Skip to content

Climate Data Assimilation

Contacts, for more information:

Due to insufficient observations and an incomplete understanding of physical processes, climate models always contain some biases, and they may produce climate features and variability which are different from the real world. For understanding climate variability and predictability on seasonal-interannual to decadal scales, GFDL scientists use coupled model dynamics to extract observational information from the earth observing system and reconstruct the historical and present states of the earth climate system.

Climate Data Assimilation optimally integrates pieces of observational information and produces a balanced and coherent climate estimate and prediction initialization by maintaining the instantaneous flux exchanges among the coupled components. Optimizing model parameters by using observations through coupled data assimilation is expected to mitigate model biases and enhance model predictability.

GFDL’s Climate Data Assimilation (CDA) uses global [comprehensive] climate models to interpret a broad array of Earth observations, in order to generate detailed, accurate, and physically-consistent estimates of the state of the global ocean, atmosphere, land, and sea ice. GFDL continues to advance the state-of-the-art in CDA — using its climate reanalyses to evaluate next-generation models, initialize and evaluate climate predictions, and to inform scientific research on climate variability and change.

Related Links

Research Highlights