[A-CG41] Satellite Earth Environment Observation

[ACG41-16] Quantification of Constrained Scales in the Ocean with an Ensemble Analysis

*Kenneth Andrew Peterson¹, Gregory C. Smith¹, Kamel Chikhar², Andrea Storto³ (1.RPN-EM, ECCC, 2.MSC, ECCC, 3.CNR)

Keywords:Ensemble Prediction, Constrained Scales , Data Assimilation, Altimeter Observations

Numerical models used in operational ocean prediction systems typically resolve finer scales than can be constrained through the assimilation of conventional satellite measurements. This results in unconstrained variability contributing to larger model error. An ensemble of ocean analyses, if correctly constructed, could provide a means to remove uncertainty associated with features having length scales that cannot be constrained by observations.

Here we use an eddy-permitting ocean prediction system to demonstrate that the ensemble mean can be used as a filter to remove unconstrained variability and reduce forecast error. It is demonstrated that the limits separating length scales of constrained and unconstrained variability can vary over the global domain, and that these separation scales are a product of the analysis system, not imposed by lengthscales associated with the ensemble perturbations. A further demonstration is made of how the removal of the unconstrained scales reduce errors in surface currents when compared to drifting buoys.

These findings support the use of ensembles as a means to account for errors due to unconstrained variability found in deterministic ocean predictions.