We analyze velocity fields of the Great Red Spot (GRS) and Oval BA that were previously extracted from Cassini, Galileo, and Hubble Space Telescope images (Asay-Davis, X.S., Marcus, P.S., Wong, M., de Pater, I. [2009]. Icarus 203, 164–188). Our analyses use reduced-parameter models in which the GRS, Oval BA, and surrounding zonal (east–west) flows are assumed to have piece-wise-constant potential vorticity (PV), but with finite-sized transition regions between the pieces of constant PV rather than sharp steps. The shapes of the regions of constant PV are computed such that the flow is a steady, equilibrium solution of the 2D quasigeostrophic equations when viewed in a frame translating uniformly in the east–west direction. All parameter values of the models, including the magnitudes of the PV, areas of the regions with constant PV, locations of the transition regions, widths of the transition regions, and the value of the Rossby deformation radius, are found with a genetic algorithm such that the velocity produced by the equilibrium solution is a “best-fit” to the observed velocity fields. A Monte Carlo method is used to estimate the uncertainties in the best-fit parameter values. The best-fit results show that there were significant changes (greater than the uncertainties) in the PV of the GRS between Galileo in 1996 and Hubble in 2006. In particular, the shape of the PV anomaly of the GRS became rounder, and the area of the PV anomaly of the GRS decreased by 18%, although the magnitudes of PV in the anomaly remained constant. In contrast, neither the area nor the magnitude of the PV anomaly of the Oval BA changed from 2000, when its cloud cover was white, to 2006, when its cloud cover was red. The best-fit results also show that the areas of the PV anomalies of the GRS and of the Oval BA are smaller than the areas of their corresponding cloud covers at all times. Using the best-fit values of the Rossby deformation radius, we show that the Brunt–Väisälä frequency is 15% larger at 33°S than at 23°S. As expected ( Marcus, 1993), the best-fit results show that the PV of the zonal flow has “jumps” at the latitudes of the maxima of the eastward-going jet streams. However, a surprising result is that a large “jump” in the PV of the zonal flow occurs at the location of a maximum of the westward going jet stream neighboring the GRS. Another surprise is that the jumps in the PV of the zonal flow do not all have the same sign, which implies that there is not a monotonic “staircase” of zonal PV from north to south as was anticipated ( Marcus, 1993; McIntyre, 2008).