The Magical World of Climate Change in Iowa
The National Climate Assessment (NCA) has focused attention on Iowa, which is probably good, given how it is apparently in the public interest to examine what is being claimed about climate change in the Hawkeye State.
On January 1, 2011, the Iowa Climate Change Impacts Committee released a "Report to the Governor and the Iowa General Assembly" entitled "Climate Change Impacts on Iowa: 2010." Information from this report was used in the NCA section on agriculture.
Tom Philpott at Mother Jones wrote about the NCA's take on this Iowa climate change report in his subtly (note sarcasm) titled article "Our Alarming Food Future, Explained in 7 Charts." The following figure was republished from the NCA:
Some major problems exist with this plot and the interpretations of it by the NCA, and -- of course -- Mother Jones.
Philpott at Mother Jones wrote this about the figure above:
"Iowa is hemorrhaging soil. A while back, I wrote about Iowa's quiet soil crisis. When heavy rains strike bare corn and soy fields in the spring, huge amounts of topsoil wash away. Known as 'gully erosion,' this kind of soil loss currently isn't counted in the US Department of Agriculture's rosy erosion numbers, which hold that Iowa's soils are holding steady. But Richard Cruse, an agronomist and the director of Iowa State University's Iowa Water Center, has found Iowa's soils are currently disappearing at a rate as much as 16 times faster than the natural regeneration. According to the National Assessment, days of heavy rain have increased steadily in Iowa over the past two decades, and will continue doing so."
So the point is that climate change is increasing the number of heavy rain events each year in Iowa, and that this is eroding soil and threatening agriculture -- and, by extension, the food security of Americans.
Follow along as we probe the source of this figure and its flaws. First off, Philpott is reproducing the figure directly from the NCA, who published it with the following caption:
"The figure shows the number of days per year during which more than 1.25 inches of rain fell in Des Moines, Iowa. Recent frequent occurrences of such events are consistent with the significant upward trend of heavy precipitation events documented in the Midwest. (Figure source: adapted from Takle 2011)."
The reference "Takle 2011" is the Iowa climate change report above. Here is Takle's original figure from the Iowa climate change report (of note, Takle is apparently also "a lead author of the National Climate Assessment report"):
You may look at the figure published in the NCA, and the source figure from the Iowa climate change report, and say they look the same. Close, but not quite. Look at the three years with zero precipitation events >1.25 inches. In the NCA, the first one occurs on or about 1950, with the next two spaced what appears to be a year apart in the late 1950s. Then look at Takle's original figure. These three points appear to come as follows: the first one in the mid-1940s and the next two in the early-to-mid 1950s. In other words, one of these graphs is wrong -- and is temporally shifted along the x-axis. There are a couple other differences as well that are odd (e.g., look closely just after 1980 and in the early 1900s between the two plots).
Being curious, I went directly to the source data at NOAA. Sure enough, Takle's original data had been shifted along the x-axis. As well, in the source NOAA data, the years 1945, 1946, and 1947 are listed as "missing" on an annual basis, yet both Takle and the NCA appear to include these years. How? Where could they possibly have gotten the missing data from? If a year is incomplete in the database, it should be omitted entirely from the analysis.
The figure below shows the data direct from NOAA that apparently should have appeared in the Iowa climate change report and the NCA.
By the way, the trend in the data is not statistically significant since 1895 using parametric methods (p>0.1), and gets even less significant (p~0.2) with non-parametric approaches. In other words, there is no significant trend in the number of days of heavy precipitation over time at Des Moines. Funny that neither the NCA, the Iowa climate change report, nor Mother Jones mention that fact. Actually, the NCA strongly suggests the trend is significant, when it is not.
Also recall Philpott's statement that "according to the National Assessment, days of heavy rain have increased steadily in Iowa over the past two decades." There is absolutely no significant temporal trend for the number of days of heavy rain over the past two decades in this dataset. Same applies if we look for a trend since 1960, or 1970, or 1980. None whatsoever.
But there is no reason to believe that Des Moines is representative of Iowa when it comes to heavy precipitation events. Indeed, over at Sioux City in the western part of the state, the time trend for heavy precipitation since 1895 is highly non-significant (p=0.50). The Iowa climate change report also attempts to show that heavy precipitation events are increasing at Cedar Rapids. There has been no significant trend since the mid-1940s. How is that evidence of climate change? In Dubuque, there is no significant trend since before 1910, and the correlation after 1940 is negative. No significant trend since the early 1900s at Iowa City, either. The record doesn't start until 1952 in Waterloo, but absolutely no trend in the number of heavy precipitation events there during the past 60 years.
Iowa's climate change report claims that "Iowa's annual average temperature has increased since 1873 at a rather modest rate." That may be true, but there has been no significant trend since 1900. Apparently "temperatures have increased six times more in winter (0.18°F/decade) than in summer (0.03°F/decade)." Well, that is interesting. I get no significant trend in Iowa's average winter temperatures since the NOAA database begins in 1895. Nor is there one in summer. Indeed, the correlation for summer temperature trends in Iowa since 1895 is massively non-significant (p=0.77).
It is oxymoronic to claim temperatures have increased when the change is statistically insignificant -- which, by definition, means there almost certainly has been no change. This is basic statistics, and is taught in introductory college statistics courses.
Another gem in Iowa's climate change report:
"The annual number of GDDs [growing degree days] in Iowa has varied from location to location across the state over the last 116 years, with increases in Cedar Rapids and Des Moines and a decrease in Waterloo. The most recent 40 years also show site-specific trends, with slight increases in Cedar Rapids and Des Moines and slight decreases in Ottumwa and Mason City."
I get no significant trend in growing degree days either during the last 116 years or the last 40 years at Des Moines. And at Cedar Rapids, there is an almost perfectly non-significant trend for GDDs (p=0.91) since records began, not the "increase" claimed in the climate change report. No significant trend over the past 40 years at Cedar Rapids, either. Over the last 40 years at Mason City there is most certainly no trend, not a "slight decrease." Same goes for Ottumwa.
Last, but not least, The Gazette newspaper in Cedar Rapids reported the following last week when the NCA was released:
"Crop insurance payouts set records in 2011 and 2012, said another briefing speaker, fifth-generation Iowa farmer Matt Russell, who also serves as State Food Policy Project coordinator at Drake University in Des Moines. 'We're getting the wrong weather at the wrong time,' he said, noting that last year's growing season included the wettest May on record, one of the coldest Julys and one of the driest Augusts."
Wait a minute. Last year's growing season in Iowa included "one of the coldest Julys"? Well, 2013 did have the wettest May on record, and one of the driest months of August (ranked 6th driest), but here is the record of July temperatures since 1895.
One of the coldest months of July? No chance. The coldest July on record was in 2009 at 68.0°F, or 5.9°F below the 20th century average. In 2013 the temperature was 72.3°F, only 1.6°F below the 20th century average and only the 36th coldest July in a 119-year record. The standard deviation of July temperatures in Iowa since 1895 is 2.7°F. In other words, July 2013 was only very slightly colder than average. It is just a ridiculous exaggeration to claim that last July in Iowa was "one of the coldest."
And we are supposed to frame coherent science policy based on these types of analyses? No chance.
Apparently coverage of the NCA is waning fast in the mainstream media. One can see why, given the historical collusion between most of the MSM and the climate alarmists. As the flailings of the NCA and its supporting documents are progressively exposed by the alternative media, there will continue to be efforts made to sweep the fundamental flaws in this assessment under the rug.
More reason than ever to continue to keep the heat on the NCA and its media proponents.