INTERCOMPARISON MEASUREMENTS BY THE GROUND LEVEL AND
ELEVATED DISDROMETER
Branislav Chvíla1 and Boris Sevruk2
1
Slovak Hydrometeorological Institute, Jeséniova 17, 833 15 Bratislava, Slovakia,
tel. +421 2 5941 5162, fax. +421 2 5477 4419, e-mail: branislav.chvila@shmu.sk
2 Former: Swiss Federal Institute of Technology ETH, Institute of Climate Research, Winterthurerstr. 190,
CH-8057 Zürich, Switzerland, tel. + 41 1 635 5235, fax. + 41 1 362 5197, e-mail: boris.sevruk@env.ethz.ch
Abstract
The number of raindrops of various sizes was measured using ground level and elevated
disdrometers during the summer season 2007 at the experimental station of the Slovak Hydrometeorological
Institute in Jaslovske Bohunice. The number of raindrops registered by disdrometers in 1-minute intervals
varied from a few units up to the several thousands of raindrops per minute, which corresponds to the
precipitation intensity between trashes and as much as 2.62 mm.min-1 respectively. The ground level
instrument protected again in-splash showed systematically a greater number of raindrops. The average
difference for a total of 615 15-minute intervals related to the elevated gauge measurements was about 9 %.
The difference between the corresponding precipitation amounts was 33 mm, i.e. more than 16 %. In spite of
losses due to wind, the disdrometers still seem to overestimate the rain amounts, particularly during higher
intensities, as resulted from the intercomparison measurements using other ground level and elevated,
Hellmann and automated weighing gauges installed at the site. They showed smaller differences (4-6%)
between the corresponding precipitation amounts. The effect of wind speed is evident either in the increasing
difference of the recorded number of raindrops and precipitation amounts between the paired disdrometers
with the increasing wind speed. The relative difference increases with wind speed from about 13 % up to
22 %. The preliminary results show that in the case of rains composed of smaller raindrops the wind effect is
greater than in the case of rains with bigger raindrops, which agrees well with the theory.
Keywords: disdrometer, intercomparison, precipitation measurement, wind effect.
INTRODUCTION
The Slovak Hydrometeorological Institute (SHMI) continues with intercomparison
measurements of precipitation recording systems by comparison of a new laser-optical
disdrometers. This paper follows the previous works of both authors in the field of study of
systematic and random errors of new recording precipitation gauges [1,2,3,4,5]. The aim of this
study was to check the operational reliability and accuracy of the new system, particularly the
influence of the ambient meteorological conditions (wind and temperature).
METHODS
The intercomparison measurements have taken place at the experimental station in
Jaslovské Bohunice (Figure 1), which is located at the plain field about 60 km northeast from
Bratislava, Slovakia, since July 2007. The site is equipped by the disdrometers Parsivel by Ott
Messtechnik (Figure 2), the recording weighing gauges TRwS500 by MPS System Ltd., and the
reference manual standard gauges of Hellmann, which is measured with the resolution of 0.1 mm
three times per day. All of the gauges are installed in the pair, one with the orifice rim at ground
level protected again in-splash as recommended by the WMO, and the other at the 1 m height
above the ground, which is the standard for precipitation measurements in Slovakia.
1
Figure 1. The pit and elevated precipitation gauges and the wind speed sensor at the experimental station at
the observatory of SHMI in Jaslovské Bohunice, Slovakia.
Parsivel is a modern, laser-based optical system for measuring and classifying all types of
precipitation. Baseline are measured the size and velocity of the fallen precipitation elements, from
which are derived the compared amount of precipitation and the particle’s size spectrum [6]. The
resolution of the weighing gauge is 0.001 mm.
Figure 2. The elevated disdrometer installed at the experimental polygon in Jaslovské Bohunice.
The wind speed was measured using the instrument installed also at 1 m height above the
ground. The weather phenomena and clouds observations from SYNOP messages were used to
separate between the convective and non-convective precipitation as well as between liquid and
solid ones.
The analysis was based on 15-minute intervals each consisting of (i) the total number of
recorded precipitation particles, (ii) corresponding precipitation amounts for distrometers and
weighing gauges, (iii) average wind speed and wind direction, (iv) mean ambient air temperature.
In addition, the type of precipitation genesis, stratiform or convective, was assessed. Events with
2
recorded precipitation at times when no precipitation was observed were eliminated. The specific
cases when the difference between the number of recorded particles by the pit and elevated
distrometers did not agree with the computed precipitation totals will be separately analyzed.
The differences between the pit and elevated disdrometers were subdivided according to
the wind speed, wind direction, air temperature and the type of precipitation genesis. They were
compared with precipitation amounts as recorded at the same time by weighing precipitation
gauges.
RESULTS AND DISCUSSION
The number of registered rain drops in the one minute recording intervals varied from few
units up to several hundreds, in the extremes up to thousands of raindrops. It corresponds to the
precipitation intensities from trashes to more than 2.5 mm.min-1. The comparison of the total
number of rain drops and related precipitation amounts between both disdrometers showed that
the ground level instrument gave usually a greater number of rain drops as elevated one. This
difference increased with increasing wind speed as is shown in Table 1. The relative difference of
the registered number of raindrops varied from 4.5 % for the group of calm up to approximately
14 % for the group of the strongest winds. The relative difference of the precipitation totals varied
from 13 % to 22 %. It agrees well with the theory, that the falling precipitation particles are more
affected by stronger winds like by the breeze and consequently, the elevated instrument registered
less precipitation than that in the pit. Evidently, the difference is generally great for the calm group
too. This could be explained either by still possible in-splash or due to the failure of one of the
instruments.
Table 1. The increase of the percentage difference of the recorded number of raindrops and the
corresponding precipitation amounts between the ground level and the elevated disdrometer, related to the
elevated one, with increasing wind speed. Jaslovské Bohunice, Slovakia, July-October 2007.
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
WS group
WS avg
n
N0
N1
RA0
RA1
DN
DRA
[m.s-1]
[m.s-1]
[]
[]
[]
[mm]
[mm]
[%]
[%]
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Calm
0.0
93
202183
193008
31.08
27.51
4.54
12.98
0–1
0.5
124
287235
272841
45.15
39.39
5.01
14.62
1–2
1.5
127
354959
327848
47.75
40.09
7.63
19.11
2–3
2.5
120
257319
230913
49.40
42.33
10.26
16.70
3–4
3.5
70
165716
144959
33.50
28.83
12.52
16.20
>4
4.8
81
179919
154779
21.54
17.66
13.97
21.97
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
WS = wind speed
avg = average
n = number of considered cases
N0 = number of rain drops registered by disdrometer at ground level
N1 = number of rain drops registered by disdrometer at 1 m height
RA0 = rain amount of disdrometer at ground level
RA1 = rain amount of disdrometer at 1 m height
DN = difference between registered rain drops number of disdrometer 0 and disdrometer 1 (relative related to elevated gauge)
DRA = difference between rain amount of disdrometer 0 and disdrometer 1 (relative related to elevated gauge)
The similar comparison was made also for the non-convective and convective precipitation. The
results are showed in Tables 2a and 2b. Overall there are not significant differences in the results
between these two groups of precipitation genesis, nevertheless in the cases of light winds the
difference is smaller for the convective precipitation. It can be interpreted in a way that the bigger
and heavier drops of convective precipitation are less affected by the wind. However, number of
considered events in the group of convective precipitation was not statistically significant.
3
Table 2. The increase of percentage difference of the recorded number of raindrops and the corresponding
precipitation amounts between the ground level and elevated disdrometer, related to the elevated one, with
increasing wind speed. Jaslovské Bohunice, Slovakia, July-October 2007.
a) Stratiform precipitation only.
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
WS group
WS avg
n
N0
N1
RA0
RA1
DN
DRA
[m.s-1]
[m.s-1]
[]
[]
[]
[mm]
[mm]
[%]
[%]
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Calm
0.0
74
110821
105481
12.67
10.83
4.82
16.99
0–1
0.5
108
243247
230564
36.66
32.09
5.21
14.24
1–2
1.5
114
327395
301274
43.60
36.63
7.98
19.03
2–3
2.5
96
208966
189714
34.80
30.43
9.21
14.36
3–4
3.5
54
125318
111272
22.58
19.62
11.21
15.09
>4
4.8
69
156253
134060
19.12
15.59
14.20
22.64
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
b) Convective precipitation only.
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
WS group
WS avg
n
N0
N1
RA0
RA1
DN
DRA
[m.s-1]
[m.s-1]
[]
[]
[]
[mm]
[mm]
[%]
[%]
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Calm
0.0
19
91362
87527
18.41
16.68
4.20
10.37
0–1
0.4
16
43988
42277
8.49
7.30
3.89
16.30
1–2
1.4
13
27564
26574
4.15
3.46
3.59
19.94
2–3
2.6
24
48353
41199
14.60
11.90
14.79
22.69
3–4
3.6
16
40398
33687
10.92
9.21
16.61
18.57
>4
5.0
12
23666
20719
2.42
2.07
12.45
16.91
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
WS = wind speed
avg = average
n = number of considered cases
N0 = number of rain drops registered by disdrometer at ground level
N1 = number of rain drops registered by disdrometer at 1 m height
RA0 = rain amount of disdrometer at ground level
RA1 = rain amount of disdrometer at 1 m height
DN = difference between registered rain drops number of disdrometer 0 and disdrometer 1 (relative related to elevated gauge)
DRA = difference between rain amount of disdrometer 0 and disdrometer 1 (relative related to elevated gauge)
Because the disdrometer has not symmetric construction in circle (like the conventional rain
gauge orifice) the effect of wind direction on the results of the elevated instrument was also
analyzed (see Table 3). The major differences of the registered number of raindrops between the
pit and elevated disdrometers occurred in two quadrants oriented across the lengthwise axial line
of the instruments. The greater precipitation totals were measured in these two quadrants. Though,
the difference of the rainfall amounts is about 3 – 5% smaller in these quadrants than in the upright
ones. This can be explained by the splashing of the drops on the covers of the sensor’s arms and
the effect of eddies around them. If the wind flows across the lengthwise axis the smaller splashed
drops are going out of the measuring area in contrast to the splashed drops in the case of the
parallel wind direction. The reverse difference in the precipitation amounts can be explained with
the same principle. The sensor’s arms create obstruction for the falling particles, and during the
precipitation events when wind flows along the lengthwise axis of the instrument, the windward arm
keeps some portion of these precipitation particles and therefore sensor records smaller amount of
precipitation.
The effects of ambient air temperature on differences are presented in Table 4. The relative
difference in the registered precipitation particles varied between 2.3 % and 5.5 % in the
precipitation amounts between the different air temperature groups. There is not apparent relation
between the magnitude of the differences and the magnitude of the air temperature.
4
Table 3. The effect of wind direction on the number of registered rain drops and respective rain amounts and
the relative differences of them, all between the pit and elevated disdrometer, related to the elevated one.
Jaslovské Bohunice, Slovakia, July-October 2007.
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
WD
WD avg
n
WS range
WS avg
N0
N1
RA0
RA1
DN
DRA
[deg]
[deg]
[]
[m/s]
[m/s]
[]
[]
[mm]
[mm]
[%]
[%]
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
360 – 090
35
62
0.3 – 5.3
1.9
90229
82579
11.02
9.15
8.47 20.44
090 – 180
129
53
0.1 – 5.2
1.6
162284
143944
28.92
25.15
11.30 14.99
180 – 270
239
81
0.1 – 3.6
1.6
145170
135516
24.42
20.66
6.65 18.20
270 – 360
303
320
0.1 – 7.2
2.7
835721
757926
131.20
111.91
9.30 17.24
Calm
93
0.0
202183
193008
31.08
27.51
4.54 12.98
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
WD = wind direction
WS = wind speed
avg = average
n = number of considered cases
N0 = number of rain drops registered by disdrometer at ground level
N1 = number of rain drops registered disdrometer at 1 m height
RA0 = rain amount of disdrometer at ground level
RA1 = rain amount of disdrometer at 1 m height
DN = difference between registered rain drops number disdrometer 0 disdrometer 1 (relative related to elevated gauge)
DRA = difference between rain amount of disdrometer 0 and disdrometer 1 (relative related to elevated gauge)
Table 4. The effect of ambient air temperature on the number of rain drops and respective rain amounts
registered by pit and elevated disdrometer. Jaslovské Bohunice, Slovakia, July-October 2007.
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
T group
T average
n
N0
N1
RA0
RA1
DN
DRA
DRA
[°C]
[°C]
[]
[]
[]
[mm]
[mm]
[%]
[mm]
[%]
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
5 – 10
9.1
149
439858
407076
68.57
60.47
7.45
8.10
13.40
10 – 15
12.8
241
692304
630772
101.17
85.98
8.89
15.19
17.67
15 – 20
17.4
167
256387
231519
49.76
41.78
9.70
7.98
19.10
20 – 25
22.0
47
45779
42307
6.90
5.98
7.58
0.92
15.38
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
T = ambient air temperature at 2 m height
n = number of considered cases
N0 = number of rain drops registered using disdrometer at ground level
N1 = number of rain drops registered using disdrometer at 1 m height
RA0 = rain amount of disdrometer at ground level
RA1 = rain amount of disdrometer at 1 m height
DN = difference between registered rain drops number of disdrometer 0 and disdrometer 1 (relative related to elevated gauge)
DRA = difference between rain amount of disdrometer 0 and disdromerer 1 (relative related to elevated gauge)
Wind speed average in all groups varied between 1.8 and 2.3 m.s -1.
The intercomparison of the registered precipitation amounts between the disdrometers and
weighing rain gauges showed that disdrometers systemically overestimate the precipitation totals.
Significantly overestimation is obvious for the pit disdrometer as is shown in Table 5.
Table 5. The dependence of the relative difference of rain amounts registered by disdrometers and weighing
raingauges on the wind speed. Jaslovské Bohunice, Slovakia, July-October 2007.
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
WS group
Calm
0 – 1 m.s-1
1 – 2 m.s-1
2 – 3 m.s-1
3 – 4 m.s-1
>4 m.s-1
All speeds
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
n[]
88
118
127
120
70
81
604
ground level
DRA [%]
30.32
28.61
25.59
36.46
33.41
43.48
31.83
elevated
DRA [%]
15.28
12.41
10.32
22.77
20.77
34.14
17.46
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
WS = wind speed
n = number of considered cases
DRA = relative difference between rain amount of disdrometer
CONCLUSIONS
The effect of wind speed and direction on the measurements using disdrometers was
evident in the increasing difference of the recorded number of raindrops and precipitation amounts
between the paired ground level and elevated disdrometers with the increasing wind speed. In the
case of rains composed of smaller raindrops the wind effect was greater than in the case of rains
5
with bigger raindrops, which agrees well with the theory. Differences occurred also between
different quadrants of wind directions, the oriented across the lengthwise axial line of the
instruments and other ones parallel to it. The greater precipitation totals were measured in both of
the quadrants oriented across the lengthwise axial line. The difference of totals in these quadrants
is about 3 – 5 % smaller than in the upright ones. In spite of losses due to wind, the disdrometers
still seem to overestimate the rain amounts, particularly during higher intensities, as resulted from
the intercomparison measurements using other ground level and elevated, Helllmann and
automated weighing, gauges installed at the site. The effect of type of precipitation, convective or
stratiform, on differences was small.
ACKNOWLEDGEMENT
Many thanks to Ott Messtechnik company for lending the Parsivel disdrometers to Slovak
Hydrometeorological Institute in order to provide the intercomparison measurements and study of
reliability and accuracy of these equipment.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
Chvíla, B., Ondráš, M., Sevruk, B., 2002: The wind-induced loss of precipitation
measurement of small time intervals as recorded in the field. In: WMO/CIMO Technical
conference 2002. WMO Instrument and Observing Methods Rep. No. 75, WMO/TD-No.
1123, Geneva. CD ROM edition.
Chvíla, B., Sevruk, B., Ondráš, M., 2005: The wind induced loss of thunderstorm precipitation
measurements. Atmospheric Research, 77: 29-38.
Chvíla, B., Sevruk, B., Ondráš, M., 2005: Intercomparison measurements of recording
precipitation gauges in Slovakia. In: WMO/CIMO Technical Conference 2005. WMO IOM
Report No.82, WMO/TD-No. 1265, Geneva. CD ROM Edition.
Chvíla, B., Sevruk, B., 2006: Effect of small intensity precipitation on wind induced error of
precipitation measurement. In: Proceedings of 7th international workshop on precipitation in
urban areas, St. Moritz, Switzerland, 7-10 December 2006.
Sevruk, B., Chvíla, B., 2003: Error sources of precipitation measurement using electronic
weight systems. Atmospheric Research, 77.
Nemeth, K., Hahn, J.M., 2005: Enhanced precipitation identifier and new generation of
present weather sensor by OTT Messtechnik, Germany. In: WMO/CIMO Technical
Conference 2005. WMO IOM Report No.82, WMO/TD-No. 1265, Geneva. CD ROM
Edition.
6