“Spencer: Using hourly surface data to gauge UHI by population density“. Dr. Roy Spencer has realised that “forsaking blindingly technical statistics” isn’t a practical position and has come up with an analysis, using one(!) year of data, that correlates “warming bias” of  station temperature records with population density (Urban Heat Island!). Population density is “presumed to be related to how much the environment around the thermometer site has been modified over time” (emphasis mine). That’s a rather big presumption. There are plenty of other simplistic adjustments in Dr. Spencer’s data, such as a blanket 5.4°C per 1000m increase in station elevation adjustment. Dr. Spencer has shown before that he has problems using statistics correctly, so it will be interesting to see if this stands up, but given Dr. Spencer’s track record I’m going to bet on “confirmation bias“.

Dr. Spencer makes a dangerous statement though: “Note that the philosophy here is not to provide the best adjustments for each station individually, but to do adjustments for spurious effects which, when averaged over all stations, will remove the effect…” (emphasis mine). He’s setting himself up as a target for exactly the same weather station correction nitpicking that Anthony has played with the USHNC’s weather station data. Will Anthony hold him to the fire?

Anthony inadvertently answers this question with “I believe this is a truly important piece of work” in spite of all dodgy assumptions Dr. Spencer admits to. Also, his blogging about Spencer’s draft should be considered as an “early peer review.” I will support Anthony’s hope that “Dr. Spencer will submit it to a journal” though. I think I’ll enjoy watching what is effectively another “sciency” attempt at Anthony’s failed surfacestations.org project be subjected to legitimate scientific scrutiny.