A S CAD E S T MO S P HERI C E SO URCE S RO G RAM Prepared for U.S. DEPARTMENT OF THE INTERIOR BUREAU OF RECLAMATION DIVISION OF ATMOSPHERIC WATER RESOURCES MANAGEMENT DENVER, COLORADO UNDER CONTRACT NO. 14-06-D-6999 STATE OF WASHINGTON DANIEL J. EVANS GOVERNOR DEPARTMENT OF ECOLOGY JOHN A. BIGGS DIRECTOR INTERIM REPORT JANUARY TO JUNE 30, 1972
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
ASCADESTMOSPHERIC
ESOURCESROGRAM
Prepared forU.S. DEPARTMENT OF THE INTERIOR
BUREAU OF RECLAMATIONDIVISION OF ATMOSPHERIC WATER RESOURCES MANAGEMENT
DENVER, COLORADOUNDER CONTRACT NO. 14-06-D-6999
STATE OF WASHINGTONDANIEL J. EVANS
GOVERNOR
DEPARTMENT OF ECOLOGYJOHN A. BIGGS
DIRECTOR
INTERIM REPORTJANUARY TO JUNE 30, 1972
STATE OF WASHINGTON
DUPAI^TMEOT @F IS(C)@1L@(C)YDANIEL J. EVANS JOHN A. BIGGS
GOVERNOR DIRECTOR335 GENERAL ADMINISTRATION BUILDING OLYMPIA 98501
August 10 19 72
Dr. Archie M. Kahan ChiefDivision of Atmospheric Water
Resources ManagementBureau of ReclamationBuilding 6 7 Denver Federal CenterDenver Colorado 802 25
Attention D-1200
Dear Dr. Kahan
According to the provisions of the U. S Bureau ofReclamation Contract No. 14-06-D-6999 dated 1 July 19 70Subparagraph 7b we are herewith forwarding five copies ofan Interim Progress Report This report covers researchefforts on the Cascades Atmospheric Water Resources Programduring the period 1 January to 30 June 1972
Sincerely
Office of Technical Services
<^.S-^- o^--^" ^’. -"^,^L^^ -yStuart E Shumway MeteorologistTechnical Assistance Division
SES ss
Enclosures ( 5 )
Section 1
Activity for the Report Period in Each of the Topics
Stated in the Work Plan
January 1 to June 30, 1972
Task #1: Aircraft Research Instrumentation
This task’s projected goals were largely met in the previous
report period. During the present period the storage monitor scope for the
5 cm weather radar was installed in the rear of the aircraft near the flight
scientist’s transparent observation dome. The experimental telemetry
system continues to be postponed due to budgetary limitations and licensing
difficulties.
Considerable progress has been made in other areas of aircraft
improvement which were not specifically listed under this task. For example:
(1) The Mee Industries ice nucleus counter which was intermittent
at best during the last report period was substantially
improved by a series of minor modifications. While considerable
doubt still remains concerning the reliability of the actual
numerical values of the counts recorded, the mechanical
reliability is much improved. Studies continue into the
reasons for the large variations in count which can be obtained
with preconditioning adjustments.
2
(2) To augment the Mee ice nucleus measurements millipore
filters can now be exposed from the aircraft during a flight.
The filters are developed in the laboratory using a modification
of the Bigg-Stevenson technique.
(3) The radar altimeter, which was previously somewhat intermittent,
has been greatly improved since we acquired complete test
and calibration sets. It now operates reliably up to 10,000
feet above terrain.
(4) The airborne wind speed and direction sensor is nearing the
airborne test stage. During this report period "bread-board"
versions have been flown and aircraft ground speed and drift
have been calculated. The further electronic calculation of
wind speed and direction is not expected to provide any great
difficulties although considerable testing is anticipated
before the unit’ is fully operational.
We expect this device to improve our targeting and
evaluation of seeding effects since the winds at flight level
will be monitored in real time. In addition, the ability to
obtain wind speed and direction over a wide area and in the
vertical anywhere the DME-VOR signals can be received will
add a valuable dimension to the aircraft’s capability.
(5) The aerosol conditioning chamber has been substantially
enlarged and improved and the sample inlet has been modified
to provide both deicing and isokinetic sampling. However,
the large heated venturi, which is also necessary for
isokinetic sampling, remains to be installed.
3
The B-23 aircraft continued to operate faultlessly throughout
the remainder of the 1971-72 winter field program.
Tasks #2 and #7: Operation and Maintenance of the Ground Stations
The network of manned and automatic ground meteorological stations
remained as described in the previous report. Operational reliability was
very good throughout the winter largely due to the careful preparation and
calibration of the instruments during the previous summer together with a
regular program of service in the field.
Task #4: Rawinsonde Measurements
The rawinsonde tracker (on loan from NCAR) and launch facility
at Greenwater, Washington, operated satisfactorily during this report
period. As before, the BuRec small computer terminal was an important link
in the rapid analysis of the rawinsonde raw data.
Tasks #3 and #8: Radar Development and Operations
Introduction
The mission of the radar portion of the Cascade Project is
(1) To obtain remote measurements of precipitation, its intensity and fall
velocity spectra and to relate these measurements to the physical structure of
the cloud systems and (2) To study "control conditions" and compare them
with the seeded situations in order to define adequately the effects of
seeding on hydrometeor size^and trajectories in a storm system. In pursuit of
these goals a low power pulse Doppler radar system with a single range gate
4
was interfaced -to an incoherent radar (a T-9) so that a common receiving
antenna and receiver hardware could be used.
System Operation
Pulse Doppler operation began early in January after the development
of solid state pulse control features. These were interfaced to the
incoherent radar synchronizer. Tests showed that the peak power of 20 watts
developed by the traveling wave tube was not sufficient to fire the transmit-
receiver switch in the T-9 therefore, operation of two antennas was necessary.
A 6-foot parabolic dish was installed on the ground approximately 3 meters from
the incoherent radar and was equipped with a heater for snow removal. This
arrangement proved satisfactory and was used the remainder of the winter
season. Both radar systems the high power incoherent and the low power pulse
Doppler, could be operated simultaneously when the pulse repetition rate was
harmonically related.
Data Processing
One of the most important developments during the period was the
evolution of the data processing system. Early observation of spectra from
CW and pulse Doppler returns showed the variability of the raw spectra of
the fall-velocities of the hydrometeors. The spectra had a phase flicker with
a time constant around 16 ms due to the shuffling of the scatterers in and out
of the beam. Slower variations due to meteorlogical changes were observed
over a period of minutes. The problem of reducing this type of data is a
critical one because internal variables must be found that are independent of
5
the randomness of the signal. In order to do this some specialized electronics
were developed that could be used with available pieces of equipment to put
the raw spectra on digital tape. Once the link with the large digital
CDC-6400 computer had been established the raw spectra could be averaged and
processed using standard Fortran programs. This provided a great advance
in both the reduction and interpretation of the data.
Video Processing
Early video processing with CW used a Burr Brown instrumentation
amplifier and fourth order active filter’s. With the advent of pulse Doppler
work in early January the system was found inadequate to accept the wide band
signals of pulse operation and a redesign proved necessary. Later, with the
use of higher power and a reduced pulse repetition rate, simple video amplification
and filtering was too insensitive and a further modification was installed. This
is a "box car" type of detector adapted from the AFCRL "Porcupine" radar
circuits which were kindly supplied by Dr. Ken Hardy. The use of the pulse
integration scheme enormously increases the detector sensitivity.
Installation of the High Power Klystron
With suitable modification of wave guide plumbing, cooling and
mounting hardware a V-82, 7 kW pulse klystron has been adapted to replace the
magnetron of the incoherent radar. Modification of the support electronics
resulted in a simplification of peripheral equipment and gives more reliable
operation.
Future Plans
The next major improvement in the radar operation will be the
installation of the digital data recording system in the radar van. This will
allow instantaneous observations of spectra and eliminate the re-recording
step. Furthermore data reduction will be greatly improved as the digital
tape can be brought from the radar site to the University computer center
and reduced within 24 hours or less of the data acquisition. This type of
operation would greatly improve the "feed back" between measurement and
interpretation.
Task #5: Theoretical Models
We are continuing to use two rather different computer models as a
guide to both field operations and the analysis of the measurements. The
more detailed model, which attempts to model both the airflow and microphysics
over the Cascades, has been described in a recent report to the Bureau (see
"A Theoretical Model for Orographic Precipitation and its Modification by
Artificial Seeding with Ice Nuclei" by A. B. Fraser, P. V. Hobbs and R. C.
Easter, April 1972). Since this report was written most of the developments
have centered on the microphysics portion of the model, although some efforts
are being directed toward the inclusion of the capability to handle a layer of
air with neutral stability.
The most recent improvement to the cloud microphysics part of the
model has been the addition of the aggregation process to the crystal growth
equations. Previously, only growth by riming and by diffusion from the vapor
phase had been considered in the model, but observational evidence shows
aggregation to be a common occurrence in winter storms over the Cascades. The
7
model now computes the percentages of crystal mass due to riming and
aggregation to provide information on the relative importance of these
two growth mechanisms and to classify the crystal for determination of its
fall speed. A fourth-order Runge-Kutta scheme has been employed to solve the
growth equations. The debugging of the new computer program has just been
completed so that new results will soon be available.
While our goal is to use our field measurements as input data in
our sophisticated airflow and microphysics model in order to obtain outputs of
the effects of seeding of precipitation in real-time, most of the operational
targeting and decision making during this report period made use of our simpler
operational model.
The operational model, in its present form, uses rawinsonde data and
surface observations of snow crystal type to construct backward trajectories
of rimed and unrimed snow crystals. The model is based upon experimental
observations which indicate that specific crystal types (e.g. dendrites
needles etc. ) grow only in particular temperature regimes. Thus if, for
example, a dendrite is sampled at the surface, the observer can be quite
confident that the crystal grew from the vapor phase at temperatures between
-12 and -17C. The operational model assumes that any additional increase in
the mass of the crystal below the -12C isotherm is due solely to riming.
Diffusive growth of the ice crystals is not considered in the computations.
The computer program simply identifies the level where diffusive growth is
assumed to stop and then calculates trajectories from the surface to that
height. Thus only a segment of the entire trajectory is actually calculated.
Special techniques for the reduction of freezing nucleus and
silver concentrations and replica information remained the same as reported
previously (Hobbs et al. 1971). Freezing nucleus concentrations were
determined by the drop freezing technique. Replicas of snow crystals were
examined with a stereomicroscope and the dominant crystal type(s) present
were recorded along with the degree of riming and an estimate of aggregation.
Special effects were also noted. The replicas which were collected on
2 x 2 inch glass slides were then projected and the crystals sized. Silver
concentrations are to be determined by the Bureau of Reclamation using an
atomic absorption spectrophotometer.
3. A review and preliminary evaluation of the effects of seeding on
March 3, 1972.
As an example of our methods for evaluating the effects of seeding
we describe here the results of a preliminary analysis of results obtained
on March 3, 1972.
Aircraft operations
Prior to any seeding extensive measurements were made from the
aircraft of the extent and structure of the clouds. The clouds were continuous
only over the Cascades. Over the Puget Sound Basin there was 8/10 broken
cumuliform cloud, and there were no clouds further east than 15 miles west of
Ellensburg. Over the Cascades the cloud tops were initially from 11,000 to
12,000 feet and they contained comparatively high liquid water contents at
8,000 ft west of the crest.
17
The theoretical model predicted that artificial ice nuclei should
be dispersed about 7 miles west of Hyak at an altitude of 8 ,000 ft. Therefore,
the aircraft flew at an altitude of 10 ,000 ft along a roughly square track
with sides 7 miles long centered on this point and 10 gm Agl ejection flares
with 2,000 ft delay and 3,000 ft dispersal distances, were ejected at each
corner of the square. Seeding began at 1251 and ended at 1440 PST and total
of 340 gm of Agl was ejected during this period.
From 1250 to 1330 PST there was a general decrease in cloud height
in the region of seeding but by 1400 PST the cloud height was increasing
again in this region. At 1420 PST the first ice particles of the day were
observed visually from the aircraft and detected with the optical ice crystal
counter in the downwind leg of the seeded track. Following seeding the
aircraft flew further downwind and over Kechellus Dam where the clouds appeared
to be glaciated since a subsun was observed.
Subsequent analysis of the continuous particle sampler showed that
prior to seeding the clouds west of the Cascade crest were composed almost
entirely of water droplets with only occasional rimed irregular ice particles
and frozen droplets. Samples taken in the seeded area from 1249-1405 PST were
also only droplets. However, after 1410 PST the particles collected were
predominately ice, irregular crystals and hexagonal plates dominated but
occasional stellars and broad-branched crystals were also found.
The data strongly suggests that the artificial seeding on this
occasion caused glaciation. Moreover, the plume-of-effect was roughly as
predicted by the theoretical model.
Radar operations:
The radar measurements at Hyak on March 3, 1972 began at 1020 PST
when the incoherent radar was turned on to survey the sky. At that time there
18
was no precipitation at the radar station. At 1052 PST the Doppler radar
was switched on and vertical probing began. At this time a few heavily rimed
dendrites were falling at Hyak. Very light precipitation continued until
1200 PST when the incoherent radar was again used for a sky survey.
Precipitation was observed at altitude to the northwest. By 1207 PST the
first weak signals were observed on the Doppler and the range gate was set
out to 500 m. Particles with fall-velocities from 1. 5 to 2.0 m sec were
observed while heavily rimed stellars and graupel fell at the station. By
1225 PST the average fall-velocity had decreased substantially and lightly
rimed stellars and dendrites predominated at ground level.
Subsequently, heavily rimed fast-falling crystals were observed at
the station until 1430 PST. At 1315 PST and 1330 PST the Doppler measurements
indicated that the average fall velocity fell below 1 m sec for short time
periods (see Fig. 2). However, these low fall velocities may have been due to
a data processing problem associated with small returned signals in very light
snow. However, at 1330 PST the slide replicas did indicate some decrease in
riming which would have caused lower fall velocities.
Starting at 1350 PST there was a steady increase in the average
reflected power from the precipitation (Fig. 2). This marked the beginning
of a rather unique event in the winter’s radar studies which is now described.
At 1415 PST the incoherent radar was turned on for another survey.
Snow generating areas could be seen and wide spread precipitation was observed.
At 1420 PST heavily rimed columns and dendrites 1 mm in size were collected
at the ground. The Doppler velocities appeared to be lower and the range gate
was probed for different velocities with height. However, there appeared to be
1300 1400 1500
Time PST)
^ - ’^X;^^^^^^- rw
19
little change with height at this time, indicating little growth over the
range surveyed (1300 m). By 1448 PST the spectral characteristics had
stabilized and appeared as a nearly single fall velocity of about 0. 75 in sec
(Fig. 3). At about 1518 PST the range gate on the pulse Doppler was moved
out to the region of snow generation which appeared to be between 1. 8 and
1.95 km above the radar. By about 1540 PST the fall. velocities had increased
again (Fig. 3) and an observation of the generation layer showed that it
had moved down to about 0. 5 km above the ground. Conditions remained fairly
stable until 1655 when the radar was switched off.
The marked shift to small fall velocities observed between about
1432 and 1529 PST (see Fig. 3) was accompanied by largely unrimed crystals
reaching the ground at Hyak (Fig. 4). Similar changes in riming characteristics
were observed during this period at Alpental and Keechelus Dam which are
within the target area (Fig. 4). It should be noted that the predicted
effect of our heavier seeding is to reduce the riming and therefore the fall
velocity of precipitation particles and that seeding was carried out from the
aircraft situated about 7 miles upwind of Hyak between 1251 and 1440 PST.
It can also be seen from Fig. 4 that very sharp increases in precipitation
occurred at Hyak Alpental and Keechelus Dam from about 1430 to 1600 PST.
According to our transport model the time period during which snowfall in our
target area should have been affected by the artificial seeding was from about
1300 to 1515 PST.
Freezing nuclei in precipitation
Snow samples collected on the ground at Hyak, Alpental and Keechelus
Dam on March 3, 1972 were analyzed for their freezing nuclei concentrations in
2
Foi Velocity (m/)
F a 3--L; ep; eati-i i.ci;. ^;-; ^: Dopp Sei- ,jJa’ .e iuci ty spect ra ofpreci pi tat ion elements from \H08 to 1608 PST on March 3 1972The radar was located at Hyak and ai rborne seeding wi th Agwas carried out. 7 mi les upwi nd (west) of Hyak between 125and 4^0 PST Note the shi ft to smal ler fal veloci t ies
between about 1432 and 1528 PST
"’-’"’’-"-
K i*-(R) .\(R) (R) >t X X(R)<
n(R) (C) x
(R) A A X X^(R)0 (C) (R)(5
^HYAK
^x ^ ^G>@ (j) (R) ^ X
FDA(R)^7 A A A
* * ?K D:+ n n X
K ^ ;^ A (j"^(C) C?(R)
ALPENTAL
N
A+ FDPO
A
* "^ X ^(R) X A (R) K X (C) (R) ^ (R) A ^(R)(R)N H N H333e
^^m-noo 1200 1300 1400
Pocific Stondord Time
1500 1600
Fig. 4 Preci pi tat ion rates and characteri st cs of preci pi tat ion observedat three ground stat ions n the target area on March 3 1972 (SeeT,-ih)p ^ fnr 1-p,, m cwmhnic
TABLE h
KEY TO SYMBOLS IN FIGURE k
Symbol
0(53^B Densely rimed
XA*7
<(R)Dn+(R)
Explanation
Unrimed
Light riming
Moderate riming
Dendri tes
Radiating assemblageof dendri tes
Stel lars
Side planes
Assemblages of sectors
Columns
Capped columns
Broad branched crystals
Undetermi ned
Symbol
G
GS
FD
MD
HD
N
A
Explanati on
Graupel
Graupel -l ike snow
Few crystals on sl idethat are composed oftwo paral lei planecrystals separated bya water drop at thei rcenter.
- Half crystals ons ide are composed oftwo paral lei planecrystals separated bya water drop at thei rcenter.
^A1 crystals ons ide are composed oftwo paral lei planecrystals separated bya water drop at thei rcenter.
Nothing on sl de
Aggregates of icecrystals
20
the manner described by Hobbs et al. (1970). The results are shown in Figs.
5 7. It can be seen that increases in the concentrations of freezing nuclei
were observed at all three ground stations during the predicted time of
arrival (1300 to 1515 PST) of the plume-of-effect due to the artificial
seeding. It seems likely that this increase was due to the Agl dispersed upwind.
However, final confirmation of this conclusion must await analysis of the
concentration of silver in the snow samples on March 3. (Silver analysis is
being carried out by the Bureau but the results have not been received at
the time of writing. )
Distribution of precipitation:
It can be seen from Fig. 4 that there were marked increases in
precipitation at Hyak, Alpental and Keechelus Dam between about 1400 and
1800 PST. The predicted period of the direct effects due to seeding on
precipitation at these ground stations was 1300 to 1515 PST. The observed
precipitation pulses at several stations in the Cascades and their relationship
to the predicted period of the plume-of-effect due to seeding are shown in
Fig. 8. It would appear from these results that the increases in the rate of
precipitation which started at about 1400 PST was fairly widely spread and
may not have been due entirely to the seeding. For example, Nagrom is situated
about 10 miles south of the predicted plume-of-effect due to seeding, but
it shows an increase in precipitation between 1300 and 1700 PST. We plan to
obtain precipitation records from stations in the North and South Cascades
which are well-removed from our target area in Snoqualmie Pass in order to
further elucidate any effects which may have been due to seeding on March 3.
Kwm_y
I
1100 1200 1300 1400 1500 1600
Poclfic Stondord Tim*
Concentrations of freezing nuclei per gram of snow collected atHydk on March 3, 1972. The time period during which artificialseeding upwind was predicted to affect the snowfall was 1300 to
1515 PST.
""TO
1200 1300 1400 1500
PocHIc Stondard Time
1600
-24*C
-21C
:: i8c
S’C
12*C
Fig. 6 Concentrations of freezing nuclei per gram of snow collected at
Alpenta) on March 3, 1972. The time period during which arti ficial
seeding upwind was predicted to affect the snowfal was 1300 to
15)5 PST,
Concen^.ons of .e^,n. ^^,^r ^ ^^^^e^S^al^ed?^^ ^ predicted to affec. the sno.fal
Fig. 8. Precipi tat ion rates at s ix s tations n the Cascades on March 3 1972. The pred ctedperiods of the piume-of-effect due to ai rborne seeding wi th s lver-iodide upwind areenclosed between the arrows
21
Conclusions
A final conclusion on effects due to seeding on March 3, 1972
must await retrieval of further data (e.g. silver content of snow samples
precipitation records from more stations in Washington State) and more
detailed analysis. However, the following features do indicate successful
artificial modification:
(a) Aircraft observations documented a substantial increase in
ice in the clouds over the target area following seeding.
(b) The fall velocities of precipitation particles in the target
area, as measured by Doppler radar, showed a substantial
decrease during the predicted period-of-effect of seeding.
Moreover, the fall velocities increased again at almost
the same time the plume-of-effect from seeding was predicted
to cease. These changes are exactly those predicted to
occur due to glaciation of the clouds by artificial seeding.
Such sharp changes in the fall velocities of precipitation
particles have not been observed previously by Doppler
radar in the Cascades.
(c) The ice crystals collected on the ground in the target area
were heavily rimed before and after the predicted period-of-
effect due to seeding, but during most of the period-of-effect
they were unrimed. This observation also shows that the clouds
were glaciated during the seeding period.
(d) Measurements of the concentrations of freezing nuclei in
snow collected in the target area show increased concentrations
during the predicted period-of-effect of seeding.
22
(e) Snowfall in the target area increased during the predicted
period-of-effect of seeding.
On the other hand, the following features indicate that some natural
changes might have been taking place during the period of seeding:
(a) Changes in cloud height were observed from the aircraft.
(b) The barograph trace at Alpental indicates that a weak post-
frontal impulse might have passed through our target area
at about the time of seeding, although this does not appear
on the synoptic charts.
(c) The increases in snowfall rate in the target area extended
beyond the predicted period-of-effect due to seeding.
(d) Increases in snowfall rates occurred at some stations
outside the predicted area-of-effect of seeding.
This case study illustrates some of the difficulties encountered
in attempting to ascribe effects to artificial seeding even when (perhaps
particularly when) very careful measurements and analysis are made.
4. Observations of seeding effects with the Continuous Particle Sampler
on January 31, 1972.
On a number of occasions the effects of artificial seeding on cloud
structure were traced with the aircraft. The seeded regions are most often
detected by the real-time measurements of large increases in crystal concentrations
with the University of Washington’s Optical Ice Crystal Counter (OIC) or
by visual observation of the glaciated region. However, in thick cloud optical
23
effects are often difficult to observe and the crystals may be too small to be
detected by -theOIC. In these cases the effects can often by traced by a post-
flight analysis of the Formvar replicas from the continuous particle sampler
(CPS).
In our experience the seeded areas can often be distinguished from
even naturally glaciated areas by the predominance of regular crystals. In
addition, in recently seeded areas the crystals are smaller than average and
are present in high concentrations.
Flight 101 on January 31, 1972 (CPS Run No. 20) provides a good
example of a penetration of a recently seeded area. Agl seeding with both
drop flares and flares fixed to the aircraft had been in progress for 51
minutes as the aircraft flew a triangular pattern. CPS Run No. 20 was made on
the downwind leg of the seeding pattern. From the wind speed and the length
of the seeded legs it is estimated that the volume had been seeded about
5 minutes prior to CPS Run No. 20.
The observations which comprised about 1000 meters of cloud, are
summarized schematically in Fig. 9. The ice crystals found were very thick
plates and short columns with occasional irregular crystals. The ice crystal
areas contrasted strongly with sections of unglaciated cloud where the crystal
count went essentially to zero and the droplet concentration rose sharply.
The ice crystals were almost entirely between 50 and 150 ym in size
and the droplets from 10 to 30 pm in diameter. Only occasional counts
were observed on the QIC due to the small size of the crystals.
This run is especially valuable for the determination of ice crystal
concentrations due to seeding. The CPS can only rarely be used to obtain
Gtocifd Oroptt Gtociofdora -<
Oraprtfa
400 500 600
Ortonce ( mtrs
Fig. 9 Fl ight No. 10i on January 31 1972Fl ight al ti tude 10,600 feet.
Run No. 20. 13^5 PST Ai r Temperature-21 .3C
24
accurate measurement of concentration. Generally, the CPS is fitted with one
of our decelerators in order to reduce crystal fragmentation but in this case
the collection efficiency is unknown. On Flight 101 a decelerator was not
used but the crystals were sufficiently small and robust that virtually all
of them were collected intact. Thus on this occasion we estimate that our
deduced concentrations of ice crystals are correct to within +/- 20%. The value
of 2500 ice crystals per liter (Fig. 9) should provide an interesting starting
point in estimating the nucleating activity of our Agl flares using standard
plume diffusion theory from a time source.
5. Measurements of fall-velocities of snow crystals
One of the major problems in modelling the trajectory behavior of
snowflakes is the lack of detailed data on their fall speeds. Our model
requires this information not only as a function of the type and size of the
frozen hydrometeors but also as a function of their degree of aggregation
and riming.
To help remove these uncertainties a field project was begun last
winter to measure the sizes masses and velocities of frozen hydrometeors. A
brief description of this work is given below.
Instrumentation:
The instrument used for measuring the fall speeds of the ice
crystals is shown in Fig. 10. It consists of a transmitter and a receiver.
The transmitting section uses lights and fiber optics to transmit two planes
of parallel light separated vertically by 4.1 cm to similar fiber optics
PtWtOBMittlpitWtub
F g ,0 instruct for n^asuri ng fal veloci ties of preci pi tat ion parti cles (Dimens ions in en,.
25
in the receiving section. Light from the receiving fiber optics is passed
into photomultiplier tubes. Any decrease in the transmitted light caused by
a snow particle falling through the two planes of light is detected by the
photomultiplier tubes and can be displayed on a storage oscilloscope. The
time difference between signals from the upper and lower fiber optics is
recorded and can be used to determine the fall velocity of the snow particle,
Operational procedures
The instrument is located at the bottom of a tower at Keechelus
Dam which is designed to minimize horizontal wind effects. When a snow
particle falls between the two sets of fiber optics it is caught on a piece
of plastic wrap suspended on a frame. This frame is then taken from the
instrument, positioned under a microscope where a microphotograph is taken
of the particle. During this time the section of plastic having the particle
on it is stretched over a cold plate to keep the particle from melting. After
the microphotograph is taken the particle is melted into a water drop which
is also photographed. The mass of the original snow particle is then determined
from the droplet diameter.
Results
During the past few months the fall velocities sizes and masses
have been measured of over 400 snow particles. These data are sufficient to
give a fair determination of size versus mass and size versus velocity for the
following types of particles lump cone and hexagonal grauple, grauple-
like snow of lump and hexagonal type, radiating assemblages of dendrites
densely rimed columns, aggregates of sideplanes radiating assemblages of
dendrites sideplanes columns and bullets and dendrites. For the following
example, lump graupel is chosen to illustrate data handling methods and results.
26
Fig. 11 shows a log-log plot of mass versus diameter for lump
grauple. The data points are divided into six groups of similar density
types numbered 1 through 6 in Fig. 11, with group 1 having the lowest
average density and group 6 the highest. The empirical equation relating
mass M (in 10 grams ) with diameter D (in mm) for group 3 is
M 6. 5D3
Fig. 12 shows a plot of fall velocity versus diameter for lump
graupel. This data is also divided by density with some groupings combined.
First a line was established for the group containing the largest number
of data points (group 3). Then on the assumption that the velocity of grauple
in the other density groups would be a similar function of the diameter,
lines for the other groups were drawn in parallel to the line for group three.
The table in Fig. 12 lists the different groups the average density of the
particles in the groups and the equations of velocity versus diameter.
Fig. 13 is a plot on linear paper of fall velocity versus diameter
for lump graupel in the different density groups. For comparison the dashed
line shows a curve of velocity versus diameter for lump graupel of density
0.12 gm cm drawn from results obtained by Nakaya.
When the data collected so far is combined with similar measurements
which we expect to obtain next winter, we should be able to provide detailed
information of the kind given above for lump graupel for a wide variety of
crystal types and also to show the effect of riming and aggregation on the
fall velocity.
LiJ------- *--’ * ’ "J-------I---I--I--I I lii.i------I---I--I- > t *.tJK) 100 1000
M ( to uftif < io’5 m)
Fig. Di ameter versus mass for lump graupel <,
Oiawter 0 mm)
Fi q. 12 Fd vc ioci t-’y ,’t-i’su-a d dme ie. or Limp qraupe
3.0
uQ
"5
2.5
s.6 2.0>
. 1.5
1.0
0.5
0 2 3
04mtr Oirnrn)
5
13 Fal veloci ty versus cii ariiei.er or lump graupe p s the average dens ty (in gm cm
of the crystals n eaci-! group
27
Task #13: Instrument Research and Development
1. The University of Washington Optical Ice Crystal Counter
A description of the University of Washington’s Optical Ice Crystal
Counter (QIC) in its present form is given in Appendix D, which is an
extended abstract of a paper accepted for oral presentation to the International
Cloud Physics Conference, London, August 21-26, 1972.
While this device is already providing us with very valuable
real-time data on ice crystals in clouds the minimum size of crystal which it
can detect is still not known precisely. During late March 1972, Mr. F. Turner
of our research group visited Elk Mountain Observatory in Wyoming with the
intention of obtaining ground-based measurements of ice crystals in cap clouds
with the QIC and correlating these with slide replicas of the crystals.
Unfortunately, cap clouds were absent during the period of his stay so no
measurements were obtained. Efforts to simulate cloud conditions in our cold
laboratory have generally produced either very small (< 100 pm) crystals which
are generally not detected by the QIC, or rather larger aggregates which are
counted. Presently we are making aircraft measurements in natural clouds and
correlating the observations with the crystals collected by the continuous
particle sampler. The problem is complicated by the fact that we expect the
minimum detectible size to be somewhat dependent on crystal habit and degree
of riming. At the present time we can only say that it counts practically no
crystals smaller than 100 urn and practically all crystals larger than about
700 pm. Water droplets, even as large as several mm in size, are not counted.
A good example of the useful data provided by the University of
Washington’s Optical Ice Crystal Counter and the Mee fast response Ice
28
Nucleus Counter (INC) in delineating the effects of artificial seeding
occurred on February 23, 1972.
During this flight (Flight No. 105) a short isolated seeding
experiment was performed near cloud tops. The absence of ice crystals in the
clouds prior to seeding was verified by the QIC, post-flight analysis of
formvar replicas and by the presence of strong water droplet optics in the
form. of a glory and cloud bow. 180 grams of Agl were injected into the
cloud tops between 1237 and 1255 PST from pryrotechnic flares fixed to the
aircraft. Ice crystals were first observed in the seeded volume at 1250 PST,
and after 1252 PST a sub-sun was almost continuously visible. Fig. 14 shows
a portion of the flight record from 1313 to 1321 PST when the aircraft was
flying in and out of the seeded volume near cloud tops. The correlation
between the measured bursts of ice crystals and ice nuclei as detected by
the QIC and the INC is gratifying. Post-flight analysis of the formvar
replicas made at 1318 and 1321 PST showed primarily hexagonal plates and
plate-like fragments.
2. A Ground-based Snowflake Counter
A prototype snowflake counter was constructed and tested during the
winter of 1971-72. This was intended to be a simple approach to the problem
of measuring the number of snowflakes per unit volume of air. The sensor
(Fig. 15) consists of a horizontal sliding trap door located 20 cm above a
planar light source and photo transistor array. An electric motor pulls the door
open allowing saowflakes to fall into the volume below. The door is held open
for 2 3 seconds after which a spring pulls it rapidly closed. The light
Illnl ^(b) Ice nucleus counts
(o) Ice crystal concentration
1321 1319 1317 1315
Time
Fig. 1A Simultaneous ice nucleus counts Mee Counter) and ice crystal concentrations
(U. of W. Optical Counter) in a cloud seeded with Agl over Cascade
Mountains on February 23. 1972.
To Electronics
DriveMotor
--i ser
^Open closesense switch
Sliding’Trop door
"9WOReturnspring
Worm
20 cm
Reguloted ^ \Supply / \
Light
To Electronics
Phototronsistor
Snowflokesdetected inthis region(area 50 cm2)
Fig 15 Schemati c representation of prototype snowflake counter
29
source and photo sensor are mounted on the underside of an aluminum plate on
2opposite sides of a rectangular cutout (Area 50 cm through which the snow
is allowed to fall. Immediately the door closes the electronics are activated,
thereby counting the number of flakes within the enclosed volume (1 liter).
This procedure is repeated every 37 seconds. The electronics (Fig. 16) used in
this design consist of the photo transistor feeding a high-gain amplifier which
triggers a one-shot multivibrator which provides a constant pulse-width signal
for each particle sensed. A long-term integrator is allowed to sum all the
pulses encountered during the time that the door is closed, but is reset to
zero each time the door is opened. After a suitable time to count all the
particles within the volume has elapsed, the total number of snowflakes counted
is recorded, the trap-door opens again, and the cycle is repeated.
Figure 17 shows an example of the data obtained with the snowflake
counter on March 3, 1971, averaged over 15 minute intervals. Also shown in this
figure are measurements obtained with the heated tipping bucket and the E.
Bollay Assoc. optical snow rate sensor. It can be seen that on this occasion
the results are in reasonable agreement as far as fluctuations with time are
concerned. However, it should be noted that in general it is not to be
expected that the precipitation rate will vary as the concentration of snow
crystals in the air since the former quantity also depends on the mass and
fallspeeds of the crystals.
Despite reasonably encouraging results with this device in its present
form it has proved to be mechanically too complex and despite weatherproofing
and heating not sufficiently robust for severe winter environments. Several
alternatives for obtaining the concentrations of snowflakes at ground level
are being considered.
3. Keilly Probe
We have built, with some changes, the NCAR-modification of the Keilly
Photo-transistor
CD-) )-
Open-closeswitch
1
High ^\.
7^----U----------------------------------n
gain .>----amp. ^"
OneShot
Long-termIntegrator
reset
Sample andHold
&.4 secDelay
’Gote--1---o To
J Recorder
Fi g. 16 Schemati c of e lectroni cs for snowflake counter
Concentration of snowflakes from Universityof Washington counter
Precipitation rate from tipping bucket
Snow rate from E. Bollay Assoc. opticalsnowrote sensor
60
40
30
ou(A
0
8?o
o00)
0A
(A
.C0
0
0
1430 1445 1500 1515 1530
Pacific Standard Time
1545 1600
Compari son of concentration of snowflakes determined by Uni versi ty of Washington ’s counter
1 40 INTEGER DO1 50 REAL K1 , IN1 60 LMAX-01 70 I RUN-01 80 IRIME-01 90 HAXVO200 JSAVE-02 10 IFIN-0220 I DATA-0330 KI-2.8704E+6/980.665840 PRINT*"THIS PROGRAM COMPUTES THE TRAJECTORIES OF RIMED AND"250 PRINT*"UNRIMED SMOV CRYSTALS OF VARIOUS TYPES."260 1 16 PRlNT*"ENTiai THE STATION HEIGHT IN METERS"870 READ* ZSFC280 1 FORMAT( F6. 1 )
290 GO TO 4000300 PRINT*"ENTER THE OBSERVED CRYSTAL TYPE ACCORDING TO THE"310 PRINT*"FQLLQIMNfl LEGEND,"}20 PRINT*"IF OENDRITES PUNCH 1"i30 PRINT* "IF STELLARS PUNCH 2"140 PRINT* "IF PLATES PUNCH 3"150 PRINT* "IF SCROLLS PUNCH 4"160 PRINT* "IFSHEATHS/NEEDLES PUNCH 5"i70 PRINT* "IF BULLETS PUNCH 6"180 PRINT* "IF COLUMNS PUNCH 7"190 PRINT* "IF SIDE PLANES PUNCH 8"00 PRINT* "IF PRISMS PUNCH 9"l0 PRINT*"IF ASSEMBLEflES OF"20 PRINT*" PLATES430 PRINT*"IF ASSEMBLEGES OF"440 PRINT*" SECTORS450 PRINT*" IF SECTORS460 PR I NT*"IF GRAUPEL465 PRINT*"IF AQOREQATES_____470 4000 PRINT*"CRYSTAL TYPE?"480 READ* I TYPE
PUNCH 10"
PUNCH 1 1"PUNCH 12"PUNCH 13"PUNCH 14"
490 CALL SNOCAR(VTERM*TMAX* ITYPE)500 IFIN-05 10 IF( IDATA ,OT. 0) GO TO 134520 PRINT* "IS THE SOUNDING DATA FROM UIL?"530 READ 6* IDAD540 6 FORMAT(A3>
550 PRINT* "ENTER THE NAME OF THE FILE WHERE SOUNDING DATA IS560A STORED^570 READ 2* ZAF< 1 )
580 IF( IDAD .Efl. "YES" ) GO TO 132590 PRINT*"ENTER THE LOWEST PRESSURE IN THE SOUNDING"600 READ* PLOW610 PLOV-PLOV+80,’on n n taPi on
04U tje *-*650 CALL OPEWFO^ IH)
660 DO 3 J-l*76 39
670 LMAX"J680 IJ690 1F( 1DAD *EQ. "NO" ) GO TO 1 35
770 5 FORMAT<2F7.0-F6.1 l8X,F5. l <9X,F4.0, lX,F4.0)
780 JLOV-Pd )
790 IFC JLOW- *LE. ILO^) GO TO 4
800 133 TD< I )"T< I ) -TD( I )
810 T< I )"T< D+273. 16820 S( I )S< I )*514830 Z( I )ZZ( I )
840 ZZ( I )"ZZ< I )*100.850 E-6. l l*( lO<0**< 7.5*( TD( I ) )/(TD( I )+237.3) ) )
860 < I )0.622*E/<P( I )-E)870 IF< IFIM OT 0) CALL CLOSEF< 1 )
880 IF< IFIN .GT. 0) GO TO 134890 3 CONTINUE900 4 CALL CLOSEFC 1 )
?20 ^PRTSr^ENTER THE OBSERVED SURFACE ^IND AZIMUTH IN DEGREES-
940 PRINT"ENTER THE OBSERVED SURFACE WIND SPEED IN KNOTS-
960 PR^NT^I^A COMPLETE TRAJECTORY WITH 300M STEPS IS DESI^D"970 PRIMT^PUNCH A OWE. IF A REDUCED PRINTOUT IS DESIRED, PUNCH-
980 PHIMT<**A ZERO***990 READ* I FLAG1 000 Z( l )"ZSFC1 010 S( 1 )"S( 1 )*.5l41020 1 18 DO 100 N-l^LMAX1030 IN1 040 IF( T<M) .LT. TMAX) GO TO 101
1050 100 CONTINUE1 060C TMAX IS BETWEEN T( I ) AND Td-D1070 101 TLAPSE"( T< I )-T( I-1 ) )/(ALOG<P< I )/P< I -1 ) ) >1080 XLNP-CTMAX-T< l-l ) )/TLAPSE+ALOG<P< I-D )
1090 PP-EXP<XLMP)100C COHPUTE HEIGHT AT TMAX (M) HEIGHTZSTOP1 10 TV( I )-T< I >-*-( I )<"166.667120 TV( I-l )T( I-l >*<< I-l )*166.667
Ito PRiMT^TSE’HT. (FT) OF THE BASE OF THE CRYSTAL GROWTH LAYER-
170 PR1NT ZSl80 IFdDATA GT. 0) 60 TO 109190 DO 102 K>140200 L-K1210 IF<P(K) .GT* 841 . .AMD. P<K) .LT. 859. ) GO TO 103
220 IF<K EQ. 40) PRIMT^ERROR IN INPUT PRESSURE-1230 IF(K EQ. 40) 00 TO 10001240 102 CONTINUE1250C L IS INDEX OF 850 MB WIND,260C COMPUTE WIND COMPONENTS IN 150 METER INTERVALS,270 103 DZ-1 50.280 HZ-ZSPC-DZ/2.
,i:)ft<W tm Httf^HVU13M U-K336 KWUC-K 40
340 HZ-HZ+DZ350 TOP-2( I )
360 H2-HZ370 IF(HZ .OT. TOP) HZ-HZ-DZ380 IF(H8 .GT. TOP) GO TO 104390 U<K)-0.0400 V<K)-0.0
460 IFCADDI Eft* 180. ) GO TO 105470 IFCDDI LT. 180* ) GO TO 810480 DDI-DDI-360.490 GO TO 819500 818 IF(DDI+180) 817,819,819510 817 DDI-DDI+360*580 819 DD"D(LOV)*(DD1*(HZ-Z(LOV) ) )/( Z< I ) -Z(LCV) >530 ZF(OD .GT. 0) GO TO 822540 DD-DD+360.550 GO TO 823560 822 IFCDD .LT. 360) 80 TO 823570 DD-DD-360580 823 SS"S(LOlir)^<SSl*(HZ-Z(LOW) ) )/( Z( I )-Z(LOV) )
590 DX-DD600 IFCDX LT. 90 AND* DX .GE. 0. ) GO TO 106610 IFCDX LT* 180 AND* DX GE* 90* > GO TO 107620 IFCDX .LT* 270* AND* DX GE* 180 ) GO TO 108630 THETA-(DX-870. )/57.284640 U(K) "SS*COS< THETA)650 V(K)-SS<dSiN( THETA)660 GO TO 105670 106 THETA-(90.-DX)/57.284680 U<t()-SS*COS< THETA)690 V<K>-SS*SIM(TKETA)700 80 TO 105710 107 THETA-(DX-90 )/57.284720 U(K)--SS*COS( THETA)730 V(K)SS*SIN(THETA)740 GO TO 105750 108 THETA-(870.-DX)/57.a84760 IKK)"SS*COS(THETA)770 V(K)-SS*8IN(THETA)780 105 CONTINUE790 104 CONTINUE800C COMPUTE FALL VELOCITIES FOR LARGE AND SHALL CRYSTALS810 109 DO 1 10 I-l-S820 IF( I EQ 2 AND ITYPE EQ* 13) GO TO 1 10830 HZ-ZSFC-DZ/8.840 DO 1 12 K-1,KMAX850 HZ-HZ+DZ860 MAXVI870 IF(HZ GT. ZSTOP) GO TO 1 19880 IFC IRIHE EQ* 1 ) QO TO 1 1 1890C HERE IS THE UNRIMED FALL VELOCITY900 VT-VTERM( I )
910 GO TO 1 12920 1 1 1 VLAPSE-VTERM( p/(ZSTOP-ZSFC)930 VT"VLAPSE*< ZSTOP-HZ ) t-VTERM< I )
8890 84 FORMATC IH SMILES- AND*** IX<F7.2, IX,"DEGREES FROM THE2300A SURFACE STATION")8310 81 FORMATC 1H *4X/<"IF THERE IS RIMING* THE TRAJECTORY FOR" )
8380 88 FORMATdH ^-THE LARGEST ICE CRYSTALS ORIGINATE AT^ IX,8330AF7.2)8340 IFCMAXV S6 l > PRINT63, R( JSAVE)8350 IFCMAXV .EQ* 8) PRINT 82, R< JSAVE)8360 PRINT 84 AZ( JSAVE)8370 IFC IFLAG BQ l > PRIMT 858380 IFC IFLAG E 0) PRINT 888390 88 FORNATdHO^-FOR THE ABOVE CONDITIONS* THE TRAJECTORY’S /8400&-POSITIONS ABE AS FOLLOWS--//)8410 85 FORMATC IHO*-FOR THE ABOVE CONDITIONS* THE TRAJECTORY*’^/*8480**’POSITIONS AT 300 M INTERVALS ARE AS FOLLOWS*"*//)8430 PRINT 878440 87 FORMATdH *2X,"AZI FROM"*4X*"NA MI"*4X*"MSL HT"*4X*8450A-FALL TIME"*8X*"A/C MAG<’*8X*MA/C RANGE** */*3X*"TAR6Er’*8460A4X*"TO TARGET <FEET) (MIN) HEADING"*2X*8470A-FROM SEACN Ml )-*/)8480 DO 810 J-1 *JSAVE8490 CALL AKGLC(X*YJ*AZ*R)8500 TOTAL( J)-TOTAL< J)/60.8510 Z( J)-Z<J)*3.288580 810 CONTINUE8530 CALL HEAO(AZR*JSAVE*HED*AR)8540 iFdFLAG .EQ* 0) GO TO 1808550 213 DO 181 J-l^JSAVE-28560 181 PRINT 886AZ< J)R( J)Z( J)TOTALC J)*HED( J)AR( J)
8570 886 FORMATdH ,6F10.1 )
8580 GO TO 1318590 180 K-28600 IS-58610 DO 183 11-1,36- IS
1 3090C AMY MORE TRAJECTORIES TO BE CALCULATED?3100 PRINTS IF MORE TRAJECTORIES ARE TO BE CALUCUALTED* PUNCH A"
31 10 PRINT-THO* IF NO MORE- PUNCH A ONE."3120 READ* IRUN3 130 IFC IRUN EQ. 1 ) GO TO 10003140 PRINT*" IF THE SAME iTIND DATA IS TO BE USED< PUNCH A ONE-"3190 PRINT*"IF NEW IMD DATA IS TO BE INPUT PUNCH A ZERO."
1 3160 READ* I DATA3170 IF< IRUM .OT. l > GO TO 1 163180 1000 END3 190 SUBROUTINE SNOCAR<VTERM*TMAX, I TYPE)
| 3800 DIMENSION VTERMC2)3210 109 IFdTYPE .Sa. D GO TO 1003220 IFdTYPE EQ. 2> 00 TO 101
1 3230 IFdTYPE Efl. 3) 80 TO 1023240 IFdTYPE EO. 4) 00 TO 1033250 IFdTYPE Ett 5) QO TO 104
1 3260 IFdTYPE EQ 6) GO TO 1053270 IPdTYPB Btt 7) 00 TO 1063280 IFdTYPE .EQ. 8) 00 TO 107
’ K > tir/ ITVP’f .’ ^ Ql FQ TO
3318 (y< ITYPe BB. 14> BO TO 1 15 433330C SECTORS3340 TMAX-261 .3350 VTERM( 1 >-.243360 VTERM(2).A93370 00 TO 1083380C DENDRITES3390 100 TMAX-2613400 VTERM( D- .243410 VTERM(2)-.493480 GO TO 1083430C STELLARS3440 101 TMAX-2613450 VTERMC D-.243460 VTERM(2 )-1 .03470 GO TO 1083480C PLATES3490 102 PRINT*"PUNCH .MAX TEMPERATURE FOR CRYSTAL HABIABASED3500 PRINT<"ON SOUNDING AND AIRCRAFT DATA"3510 READjTMAX
3520 TMAX-273.+TMAX3530 VTERM( 1 )-.243540 VTERM(8>.83550 00 TO 1083560C SCROLLS3570 103 TMAX-265.3580 ITYPE-03590 GO TO 1083600C SHEATHS/NEEDLES3610 104 TMAX-269.3680 VTERM( l )-.243630 VTERM<2)1 .03640 GO TO 1083650C BULLETS3660 105 TMAX-248.3670 VTERM( 1 )-.53680 WERM(2>"1 .03690 GO TO 1083700C COLUMNS3710 106 TMAX-S68.3720 VTERMC D-.53730 VTERM(2)"1 .53740 GO TO 1083750C SIDE PLANES3760 107 TMAX2533770 VTERM( D-.243780 VTERM(2)493790 GO TO 1083800C PRISMS3810 1 1 1 TMAX-248.3820 ITYPE-03830 GO TO 1083840C ASSEMBLEQES OF PLATES3850 1 12 TMAX-255.3860 VTERM( 1 >>.243870 VTERM(2)-.49
3880 00 TO 1083890C ASSEMBLEGES OP SECTORS3900 1 13 TMAX-255.3910 VTERM( 1 ).243920 VTERM(2)493930 GO TO 1083940C ORAUPEL3950 1 14 C2"-.01930
,; .-3990 Bl.363 44
4000’ 0-9806554010 XNEV-.1424020 RHO-1 15E-34030 XK2 .*G/(RHO*XNEW’-XNEW)4040 S-1 .274050 PRINT,"WAT IS 8RAUPEL DIAMETER IN CM?"
4060 REA’D^ 04070 IF( D .GT. .3 .OR. D .LT. .02) ITYPE0
VTERM<2)2GO TO 108108 IF< ITYPE .8T. 0) GO TO 1 10
PRINTS-NO DATA AVAILABLE- TRY ANOTHER CRYSTAL
READ- I TYPEGO TO 1091 10 RETURNENDSUBROUTINE ANGLE<X,Y-K,AZ,RANGE)DIMENSION X(200)-Y(200>- AZ(200)-RANGE(200)TANfl"ABS<Y<K)/X(KTHETA-ATAN(TANG)*57.284IF<Y(K) .LT. 0. .AND. X(K) .LT>. 0. ) GO TO 100
AND. X(K) .GT. 0. ) GO TO 101IF<Y(K) .LT. 0.IF(Y(K) GT. O.PHI-870.+7HETA00 TO 103100 PHI-270.-THETAGO TO 103101 PHI-90.+THETAGO TO 103102 PHI-90.-THETA103 AZ(K)PHITHETA-THETA/57.284RANGE( K)-Y< 10/SIBK THETA)/1852RANGE< K) <*ABS< RANGE( K) )