Evaluating Observed Differences in Co-Located Rainfall Measurements at a Highly Exposed Field Research Station in Scotland, UK

Rainfall measurements worldwide are systematically deficient, to an extent which is currently unknown. The most widely used device to measure rainfall globally is the tipping bucket rain gauge (TBR). The precision and accuracy of TBR measurements vary, and as such, calibration procedures are dependent upon the organisation or institution operating a network. Rainfall datasets may therefore be heterogeneous and not easily comparable. The measurement of other meteorological variables, such as temperature, is regulated, and procedures are in place which assure standardisation. With TBRs there are no universally adopted standards assuring users that rainfall datasets are heterogeneous and interoperable. CEN (2012) recommends a procedure for dynamic calibration of rain gauges, but this has not yet been released as an official international standard.

Without a unanimously applied method for assessing and correcting the instrumental error, it is difficult to evaluate the environmental error, which is also referred to as the ‘catching' error. Nevertheless, since National Meteorological Services currently use a conventional static calibration, it is a worthwhile exercise to initially use this same method of rain gauge calibration which has been in operation for many years.

The most commonly cited type of this environmental error is that which is caused by wind, in the phenomenon known as ‘wind-induced undercatching'. Rainfall measured by a rain gauge at a height should be a true representation of what would have actually hit the ground if the gauge was not present. The trajectories of precipitation particles become distorted in a wind through the displacement and acceleration of wind flow over the top of the gauge as caused by the aerodynamic blockage by the gauge body (Goodison et. al., 1998). One method of reducing this error is by using rain gauges with an aerodynamic profile (Strangeways, 2004). Another complication when attempting to compare like-for-like rainfall measurements is caused by the random spatio-temporal variation of a rainfall field. For instance, if two rain gauges are situated some distance apart it is impossible to be assured that the rainfall they are both measuring is homogenous at the two positions in space and time.

To achieve a fair comparison in a field measurement campaign, it is necessary to deploy a number of identical instruments which are effectively co-located. A detailed rainfall experiment is set out to study the effects and understand the causes of the catching errors (Pollock et al., 2014). Four sites are instrumented, representing typical rainfall regimes common throughout the United Kingdom. By comparing high resolution rainfall data measured in a Reference Rain Gauge Pit against conventional UK ground-based mounting practices, it is possible to study the differences in ‘catching type' errors due to environmental factors such as wind. Data from Talla Reservoir in the Scottish Borders are presented. The site was chosen for its strong exposure to wind and because it is a situated in a very wet location. Measurements from aerodynamic and conventional cylindrical gauges, mounted at a variety of different heights, are compared across a number of rain events. Data obtained from this site are analysed and discussed in a statistical context, and the results are presented.