re: #228 Cineaste

You didn’t bother to read the article did you? They are very plain about the adjustments. The temperature monitoring stations were moved at various points - some were closed and others were opened. These scientists, when they looked at the data, realized that the temperatures being recorded were not recorded at the same elevation. The oldest block of temperatures were recorded near sea level, the next chunk were recorded 124 meters above sea level. Temperatures drop when you rise in elevation so you have to offset the data (either slightly lower the old data or slightly raise the new data). They calculated the amount of the offset by looking at the average variation during overlap years. For several years both monitoring stations were open and they compared the data from each and saw the average difference and applied that to the old data set.

Actually I did read the article and there’s no need to sneer at me. I said that I understand adjustments. I have an analytical career and do similar calculations all the time. However, I am not seeing evidence of a one-off adjustment based on moving a station, but rather what appears to be a complex formula that tends to adjust old temperatures down and new temperatures up. I also am going to have to look for what was used as a control in order to properly ground the difference between 0m and 124m, or whatever the move(s) were.

You can’t just move the station, see a drop in temperatures, and subtract it out (which clearly NIWA has not done, as I said the formula is more complex). What you do need to do is leave something behind at 0m to establish a forward baseline against which to evaluate the new, 124m station, otherwise you can’t distinguish between the changes due to the move and actual changes in temperature. I presume something like this was done, but I haven’t read anything about it yet. I would be concerned if the adjustments were made simply based on some global constant linking the two altitudes, this would ignore all kinds of local effects.

I suppose the systematic nature of the adjustments could be due to station moves generally being from lower to higher altitudes. I don’t know enough about them to speculate on why this would be the case, but it might explain a lot of the variation (and the adjustments would be completely defensible assuming the existence of proper controls).