Optimum sampling designs for a glider-mooring observing network
Ocean observing systems (OOSs) constituted by moorings and gliders are becoming relevant in oceanographic and climate studies. In these observing networks, the temporal variability is captured by mooring observations, while the spatial variability is obtained from gliders sampling in the surrounding area. The advent of this observing capability brings the need to find optimal procedures to sample a given ocean region with a glider in the presence of a neighboring mooring, in order to maximize the information content of the data collected by this observing network. Different criteria (e.g., A, G, or E optimality) commonly used in the geosciences to obtain an optimum design lead to different sampling strategies. The question of which criterion performs better for optimal design in the marine environment remains open. This work investigates optimal procedures to sample a given ocean region with a glider in the presence of a mooring. Specifically, observing systems simulation experiments (OSSEs) are carried out in the Ligurian Sea (western Mediterranean) in August 2010 to study the different sampling strategies. Three criteria, which respectively aim at minimizing the trace (A optimal), maximum diagonal value (G optimal), and maximum eigenvalue of the error covariance matrix (E optimal), are considered. The resulting temperature field estimations are evaluated against a control field at 50-, 100-, and 150-m depth. The results indicate that the most appropriate strategy for environmental characterization using gliders employs the A optimal criterion, minimizing the mean uncertainty over the study area.
SourceIn: Journal of Atmospheric and Oceanic Technology, vol. 29, 2012, pp. 601-612.