Posted by drj | Filed under Uncategorized
More or less on a whim I split the GHCN data into two sets: Those stations marked as being at an airport; those stations not marked as being at an airport. This is easy to do because the v2.inv file puts an ‘A’ in column 81 (counting from 0) for airport stations.
Here’s the airport versus non-airport comparison for ccc-gistemp:
Certainly for the most recent 50 years it doesn’t seem to matter much whether you use exclusively airport based measurements or exclude airport based measurements (considering the global anomaly).
My earlier post about the 1990s station “dropout” used a similar technique of splitting the input data into two sets.