The West Coast of Washington and the NE and SW corners of Wyoming are regions of the contiguous United States where NEXRAD coverage is incomplete. One approach to addressing these gaps is to install additional NEXRAD-class radars. Another potential approach is to install small radar networks of the type being investigated in the CASA project. This paper compares these two approaches. We provide a meteorological and user-need assessment of present radar coverage in these regions (based on a recent feasibility study led by J. Brotzge [1]) as well as an objective assessment of the radar-coverage that would be achieved using the large radar and small radar approaches.
3. Introduction
• Severe tornado events in northeastern
Wyoming and a three-day coastal storm system
in Washington raised awareness of the radar
coverage gaps if the observation current
system (NEXRAD)
• These coverage gaps have several causes: the
distance between radars (345 km in the
western United States); radar beam blockage
due to mountains and other terrain.
• A study that assesses the meteorological need
for and the feasibility of deploying additional
radars to augment the current radar observing
system in northeastern and southwestern
Wyoming and coastal Washington was funded
by a contract from NOAA to CASA through the
University of Oklahoma with Dr. Jerry Brotzge
as Principal Investigator [1]
This work was primarily supported by the Engineering Research Center Program of the
National Science Foundation under NSF Award Number 0313747. Any opinions, findings
and conclusions or recommendations expressed in this material are those of the author
and do not necessarily reflect of the National Science Foundation.
4. Wyoming
• The gap regions in Wyoming lack radar
coverage because of the distance
between radars and mountain
blockage. In Campbell County, no radar
coverage is available below 3 km AGL,
and only 24% of Johnson County and
39% of Sweetwater County have
coverage below 3 km AGL.
• Low-level radar coverage is needed for
routine detection of snow, flooding
rains, and low-topped thunderstorms
(Top) The County Warning Forecast Areas of Wyoming. Figures courtesy of the NWS
(http://www.weather.gov/mirs/public/prods/maps/state_list_cwfa.htm). (Bottom) The existing radar coverage at
heights 1 km (yellow), 2 km (red) and 3 km (blue) AGL. Interstate highways are marked in red and county outlines
are marked in black.
5. Washington
• Radar coverage gaps in coastal
Washington are primarily caused by
beam blockage. Much of the region is
blocked by the Olympic Mountains to
the west, and so there is virtually no
coverage over the ocean where the
majority of western Washington's
weather hazards originate
• Flooding and severe (non-thunderstorm)
winds are the primary weather threats
to western Washington state. While the
number of severe weather events is
relatively low, these events are typically
large in scale, duration, and severity.
• High vulnerability in the coastal counties
Jefferson, Grays Harbor, Pacific because
the lack low-level (< 3 km AGL) radar
coverage
The County Warning Forecast Areas of western Washington. Figure courtesy of the NWS:
(http://www.weather.gov/mirs/public/prods/maps/state_list_cwfa.htm). (Bottom) The existing radar coverage at
heights 1 km (yellow), 2 km (red) and 3 km (blue) AGL. Interstate highways are marked in red and county outlines are
marked in black.
6. Long-range radars Vs. short-range radar
14.0°
NEXRAD
Operating frequency
WSR-88D
2.7 – 3.0
CASA
9.41 GHz
CASA 11.0°
GHz E-SCAN PANELS
Wavelength 10.0 cm 3 cm
Antenna diameter 8.53 m 1.20 m
Antenna gain
Antenna beamwidth
45.5 dB
1.0º
38 dB
1.8º 9.0°
Range gate spacing 250 m 100 m
Maximum rotation rate 36 deg s-1 35 deg s-1
Acceleration rate 15 deg s-2 50 deg s-2
Average transmitter 1.56 kW 9 W/pol
power
Peak transmitter power 750 kW 7.5 kW/pol
7.0°
Min. Detect. Reflect. -23 dBZ -2 dBZ
(10 km) CASA IP1
Pulse repetition 318-452 1.6, 2.4 kHz
frequency 318-1304
pulses/sec 5.0°
Pulse width 1.6, 4.5-5.0 660 nsec
μ sec
Dual-polarization For 2010- Yes
2012 3.0°
2.0°
1.0°
PR1 First Off the Grid
(Stefani) Radar Radar Tested At
“Storm signatures are detected with higher temporal Chill/CSU
and spatial resolution with PR1 data and the radar
seems to be functioning correctly” - Israel Matos (MIC)
7. Possible radar solutions Wyoming
WYOMING STATE
165 km
Domain A:
Area: 165x165 km2
Counties: Short Range Long Range
165 km
• Sheridan
• Johnson
• Campbell Rmax=40km Rmax=230km
Domain B: Rspacing=35km Rspacing=225km
Area: 210x141 km2
Counties:
3dB 3dB
• Sweetwater el _ min 0.9o el _ min 0.5o
210 km
2 2
LEGEND No cone of silence area No cone of silence area
141 km
2km<H<3km Triangular Grid Triangular Grid
1km<H<2km
H<1km
Vertical Coverage Wyoming Domain A (160km x160km)
Wyoming Domain A Height H:1km Coverage: 1e+002%
4 0
0.9
3.5 20
Wyoming – Domain 0.8
Heigh above Ground level (km)
3 40
160km
A 2.5
0.7
60
• Small nodes: 18 0.6
• Large nodes: 01 Range in km
2 80 0.5
0.4
1.5 100
Vertical Coverage 0.3 Coverage in Plan view- Wyoming Domain A
1 120 Wyoming Domain A(160km x 160km) Height H:1km CoverageS: 69%
4 0
0.2
0.5 140 0.9
162 km 3.5 0.1 20
0.8
0 160
Heigh above Ground level (km)
0 5 10 15 20 25 30 35 0 3 50 100 150 40
Distance (km) Range in km 0.7
2.5 60
0.6
Range in km
2 80 0.5
Wyoming – Domain 0.4
1.5 100
B
138 km
0.3
• Small nodes: 16 1 120
0.2
• Large nodes: 01 0.5 140
0.1
0 160
0 50 100 150 200 0 50 100 150
Distance (km) Range in km
8. Possible radar solutions Washington
Short Range Long Range
WASHINGTON STATE
Rmax=40km Rmax=230km
Domain C:
Area: ~360x160
km2
Rspacing=35km Rspacing=225km
360 km
Counties:
3dB
• Lewis
el _ min 0.9o el _ min 3dB 0.5o
• Pacific
• Thurston
2 2
~60 nmi
160 km
• Pierce
• King
No cone of silence area No cone of silence area
• Kitsap
• Mason
Triangular Grid Triangular Grid
• Grays Harbor
LEGEND
2km<H<3km
Washington Domain C(360km x 160km)
1km<H<2km
Vertical Coverage - Washington Domain C(360km x 160km) Height H:1km Coverage: 83%
4 0
H<1km
0.9
3.5
50
0.8
3
100
0.7
Heigh above Ground level (km)
2.5
0.6
150
Range in km
Washington Domain C(260km x 140km)
Vertical Coverage - Washington Domain C(260km x 140km)
2 Height H:1km Coverage: 37%
0.5
4
0
200
3.5 0.9
0.4
1.5
50 0.8
141km
3
Heigh above Ground level (km)
250 0.3 0.7
1 2.5
100 0.6
Range in km
0.2
2 300 0.5
0.5 150
1.5 0.4
0.1
0.3
1 350
256 km
0 200
0 5 10 15 20 25 30 35 0 50 100 150 200 250 300 350 0.2
Distance (km) 0.5 Range in km
0.1
250
0
0 50 100 150 200 0 50 100 150 200 250
Distance (km) Range in km
9. Idealized Radar Calculations
Number of Radars vs. Domain Size Coverage Percentage vs. Altitude
for Short-range and Long-range Radar Spacing for Short-range and Long-range Radar Spacings
100
35
Short Ideal Grid 90
Long Ideal Grid
30 80
Coverage Percentage (%)
Short Ideal Grid
Number of Radars
25 70 Long Ideal Grid
60
20
50
15
40
10
30
5 20
0 10
50 100 150 200
Domain Size (km) 0
0 0.5 1 1.5 2 2.5 3
Altitude (km)
a) b)
a) Number of radars needed to populate a given domain size for short-range (35 km) and long-
range (225 km) radar spacing. The domain considered is square with the side of the square
defined by the "domain size" in kilometers. b) Percent coverage evaluated at a given altitude
for short-range (35 km) and long-range (225 km) radar spacing. Percent coverage is evaluated
at the center of the radar beam, 0.9 deg and 0.5 deg for the short-range and long-range
radars, respectively, and takes into consideration the curvature of the earth.
10. Radar Coverage Simulations
• Illustration of height of the radar beam (AGL) for
3 km
a) "smooth-earth" coverage calculations and b)
those over terrain. Note that when height is given 2 km
as AGL the height is relative to the terrain. 1 km
• The radar beam is considered blocked when the
path integrated occultation is greater than 50%.
a)
• The blockage calculations were performed using
the University of Oklahoma’s Advance Regional
3 km
prediction System (ARPS).
2 km
• Geospatial analysis and visualization was 1 km
performed using Geographic Resources Analysis
Support System (GRASS) Support System (GRASS)
and Geographic Information System (GIS).
b)
11. Radar Coverage Wyoming Domain A
WYOMING STATE + + + + WYOMING STATE
+
Domain A:
Area: 165x165 km 2
+ + + Domain A:
Area: 165x165 km2
Counties: + + + Counties:
• Sheridan • Sheridan
• Johnson + + + + + • Johnson
• Campbell • Campbell
+ + + +
# Radars
LEGEND
Domain A • Short : 21
• Long : 01 +
2km<H<3km
1km<H<2km
H<1km
WYOMING STATE
Domain A:
Area: 165x165 km 2
Counties:
• Sheridan
• Johnson
• Campbell
Coverage Percentage vs. Altitude
for Wyoming Domain A (165km x 165km)
LEGEND Long-range radars: 1 Short-range radars: 21
100
2km<H<3km
1km<H<2km 90 Altitude Coverage (%) Coverage (%)
H<1km
80
(m) Long (1 radar) Short (21
Coverage Percentage (%)
radars)
70 100 1% 7%
60 Short (Smooth) 500 25% 71%
Long (Smooth) 1000 61% 90%
WYOMING STATE LEGEND 50 Short (Terrain) 2000 96% 95%
Long (Terrain) 3000 99% 97%
Domain A: 2km<H<3km 40
Area: 165x165 1km<H<2km
km2 30
Counties: H<1km
• Sheridan 20
• Johnson
10
• Campbell
0
0 0.5 1 1.5 2 2.5 3
Altitude (km)
12. Radar Coverage Wyoming
WYOMING STATE
WYOMING STATE
Domain B:
Domain B: Area: 210x141
Area: 210x141 km 2 km2
Counties: Counties:
• Sweetwater • Sweetwater
# Radars
• Short : 21
+
LEGEND • Long : 01
2km<H<3km
1km<H<2km
H<1km
Domain B
+ +
+ +
+ + +
+ + +
+ +
+ + + + +
WYOMING STATE +
+ + +
Domain B:
Area: 210x141 km 2
Counties:
• Sweetwater
Coverage Percentage vs. Altitude
for Wyoming Domain B(210km x 141km)
LEGEND Long-range radars: 1 Short-range radars: 21
100
2km<H<3km
1km<H<2km
90
H<1km
80
Coverage Percentage (%)
Altitude Coverage (%) Coverage (%)
70 (m) Long Short
Short (Smooth) (1 radar) (21 radars)
60 Long (Smooth) 100 0% 5%
Short (Terrain) 500 14% 80%
50 1000 49% 98%
WYOMING STATE Long (Terrain)
LEGEND 2000 93% 99%
Domain B: 40
2km<H<3km Area: 210x141 km2 3000 100% 99%
Counties: 30
1km<H<2km • Sweetwater
H<1km 20
10
0
0 0.5 1 1.5 2 2.5 3
Altitude (km)
13. Radar Coverage Washington
WASHINGTON
Domain C:
Area: ~360x160
km2
Counties:
• Lewis
• Pacific
Domain C • Thurston
• Pierce
• King
+ + + + • Kitsap
+ + + • Mason
~60 nmi + + +
+ + • Grays Harbor
+ + + +
+ +
+ + + + # Radars
• Short : 27 +
WASHINGTON STATE + + +
+ + • Long : 01
LEGEND
Domain C:
Area: ~360x160 km 2 2km<H<3km
Counties: 1km<H<2km
• Lewis, Pacific, Thurston, Pierce H<1km
• King, Kitsap, Mason, Grays Harbor
Coverage Percentage vs. Altitude
for Washington Domain C(360 km x 160 km)
Long-range radars: 1 Short-range radars: 27
100
90
80 Altitude Coverage (%) Coverage (%)
Coverage Percentage (%)
ATE 70 (m) Long Short
LEGEND (1 radar) (27 radars)
WASHINGTON STATE
60 100 1% 3%
2km<H<3km
Domain C: 500 15% 45%
Area: 1km<H<2km 2
~360x160 km 50
ston, Pierce Short (Smooth) 1000 37% 67%
Counties: H<1km 2000 74% 72%
n, Grays Harbor 40 Long (Smooth)
• Lewis, Pacific, Thurston, Pierce
• King, Kitsap, Mason, Grays Harbor
Short (Terrain) 3000 94% 73%
30 Long (Terrain)
20
LEGEND
10
2km<H<3km
0
1km<H<2km 0 0.5 1 1.5 2 2.5 3
Altitude (km)
H<1km
14. Cost Considerations
• For long-range radar, site selection, land
preparation, site access, tower, communication
and electrical power, represent a complicated
process that in most cases takes from one to 3
years.
• An estimate for the up-front cost for a long-
range S-band radar (including the costs to buy
and install the radar as well as the land and
supporting infrastructure) is $10M. This figure
was cited in 2008 in a National Research Photo of a state-of-the-art phased array radar system.
Council report (NRC 2008) based on a Lincoln Phased arrays cost $1M/m2 today.
Laboratory estimate. This estimate is derived
based on the fact that the 156 radar WSR-
E-SCAN PANELS
88D/NEXRAD radar network cost $1.56 billion
to deploy between 1990 and 1997, or ~ $10M
per radar.
• For short-range radar, CASA project has set an
aim-point of $200k as the up-front cost of each
radar in a dense radar network and per-site
recurring cost of $20k per radar per year. Artistic concept of CASA radar panels attached to a
cellular telephone tower and the sides of a building.
15. Summary
• Additional weather radar, strategically placed along and near critical weather-sensitive
industries and infrastructure in radar gap regions, may improve public safety and reduce
weather-imposed economic loss. Beyond simply filling gaps in existing coverage,
additional radar capabilities such as rapid scanning, higher spatial resolution and multi-
Doppler coverage and enhanced radar products such as provided by dual-polarization
have the potential to significantly improve current observing and predicting capabilities.
• Forty-two small X-band radars (21 radars in NE Wyoming and 21 radars in SW Wyoming)
or two large S-band radars (one in each region), could provide this coverage. The 42
short-range (X-band) radars, deployed strategically along critical infrastructure, would
provide extensive multi-Doppler coverage of wind and rain at low-levels (below 2 km
AGL), thereby enabling real-time monitoring and improved prediction of warm-season
severe thunderstorms and low-level winter weather. Two long-range weather radars
could provide equivalent coverage at and above 2 km AGL.
• A network of 27 short-range radars, deployed along the coast, would provide multi-
Doppler coverage as low as 1 km AGL along the coast and up to a distance of 40 km from
shore. Over terrain, high resolution radar observations would help quantify locally intense
terrain-forced precipitation improving QPE and identifying low-level wind hazards.
16. References
Jerry Brotzge, Robb Contreras, Brenda Phillips and Keith Brewsterl, “Radar Feasibility Study,” January 31, 2009.
D. J McLaughlin, et al, “Distributed Collaborative Adaptive Sensing (DCAS) for Improved Detection, Understanding and Predicting
of Atmosphere Hazards,” in Proc of 85th AMS Annual Meeting 2005, San Diego, CA.
Brotzge, J., K. Brewster, V. Chandrasekar, B. Philips, S. Hill, K. Hondl, B. Johnson, E. Lyons, D. McLaughlin, and D. Westbrook, 2007:
“CASA IP1: Network operations and initial data”. Preprints, 23rd International Conf. on Interactive Information Processing
Systems (IIPS) for Meteor., Ocean., and Hydrology, AMS Conf., San Antonio, TX.
Gorgucci, E., and V. Chandrasekar, 2005: “Evaluation of attenuation correction methodology for dual-polarization radars:
Application to X-band systems”, J. Atmos. Oceanic Technol., 22, 1195-1206.
Brewster, K., E. Fay and F. Junyent, 2005: How will X-band attenuation affect tornado detection in the CASA IP1 radar network?,
32nd Conference on Radar Meteorology, Albuquerque, NM, AMS, Boston. Conference CD, Paper 14R.4.
Brotzge, J., D. Andra, K. Hondl, and L. Lemon, 2008: “A case study evaluating Distributed, Collaborative, Adaptive Scanning:
Analysis of the May 8th, 2007, minisupercell event”. Preprints, Symposium on Recent Developments in Atmospheric
Applications of Radar and Lidar, AMS Conf., New Orleans, LA.
Cheong, B. L., K. Hardwick, J. Fritz, P. S. Tsai, R. Palmer, V. Chandrasekar, S. Frasier, J. George, D. Brunkow, B. Bowie, P. Kennedy,
2007: “Refractivity retrieval using the CASA X-band radars.” Preprints, Proceedings of AMS 33rd Conference on Radar
Meteorology, Cairns, Australia.
Leone, D., R. Endlich, J. Petričeks, R. Collis, and J. Porter, 1989: “Meteorological considerations used in planning the NEXRAD
network.” Bull. Amer. Meteor. Soc., 70, 4–13.
18. 500 current radar units reach end of life 2012 – 2025
Red: sustain, replace legacy radars; Blue: large MPAR solution (~ same coverage)
10000 Small CASA Radars - Acquisition:
9000 10,000 radars x $100k per radar = $1B 2012 – 2016
8000 O&M will be a key factor – to be modeled
7000
Legacy O&M Cost
6000 Large MPAR Acquisition: Legacy Replacement Cost
Millions
Total Legacy Cost
5000 320 radars x $10M per radar = $3.2B 2012 - 2016 Legacy and MPAR O&M Cost
MPAR Acquisition Cost
4000
Total Cost with MPAR
3000
2000
1000
0
11 13 15 17 19 21 23 25 27 29
20 20 20 20 20 20 20 20 20 20
Year
1/13/2012 1:45:46 PM HiLo1 ECE697V Fall 2008 18