Eighty one hectares of catchment, a rain gauge in the middle of it, and a stream that carries about fourteen per cent more water out of the bottom than the sky puts in the top.
I monitor a stream for a rivers trust. It is volunteer work and it is the most satisfying thing I have ever done, and I am going to explain the whole method because the method is the point.
A catchment is the area of ground that drains to one place. You find it on a map by following the high ground: everything inside the ring runs downhill to your stream, everything outside it goes somewhere else. Mine is eighty one hectares, and I have had that checked twice, once by a hydrologist at the trust and once by me on the free LiDAR data, and we agree to within about half a hectare.
Nearly all of it is wood. Old broadleaf on a clay cap with sandstone under it, on a gentle slope.
Rain falls on those eighty one hectares. Some of it evaporates, some is taken up by the trees and breathed out of the leaves, some soaks away into the rock, and the rest comes down the stream past my gauge.
So the stream can carry less water than the rain that fell. It is supposed to. Around here a wooded catchment puts maybe half of its rainfall into its stream over a year and the rest is lost to the air and the ground.
It cannot carry more. There is nowhere else for water to come from.
My instruments are not sophisticated and I want that on the record, because the first thing anybody says is that my kit is wrong.
Rainfall: a tipping bucket rain gauge in a clearing near the top of the catchment, on a post, sited to the standard, which matters more than the gauge does. It logs every tip and a tip is a fixed small amount of rain.
Flow: a V notch weir plate in a concrete throat in the stream, about four hundred metres above where it leaves the wood. Water backs up behind the plate and goes through the notch, and the depth of water over the notch tells you the flow, because the relationship between the two is fixed by the geometry of the notch and is in every hydrology textbook there is. A pressure sensor in a stilling well logs the depth every five minutes.
That is it. A bucket that tips and a triangular hole with a ruler behind it.
Over six full years the rain gauge has recorded an annual average of 782 millimetres.
Across eighty one hectares that is about 633,000 cubic metres of water a year arriving on the catchment.
The weir has passed an annual average of about 722,000.
That is fourteen per cent more water leaving than arriving, every year, six years running, and the yearly figures are 11, 13, 16, 12, 18 and 14 per cent, so it is not one bad year carrying an average.
And it is not fourteen per cent more than the rain that ran off. It is fourteen per cent more than the rain that fell, before anything is taken out for the trees and the ground, which between them should be swallowing roughly half.
So the honest number is that the stream carries something like two and a quarter times the water this catchment has any business producing.
Everything sensible, in the order I did it.
The rain gauge reads low. This is the obvious answer and it is a real problem with tipping buckets in heavy rain, where the bucket is moving while water is still coming in. I put a second gauge of a different type fifty metres away and ran both for fourteen months. They agree within two per cent.
The catchment is bigger than I think. This was my best hope for two years. Water does not respect a map: ground water can cross a surface watershed through the rock and feed a stream from outside its catchment, and that is not exotic, it is ordinary. But it has a limit, and the hydrologist from the trust went through it with me properly. To make up this much you would need to be draining something like two hundred hectares, and the neighbouring catchments on both sides are gauged by the water company and neither of them is missing any water. Their numbers are normal. There is no hole in anybody's budget but mine.
The weir is wrong. A weir over reads if the approach is silted, or the plate is not level, or the crest is damaged, or the flow is not properly backed up behind it. So I had the whole throat rebuilt in 2022 by the contractor the water company uses, with a new plate, surveyed level, and a proper stilling basin. It cost the trust nine hundred pounds and I felt sick about asking.
The new weir reads the same as the old one.
The sensor is wrong. I have checked it against a dipped staff gauge in the well more times than I can count. It is right.
The part I did not expect.
If the stream has extra water in it, the extra has to be coming in somewhere, so I gauged the tributaries. There are four, and you can measure a small stream with a bucket and a stopwatch if you are patient, which I was, on thirty one separate days.
They add up. The four of them plus the main channel above the top tributary come to the flow at the weir, within the error of a man with a bucket. Nothing is appearing between them.
So I went up each one. And the same thing is true of every one of them, all the way to the top, at every point I could measure: a tributary a hundred metres long is carrying more water than the ground it drains can be giving it, by roughly the same proportion as the whole wood.
It is not a spring. A spring would be a place. I have looked for a place for four years.
It is everywhere, a little bit, all through the wood, and it is about fourteen per cent.
What I have written in the annual return, for six years, is the flow data and nothing else. The trust wants flow and temperature and invertebrate counts and I give them all three and they are good data and somebody uses them.
I have never once submitted the rainfall comparison, and the reason is not that I think anybody would laugh.
It is that the first thing a hydrologist will do with that number is look for the water, and the only way to look for water in clay is to drill, and I have stood in that wood on a still night in February and listened to it, and it is the wettest quiet place I have ever been in. You can hear it moving under the leaf litter everywhere, all at once, in ground that has no reason to be wet, forty metres above the stream.
I am sixty-four. The gauge will outlast me and the data will be there and somebody cleverer can have it.
What I have stopped doing is going up there after dark, and I could not give you a reason for that which would satisfy anybody, including me.