The 3-D structure and processes of the atmosphere are relatively well known thanks to a dense surface observation network and imaging as well as profiling satellite sensors. The situation is somewhat different for the oceans. Here routine observations at the same location have been limited to ships or moorings and satellite sensors usually give only information about the ocean surface. During the last decade drifters and floats have greatly increased the data density which is, however, still small compared to land.
One way to get along with this caveat is to combine data of a numerical model with observations, i.e. to perform an ocean re-analysis, for an improved understanding of the ocean and for improved initialisation of coupled global climate models. Unlike for the re-analysis of the atmosphere, where the observations are mainly used to get the model results closer to reality, the character of an ocean re-analysis is different: Here the numerical model is used to carry out a model-based synthesis of the observations to fill gaps in the observation data and to obtain a consistent 3-D picture of the oceans.
The Ocean Synthesis Directory / EasyInit gives an overview about these activities.
Organized by CLIVAR GSOP and GODAE OceanView, the Ocean Reanalysis Intercomparison project (ORA-IP) was initiated in 2011 to facilitate a first step toward evaluating ocean reanalysis products by estimating the noise from the ensemble spread and to gain insight into ocean variability by considering ensemble means. The repository here makes the data used for the inter comparison available to everyone >>