The problem of comparison in Illinois is that, using Chiefio’s list of GISS stations that are actually used in their global analysis, the ones used in Illinois are Chicago, Peoria and Moline. The problem arises in that there are two stations possible for Chicago (O’Hare and Midway) and that each only provides a partial record, while for Moline the data also exists for only part of the period, though there is a nearby station at Davenport that might provide equivalent values when Moline wasn’t there. To work out what correction might be needed I looked at the differences between the two sets of stations:
As I commented at the time, there is a very slight trend line increase in difference with time for Chicago, but it is not great, and the average temperature difference in Chicago is that Midway is 1.26 deg F warmer, and Moline is 0.97 deg F cooler than Davenport. It is a bit of a kludge to put the two together, but with that adjustment to allow using a full set of data for the 3 GISS stations, the difference between the GISS stations and the USHCN stations using the homogenized data showed a fairly consistent (no significant statistical change over time) 1.13 degrees lower for the GISS sites. Which is not surprising at least as regards the difference, really, given that the 3 GISS stations are all located in the Northern part of the state.
Turning to the TOBS data, however, there is a consistent increase in the GISS temperatures, relative to the USHCN.
This wasn’t there with the homogenized data, and since the GISS data has remained the same this suggests that homogenizing the data introduces a warming trend that was not there in the original data.
Looking at the TOBS average values for the state over the interval,
The annual increase in temperature is very small (0.3 deg per century) relative to that found with the homogenized data, which was 0.9 deg per century. This is fairly significantly less than that reported for the entire county, and the highest temperatures were 1921 and 1931, with the dust bowl years having consistently higher temperatures than today.
It was at this point back in June that Stuart Staniford commented on my use of statistics. So it is perhaps appropriate that I should explain a little more of what I am doing, and why. Most of my life was spent in finding how to understand and use high-pressure water in removing material. And while this can be done today, I suppose, with sophisticated computer models, back when I started a calculator with 9 memories and a printer cost $3,000. Thus, in order to analyze what caused the behavior of the jet to change, I used to run rather large full factorial experiments. I would change a number of parameters within the experiments, and the first thing one did, in the analysis (which got a whole lot more intricate downstream), was to find out which parameters caused a greater change in result. It was using very simple first order correlations to see what was most important (and incidentally it showed that flow rate can be more important than pressure – to the chagrin of some of my colleagues). I am fundamentally, as I acquire the data around the country, doing the same thing again. Trying to see what factors most control local temperatures.
And the most obvious is where the place sits between the North Pole and the Equator. Looking at the data from each of the states, the latitude of the station strongly controls the relative temperature at that site.
The correlation here is so good, that were I to get it in a lab test, I’d go back and check to make sure there wasn’t some hidden bias somewhere, but apart from the high level of correlation, it is consistent with what I have found elsewhere. It does make me wonder though if this is because of the relative flatness of the state.
Looking at longitude, without any significant mountains on either side, there isn’t any clear trend over the state:
However, and this was the surprise when I first looked at the state, when one looks at the effect of elevation change, this is very evident, and that does not change when one uses the TOBS data.
There are thus two factors that need to be considered when estimating average temperatures, based on station data, one is the relative latitude of the average point relative to that of the stations contributing to that average, and the other is elevation. I make this point, which is to a degree intuitive, since basing the average temperature of Illinois on the three GISS stations would, without such corrections, give an inaccurate estimate of the average state temperature.
The second surprise that I found in the original data, related to the effect of population on temperature (essentially a look at the Urban Heat Island effect). Until now in the less populous states, where there were many stations with few inhabitants there was a clear log relationship.
In Illinois, where the majority of the stations were located in larger communities, this correlation is not as clear.
Most of the states as I move from Missouri east are going to have the same sort of population distribution, so it should be interesting to see how this all works out.
Posts
