Ocean Data Synthesis
The key objective of the ocean observing system for climate is to characterize the state of the ocean in adequate detail to monitor and understand climate variability and change in the ocean as well as its interaction with the atmosphere. To this end the observing system produces a rich collection of datasets, which, when combined with space-based observations, yield extensive coverage of the ocean surface. Yet despite the large number of subsurface measurements made by profiling floats, XBTs, repeat hydrography, and other platforms, the vast interior of the ocean, which cannot be seen directly from space, remains largely unsampled. It is also difficult to cross-calibrate thousands of in situ measurements with adequate accuracy to ensure the consistency among them desired for robust conclusions about the state of the complete ocean.
In contrast, computer models are capable of describing all of the ocean in consistent fashion, though their imperfections lead them to do so with limited accuracy.
To address the limitations of both models and measurements, scientists have developed methods for continuously correcting model errors with measured data. Such data assimilation methodologies not only keep the model from drifiting away from the truth, but also utilizes the model to interpolate between measured data points, thus effectively increasing the spatial and temporal resolution of the measurements. Doing so also corrects any inconsistencies among measured data points by allowing the model to derive a representation of the state of the ocean that improves upon estimates based solely on either sparse observations or numerical simulations alone. In this manner, the data assimilation model serves as an integrator of disparate observational datasets, knitting them together in a self-consistent manner for subsequent analysis of ocean climate phenomena that may be difficult to infer from observations alone. For example, the model can be queried for ocean heat transport in particular regions, or the model can be used to evaluate the observing system itself by withholding individual observations from the data assimilation scheme and noting its impact on calculated quantities.
To these ends the Ocean Climate Observation Program sponsors the operational Global Ocean Data Assimilation System (GODAS), developed and operated by the National Weather Service Climate Prediction Center, which creates and disseminates a suite of ocean products derived from synthesis of many of the datasets collected by the observing system.
Improvements in data assimilation methodologies are a focus of numerous research groups worldwide, many of whose activities have been coordinated under the Global Ocean Data Assimilation Experiment (GODAE).