Efficient objective analysis of dynamically heterogeneous and nonstationary fields via the parameter matrix

We develop a generalized approach for the objective analysis of nonstationary, heterogeneous fields. An algorithm is presented that uses an anisotropic, time-dependent correlation function with correlation parameters that vary in space/time and a time-dependent trend surface for efficient objective analysis of dynamically heterogeneous and nonstationary fields. The algorithm, which we term the “parameter matrix algorithm”, is applied to two data sets. The first is tropical Pacific sea surface temperature (SST) derived from satellite AVHRR data and Pan-Toga drifting buoys. The SST appliclication illustrates how the parameter matrix is used for the computationally efficient objective analysis of the tropical Pacific SST from 30°S to 30°N at 0.2° resolution (over 290,000 grid points) using approximately 350,000 data points from 12 2-day satellite SST composites. The second example uses data from the Anatomy of a Meander/ BIOSYNOP experiment in the Gulf Stream ring and meander region and illustrates that an objective analysis using the parameter matrix can yield a more accurate representation of oceanic features than typical objective analysis techniques.