Search This Blog

Tuesday, September 10, 2013

An independent confirmation of global land warming, not using temperatures from meteorological stations

It has been hypothesized in some scientific publications (e.g., Pielke et al, JGRA, 2007, doi:10.1029/2006JD008229) and sometimes outright asserted (by AGW "skeptics" or in some media outlets; e.g., here, here, and here) that temperature analyses like GISTEMP, HadCRUT, or NCDC were not reliable, and the global warming, seen in these analyses over the last century, using direct temperature measurements from meteorological stations was largely just an artefact of faulty measurements, data adjustments (homogenization), the urban heat island effect, and other factors.

Gilbert P. Compo, Prashant D. Sardeshmukh, Jeffrey S. Whitaker, Philip Brohan, Philip D. Jones, and Chesley McColl have just published a new study (for the abstract: Compo et al., GRL, 2013, doi:10.1002/grl.50425), where the land near surface air temperature is derived using a state-of-the-art data assimilation system (20th Century Reanalysis). The temperature is calculated from other variables (surface pressure) and parameters in the reanalysis. The measured land surface air temperature is not part of the input data. Thus, any possible contamination of the measured land near surface air temperature by other factors than the state of the atmosphere does not have any effect on the land near surface temperatures calculated in the reanalysis system. The correlation between the land near surface temperature, calculated from other variables in the data assimilation system, and the land near surface temperature data series, derived using direct measurements over land is very high. This increases the confidence in that the statistically significant global warming trend of the atmosphere near the surface, seen in the GISTEMP, HadCRUT, or NCDC analyses over the last century is real, and those data sets can be used as reliable references for other scientific analyses.