After an extensive experiment, here are my conclusions on light aquifers:
1. An open tile will receive water if it is N, S, E, W, or down from a light aquifer tile. There is no diagonal transmission.
2. The amount of water received by an open tile is not affected by the number of adjacent aquifer tiles. One is just as good as the maximum of five.
3. The rate of water generation is quite variable. It probably uses a low-chance-per-day system similar to strange moods. The average is four units of water per month, but over a six-month period I've seen the average go from as low as two to as high as six.
4. Therefore, the most efficient way to get water out of a light aquifer is to put it in the ceiling. That way, every open tile will produce water.
The experiment used a small fort with a two-layer light aquifer. I dug out twelve different open tiles and surrounded each with a diffent configuration of aquifer tiles and constructed walls. I put a measuring column 10 z-levels high under each open tile. All the columns were closed at the bottom with hatches linked to a single lever. I left the hatches open while digging the configurations, to avoid premature water collection. Then I closed all the hatches on the first day of autumn 292. I collected data six months later on the first day of spring, and again twelve months later on the first day of autumn 293.
Here's the raw data:
O = Open Tile
A = Aquifer Tile
# = Constructed Wall
1. Single aquifer tile with two empty orthogonal tiles running N/S. No ceiling. (Patrik's suggestion)
###
#O#
#A#
#O#
###
1a (north tile) at six months: 25
at twelve months: 50
1b (south tile) at six months: 16
at twelve months: 37
2. Empty tile with two orthogonal aquifer tiles. No ceiling
#A#
#OA
###
At six months: 13
At twelve months: 47
3. Empty tile surrounded on eight sides with aquifer tiles. No ceiling.
AAA
AOA
AAA
At six months: 19
At twelve months: 48
4. Empty tile only touching the aquifer via the ceiling.
###
#O#
###
At six months: 24
At twelve months: 53
5. Empty tile touching two orthongonal and three diagonal aquifer tiles, all continguous in an L shape. No ceiling.
AAA
#OA
##A
At six months: 22
At twelve months: 46
6. Empty tile touching four orthogonal aquifers and a ceiling.
#A#
AOA
#A#
At six months: 25
At twelve months: 46
7. Empty tile surrounded by eight aquifer tiles, plus a ceiling.
AAA
AOA
AAA
At six months: 21
At twelve months: 44
8. Empty tile touching a single orthogonal aquifer tile. No ceiling.
###
AO#
###
At six months: 28
At twelve months: 57
9. Empty tile touching a single diagonal aquifer tile. No ceiling.
##A
#O#
###
At six months: 0
At twelve months: 0
10. Empty tile touching four diagonal aquifer tiles. No ceiling.
A#A
#O#
A#A
At six months: 0
At twelve months: 0
11. Empty tile touching four orthogonal aquifer tiles. No ceiling.
#A#
AOA
#A#
At six months: 38
At twelve months: 68 (nearly filled the column)
Summary of six-month data for tiles touching orthogonally:
25, 16, 13, 19, 24, 22, 25, 21, 28, 38
Mean: 23.1
Standard Deviation: 6.54
Mean per month: 3.85
Summary of twelve-month data for tiles touching orthogonally:
50, 37, 47, 49, 53, 46, 46, 44, 57, 68
Mean: 49.7
Standard Deviation: 7.92
Mean per month: 4.14
Three of the four configurations that produced the most water over twelve months involved only a single aquifer tile, hence the conclusion that extra orthogonal aquifer tiles yield no benefit.
The fort with the 12-month data is
here. The hatches and the lever that controls them are all on z97. You're welcome to dump out the collected water and run it again, or fiddle with the configurations.
Edit: One more detail: Light aquifers do not act as drains.
Author
Topic: Light Aquifers: The Definitive Guide (!!Science!!) (Read 5376 times)