The strange choice of an elevated design of these Davis AWS gauges
will lead to under-reading owing to wind eddying effects, as Will
correctly points out. The difference between this type of gauge and a
standard adjacent checkgauge (rim at 30 cm) will vary significantly
with both site and weather conditions, so a windy wet day in an
exposed site will lead to greater losses than a day's rain with light
winds - the differences could then be expected to vary from almost nil
to perhaps 30% or more, and adjusting the tip capacity may not be the
right answer.

Sheltered sites may see much smaller differences between these gauges
and a standard checkgauge - but of course both may be under-recording
if the site is seriously under-exposed, particularly to the main rain-
bearing winds.

I'd suggest before tinkering with the calibration settings themselves:

1. Maintain a standard five-inch checkgauge in a reasonable exposure
to provide the benchmark against which to judge the accuracy of the
tipping-bucket readings. They're not expensive, perhaps £150 for a new
one, less second-hand (check on e-bay): it'll last 50 years with a bit
of care. Don't forget the measuring cylinder too.

2. Do the drip test as outline in my earlier post. If this comes out
OK, and 5% is probably the limit of accuracy without laboratory
equipment, look elsewhere for the reasons for any under-reading.

3. If your exposure is limited, check the readings from your
checkgauge with that of a local gauge in a standard location. Readings
can vary significantly over short distances in hilly locations, and in
showery situations, so compare over weeks and months rather than days.
There's enough COL sites in the major population areas to get an idea
of the true fall in your locality, or ask for help on this site. If
you feel your readings show a previously unsuspected dry spot in your
neighbourhood, check against large-scale average annual rainfall maps.
If your percentage of average over 3-6 months is significantly
different from nearby gauges, your gauge is under-reading, probably
owing to sheltered exposure.

4. If the checkgauge readings look OK, the calibration test is OK and
you're still under-reading on your TBR, it's your exposure - most
likely the height that's causing the problem. Use the elevated gauge
readings only for time/intensity readings and use the checkgauge as
the absolute values, or if you want automatic gauge readings to agree
more closely, disconnect the elevated gauge and plug in a standalone
0.2 mm unit instead - mounted on the ground.

HTH.

Stephen Burt
Stratfield Mortimer