Scaling property estimation of thermal sea surface turbulent structures on NOAA IR-imagery

A new approach for visualization of SST structures and investigation of sea surface dynamics based on analysis of infrared NOAA/HRPT images is considered. An infrared sea surface image may be considered as a set of oriented textures on the temperature field. The oriented texture is a two dimensional image with dominant local orientation of the IR radiation contrast in every point. Calculation of the oriented textures is based on estimation of the dominant orientations and their statistical significance from field of local radiation gradients. Due to dynamical properties of sea surface the thermal contrasts are stretched along flow currents. So, the nature of the oriented textures is related to the sea flow dynamics. It was verified that statistically significant orientations of the radiation contrasts have good agreement with location and orientation of sea currents and they may be considered as the sea flow markers. To estimate local properties of sea flow dynamics the fractal parameters of the oriented textures are evaluated for curves of the most probable contrast orientations calculated as the integral curves of the contrast orientation field. The box-counting method is used to estimate the local fractal dimension of integral curves and the main scale of the turbulent structures, as the result. Analysis of satellite images of the Okhotsk Sea and Japan Sea were used for verification experiments. The verification experiments showed acceptable performance of the method for evaluation of the turbulent structure features