Aviation fuel is metered at four points between the tanker and the wing and reconciled every morning, because an unexplained loss is either a leak, which is a prosecution, or a theft, which is a dismissal. Tank 2 loses six hundred litres between midnight and four.
I run a fuel farm at a regional airport. Seventeen years, the last nine as supervisor.
The reconciliation is the whole of this, so here is how a tank is actually counted.
Fuel comes in by road tanker through a meter. It goes out to the aircraft through another meter on the bowser, and the bowser is metered again when it fills. Every morning at six you dip each tank by hand with a rod, and you also read its gauge, and the two have to agree, and then you correct the volume to fifteen degrees.
That last part is the one people trip over. Fuel expands and contracts with temperature, so a tank that is cold at four in the morning genuinely holds fewer litres than the same fuel did at four in the afternoon. So nobody in this trade counts litres. You count standard litres at fifteen degrees, worked out from the temperature probes in the tank, and then the arithmetic closes.
You also dip for water every single day, because water sits under the fuel and a tank with water in it fuels nothing.
At the end of the month the gains and the losses go on a return. An unaccounted figure under about two tenths of a percent is considered normal in this industry and nobody looks twice at it.
Tank 2 loses six hundred litres a night.
Corrected to fifteen degrees, so it is not the temperature. I need that said early because it is the first answer everybody reaches for and it is a good answer, and it is the reason it took me two years to believe the figure.
It happens between the midnight reading and the four o'clock reading, which we only take because I started taking it.
It is six hundred litres whether the tank is full or a quarter full.
What has been ruled out, and the company has spent real money on most of it.
The meters. Calibrated, swapped, and proved against a certified measure.
The gauge. It is a radar gauge and it agrees with a rod in my own hand, and a rod in my own hand is the oldest and least arguable measurement on this site.
The line. There is one outlet from tank 2 and it runs to the bowser fill point. It has been pressure tested twice, by a contractor, overnight, and it holds. A line that holds pressure for eight hours is not a line that passes six hundred litres.
The ground. We have four monitoring boreholes round the farm, sampled quarterly by a laboratory because aviation fuel in groundwater is a criminal matter. Eleven years of clean results. The interceptor has never shown product.
The vent and the vapour. If it were leaving as vapour the loss would follow the temperature, and it does not, and the figure is the same on a still December night as on a mild October one.
People. There is a camera on the gate and a camera on the farm and a key system that logs every entry. No vehicle, no person, no second set of dips.
What I did in 2022, which is the thing I cannot get round.
I fitted a second flow meter on the outlet from tank 2, on my own initiative and with a maintenance engineer who did it as a favour, because the outlet is the only hole in the tank that fuel can get out of.
I logged the tank level and that meter together for eleven nights.
On nine of those nights the level fell and the corrected volume dropped by about six hundred litres.
The meter on the only pipe out of the tank registered no flow at all.
Not a trickle. Zero, on a meter that will see four litres a minute.
What I do about it now.
I put it on the return as unaccounted loss, which is honest, and the figure runs at about a tenth of a percent of throughput because we move a lot of fuel, and in nine years no auditor has asked me a single question about tank 2.
I have stopped raising it. I raised it three times. The last reply, from a regional manager in 2020, said that persistent small discrepancies in a farm of our age are normal, which is true, and that I should focus on the daily dip discipline, which I have never once missed.
The thing I changed is how much I leave in it.
We ran tank 2 down for an internal inspection in March. It was down to about seven hundred litres, below any level I had ever seen in it, which is a puddle in a tank that size.
That night it took six hundred.
The inspection went ahead in the morning. I went in through the manhole myself with a lamp because I am the one who signs the confined space entry and I was not going to send a lad in.
It is a clean tank. No scale, no water, no product on the floor beyond what a hundred litres looks like spread over that area.
We refilled it that afternoon and I have kept it above eighty percent ever since, and I am aware that makes no difference to the six hundred, and I do it anyway.