A New Statistical Modeling Approach to Ocean Front Detection from SST Satellite Images

• DOI: 10.1175/2009jtecho684.1
• Publication Date: 2010-01-15T22:45:50Z
• Authors: Jo Hopkins; Peter Challenor; Andrew G. P. Shaw
• Keywords: Advanced very-high-resolution radiometer; Position (finance)

Abstract Ocean fronts are narrow zones of intense dynamic activity that play an important role in global ocean–atmosphere interactions. Owing to their highly variable nature, both space and time, they notoriously difficult features adequately sample using traditional situ techniques. In this paper, the authors propose a new statistical modeling approach for detecting monitoring ocean from Advanced Very High Resolution Radiometer (AVHRR) SST satellite images builds on previous “front following” algorithm. Weighted local likelihood is used provide smooth, nonparametric description spatial variations position, mean temperature, width, temperature change individual front within image. Weightings provided by Gaussian kernel function whose width automatically determined cross-validation. The model fitting allows estimation uncertainty each parameter be quantified, capability not possessed other algorithm shown robust noise missing data image, problems hamper many existing front-detection schemes. general could with remotely sensed datasets, output, or assimilation products.