During the early1960s global or at least NH analyses became available for systematic scientific investigations of the general circulation of the atmosphere. The analyses were mainly based on operational numerical analyses produced by the meteorological services. Aksel Wiin-Nielsen and his co-workers at NCAR were probably the very first to undertake general circulation studies by using analyses from JNWP later NMC. An interesting early result was the finding that the general circulation can operate in two modes, either transporting kinetic energy up-gradient or alternatively, but less common, down-gradient. Following the huge improvements in modelling and data-assimilation the operationally produced analyses have undergone a colossal development creating exiting research possibilities.

However, because of the ongoing changes in the operationally produced analyses it was found necessary to redo the NWP calculations using a procedure whereby both the model and the data-assimilation were kept fixed. Moreover in the re-analysis additional observations have become available and observations are more carefully checked. Such data sets at high temporal resolution covering several decades have been produced by ECMWF and by some of the major national meteorological services. The re-analyses are now central research tools and are also being investigated to explore climate variations in a systematic way. Recently, a global data set, the NOAA-CIRES 20th Century Reanalysis (20CR), from 1871 until present time was released with analyses for every 6 hour, an undertaken that in many ways is an extra-ordinary achievement. Remaining difficulties are related to the large changes in observational data and data coverage in particular the changes after WW II with the general introduction of upper air data and in 1978/79 with the commence of a global space based observing network. The talk will put the re-analysis in a long-term perspective and outline some of the main opportunities and challenges for future work