First of all, you manually place the APs on each floor map. As far as I know WCS only deals with 2-D signal spread.
Having said that, there are few site survey/design tools that take multiple floor architecture into consideration when they propose location for APs - for example, AirMagner and Wireless Valley.
You have touched on a known limitation of the WCS.
The WCS can only build RF from APs mounted on that floor. This is unfortunate because if you happen to have a design where you are intentionally using APs from adjacent floors to provide coverage, the WCS heatmaps will shows holes in coverage that are not, in actuality holes.
Also, if you are in a multi-floor building, and are using 1131AG series access points and you are using them in a ceiling-mounted orientation (the only orientation currently supported on the controllers), you will notice that these APs have a somewhat significant downward RF coverage pattern.
For reference, please check out the elevation plane RF pattern under the "Cisco Aironet 1130 Series Integrated Antenna" section of the following document for details: http://www.cisco.com/en/US/products/hw/wireless/ps469/products_data_sheet09186a008008883b.html
The RF characteristic of the 1131AG can often result in RF signal travelling to adjacent floors (both above and below, but mostly below). This is significantly different behavior from the 1000 series APs (designed by AirSpace) that had a more horizontal coverage pattern and did not result in as much floor-to-floor signal leakage. As you may already be aware, the controllers were originally designed to work with the 1000 series APs and so this was not as big of a problem until the other Cisco APs were supported.
The good news is that the RRM (Radio Resource Management) algorithm is able to adjust/account for RF signals from adjacent APs regardless of what floor they are on.
So the heatmap, while cool to look at, is not necessarily going to give you a true picture of your RF coverage.
Also, if you are doing location, there may be some issues regarding vertical accuracy since location tends to be addressed two-dimensionally as well.
One thing that I would like to see (and others whom I have spoken to have also agreed) is a version of the 1131AG access point that has a more horizontal coverage pattern - perhaps similar to that of the 1000 series APs. This would help contain the RF to a single floor and would utilize RF energy more efficiently in single-floor applications as well since the AP would not be blasting RF energy into the ground where there really aren't many users.
For us this is actually even worse, because we have a lot of legacy APs of the 1121 and 1200 variety, the latter with various external antennas and orientations, many of which result in significant inter-floor bleed, which was quite intentionally used in our older designs.
I think the proper solution is to give WCS 3D capability, and there seem to be some hints of progress in that direction. Your suggestion of a properly "flat" (and consequently high-gain) 1131 would be helpful as well, so I'll second your call for one.