11 December 2009
More on Weather Station Unreliability
Joseph D'Aleo, AMS Fellow, CCM has posted an interesting piece called How bad is the global temperature data? at Watts Up With That? Among the problems, see what has happened to the number of reporting weather stations in the NOAA GHCH, which we now know to be the source of the raw temperature data for NASA's GISS and for Hadley's CRU:
Given that the rural stations are the only ones perhaps not suffering from heat island effects, the greater fractional decrease in rural stations will certainly bias the land surface temperature record upward. The last thing that should be happening is a decrease in the number of rural stations. A decrease in the number of urban and suburban stations is good. It is good that there are more suburban than urban stations now, but this is hardly able to make amends for the severe lack of rural stations, especially given that a much larger fraction of the Earth's surface is rural than either urban or suburban. This higher weight of urban and suburban stations is likely to bias the temperature upward strongly.
D'Aleo also points out that the Pipogenus Dam weather station site in Maine, which had been reporting data since 1925, was officially closed in 1995. However, the NOAA USHCN database of land surface temperature data reports data from the date of closing to at least 2006. Not surprisingly, the data after 1995 averages higher than the data before the station was closed! See that data below:
The survey of the climate stations in Maine found that everyone of them had microclimate or urbanization biases of the temperature data.
As I have noted before, the land surface temperature data is simply not reliable.