Sensibility to noise of new multifractal fusion methods for ocean variables

Abstract

The repeated observation of the same signatures of mesoscale and submesoscale features in different ocean variables indicates that some common, non-linear processes affect them to a significant extent. A new method to exploit these common signatures to improve the quality of a noisy variable (i.e. increasing the signal-to-noise ratio) using another variable as template has recently been introduced. The method is based on superimposing the multifractal structure of singularity exponents from the template variable to the variable to be enhanced. In this paper, we will discuss the sensitivity of this method to the presence of noise of different types and amplitude. Our results indicate that multifractal methods can be a key to enhancing the existing databases of remote sensing images and give hints about non-linear dynamics of the ocean.