The synergy between satellite measurements of Sea Surface Temperature (SST) and Sea Surface Height (SSH) has been studied in the last decades. Among the main approaches, we find methodologies based on (i) global analysis using empirical orthogonal functions (cf. Leuliette and Wahr 1999), (ii) local analysis using surface quasi-geostrophic dynamics (cf. Isern-Fontanet et al. 2006) or (iii) local patch-based regressions introducing a latent variable (cf. Tandeo et al. 2013). This last approach is able to identify and track (temporally and spatially) different relationships between local SST and SSH that correspond to different dynamical modes at the surface of the ocean, characterized by geostrophic/advective displacements with different amplitudes. In this work, Markovian priors are combined with remote sensing data including chlorophyll-a and sea surface salinity along with in-situ measurements of the mixed layer depth, to physically constrain the identique location of the dynamical modes. Then, using this information of surface dynamics and daily microwave SST data, we make estimation of mesoscale geostrophic surface currents. We compare them to the reference Aviso product which only use altimetric data. The results show accurate identification of eddies and ocean fronts in the Brazil-Malvinas confluence region in the Southwestern Atlantic.