In a companion paper we have shown that at the mesoscale, the unusual across front sampling strategy (AL) is more accurate than the usual across front sampling strategy (AC) on hydrographic and bio-chemical properties when the front is sufficiently developed. The cruise design based on the optimal reconstruction of the density fields does however not imply that derived variables like quasi-geostrophic (QG) vertical velocities and temperature fluxes also exhibit minimal associated errors. Here we present results of optimized sampling strategies for diagnostic QG vertical velocities (w_QG) and temperature fluxes (θ_QG) derived from the omega equation. Results are illustrated in the same framework, for the Iceland-Faroes oceanic frontal area and for a control model and reveal that at these spatio-temporal scales, the unusual AL may also provide better estimations of vertical velocities and temperature fluxes compared to the classic AC, especially when the front is sufficiently developed.