একটি খসড়া ডেমো

ছাদের এক বৃষ্টিমাপকের হিসাব

মেনু

The Roof Notebook · উপকরণ

The Roof Notebook

২০টি ফাইল

research.md৫ KBডাউনলোড

Rain gauges, and why a tin is not one

Notes gathered between 20 July and 3 August, mostly from the library’s copy of a meteorology primer, two afternoons with Mr Haque, and a lot of measuring the tin.

What a gauge is

A standard rain gauge is a funnel over a narrow measuring tube. The funnel’s mouth has a known area — the common one is 203 mm across — and the tube’s is a tenth of that, so every millimetre of rain over the mouth stands ten millimetres deep in the tube. That magnification is the whole trick: it lets a reader see a tenth of a millimetre.

Three things make a reading comparable with anybody else’s:

  1. The mouth. A sharp, level rim of known area. Rain that hits the rim’s edge either goes in or it does not; a rounded or ridged rim makes that a coin toss.
  2. The height. The mouth sits about 30 cm above the ground. Lower, and splash from the ground gets in; higher, and wind curls round it and carries drops past.
  3. The hour. Readings are taken at a fixed hour, and the day’s rain is the rain in the twenty-four hours before it. Our met office reads at 09:00 local time.

What the tin is

Gauge Tin
Mouth 203 mm, sharp rim 220 mm, rolled and ridged
Base narrow tube 196 mm
Sides — taper inward
Height above roof 30 cm 0 cm, on the slab
Reading graduated tube a steel ruler, later warped
Hour 09:00 06:00

The taper is the part that matters most. Because the tin is narrower at the base than at the mouth, a millimetre of rain over the mouth stands more than a millimetre deep at the bottom and less than that near the top. A reading of 41 mm in a nearly full tin is not the same amount of rain as a reading of 41 mm would be in a tin that was nearly empty the day before — except that it is, because I emptied it every morning. Emptying it every morning is the one thing I did right without knowing why.

Working out the factor

For a vessel whose sides taper evenly from radius R at the mouth to r at the base, the volume up to height h is a frustum’s. I did the sum on the back of the 24 July page (typed up as frustum.py) and got a factor of about 1.25 at shallow depths and 1.12 near the brim. In other words the tin overstated the rain by somewhere between twelve and twenty-five per cent, depending on how much there was.

mouth   R = 110 mm
base    r =  98 mm
depth   H =  95 mm
area at mouth   = π R²  ≈ 38 013 mm²
area at depth h = π (r + (R − r) h / H)²

I never applied it. See below.

Splash and the slab

The tin sat directly on the roof slab. In heavy rain, drops hitting the slab around it bounce, and some of the bounce lands in the tin. Mr Haque estimated that on the worst nights this could add “a few millimetres, more if the slab is wet already, which it always is.” There is no correcting for this after the fact. The only fix is to raise the tin.

I did not raise the tin, for the same reason I did not apply the factor.

Why none of it was corrected

“Your numbers are wrong. Your days are right.”

— Mr Haque, 2 August, after reading the notebook for about twenty minutes

Every correction above would have made the numbers closer to the met office’s. None of them would have made the days more like themselves. A dry day was nil in the tin and nil in the gauge. A downpour was a downpour in both. The warped ruler added roughly a millimetre to every reading after the sixteenth, the same millimetre every day, so the difference between two days after the sixteenth is as true as it ever was.

What I wanted from the notebook was the shape of the season, not its total. For a shape, consistency beats accuracy. That is the sentence the finished piece is built around, and it took three weeks of these notes to earn it.

Still open

  • Whether the tin’s lid, left leaning against the tank, sheltered it from the west on windy nights. Probably on the fifteenth, when the neighbour’s tank caught what the tin did not.
  • What the overflow on the fourteenth actually was. The tin holds 95 mm. It was full at three and still raining.
  • Whether Tanvir was right about the leak. I filled it from a jug on 30 July and left it a day; it lost nothing.

Sources

  • A Field Guide to Weather Instruments, ch. 4, “Precipitation”, and the table of standard gauge dimensions.
  • Two conversations with Mr A. Haque, 28 July and 2 August. See interview.md.
  • The tin, the ruler, and a kitchen jug of known volume.

khosra · khata 30723e3a · content bbf50b9e · 228 publications, 4256671 bytes · built in 1.41295s