Finding water

A borewell is the least reversible thing we have done so far. You choose a spot, a lorry arrives, and within a day there is either water under your house or a very expensive dry hole in the ground. Nobody gives you the money back for the second one.

So the question of where is worth taking seriously. We took it seriously twice, in two completely different ways, and got two completely different answers.

The coconut

The first man works the way this has been done here for a very long time. He walks the land holding a coconut in his open palm, and where the coconut stands up on its end, there is water below.

It is easy to be rude about this and I would rather not be. He was unhurried, he clearly knew the area, and he was quite certain. Plenty of working borewells across Karnataka were sited exactly this way, and the people who own them are not fools.

What he gave us was one point. No second choice, no reasoning we could follow, no sense of how deep or how likely. One spot on the ground, and a fee.

The survey

The second was a proper groundwater survey — instruments, readings taken at intervals across the plot, a man who was interested in the rock underneath rather than the surface of it.

What came back was four points, spread across the property, each one a candidate. Not a promise — nobody sane promises water — but four places worth drilling, with the reasoning attached.

The part that decided it

The two sets of points did not agree at all. Not "roughly the same corner". The coconut's single spot was nowhere near any of the survey's four. One of them was wrong, and there was no way to tell which from the outside.

Two things settled it for us.

The first was simply optionality. If you drill a surveyed point and it comes up dry, you still have three more candidates and a reason to believe in them. If you drill the coconut's point and it comes up dry, you have nothing — no fallback, and nothing to think with.

The second was the pricing, which we did not expect. Both men charged the same. One point and four points, no reasoning and full reasoning, for the same money. Once you notice that, the choice makes itself.

We went with one of the four.

Drilling

The rig came on the 9th of July — a truck with a mast, purple and green and decorated the way these lorries always are.

It went down 325 feet in the end, and the way that is charged is worth knowing if you have never done this: the price per foot steps up with depth. The first 250 feet were ₹95 a foot. Everything past 250 was ₹105. The deeper you go the more each foot costs, which means a borewell's bill is not linear — a dry hole at 400 feet hurts far more than twice as much as one at 200.

Water did not arrive once. It arrived three times.

The first two were worth having. The third was the one worth stopping for — you could see the difference in how hard it came up. They pushed on another fifteen feet, and by 325 feet the rock had become almost impossible to drill. That is where it was called off: not at a planned depth, but at the point where the ground stopped giving.

When they stopped, the driller called the yield at four to four and a half inches.

If you are not from here, that unit will mean nothing, so: borewell yield in this part of the country is quoted in inches, and it is not an instrument reading. It is the driller judging the water coming out of the hole — how thick the flow is, how hard it is arriving — from long practice. The rule of thumb people use locally is roughly a thousand litres an hour to the inch, which would put this somewhere around four thousand. Treat that as what it is: a rule of thumb resting on an estimate.

By local reckoning, four and a half is a good borewell. Not a remarkable one — people do strike six and eight — but comfortably the right side of the line you worry about.

All three sources feed the same well, which is worth understanding. The steel casing only goes down as far as the rock. Everything below that is open hole, so each water-bearing fracture the bit passed through — 190, 240, 310 — drains into it. You are not tapping one aquifer. You are collecting from every crack you went through.

And you do not get a gentle sign that you have found it. Grey water comes up out of the hole under pressure, mixed with rock cuttings, and runs off across the site in whatever channel it can find. Everyone stands around looking at it. It is a genuinely good moment.

What it cost

₹1,68,175, all in.

Here is the part we did not see coming, and it is the single most useful thing in this entry:

The casing cost four times what the drilling cost.

Everybody quotes borewells by the foot, and so everybody budgets by the foot, and the per-foot number turns out to be under a fifth of the bill. The pipe that holds the hole open is the real expense: 39 feet of 10-inch PVC at the top, and then 185 feet of 7-inch mild steel casing below it at ₹590 a foot.

Two things drove that number.

The first is how deep the casing had to go. You case a borewell through whatever is too loose to stand on its own and stop once you reach rock that will hold itself open. Here the rock was not found until about 180 feet — so the steel had to run the whole way down to it. The soft, weathered material on this site goes a very long way before it turns into anything solid, and the casing bill is simply that fact converted into rupees.

The second is how thick the steel is. Ours is 5 mm wall — roughly twice the gauge normally used for a borewell of this size. Over 185 feet, that choice alone is a large part of why this line reads ₹1,09,150. It buys a casing far less likely to buckle or corrode through, in a hole where an awful lot of loose ground is leaning on it for a very long time.

If you are budgeting for a borewell, ask two questions before you ask the per-foot drilling rate: how deep is the casing expected to go, and what gauge. Nobody volunteered either to us.

What this has to do with a house made of earth

Two things, and neither is obvious.

The first is that our soil report found no water table within 3 m of the surface, which we were glad about — an earth wall's enemy is damp rising into it from below. That is an entirely different question from this one. The water we just found is 300-odd feet down, in rock, behind 185 feet of steel. Those two facts are not in tension; they are about two different worlds.

The second is that rammed earth is thirsty to build. Every lift has to be mixed to a specific moisture content before it goes into the formwork, and there will be a great many lifts. There is a line on this very bill — a water tanker, ₹1,500 — for the water needed just to drill the hole.

A four-inch well, if it holds up, is comfortably more than a household needs and leaves a real margin for wall-building on top. That margin is the whole reason this had to happen before the walls and not alongside them. Having our own supply is not a convenience here. It is close to a precondition.

What we still do not know

We have water from three separate depths and a driller's figure of four to four and a half inches. What we do not have is a proper yield test, and the difference between those two things is the part worth being careful about.

The number you are given on the day is taken while the rig is still blowing compressed air down the hole, at the end of a monsoon, from a well that has never been pumped. It is a genuine signal and it is not a promise. What actually governs how you live is the sustained yield — what the well gives hour after hour, with a pump on it, in April, after three dry months. That figure is routinely lower than the one called out on the day, and a good many people have been disappointed by the gap.

So the real test is a pumping test, later, ideally at the end of a summer rather than the end of a monsoon. Nor do we have a water quality test yet. Both are on the list.

What this proves, honestly

Not much, and it would be dishonest to claim otherwise.

We drilled one hole, at one of four surveyed points, and found water. That is a single trial. It does not prove the survey was right and the coconut was wrong — if we had drilled the coconut's point we might well be writing the same entry with a different photograph.

What we can say is narrower and, we think, more useful. For the same money, one method gave us a single instruction and the other gave us four options and its reasoning. When you cannot verify either, take the one that leaves you somewhere to go if it is wrong.