[9-1] Invited

Data assimilation for ocean climate studies

James A. Carton (1), Gennady A. Chepurin (1), Steven G. Penny (1,2), and David Behringer (2)

(1) University of Maryland, College Park, MD USA
(2) NOAA/NCEP, College Park, MD USA

The atmosphere is a rapidly moving fluid with Rossby numbers approaching unity and Rossby radii approaching basin scales. Much of atmospheric variability results from internal flow instabilities. Surface boundary conditions of topography, heat, and water flux provide loose constraints while diabatic processes are major energy sinks and sources. In contrast, the World Ocean is a slowly moving fluid maintaining generally small Rossby numbers and with Rossby radii an order of magnitude smaller than basin scales. Much of the variability of the ocean mass and momentum fields is introduced through momentum exchanges at the surface, while internal mixing rates are generally low. Internal instabilities are important only in geographically limited regions such as western boundary currents. Flows in the ocean are also much more severely constrained by basin topography, and particularly by key passages such as the Straits of Gibraltar. These fundamental differences between atmospheric and oceanic dynamics as well as between their observing systems need to be accounted for by changes in data assimilation systems. Yet the history of data assimilation applied to the ocean is mainly one of borrowing techniques developed for the atmosphere. The first part of this talk is a more detailed discussion of the differences that give rise to differences in assimilation/estimation in these two fluids.

The second part of this talk discusses several ocean data assimilation development issues in the context of a new generation of the Simple Ocean Data Assimilation (SODA). I will focus on three issues. The first issue is an effort to address deficiencies in forecast error modeling through implementation of Penny's 2014 hybrid ensemble Kalman Filter. Results from observing system simulation experiments will be presented. The second is to address biases introduced through changes in the observing system and due to deficiencies in mixing processes. Here also results from observing system simulation experiments will be presented. The third is to specifically focus on problems associated with estimating global sea level in the era before satellite altimetry was available.