Impact of spatially-dense in-situ observations on ocean forecasts of mixed layer and thermocline depth

Abstract

ABSTRACT Ocean models, such as the Navy Coastal Ocean Model (NCOM), rely on ocean observations through data assimilation to predict the evolution of ocean phenomena accurately. The accuracy is highly dependent on the type and quantity of observations available. This work attempts to examine the impact of spatially dense in-situ profile observations on the forecasts of mixed layer depth (MLD) and thermocline depth (TD), using a three-dimensional variational data assimilation algorithm. To do this, a twin data assimilation experiment is used to simulate the impact of float deployments on the forecast of MLD and TD in a specific area of interest (AOI) of the North Atlantic. It is found that during the northern hemisphere summer the model MLD is sensitive to the sea surface temperature (SST) observations rather than the profiling floats. TD is more strongly impacted by the assimilation of float observations, and the experiment with the most floats produces the best results. During the northern hemisphere winter the profiling floats have a greater impact on the model simulation of MLD. Finally, the overall lack of substantial improvement in the winter months for the highest density of profiling floats is determined to be due to the data assimilation methodology itself.