Figure 10: Zonal averages of water vapor column abundances from mid-spring to mid-summer for both hemispheres. The top four panels show column abundances for all years, the dots are data points averaged in bins of 2° latitude and 15° Ls, while the curves represent the smoothed bins. The top row illustrates the synergistic retrievals, while the middle row shows the corresponding MCD prior column abundances for each hemisphere. Curves representing  the same seasonal period for both hemispheres have identical colors, with the SH Ls interval listed first. The two bottom panels compare the synergy and the MCD averages from Ls=255°-315° for the SH and Ls=75°-135° for the NH, covering the sublimation season for both hemispheres.
The MCD shows a decreasing trend for all seasons in the extreme high latitudes poleward of the CIA peak, as expected due to the polar cap circulation known as the polar cap breeze (Haberle & Jakosky, 1990), the Martian equivalent to the terrestrial sea breeze. The effect is expected to be stronger in the NH where the more massive ice cap generates a larger temperature gradient. In the retrieved synergy data, the CIA does not always decrease poleward of the cap edge in the SH, and most noticeably continues to increase even beyond 80°S for observations during Ls=285°-300°. This could be due to averaging of data from multiple years (Pankine et al., (2010) reported high interannual variability of this behavior over the NPC), imperfect coverage of this region and season, or perhaps a variable polar cap breeze in mid-summer is not effectively transporting water vapor off the polar cap.
The NH is as expected far wetter than the SH. The CIA increases monotonically from the equator, and does not remain constant across large regions, as in the SH. Distinct maxima are visible with decreasing abundances northward of 80° latitude for all seasonal intervals, in agreement with the model. The overall maximum is observed at 80°N in the Ls=105°-120° interval, same as in the SH, and reaches a peak value of 60 pr-μm. The highest column abundance obtained by the MCD is in the interval Ls=90°-105° and reaches 83 pr-μm. The locations of the CIA peaks are found just south of the polar cap edge, with a clear decreasing trend for all seasons in the extreme high latitudes poleward of the CIA maximum, as expected due to the effects of the polar cap breeze. The sublimation onset is observed to occur later than what is predicted from the MCD, where during Ls=60°-75°, the synergy finds a gradually increasing latitudinal trend with a modest peak at 65°N of just below 20 pr-μm, while the MCD already estimates a significant maximum of 30 pr-μm at 70°N.
In the bottom two panels of Figure 10, seasonal averages of the intervals Ls=255°-315° for the SH and Ls=75°-135° for the NH (covering the main sublimation period for both hemispheres) are shown to provide comparisons between the general trends in meridional CIA gradients from the synergy and MCD. The CIA absolute values are interesting to compare, but even more so the meridional variation. The summer sublimation maximum in the MCD is quite easily adjusted by tuning model parameters, while the change with latitude is subject to convection, transportation and possible surface exchanges, and not so straightforward to modify to obtain the desired output. In the south the trends are nearly identical, with the synergy only yielding slightly smaller average abundances in the 10°-30°S and 50°-70°S regions. In the north, the MCD deviates from the synergy most significantly in two places; at 20°N and at 50°N, where in both instances the MCD gradient distinctly increases with respect to the synergy. The “double-hump” shape of the CIA is also much more prominent in the MCD. The difference between the MCD and synergy is small towards the equator for both hemispheres, which might be indicative that the influence of the CIA sublimation peak diminishes at lower latitudes.
Seasonal differences in the PI appear small in the MCD model compared to observations, as can be seen for all seasons in Figure 11, where all the curves are more or less stacked on top of each other. In our retrievals the partitioning exhibits a wave-like behavior in both hemispheres, oscillating around PI=0.5 in the south and around PI=0.65 in the north. The shape of the MCD PI curves resemble those of the CIA seasonal averages, and do not have the same wave-like quality that the synergy finds. As the synergy yields very stable column abundances, for low/mid latitudes for all seasons, the partitioning varies greatly, particularly in the southern mid summer. However, the number of data points in the SH are far fewer than for the NH, and the averages from this region should therefore be considered somewhat less precise. This disagreement is also visible (to a lesser extent) in the NH, indicating that the discord is likely not purely a result of poor sampling in the south. In the NH there is a clear tendency for the partitioning to suddenly increase poleward of 80°N while the total water content decreases. The MCD PI on the other hand has been steadily increasing from the mid latitudes, and during late spring the PI even decreases north of 80°N. In the north, no stable PI gradient is observed as the MCD suggests. The synergy finds a highly variable PI for all latitudes and seasons, but with no clear meridional tendency.
These differences between the MCD and synergy are highlighted in the sublimation season averages for the PI in the two bottom panels of Figure 11, which clearly show the observed wave-like behavior being consistently higher than the estimated stable MCD PI. While the MCD indicates that around 40% of the water column is kept near the surface at all latitudes and seasons, the synergy finds that number to vary from 40-60%, with local maxima at equator, 50°S and at the pole. This trend is very similar to what is observed in the north, albeit with a smaller wave amplitude. The MCD PI here is not as stable as in the south, and displays a fairly constantly increasing gradient from the mid latitudes (PI=0.4) towards the north pole (PI=0.75). The synergy finds that the PI never goes below 0.6, indicating that most of the column is always kept close to the surface, but can vary rapidly. This leads to the synergy and MCD finding similar PI values only in the north polar region.