Understanding evaporation #1

One of a series; also see editions: #2, #3, #4

Evaporation is that weird cousin in meteorology. The Australian Bureau of Meteorology lists over 18,000 present and past rainfall recording stations in its huge database, but only about 600 stations that have ever measured evaporation, and just 180 of those still operate.  But wait, measure evaporation … what?  Ok, step back a bit…


Before moisture can precipitate from the atmosphere to fall as rain or snow, it first has to get up there.  That happens mainly by evaporation from the ocean surface (70% of the globe), but also by evaporation from the land, pedantically called evapotranspiration — the combination of evaporation from the ground and transpiration from plants.  In the longer term, global average precipitation and global average “evaporation” (broadly defined) are of course exactly equal, at around 2.7 mm/day.

Evaporation is a primary hydrological variable. The fundamental equation of hydrology goes:

runoff   =   rainfall  –  evaporation  –  storage

…where of course storage is the time rate of change (net new storage¹ per interval).  It’s interesting that we put massive effort into measuring rainfall, runoff and storage, but hardly any into that other parameter².


“Class A” evaporation pan (Wikipedia)

So how do you measure evaporation?  Well, if you put some water in a bucket, put it out in the weather and measured the water depth each day you’d pretty much have it.  In fact we just described the standard instrument.  The “bucket” is a galvanised steel tank 1.2 m in diameter and 25 cm deep, and “out in the weather” is sitting on a nice wooden platform, but that’s about it.  Simple and obvious, yes; but troublesome:

  • Birds and critters will come to drink and bathe, interfering with your measure.  Better add a wire cover (a “bird-guard”)³ … but that’s gong to block some wind and sun, altering the answer (best make it some standard design).
  • Better not be any trees or shelter nearby at all. This instrument is very sensitive to sunshine and near-surface wind.
  • Some days it’s going to rain.  We want total evaporation (not net evaporation), so we’ll need to measure rainfall separately and correct for that.  In practice that’s not as trivial as it sounds, and on days of heavy rain the thing is going to overflow; whoops.
  • Most importantly, that water is out in the sun in an elevated tank.  It’s going to get hot (hotter if the tank is lined with gunk, or even if the outside is dirty), which is going to change the answer.
  • What is this thing actually measuring anyway?

It’s all not quite as simple as it seems.  In fact a class A evaporation pan (what the thing is called) is a notoriously difficult instrument to operate reliably and repeatably, and gives results of dubious applicability anyway.



An evaporation pan measures, well, pan evaporation, which may or may not be closely related to meteorological evaporation and evapotranspiration⁴. At best it models evaporation from a small, shallow pond or tank. For larger, deeper water bodies pan evaporation overestimates actual evaporation — by as little as 10% in tropics increasing to over 30% further south⁵. That’s thought to be mainly due to the different diurnal (daily) temperature cycles: the pan heats up; the lake doesn’t⁶.


Evaporation is fundamental to drought; fundamental to agronomy, forestry and ecology; and fundamental to runoff and to groundwater recharge.  It’s likely to be seriously affected by a warming climate, yet it seems little understood, even by some experts.  More next time…


Next:   edition #2



1.  That is both surface storage and underground storage:  surface storage in dams, lakes, ponds, streams, puddles, or frozen in ice and snowpack; and underground storage in soil moisture and various other forms of groundwater.

2.  In fact storage evaporation is sometimes just estimated (very inaccurately!) by subtraction.

3.  Believe it or not the standard when these things were first introduced here (from the US … where else) in the 1960s was no bird-guard.  Why?  Because it might interfere with the measurements of course … by partially blocking wind and sunlight.  Bird-guards did not become universal in Australia until about 1975.

4.  There are other instruments.  Before the 1960s the standard evaporation instrument in Australia was the “Australian sunken tank”, an evaporation pan set in the ground so that it didn’t get hot and didn’t alter surface wind flow.  Nice idea, except sticks and rubbish would blow in on windy days and create havoc.  Then there is the whole class of things called lysimeters: great big tanks full of soil and plants with instruments (e.g. a giant weighing machine) to directly measure soil moisture and hence evapotranspiration.  Not really your practical, every day, meteorological gadget.  And, more recently, there is Fluxnet.

5. See the classic paper: Hoy, 1977 (below).

6. Actual evaporation is also slightly lower from sea water due to the salt content, but not by much — only a percent or two.



  • Hoy, R.D. “Pan and lake evaporation in northern Australia” Proceedings of a Hydrology Symposium, Brisbane, 1977, Institution of Engineers, Australia.