Reanalysis is the reprocessing of archived observations through a fixed forecast model and data assimilation system to produce a consistent gridded record of past conditions.12

ECMWF says reanalysis combines model data with observations from across the world into a globally complete and consistent dataset using the laws of physics.1 NCAR’s Climate Data Guide adds that reanalyses are created with an unchanging assimilation scheme and models that ingest all available observations every 6-12 hours over the period being analysed.2

Sources agree that the record is not free of artefacts. NCAR warns that the changing observation mix can produce artificial variability and spurious trends, and the ERA5 documentation states that spurious changes will still be included in the reanalysis because of changes in the observing system.23

In SWEET, Reanalysis is declared equivalent to Reanalyze, which is a subclass of DataAnalysisService.4 That is a structural statement and carries no definition text. Retained sources describe atmospheric reanalysis, so they do not establish how the term applies to ocean, land or other Earth system products, and they give no accuracy figure for any reanalysis.

Footnotes

  1. ECMWF, “ERA5: the new reanalysis of weather and climate data”. https://www.ecmwf.int/en/about/media-centre/science-blog/2017/era5-new-reanalysis-weather-and-climate-data ↩ ↩2

  2. NCAR Climate Data Guide, “Atmospheric Reanalysis: Overview & Comparison Tables”. https://climatedataguide.ucar.edu/climate-data/atmospheric-reanalysis-overview-comparison-tables ↩ ↩2 ↩3

  3. ECMWF Copernicus Knowledge Base, “ERA5: data documentation”. https://confluence.ecmwf.int/spaces/CKB/pages/76414402/ERA5+data+documentation ↩

  4. ESIP SWEET ontology (GitHub mirror), “reprDataServiceAnalysis.ttl”. https://raw.githubusercontent.com/ESIPFed/sweet/master/src/reprDataServiceAnalysis.ttl ↩