ballvalve: there are some designs out there that are based on passive cooling. One design does use the "water tower" concept by placing a large tank of water above the core inside of the containment (reactor) building. Although gravity always works, there are challenges with this too. Remember that a plant during normal operation will be at 2200 psia or so (for PWR reactors)..the GE is a BWR and typically runs around 1000 psia. You get about 0.5 psi for every foot of head, so to be able to inject under high pressures, this water tower would have to be close to 0.5 - 1 mile high (not practical). The other issue is the amount of water needed. Not only do you need enough water to fill the reactor vessel to the top of the core, but you really need enough to flood the entire reactor building to the level. If there is a break in the piping, the water that you put in just goes back out and onto the floor. If your containment is intact, you just flood the thing to the top of the core level and go into "boiling pot" mode. This might require 500,000 to 1,000,000 gallons of water or more. The other issue is that this tower and piping would need to be "safety-grade" meaning that it could withstand earthquakes, tornados, huricanes, etc. The other issue is that the containment outer shell is the last line of defense with regards to a fission product barrier (I'm talking about typical PWR reactors here). This means that you don't want penetrations through this wall to the outside. In the current designs that use a water tank above the core for passive cooling, they need to get the system pressure down to were the head from the tank can overcome the system pressure in order to inject. They use 4 valves that are about 14" in diameter that pop open and release to the containment. If your containment has failed, then this would be released to the environment. These are the reasons why active systems are typically used. In the design of the reactors where this event is going on (BWRs), the building that was damaged is not designed to be a real barrier. The real containment barrier is inside the building. On typical PWR (usually plants where you see a dome), the outer wall is typically 4' thick concrete, heavy rebar, plus a steel liner. Some newer designs even use a double containment ( two shells, each 4' thick, with a gap between them). The other aspect to this is to what levels do you design to? Typically, we will look at historical information and design for the worst + additional margin. Now say you build somewhere that has never recorded a quake bigger than 2.0. What do you design for? Do you design for a 4.0...but then maybe a 5.0 will hit? Maybe you design for 8.0 and an 8.9 hits??? Each point you go up on the scale is 10x the magnitude, so building to withstand an 8.0/9.0 in an area that has never seen over 2.0 is not very practical. Nature typically doesn't give us an upper bound for disasters. All you can do is look at the historical information and do your best to design based on that information. There is always some chance that something will happen that is worst than anything ever recorded (like this event).