|
Items in cart [0] |
Questions? Call 888-624-8373 |
PAPERBACK + PDF PAPERBACK PDF BOOK PDF CHAPTERS Free Resources |
||||||||||||||||
|
||||||||||||||||
|
|
|||||||||||||||
| Copyright © 2009. National Academy of Sciences. All rights reserved. Terms of Use and Privacy Statement |
Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 45
4
T~ r ~ .
OOlS and ec 1nlques for
Advancing Our Understanding
The past few decades have seen the development of a multitude of new tools for
measuring and modeling physical processes of cloud and storm systems. It is becoming
feasible to carry out detailed studies of the chain of physical events in the evolution of a
cloud system. This will lead to restore definitive assessments of the effects of seeding,
refinements of physical hypotheses, and "prospecting" information about suitable seeding
targets. Allis chapter identifies important developments in observational technologies and
modeling and data assimilation capabilities and discusses how these new tools and
techniques can best be applied to studies of enhancing atmospheric water resources and
mitigating hazardous weather.
MEASUREMENT AND OBSERVING TECHNOLOGIES
Several large weather modification research programs were carried out in the late
1960s and early 1970s, including the National Hail Research Experiment aimed at hail
suppression, the Sierra Cooperative Pilot Project aimed at snowpack enhancement, and
the High Plains Experiment aimed at warm-season rainfall enhancement (among others
discussed in Chapter 2 and Appendix A). These experiments contributed to the
development of many new observational instruments and facilities such as the Wyoming
King Air reseat ch aircraft, the NCAR CP-2 dual-wavelength radar, the CHILL dual-
wavelength and Doppler radar systems, NCAR and NOAA Doppler radars, and the
NCAR Portable Automated Mesonetwork. These systems defined the state of the art at
the time and contributed much to our current understanding of precipitation processes.
Although weather modifications research has declined since that time, observing
technologies with which the field could benefit have continued to advance. Cloud-
seeding research activities Carl now employ revealing measurements that were
unavailable in earlier decades, particularly in terms of remote ser~sing. The new
observations offer more accurate and higher resolution precipitation measurements and
three-dimensional depictions of the structure, airflow, and hydrometeor composition of
clouds before and after seeding.
45
OCR for page 46
~6
CRITiCA L I~SSIJES IN AREA THEA AlODIFI(-'/ TION RESE-ARCIJ
Several remote-sensing advances of great potential value to cloud seeding were
fostered by urgent needs in other fields? including requirements for improved severe
storm warnings, detection of aircraft icing conditions, and better understanding of the role
of clouds in climate change. Some of these new observing, technologies leave had cursory
initial demonstration uses in actual weather modification experiments but none have as
yet been used as integral components of experiments designed to test and evaluate
specific scientific hypotheses. Thanks to continuing development in other fields these
technologies are reaching a level of maturity float makes their wider use in cloud-seeding
research and operations feasible and attractive. The following observational tools ar
likely to provide contributions to future weather modification steadies.
Doppler Radars
At the time of the major weather modification field studies mentioned earlier ~ the
use of Doppler radar was embryonic, the performance characteristics of Doppler radars
were still topics of research' and multiple Doppler networks were just emerging In the
subsequent decades attendant research led to operational deployment of Doppler radars
for precipitation measurement, severe weather detection and warning (the Next
Generation Radar, or NEXRAD, network), and for detection and warning of hazardous
wind shear at airports. Serafin and Wilson (2000) describe the status of these operational
systems. These radars produce data that are of research quality and the data are becoming
available in real time (for instance, through the Collaborative Radar Acquisition Field
Test CRAFTY.
Another major airborne instrument development has been the advent of airborne
Doppler radars flown on NCAR and NOAA research aircraft as well as on the NASA
ER-2. These radars have produced information of unprecedented accuracy and resolution
in precipitating systems' leading to improved understanding of the structure of and air
motion fields in hurricanes (Heymsfield et al, 2001), severe storms, and even in optically
clear air (Wakimoto and Liu, 19983. New understanding of the genesis and evolution of
tornadoes and the intensity of hurricanes has been gained from these observations. Highly
mobile ground-based radars have also demonstrated their utility for high-resolution
measurements in the challenging conditions prevalent in severe storm environments
(Wurman and Gill, 20009.
Atmospheric Profiling
Much progress has been made in the arena of atmospheric profiling? and sensitive
wind profilers now are available commercially. These devices measure profiles of
tropospheric winds continuously and when coupled with acoustic sounders' also measure
profiles of temperature (May et al., 1990~. Ground-based GPS receivers can routinely
measure path-integrated water vapor. Progress has also been made in optical sensing of
the atmosphere. Differential absorption and Ra~nan-scattering lidar are capable of
measuring water vapor profiles (Ismail and Browell, 1994; Melfi and Whiteman, 19854.
Solid-state and reliable Doppler lidars have been used very effectively for measurements
of winds and turbulence (Poon and Wagoner' 1995 J. Scientists have recognized the
OCR for page 47
TOOLS A .VD TECI-I .VIQZ;rES FOR A D VA NCIN(r OUR Z!,?~7DERSTA .VDI.INTG
47
importance of better water vapor measurement techniques and completed the most
comprehensive research project ever attempted to better characterize the three-
dimensional structure of water vapor (described at
~. Research interests in
profiling the atmosphere have become so active that a special issue of the Journal of
Atmosuhe'^ic and Oceanic Technology has been devoted to the topic (JAOT, 20024.
~ . .
Microwave Radiometry
In glaciogenic seeding the objective is to use a seeding agent (nuclei or dry iced
to convert tipsy supercooled water droplets to ice crystals, plaice grow rapidly and
precipitate out of the cloud. Thus, locating regions of high concentrations of supercooled
liquid in natural clouds is of paran~ount importance. A promising tool for this
"prospecting" work is the dual-channel microwave radiometer, which retrieves the
patl~-integrated total amount of liquid water and water vapor along its beam by
simultaneously measuring emissions from vapor and liquid at frequencies near 21 GHz or
23 GHz and 31 GHz (Westwater, 19939. Ground-based, unattended vertically pointing
microwave radiometers have been used for monitoring aircraft icing conditions aloft and
in atmospheric radiation climate research programs. These units' based on technology
developed in the l980s, are now commercially available, as are newer ones that monitor
additional frequencies to provide coarse vertical profiles of cloud liquid water content
and temperature. The ability of a scanning microwave radiometer to observe cloud-
seeding opportunities was demonstrated by the NOAA/ETL in the Sierra Cooperative
Pilot Project orographic snowpack enhancement experiment (Snider and Rottner7 l 9824.
Aircraft-~ounted microwave radiometers are also now available and Inlay be suitable for
cloud-seeding activities.
Polarimetric Radar
Polarization-diversity (dual-polarization) radars measure signals backscattered
from targets in two orthogonal orientations to discriminate between water and ice in
clouds, detect hail, identify the types of particles present (see Plate 6), and attain more
accurate estimates of rainfall rates using differential phase (KDp) methods (Bring) and
Chandrasekar, 2001~. These capabilities are of great potential value in assessing cloud-
seeding experiments For individual cloud studies, polarimetric particle classifications
have the potential to reveal the transformation of supercooled liquid water droplets to ice
crystals in glaciogenic seeding arid the development of large drops in hydroscopic
seeding They can also follow the movement and dispersion of seeding aerosols using
microwave chaff tubers as tracers (as discussed later). I hree-dimensional depictions ot
these processes may be observed as they occur using ground-based or airborne
r~olarimetric radars. The particle classifications also can r efine conventional
,~ .. . . . . `% .. .. . . . ~ . ~ . . ~ . .. . .
retlectl~lty-based ralntall estimates by 1dentllylng regions ot echo that are not rain or
contain rain with contaminations of hail, snow, ground clutter, or insects. The new
differential phase estimations of rainfall rate offers a method for measuring the
ground-level result of seeding that is free from several factors that have historically
degraded the simple reflectivity-based estimates of precipitation. The method avoids or
OCR for page 48
~8
CRITICAL I~SSlJES IN TLEATI-IER AlODIFICATIONRESEARCH
minimizes problems related to hardware calibration errors, attenuation, partial beam
filling' partial beam blockage, the presence of hail, and variability of drop size
distributions (Zrnic and Ryzhkov, 19964.
Polarization-diverse radars are available only in the research community, but
their numbers are expanding. Most dual-polarization research in the Ur~ited States has
been conducted with the large S-band (3 GHz) weather surveillance radars, such as those
at NCAR, NOAA's National Severe Storms Laboratory, and Colorado State University.
NOAA's Environmental Technology Laboratory uses polarimetric methods with much
smaller millimeter-wave radars (35 GHz) for cloud hydrometeor identifications and at X
band (9 GHz) for chaff tracer tracking and differential-phase rainfall estimations. Even
smaller' highly mobile polarization-diversity ~nillimeter-wave radars are operated on
trucks by the University of Massachusetts and on research aircraft by the University of
Wyoming. The technology now exists to inexpensively upgrade radars to multiparan~eter
capability; and the national network of operational S-band weather surveillance radars
(WSR-88D or NEXRAD) may be upgraded to include polarimetric capabilities by the
end of this decade, depending in part on results of the Joint Polarization Experiment
demonstration in Oklahoma in 2002-2003 (NRC, 20029.
Millimeter-Wave Cloud Radar
Millimeter-wave cloud radars use wavelengths of 8 mm or 3 mm that are more
than an order of magnitude shorter than those of S-band weather surveillance radars.
Lhermitte (1987, 1988) pioneered the use of 3 Alum wavelength for sensitive and high-
resolution observations of developing clouds and precipitation. Use of this short
wavelength offers unique opportunities for both airborne research (Leon and Vali, 1998;
Pazmany et al., 1994) and ground-based studies (Martner et al., 2002~.
The primary attributes of these radars are superb sensitivity and resolution (~50
m), which enable them to detect very weak targets, such as non-precipitating clouds with
remarkable detail and without the need for large antennas and powerful transmitters The
small size and weight of their hardware components makes mobility highly feasible.
Trailer-mounted, truck-mounted, and airborne versions are now in operation arid the first
space-borne cloud r adar (CloudSat) will be launched in about 2005. The main
disadvantages of millimeter-wave radar are severe attenuation by liquid water clouds and
rain and limited range coverage. Thus cloud radars are best suited for short-range
observations of the fine-scale structure of clouds, snowstorms, arid weak rainfall.
These radars can possess all the scanning' Doppler, and polarization-diversity
capabilities that have been developed originally for the much larger microwave radars. A
decade of research at NOAA/ETL on polarimetric identification of cloud hydrometeors
with millimeter-wave radar (for the purpose of remote detection of aircraft icing) has
derived hydrometeor polarimetric signatures (Figure 4.1) that have obvious applications
to cloud-seeding experiments (e.g., Reinking et al., 20023. Short-wavelengtl~ cloud
radars, especially airborne units, hold great promise for revealing the physical
transformations in the seeded regions of clouds. Longer wavelength radars, however, are
OCR for page 49
TOOLS A ND TECI-~.VIQ lJES FOR A D VA NCIN( r O UR UlN'DERSTA NDI.l\TG 49
w40 .
:.
.~
or
.
.
+~D :~;
at.,
. i
.
<^, ~ hi: Pl~3 i/
'I, ~.~%~.
~ ~~ ~ J
. .
- ~ - ~~.~"
· s .~
~ ^~ Is—~~ ~ ~~ ^ N—W ~.-
. .
X
a!
· 4,,:
· <
FIGURE 4.1 Depolarization ratio as a function of antenna elevation angle, showing signatories of
various hydrometeor types obtained with scanning milJimeter-wave cloud radar. Each signature
type has been matched to theoretical model simulations arid verified with in site particle sampling.
SOURCE: Reinking et al. (20001.
likely to restrain the primary tool for observing and assessing the ultimate desired result of
seeding in terms of precipitation reaching the ground.
Combining simultaneous cloud radar and radiometer observations of clouds
overhead to retrieve estimated profiles of hydrometeor mass content, median size, and
concentration has become a routine procedure at the U.S. DOE CART sites and in other
cloud/climate research experiments. Millimeter-wave radar data are combined with
microwave radiometer data for retrievals in liquid clouds, such as stratus (Frisch et al.,
1 9951, and with infrared radiometer data for retrievals in optically thin fee clouds, such as
cirrus (Matrosov et al. 1992). Retrievals of properties in mixed-phase clouds are more
problematic. These kinds of active/passive remote sensing combinations could benefit
eloud-seedi~g research' particularly if the theory and technology can be extended to
scanning applications.
Perhaps the most impressive demonstration of the combined use of cloud radar
arid microwave radiometers in a cloud-seeding experiment is the ease described by
Reinking et al. (2000~. Earlier numerical modeling simulations by Bruintjes et al. (1994)
indicated that under certain wintertime stability and airflow conditions, the mountains of
central Arizona initiate the development of a strong gravity wave, which produces
sustained updrafts that condense vapor into significant amounts of supercooled liquid
water. This orographieally induced standing wave of supercooled liquid represents an
attractive target for glaciogenic seeding to increase snowpack on the downwind
Mogollon Rim, which is the state's major water supply source. A field program
OCR for page 50
so
CRITICS L I.SS(.JES IN IDEA TlIER AlODIFICA TION RESEARCH
incorporating ground-based remote sensors and aircraft observations was established in
1995 to investigate the n~odel predictions. Plate 7 shows a prominent wave across the
Verde Valley as observed by a scanning cloud radar and strong accentuation of liquid
water content in the ascending part of the wave measured by a steerable microwave
radiometer, thereby confirming the model prediction.
GPS and Radar Cell Tracking Software
In recent years cloud-seeding operations have relied heavily on sophisticated
real-tine displays of the radar reflectivity of storms and the location of seeding aircraft to
manage and assess seeding operations. Although there are many cell-tracking programs,
such as the one described by Rosenfeld (1987~' the TITAN software package developed
at NCAR is most used among these systems (Dixon and Wiener, 1993) This software
objectively identifies discrete storm cells, follows their movement and development, and
keeps statistics (Plate 84. In addition to providing guidance for real-time operations,
TITAN is used extensively in subsequent analysis to examine the effects of seeding, in
terms of reflectivity enhancements, on treated stolen clouds. It has become an important
tool in many operational convective cJoud-seeding operations and represents a valuable
aid for automating the display and analysis of radar data. TITAN has evolved since 1993
and has several features that are specifically aimed at weather modification applications.
Among these are the ability to distinguish independent cells within merged cells, and the
use of an altitude threshold that mitigates the effects of the Earth's curvature. In weather
modification research an annulus between 15 km and 90 km is usually used as the region
in which echoes are reliably tracked.
For TITAN to be effective, accurate location of seeding and research aircraft is
essential. This was a significant impediment to many weather modification studies in the
past. The advent of the GPS now provides a superb and inexpensive tool for this purpose
(Plate 7i. In addition ground-based GPS receivers, in combination with other co-located
routine temperature and pressure measurements, are now available as a national network
(Ware et al., 2000) for measurements of column-integrated water vapor, a necessary
measurement in weather modification research. Dense networks of such measurements
could be cost-effectively deployed in future experiments. Finally, GPS tracking is now
used with radiosondes to provide very high-resolution vertical profiles of temperature?
humidity, and winds (Hock and Franklin, 1999; Aberson and Franklin, l999~.
Satellite Imagery
Satellite-borne instrumentation provides horizontally contiguous observations of
water vapor fields, aerosol amounts and particle sizes, cloud-top temperature, particle
size and thermodynamic phase, and to a limited extent in-cloud processes and
precipitation over a large aerial extent. For instance, the Tropical Rainfall Measuring
Mission (TRMM) includes precipitation radar, a microwave imager, and a visible-
infrared radiometer, all of wl~ich will help improve modeling and prediction of rainfall
processes. CloudSat, an upcoming multisatellite, mu]tisensor mission, will utilize a
millimeter-wave radar to profile the vertical structure of clouds, and measure the profiles
OCR for page 51
7 00LS AND TECH IQZJES FOR AD VANCIN(~r OUR {!~7\~DERSTANDI.INTG
5!
of cloud optical properties cloud liquid water, and ice-water content. These data can be
used to evaluate and improve the way clouds are para~neterized in models. The Global
Precipitation Measurement (GPM) Microwave Imager will utilize a series Ott passive
microwave radiometers to provide near-global measurements of precipitation.
These capabilities have opened a resew era in cloud physics and could provide
many new opportunities for assessing the effects of weather modification. Satellite
observations already are playing an important role in studies of inadvertent weather
modification by tracking plumes of industrial pollution arid their ejects on precipitation
suppression, as well as hydroscopic effects of salt aerosols that aid in restoring
precipitation. Rosenfeld and Lensky (1998) developed a new methodv]ogy for using
T RMM and flee Advanced Very High Resolution Radiometer sensors to infer the
microstructure e of convective clouds and their precipitation-forming processes with
height.
1
Ir'Situ Measurements
Robert Knollenberg pioneered the development of laser based measurements of
the particle size distributions in clouds. These revolutionary devices usually mounted on
the tips of research aircraft wings, use laser light to image and count particles.
Knollenberg probes rapidly became the tools of choice for cloud physics researchers.
These Particle Measuring Systems, Inc. (PMS) probes (Knollenbe~g, 198] ~ together with
hot-wire liquid water probes (King, 1978) have been the principal instruments for
characterizing aerosol and cloud particle properties for the past two decades. They are
useful for understanding the types and numbers of hydrometeors and their evolution.
They have also been used to develop interpretative algorithms for ground-based radar
measurements. I n many weather modification experiments the probes have been
deployed to observe the hydrometeor evolution that takes place before and after seeding.
Through the years new probe designs have evolved, and they now cover a wide
range of particle sizes. Some designs use forward scattering to detect very small particles,
including aerosols. At present, however' no single instrument can provide simultaneous,
accurate information about cloud particle spectra and liquid water content. A combination
of instruments is needed, and this situation seems unlikely to change in the near future.
The Passive Cavity Aerosol Probe measures the size distribution of aerosol
particles between 0.1 Em and 3 Em diameter in 15 size channels. The Forward Scattering
Spectrometer Probe (FSSP-100) measures cloud droplet distributions between 0.5 Am
and 47 Em diameter in 15 size bins. Anothe1 version of this probe (FSSP-300) with
higher size resolution for aerosol and cloud droplet sizes between 0.3 him and 20 aim
diameter has also been used extensively. The Fast-FSSP (Brenguie~ et al., 1998), an
improved version of the FSSP-100, provides better sizing of the droplets and more
accurate determination of the concentration of particles.
Several optical array probes have been developed to measure the concentration
and sizes of larger particles. The technology in use currently is the Optica] Array Probe
(OAP-260X) which measures the concentrations and sizes of particles between 40 Em
OCR for page 52
52
CRITIC.'A L I~SSlJES IN [YE-A TI--IER AIODIFICA PlON RESEARCf-f
and 640 Em diameter. Optical arrant probes have also been developed to provide two-
dimensional images of hydrometeors, with a resolution of 25 Bin for cloud particles and
300 Bins for larger hydrometeors such as large ice crystals and raindrops.
The Cloud, Aerosol and Precipitation Spectrometer (CAPS) (Baumgardner et al.,
2000) instrument consists of five sensors: the aerosol and cloud droplet spectrometer
(CAS) (0 35 Am - 50 Am diameter), the cloud imaging probe (CTP) (25 M-1550 Am
diameter), the liquid water detector (0.01 gm~3-3 gnat), tile air speed sensor, and a
temperature probe. The CAS measures the conventional fo~ward-scattering light from
single particles but also the back-scattered light that provides an estimation of the aerosol
refiactive index. in addition, the sample volume is defined similar to that used in the
FSSP-300X (Baumgardner et al., 1992~. These improvements provide an extended size
range of particle measurement that covers much of the accumulation mode aerosols arid
up to small drizzle drops in clouds. Due to the improved electronics nanny of the
limitations associated with the FSSP-100 have been overcome. The principal
improvements of the CIP are added stability against vibration' decreased response time,
and decreased dead tickle that provides for better resolution, sizing, and more accurate
particle concentrations. The liquid water content detector uses technique described by
King ~ 19783. Preliminary results using the CAPS have shown increased capability
compared to the conventional PMS probes.
A new generation of particle spectrometers uses optical response rather than
direct single-particle collection. The Gerber Particle Volume Monitor (Gerber et a].,
1994) measures the liquid water content, drop surface area, and effective radius. The light
scattered in the forward direction by an ensemble of drops is optically weighted and
summed on a photodetector. The Cloud Droplet Spectrometer (CDS) (Lawson and
Cormack, 1995) measures the forward-scattered light from an ensemble of drops Tl~e
CDS also computes drop size from the raw scattered light by inverting the measurements.
The measurement has inherent advantages to overcome the limitations of single particle
sizing and counting methods. Lawson et al. (1996) describe preliminary measurements
with this instrument.
Another instrument, the Cloud Particle Imager (CPI) uses innovative new
technology to record high-definition digital images of cloud particles and measure
particle size, shape, and concentration (Lawson, 1997; Lawson and Jensen, 19981. Tl~e
high quality of the CPJ images supports the generation of individual size distributions for
different types of particles (see Figure 4.2~. Due to varying depth of field (depending on
the size of the particles), the imaging sample volume of the CPI varies from about 0.002
cm3 to 0.2 cm3. A drop-off in particle detection efficiency starts at about 25 ~m, thus the
small end of narrow particle distributions (such as a typical distribution of cloud drops)
will be undercounted. Research is ongoing to interpret the measurements from this
instrument and its operational limitations. Korolev et al (1999) described some recent
measurements using this instrument.
Another important parameter is the measurement of LWC. While LWC can be
calculated from the FSSP, the most widely used instruments have been the Johnson-
Williams and CSIRO-King probes. The LWC is determined from the cooling effect of
cloud droplets impinging on a treated sensor element thatis exposed tothe airflow
OCR for page 53
TOOLS AND TECF-INIQZ,rL-S FOR ADI'ANCI.~(, OUR lJ,N'DERSTANDl.iN~G 53
::
~~ ~~ ~~ ~ ~ ~0
,] PleCes of ~~.~d o.~-~.~-~S
I:'
: ~
FIGURE 4.9 Particle images from the CPI instrument SOURCE: Lawson, et al. (19981.
Outside the aircraft. Limitations exist for all instruments measuring LWC, but for the
King probes, errors occur when droplet diameters become greater than 50 Em as droplets
break up on the sensing element and are removed by the airflow before they evaporate
completely; this causes an underestimation of liquid water. Large quantities of ice
particles also are a limiting factor (FIeishauer et al., 2002~. The Gerber and CDR probes
are also used to measure LWC. A comparison of more than 20 different types of probes
(Strapp et al., 2000) indicated that the Nev~orov total-wate~-content probe (Korolev et a].,
1998) is the most accurate }~ot-wire estimate of LWC in water-only clouds Title large
do oplets.
T
racers
A difficult problem that has plagued matly cloud-seeding experiments and
operations is the question of whether the seeding material actually reaches the targeted
regions of cloud, and whether it arrives there in effective concentrations. This is
especially true for g~ound-based seeding operations, but it also applies to seeding from
aircraft. Tracer techniques offer valuable information on nucleant transport and
OCR for page 54
54
CRITICAL I~S~Sl.JES IN TT'F,A TI--IER AlODIFIC,4 T10~ REtSE-ARCI!
dispersion. The tracer is released together with the seeding material, and its location and
concentration is subsequently measured as a proxy for the t~ucleant.
The Almost widely used tracer for cloud seeding is SF6, an inert, anthropogenically
produced compound that can be detected in incredibly small concentrations (Stith and
Benner? 1987) but requires in situ sampling, which can be difficult. Other i'? Mu
techniques include airborne ice-nuclei counters and chemical analysis of the silver
content (i.e., seeding material) in snowfall.
A particularly promising remote-sensing tracer method uses radar to track
microwave chaff, which consists of very thin aluminum-coated glass fibers cut to half the
wavelength of the observing radar. Chaff fibers released with or without seeding material
show by direct measurement the actual transport and dispersion occurring within clouds.
T lee fibers can be detected by radar in extremely small concentrations. Tl~e depolarization
of the radar signal (the depolarization ratio) caused by the chaff allows it to be isolated
from the signal of cloud intensity (reflectivity) and to be effectively tracked (partner et
al., 1992; Reinking and Martner, 19961. The volume treated and the location of treatment
effects thus can be identified and assessed in relation to flee total cloud volume. The
concentration of chaff fibers can be computed from the radar measurements to yield
information about diffusion rates. Although the chaff fibers fall faster than silver iodide
aerosols (i.e., the seeding material), they provide a good approximation of the aerosol
movement for several minutes after a release. This allows a polarization-diversity radar to
observe and provide three-dimensional depictions of seeding aerosol movement to a
treated cloud, as shown in Figure 4.3. Chaff tagging offers additional opportunities to
remotely sense microphysical changes between tags. For instance, using such tagging, ice
particle production arid enlargement by seeding has been followed from the source to
snow on the ground (KIimowski et al., 1998; Reinking et al., 1999, 20004.
1 1
All of these tracer methods have had modest demonstrations in weather
modification research experiments, such as the 1993 North Dakota Tracer Experiment, a
summer convective cloud-seeding tesearch experiment that emphasized the use of a
variety of tracer methods (Stith et al., 1996~. But none has yet gained widespread, routine
usage. Nevertheless, tracers ale likely to be an important part of future seeding research
because they offer vital observations of both the seeding material deliver>" arid else cloud
r espouse.
MODELING AND DATA ASSIMILATION
Numerical modeling should be a lcey component of weather modification
research. Computational resources are now probably sufficient to allow realistic cloud-
resolving simulations with short-term predictive value. A properly constructed simulations
model is internally self consistent, complete in spatial and temporal coverage, and
suitable for comparison with datasets. Such a model also can be the basis for a data
assimilation process, which allows incomplete observational data from various sensing
systems to be used to initialize a model's predictions. To h~11;11 these needs the
microphysical processes relevant to weather modification need to be carefully
incorporated and tested in the models, a process that is well under way. The
OCR for page 55
OCR for page 57
OCR for page 58
OCR for page 59
OCR for page 60
OCR for page 61
OCR for page 62
OCR for page 63
OCR for page 64
OCR for page 65
OCR for page 66
Representative terms from entire chapter:
field studies
TOOLS A ND TECI-I.VIQZ;TES FOR A D I7A NCIN( r O l;R U~7~7DERSTA NDI.IN7G 55
7~-~'ng Cumulus
5~0~ Y[_'`/'J ~ ~
~ ,X ~ I,
.,, ~ I'
'
56
CRITICS L I.SSZJES IN [VEA TI-~lER AlODIFICA T10N RESEARCI!
establish or refine physical hypotheses. They offer the only opportunity to see the effects
of cloud seeding on identical (model) cloud situations, one seeded and once not seeded.
They may be used to recreate cloud-seeding experiments from the past to help in the
evaluation of those cloud-seeding effects. alley can be used to simulate the dispersion
trajectories of seeding material, provide r eal-time forecasting in support of field
experiments and operations' examine flee potential effects of cloud seeding outside of the
seeded area, and aid in the statistical analysis of weather modification experiments.
The following sections review the history and methodology of modeling related
to weather modification and evaluate future capabilities and needs. During flee last 20
Yeats cloud and storm modeling have been pursued most seriously tar basic research and
application to prediction and warning and to a lesser extent for application to weather
modification. 1~ an important review article Orville (1996) surveyed the progress of
modeling related to weather modification to that date. A snore recent review has been
presented by Khain et al. (2000), and a substantial account of the NASA-Goddard
modeling activities is given by Tao et al. (20031. The following account is based partly
on these surveys.
Cloud Modeling History Old Methodology
Cloud mice ophysics and dynamics have developed ghostly from different
academic bases. The discipline of cloud mice ophysics was developed mainly by
physicists, while cloud dynamics tended to be a blanch of fluid dynamics developed
mostly by engineers, meteorologists, and oceanographers. A few scientists focusing on
cloud processes have attempted the difficult task of combining these sources of
knowledge. The theoretical bases of both dynamical and cloud mic~^ophysical processes
have existed for some 30 to 40 years. Computing facilities and techniques, however, were
much too limited to allow realistic model simulations until fairly recently. Early models
of microphysical processes tended to be based on assumed particle trajectories, Title
almost no dynamic contents while early cloud dynamics models contained only the most
limited microphysical parameterizations. As computing hardware and numerical
technology evolved, the dynamical and microphysical simulations advanced and became
mutually accessible.
An early but sometimes still used form of modeling is based on the plume
theories for convection developed by fluid dynamicists in the 1940s and 1950s, Fist
applied to prediction of nuclear bomb effects (Mo~4on et al., 1956~. A few one-
dimensional equations are applied, representing the budgets of mass, buoyancy, moisture,
and momentum in a cloud. These one-dimensional steady-state models ate based on
ordinal y differential equations, and they have coupled microphysics and dynamics
(Simpson et al, 1965; Simpson and Wiggert, 1969; Cotton, 1972) In the more modern
versions a realistic environment may be assumed, with natural convection forced by
condensation heating and freezing Cylindrical or slab symmetry normally is required'
which limits or neglects the effects of mean shear. Microphysical processes may be
simulated, but neither the distribution of seeding agents nor the trajectories of
precipitation particles can be realistically followed. A list of such models' designated as
TOOLSA~D TECIl.~[QZJESFORADVANCIN(r Ot;R ti,N'DERSTA~D{NG
57
"one-dimensional steady state" or'`one-dimensional time dependent" is given by Orville
(19964.
The first non-steady numerical simulations of cloud convection date front the
1960s (Ogura, 1963; Orville, 1965) and were two dimensional, usually slab symmetric.
Precipitation was introduced with varying levels of sophistication in flee late 1960s' and
attempts at thunderstorm simulation were made by Takeda (19713. The importance ofthe
third dimension followed the clarification of the important differences between two- and
thtee dimensional turbulence- by Fjortoft (1953) and Kraichnan (1967~. The first three
dimensional simulations of boundary layer stratocumulus, cumulus, and deeper
convection were presented in the mid-1970s. Those which produced the greatest impact,
however, were the Klemp and Wilhelmson ~ 1978) and subsequent simulations (see
review by Klemp, 1987), which showed how shear could contribute to convective
dynamics and produce tl~understorn~s with strong rotation and other observed '~supercell"
characteristics. "Bulk" microphysics were used, with just two categories of liquid water:
cloud and rain. The transformation front cloud water to rainwater involved crudely
simulated processes of autoconversion and collection.
Models aimed at more accurate simulation of microphysical processes (usually at
the expense of dynamic reality) were also being developed. These included the Orville
and Kopp (1977) hailstorm model and later the Orville and Chen (1982) simulations,
oriented specifically to cloud seeding. In the latter the n~ic~ophysical module—though
still confined to "bulk" processes—contained four categories of cloud ice with fairly
complex conversion algorithms, but the domains remained two dimensional. The correct
simulation of the thermodynamic effects associated with precipitation processes-
melting' evaporation, arid recycling of ice and water particles into new cloud updrafts is
usually dependent on having three dimensions and fairly high resolution.
Since Orville's (1996) report, it has become possible to incorporate more detailed
cloud physics algorithms into three-dimensional dynamics simulations. The original
single moment bulk schemes were expanded to two moment schemes (Meyers et al.
1997), allowing noose freedom for the distributions of hydrometeors to respond to
ohvsical processes. A method used freoue~tlv now is to define the mass distribution of
~ 1 1 ~
particles by blnS covering size ranges, With each On larger by some factor shall the
previous one. The particles in each bill ate allowed to grow or shrinl; by condensation,
evaporation, deposition, and coalescence; to freeze or melt; to settle gravitationally; and
to shed water or break up into smaller drops. Thus the number of particles in each bin
mar increase or decrease with tinge. This method obviously requires greater computer
memory and speed than for the bulk process assumptions. These simulations were first
done in a zero-dimensional mode that follows a supposedly uniform parcel up or down
(Berry and Reinhardt, 19741. Later the models were pursued in two or three dimensions
in the context of cumulus clouds (Kogan, 1991) or shallow cloud-topped mixed layers
(Kogan et al., 1995), for which the microphysics consists of purely liquid water
processes. More recently simulations have been carried out for deeper clouds with large
drops, freezing processes, and simulated seeding with cryogenic or hydroscopic agents
(Khan et al., 2000, 2001; Khain and Sednev7 1995, 1996; Reisin et al., 1996a,b; Tao et
al.' 2003; Tzivion et al., 1994; Yin et al., 200Oa,b, 2001;~. Bin models also recently have
been applied to marine stratocumuli (Feingold et al., 1999; Jag et al., 2000, 2001,
sS
CRITICA L ILSSIJES IN [VEA TI-lER AlODIFICA TION RESEARCI-f
20021. As illustrated in Figure 4.4, the Goddard Cumulus Ensemble model, as well as
several other cloud models, can simulate multicell convective systems and be nested in
the framework of larger-scale models and observational systems (Tao, 200~.
(/ Passive Radiative'
\: Model ~7
\~ I inhtnino J
I
i;.; ~
Goddard Radiation
_
LW and SW
Radiation
;,
O
.0
en
a'
to
an,
Scale
Interaction
Sea Surface Fluxes
TOGA COARE
~ ,
_
~.~ ~ 1
,~
1'
GCE Model
T. Q. U,V,W, P. Ke
c,
O I
. ,
(q,
N I
Call
m' an
at,
(L,
i
GCM and
Climate Model
2-l\Aoments
Microphysics
qc, qr, pi, qs, qg, qh
( - - 3
Soil/Vegetation
(7 layers) J
PLACE
FIGURE 4.4 Schematic diagram showing the characteristics of the Goddard Cumulus Ensemble
(GCE), a cloud-resolving model that includes explicit representation of warm rain and ice
microphysical processes. Its main features are described in Tao et al. (20031. Arrows with solid
lines indicate a two-way interaction between different physical processes and arrows with dashed
lines indicate a one-way interaction. SCM stands for Single Columns Model, a one-dime~sional
model with all GCM's physical processes PLACE stands for Parameterization for Land-
Atmosphere Cloud Exchange, a detailed interactive process model of the heterogeneous land
surface and adjacent near-surface atmosphere. The model variables include horizontal (u, v) and
vertical velocities (w), potential temperature (T), perturbation pressure (p), turbulent kinetic
energy (Ke), and mixing ratios of all water phases Water vapor (Qj, liquid (cloud ~vater/qc, rain
drops/qr), and ice (cloud ice/qi, s~ow/qs, graupel/qg, hail/qh)~. Recently, detailed spectral-bin
microphysical schemes were implemented into the GCE model. The formulation for the explicit
spectral-bin microphysical processes is based on solving stochastic kinetic equations for the size
distribution functions of water droplets and several types of ice particles. Due to extensive
computation, this microphysical scheme can only be run on the two-dimensional version of the
model SOURCE: Wei-Kuo Tao, NASA/Goddard Space Flight Center.
TOOLSAND TECHNI QUES FORADVANCI NG OUR UNDER~ANDI NG
Cur r ent Status and Pr aspects
59
The most fully reported cloud simulation model relevant to nucleation,
preci pi tati on, and weather modi f i cati on studi es are the model s of the two I srael i groups,
one at the University of Tel Aviv developed by Tzivion and associates, the other at the
Hebrew University of ~Jerusalem, developed by Khain and associates. The group at Tel
Aviv focused more on the hydroscopic seedi ng agents, whereas at Jerusalem they focused
more on the effect of variations in the natural and anthropogenic aerosol on the
precipitation formation process. Yin et al (2001) found that seeding with hydroscopic
flares produces changes in the hydrometeor distribution, with resulting changes in the
radar reflectivity-rat nfal I rate relationshi p. Such changes are significant si nce radar is the
primary evaluation tool for precipitation enhancement projects. Khain et al. (1999) report
on si mulations of cold season clouds over an eastern M editerranean coastal zone i n
conditions of large-scale convergence that lead to significant precipitation. They
concentrate attention on the effects of varyi ng amounts (100, 500, and 1000 CCN cm~3),
verti cal di stri buti ons (uniform or decreasi ng upward), and types (sodi urn chl ori de and
ammonium sulfate) of condensation nuclei. They found that although most of the rain
forms from melted snow or graupel, the larder droo sizes Generated bv the cleaner air
, .. . . . . . .
_ - _ _ , ,
_ . _ ,
smear cc~x counts' produced rain much faster and that the total amount of rain was
sensitive to the nucleus type (greater for ammonium sulfate). Neither of the results of the
two groups coul d have been obtai ned by exi sti ng bul k model approaches.
Other model i ng groups have adopted approaches to mi crophysi cal model i ng
similar to that of the Israelis. A m for contributor is the NASA Goddard group, whose
cl oud-model i ng results were recently summari zed by Tao et al . (2003~. The pri mary
emphasi s of the Goddard group i s cl ouds and preci pi tati on as mat or i nputs to gl obal and
regional climatology, but here too the microphysical interactions are often crucial . For
exampl e, the f ormati on of I ong-l i ved real dual ci rrus sheets i s cri ti cal to the radi ati on
budget, which then feeds back into the cloud dynamics. Also precipitation efficiencya
the fraction of cloud liquid water that reaches the ground as rains is important both for
climatological and weather-forecasting purposes, and it apparently is strongly dependent
on mi crophysi cal processes. Tao et al . (2003) report on three versi ons of mi crophysi cal
si mul ati on, i ncl udi ng i ce processes, two of them rather sophi sti cat ed bul k model s and one
a hi n model . M ost of the results shown are compari sons of model s with each other, rather
than with observati ons. Compari sons of hi n model results wi th hi gh and I ow CCN counts,
in this case for entirely liquid clouds, indicate considerably greater rainfall for the clean
al r case.
Despite the progress that has been made, model predictions of hydrometeor
evolution are not sufficiently accurate to inspire great confidence. Errors arise from
I i mited revel ution, i Insufficiently accurate physics, and i nadequate observati ons. Bryan et
al. (2003) point out that thetypical resolution of simulated cloud and storm models, about
1 km, is insufficient to resolve the inertial range and predict dissipation. This is important
because condensation, freezing, and coalescence appear to be dependent on at least the
statistical structure of small-scale turbulence as principally defined by the dissipation
rate. Resolution of order 100 m is found to be necessary for fairly accurate dynamical
simulations, which stretches computer capabilities close to the limit, even without the
best treatment of hydrometeors. Obs~vati onal I i mitati ons i ncl ude the revel uti on of
60
CRITI CAL I S9UES I N WEATHER MODI Fl CATI ON RESEARCH
humidity measurements and the very li mited observational knowledge of the size and
compost ti on di stri buti on of condensati on nucl ei and the di stri buti on of temperatures at
which freezing nuclei become effective. New methods of remote sensing may
significantly i mprove the humidity observations, but the nuclei are only observable i n situ
from i Instrumentation at ground stations or on a few research al rcraft, although alternative
methods of nuclei retrieval are being explored. The model physics are again subject to
computer I i mi tati ons (and cl everness of desi gn), but model i ng of the i nteracti on between
ice and water speciesO and even between water drops themselves, whether the same size
or notOrestson largely untested hypotheses.4 Accurateprediction of the hydrometeor
distri button development is critical to getti ng the dynamics-mi crophysics i nteracti on
correct, since hydrometoors determine (through sedi mentation) the location and ti mi ng of
latent heat release and precipitation loading impacts on cloud dynamics. It is exactly
these detai I s of the hydrometeor di stri buti on devel opment that cl oud seedi ng tri es to alter.
Thus, while bin models have many degrees of freedom and thus can simulate many
physical situations realistically, much of the knowledge necessary to specify parameters
needed i n thei r i mpl ementati on i s sti 11 1 acki ng.
Data Assi mi ration, M odel I nitial ization, and Advanced Forecasti ng Systems
Methods of optimally assimilating observed data and generating a series of fields
sui tabl e f or i ni ti al i zi ng a predi cti on model have al ways been cri ti cal parts of I arge-scal e
numerical weather prediction, but at the convective scaJ es, models have been under
development for only 10 to 15 years. The potential for assimilation of fin~scale Doppler
radar data, and from it establishing the dynamic and thermodynamic fields, was a major
el ement of the proposal for the Center for Anal ysi s and Predi cti on of Storms, one of the
first of the NSF science and technology centers. Most of the methods developed or
adapted by the Center~; scientists and others are variational in nature, involving
minimization of the integral of an error function. Among the most sophisticatecl is the
adj oi nt method. The adj oi nt of a set of predi cti ve equati ons i s a si mi I ar set whi ch predi cts
backwards the weightings of variables at a previous time which contribute to the change
of a variable at a given position and current ti me. This al lows, i n pri nci pie, opti mal
utilization of current and previous data to produce an initial state for a future prediction.
The adjoins method has shown fairly good success in obtaining three-dimensional
i nitialization from single Doppler radar data (Sun and Crook, 1997, 2001; Xu et al .,
1994), but it is rather expensive, often requiring the equivalent of 50 to 100 time
integrations for a few minutes each. Methods for speeding the convergence are under
active devel opment.
~For example, in a model with many different bins of ice and water species, the rate at which ice
particles (of size 1 mm to 2 mm) combine with water droplets (of 1/8 mm to 1/4 mm) is a
parameter that must be specified. This is a function of drop-size distributions, turbulence,
temperature, the hydrodynamics of sedimentation, and, to a lesser extent, electrification of the
cloud. Similar rate constants must be specified for all pairs of particle bins.
TOOLS A ND TECHNIQI J,ES FOR A D I7A NCirN( r O UR Z>Tl\7DERSTA NDI.iNtG
Future Prospects
61
Models and data assimilation offer the possibility of greatly ameliorating the
difficulties of past statistical verification described in this report. With today's itnproved
statistical techniques and sophisticated models sources of uncertainty can be explicitly
accounted for' and treatment and control experiments can now be compared spatially and
temporally. The computational facilities and human resources necessary for work in these
areas exist and can be rapidly developed at a number of governmental (e.g. NCAR,
NOAA' NASA) and non-governmental laboratories and university groups for application
to weather modification. Development of a cloud and precipitation model suitable for
planning and testing seeding experiments may be feasible using the cutting edge of
current simulation modeling. However? for real-time modeling studies that run
coincidentally with field experiments a model would need to run faster (and therefore
nary be confined to a spatially coarser mesh and have less physical complexity) and
would require data assimilation and initialization techniques that include microphysical
paran~eters. Again, the techniques used for storm analysis and experimental prediction
help point the way, although they have not been applied to the newer methods for
observing water substance and please, and methods need to be developed for rapid
assimilation of these data types.
Model forecasts are always uncertain. Increasingly, predictions of large-scale
models are presented as probabilities or ensembles. These probabilistic forecasts attempt
to account for the uncertainties inherent in initial conditions, boundary conditions' and in
the models themselves (especially the model parameterizations of subgrid-scale }physical
processes). Similar approaches should be used to quantify the uncertainty in simulations
of weather modification experiments, including uncertainties related to the experimental
treatment.
LABORATORY STUDIES
Laboratory investigations play an integral role in advancing the understanding of
cloud physical processes. The high degree of measurement capability? repeatability, and
control over experimental conditions in the laboratory allows r esearch on detai led
processes that is not possible in the free atmosphere.
Rogers and DeMott (1991) provide an excellent overview of the state of cloud
physics laboratory work as of 1990. The most significant development in cloud physics
laboratory studies since the early 1990s is the successful use of electrody~amic levitation
chambers, in which nucleation and vapor deposition properties of individual, freely
suspended hydrometeors earl be studied in a fully controlled environment (Straw et al.,
2000; Swanson et al., 19994. Other important reseal ch continues on drop-drop
interactions (Beard et al., 2001), on primaty fee crystal habits arid the impacts of growth
and evaporation cycles (Bailey and Hallett, 2002), on nucleation coefficients of liquid
and ice phases (Bailey and Hallett, 2002; Shaw and Lamb, 1999; Xue and Lamb, 2002),
and on the growth of ice crystals in a water-saturated environment (Fukuta and
Takahashi, 19991.
62
CRIT1C,4L i`SSIJES IN [YF,A TI--IER A1ODIFIC,4 TION RE5~EARCIT
BOX 4.1
Hurricane Modeling and Prediction
As noted in DeMaria and Gross (2002), hurricanes present a particularly
difficult modeling challenge in which a fairly sn~all-scale, circularly symmetric
disturbance (the storm) is embedded in a larger-scale surrounding flow. The lack
of computer power and adequate observations, especially over the oceans' needed
to properly represent initial conditions have been among the greatest difficulties
in hurricane modeling
More than 20 different types of hurricane models have been developed since
1959. Current hurricane simulations are limited to a resolution of about ] 0 km.
with highly parameterized convection schemes. Using nested grid techniques,
higl~er-resolution (~1 km), mixed-phase bulk microphysics models can be
applied to small, critical regions ifs a hurricane, but until these high-resolution
models can be applied to the entire domain of the storm system, only very basic
aspects of hurricane modification theories can be tested.
Since the 1950s hurricane modeling has been divided into track-forecast
models aimed at predicting where the storm will strike land, and intensity-
forecast models aimed at predicting the strength and extent of the storm's winds
and consequent effects on the ocean (i.e., storm surge). Accurate track
predictions require three-dimensional models that can account for the full range
of interactions between the storm and its environment. Despite considerable
advances in modeling hurricanes, the skill of track forecasts frown a numerical
model have only very recently overtaken that of statistical Precast methods
(Emanuel, 20029. Average (24-hour) track errors remain above 70 miles for all
models (DeMaria and Gross, 2002~.
Modeling and forecasting the intensity of a hurricane remains art unresolved
challenge The present generation of models may not have enough horizontal
resolution to capture the fu]] intensity of extreme storms. However' new three-
dimensional storm models (coupled to upper ocean models) should lead to better
understanding of the factors that control hurricane intensity (Emanuel, 19999.
Many other aspects of the hurricane system are not yet adequately modeled,
including the areal extent of storm wields, the storm surge, and precipitation
especially flooding rainfall.
Improvement in theoretical and numerical modeling of hurricanes will
undoubtedly remain a high national priority because of the value of predicting
their behavior with increasing accuracy. Whether or not we can learn enough to
consider modifying hurricanes to mitigate damage remains to be seen. Certainly
any attempt to modify hurricanes must be dependent upon whether their behavior
with and without modification can be predicted accurately and reliably. Even
alien, any serious consideration of l~urricane modification will raise grave arid far
reaching issues of public policy with both ethical and economic implications.
TOOLS AND TECI-~.VIQl;TES FOR ADI'ANCIN(, Ot;R Zi,N'DERSrTANDI.IN:G
~3
List et al. (1986) and Rogers arid DeMott (1991) identified the need for a large
national laboratory facility to study difficult simulation experiments such as the
interactions between particles in the presence of aerosols or gases and electric fields.
Suel~ a facility has not yet been created, nor is fleece even any ~neel~anism for long-term
planning and funding of laboratory cloud physics research. As a result the number of
cloud physics laboratory facilities in the nation has decreased in recent years and there
has been little influx of new talent. There is currently no coordinated etfott to address the
overall process of precipitation Coronation; rather? individual researchers address parts of
the problem as permitted by their existing taeilities. In particular, there appears to be no
on:,oi~g investigation of fee or fee interactions, and only limited facilities to study n~ixed-
phase processes.
There are, of courser constraints on the types of problems that can be addressed
through laboratory steadies; thus flee greatest progress can be made when laboratory
studies are linlced to theoretical and numerical modeling studies and observational world.
D ~ — ~ ~ ~
_ ~ _ _ J ,
FIELD STUDIES
Physical concepts, laboratory findings, and n~nerical models must ultimately be
tested in the field. Field studies have the unique capability of concentrating analytical and
technical tools on a specific problem in a given time and space domain. Progress in
understanding the chain of physical processes leading to precipitation or underlying
severe weather has isolated key uncertainties, as identified in earlier sections. These
uncertainties constitute goals that can be addressed in a hierarchy of field studies. Such
studies progress from limited activities that can build on othet at~nospherie field
programs to dedicated large-scale weather mo~eat~on experiments. Ordeal
uncertainties inherent in the exploitation of atmospheric resources and mitigation of
weather hazards (Box 2.2) need to be addressed if larger-scale dedicated weather
modification experiments are to make substantial advances. Such field studies must be
founded upon testable physical hypotheses and must advance stepwise from the
si~nplif~ed to the more complex. It should be noted that scientists at the Mazatlan
workshop (discussed in Appendix A) identified a number of specific, testable hypotheses
that could form a useful basis for fixture field experiments (WMO, 2000 J.
Because many of the roadblocks impeding progress in weather modification are
part of the wider research problems facing atmospheric science as a whole, these studies
may be pursued on a broad front. Cloud formation, precipitation generation, and the
dynamics of severe weather are all of interest to a large number of atmospheric scientists
Opportunities thus abound for the pursuit of basic studies of critical concern to weather
modification What is lacking is a centralized program to coordinate this research as a
national elfin in atmospheric resource enhancement and weather hazard mitigation. Such
a program could coordinate modeling? Iaborato~y, and field studies that range from
modest "piggyback" experiments to full-blown, dedicated field studies for testing and
demonstrating weather modification procedures.
These field studies need to be sustained arid would benefit from centralized long-
lived facilities. Such centralized and essentially permanent facilities exist at NCAR,
64
CRITICA L ISSUES IN AREA TlIER AIODIFICA TION RESE-ARCI!
NOAA/ETL~ and the U.S. Southern Great Plains CART established on the
Oklahoma/Kansas border by the DOE ARM Program. NCAR has a long history of basic
and applied research in weather modification with advanced compute and observing
facilities designed to serve the atmospheric research community. Similarly, NOAA/ETL
has contributed significant funding toward weather modification research efforts in the
past. Else CART/ARM site has an extensive array of observing systems detailed in Table
4.1. NASA is planning as part of the GPM to significantly enhance the CART/ARM site.
This array of observing systems with its attendant ir~frastructure presents an
unprecedented opportunity to pursue fundamental questions facing the weather
modification community. While the Oklahoma/Kansas location will not address all
problems of weather modification research, fundamental questions involving flee
formation of precipitation, the distribution and nature of cloud liquid water and ice in
large convective storms, and a host of other more sophisticated experiments, which could
involve actual treatment, are among important problems that can be tackled. The
combined capabilities at NCAR, NOAA/ETL, and the CART/ARM/GPM site constitute
an opportunity that may only require financial and logistical coordination by a central
agency to provide a powerful base for weather modification field studies.
A number of other operational networks and facilities are available that can
advance studies in weather modification; for instance,
· operational facilities of the National Weather Service (NWS) could be used to
conduct comparative, parallel climatological studies in different geographic regions;
· the national operational Doppler weather radar network (NEXRAD) might be
useful in characterizing cloud and precipitation climatologies in neighboring treated and
untreated regions in operational weather modification programs;
. the Oklahoma Mesonet (Brock et al., 1995) provides l~igh-resolution
meteorological data for research, educational, operational' and commercial purposes; and
the Automated Surface Observing System? operated by the NWS and the Federal
Aviation Administration, is a highly sophisticated surface network that provides high-
.
... . .. . . . ~ . . , . . . . . ~ ~ . .
quality data routinely at approx1rmately 1,()()~) sites (mostly at alrpo1^ts) across the United
States.
Ongoing operational programs in weather modification can be improved by the
addition of research components. Ultimately, however, major issues of atmospheric
resource use and hazard mitigation must be addressed by a sustained research effort. Such
a sustained effort ideally rests on an infrastructure of administrative, logistical, numerical,
laboratory, and field support coordinated under a single program
TOOLS AND TECI-I.VIQUES FOR AD VANCIN(.7 OUR l.),N'DERSTANDING
TABLE 4.1 ARM/CART Site Instruments
65
Purpose or parameter
measured
System (if applicable) Instrument
Aerosols Aerosol observation n/a
system
Additional systems Cimel sunphoton~eter
Multifilter rotating sl~adowband
Radiometer
Raman lidar
Atmospheric profiling Balloon-borne sounding system
Microwave radiometer
Radars lidar
50 MHz r adar wind profiler and
radio acoustic sounding system
MASSE
915 MHz radar wind profiler and
RASS
Clouds Belfort laser ceilometer
Micropulse lidar
MilJimeter-wavelength cloud radar
Microwave radiometer
Video time-lapse camera
Whole-sky invader
Narrow field-of-view sensor
Raman lidar
Atmospheric emitted radiance
interferometer
Absolute solar transmittance
interferometer
Cimel sunphotometer
Infrared thermometer
Microwave radiometer
Narrow field-of-view sensor
Rotating shadowband spectrometer
Shortwave spectrometer
Solar radiance transmission
interferometer
Multifilter rotating shadowband
radiometer
MFR (upwelling)
Pyranometers
Radiometers
MFRSR-related
Broad-band
instruments
Radion~etric
instrument systems
Pyrgeometers
Pyrhe]iometers
UV-B radiometer
UV spectroradiometer
Solar infrared radiation station
66 CRJ7~(-~,4 L ISSlJES IN IDEA TI-IER AlODIFR-A TIO.V RESEARCI-f
Surface energy flux Eddy con elation system
Energy balance Bowen ratio station
Infrared thermometer
Soil waters and temperature system
Surface meteorology Chilled mirror
Surface meteorological observation
system instruments
60-m tower: temperature and
humidity sensors
Temperature, humidity, wind, and
pressure sensors
Instruments of Radiometers Solarinfrared radiation station
extended facilities of Multifilter rotating shadowband
the CART/ARM site radiometer
Surface energy flux Eddy correlation systems
Energy balance Bowen ratio
stations
Soil water and temperature system
Surface n/a
meteor ological
observation system
instruments
Instruments at Balloon-borne sounding system
boundary facilities of
the CART/ARM site Microwave radiometer
Vaisala ceilometer
Atmospheric emitted radiance
interferometer
Temperature, humidity, wind, and
pressure sensors
Instep uments at 91 5-MHz radar wind pi ofiler
intermediate facilities Radio acoustic sounding system
of the CART/ARM
site
