Observation and model resolution implication to ocean prediction

We must reconcile ocean modeling capability growing exponentially while ocean observation density has not maintained pace (1) leading to seemingly degraded forecast skill when model resolution is increased (2). Ocean forecasting skill requires satellite and in situ observations continually correcting numerical model conditions(3). Observations constrain positions of larger ocean model features, while smaller features are unconstrained. We show the separation of constrained and unconstrained features as a function of spatial scale and demonstrate observation density controls the boundary between scales. Constrained scales have deterministic skill in predictions, and unconstrained scales provide skill in predicting areas of higher expected errors. Separating scales allows us to reconcile the issue.