Work Package 11: Model performance assessment and initial fields for scenarios

Leader: Christoph Heinze

Objectives
- To provide interfaces between the carbon cycle climate models and observational data sets to be
compiled in CARBOOCEAN.
- To determine, how well biogeochemical ocean general circulation models (BOGCMs) are able
to reproduce carbon cycle observations from the real world with respect to temporal and spatial
distributions.
- To define criteria for model goodness in form of a cost function, which can serve as benchmark
for model/data intercomparison to trace the model improvement throughout the project.
- To establish initial conditions for future scenarios with the BOGCMs and a quality check of
these initial conditions.

Description of work
- Observational data are reformatted to interface the model results (link to WP 1 and WP 19)
(partners 1, 3, 6).
- Model distributions are interpolated onto the measurement position in space and time and the
difference between observed and simulated data analogs is determined (partners 1, 3, 6, 11, 33,
34).
- Suitable cost functions are developed which serve as a measure for the goodness of the
simulation with respect to observations (least squares, bias etc.). (partners 1 and 3)
- The goodness of present day BOGCM simulations is determined for the BOGCMs which will
later on be used in the future scenarios and the goodness indices for these models are
intercompared. (partners 1, 3, 6, 11, 33, 34).