Anisotropy and Inhomogeneity in Drifter Dispersion. Issue 12 (8th December 2019)

• Record Type: Journal Article
• Authors: Huntley, Helga S.; Lipphardt, B. L.; Kirwan, A. D.
• Is Part Of: Journal of geophysical research. Volume 124:Issue 12(2019)
• Page Start: 8667
• Page End: 8682
• Publication Date: 2019-12-08
• Subjects: Oceanography -- Periodicals; 551.4605
• DOI: 10.1029/2019JC015179
• Languages: English
• ISSNs: 2169-9275

Abstract

Abstract: Ocean flows are known to be locally anisotropic and inhomogeneous. Nonetheless, the ocean's statistical dispersion properties are traditionally assumed to be isotropic and homogeneous. Here, we investigate the effect of local anisotropy and inhomogeneity on dispersion statistics, using a unique data set of roughly 300 near‐surface drifters that were launched within 10 days in the summer of 2012. The unique launch strategy based on nested triplets resulted in an unusually large number of nearly colocated drifter pairs. Thus, this data set is ideally suited for an estimate of the directional bias and inhomogeneity effects inherent in drifter pair statistics. Several metrics are proposed to assess the time evolution of anisotropy and inhomogeneity effects at multiple initial separation scales (100 m, 200 m, 500 m, 1 km, and 10 km). Locally, statistically significant anisotropy and inhomogeneity are observed at all scales, although anisotropy is noticeably less at 10 km, suggesting that the energetic processes driving anisotropic dispersion operate primarily at smaller scales. Moreover, averaged over a sufficient variety of different flow regimes, the signature of both anisotropy and inhomogeneity in dispersion metrics lessens. These trends hold generally across all scales, with longer time scales associated with larger spatial scales. The results indicate that oceanic dispersion is statistically isotropic and homogeneous over large swaths, but for an application in a specific location, local anisotropy and inhomogeneity matter. What size swath is large enough is situation dependent: For this specific data set, statistics had to be evaluated over multiple deployments, giving a required area greater than 150 km 2 . Key Points: Local anisotropy is observed across all scales in the initial evolution of pair separations Averaged over multiple deployments, the observed local anisotropy tends to average out Local inhomogeneity is observed in pair separation statistics for at least a week