GFDL - Geophysical Fluid Dynamics Laboratory

Reanalysis Ideas — Current Limitations

Some Ideas For Future Reanalysis Efforts

John Lanzante

Current Limitations

  • The LKS adjusted data are limited to 87 stations for the nominal time period of
    1948-1997. The adjustment process is tedious and very time consuming, with a
    subjective component. It is not practical to apply this methodology to other
    stations or even to extend the existing ones beyond 1997. There is no other
    proven, general-purpose method which can be applied to radiosonde station data.

  • It is very likely that LKS did not adjust enough (i.e., failed to adjust enough
    stations/vertical levels) so that the spurious cooling bias is even larger than
    estimated. This statement is supported by three independent pieces of evidence:

    1. By design, LKS selected only the most obvious potential inhomogeneities
      for consideration; more ambiguous cases were not scrutinized.

    2. Even after homogeneity adjustment, the LKS radiosonde temperatures
      have a non-trivial spurious cooling trend relative to the
      Christy/Spencer MSU data.

    3. The new Wentz et al. MSU temperature products have a considerably
      larger warming trend than the comparable Christy/Spencer products.

  • In summary, current radiosonde temperature time series (LKS as well as
    all other such products) are sub-standard both in their own right for
    climate monitoring and for use as input to the next~ 50-year Reanalysis.

Back to the main page