Natural and man-made distributions of tensioactive substance concentrations in the sea surface features exhibit self-similarity at all radar reflectivity levels when illuminated by SAR. This allows the investigation of the traces produced by vortices and other features in the ocean surface. The man-made oil spills besides often presenting some linear axis of the pollutant concentration produced by moving ships also show their artificial production in the sea surface by the reduced range of scales, which widens as time measured in terms of the local eddy diffusivity distorts the shape of the oil spills. Thanks to this, multifractal analysis of the different backscattered intensity levels in SAR imagery can be used to distinguish between natural and man-made sea surface features due to their distinct self-similar properties. The differences are detected using the multifractal box-counting algorithm on different sets of SAR images giving also information on the age of the spills. Different multifractal algorithms are compared presenting the differences in scaling as a function of some physical generating process such as the locality or the spectral energy cascade .