A STUDY/COMPARISON OF A STANDARD
RAIN GAUGE v THE DAVIS VPro2

I have had a davis Vpro2 for a few years now and whilst a very good
bit of kit, I always had a slight nagging feeling that the rainfall
amounts under recorded and a certain knowledge that the number of rain
days over recorded - but to be fairly sure you need a standard copper
rain gauge to be able to make that comparison and whilst weatherlink
does not allow for 0900-0900 readings, the console does give a 'last
24hrs' reading for rainfall - so a direct comparsion is fairly easy to
make.
when i then read Stephen Burts quite excellent study comparing all
aspects of the davis to standard instruments - it nagged away at me
even more and a standard rain gauge came top of my things to have.
I don't know if the person who supplied me with the copper rain gauge
is likely to read this (on this site) - but thankyou for the loan of
the gauge.

The results for the year are below:

The period of this 'study' is 1st August 2010 to 31st July 2011.
Employed were a ‘Standard’ Snowdon copper 5” rain gauge and the Davis
VPro2.
The rain gauge was sited 30cm above ground and the VPro2 at 4ft (as
per Standard, for correct air temperature readings) – this in itself
places the VPro2’s rain gauge at some 5ft above ground and immediately
at a non-standard height.
Both were located in an open and exposed location and were some 4
yards distant from each other.

Three readings are provided:
1.) The rain gauge – with readings at the morning 0900 – 0900 hrs
obs
2.) The Vpro2 ‘normal’ readings of Midnight - Midnight
3.) The Vpro2 ‘last 24hours’ readings – which can be found on the
console when ‘daily rain’ is selected

The individual monthly totals are thus:
Figures in brackets are ‘rain days’ and ‘wet days’

Month Rain Gauge VPro2
‘normal’ VPro2 ‘last 24 hrs’
AUG 10 59.0 mm (15 – 9) 56.8 mm
(19 – 10) 57.4 mm (21 – 9)
SEPT 10 87.9 mm (17 – 15) 83.0 mm
(20 – 13) 83.2 mm (20 – 15)
OCT 10 89.4 mm (19 – 15) 82.8 mm
(24 – 14) 83.8 mm (25 – 15)
NOV 10 128.1 mm (23 – 18) 104.6 mm (27
– 14) 104.4 mm (25 – 18)
DEC 10 24.3 mm (13 – 8) 20.0
mm (16 – 4) 20.6 mm (16 – 5)
JAN 11 142.6 mm (18 -13) 130.8 mm
(23 – 13) 132.6 mm (24 – 13)
FEB 11 174.7 mm (21 – 21) 161.0 mm
(24 - 20) 162.6 mm (25 – 20)
MAR 11 59.9 mm (11 – 9) 47.4
mm (18 – 8) 57.4 mm (16 – 9)
APR 11 43.7 mm (9 – 3) 46.2
mm (9 – 6) 36.2 mm (10 – 3)
MAY 11 112.6 mm (24 – 18) 111.6 mm
(22 – 16) 111.6 mm (24 -17)
JUNE 11 65.8 mm (19 – 12) 63.4
mm (20 – 11) 63.4 mm (17 – 11)
JULY 11 69.6 mm (13 – 10) 66.0
mm (13 – 8) 66.2 mm (13 – 10)

Totals: 1057. 6
mm 973.6
mm 979.4 mm

Rain + Wet days: 202 -- 151
235 -- 137 236 -- 145

The standard rain gauge (SRG) recorded +7.98% more rain than the VPro2
‘last 24 hrs’.
The yearly difference between the two is quite possibly far less than
what was expected, particularly considering the open exposure and the
frequency of high winds at the site, although I do make some
observations regarding the effect that the wind may or may not have
had on the readings below .

On individual days the SRG would generally recorded greater rainfall
than the VPro2, although the margin between the two would differ
noticeably.
On days of less rainfall, eg: 3mm, the VPro2 may record anything from
2.2mm to 3.4mm and yet on occasions such as the wettest day recorded
in this period, 15.1.11., the SRG recorded 48.7 mm compared to 46mm
with the VPro2 – when one may have expected a greater margin.
Yet on 3.2.11. with winds upto 67 mph the VPro2 recorded 28.4 mm
compared to 27.1 mm in the SRG, when under such circumstances the SRG
would have been expected to have considerably more than the VPro2.
It would be mere speculation to suggest that in such strong winds the
Vpro2 recorded erroneous tips caused by the buffeting of the wind.
Then just three days later on the 6th with winds to 56mph the SRG
recorded 34mm compared to 28.4mm of the VPro2.
February was probably the month that provided for the most unusual
readings in the year. During the period 11-20th the VPro2 recorded
more rainfall on each and every day than the SRG, totalling 35.2mm to
28.7 mm of the SRG – although it must be said that for the month as a
whole the SRG still recorded more rainfall than the VPro2.
Whilst the SRG recorded more rain in each individual month than the
VPro2 (using the last 24 hrs) the monthly margins also differed
markedly:
MOST: November with 22.7%
LEAST: May with 0.9%

Indeed it is quite easy to detect a trend in that the ‘Summer months’
of May - July saw the smallest % margins and the greatest margins fell
more in the winter months (apart from December which saw snow and a
prolonged cold spell).
This might be easily explained by the winter months being that bit
more windy, but this site is exposed to the wind anyway and the mean
monthly wind speeds did not vary significantly to make this a valid
explanation.
Indeed MAY 2011 is now the record holder for the month with the
highest monthly mean wind speed at this site!
I don’t offer an explanation as to why the monthly margins do differ
by over 21%, but feel that the most likely explanation (wind) should
be placed into context (and possibly discounted).

However, the difference in the number of rain and wet days is
significant, although expected.
The ‘Last 24hrs’ having 16.8% more rain days than the standard gauge.
Whilst the rain/wet days are shown above I will now also include the
full data:

TOTAL RAIN
WET 5mm+ 10mm+ 20mm+ 25.4mm+

SRG 1057.6mm 202
151 69 31
7 5

VPRO2 973.6 mm 235
137 66 28
8 3

LAST 24 Hrs 979.4 mm 236 145
68 30 5
4


Possibly surprisingly is that whilst the rain/wet days differ
markedly, is that the 5mm, etc, days do not.
But a look at the individual days data does reasonably easily explain
the difference:

Days on which the SRG recorded a ‘Trace’ and the VPro2 0.2mm = 12
days
Days on which the SRG recorded nil rain and the VPRo2 0.2mm = 18
days
Days on which the SRG recorded nil, trace or 0.1mm and the VPRo2 0.2mm
+ = 5 days
Days on which the SRG recorded 0.1mm and the VPro 2 0.2mm = 7 days

The opposite to the above was far less:

Days on which SRG recorded 0.2mm or greater when the VPro2 recorded
nil = 8 days
Three such days occurred in December during the cold spell and when
snow fell, which the VPro2 did not record until a thaw set in.

In October the VPro2 recorded 0.2mm and the SRG nil on four occasions
when I marked ‘DEW’ as the reason for that VPro2 reading.
Ice forming in the VPro2 rain bucket on cold, rainless nights is also
another frequent source of additional rain days for the VPro2.

During snowfall events in November and December the difference between
both varied:
November saw the SRG record 11.3mm with the VPro2 having 5.4mm – but
by the month’s end there was still un melted snow in the VPro2.
Overall, including both November and December, the final amount for
both once the snow was melted was:
SRG = 15.9mm
VPro2 = 13.2 mm

Conclusions:
The VPro2 results compare favourably, although at times erratically
with the SRG.
Based on yearly results, the margin of error in the total rainfall
would be at a level that would be expected and probably tolerated.
However, the rain/wet days is too great a difference to be
acceptable.
However, it would be slight guess work to amend daily amounts and to
some extent the monthly amounts due to the erratic margins
experienced.
For accuracy (and peace of mind) a SRG is required though - but to buy
new is an added expense (£250-£300) that is not insignificant.

I should add that my site is a small village in Cumbria half way (5-6
miles) between Shap and Appleby