Temporal Sampling: How Many Sections Are Needed to Quantify the Mean Transport and Structure of a Meandering Current?

Abstract

Two years of observations from an array of 16 inverted echo sounders deployed south of Australia near 51°S, 143.5°E are combined with hydrographic observations from the region to estimate the differences in baroclinic transport, as well as temperature and velocity structure, that result from trying to estimate the true mean using a limited number of snapshot sections. The results of a Monte Carlo?type simulation suggest that over a 350-km distance, completely spanning the Subantarctic Front (SAF) at most times, a minimum of six temporally independent sections would be required to determine the baroclinic transport mean (surface to 4000 db) of the observed 2-yr period to within an accuracy of 10% when the sections are averaged in either an Eulerian (geographic) or stream coordinates manner. However, even with 10 sections during a 2-yr period the details of the mean velocity and temperature structures obtained can be quite different than the ?true? 2-yr mean structure, regardless of whether the sections are averaged in either Eulerian or stream coordinates. It is estimated that at least 20 independent sections would be required during a 2-yr period in order to determine the baroclinic velocity structure to within an accuracy of 10%, irrespective of whether they are averaged in Eulerian or stream coordinates. Implications for future sampling strategies and for inverse modeling analyses are discussed.