German Weather Service meteorologist Christoph Hartmann writes what I think is a surprising essay on measuring precipitation, and the errors in doing so. Indeed Hartmann says precipitation may be understated by up to 50%, or much more at some locations.
As Hartmann explains, measuring precipitation is by no means an exact science, and results have to be taken with a lump of salt.
There are many sources of errors, and in his essay here he looks at just two main sources: wind and instrumentation.
But first, let’s take a look at how precipitation is measured. In his previous essay he described two types of precipitation measuring gages. In Germany precipitation is measured with the unit of liters/m², e.g. 25.4 liters is an inch of rain.
Two methods of measuring precipitation
Hartman explains that precipitation is generally measured by a rain gage with a known opening area, for example 200 cm² in Germany, which is positioned 1 meter above the ground surface. The gage funnel catches the precipitation and leads it to either
1) a graduated measuring tube or a
2) an optical drop counter
With the measuring tube system, the tube is graduated and the amount of precipitation can be simply read off. With the optical rain gage (drop counter), the amount of precipitation is derived from the number of drops. If the precipitation is snow or ice, then the measuring tube or optical gage are brought inside and the captured precipitation is melted and measured.
Hartmann explains that the biggest sources of error are wind-related. This is easily seen when measuring snowfall. Just before a snowflake falls into the gage, air turbulence sucks it back out tosses it overboard. Just taking a look around after a blizzard, it’s easy to imagine how difficult it is to measure snowfall. Places exposed to wind are barren, while other places are covered by meter-deep snowdrifts. How much snow really fell?
Hartmann says measurement errors of up 400% can occur over time when measuring powdery snowfall in alpine, polar or windy areas.
One way to reduce error is to place the instrument in a wind-protected area. By measuring the wind speed, it is then possible to adjust precipitation measurements. But Hartmann writes:
Wind effects lead to an under-estimation of the actual fallen precipitation. The level of deviation depends on the speed of the wind and the type of precipitation.
Because wind speeds are factored into precipitation measurements, climatological precipitation trends without taking changes in wind speeds into account should always be deduced very carefully.
The second problem encountered arise from the two above described measurement instruments, especially with the optical rain gage, writes Hartmann. With frozen precipitation, the gages are heated up in order to melt the precipitation. But this involves evaporation. And under torrential rains, the optical gage becomes much less accurate. The result, writes Hartmann:
Under equal precipitation amounts, the optical gage measures less precipitation than the measuring tube, both in summer and in winter.
So if two different stations use different instruments, them they will show different precipitation amounts even when the actual precipitation is the same. In summary, Hartmann writes his stunning conclusion:
In total these two sources of errors lead to a precipitation deficit of 5 to 15% for liquid precipitation, and between 20 and 50% for solid [frozen] precipitation. In very windy locations, the deficits are substantially more.
Because instruments measure less precipitation than what actually falls, it means we have a worldwide precipitation deficit solely because of the measurement method.
What does it all mean? Are many of the reported droughts solely the product of faulty readings? And we all thought that the network of temperature measurement stations was a mess. This is a huge open floodgate to potential climatological data manipulation and bogus assertions. See here for example: motherjones – the coming mega-drought (h/t NTZ reader DirkH).