Reanalyses Suitable for Characterizing Long-Term Trends

Abstract

Reanalyses are, by a substantial margin, the most utilized climate data products, and they are applied in a myriad of different contexts. Despite their popularity, there are substantial concerns about their suitability for the monitoring of long-term climate trends. This has led to calls for a truly ?climate quality? reanalysis that retains long-term trend fidelity. The authors contend that for such a reanalysis to be achieved, a substantial rethinking of the current strategy for producing reanalysis products is required. First, the problem must be defined clearly. Second, the methodology that is employed must be reconsidered so as to minimize potential non-climatic artifacts and robustly ascertain the inevitable residual uncertainty. Finally, a set of validation data and metrics must be constructed that the community can use to compare and unambiguously assess the claims of climate quality. The purpose of this essay is very much to initiate discussions to this end rather than to prescribe solutions.