunch of alarmist assertions which they themselves don’t test (even though they easily could) but that have been tested by others and shown to be false. Not that this dissuades Environment America from going on to call for immediate action to restrict greenhouse gas emissions for the sake of preventing climate change—a climate change that (regardless of the cause) has led to a greater amount of precipitation falling across the United States over the course of the past 50 years. And considering that the United Nations has proclaimed that the demand for clean, fresh water will be one of the greatest pressures on humanity during the coming future, you would think that more precipitation would be hailed as a trend for the better.

Apparently, we now know what it takes to get the attention of the New York Times—alarmist claims backed by loosey-goosey science as opposed to sound science not accompanied by dire warnings and calls to action. Obviously, it is the call to action that is more important than the scientific details that underlie it. Such is the sad sign of the Times.


Andreadis, K. and D. Lettenmaier. 2006. Trends in 20th century drought over the continental United States. Geophysical Research Letters, 33, 2006GL025711.

Brommer, D.M, Cerveny, R.S., and R.C. Balling, 2007. Characteristics of long duration precipitation events across the United States. Geophysical Research Letters, 34, L22712, doi:10.1029/2007GL031808.

Davis, R.E., Lowit, M.B., Knappenberger, P.C., and D.R. Legates, 1999. A climatology of snowfall-temperature relationships in Canada. Journal of Geophysical Research, 104, 11985-11994.

Karl, T.R., Knight, R.W., Secular trends in precipitation amount, frequency, and intensity in the United States. Bulletin of the American Meteorological Society, 79, 231-241.

Michaels, P.J., Knappenberger, P.C., Frauenfeld, O.W., and R.E. Davis, 2004. Trends in precipitation on the wettest days of the year across the contiguous United States. International Journal of Climatology, 24, 1872-1882,

Semenov, V.A, and L. Bengtsson, 2002. Secular Trends in Daily Precipitation Characteristics: Greenhouse Gas Simulation with a Coupled AOGCM. Climate Dynamics, 19, 123-140.

No Comments

No comments yet.

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.

Powered by WordPress