AN ABSTRACT OF THE THESIS OF
r
Richard Williarn Couch for the M. S. in
i5%r..i
Date thesis is presented
Title
Meteorology
May 15, 1963
SURFACE OBSERVATIONS OF THE ELECTRICAL
CHARGES RETAINED BY PRECIPITATION
Redacted for Privacy
Abstract approved
professor)
After cornparing theories of precipitation electrification,
this paper presents the design of a precipitation charge electrorneter and a point discharge current rneter. Data frorn these instrurnents make possible a statistical cornparison of cloud heights
and
electrical charge
on
precipitation. A 50-rninute period
of
cornparison shows a 37 percent dependency of precipitation charge
on cloud
heights. Analysis of five independent sarnples of charge
rnagnitude and drop size show a 45 percent interdependency.
Major lirniting factors include rnodifications of the drop charge
during descent and sampling of the drop size distribution at a rernote location frorn the precipitation charge measurement. Fur-
ther research should include a rnethod of rneasuring the variations
in precipitation charge during descent and sarnples during rnany
weather situations.
SURFACE OBSERVATIONS OF THE ELECTRICAL CHARGES
RETAINED BY PRECIPITATION
by
RICHARD WILLIAM COUCH
A THESIS
subrnitted to
OREGON STATE UNIVERSITY
in partial fulfillrnent of
the requirements for the
degree of
MASTER OT' SCIENCE
June,
1963
APPROVED:
Redacted for Privacy
AssociEte Pr
ssor of Physics
In Charge of Major
Redacted for Privacy
Chairman of
partrnent
Physics
Redacted for Privacy
uate Schoo
Date thesis is presented
May 15, 1963
Typed by Jolene Hunter 'W'uest
ACKNOWLEDGEMENT
The nature of the work presented in this thesis depends of
necessity on the efforts of rnany able and dedicated research parti-
cipants. In cognizance, the author wishes to sincerely thank Dr.
Fred W. Decker whose vision and forethought provided the opportunity and facilities for this researchi John D. Pembrook for illus-
trative photographs and technical aid; John V. McFadden for constructional assistance; John C. Plankinton, Jr. , Robert C. Larnb,
Jarnes W. Sears, and Donald M. Takeuchi for help in data reduc-
tion; Mrs. Patricia Pernbrook and Mrs. Laura Srnothers for editorial
and typing assistance; and
all the rnernbers of the Atrnospheric
Science Branch of the Science Research Institute, Oregon State
University for rrrany helpful discussions and for providing a stirnulating environrnent in which to work.
TABLE OF CONTENTS
Page
INTRODUC TION
I
THEORY
z
INSTRUMENTATION
5
SYNOPTIC CONDITIONS
t7
DATA
19
DISCUSSION
Z3
SUMMARY
34
CONCLUSIONS
35
SUGGESTIONS FOR FUTURE RESEARCH
38
BIBLIOGRAPHY
39
APPENDIX
4t
LIST OF FIGURES
Figure
I
Z
3
4
5
PRECIPITATION CHARGE ELECTROMETER
CONSTRUCTION
6
PRECIPITATIONCHARGEELECTROMETER
CIRCUIT
8
PRECIPITATION CI{ARGE ELECTROMETER
SYSTEM
IO
PRECIPITATIONCHARGEELECTROMETER
IN THE T'IELD
11
POINT DISCHARGE CURRENT MEASURING
SYSTEM
IZ
6
POINT DISCHARGE CURRENT AMPLIFIER
13
7
RADAR ANTENNA AND POINT DISCHARGE
ANTENNA ATOP McCULLOCH PEAK
15
RHI RADAR PHOTOGRAPH OF THE LEADING
EDGE OF'A LINE OT' PRECIPITATION
ZZ
8
9
IO
I
Page
I
IZ
A COMPARISON OT' CLOUD HEIGHT AND THE
CUBE ROOT OT' THE PRECTPITATION CHARGE
RATIO
?4
A COMPARISON OT' FUNCTIONS OT' PRECIPITATION HEIGHT AND PRECIPITATION CHARGE
DURING A FOUR DAY PERIOD
27
A COMPARISON
OT' RAINDROP DISTRIBUTIONS
DETERMINED T'ROM T'ILTER PAPER AND
ELECTROMETER MEASUREMENTS 1650 PST
? Mar 63
30
A COMPARISON OT' RAINDROP DISTRIBUTIONS
DETERMINED FROM FILTER PAPER AND
ELECTROMETER MEASUREMENTS OO25 PST
30 Mar 63
3I
SURF'ACE OBSERVATIONS OF' THE ELECTRICAL CHARGES
RETAINED BY PRECIPITATION
INTRODUCTION
Firrn establishrnent of the physical principles of cloud
and
precipitation forrnation depends on rnanrs ability to critically observe these processes at the proper tirne and place. Measuring
instrurnents carried by airplanes and sounding balloons suffer constant rnotion and hence have provided variable results. Radar,
ernploying short wave length electrornagnetic radiation, provides
an
ability to peer inside a cloud frorn a position rniles away.
The
patterns presented by radar, however, are those of precipitation
particles already on their trek towards earth. Consequently much
of the precipitation process including the initial forrnation and the
factors controlling the forrnation, rernain hidden.
A different approach is the exarnination of the fallen precipitation for properties retained since forrnation. For exarnple,
rnicro-analysis of the entrapped particles in a raindrop can contribute knowledge of the origin of the condensation nuclei. The tern-
perature variations during growth contribute to the history rrfrozenrl
in the structure of a snow crystal (13). This thesis reports on the
exarnination of the electrical charge as the historical residue retained by precipitation.
THEORY
Past studies tended to treat the elecLrical processes of the
atrnosphere as a separate entity. However, they are inseparably
related to the condensation and precipitation processes; indeed the
difference between cause and effect has not clearly ernerged (14).
Setting aside the question of the cause and effect relationship prior to the onset of precipitation, this study concerns the
rnechanisrn of continuing charge separation in precipitating curnuli
and to a greater extent in thunderstorrns. During the last half
century the rnany theories put forward to explain thunderstorrn
electricity have in no case withstood the accurnulation of rnore detailed inforrnation concerning storrn behavior.
Sirnpson (13) accounted
for production of electricity by
the
rnornentary contact of ice crystals during collision. The ice crys-
tals were assurned to receive a negative charge, the cornpensating
positive charge being carried upward on cloud droplets.
Workrnan and Reynolds (I7) proposed in 1948 a theory of
electrification based on experirnents which discovered that during
the freezing of dilute aqueous solutions a potential difference developed across the ice-liquid interface.
With the rnajority of solutions tested, the ice became nega-
tive with respect to the iiquid.
3
Mason (12) discounts these theories. Without denying thern
as possible, he observed that they do not account for the charging
rate required by observed thunderstorrrrs. Thus, Mason proposed
a new theory based on the preferential rnigration of positive ions
under the influence of a ternperature gradient.
H* and OH- ions, forrned by dissociation of a srnall fraction
of the ice rnolecules, separate under the influence of a ternperature
gradient. This process depends on the fact that the concentrations
of positive and negative ions increase rapidly with increasing tern-
perature with the OH- ion diffusing rnore slowly through the ice
crystal than does the H* ion. Thus, a steady ternperature differen-
tial across a piece of ice will result in a greater concentration of
negative ions at the warrn end. In the case of afreezing raindrop,
a shell of ice forms on the outside of the drop. The interior proceeds to freeze at a rate deterrnined by the dissipation rate of the
latent heat. In this rnanner a radial ternperature gradient forrns
as the drop cools to the air ternperature. Under these conditions
positive charges rnigrate toward the surface and leave a negative
charge at the center of the drop. The drop bursts because of expansion forces created as the liquid interior freezes, and the
splinters carry away the positive charges on the ice surface.
Hailstones, warlrrer than the supercooled cloud droplets
4
through which they fall also separate charges with the negative
charges residing on the hailstone. Charge separation occurs then
as basically a therrnoelectric effect.
Possibly the greater separation of electrical charges results from a precipitation particle rernaining in a region of sub-
freezing ternperatures during the growth process. The height of
the source cloud provides a suggested determinant of the tirne
spent by a precipitation particle above the freezing leveI.
Masonrs theory and others previously described postulate
negative charges on the frozen particles. If, then, these charges
are retained, or at least only proportionately rnodified during des-
cent, Eleasurernents of the charge polarity would indicate whether
or not the rain resulted frorn rnelting an ice forrn of precipitation.
A cornparison of the negative to positive charge ratio of
precipitation with cloud height above the freezing level should disclose whether earth based instrurnents can discern the initial charge
pattern.
INSTRUMENTATION
To test the above hypothesis and to provide additional inforrna-
tion concerning the electrical properties of the atrnosphere, the au-
thor designed, and with the aid of Atrnospheric Science Branch personnel, constructed a precipitation charge electrorneter and a point
discharge current rneter.
The precipitation charge electrorneter, shown schernatically
in Figure 1, consists basically of an electrostatically shielded collector, charge sensor, amplifier, and chart recorder. The precipitation falling into the collector deposits its characteristic charge on
the surface. There a static charge sensing electrorneter detects it.
An arnplifier then provides sufficient gain to drive the writing pen
of the chart recorder.
The wooden box containing the collecting funnel and rneasur-
ing apparatus has a total covering of copper screen, soldered at all
searns, to provide good electrostatic shielding. Low conductivity
polystyrene spacers separate the collector, a standard U. S. Weather
Bureau 8" tipping-bucket raingage, frorn the box. A truncated rnetal
cone placed over the funnel opening provides collirnation, splash pre-
vention, and additional electrostatic shielding.
The static field sensor, a high input irnpedance (tg-14
arrlperes grid current) electrorneter tube, converts the varying
6
Preclpltatlon
(
I
t
I
8-lnch Faln
GagE
lbnnel
*etal Electrostatl.c
,/
Shleld and Coll$oator
Polystyrreao
Saparator
Blectrostatlc
Srleldlng
hooden Box
---i
!&tai. rl---)
tuLeLd /
Tin6 !'larkor--+
and Zero Set
Electrornetar
I\rbs
Anpllfler
To
Carn
Motor
-
Earthlng
nod
FLg.
I
-f '--
l-
Screea only on Botton
PRECIPITATION C,HAIIG!: nLE0TR0METER CONSIRUCTION
7
static electric field to a varying electric current. A differential
arnplifier arnplifies the varying current and then converts it to a lower
irnpedance to drive the rernotely located chart recorder.
The cornbination of the low voltage plate operation of the elec-
trorneter tube and the ordinarily low voltage transistor arnplifier and
driver perrnit cornrnon battery operation with srnall power
dernands.
The rninirrrurn ternperature drift design of.this circuit (Figure 2)
should provide hundreds of hours of trouble-free field use under nor-
rnal arnbient operating conditions.
A small electric rnotor drives a carrr operated switch which
earths the input of the measuring circuit every fifteen seconds.
This provides a tirning rnark, a zeto level on the chart of the record-
er, and drains any residual charge on the collector to earth. An
earthing lead attached to the box shielding connects to a rod driven
into the earth at the location.
The circuit has an input tirne constant of approxirnately
fourteen seconds and a response tirne of two rnilliseconds. The input conductance and capacitance define the time constant, while the
recording pen deterrnines the response tirne. Lowering the input
resistance resulting in a tirne constant of 0.7 seconds, and reducing
the collecting aperture frorn 50 square inches to 25 sguare inches
reduced the static charge build-up during periods of intense rainfall.
r+
t{.|
0
e
zN
my
@Y!
loN
N
0
o
tn
z
N
h
i_d
x
-9
o
@
-t
F{
H
Ef
o
d
H
o
Fr
'i:{fr
31
o
&
'a)e{
l
ftl
J
Ll
.
CJ
d
,t
,o
'z
o
H
E:'
F
H
p{
,()H
3
C.
a
ol
u0
r{
ft
I
9
The precipitation electrorneter has a sensitivity of ?.0
x lO-4 eso
(0.7 x 10-13 coulornb) and a range of *3.3 x l0-2 esu (1. I x 10-I1
coulornb). The accuracy of the electometer-recorder systern is
4 10 percent of the chart reading.
The 8" tipping-bucket raingage also provides a rneasure of
the total rainfall on a separate recorder as shown in Figure
3.
Figure 4 shows the precipitation charge electrorneter in the
field.
f igure 5 outlines the rnethod of rneasuring the corona or
point discharge current. A high-gain direct current arnplifier arnpli-
fies the voltage drop across a one ohrn resistor located approxirnate-
ly four feet frorn the top of a 40 foot antenna. The top of the antenna
has a sharp point, thus increasing the electric field gradient and
facilitating the flow of current into the atrnosphere. The signal
generated across the one ohrn resistor travels to the high-gain D.
C. arnplifier via coaxial cable. The arnplifier provides sufficient
power to actuate a chart recorder with a sensitivity of 0. 5 rnilliarrlpere s.
I'igure 6 shows schernatically the electronic circuit of the
arnplifier. High gain circuits of this type generally suffer electrical drift, caused largely by therrnal variations in the surrounding
environrnent. To cornpensate for the inherent drift, the design
'-Lut
/
a a!
j
CIi.i:
i- t*::.i_.ii.ETt:
.iY:TEl{
li
fJ
11
Flg. b
PRECIPITATfON CHARCE ELECTR0IN THE FTELD
METER
POINT DISC}IARGI]
CUBREIIT
eCharge Dlatrlbutlon
ln G\rmuloniobl
l'(
I ,"r" L,barher
Electrlc Fleld
Gradlent
Pousr
Sspply
I{tgh Gala Ampllfler
Becorder
Flg. 5
ts
POINT DISCHAROE CURRENT MEASURING SXSTd,l
N
2N
trl j
I
Flg. 6
POINT DISCHARGE CURRENT A}{PLI!'IER
\,H
t4
includes a negative feed-back loop. This results in a long-terrn
stability of better than + 0. 5 rnillivolts per hour. A carn-operated
switch earths the input approxirnately every ten seconds to provide
a zeto reference rnark on the recording chart. The systern can resolve changes in point discharge current of two rnicroarnperes.
Esterline Angus chart recorders with a full scale deflection of +0.5
rnilliarnperes recorded the output of both the precipitation charge
electrorneter and the point discharge current rneter.
Observers took raindrop distribution rneasurernents using the
rnethylene blue irnpregnated filter paper technique. Assistants
counted by hand the dyed blots caused by the raindrops and deter-
rnined their sizes using a transparent overlay size gage.
A 3. Z crn wavelength Radar Set AN/CPS-9 recorded precipitation echo heights above the precipitation electrorneter. This radar,
located atop McCulloch Peak approxirnately 5.7 rniles frorn the pre-
cipitation recorder supplied range height (RHI) indications of precipitation echoes at approxirnately ten rninute intervals on March Z ar,d
March 27 through 30, 1963. Figure 7 shows the AN/CPS-9 antenna
and the 40 foot Point discharge current antenna.
On March 29 and 30 a vertically pointing 3.2 crn radar OS/
APR -21 atop Covell HalI on the Oregon State University carnpus
provided continuous precipitation echo heights (l 5). Photographic
L5
Fi.g'
7
IiADAii ANTENTJA AND POINT IiSCHARGE
ANTENNA ATOP McCULLOCH PEAK
t6
scope display recorders on both radars provided perrnanent records
of the precipitation echo heights.
The U. S. Weather Bureau office at Salern, Oregon furnished
atrnospheric soundings at l50OZ and 23002 on the days of observation.
Observers rnade rneasurements of wind velocity and direc-
tion, temperature, pressure, hurnidity, and rainfall on site.
continuor-r,s
The
recording weather instrurnents in the Physics Chernistry
building on the Oregon State University carnpus provided supplernen-
tary data.
t7
SYNOPTIC CONDITIONS
Cornprehensive data obtained on March 2, 1953 and during
the period
frorn March
Z7 through
30.
1963 allowed
reliable analysis
and cornparlson.
Synoptic analysis for March 2 showed that a cold front rnoved
across Oregon two days prior to the investigation and brought a
rnaritirne polar air rnass which dorninated throughout the investigation period. A high pressure systern off the southern coast of Oregon,
acting with a low pressure area on the southern Alaska coast, resulted in upper winds frorn the west-northwest.
The 23002 Salern sounding showed a rnoist air rnass with
conditional instability in the lower layers and a freezing level of
3000
feet. Records
showed
a
43oF.
ternperature with 84 percent re-
lative hurnidity, light and northerly winds, and cloudy skies with
rain showers west of the Cascades.
On the rnorning of March 27, a low pressure area off the
southwest coast of Oregon began to intensify and rnove toward Oregon,
replacing a rnaritirne polar air rnass. The resulting low and the
indistinct occluded front dorninated the synoptic situation until frontal passage at 1800 PST. Air flowed generally frorn the south to
southwest in the rnorning. The afternoon sounding revealed a rnoist
l8
but fairly stable atrnosphere with a freezing level of 6000 feet. Sur-
face ternperatures continued slightly below seasonal, with littIe tern-
perature change after frontal passage. Clouds with rnoderate precipitation predorninated during the afternoon and early evening.
At approxirnately 2300 PST another occluded front passed
over Oregon while still another passed over at 1100 PST on March
2,8. The passage of many lines of precipitation on the 28th provided
the significant difference in weather frorn the previous day. The
barorneter rernained low under considerably unstable conditions with
values averaging 1/Z inch lower than on the days prior to the Z7th.
Strong surface winds blew with gusts to 70 rniles per hour in places.
The I l0OZ sounding showed dryer air aloft with conditional instability
in the lower layer and a freezing Ievel of 3600 feet.
The weather of t}:e Z9th rernained under the influence of rnari-
tirne polar air. Afternoon and evening winds blew generally frorn
the west. The 23002 soundings showed conditional instability present with considerable rnoisture aloft and a freezing level of 3600
feet. In the late evening of t}:e Z9th and early rnorning of the 30th,
the barorneter showed large instability due to line passages. Short
heavy showers predorninated until late in the rnorning of the 3Oth.
The freezing level dropped to 3400 feet at the tirne of the IIOOZ
sounding.
I9
DATA
The Esterline Angus chart recorder graphically recorded
the rnagnitude and polarity of the precipitation charge. For anal-
ysis, the author sectored the recording into two-rninute periods
with readings of the rnagnitude and polarity taken at two-second in-
tervals, thus giving sixty possible sarnples per period. Extension
of the interval to four rninutes provided a rrlore reliable statistical
sarnple at tirnes of light or questionable precipitation. A tabulation
then facilitated analysis of the data in each two-rninute tirne sequence under the following headings:
I) Total nurnber of positive drops
Zl Total nurnber of negative
droPs
3) Total nurnber of positive drops of a given
rnagnitude
4l Total nurnber of negative drops of a given rnagnitude
5) Total nurnber of drops of an absolute rnagnitude
6) Largest positive charge per period
7l Largest negative charge per period
8) Ratio of negative to positive drops per period
A sarnple of this data is shown in Appendix L
On March Z
t.lne
tabulation shows rnaxirnrtrn positive and
ZO
negative charges of +11.0x10-3 esu and. -43. ?xl0-3 ."o respectively.
During the line passage on March 30 at 0100 PST +46.0x10-3 esu
and -35. 0xl0-3 .su appeared as the corresponding rnaxirna. Maxirrrurrr
positive and negative charges for the five days exarnined aver-
aged approxirnately *ZO. OxI0-3
esu. The positive and negative
charge for the total rain analyzed for the five day period averaged
approxirnately *4.5xI0-3 esu and -4. 2xI0-3
respectively. The
""rr
analysis sarnpled only the heavier showers, therefore, resulting in
higher averages than norrnally expected for continuous period sarrp-
ling. In general rnore positive than negative rain fe11, apparently
agreeing with observations of other researchers in atrnospheric
electricity (10)(I3).
The author read point discharge currents frorn the recording
in a sirnilar rnanner but plotted thern directly as a graph to give
a
tirne-cornpressed picture of the entire recording period.
Tirne lapse photographs of the AN/CPS-9 radar RHI scope,
taken at the sarne tirne as the charge recordings, furnished direct
rneasurernents of the precipitation echo heights. A continuous fikn
provided by the vertically pointing radar recorded precipitation
echo heights on the evening of March
30, 1963. The analyzer
sarnples this filrn at ten-second intervals and plotted it to give a
continuous tirne-cornpressed picture of the precipitation echo
zt
height.
Assistants in the Atrnospheric Science Branch counted and
tabulated the raindrops. Norrnalization of these distributions by the
ratio of the filter paper and charge collector areas and the ratio of
the exposure tirnes perrnitted statistical cornparisons. The author
then plotted a function of the ratio of the nurnber of negative drops to
the nurnber of positive drops against the cloud height for a',continuousrf 50 rninute period on March
30, L963. Fourteen selected sarn-
ples of charge ratios covering the four day period frorn March 2'l to
March 30, t!63 appear plotted against precipitation echo height in
Figure
10.
The statistical techniques of linear regression analysis provided a rnethod of arriving at a quantative rneasurernent of the inter-
relationships between the two variables (5). Regression line para-
rreters deterrnine the relationship of the equation. The square of
the correlation coefficient gives a rneasure of the relative arnount
of variation in the dependent variable that is explained by the independent variable.
The author corrrpared a function of the absolute charge on the
raindrop to the raindrop size distribution and after deterrnination of
atrbest fit", he took five sarnples on five different days to cornpare
statistically using the sarne relationship.
I::1 a'irrl *
T.i71I:
-', I
., -:':: l1'iii:,1.,t'll :l
r, ::,'l; -':TA:' l':
"
T:rI
N
N
23
DISCUSSION
Point discharge currents, recorded for the entire period, for
the tirnes of precipitation charge recordings, March 27 through
March 30, 1953 show a point discharge current very nearly steady
at approxirnately 9 - I4 rnicroamperes. Positive current flowed
into the pole, i. e. , electrons flowed frorn the rnountain to the at-
rnosphere. The current did not reverse and showed only slight increases and decreases associated apparently with wind variations.
During this tirne, precipitation charges had varying rnagnitudes, in-
tensities, and polarities. This agrees with the conclusions of Gunn
and Devin (9) that no relationship exists between point discharge
current and rain charge.
Observers reported hai1, snow, and the sounds of thunder on
the evening of March 28, 1963. During this tirne exceptionally large
point discharge currents exceeding 600 rnicroarnperes were rrreas-
ured. Multiple lightning strokes with norrnal field recovery tirnes
occurred, but no evidence appears of expected field revergais (It).
Possible explanations of these rneasurernents include these
two: t. The orographic and rnaritirne in{luences tend to diffuse
the
charge centers which occur in rrnorrnal" curnulonimbus. 2. In ef-
fect the corona point atop McCulloch Peak provides the terrninating
Preclpltatlon
Thousands
20
Echo Helght
of feet
A
Flg.9
@MPARISON OF CIOUD HEIGHT AND THii.I
qJBE ROOT OT THE PRECIPITATION
CTIARGE ITATIO
1g
(+f
18
3
L7
2
No
Preeipitatlon
PACITIC STAI{DARD TII.,IE 30 }IARCU 1963
5
25
point for the entire rnountain. One rnay expect that rnountain cur-
rents would have a certain arrrount of t'electrical inertiar', a distributed resistance-capacitance effect which could tend to support the
current flow in the absence of a field gradient and rnight even sustain is through a short-time field reversal. The data shows no unusual variations in precipitation cha::ge during this period.
operation of the vertically poi.nting radar enabled a continuous 50-rninute cornparison of the precipitati.on echo heights and pre-
cipitatron charge polarity ratio. The 50-rninute cornparison period
covers the passage of a line 15 rniles wide rnoving at a line velocity
of 15 miles per hour across the site of the precipitatron charge
electrorneter. Figure 8 shows the leading edge of this line March
30 at OOZ6Z as
it appeared on the RHI scope of the AN/CPS-9 radar
set.
Figure 9 shows echo height with tirne cornpared to the cube
root of the ratio of the nurnber of negative drops to the nurnber of
positive drops. The ratio of the nurnber of negative to positive
drops appears instead of the nurnber of a given rnagnitude since
the author believes that the rnagnitude depends rnore on the trajec-
tory than does the nurnber of drops of a given charge sign. For
exarnple, a large positive drop will, on the average, break into
about the same number of srnaller drops while falling as
will
a
Z6
large negative drop. The ratio, therefore, rernains the sarne.
However, the loss or gain of charge by a drop in falling through
a
gradient field depends on size, thereby changing the original ratio.
The cube root in this analysis provides a convenient ratio only.
Other powers would rnerely tend to reduce or exaggerate the con-
tour. Statistical cornparison at 22 pol'rfis gives a correlation coefficient of 0.61, the charge ratio, therefore, showing a 37 percent
dependency on cloud height.
In cornparing the two curves, the author cornputed rnean falI
tirnes for the measured drop sizes. Frorn heights
of.
?0,000 feet
drops fall to the ground in 15 rninutes. Offsetting the two curves by
this arnount of tirne destroys all vestiges of correlation. Best fit
occurs with the curves within +Z rninutes of each other, a value less
than the uncertainty of synchronization of the tirners on the charge
recorder and the radar. This indicates cloud or ceIl persistence.
The cloud evidently rnaintains its vertical identity for periods great-
er than the precipitation fall tirne.
Figure 10 shows a cloud height-charge cornparison for
14
tirne periods over four days. The lines connecting the points perrnit
cornparisons only, since the points are cornpletely independent. Not
randornly selected, these sarnples occurred when the charge rrreasurernent allowed a large nurnber of sarnples for a given period and
f"
I
I
x<
GuoTxclrnjr lq8TeH
uoTlE?TdTcord puB eSrBr{c uoTnB+TdTcaJd
69
06
TZ-5TTO
69 ur^I 06
w}{
6T-TITO
69 uv,.I 06
69-L900 69 uvt{ 06
f|-olao
69 uw[ 06
€9
69
{ffi
wt{
6Z
og
69 uTr( 06
9e-Z,eOO
ze-o€oo
+rz-a,oo
a€-6z0z
E-r Q
ido
(JH
HlI
hl frl
I4o{
TH
o<
HA
e{
<J6
tstp
HO
O.h
H
(J<
r"l
ri()
tu2
H
E{€
O=)
a
J)
aaI7
o(,
HlI
E-{ <
OE
z<)
69 uvn 6z
hFr
O E-t
T:'.6T0Z 69 uvr^I 5z
9T-6t02
€9 WA 6Z
dofrl
<
O. o{
Eg{
o
C)A
c)
5
l\2
90-9002
69 uw{ 8z
d
h
u0
F{
o
C)H
.ff O{
HH
au
zr-8097
Lz,
R
C9 Er,{..
€9 Wvt
l7-Zr9r
80-7097
Z7
z8
when the radar precipitation height rernained constant over the
charge recorder for periods of tirne comparable to the precipitation
fall tirne. The comparison used rrnorrnalizedrr cloud heights for
the
four day period determined by dividing actual height by the freezing
level height deterrnined frorn the Salern, Oregon radiosonde sound-
ings. After scale adjustrnent, statistical analysis showed a correlalation coefficient of 0. 74 or a ch.arge ratio dependency on cloud
height - freezing level
be
of. 54
percent and the following equation of
st fit:
L
F"J2
= 5.4 + 5.2
/N \ r
tq/
where:
H
C
cloud height
H.
t
freezing level
N
nurnber of negative drops
N
n
p
nurnber of positi-ve drops
According to the equation, if the freezing level should drop
to the surface, all ttre drops should have negative charge. In general
this does not occur (4). Nor does there always appear a rnixture of
negative and positive drops. A rnore satisfactory equation would have
the forrn:
(ffi)
"=B*Kcffi) '
29
with the constants A, B, C, D, and K dependent on geographical location and atmospheric conditions.
The charges on raindrops measure about one order of rnagnitude smaller than that necessary to produce corona on a spherical
drop (8). However, the charges approxirnate those expected on irregularly shaped drops. If this is the case, the absolute charge of
each drop would relate to the drop
size. Figures 11 and 12 show a
cornparison of the distribution of the cube root of the absolute charge
and the drop size
distribution. An alternate rnethod giving
the
same results would compare the drop area (or diarneter cubed) with
the absolute charge.
Because of the sarnpling technique used to
arrive at a
charge distribution, the nurnber of drops varies linearly with the tota1 counted
in the filter paper technique. The resulting constant of
proportionality corrects the number scale. The fact that the drops
splash on irnpact decreases the reliability of the drop size distribu-
tion on the small drop (Lefthand) end of the scale. The drop size
distribution at the large diarneter end of the scale depends on the
chance encounter of one
or two large drops, a f.act which also de-
creases its reliability.
Statistical cornparison of the curves of
Figures
11 and
l2 only in the more reliable center area of the distri-
bution gives correlations of .98 and .96 respectivelyr or better
than 90 percent interdependency. Sarnples taken on five different
30
Flg. L1 A COMPARISOII
0F n[INDn0P
DISTMBUTIONS DETEruINED
f,:I.trCTfiOMETER UEA SIRU.IEI{TS
15'O PST 2 MARffi 63
103
102
H
x--+(
TITTEB PAPER
ELEff8O{ETER
1.0
L.2
lrlrlrl,l,l
L.4
1.6
1.9
2.O
Qesu
DNOP
.8
L.2
1,.6
2.A
2.1
r
1g-1
DIAIIETER
2.8
(m)
3.2
3L
Flg. 12 A COMPARIS0N 0f
NAINDNOP DISTIUEI'TIONS
DEf,EAX'IINED
I'noU fItTEB
PAPEE AND EISCTNOMMEN
MEAqinE,rE{TS OO25 PSI
30 MARSI 63
H
SILTER PAPEE
x----x
EtECfnOUEf,ER
116,2?A,2t4,2f8,3i
Qecu
x LO-l
DnOP DIAMBIEB
2.6
3,A
(m)
3Z
days give a correlation for the total of
r
= 0.67
or a 45 percent inter-
dependency.
Certain of the sarnples, although taken sirnultaneously, involved separation by approxirnately 100 hundred feet. At the recording rate used on the precipitation electrorneter recorder the
larger drops tended to cover the srnaller drops during periods of
intense rain. Therefore, the strength of the correlation depends
sornewhat on the proxirnity of the sarnplers and on
rainfall intensity.
An equation which describes the suggested charge - drop
size relationship has the forrn:
Nd = A
loBclol
1
3
where: N,
o - the nurnber of drops of a given diarneter
A= A constant dependent on sarnpling rate
B= A proportionality constant between charge rnagnitude
and drop size
O
= Absolute rnagnitude of the charge
= Atrnospheric condition
factor
The constancy of charge with drop size indicates that either the drop
will support no rnore charge without corona discharge or that
the
drops experienced a constant charging rate during forrnation.
Drop charge distributions show a birnodal distribution in
four of the five cases exarnined. This agrees with Sarrnah's (15)
findings on drop distributions in western Oregon.
33
A qualitative cornparison of radar precipitation echo heights,
rain intensity, and charge ratio shows that in general the greater
rain intensity and the larger negative-to-positive charge ratios
occur near the leading edge of the precipitation echoes.
34
SUMMARY
Present existing theories of precipitation electrification suggest that the larger frozerr precipitation particles acquire a negative
charge during forrnation. The author designed a point discharge
current rneter to provide inforrnation to show that precipitation
charges are independent of point discharge current. He also designed and constructed a precipitation charge electrorneter to rneasure
the charges retained by precipitation at the earthrs surface.
The measured charges cornpared with cloud heights for a
continuous 50-rninute period and as sarnples over a four day period
show a statistical interdependency of.37 percent and 55 percent res-
pectively. A cornparison of charge rnagnitude and drop size
data
frorn five sarnple periods show a 45 percent interdependency.
Qualitative observations show that: l) the drop size variations
as
deterrnined frorn charge rneasurernents are birnodal; and 2) that the
rrrore intense rain and the higher percentage of negative drops occur
near the leading edge of the precipitation echoes.
35
CONCLUSIONS
In general, the prototype instrurnents perforrned satisfactorily
in rneasuring the atrnospheric electric parameters of precipitation
charge and point discharge current. The instrurnents need irnprovernents in dynarnic range and general construction.
Of the cases analyzed, the electrical charges retained by
precipitation have a distribution in tirne and polarity showing a de-
finite relationship between cloud height and charge ratio
and between
drop size and absolute charge rnagnitude. This indicates that the
surface observations of the electric charges retained by precipita-
tion are dependent on the source of the precipitation. If substantiated,
these findings will provide the research meteorologist and cloud
physicist with a new tool for studying the rnechanisrns of precipitation.
Referring to the radar photograph (Figure 8) of a precipitating
cell and rernernbering that in general the radar echo shows only the
precipitation and not necessarily the cloud structure surrounding
the precipitation, the reader can envision a cloud forward and above
the echo. If , then, air flows inward and upward at approxirnately
the 5000 foot level, the precipitation particles at the forward edge
of the cloud would tend to be sustained at higher elevations longer
36
and grow larger than those toward the
rear of the cloud. Likewise,
the reader rrray envision a cloud arranged syrnmetrically about the
precipitation echo with air entering the bottorn and rising through
the center to sustain large drops during the cloud growth stage.
Shearing of the cloud top during the rnature and dissipation stages
will tend to carry the precipitation downward and toward the front
of the
cloud. Either rnechanisrn could explain the drop size
and
drop distribution observed on the surface.
These results could substantiate either the Byers l2l or
Ludlarn (I) rnodels of the thunderstorrrr. Ffowever, atrnospheric
conditions reported by other researchers reveal cloud heights of
30,000 to 50,000 feet, bases near 4,000 to I0,000 feet and thicknesses
of thousands of feet. Clouds in western Oregon during this study had
heights typically 8,000 to 2O,000 feet with bases 800-Z5OO feet high
and precipitating cloud layers sometirnes only hundreds of feet in
depth. The presence of rnaritirne and orographic influences,
the
differences in raindrop size distribution, and the electrical evidence
previously cited all tend to show a dissirnilarity in cloud structure
between that generally cited in the
literature and that studied at
Oregon State University.
For these reasons the author refrains frorn fitting the evidence to an established rnodel, preferring to look upon
it as a
37
coastal cloud phenornenon - a class in itself - until further research
rnore sharply delineates the rnodel which the evidence prescribes.
A precipitation particle, during its flight towards earth rnay
colide, splinter, fteeze, pass through regions of strong electric
gradients, encounter ions, and rrrany other environrnental anornalies,
all of which tend to modify the original origin-dependent electric
charge.
Geographical differences in atrnospheric conditions previously
described would increase the probability of rnodifications to the
original charge patterns. For this reaeon, one probably cannot
extrapolate these results to other geographical areas without rnodification.
38
SUGGESTIONS FOR FUTURE RESEARCH
The author believes the irnplications suggested by this research
warrant verification. To deterrnine the rnodifications of the charge
on the drop while falling through the atrnosphere, the author proposes
to construct at least three precipitation charge electrorneters, placing thern in a line on the windward slope of a rnountain such that, as
nearly as possible, they would sarnple the sarne precipitation at different heights. Under certain conditions, this could achieve sirnultaneous rneasurernents of the precipitation charge both above and
below the freezing level. This could also provide rneasurernents at
the precipitation origin or place of forrnation inside the cloud as it
encornpasses the rnountainside recording
sites. This rnethod would
at least partially elirninate rnany of the questionable pararneters in
the above research.
For better statistical data collection, an apparatus sirnilar to
that used by Gunn (7) would irnprove the raindrop size-charge relationship data. His apparatus sirnultaneously rneasured raindrop
charge and size. The drops passed through an induction ring and
then irnpacted on a continuously rnoving filter paper belt.
The rnaritirne-orographic affects on the point discharge cur-
rent cause anornalies which need further investigation.
A rnost intriguing speculation arises in the question, 'rWl:at
natural weather rnodification occurs because of point discharge currents ernitted by the rnountains of the Coast Range as rnaritirne air
rnasses waft over thern?rr This could provide interesting research.
39
BIBLIOGRAPHY
l. Battan, Louis J. Cloud physics and cloud seeding. New York,
Doubleday, 1962. I44 p.
Z. Byers, Horace Robert. General rneteorology. New York,
McGraw-Hill, 1959. 540 p.
3. Chalrners, J. AIan. Atmospheric electricity. London, Pergarnon, 1957. 327 p.
4. Chaprnan, S. Thundercloud electrification in relation to rain
and snow particles. In: Horace Robert Byersr Thunderstorrn
electricity. Chicago, University of Chicago, 1953. p. ZO7230.
5. Duncan, Acheson J. Quality control and industrial statistics.
Flornewood, Illinois, Richard D. Irwin, 1955. 344p.
6. I'Ietcher, N. H. The physics of rainclouds. Carnbridge,
Carnbridge University, 1962. 385 p.
7. Gunn, Ross. Electronic apparatus for the deterrnination of the
physical properties of freely falling raindrops. Review of
Scientific Instrumentation ZOz29l-296. 1949.
8.
. The frde electrical charge on thunderstorrn rain and its relation to droplet size. Journal oL
Geophysical Research 54:57-63. 1949.
9. Gunn, Ross and Charles Devin, Jr. Raindrop charge and
electric field in active thunderstorrns. Journal of Meteorology lOz?79-284.
1953.
10. Hurnphrey, W. J. Physics of the air. New York, McGraw-Hill,
r94o.
675 p.
11. Kitagawa, N. and M. Brook. A cornparison of intracloud and
cloud-to-ground discharges. Journal of Geophysical Research 65:1189-1201. 1960.
40
LZ. Mason, B. J. Clouds, rain and rainrnaking. Carnbridge,
Carnbridge University, 1962. 145 p.
13.
don
. The physics of clouds. London, C1arenPress, 1957. 481 p.
Vonnegut and A T. Botka. Results of an
experiment to deterrnine initial precedence of organized
electrification and precipitation in thunderstorrns. In:
Leslie G Srnithts Recent advances in atrnospheric electricity. New York, Pergamon, 1958. p. 333-360.
14. Moore, C. 8..
15. Sarrnah, Suryya K. A radar and synoptic study of rain at a
point. Master's thesis. Corvallie, Oregon State University,
1963. numb. leaves.
16. Sutton, Robert IvV. Design criteria, construction and evaluation of a rnediurn power weather radar systern. Masterrs
thesis. Corvallis, Oregon State University, 1959. 52 nurnb.
leave s.
17. Workrnan, E. J and S. E. Reynolds. Structure and electrification. In: Horace Robert Byersr Thunderstorrn electricity. Chicago, University of Chicago, 1953. p. 139- 150.
Number and Magnitude of Charges
Time
Positive
PST
30 Mar 63
0047-49
Q x 103 esu
1 II 1 1 1 2 n o
3S-.2 •33.0 '30.8 12&.t 1
.2CV|iy.z]?2.olf8\nC[/^t/\'3
Negative
1
8 8
it \f-V j2-2
j
0
<R
Z.2
H
W
£•6
0.8
/
O
1
I
//•O /JZ
IS.f
nc)\ifs 220\21.z' 2LM 28i\*°-8 53^ 3**i
1
-
49-51
51-53
53-55
0
1
IH
3
0
£-
V
0
O
II
O
s
O
/
£•
V
0
/
0
/
55-57
0101-03
0
03-05
05-07
0
1
0
. .
/
loo
2
O
H
2
0
6
3
/
/
2/3
O
2
8
0
2
/
O
O
/
2
1
2
2
/
2
O
/ ir
2
O
O
/
Z
Z \3
3
lO
2
2
O
0
V
11-13
/
0
0
t .1o
.0040-42
42-44
44-46
L
46-48
APPENDIX I
O
/
/
loot
/
1 \l
2
/
i / ;0
1 •
1
0
3
/ \s
Z
3
/
O
/
O
O
Zj z
S
\9
3
2
/
O
O
0
l
Z
3
£
0
J
1
0
2
0
1
/
V
1
/
"
0
1
I
/
1
"
i
2
2
2 |/
3
8
t
2
/
0
0
Z
/o
y
i
2
0
6
J
2
O
0
n
5
0
/
I
0
/
0
j
i
!
0
1
O
07-09
0! 1
7
/
z\o\s
09-11
O
O
57-59
59-01
/
2
2
O
/
v ;2
1
Sample Data of the Electric Charges on Precipitation
2
0
i.