Font Size: a A A

Mesoscale variability and Lagrangian statistics in the tropical North Atlantic

Posted on:2010-11-04Degree:Ph.DType:Dissertation
University:University of Rhode IslandCandidate:Zhou, LongFull Text:PDF
GTID:1440390002981579Subject:Physical oceanography
Abstract/Summary:
Daily sea level anomaly (SLA) maps during a period of three years (2003-2006) are computed from multiple satellite altimeter along-track measurements via objective analysis with a specified data-derived space-time correlation model. The mapping results are verified with in-situ hydrographic data and have optimized resolutions. To provide background information on dispersion studies at surface and depths, in the same domain as the Lagrangian Isopycnal Dispersion Experiment (LIDEX) in the tropical North Atlantic, the mesoscale variability as well as statistical properties of coherent eddies are investigated based on geostrophic velocities derived from SLA field. The spatially and temporally averaged total EKE derived from SLA is caused by primarily coherent eddies (60%) and currents (24%).For the particle dispersion at the ocean surface in the tropical North Atlantic, we use the stirring velocities of the mesoscale eddies derived from satellite altimeter measurements to determine particle motions. A kinematic numerical model with a daily time step is used to simulate the horizontal dispersion of advected particles by the two-dimensional eddy flow field within 1°N to 13°N and 18°W to 38°W. The altimeter-based simulation is validated with 2 in-situ datasets of surface drifting buoys.A standardized method is proposed to construct mean flow field and decompose turbulent eddy velocities from the mean flow based on optimal numbers of independent observations for all Lagrangian datasets. The Lagrangian statistics of turbulent flow field are determined for all three datasets. It is found that the turbulent diffusivity is dominated by the mesoscale eddy stirring process.Spatial variability is investigated in two sub-regions of the study area. Both numerical and in-situ statistics show anisotropy features such that zonal component is stronger than the meridional counterpart in every sub-region and across the domain. Temporal variability of the mesoscale turbulent field reveals a seasonal cycle in the Lagrangian statistics. Based on 900 dispersion experiments of simulated mesoscale eddy stirring, the absolute diffusivity can be scaled in a linear relationship as the turbulent velocity variance times 4.4 days (zonal) and 2.4 days (meridional). The results consistent with other studies at the sea surface.
Keywords/Search Tags:Tropical north, Lagrangian statistics, Mesoscale, Variability, SLA, Surface
Related items