Recent Feasibility Studies by

NCAR/RAL of Weather
Modification Programs:
Emphasis on programs with
hygroscopic seeding

Dan Breed
with colleagues Roelof Bruintjes,
Vidal Salazar, and Tara Jensen

International Symposium on Weather Modification

JCSEPA – Tsukuba, Japan
31 January-1 February 2007
 Hygroscopic Seeding Programs

Mexico* – 1996-1998 (1999 – different location)

United Arab Emirates (UAE)* – 2001-2004
(current operational seeding program)
Saudi Arabia – 2004
(current research program in central Saudi Arabia)
Indonesia – 2005
Related Studies: Flare characterization,
Stratocumulus seeding case

*denotes randomized seeding experiments

Hygroscopic Seeding with Flares

RESULTS FROM SOUTH AFRICA

• Recognized altered cloud properties and

enhanced precipitation in storms affected
by the plume of a paper mill.
• Developed a flare to introduce
hygroscopic particles into the clouds at
cloud base – similar to the paper mill
plume.
• Devised a randomized seeding
experiment to detect changes in individual
storms (radar analysis using software like
TITAN).
Supporting Observations – cloud
droplet distributions
South Africa: Argentina:
1000

NATURAL
SEEDED
100

SURROUNDING 10

REGION
1

0.1

0.01
SEEDED
0.001
0 10 20 30 40 50
DIAMETER ( µm)
Flare Test Facility
Flare particle distributions
b. 6
10

South African
5
10 ICE 1090 65%
ICE 1075 70 %

4
5MgAl 1130
10
Peak Conc.

dN/dlogD [cm µm ]
-3
-3
3
10

2
10

1
10

0
10

-1
10
-1 0 1
10 10 10

DIAMETER [µm]

PCASP comparison of different flares

Condensate after 30 min: 0.02 % in raindrops vs. 82 - 85% in raindrops

Model comparison: “natural” aerosols and seeded aerosols

 Recent Results of Flare Experiment in
StratoCumulus (west of Monterey, CA)

Flight track of the CIRPAS Twin Otter during
the seeding experiment on June 28, 2006. The
flight level was at 350 m. Numbers correspond
to plume crossings. The two dashed lines
indicate the initial flare burn lines conducted by
the Cheyenne II aircraft.
PCASP concentrations: Numbers indicate
plume crossings and vertical lines indicate the
boundaries of the plumes. Background areas
are designated by letters.
West plume shows no noticeable
change.

East plume broadens consistently to

4-5 times the number of large droplets.
No drizzle forms however.

East plume distribution (14 and 15)

shows broadening over the
background distribution (H).
* Monclova
Mexico design the same as the
South African experiment
• Cloud/storm characteristics similar: cloud
droplet distributions, storm evolution, size,
and maximum reflectivities.
• Differences: storm lifetimes somewhat shorter
in Mexico. Wind speeds and windshear less
than South Africa also.
• Same case selection criteria (same pilots).
• Same seeding technique and same flares.
• Same evaluation technique (quartile analyses)
using TITAN software.

• 94 cases over two seasons (43 seeded, 51

non-seeded).
 Rain Mass Results
Mexico Experiment
800
700
600
Rain mass (kton)

Thin red line – control (unseeded)

500
Thick blue line – seeded
400
300
200
100
0
-10 -5 0 5 10 15 20 25 30 35 40 45 50
Time from decision South Africa Experiment
500
450
400
Rain mass (kton)

350
300
250
200
150
100
50
0
-5 5 15 25 35 45 55
Time from decision
Mexico Experiment – All Storms
Not Seeded Storms

Log (Rain Mass + 1)

0 1 2 3

-20-15-10 -5 0 5 10 15 20 25 30 35 40 45 50 55 60
Minutes From Decision

Seeded Storms
Log (Rain Mass + 1)
0 1 2 3

-20-15-10 -5 0 5 10 15 20 25 30 35 40 45 50 55 60
Minutes From Decision
 Aerosol and non-aerosol days
classification

Typical non-aerosol day Typical aerosol day

MEXICO RANDOMIZED EXPERIMENT
ONLY STORMS WITH REFLECTIVITY
Not Seeded Storms
Log of Rain Mass > 0
0.0 1.5 3.0

-20-15-10 -5 0 5 10 15 20 25 30 35 40 45 50 55 60
Minutes From Decision

Seeded Storms
Log of Rain Mass > 0
0.0 1.5 3.0

-20-15-10 -5 0 5 10 15 20 25 30 35 40 45 50 55 60
Minutes From Decision
 Feasibility Study for
Rainfall Enhancement in the
United Arab Emirates: Phase I
(2001 and 2002)

Phase I consisted of studies to address two fundamental

questions for the UAE:

1. Is the frequency of cloud occurrence sufficient for

investing in a cloud seeding program?

2. Are the clouds that do occur amenable to hygroscopic

seeding?
• Storm climatologies/statistics using radar data –
summer and winter
• Aerosols and trace gases
• Cloud microphysics
 The results from 2001-2002 supported
proceeding with a Phase II

• Randomized hygroscopic seeding experiment

• Target the Hajar (Oman) mountains during the

summer convective season

• Evaluate storms using radar-derived rainfall

from the Al Ain weather radar with improved
signal and data processing.

• Utilize two aircraft: Research 1

(seeding/research); Research 2 (seeding)

• Develop training program in research and

seeding operations.
COMBINED SATELLITE AND RADAR
(UAE randomized seeding experiment)
 Results
• Estimated that ~250 cases needed to show 25%
increase (per TITAN track)
• 134 “treated” cases – but only 96 met criteria for
acceptance
• 45 seeded, 51 not seeded
 Reasons: Treatment Problems

Long distances to storms, both
from aircraft base and within
seeding area.
Box plots of time interval (in minutes) from the
start of the track to decision time for 2004
cases. A negative time indicates a decision
time prior to the storm developing a track
(reaching 30 dBZ).
 Reasons: Stability and cloud
development

• Capped clouds
• Weak or sporadic updrafts
• Cycling of cells – possibly pumping particles aloft
• Recycling particles into subsequent cells.
Reasons: Microphysical – recycling
 Saudi Arabia (Asir region)
Aerosols and Composition (Saudi Arabia):
In-situ measurements, satellites and models
Vertical measured size NAAPS (model and
distribution of aerosols satellite) optical depth:
between 0.1 and 3 µm Sulfate, dust and
diameter biomass aerosols
July 16 2004

c d
ZN
4
4
10
3.6
Sulfate
Altitude (m )

3.23
10
2.8
2.4
2
2
10
1.6
1.21 Biomass
10
0.8
0.4
0
0
10

Dust
5

5
2

0
5

0
3

5
1

5
6

5
5
0 .3

1 .0
0 .2

0 .6
0 .1

0 .2

0 .4

0 .8

1 .3
0 .1

0 .1

1 .7
0 .1

2 .2
2 .7

PCASP PCASP Channel size

Total Concentration [cm-3] concentration [microns]
Near cloud base measurements

SPP Concentration [cm ]

-3
1000 20

2DC shadowr [l ]
800 15
600
400 2DC
10 Extremely high droplet
200 SPP 5 concentrations during

-1
0 0
incursion of biomass
14 14
12 12 smoke from Africa
10 10
associated with
[µm]

[µm]
8 8
6 6
4
Mean Diameter MVD
4 decrease in mean sizes
2 ER 2
0 0 of droplets
4 0.4
Standard Deviation

3.5 LWC
3 0.35

Disperssion
LWC [gm ]
2.5
2 0.3
1.5 STD

-3
1 Disspersion 0.25
0.5
0 0.2
Microphysical

SPP Concentration [cm-3 µ -1]

50

40 10
2
measurements in clouds
SPP Bin Size

30 10
1
in Saudi Arabia during
20
100 July and August 2004
10 -1
10
0
04
004

4
04

004

00 4

00 4

00 4

004
004
04

04
4

200
200

, 20
, 20

, 20

, 20

3, 2
1, 2

1, 2

1, 2
9, 2

5, 2
2, 2

,
st 5
16,

st 4
y4

y7

st 2
y1

y3
y1

y2
y1

us t
us t
J ul

u
J ul

u
y

Aug
J ul

J ul
J ul

J ul

J ul
Jul

A ug

Aug
A ug

Aug
Lake Matano

40 km 60 km

Lake Towuti

Watershed Area
2477 km2
Sulawesi microphysical measurements
 Sulawesi precipitation process

70 isolated storms over two-week period

• Tropical/maritime environment
• Efficient warm rain process
 Smelter plume

INCO

• Five plume intercepts

• Near cloud base penetrations in clouds close to the plume
 Smelter plume

Cloud penetrations partially influenced

by smelter plume

• Extremely high aerosol concentrations in plume

• Nonetheless, nearby cloud droplet distribution is
still somewhat broad (also evident in other examples)
 Summary
• Mexico – replicated South African results
– Why do storms last longer?
– If drizzle drops are the result of seeding, how does that make ice
processes more efficient?
• UAE – no clear effects from hygroscopic seeding; importance
of cloud thermodynamics/dynamics and background aerosols
– Are these the real or only reasons?
• Saudi Arabia – variability in aerosol background; narrow
droplet spectra and deep convective clouds (ice processes)
– Is knowing the aerosol characteristics and typical droplet spectra
adequate to justify hygroscopic seeding?
• Indonesia – efficient warm rain process; even polluted clouds
show a warm rain process
– How would plume-affected clouds respond to seeding
(hygroscopic or glaciogenic)?