Recognition of spatial structures and ensemble averages in the benthic boundary layer

Abstract

Simultaneous high‐frequency spatial and temporal measurements of benthic boundary layer (BBL) velocity and sediment concentration generate complex data sets. In order to investigate the various spatial and temporal structures, a pattern recognition method is, for the first time, adapted and used to investigate different and similar structures (patterns) of velocity and acoustically measured suspended sediment concentration profiles in the lowest 2 m of the BBL at Tiana Beach, Long Island. This layer is subject to interactions between differently scaled forcing fields which result in structures difficult to predict using existing time average or boundary layer methods. Under the assumption of no a priori knowledge about either the spatial or temporal structures in the BBL, a deterministic mathematical methodology is implemented to recognize spatial structures in the data and classify them in an optimum number of classes. The spatial pattern analysis in this paper indicates that ∼ 70% of the above mentioned data are represented by only 5–20% of its size. Spatial patterns different from those predicted from theory or experience are recognized.