2011-01-26 06:01:00GISS weather data: “homogenization”
John Hartz
John Hartz

For the past couple of days, I've been engaged in a marathon blogging session on the comment thread to Dlingpole's "Oh no, not another unbiased BBC documentary about 'Climate Change'…" posted on The Telegraph.


One of the Anti-AGW bloggers referenced his personal blog site wehere he has posted a number of "homogenized" graphs of GISS weather data for specific locations around the world.


What does this guy mean when he says he has "homoginized" the data?

Is the term "homoginzed" commonly used by statisiticans? 

2011-01-26 06:42:58Info
Julian Brimelow


IIRC, climatologists "homogenize" the station temperature data for a number of reasons.  Unfortunately, obtaining decadal or longer record from a sigle site is actually quite difficult.  Observation times change, instruments change, instrument sheters change, the natural and anthro landscape surrounding the stations change.

Usually most of these changes (not all) are discernible in the record by using statistical techniques and they can then be corrected for-- this process is called data "homogenization".

Now, and this is the important part, to do the homgenization properly you need to have access to the metadata and records detailing when the changes were introduced.   Your "friend" probably does not have access to this information to homogenize the data properly.  The GHCN has already done this, as have weather agencies around the world, so I am not entirely sure what they mean when they say they have homogenized the data.

As far as I can tell, "homogenization" is not the same as adjusting for the UHI, although doing so may fall under the general umbrella of "homogenization".  Homogenization of the data is not a trivial exercise, ask him if he used the same technique as Menne andWilliams (2009).




And regardless, this is probably all an elaborate red herring.  Show him how incredibly well the satellite, radiosonde and global SATs agree.



[Source: http://www.ncdc.noaa.gov/sotc/upper-air/]


Global radiosonde data show 2010 to be the warmest on record (1958-present).

Maybe John should add those RATPAC data to this graph:





2011-01-26 08:09:39

Guess what skeptics would say if, for example, data were not homogenized after the met station had move from high to low ground. And what would they say if UHI was not taken into account. And what would they say if the time of observation was not taken into account. And what if ... do I need to contnue?