Let me introduce myself. I received a Ph.D. degree in Biochemistry and Molecular Biology (Purdue University) and did postgraduate training in Biophysics (The Albert Einstein College of Medicine). I ran a research lab at a major university medical center for 32 years and recently retired. I served as a research consultant for the National Institutes of Health for 8 of those years and am presently on the editorial board of a journal in my field.

Several years ago I began reading the literature on climate change that was appearing in Science, Nature, and other peer-reviewed journals. I did so because I was concerned at the alarmism I was seeing in the media regarding “global warming” and the dire predictions of some in the scientific literature. I consider myself a scientific skeptic and want to be convinced by the data before I accept something as “true” (see Freeman Dyson at edge.org on skepticism in science).

http://www.edge.org/3rd_culture/dysonf07/dysonf07_index.html

As a biologist, I am aware of a number of cases in which science has been led in directions not based on hard evidence. Examples include Malthus and the Malthusian Theory, Lysenkoism in the old Soviet Union, and eugenics in the US and elsewhere (see the excellent archive at Cold Spring Harbor for examples of such “science”). I suspect not one in fifty Americans alive today is aware that nearly 30,000 were sterilized in the early part of the 20^{th} century because they were deemed “genetically defective”.

http://www.dnaftb.org/eugenics/

To quote from the introduction to the Cold Spring Harbor website “…..It is important to remind yourself that the vast majority of eugenics work has been completely discredited. In the final analysis, the eugenic description of human life reflected political and social prejudices, rather than scientific facts.” We must try and ensure that we don’t repeat this process with the issue of climate change.

In my reading on climate and climate change I read the work of Stott, Shaviv, Hansen, Veizer, Feddema, Von Storch and numerous others, too many to mention. I also discovered the work of the Pielkes, first that of Roger Junior and later that of his father, Roger Senior. Subsequently, I have read the works of Prins, Rayner, Sarewitz, and others regarding the fallacy of focusing only on CO_{2}, as does the Kyoto approach. I believe their conclusion is correct, Kyoto is a failure and a new approach is badly needed.

Earlier this year our local paper, the Kansas City Star, ran an article on temperature anomalies in Kansas City. I had read the definition by Roger Pielke Sr. of an anomaly as being defined by 2 or more standard deviations from the mean and wanted to examine the data for myself I then did some web research and found the NOAA data for KC for April for the past 20+ years. The graphed data from NOAA showed a trend line with a slight slope.

This slope seemed of questionable significance and I wrote the following email to a NOAA contact:

Lyn Yarbrough said the following on 4/13/2008 9:23 AM:

1. I did a plot of the April data from 1972 to 2007 and the trend line (green) showed a positive slope. What is the correlation coefficient for the trend line for this time period? I serve on the editorial board of a journal of Biochemistry and Biophysics and I would reject a paper that showed such a “trend” because it appears to me that the correlation coefficient would be so low as to be meaningless from a statistical point of view. Why not show the correlation coefficient of any “trend” so that viewers could evaluate the significance for themselves?

2. Are means and standard deviations available for daily high and low temperatures for April over the above time period? The local newspaper had an article today stating that the low temperatures we are seeing this year in KC are not anomalous. From my reading about climate (the scientific literature and blogs such as that of Roger Pielke Sr.) I understand that an anomaly is defined as a temperature that falls outside two standard deviations of the mean. Is this correct?

3. The table format presumably intended to format the data and rank from both the highest and lowest temperature. Both appear to be the same

thank you, Lynwood Yarbrough

I received the following response to my inquiry:

Lyn,

1. At present, we do not compute the correlation coefficient for the trend lines on these plots. We have talked about including them for those who are more statistically savvy. This website was originally designed to be interpreted by a layperson. We will likely include more statistics with these plots in the future.

2. The Climate at a Glance website doesn’t utilize daily data, only monthly data. So, I wouldn’t have these statistics handy. We utilize the term “anomaly” to mean anything which departs from the mean (positive or negative). So, certainly, 2 std from the mean would also be an anomaly, but I’m not sure if it is more accurate than the definition we use here at NCDC.

3. The table format allows a user to shorten the period of record for ranking purposes. You will always have a full period of record comparison and whatever the user selects. The default is the entire period of record. So, in that case, you will see two sets of the same rankings for the same period of record. I hope that makes sense.

Thank you for your input. It will help to make our products more useful in the future.

Following this response, I downloaded the data and did a linear regression and the R value obtained was 0.157.

As I understand, the square of this value (coefficient of variation) represents the observed variation due to the independent variable (year). R square is thus equal to approximately 0.03. I then did a “t test” of the temperatures for the first ten years of the period (1973-1982) and compared them with the last 10 years (1998-2007). The results are below.

The results of an unpaired t-test performed at 16:12 on 7-MAY-2008

t= -1.09

sdev= 3.19

degrees of freedom = 18

The probability of this result, assuming the null hypothesis, is 0.29

I am no statistician and have no formal training in statistics. However, it appears fair to conclude that, based on the above data, there has been no statistically significant change in average April temperature for Kansas City over the time period for which data was obtained. I have examined NOAA graphs for other months during this time period and, although I have not analyzed the data, most months show “trends” similar to that above.

I think that these data show clearly that there should be changes in the way in which NOAA presents temperature data and any “trends” in data. At minimum, the NOAA graph should show the calculated correlation coefficient for any trend line shown. I believe that there are few, if any, journals in my field of science and medicine that would publish a report purporting to show a significant trend based on data such as these.

Finally, the NOAA contact noted that “We utilize the term “anomaly” to mean anything which departs from the mean (positive or negative). So, certainly, 2 std from the mean would also be an anomaly, but I’m not sure if it is more accurate than the definition we use here at NCDC.” Since the temperature data is presented to tenths of a degree, almost every year would be anomalous by this definition since few years show exactly the same temperature for any given month. Perhaps there needs to be a little more rigor in data analysis and the definition of climate change at NOAA. See:

Pielke, R.A. and N. Waage, 1987: A definition of normal weather. Natl. Wea. Dig., 12, 20-22.