---

Starting in the early 1990’s, in collaboration with several other folks, I worked on the problem of long-term temporal homogeneity of radiosonde station time series in the context of studying climate variability. The concern is the introduction of artificial inhomogeneities, particularly abrupt jumps, in climate series when instrumentation or measurement practices are changed.

Some progress (see below) has been regarding temperature data. While we’ve gotten some encouraging results, we’ve reached somewhat of a impasse as far as taking it to the next step. Below I’ve sketched out some ideas aimed at attacking this problem using an analysis/assimilation scheme. Success in this endeavor could be benefitial towards improving both station time series as well as future Reanalysis. These ideas are not well-refined but are meant as a starting point for discussion as to their feasibility.

• The Problem
• Some Progress
• Current Limitations
• A Path To A Solution?
• A Strawman For Homogenization?
• Questions
• New Ideas (Jan 2012)