A few months back I downloaded the raw records for Missouri--all the stations. I then compared nearby cities across the state to see how repeatable the data is. After all, two nearby cities ought to have annual temperatures which are quite nearly identical. Unfortunately that is not what one finds when one compares nearby cities.
So, lets look at Bowling Green, MO and Warrenton, MO. I subtracted the annual temperature of Warrenton from Bowling Green's. The two towns are 40 miles apart and the annual temperature difference varies from -4 to +5, meaning that the temperature profile between the two towns, measured by the thermometers is highly eratic and varies by a swing of 9 degrees fahrenheit.
These kinds of temperature differences, and reversals of which town is hot should not be happening over a distance of 40 miles. Temperature differences cause thunderstorms--that is what a cold front coming through Missouri causes. But, if we are to believe the temperature record, we are supposed to believe that a 5 deg F temperature difference existed between Bowling Green and Warrenton for an entire yea without any thunderstorms or winds arising from that temperature difference. If you believe that I have some swamp land to sell you.
A few months ago a friend of mine loaded the Missouri temperature data into a program used in the oil industry to evaluate 3D cubes of seismic data. I just got that data today and have begun to analyze it. We loaded the temperature data for each city and contoured each year and made a 3D cube of it over Missouri, over X, Y and year. I thought I would share a couple of time slices (years). They illustrate the problems quite effectively.
The first is from 1972. It shows an isolated cold area in SW Missouri. Doniphan was quite cold, Marble Hills slightly cold. When we contoured the data, it made the cold area you see (blue). Now, if there is no error in the data, we are forced to believe that the coldest part of the state in 1972 was near Doniphan in the SW part of the state. There is only one problem. I drove through there that Christmas on my way to a fraternity convetion. It got colder as we went northeast. Sure that was only during the winter months, but at least on that drive, it doesn't verify the map.
The second map is from 1950. If we are to believe the data, the coldest part of the state is the SE part of the state--Caruthersville. Yes, all that cold air avoided the northern part of Missouri, going either through Kansas, Oklahoma and Arkansas to end up at Caruthersville, or it went through Illinois, Kentucky, Tennessee and Mississippi to end up there. This data makes no sense.
One final thing. I took each pair of nearest cities, marching across the state and captured the maximum positive and maximum negative annual temperature difference between each city and its two nearest neighbors. Remember these temperature differences should be small if the the yearly temperatures vary together, as they ought to over small distances. But, the differences are anything but small.
The data upon which we base our conclusions of global warming is so bad, so noisy and leads to silly conclusions(like SE Missouri being the coldest for the year 1950, that to beleive anything about what the climate is doing from the thermometer data is quite silly.