Andrea Steinberger has always been fascinated by the ocean. Her favorite hobbies include surfing, sailing and scuba diving, so it’s no surprise that her love of the water has influenced her undergraduate studies. After taking Physical Oceanography with Dr. Francois Primeau, she started working with him on the relatively new area of forecasting and now-casting the ocean. Since completing this research project, she feels prepared and enthusiastic for her graduate studies in physical/chemical oceanography, which she hopes to begin next fall. Andrea has also been conducting studies for the Southern California Coastal Water Research Project since 1999, and is currently working on acquiring a skipper’s license and becoming a master diver.

A new method for estimating advection and diffusion in an ocean basin from Lagrangian data is developed using a simple box-model approach and the method of maximum likelihood. This method estimates both advective and diffusive flow concurrently, which is not possible using existing methods. Computer simulations of one- and two-box ocean models are used to generate synthetic Lagrangian data on which the method is tested. In these models, true flow parameters are defined then recovered by maximizing the likelihood function of the one- and two-box probability models. The relative likelihood function is used to extract information regarding the consistency of the parameter estimates, while ensemble means are used to determine the method bias. For both the one- and two-box probability models, the method returned flow estimates that were consistent and unbiased in the limit that the number of Lagrangian drifters, N, became large. Limitations to concurrent estimation of the advective and diffusive flows were observed for small N, but were absent for large N. Techniques used in this study show capacity for use in planning Lagrangian drifter studies. The method shows potential for use in optimizing data assimilation in ocean circulation models, which are subsequently used in climate models to study global climate change.

Francois W. Primeau School of Physical Sciences

The positions of thousands of surface and subsurface drifters deployed throughout the oceans are tracked daily in order to monitor the oceans’ circulation. The drifter trajectories, when combined with an ocean circulation model during data assimilation, provide insight into such significant variables as temperature and velocity. Data assimilation methods require the Lagrangian drifter position data to be converted into an Eulerian velocity field on the fixed model grid. An unfortunate result of this conversion is that some information within the data gets thrown away. In an effort to extract more of the useful information, Andrea Steinberger successfully tested a new method, using the drifter data to estimate a sub-gridscale eddy-diffusivity in addition to the large-scale flow field. Focusing on idealized models, Andrea developed an understanding of the limitations and pitfalls of the method’s implementation. Encouraged by Andrea’s results, our group is currently extending this method to more complex ocean models.