Project aims

It is essential to estimate the quality of the decadal projections within MiKlip for all involved model components. The VALOCEAN work package will evaluate the oceanic part of the model hindcasts and will aim at a process-oriented validation to enhance the understanding of the processes represented in the model. For the model validation/evaluation a variety of oceanic observational data sets will be assembled and time series of key parameters will be constructed. These key parameter sets will be compared with the hindcast simulations and with the oceanic reanalysis data sets, which are used for the model initialisation.
 

Project structure

It is important for the quality of the MiKlip decadal predictions that the model not only reconstructs the mean state of the ocean but also realistically reconstructs the temporal evolution. Consequently, the comparison of model outputs with observed time series is an important measure in the assessment of the MiKlip decadal prediction system. One of the main tasks therefore is to collect available oceanic observations and integrate them into the central evaluation system (CES). Based on these observations key oceanic indices will be constructed. These indices shall represent various aspects of the ocean dynamics through time and space integrating properties. Such properties are e.g. ocean heat content (OHC), ocean freshwater content (OFC), mixed layer depth (MLD), the Atlantic multi-decadal oscillation index (AMOI) and other indices which represent the strength of the subpolar and subtropical gyre. It is planned to use homogenized and quality assessed data sets from the Argo and satellite observation programmes (SST, SSH) but also historical data like the world ocean data base, as well as time series from moorings (i.e. TAO/TRITON).

Realistic reanalyzes are prerequisite for the initialization of the MiKlip decadal prediction system. Thus, oceanic reanalyses and hindcasts have to be evaluated for their temporal evolutions. The spatial focus will be on key regions such as the subpolar Atlantic and the tropical oceans. The investigations will identify differences and their possible implications for ocean dynamics.

The work in VALOCEAN is divided into four tasks and is listed below.
 

Tasks of the project and deliverables

• Construct OHC and OFC data products based on multi-instrument sources (Argo data, historical measurements with appropriate uncertainty estimates (Milestone M3.1)
• Comparison and evaluation of the reanalysis data sets with observations (Milestone M3.2)
• Construction of time series of oceanic key indices and their implementation in the central evaluation system (Milestone M3.3)
• Examination of spatial patterns and the temporal evolution of multi-decadal variability in observations and model output (Milestone M3.4)

Progress so far

The work started with a vast collection of data from re-analyses (GECCO2, ORA-S4) and climatologies (e.g. Ishii et al. [2003], Levitus et al. [2012], EN4, AVISO). Re-analyses and climatologies were intercompared and evaluated for various "standard variables" such as temperature, salinity and, sea surface height in order to review  the temporal behavior of the ocean syntheses used for the initialisation of the earth system model. To evaluate further the reanalyses against the climatologies, key parameters that characterise the state of the climate system were computed such as: ocean heat content, freshwater content, mixed layer depth and dynamic height. The comparison was done in time and space (globally, regionally and in depth), and led to the identification of areas of high variability. We now focus on the North Atlantic Subpolar gyre (NASPG) region, as one of the key regions to investigate processes that contribute to multi-decadal variability, and their spatial patterns. To that end, we construct a local NASPG climate index and we are currently investigating the dynamical processes associated with the index.