Improvement in the Sensitivity of Spaceborne Precipitation Radar from TRMM to GPM

1
1P-6
The 2nd TRMM International Science Conference
Improvement in the Sensitivity of Spaceborne
Precipitation Radar from TRMM to GPM
Shinsuke SATOH , Nobuhiro TAKAHASHI, Toshio
IGUCHI National Institute of Information and
Communications Technology (NICT), Koganei,
Japan Shuji SHIMIZU, Kinji FURUKAWA, Masahiro
KOJIMA Japan Aerospace Exploration Agency (JAXA),
Tsukuba, Japan   Minoru OKUMURA NEC TOSHIBA Space
Systems, Ltd., Yokohama, Japan
?2?Definition of Minimum Detectable Rainfall Rate
(1) Radar echo signal is measured as following
signal power total receiving power
noise power Std of the total receiving
power ?t 5.57/SQRT(N) 0.568 dB (N96)
Std of the noise power ?n
5.57/SQRT(M ) 0.208 dB (M720) where N
is the averaging sampling number, and M is the
noise sampling number The mean noise power
of KaPR in 500 m range resolution under the
condition of S/N1, Te323 K (50ºC), Ta290
K (17ºC), Emissivity 0.95, Satellite
altitude 419 km (max) is 16.4 dBZ (for). (2)
Considering the judgment on the presence of rain
using the 2-? threshold in a signal
distribution for no-rain, the mean total
receiving power is expressed by 16.4
dBZ 2 ? SQRT(?t 2 ?n2) 17.61 dBZ,
and, the mean signal power is presented by
101.761 101.64 14.0 (in real number) ?
11.5 dBZ (3) The minimum detectable rainfall
rate defined at the earth ellipsoid surface
level ignoring the rain attenuation is
calculated using a Z-R relationship (Z200R1.6)
for weak rainfall. 11.5 dBZ ?
0.19 mm/hr
Fading measured by a logarithmic detector
?7?Summary
?1?Introduction

• The mean system noise level in TRMM PR is very
  stable for long term, while it includes a problem
  of the range bin offset.
• In GPM DPR, because the variable PRF technique is
  adopted to increase the sample number, the system
  noise data is sampled during the transmitting
  tern-off period of the last five pulses in a
  beam.The std of signal power with fadingTRMM
  PR (N64, M256) 0.78 dBGPM DPR (N96, M720)
  0.60 dB
• The practical sensitivity may be higher on the
  actual mitigating conditions.- over ocean
  0.7 dB less than over land- over tropics (low
  satellite altitude) 0.4 dB less than
  high-latitude- Future works Drop Size
  Distribution (Z-R relation), Mie-scattering
  effect, etc.

• The sensitivity of spaceborne radar is an
  important factor for several research fields.
• Restriction of a long distance (gt 400 km)
  measurement and the limited resources (power,
  mass, size) of spacecraft.
• Investigate the factors to determine the minimum
  detectable rainfall rate for the TRMM PR, and
  improve the sensitivity of the GPM DPR.

KaPR in 500 m reso.
KuPR KaPR in 250 m reso.
What is the real top height of precipitation?
What extent of snowfall is measured?
Fig. 1 Probability distributions of signal power
for 0.2 mm/hr (left) and 0.5 mm/hr (right) with
reference signal distributions for different
rainfall rates. The horizontal axis shows
Z-factor.
?5?TRMM PR Tx turn-off Experiments
?3?Variations in System Noise (TRMM PR)

• The system noise level is determined by the
  thermal noise and the background noise from the
  radiation of the earth surface, precipitation,
  etc.
• The variation of the thermal noise is less than
  0.15 dB, and is stable for the long period (Fig.
  2). The variation of the background noise is also
  small (lt 0.1 dB over ocean, lt 0.5 dB over land)
• The fading variation of the system noise is about
  1 dB (Fig. 3).

1B21 special averaging 500 scans June 24, 2004
Over Ocean (off Peru)
Over Land (central Africa)
Over Land
Over Sea
0.7 dBup
- 0.1 dB / 100 range bin
Fig. 2 Long-term change of the system noise and
the solar beta angle (top), and the FCIF
temperature (bottom)
Fig. 3 An example of the system noise
distribution (no-rain, over ocean, 100 orbits
data)
- 0.1 dB / 100 range bin
Range Bin
Range Bin
The cause to appear the surface echo during
transmitting turn-off operations is that only the
seed signal in FCIF is turned-off, and a noise is
amplified by pulse driven SSPA and
LNA.Therefore, the system noise has to be
sampled by turned-off SSPA, LNA, FCIF in GPM DPR.
?4?Sampling Windows (TRMM PR GPM DPR)
?6?GPM DPR Design and Concept
Fig. 4 Sampling window for TRMM PR.
Fig. 5 Range bin offset and noise offset in 1B21.

• In TRMM PR, the system noise is sampled in the
  noise-sampling window to avoid the rain echoes,
  surface echoes, and sidelobe culler echoes (Fig.
  4).
• Since the receiver (LNA) gain changes with time
  (range bin offset) caused by the T/R switching,
  the noise data in each angle bin has to correct
  the offset (Fig. 5).
• In GPM DPR, there are no extra area to sample the
  system noise (Fig. 6), because the data sampling
  window is optimized by the variable PRF (VPRF)
  technique.

Variation in the altitude of GPM core satellite
The propagation loss (Lp) is expressed by Lp
-20?LOG(4 ? ? ? R) dB , R401 km Lp
-134.047 (0.381 dB) R419 km Lp -134.428
(0 dB).
Concept of the variable PRF technique.
Fig. 6 Sampling window for GPM DPR.