Apparent inaccurate readings

The commonest support issue with rain gauges is suspected inaccurate readings. In many cases the rain gauge is actually not at fault at all, rather it’s that the user does not appreciate the local variability of rainfall or the importance of correct exposure and of regular cleaning or does not recognise the limitations of all tipping bucket rain gauges. (Rainfall is probably the most tricky of the main weather parameters to measure automatically to good accuracy.) However, if the gauge is failing to register rainfall at all or is dramatically over-estimating rainfall then see the separate Fixing Rain Gauge Faults topic.

This is not the place for a detailed description of how to measure rainfall accurately, but some key points on factors affecting accuracy would include:

  • Remember: Daily rainfall readings on Davis stations re-zero automatically at midnight. Comparing readings with manual gauges that may be read at eg 8am must take account of the different time periods used. For this reason, monthly totals are often the best comparator;
  • The gauge must be clean and level to measure accurately;
  • The gauge must be well-exposed to the sky with no nearby buildings, trees, fences etc. It’s often impossible to satisfy this requirement fully in a typical domestic situation, but the aim will be to achieve the best exposure possible in the circumstances and to remember that even residual objects in the vicinity of the gauge will potentially have some impact on recorded rainfall;
  • Rain gauges should ideally be mounted close to ground level (or protected by a wind break wall if mounted on a flat roof) – siting the gauge at any significant height will expose the gauge to wind effects and hence typically cause significant under-recording of rainfall. (This is unavoidable to an extent with an all-in-one design like the Vantage Vue where some compromise between wind and rain measurements will always have to be made);
  • Rainfall totals can vary widely over a short distance, especially during storms or showery weather – the total from another gauge even a mile or less away can provide a very misleading comparison. Checks against a reference gauge should ideally be made with the funnel openings at exactly the same height and with the funnels very close if not touching. Even separating the gauges by eg 50-100m can lead to significant differences in recorded rainfall;
  • Rainfall rates can vary considerably with different types of rainfall event, ranging from very light drizzle to torrential downpours and automatic gauges can have problems with under-recording at either end of the rainfall rate scale. In very light rain/drizzle, there can be evaporation of partially filled buckets, while extremely heavy rain eg during a thunderstorm can overwhelm the gauge to an extent because of out-splash from the cone and missed collection while the buckets are in the process of tipping;
  • The minimum amount of rain that can be measured is 0.2mm or 0.01” (depending on type). Even 0.1mm rain is sufficient to wet pavements and to be felt as fine drizzle, but will typically not register on the gauge. Conversely, a heavy morning dew can often lead to collection of 0.2mm water in the gauge, occasionally more; this is a natural phenomenon and is usually legitimately included within rainfall totals;
  • You need a significant amount of rainfall (eg 5mm or preferably substantially more) when comparing readings from two gauges. A check with 1-2mm rainfall or less is fraught with problems;
  • Measuring the liquid equivalent of snow may be problematic – this is unavoidable short of installing a rain gauge heater (which is available as a spare part for the VP2 gauge but does need access to an AC mains power supply). For small amounts of snow, eg up to 5-10cm, the snow will accumulate in the funnel and will be measured when the thaw comes – rainfall equivalent amounts will be roughly correct, but timing of the precipitation obviously will not be. For larger snowfalls then the excess snow will obviously tend to overtop the funnel and, inevitably, will escape measurement.

If you still suspect a rain gauge problem

Rain gauge maintenance – the importance of cleaning

Assuming that the gauge was installed in a good level plane and has not been disturbed, then cleaning is the only routine maintenance needed for the rain gauge. But cleaning as often as necessary is vital for rainfall accuracy.

In particular, check that the funnel and debris screen are clean and that the central hole is fully clear and unobstructed – depending on local circumstances this check may need to be made relatively often. The updated cone with provision for inserting bird spikes can sometimes help considerably – the new style cone is available as a spare part, complete with spikes, to fit existing VP2 gauges.

Check also that the buckets are clean and can tip without impediment – it’s not uncommon for spider webs, for example, to impair tipping and potentially to prevent tipping altogether – even a strand or two of spider silk is surprisingly strong. Judicious use of insecticides may be helpful in preventing recurrent problems with webs, insect nests etc.

NB Oiling the tipping mechanism is rarely necessary and can actually hinder accuracy by attracting pollen and other particulate matter.

Check operation manually

A simple procedure for checking operation manually is described in the Fixing Rain Gauge Faults topic.

Check console and software rainfall increment settings

If the readings are consistently out by 20-25% (relative to a check gauge) then check that the wrong calibration setting has not been used, eg the gauge increment may be 0.2mm (metric calibration) while the console setting is 0.01” (US units) or vice versa. The calibration setting can usually be found on a small round label either on the outside of the cone when new, or on the base of the gauge inside the rain cone, but of course may have been lost with the passage of time.

All Davis stations supplied in the UK market and specifically marked as the UK version (eg 6152UK or 6152CUK) for the past several years have all had metric 0.2mm calibration, but you may have an older station with 0.01” increment. This is also true of OV ISS units purchased separately. However, if you have purchased a cabled ISS or simply a replacement rain gauge then it’s important that it have the letter ‘M’ in its part number if you need the 0.2mm increment, ie so you would need to specify for example 6322CM for the cabled metric rain gauge or 7852M for the metric gauge. If the letter ‘M’ is missing then you should expect to receive the 0.01” gauge.

Calibration checks

The best way of checking the calibration of a rain gauge is to set up a decent quality manual gauge very close by to the Davis gauge and with its opening at the same height – proximity and height are very important to a reliable comparison. Cheap gauges eg as sold in garden centres are unlikely to be accurate enough for anything other than a very rough and ready check. Prodata sells an inexpensive quality plastic gauge known as a CoCoRaHS gauge that is an internationally-recognised standard. Even manual gauges need careful use – read and empty the gauge regularly to avoid losses due to evaporation. And run calibration trials over several significant rainfall periods – one brief shower is obviously unlikely to provide a good check.

Check that any comparison is made on data that covers precisely the same time period. For example, the Davis consoles reset the Daily Rainfall reading to zero at midnight every night whereas the traditional observation day for manual readings (which is still widely used in eg the UK) runs from 09-09 hours. As a result, if a heavy rainfall event spans the midnight period, then the nominal daily totals for the two sets of data can be very different even if both gauges were perfectly accurate. Significant differences in monthly totals can also be seen for similar reasons if a heavy rainfall event happens to continue across a month-end;

An alternative calibration check is to run a ‘drip test’. This involves allowing a known volume of water to drip slowly into the funnel and then observing the rainfall total on the console. It turns out that if 544ml (18.39 ounces US) of water are dripped slowly – eg over a period of 30-60 minutes – and completely through the gauge, a reading of 1.00” or 25.4mm should result. (Practical details: Outside of a lab, it does take a little ingenuity to think how best to drip this volume of water. One suggestion is to use a clean plastic drinks bottle with a hole made in the bottom by a needle or tack. It may be more convenient to weigh the volume of water than to measure its volume by assuming that 1ml = 1g. You may obviously need to try one or two test runs to get the drip rate right.

If, having read all of the information on this page and having carried the tests suggested, you’re still convinced that your gauge would benefit from recalibration then it is possible to adjust the calibration by means of the stops underneath each end of the Vantage Pro 2 rain gauge buckets. We strongly suggest that you only change the calibration in this way as a last resort. Users often move the stops prematurely, only to discover that there was a different explanation altogether for their presumed rainfall inaccuracy. But, by then, it’s too late. The original factory setting of the stops has been lost and it’s very difficult to return exactly to the original settings. If, nonetheless, you still wish to press ahead with changing the height of the stops, there are two key facts to be aware of:

  • Each counterclockwise movement of the stops by one flat of the hexagon raises the stop and hence increases the rainfall recorded by the gauge by approximately 1% (per flat). So one complete revolution of the stop would cause roughly a 6% increase in recorded rainfall. (Raising the stop – ie a movement in the +ve counterclockwise direction – decreases the weight of rain needed to cause a tip of the buckets and hence a given rainfall amount generates more bucket tips and thus a higher rainfall reading);
  • Each of the two stops must be adjusted by an equal amount (and obviously in the same direction);

If you do go ahead and change the stop height significantly then you may well wish to rerun the drip test after making the adjustments to the two stops so as to check the new calibration.

Last modified: Apr 13, 2020


Was this helpful?

Yes No
You indicated this topic was not helpful to you ...
Could you please leave a comment telling us why? Thank you!
Thanks for your feedback.

Post your comment on this topic.

Post Comment