Accurate measurement of rainfall

The accurate measurement of the amount of rain that falls in a set period of time is a challenge that has defied meteorologists ever since the first rain gauges were developed in the 1650’s.

Early gauges, as long as three millennia ago, were simply bowls left out in rain to see and compare how rapidly they filled up. Korean records were made relatively accurately in the fifteen century when the earliest recorded measurement network was set up; the amount of rain recorded locally was used to assess the potential taxable revenue from the annual harvest.

The earliest relatively reliable gauges, known as pluviometers (from the Latin pluvial: rain) were developed by Sir Christopher Wren in the mid seventeenth century, who invented an early version of the ‘tipping bucket’ where known amounts of rain are measured against counter balanced weights and then automatically emptied but it was a Reverend Horsley who invented what we would recognise today as a modern rain gauge, in 1722. This was a simple affair where an amount of rain falls into a cylindrical container of known size, which is then poured into a measuring cylinder at the same time each day. If we know the size of the funnel circumference we can then accurately estimate the amount of liquid which has fallen in a specified period and compare the reading with another gauge.

However, although we can certainly gain some idea of the amount of rainfall in a period, there are a number of areas and issues that can cause problems in relation to the accuracy of the measurements.

Broadly these include:

• exposure
• type of precipitation
• amount of precipitation
• wind speed

The exposure will need to ensure as far as is possible that there are no structures close to the gauge which could affect its ability to accurately record rainfall. Buildings or trees can create ‘shadowed’ areas or eddies that ensure that the rain falling doesn’t accurately represent the actual amount that has fallen. Ideally therefore a gauge should be situated in a flat open field; in practice this is rarely possible.

The type of precipitation falling will also affect the measurement made. Snow, for example, may be blowing or drifting as it falls and so not be accurately represented as it falls into the gauge; equally, in very cold weather, if the gauge isn’t equipped with an internal heater, heavy snow will probably not melt into the gauge itself, collecting in the funnel area. Over time this may mean that it builds to the top of the gauge and then falls or is blown off the top. There is no real solution to this problem, except to measure instead the equivalent area and amount of snow over an area of level ground locally, but this assumes no snow melt has occurred in the meantime. Equally problematic can be hail; this often simply bounces out of the gauge and is not collected through the funnel. Large hail may actually block the funnel and allow further heavy rain to build up and then flow out of the top of the gauge, especially in severe storm events.

How heavy the precipitation is will also affect the total received inside a gauge. If very light rain falls it may evaporate before reaching the measuring cylinder or bucket mechanism, especially on very warm or dry days with low humidities. Conversely, if the rainfall is especially heavy, perhaps during a thunderstorm or torrential shower, some of the rain drops may splash right out of the gauge.

Another factor that will need to be taken into account is the prevailing wind direction. In areas with very strong prevailing winds rainfall may tend to flow horizontally across the top of the gauge. When the wind is especially gusty too, which can often happen in the heaviest convective events, rain or hail may be blown across, rather than into, the gauge top.

As can be appreciated, a combination of all these factors may apply to the collection of rainfall on many sites, but especially those that are particularly exposed at relatively high altitudes. This means that the data which is available, even from sites that are located relatively close to each other, can potentially vary quite widely. As far as possible a location which is flat and open needs to be chosen to record accurate data but unless a heated gauge is used then some of the problems mentioned above are very hard to solve. Even in this era of digital measurement precipitation is still very much an ‘analogue’ phenomena.