Satellite radiance bias monitoring
Roger Saunders, Brett Candy, Pete Francis
(Met Office) and Tim Hewison (EUMETSAT)
Rationale: It is crucial for climate and very desirable for NWP that we understand the characteristics of satellite radiance biases. Simultaneous Nadir Overpasses (SNOs)
have a limited time and space sampling. By comparing measured radiances with those simulated from a NWP model more can be learnt about the nature of the biases, as the
space and time matchup window can be relaxed using double differences with the NWP model as a transfer standard. In addition the dependence of the biases on scan
angle, scene temperature, time of day can also be investigated. The Met Office global Unified Model 6 hour forecast fields are used for this study. The plots below show some
examples from the first year of data collection in 2009.
2. Sensor data used
1. Methodology
Measured radiances from a variety of sensors (see panel to right) are
compared with radiances computed from the model 6 hour forecast
profile x at the same location using the RTTOV fast radiative transfer
model H(x). The differences DBT(n) and DBT(m) for two sensors can
then be compared (double difference). If the frequency of the channels
are similar it is a reasonable assumption the biases will be the same
from the simulated radiances and so any variations in bias between
instruments are due to calibration problems.
For channel n
Sensor 1: DBT(n) = mean{yn - Hn(xi)}
i=1 to k obs
Monitor DBT(n) – DBT(m)
For channel m
Sensor 2: DBT(m) = mean{ym - Hm(xi)}
i=1 to l obs
7.00
3. Global mean statistics
observed-background for 2009
HIRS global stats over ocean
HIRS and SEVIRI
4.00
NOAA19
NOAA17
METOP
SEVIRI
6.00
Bias
Instrument
Channels
Source
Comments
Meteosat-8
SEVIRI
8 IR
EUMETCAST
Only when Met-9 is
unavailable
Meteosat-9
SEVIRI
8 IR
EUMETCAST
Polar-orbit satellite
Instrument
Channels
Source
METOP-A
IASI
35 IR
EUMETCAST
HIRS
19 IR
EUMETCAST
Aqua
AIRS
35 IR
NESDIS
Occasional outages
NOAA-17
HIRS
19 IR
NESDIS
Gap after mid Dec 2009
NOAA-19
HIRS
19 IR
NESDIS
From August 2009
HIRS global stats over ocean
Comments
IASI global stats over ocean
NOAA19
NOAA17
METOP
SEVIRI
3.50
5.00
Geostationary satellite
IASI
5.00
Bias
sdev
4.00
St. Dev
3.00
3.00
4.00
2.50
2.00
2.00
degK
degK
degK
3.00
2.00
1.00
1.50
1.00
0.00
1.00
0.00
-1.00
-1.00
0.50
-2.00
0.00
Channel
HIRS
channel number
Frequency (waveno.)
Frequency cm-1
Channel
HIRS channel
number
4. Observed-background and double difference plots
SEVIRIClear_MET9 Ch 5 & ATOVS_METOPA Ch 12
SEVIRI - HIRS 6.7um channel
6. Geographical distribution of 13.4µm biases
2647
19
2501
18
2394
17
2390
16
2385
15
2295
14
2261
13
2232
12
1927
11
1631
9
1558
8
1530
7
1525
5
1480
4
1402
3
1330
2
1206
1
1150
19
1143
18
1040
17
928
16
925
15
834
14
781
13
759
12
753
11
740
9
731
8
715
7
706
5
705
4
698
3
690
2
680
1
668
-2.00
5. Scan biases
SEVIRIClear_MET9 Ch 11 & ATOVS_METOPA Ch 7
SEVIRI - HIRS 13.4um channel
7. Main conclusions from this preliminary analysis are:
• Changes in NWP model affect the sensor biases but the double difference is unaffected.
• The radiance drift seen in SNOs of the SEVIRI 13.4µm channel is confirmed in this study.
• All other SEVIRI channels appear to be stable over the period of this study (2009).
• Comparisons of SEVIRI and HIRS channels are more consistent than for SEVIRI and single channels
of IASI/AIRS. It is important to convolve the IASI/AIRS channels over the SEVIRI spectral response.
• Biases of the 13.4µm channels on SEVIRI, IASI and HIRS are all different from each other.
• The 6.2µm water vapour channels have similar biases due to uncertainty in the model water vapour.
• The sensor bias becomes more positive with increasing scene temperature for all instruments.
• The biases of all 3 HIRS sensors on METOP, NOAA-17 and NOAA-19 are similar in behaviour.
• The change in bias across a scan is small for IASI/AIRS but significant for HIRS in some channels.
• For SEVIRI there is a peak in the negative bias around incidence angles of 45˚ for some channels.
Met Office FitzRoy Road, Exeter, Devon, EX1 3PB United Kingdom
Tel: 01392 886295 Fax: 01392 885681
Email: roger.saunders@metoffice.gov.uk
© Crown copyright 10/0125 Met Office and the Met Office logo are registered trademarks