WMO/CIMO/TECO-2016, Madrid, Spain, 27-29 September 2016 SESSION 2– Developments in observing technologies and system

JMA’s C-band dual-polarization Doppler weather radars with SSPAs

Naoki Tsukamoto1, Hiroshi Yamauchi1, Hidehiro Okumura2, Akihito Umehara1 and

Yusuke Kajiwara1

1 Japan Meteorological Agency (JMA), Tokyo, Japan

2 Narita Aviation Weather Service Center, JMA, Chiba, Japan

Abstract

In March 2016, the Japan Meteorological Agency (JMA) began to operate C-band

dual-polarization Doppler weather radars with solid-state power amplifiers (SSPAs) at

Kansai International Airport and Tokyo International Airport. The new radars feature a

dual-polarization function and SSPA pulse compression.

Observational data improvement is a major objective associated with the WMO

Integrated Global Observing System (WIGOS). The dual-polarization function is

expected to help improve quantitative precipitation estimation (QPE) and quality control

(QC), as well as to support the provision of new products that allow identification of

precipitation as rain, snow or hail. JMA is currently evaluating the accuracy of the new

radar observation based on comparison with surface observation data.

SSPA usage and the adoption of pulse compression reduce the radar’s peak power

from the 200 kW of conventional klystrons to just 10 kW. The set-up also supports

narrow bandwidth and improved signal detection capability, thereby contributing to the

mitigation of interference with 5 GHz-band radio LAN (RLAN).

The significant advantages of dual-polarization Doppler weather radar with SSPA will

be highlighted via analysis of observational data during the 2016 rainy season.

1. Introduction

In Japan, weather radar data are widely used for a variety of purposes including

weather monitoring/prediction, water resource management and aviation safety. Such

information therefore plays a pivotal role in protecting people from severe storms.

The Japan Meteorological Agency (JMA) deploys nine Doppler Radars for Airport

Weather (DRAWs) to support aviation safety. The first DRAW was installed at Kansai

International Airport in 1994, and has remained operational since then.

In 2016, the two DRAWs at Kansai International Airport and Tokyo International

Airport (Haneda) were replaced with new SP-DRAW (solid-state polarimetric DRAW)

types featuring SSPAs and dual polarization. Another SP-DRAW is scheduled for

installation at Narita International Airport in 2016.

The adoption of solid-state transmitters in DRAWs is based on the serious threat from

interference caused by telecommunication devices (e.g., RLAN) to weather radars

worldwide (Elena et al. 2015). The shortage of 5-GHz frequency band (C-band) space

allocated to weather radars and RLAN is also a problem in Japan.

To solve this problem, Japan’s radio regulatory authority is taking measures to

separate frequency bands allocated to weather radar and RLAN services. The resulting

radio frequency reallocation plan for weather radars in Japan will be based on nine

channels with increments of 5 MHz within narrow band-width areas, as opposed to the

conventional allocation plan based on channels with increments of 10 MHz. Accordingly,

there is a need to reduce the bandwidth, unwanted emissions and peak power of

weather radars.

During the period of transition from the conventional channel allocation plan to the

new reallocation plan, radio frequency channels may need to be temporarily changed

for some weather radars to prevent interference with other radars. However, changing

the transmission frequency of conventional radars requires very expensive waveguide

filters that allow high-power handling and very-narrow-band operation. In contrast, the

operation radio frequency of SP-DRAW can be readily changed because no such filter

is required.

To reduce operational costs, SP-DRAW also allows savings in areas such as

consumable parts (through factors such as elimination of the need to use magnetron or

klystron) and requires less space and electric power.

2. Characteristics of SP-DRAW

2.1 Purposes of SP-DRAW

Figure 1 shows the Haneda Airport SP-DRAW, whose main roles are to detect

low-level wind shear (LLWS, including microbursts (MB) and horizontal shear lines (SL))

at/around the airport and to alert pilots if threshold levels are exceeded (Fig. 2). Clutter

signals can reduce the quality of radar echo and consequently affect the reliability of

LLWS detection.

The SP-DRAW system can be used to identify weather and non-weather echoes using

dual-polarization parameters, while conventional Doppler weather radar data may be

affected by unintentional removal of weather echo because a high QC threshold is

adopted to reduce clutter contamination. The introduction of SP-DRAWs has enabled

the collection of high-quality Doppler velocity data without clutter contamination even in

weak echo regions, which improve accuracy for identification of areas of sudden wind

chan

aviat

nge. This ha

tion weathe

as led to bet

r radar usag

Fig

tter accurac

ge.

. 1 Haneda

Fig. 2 R

cy in LLWS

a Airport SP

Roles of SP

detection, w

P-DRAW to

P-DRAW

which is a m

ower

major purpoose of

2.2 Specifications of SP-DRAW and related operating parameters

Table 1 shows the specifications and operating parameters of SP-DRAW compared

with those of conventional DRAW.

Table 1. Specifications/operating parameters of SP-DRAW and conventional

DRAW

SP-DRAW Conventional DRAW

Frequency 5,360 MHz 5,280 MHz

Transmitters GaN Power FET Klystron

Peak transmitter power 5 kW for both H and V 200 kW

Antenna diameter 7 m (parabolic) 7 m

Beam width < 0.7° < 0.7°

Antenna gain > 47 dBi > 47 dBi

Signal min. < -110 dBm < -110 dBm

Side lobe level < -27 dB < -27 dB

Range side lobe level < -55 dB

Range gate spacing 150 m 150 m

Radome diameter 11 m 11 m

Antenna rotation rate 4.2 rpm (EL = < 9.2°) 4 rpm

7 rpm (EL > 9.2°)

Samples (hits) 18 – 29 32

Azimuth spacing 0.7° 0.7°

Pulse repetition frequency 1,040/832 Hz (EL = < 9.2°) 1,120/840 Hz

1,365/1,092 Hz (EL > 9.2°)

Pulse width 1 μs (range < 12 km) 1 μs

64 μs (range >= 12 km)

Transmission mode Dual (simultaneous) Single (horizontal)

Observation parameters Z, V, W, ZDR, ρHV, ΦDP Z, V, W

Transmission radio frequency can be set to any value between 5,330 and 5,370 MHz.

As high-power gallium nitride FETs are used for SSPAs, peak SP-DRAW transmitter

power can reach 5 kW for each (H/V) polarized wave. The bandwidth of transmitted

waves is quite narrow, and leakage to the adjacent channel is substantially suppressed.

Pulse widths can be set freely from 1 to 200 μs. JMA operationally uses values of 1

and 64 μs for short and long pulses, respectively. A pulse compression technique

invol

also

wave

obse

was

durin

Th

Th

detec

ZDR a

witho

evalu

can

obse

Du

rates

and h

Fo

of gr

clutte

ving the us

provides th

e transmiss

ervation mo

performed

ng the instal

he SP-DRAW

his 7-m ant

ction of LLW

and ρHV va

out radome

uation using

be set at

ervation time

ue to the tra

s differ betw

higher-eleva

or lower-elev

round clutte

er contamin

se of freque

he ability to

sion, which

de operatio

using a sta

llation perio

W parabolic

tenna allow

WS near the

alues, a cro

e; this has

g the metho

up to 7 rp

es.

ade-off betw

ween lower-

ation scans

vation scan

er removal

nation, the r

ency chirp

adjust the

is importa

on as repor

andard horn

od.

c antenna is

ws the gen

e surface. In

oss-polariza

been pro

od proposed

m for conf

ween the an

-elevation s

s (which are

ns, an anten

performanc

ate is set as

Fig. 3 S

is adopted

phase of d

ant for simu

rted by Hub

n antenna i

s shown in F

eration of

n addition,

ation isolatio

ven to exc

d by Frech e

iguration of

tenna rotat

scans (whic

e focused on

nna rotation

ce. For hig

s 7 rpm with

SP-DRAW a

in long-pu

drive signals

ultaneous t

bbert et al. (

n the far fie

Fig. 3.

sharp beam

to enable th

on value ex

ceed 34 d

et al. (2012

f optimal s

ion rate and

ch are more

n fast scann

n rate of 4.2

gher-elevatio

h high pulse

antenna

lse observa

s for horizo

transmissio

(2010). Rel

eld of SP-D

m patterns

he derivatio

xceeding 35

B with rad

). The ante

scan seque

d radar data

e easily affe

ning).

2 rpm is set

on scans w

e repetition

ation. SP-D

ontal and ve

n and rece

lated adjust

DRAW’s ant

appropriat

on of high-q

5 dB is sec

dome base

nna rotation

ncing to re

a quality, rot

ected by cl

in consider

with less gr

frequency.

DRAW

ertical

eption

tment

tenna

te for

uality

cured

ed on

n rate

educe

tation

utter)

ration

round

2.3 O

Th

minu

Dopp

meth

MB

(0.7°

The

plann

eleva

spare

is pe

struc

runw

prod

of sy

Th

Lo

the

Observatio

he major SP

ute and SL i

pler velocity

hod (Yamau

Bs are dete

°), while SLs

need to sc

ning scan

ation angle

e time for o

erformed to

ctures of pre

way. PPI sc

uces data u

ystem anom

hus, the five

ong pulse w

requisite s

n procedur

P-DRAW pr

identificatio

y data obta

uchi et al. 20

ected using

s are detec

can PPIs a

sequence

es and rang

observation

obtain RHI

ecipitation a

canning at

used for ca

malies, is con

e-minute sca

Fig. 4

width observ

sensitivity o

re

rocesses fo

n every five

ined using

006).

a plan pos

cted using P

at 0.7 and

for SP-DR

ge height

of precipita

I data, whic

and wind o

an elevatio

alibration of

nducted.

an sequenc

4 Five-minu

vation and p

of SP-DRA

or LLWS de

e minutes. T

the dual-PR

sition indica

PPIs at two

1.1° indivi

RAW. Mea

indicator (R

ation and wi

ch are espe

on the appro

on of 90° (

dual-polari

ce shown in

ute antenna

pulse-compr

AW, as the

tection invo

These forms

RF techniqu

tor (PPI) at

lower eleva

idually repr

nwhile, sca

RHI) opera

nd in and a

ecially usefu

oach to the

known as

zation para

Fig. 4 is im

a scan seq

ression tech

e peak tra

olve MB ide

s of detectio

ue and the

t the lowest

ation angle

resents a c

anning of

ation are co

around the a

ul for monito

e airport pa

bird-bath s

ameters and

mplemented

uence

hniques are

ansmitter p

entification e

on are base

hybrid mult

t elevation a

s (0.7 and

constraint i

PPIs at h

onducted d

airport. Sca

oring the ve

rallel to its

scanning), w

d early dete

for SP-DRA

e used to en

power for

every

ed on

ti-PRI

angle

1.1°).

n the

higher

during

nning

ertical

main

which

ection

AW.

nsure

each

polar

1-μs

corre

obse

short

1.5 t

and v

conv

As

marg

pract

mete

Fig.

Fig.

for c

rization is 5

pulse is

esponding

ervation, a

t-pulse obse

imes better

vertical ave

ventional rad

s the trunca

gin, a sensit

tical use b

eorological e

5 Relations

6 Relations

convention

5 kW (i.e., o

also use

to the tim

value of 6

ervation ran

r than that o

erage transm

dar, but its o

ated range f

tivity gap is

because sh

echoes nea

ship betwe

ship betwe

al DRAW (b

only 1/40th o

d to com

me of long

64 μs is a

nges (Fig. 6

of conventio

mitted powe

overall elec

for the long

observed a

hort pulses

ar the radar

een pulse w

een minimu

black line)

of those of

mplement o

g-pulse em

adopted for

6). As a res

onal DRAW

er from the n

ctricity consu

pulse widt

at 12 km. Ho

s (1 μs) h

site (Fig. 7)

width and m

um detecta

and SP-DR

convention

observation

mission (Fig

r better sen

ult, the ave

(1 μs, 200

new radar s

umption is 2

h (64 μs) is

owever, this

ave enoug

).

measurable

able reflecti

RAW (purp

al radars w

within tru

g. 5). In

nsitivity bo

rage power

kW). The o

system is th

20 percent l

s equal to 1

s does not p

gh power f

e area

ivity and ra

le line)

with a klystro

uncated ra

long-pulse-

th in long-

r of SP-DRA

overall horiz

hree times th

lower.

2 km with s

pose proble

for detectio

ange from r

on). A

anges

-width

- and

AW is

zontal

hat of

some

ms in

on of

radar

Fig.

at an

obse

2.4 Im

Th

Figur

and s

Th

inform

polar

synth

malfu

reduc

Tw

Thes

The

data

dual

Tw

As

syste

syste

and r

7 PPI plots

n elevation

ervation re

mproveme

he transmitte

re 9 gives

signal proce

he SP-DRAW

mation to f

rized waves

hesized fro

unction, ope

ced transm

wo receiver/

se devices

active syste

processors

system.

wo antenna

s SP-DRAW

em stability

em status in

receiver sen

s of (a) refle

n angle of

gion.

ent of radar

er, receiver,

an overview

essor.

W system is

flights. The

s and anoth

om the out

eration can

ission powe

/signal proc

are used to

em of the d

s used to co

controllers

W is premise

checking a

n daily app

nsitivity usin

ectivity and

f 1.1°. The

r system st

, signal proc

w of the sy

s designed w

transmitte

her eight fo

put of thes

be continu

er.

cessor sets

o generate

devices outp

onduct LLW

are operate

ed on conti

and quality

plication, SP

ng test sign

d (b) Doppl

blue dott

tability

cessor and

ystem includ

with high re

r consists o

or vertical p

se sixteen

ed on the b

s are opera

drive signa

puts observ

WS detectio

ed in the for

nuous syst

maintenanc

P-DRAW co

nals.

ler velocity

ted circle

antenna co

ding the an

edundancy t

of eight SS

polarized w

modules.

basis of rad

ated in the

als and rec

vation data

n, which ar

rm of a dup

tem operatio

ce in day-to

onstantly m

y for weak w

indicates t

ontroller are

ntenna, tran

to ensure st

SPA module

aves, and

In the eve

ar equation

form of a

ceive/proces

(the mome

re operated

lex system.

on, it includ

o-day opera

onitors tran

weather ec

the short-p

e shown in F

nsmitter, rec

table provis

es for horiz

transmits w

ent of a mo

n calculation

duplex sys

ss echo sig

nt data) into

in the form

.

des function

ation. To mo

nsmission p

choes

pulse

Fig. 8.

ceiver

ion of

zontal

waves

odule

n with

stem.

gnals.

o two

m of a

ns for

onitor

power

As m

enab

mentioned p

ble detection

Fig. 8 Tr

previously, S

n of system

ransmitter,

SP-DRAW m

m errors and

receiver, s

monitors ZD

misalignme

signal proc

DR and ΦDP

ents across

cessor and

using bird-b

s the whole

antenna c

bath scanni

system.

controller

ing to

2.5 Q

To

are e

algor

movi

data

used

no ef

used

also

proce

proce

Fig

dual-

align

Quality con

improve th

essential fo

rithm for wh

ing clutter s

and cause

d for conven

ffect in rela

d to efficient

results in th

essing data

essing data

gure 10 sh

-polarization

ment (CPA

F

ntrol metho

he quality o

or LLWS de

hich dual-po

such as sh

e false LLW

ntional Dopp

ation to mov

tly distingui

he removal

a should b

a (NOR) sho

hows a flo

n paramete

A; Hubbert e

Fig. 9 Block

ods

of Doppler v

tection, SP

olarization

ips and win

WS detectio

pler radars

ving clutter.

sh both GC

or attenuat

e applied o

ould be app

ow chart of

ers are use

et al. 2009)

k diagram o

velocity data

-DRAW inc

parameters

nd turbines

on. Although

help to mitig

Meanwhile

C and movin

ion of weat

only for bin

plied for pure

f the clutte

ed. This inv

and the tex

of SP-DRAW

a at the low

corporates a

s are utilize

impair the

h the movi

gate GC to

e, dual-pola

ng clutter. T

her echo. T

ns contami

e weather e

er discrimin

volves the

xture of the

W

west elevat

a new clutte

d. Ground

quality of

ng target in

a certain de

arization par

The use of M

To avoid this

nated by G

echo.

nation algo

utilization

differential

ion angle, w

er discrimin

clutter (GC

Doppler ve

ndicators (M

egree, they

rameters ca

MTIs some

s side effect

GC, and no

orithm for w

of clutter p

phase S(Φ

which

nation

) and

elocity

MTIs)

have

an be

times

t, MTI

ormal

which

phase

ΦDP) in

addit

to de

error

for p

proce

If th

thres

CPA(

assu

MTI

or m

MTI d

The

for o

3. SP

As

preci

comp

obse

evalu

the T

July

tion to the c

etermine w

r (RMSE) of

recipitation

essing data

he CSR, CP

sholds, the

(MTI) and S

med to be c

[flag 1]. Oth

oving clutte

data and in

threshold v

ptimization

P-DRAW ob

s part of pola

ipitation es

parison with

ervation site

uation for a

Tateno surfa

2016. Figur

conventiona

hether indiv

f fitting for Φ

echo and la

a and MTI p

PA(NOR) a

bin is assu

S(ΦDP)(MTI

contaminate

herwise, the

er whose ef

valid data a

values can b

to avoid sig

Fig.

bservation

arimetric pe

stimates u

h data from

es around e

heavy rainf

ace observ

re 11 shows

al clutter sup

vidual bins

ΦDP values o

arge for clu

rocessing d

and S(ΦDP)(

umed to be

) values fo

ed by weak

e bin is assu

ffects are d

are applied

be adjusted

gnificant effe

10 Clutter

data

erformance

sing SP-D

weighing ra

ach SP-DR

fall event th

ation site 6

s PPIs of re

ppression ra

include clu

of five bins

utter. CPA an

data.

NOR) value

e uncontam

r the bin ar

ground clu

umed to be

difficult to m

for bins flag

, and JMA

ects on wea

discrimina

verification

DRAW dua

ain gauges

RAW. This s

at produced

65 km north

eflectivity Z

ate (CSR) d

utter. S(ΦDP

in a radial r

nd S(ΦDP) a

es for a ce

minated by c

re below th

tter whose e

contaminat

mitigate with

gged with [0

is currently

ather echo.

ation algor

, the JMA is

al-polarizatio

(WGs) inst

ection deta

d hourly pre

heast of the

and the spe

derived from

P) is the ro

row. The pa

are calculat

ertain bin ar

clutter [flag

e set thres

effects are w

ted by stron

h MTI [flag

0], [1] and [2

evaluating

rithm

s evaluating

on parame

talled at five

ils the prelim

ecipitation o

e Haneda S

ecific differe

m MTI proce

ot mean sq

arameter is

ed for both

re below th

g 0]. If the

sholds, the

well mitigat

ng ground c

2]. Normal

2], respectiv

the values

g the accura

eters based

e special su

minary resu

of up to 57 m

SP-DRAW o

ential phase

essing

quare

small

NOR

he set

CSR,

bin is

ed by

clutter

data,

vely.

used

acy of

d on

urface

ults of

mm at

on 14

e Kdp

as o

(2012

(200

reflec

(a) a

respe

Fig

weig

cons

grou

desc

seen

cumu

R(W

Fig.

DRA

cros

Tate

bserved via

2), which in

1), the rain

ctivity Z_co

and (b) sh

ectively. Th

gure 13 sh

hing rain g

sideration of

nd, the tim

cent rate of

n, R(Z) rep

ulative valu

G), while th

11 PPI plo

AW with an

sses and w

no site, res

a the Hane

nvolves the

rate R(Kdp

orr. Convent

how distribu

e latter sho

hows a tim

gauge data

f the time ta

es of R(Z)

4 m/s. Wh

presents sig

e of R(Kdp

hat of R(Z) is

ots of (a) r

n elevation

white circle

spectively.

eda SP-DR

e use of co

p, Z_corr) i

tional rain ra

utions of e

ows heavy ra

me-series re

a R(WG), R

aken for ra

and R(Kdp

hile corresp

gnificant u

, Z_corr) fro

s 74% unde

reflectivity

n angle of

es indicate

RAW. In line

oefficients p

s estimated

ate estimat

estimated r

ain areas m

epresentati

R(Z) and R

indrops to f

p, Z_corr) a

pondence b

nderestima

om 16:57 to

erestimated

and (b) K

0.7º at 171

e the locat

e with the m

proposed by

d using Kdp

ion R(Z) is

rain rates

more clearly

on of rain

R(Kdp, Z_c

fall from a d

are delayed

etween R(K

ation of the

o 17:57 is 1

d.

Kdp as obs

10 JST on

ions of the

method of

y Bringi an

p and atten

also condu

R(Z) and

.

rates as

corr) at the

distance of

d by 4 minu

Kdp, Z_corr

e rain rate

1% undere

served usi

14 July 2

e Haneda

Yamauchi

d Chandras

uation-corre

cted. Figure

R(Kdp, Z_

observed

Tateno sit

1 km abov

utes assum

r) and R(W

. The one

stimated ag

ng the Ha

2016. The b

DRAW and

et al.

sekar

ected

es 12

_corr),

using

te. In

ve the

ming a

WG) is

e-hour

gainst

aneda

black

d the

Fig.

(b) R

Fig.

gaug

Z_co

4. Su

SP

anten

obse

In

assu

conti

12 As per

R(Kdp, Z_co

13 Time-s

ges (thin g

orr) (thick r

ummary an

P-DRAW gu

nna scann

ervational ca

response t

rance sys

inuous qua

Fig. 11, bu

orr)

series repr

green line)

red line)

nd future pl

uarantees a

ning. The

apacity for p

to growing

tem now

ality monito

ut with rain

resentation

and estim

lans

dequate rad

results of

practical op

demand fo

includes a

ring based

rate distri

n of rain ra

ated using

dar observa

the study

eration in v

or high-qua

a built-in

on bird-ba

ibution est

ates obser

g R(Z) (dot

ation data q

y confirme

very strong t

ality radar d

independen

ath scannin

timated usi

rved using

ted blue li

quality even

ed the exp

to very wea

data, the S

nt checking

ng to detec

ing (a) R(Z)

g weighing

ne) and R(

with high-s

pected lev

ak echo area

P-DRAW q

g function

ct issues d

) and

g rain

(Kdp,

speed

el of

as.

uality

and

during

operation.

High-quality data and frequent observation are essential for the early detection of

severe storms, which is one of JMA’s goals.

In relation to effective radio frequency usage, solid-state transmitters are very useful

because they support:

- achievement of balance between peak power reduction and high data quality;

- mitigation of unexpected emissions against adjacent channels; and

- simple frequency changes.

With this approach, electricity consumption is 20% less than with conventional radar.

Observational data improvement is a major objective associated with the WMO

Integrated Global Observing System (WIGOS). To support the achievement of this goal,

JMA will share lessons learned from its adoption of new-generation weather radars with

NMHSs.

References

Bringi, V. N. and V. Chandrasekar, 2001: Polarimetric Doppler weather radar: principles

and applications, Cambridge University Press, 636pp.

Hubbert, J., M. Dixon, S. Ellis, and G. Meymaris, 2009: Weather radar ground clutter.

Part I: Identification, modeling, and simulation. J. Atmos. Oceanic Technol., 26,

1165–1180.

Hubbert, J., S. Ellis, M. Dixon, and G. Meymaris, 2010: Modeling, error analysis, and

evaluation of dual-polarization variables obtained from simultaneous horizontal and

vertical polarization transmit radar. Part I: modeling and antenna errors. J. Atmos.

Oceanic Technol., 27, 1583–1598.

Frech, M., B. Lange, T. Mammen, J. Seltmann, C. Morehead and J. Rowan, 2013:

Influence of a radome on antenna performance. J. Atmos. Oceanic Technol., 30,

313–324.

Saltikoff, E., J. Cho, P. Tristant, A. Huuskonen, L. Allmon, R. Cook, E. Becker, and P.

Joe, 2015: The Threat to Weather Radars by Wireless Technology. Bull. Amer.

Meteor. Soc., doi:10.1175/BAMS-D-15-00048.1, in press

Yamauchi, H., O. Suzuki and K. Akaeda, 2006: A hybrid multi-PRI method to dealias

Doppler velocities. SOLA, 2, 92–95.

Yamauchi, H., A. Adachi, O. Suzuki and T. Kobayashi, 2012: Precipitation estimate of a

heavy rain event using a C-band solid-state polarimetric radar, Extended abstract,

7th European Conf. on Radar in Meteorol. and Hydrol., Toulouse, France, 201SP.