New blending algorithm to synergize ocean variables: The case of SMOS sea surface salinity maps

Abstract

Using the information of an ocean variable of a given kind to improve another variable of a different kind may be a challenging task, especially when they undergo different physical processes. Statistical methods and assimilation in numerical models had been so far the main ways to perform this type of blending, but these are relatively complicated methods that usually introduce other sources of error and uncertainty. In this paper, the existence of a multifractal hierarchy pervading the structure of all ocean scalars is exploited to introduce a new blending method. This method is not parametric and requires no knowledge about the physics governing the evolution of the scalars, provided that both scalars have the same multifractal structure. We have applied this methodology to SMOS SSS maps, using OI SST maps as template variables, observing not only a qualitative but also a significant quantitative improvement. © 2013 Elsevier Inc.