Remote detection of water property changes from a time series of oceanographic data
Abstract. A water mass analysis method based on a constrained minimization technique is developed to derive water property changes in water mass formation regions from oceanographic station data taken at significant distance from the formation regions. The method is tested with two synthetic data sets, designed to mirror conditions in the North Atlantic at the Bermuda BATS time series station.
The method requires careful definition of constraints before it produces reliable results. It is shown that an analysis of the error fields under different constraint assumptions can identify which properties vary most over the period of the observations. The method reproduces the synthetic data sets extremely well if all properties other than those that are identified as undergoing significant variations are held constant during the minimization.