Future needs in atmospheric reanalysis

Over the past decade, reanalyses of multi‐decadal series of past observations have become established as an important and widely utilized resource for the study of atmospheric and oceanic processes and predictability Reanalyses, which are produced using modern versions of the data assimilation systems developed for numerical weather prediction (NWP),also are being applied increasingly in many other fields that require a record of the state of either the atmosphere or its underlying land and ocean surfaces. Estimation of renewable energy resources, calculation of microwave telecommunication signal losses, and study of bird migration are just three examples. High‐resolution operational NWP systems provide good quality analyses for timely study of recent conditions, including climatic extremes. However, changes made to improve the operational systems introduce inhomogeneities in time series of operational analyses that limit their utility for studies of longer‐term climate variability and change. Lower‐resolution reanalyses produced using an up‐to‐date data assimilation system provide products for all but the last few years that are generally superior to those available from the archives of past operational products. The reanalysis products are, by design, more suitable than their operational counterparts for use in studies of longer‐term variability in climate, although they remain susceptible to changes in the observing system that can make accurate depiction of long‐term trends problematic.