From the beginning, I have been adamant that the Temple burn in such a way that it will conclude with the tree-like structure splitting open in four quadrants. In thinking about the most recent designs, I couldn't figure out how to handle the three distinct levels. If the entire tree splits, the topmost branches—at some 30' (9m) tall would have a tremendous amount of momentum that would tend to fling burning lumber into the crowd. For the first level, this is not an issue as it is only about 10' (3m) high so the momentum outward would be substantially smaller.
For now, I am considering the idea of the first level falling outward while the upper levels fall straight down. To avoid a tall portion of the Temple landing upright in the ground, the supports from the first to second level would fold outward: they would tie the center of the second level to an anchor on each section of the four splitting parts of the first level. The second level itself (likewise the third level) would be a solid structure that would not be designed to fall outward like the first level. Thus, the support for the second level would split outward with the quadrants of the first level while the second and third levels would tend to fall straight downward. Although the support from the second to the third level would be solid, being weakened by the fire and the direct 20' (6m) drop will at least weaken it severely, if not demolish it instantly.
Looking at a blinking yellow warning light, I got to thinking about how to mark any heavy cabling to secure the Temple during the event. In day, there is less concern as simply adding some ribbon would work. At night I thought to have a LED projection along the length of the cable. Then I figured side-emitting fiber-optic cords would be nice, lit green for optimal visibility. This led me to spiral the fiber-optics, and it would kind of look like a vine climbing, perhaps like an electronic morning glory.
On the technical side, I figured 1 watt of LED light would be plenty, but I'd like to have a redundant pair on each cable. This would mean about 10 watt-hours each per night which could be recharged by a 2W solar cell. But for the sake of using 10 watt-hours for, say, 8 nights, that's just 80 watt-hours. Quality alkaline D-cell batteries are typically 10 amp-hours at 1.5V or about 15 watt-hours each, so it would take a minimum of 6 batteries to hit that mark. A simple resistor circuit could drop a series-parallel pack wired for 4.5 volts to the necessary 3.5 volts at 280mA[1. Estimated from Luxeon Star-III specification sheet, rated for 700mA @ 3.5V = 2.5W]. The current would draw symmetrically (theoretically) from each chain of 4.5V, so the runtime on the battery would be a total of 10AH / 0.14A = 71 hours. (The discrepancy from 90 hours is because it's 280mA at 4.5V which is more like 1.3W. What will happen, though, is the light will get slightly dimmer and the battery last longer. The resistor is fixed at 1V / 280mA = 3.6Ω, so when the battery has dropped to 4V, the resistor drop is only 0.5V, and so the current is 0.5V / 3.6Ω = 140mA.)
I guess that got way too technical too fast. Oh well.