The development of new approaches to the creation of bias-corrected land surface air temperature datasets has the potential to improve the interpretation of  historical surface temperature observations. This is a problem of high scientific and societal relevance as the temperature record forms a backbone for the scientific characterization of climate variability over the past century. It is also a challenge  that is inherently statistical / mathematical in nature and where input from these communities is of paramount import. This summer program, jointly sponsored by the Statistical and Applied Mathematical Sciences Institute (SAMSI) and by the Institute for Mathematics Applied to Geosciences (IMAGe) at the National Center for Atmospheric Research (NCAR), will bring together experts from the disciplines of climate science, statistics and applied mathematics with the goal of testing, comparing and perhaps extending methods. Applications will be considered from researchers in any of these disciplines, including junior researchers and graduate students.

To allow financial support to be extended to a larger number of participants, we are asking participants, where possible, to fund their own travel. Additional financial support will mainly be awarded for local costs in Boulder (lodging, per diem) although we do anticipate a limited number of travel awards for participants without other means of financial support.

Most of the program time will be spent in small groups working on the development and application of a number of novel potential techniques identified by program participants to the homogenization of regional and global series. Plenary sessions will enable critical feedback from all participants on a regular basis. External parties will be encouraged to provide input through appropriate social media, blogging or other facilitation routes including e.g. efforts at distributed coding.

The program will leverage the framework offered by the International Surface Temperature Initiative ( Improved data holdings of monthly temperatures totaling almost 32,000 station records from around the globe will be provided. The participants will also have available a number of benchmarking datasets that exactly mimic the data availability of the real data holdings and have realistic inter-station noise and variability.

Map of station locations that will form the basis for analysis. Color denotes period of record and longer records overplot shorter records. Further details at


Registration and application for financial support closed on May 1. However if you are interested in participating please send an email to image-events@mail.ucar. edu including your name and affiliation.

If you need further information please send an email to