Microphysical variability of tropical and mid

Report
Microphysical variability of tropical
and mid-latitude rainfall as revealed by
polarimetric radar
S. Rutledge1, D. Wolff2, B. Dolan1,
P. Kennedy1, W. Petersen2 and
V. Chandrasekar1
1
Colorado State University
2 NASA/Wallops Is.
2013 AGU Fall Meeting
San Francisco, CA
9-13 December 2013
Session H42A
• Reflectivity-based rain estimation central to TRMM and
GPM.
• We will investigate the polarimetric radar derived
“structure” of rainfall at several locations around the
globe. These structures reveal regimes where the melting
of graupel and hail contribute strongly to rainfall vs.
where coalescence dominates (tropical warm rain). These
structures reflect basic differences in drop size
distributions.
• These structures have implications for the A coefficient in
Z=ARb relationships, which is how rain rate is estimated
based on TRMM PR observations.
• We will conclude by discussing how A derived from
polarimetric radar compares to that from the PR
algorithm. This comparison will be done using both the
TRMM V6 and V7 datasets.
Rainfall microphysics seen
through combination of
polarimetric variables
Kdp proportional to mass content and
mass-weighted oblateness ratio
K dp = Cl
KDP is a measure of the
difference in wave propagation
between H and V polarizations;
sensitive to non-spherical
particles
-1
ò D (1- r)N(D)dD
3
K dp µ LWC · (1- r m )
K dp µ LWC · oblateness
Mass weighted…..
b
Rain
50
Z
Z µ ò D 6 N(D)dD
0.5 deg elevation
angle
40
30
Z » D 6 Reflectivity
Z dr = 10log10 (Z HH / ZVV )
Based on computations
of Z and Kdp from DSD
Oklahoma
assumptions
0
4
2
Kdp
deg/km
a
Differential Reflectivity
6
Difference in H,V phase in degrees
r = b/a
Application of polarimetric data……
Gorgucci et al. (2006, JTECH) showed that a parameter
space formed by Kdp / Z vs. Zdr was useful for
characterizing precipitation physics.
Dm, mm
Figure on the right shows results of scattering simulations
for various Gamma DSD’s with mean diameters (Dm)
ranging from 1.5 to 3.5 mm. Variations in Dm are
evident as well-defined curving paths in Kdp/Z vs. Zdr
space.
This technique can also be used to distinguish warm
rain-coalescence situations (high freezing level and
active drop coalescence processes, smaller drop
sizes) from rain derived from the melting of graupel
and hail (larger drop sizes), as distinguished by Kdp/
Z; Zdr pairs.
For a given rainfall regime, behavior of Kdp/Z vs. Zdr
represents precipitation physics.
Smaller
drops,
large liquid
water contents
Modest Z; high
Kdp
Large
drops from
melting ice
Large Z
An illustrative example; contrasting the FNL flood case with a nearby
bow echo storm…
NLDN lightning for 5
hour period
Heavy rain, little
lightning
FNL storm
BEC storm
10 inches of rain in a 5 hour period
Lightning with
bow echo storm
Ft. Collins flood example; tropical like
heavy rain event. Z=139R1.47
Z=139R1.47
Nearby strong convective storm; Z=300R1.4
High values
of KDP/Z
indicate large
water contents
with low Z; small
ZDR, small drops
Z=300R1.4
Larger ZDR
values indicating
melting ice
particles
Normalized density of points
expressed as a percentage
Yellow 70%
Red 50%
Blue 30%
All points > 30 dBZ used
Flood (tropical like)
event distinguished from
bow echo by reduced
Z and Zdr.
Kdp/Z shifted
to higher values for FNL (flood)
case. Implies large LWC
consisting of relatively small
drops.
Polarimetric variables consistent
with Z-R forms for these events
Small drops, high LWC,
small A (FLOOD)
Large drops, large A (BOW
ECHO)
Shift to the tropics…..
TRMM LBA, Jan-Feb 1999
NCAR S-pol radar
deployed for TRMM-LBA
Documented east-west regime with 7-10 day variability (Petersen and Rutledge, 2002)
24 Feb 1999
WEST case.
Lower CAPE,
monsoon-like
regime.
26 Jan 1999
EAST case.
Stronger
convection,
higher CAPE.
WEST
Subtle
differences
between
East and
West
EAST
WEST
EAST
LARGER ZDR VALUES
IN EAST CASE
COMPARED TO WEST.
INDICATIVE OF
LARGER DROPS
(MELTING ICE)
CONSISTENT WITH
HIGHER
CAPE/STRONGER
CONVECTION IN EAST
PHASE.
IN WEST PHASE,
LARGER Kdp/Z
INDICATING
SUBSTANTIAL LWC
CONTAINED BY SMALL
DROPS.
Following Bringi et al. 2004, J. Atmos. Ocean Technol., the A coefficient in Z=ARb is
calculated via the so-called pole-tune method where a Gamma distribution is
assumed
Z = [ a¢(m ) / N w ]R1.5
Here A is a function of Z, D0 and μ. Bringi et al. argue that A derived in this manner
continuously tracks the variability in DSD.
“A” coefficients derived from the TRMM-LBA dataset using the NCAR S-pol
radar. DSD’S ASSUME BROAD RANGE OF VALUES WITHIN EACH REGIME.
KDP/LIN Z
ZDR
N-Pol in MC3E sporting its new center-fed parabolic antenna and other upgrades
Midlatitude Continental Convective
Clouds Experiment
May-June 2011
Observations from
MC3E…..
25 April
2011:
Multiple
Convectiv
e Cores
DZ
Kdp
HID
Zdr
VR
Rhohv
Kdp
DZ
24 May
2011
Severe
storm
Zdr and Kdp column
VR
Zdr
- 70+ dBZ up to 10
km
- Large (+5 º/km)
Kdp at the surface
- Signature of large
hail (in RH and
ZDR)
- Strong tilted
updraft and
divergence aloft
- Data of high
quality at
significant ranges
Rhohv
HID
Warm rain
Normalized Self-Consistency for CHILL
5
70 %
50 %
30 %
4
Colorado events for comparison;
Warm rain vs. melting hail examples
Kdp*(105)/LinZh
29 July 1997
Ft. Collins Flood
Tropical-type Sounding
3
+
2
3 July 2010
Denver International Airport
Continental High Plains
Ice based
1
+
0
1
2
3
4
N-Pol in Oklahoma, MC3E
5
Zdr (dB)
70 %
50 %
30 %
N-POL
4
Kdp*(105)/LinZh
Results consistent with
active coalescence growth
and ice-based precipitation.
Normalized Self-Consistency for N-Pol 25 APRIL 2011
5
70 %
50 %
30 %
4
3
2
Normalized Self-Consistency for N-Pol 24 May 2011
5
25 April 2011 0800-1000 UTC
MC3E Oklahoma Location
Continental High Plains
1
Resulting from higher
moisture contents, higher
freezing level, collisional
breakup, etc.
N-POL
Kdp*(105)/LinZh
0
3
2
24 May 2011 0045-0300 UTC
MC3E Oklahoma Location
Continental High Plains
1
+
+
0
0
0
1
2
3
Zdr (dB)
4
5
0
1
2
3
Zdr (dB)
4
5
Rainfall: NASA GPM GV Measurement Infrastructure for IFloods
NASA IFLOODS DEPLOYMENT
INSTRUMENTATION
Radars: Rain mapping, 4-D
precip structure, DSD, rates
• NPOL S-band transportable,
scanning dual-pol radar
• D3R radar: Dual-frequency (KAKU), dual-polarimetric, Doppler
radar.
• 3 Metek Micro Rain Radars (Kband), vertically pointing (one on
order)
Point-Network
Disdrometer/Gauges: Precip
character/reference
• 5 2D Video Disdrometers
• 16-20 Parsivel-2 Disdrometer
with MetOne 12” TB Rain Gauge
• 25 dual-gauge Met One TB rain
gauges with soil T/Q
Dual-Gauge
Net
Precipitation Video Imager
(PVI)
JW
May 2, 2013 – Cold, light rain [70 km]
TRMM-LBA
West regime
Note: Axis
change
May 26, 2013 – Convection
Very similar to
MC3E case
Squall line example from IFLOODS
Large drops from
melting graupel and
small hail
Trailing stratiform region
from same squall line
Smaller drops from
melting of aggregates
IMPLICATIONS FOR TRMM rain mapping
Normalized Self-Consistency for CHILL
5
70 %
50 %
30 %
4
Kdp*(105)/LinZh
29 July 1997
Ft. Collins Flood
Tropical-type Sounding
The precipitation physics revealed by these
polarimetric data have a direct bearing on Z-R
based rain estimation and Z based attenuation
correction
3
+
2
3 July 2010
Denver International Airport
Continental High Plains
Shift to upper left implies smaller “A”
coefficient in Z=ARb and more rain for a given Z.
1
+
0
0
1
2
3
Zdr (dB)
Towards small
A values
4
5
Shift to lower right implies larger “A”
coefficient in Z=ARb and less rain for a given Z.
Have seen clear examples of these distinctive
shifts…….which are due to microphysical
variations in the production of rainfall
Towards large
A values
Kind of the crux of the
matter…..
Comparison of A coefficient
in Z=ARb between TRMM 2A25
and those derived from S-pol
polarimetric radar
Rain physics variability was
not well captured in Version 6,
reflected by the restricted range of A
TRMM V6
underpredicted
intense rain
LBA SPOL-TRMM V7 Coefficient Conv A in Z = A*R B
100.00
10.00
For both regimes TRMM derived A has a double
peak, coalescence and melting ice.
Consistent with the range of polarimetric radar
derived A values.
1.00
0.10
LBA: Convective RR TRMM V7
0.01
0
100
200
300
Coefficient A Value
SPOL Resolution = 4km
400
100.000
500
Reasons for the improvement
Introduced NUBF correction,
increase in high rain rates
Addition of 0.5 dB to PIA;
increase in heavy rain over
land.
Changes to α in the k=αZeβ
specific attenuation
calculation
SPOL-EAST
SPOL-WEST
TRMM-EAST
TRMM-WEST
10.000
Relative Frequency of Occurrence (%)
Relative Occurrence (%)
Rainfall physics much better captured in TRMM
V7!
SPOL-EAST
SPOL-WEST
TRMM-EAST
TRMM-WEST
1.000
0.100
0.010
0.001
20
40
60
80
Rain Rate (mm hr -1)
4 km sampling
100
120
140

similar documents