Waldo, Hansen’s ROW trend, is not in Africa nor in Antarctica. Is Waldo in South America?
Here we are referring to a trend calculated according to the stated methodology of Hansen et al 1999, 2001, in which urbanization effects are supposedly removed by coercing trends to trends of rural stations. For that purpose, I am assessing the availability of rural stations which go back to the 1930s – thereby permitting a comparison of 1934 to 1998 in these other regions under Hansen’s adjustment – while also having sufficient information to permit a 1961-1990 normal to be calculated.
South America has a land area of 17.8 million km2 (about 12% of the world’s land surface) and is over twice the size of the contiguous 48. There are 7 stations with records that include all the 1930s, shown below together with their anomaly average. I noticed that 6 stations start in 1931 and I show them below separately.
Three of the stations are islands: Isla Juan Fernandez, Isla Huafo and Grytviken (South Georgia) – only 4 on the mainland. The first two (columnwise) are in Argentina, Q is in Brazil, the next 3 in Chile and the last is again South Georgia. Punta Arenas is quite far south (53 S) and provides an interesting counterpoint to the shorter records from the Antarctic Peninsula, showing strong warming since the 1940s.
Obviously several of the records have major discontinuities. We know from the U.S. HCN analyses that very slight changes in station location can create biases that Karl needs to adjust through his TOBS adjustment, MMTS adjustment, station history adjustment etc. and that in the U.S. these station adjustments are not small (actually a collation of USHCN adjustments needs to be made to get a distribution to specify the distributions that might be at work when one ventures to other countries. Is there a possibility that, for example, some discontinuity occurred at Isla Huafo or Quiaca during the decades long interval between measurements? Seems to me to be at least a possibility? If one is taking an average, this becomes relevant because the later values of Quiaca show a step increase towards the end not apparent in the continuous series – which gets reflected in the average.
A few posters are criticizing me for apparently focusing on areas with poor coverage. I don’t think that anyone can accuse me of not discussing American stations in sufficient detail – so this is a little unfair. I’ve posted extensively earlier in the year on China and Russia and passim on Australia. I’ll probably get to them again. But there’s only so much that I can do. If a critic feels that some area with good records is being given short shrift because I haven’t got to it yet, please feel free to write a post of detail equivalent to mine and I;ll post it.