Display MoreIt cannot be that stable! Nothing is. Take February, where it shows 36,000 every day for weeks. Suppose for the sake of argument the flow was incredibly stable, at almost exactly 1,500 l/hour (36,000 L/day). Problems:
1. If you did not read it at exactly the same hour and minute, it would be 35,000 some days, and 36,000 or 37,000 other days. (This meter only registers to the nearest 1,000, so the other digits would be "000.") Maybe with a computer you could record it at the same hour and minute, but I doubt you would see that kind of precision with handwritten notebooks.
2. Okay, even if you read it at the exact same time, the first day it might show up as 36,000, but if the flow was just a little below 1,500 for a few hours, the next day it would be 35,000. The day after that it might be back to 36,000. Or, if it was slightly higher than 1,500, you would see it go over 36,000 from time to time.
3. On days when the time changes to daylight savings (or back) it would be different. Someone said "maybe they are running GMT." Again, I don't see that kind of precision with handwritten notebooks. Not going to happen.
4. Why on earth would anyone go to the trouble to make it EXACTLY 1,500 l/h? What would be the point? Why bother?
5. If it was so exact in the first place, why did they arbitrarily subtract 10% from the flow rate?
If they had tuned the water flow to 1,500 l/h, for whatever reason, and it's +1% off, or +15 l/h, it takes 3 days before the difference in readings, delta, goes from 36 m3 to 37 m3:
0 d : 0
1 d x 24 h/d x 1515 l/h = 036.360 m3 => reading: 36; delta: 36
2 d x 24 h/d x 1515 l/h = 072.720 m3 => reading: 72; delta: 36
3 d x 24 h/d x 1515 l/h = 109.080 m3 => reading: 109; delta: 37
4 d x 24 h/d x 1515 l/h = 145.440 m3 => reading: 145; delta: 36
5 d x 24 h/d x 1515 l/h = 181.800 m3 => reading: 181; delta: 36
If it's 1,500 l/h +-1%, who knows how long the series of 36'es will be?
But yeah, I've assumes exact 24 h between readings and I've no idea about the possible precision for the water flow.