temporal interpolation using geostationary data: it’s ... · 25rd ceres science team meeting...
TRANSCRIPT
NASA Langley Research Center / Atmospheric Sciences
Temporal Interpolation UsingGeostationary Data: It’s About Time
D. Young, T. Wong, P. Minnis, and K. CostulisNASA Langley Research Center
J. Stassi, C. Nguyen, R. Raju, S. Sun-Mack,J. Boghosian, and E. Kizer
SAIC
P. Heck, J. KenyonAS&M
25rd CERES Science Team MeetingBrussels, Belgium
January 21-23, 2002
NASA Langley Research Center / Atmospheric Sciences
Time Interpolation and Spatial Averaging (TISA)Overview of Talk
• What is TISA ?
• TISA Status
• Monthly mean comparisons with ERBE-like
• Using geostationary data– Calibration of imager data
– Cloud property retrievals
• Results– Comparisons of monthly means w/ and w/o GGEO data
– Calibration sensitivity
– Exploring options
• Validation
NASA Langley Research Center / Atmospheric Sciences
Where TISA Fits Into CERES Processing
CERESInstrument
Data
Geolocate&
Calibrate1
ERBE-LikeInversion
2
ERBE-LikeAveraging
3
Clouds & TOA Flux
4
Grid TOA & SRB
9
MonthlyTOA & SRB
10
SARB5
GridSARB
6
TimeInterpolate
SARB7
MonthlyMeanSARB
10
GridGeoData11
ERBE-LikeProducts
TOA and Surface
Products
Synoptic & MonthlyTOA, Surface, and
Atmospheric Products
NASA Langley Research Center / Atmospheric Sciences
TISA Products
• Subsystem 10 (Data Product: SRBAVG)– Two Interpolation Methods (GGEO & non-GGEO)
– Uses Cloud Properties from GGEO Data
– Clear Sky Interpolation Using GGEO
– Total-sky Fluxes Derived Using ERBE-like ADMs
– Surface Fluxes
• SFC– Gridded SSF
– Used as input to producing SRBAVG
• GGEO– Gridded Geostationary imager data
– Used for temporal interpolation
NASA Langley Research Center / Atmospheric Sciences
Status of TOA and Surface Products
• QC reports and plots developed• Initial Validation of Flux and cloud algorithms
– Possible minor changes to clear-sky
• Transitional Format– Accommodate both ERBE and new DRMs– DRM construction and validation (# ADMS changed from 12
to 592)
• GGEO data– Calibration now automated
• Testing GGEO-radiance vs GGEO cloud albedomethods
• Validation in progress
NASA Langley Research Center / Atmospheric Sciences
ERBE-like Directional Models(Albedo as a Function of Solar Zenith Angle)
BB
BB
BB
BB
BBB
J
J
J
JJ
JJJJ
JJ
H
H
H
HH
HHHHHH
F
F
FF
FFFFFFF
0
10
20
30
40
50
60
70
0 30 60 90
Alb
edo
(per
cent
)
Solar Zenith Angle (degrees)
B Overcast
J Mostly Cloudy
H Partly Cloudy
F Clear
OCEAN
BB
BB
BB
BB
BBB
JJ
JJ
JJ
JJJJJ
H
HH
HH
HHHHHH
F
FF
FFFFFFFF
0
10
20
30
40
50
60
70
0 30 60 90
Alb
edo
(per
cent
)
Solar Zenith Angle (degrees)
B Overcast
J Mostly Cloudy
H Partly Cloudy
F Clear
LAND
NASA Langley Research Center / Atmospheric Sciences
Using Imager Data from GeostationarySatellites for Interpolation (GGEO)
• Define diurnal variations• Define cloud properties• Need calibration• LW
– Radiance - to - flux conversion– Narrowband-to-Broadband– Normalize to CERES
• SW– Radiance to flux conversion - need cloud properties– Narrowband-to-Broadband– Normalization– Second method by Haeffelin
NASA Langley Research Center / Atmospheric Sciences
CERES Time Interpolation IncorporatesGeostationary Data to Reduce Instantaneous Errors
Mean Instantaneous Interpolation Rms Errors Are Reduced By 50%For Both LW And SW TOA Flux Using Geostationary Data
B
B
B
B
BB
B
B
BB
B
B
B
B
BB
B
BBB
B
B
B
B
B
B
B
B
B
B
BB
BB
B
B
B
B
B
B
B
B
B
B
B
J
J
JJ
J
J
JJ
J
JJ
JJ
JJ
JJ
J
J
J
JJ
J
J
J
JJJJ
J
JJ
JJ
J
JJ
150
200
250
300
350
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
LW F
lux
(W/m
2)
Local Time (Day of Month)
B ERBS data
J NOAA-9 data
ERBE TSA
GOES 3-hourlyCERES TSA
NASA Langley Research Center / Atmospheric Sciences
Calibration Technique
• Calibration is tied to VIRS for TRMM– VIRS calibration appears to be stable for visible & IR
channels (Minnis et al. 2002)
– Using VIRS as standard provides consistency with CERESinstantaneous cloud properties
• Gridded geostationary imager data matched with
VIRS data in time and angle
• Separate correlations for each imager
• Separate correlations for land and desert
• Calibration trends compared with Minnis et al.
NASA Langley Research Center / Atmospheric Sciences
Recalibrated GOES-9 Visible Data
Slope = 1.51rms = 15.5
VIRS IR
Radiance (W/m2/sr/µm)
GOES-9 IR Radiance (W/m2/sr/µm)
NASA Langley Research Center / Atmospheric Sciences
METEOSAT Ocean IR
0
2
4
6
8
10
12
0 2 4 6 8 10 12
GGEO Radiance
VIR
S R
adia
nce
Feb-98Mar-98Apr-98May-98Jun-98
GGEO Calibration Stability
NASA Langley Research Center / Atmospheric Sciences
GGEO IR Calibration Stability
-5
0
5
10
15
20
Feb-98
Mar-98
Apr-98
May-98
Jun-98
Jul-98
Aug-98
Month
% D
iffer
ence
GMS-5GOES-9GOES-8METEOSAT
% Difference at 238 K
-8
-7
-6
-5
-4
-3
-2
-1
0
Feb-98
Mar-98
Apr-98
May-98
Jun-98
Jul-98 Aug-98
Month
% D
iffer
ence
% Difference at 295 K
NASA Langley Research Center / Atmospheric Sciences
GGEO Visible Calibration Stability
0
10
20
30
40
50
60
Feb-98
Mar-98
Apr-98 May-98
Jun-98 Jul-98 Aug-98
% D
iffer
ence GMS-5
GOES-9GOES-8METEOSAT
-10
0
10
20
30
40
50
60
Feb-98
Mar-98
Apr-98
May-98
Jun-98
Jul-98
Aug-98
% D
iffer
ence
% Difference at Low Radiance % Difference at High Radiance
NASA Langley Research Center / Atmospheric Sciences
GGEO Cloud Property Retrievals
• Needed for SW interpolation & for monthly clouds
• Uses IR/Vis LBTM retrievals (run as subset ofCERES cloud algorithm)
• Uses CERES surface property maps and MOAsoundings
• Properties– Cloud Amount
– Cloud Temperature
– Cloud Height (using standard 4 CERES layers)
– Optical Depth/Emittance (Daytime Only)
NASA Langley Research Center / Atmospheric Sciences
GGEO Cloud Amount 18 GMT February 6, 1998
NASA Langley Research Center / Atmospheric Sciences
GGEO Zonal Mean Cloud AmountFebruary 1998
0%
20%
40%
60%
80%
100%
-90 -60 -30 0 30 60 90
Latitude
Tot
al C
loud
Am
ount
(%
)
GGEO DayGGEO Night
NASA Langley Research Center / Atmospheric Sciences
GGEO and ISCCP Zonal Mean Cloud Amount GEO: February 1998 ISCCP: February 1986-1993
0%
20%
40%
60%
80%
100%
-90 -60 -30 0 30 60 90
Latitude
Tot
al C
loud
Am
ount
(%
)
GGEO DayGGEO NightISCCP Day+Night
NASA Langley Research Center / Atmospheric Sciences
Zonal Mean Cloud Amount Comparison GEO & VIRS: February 1998 ISCCP: February 1986-1993
0%
20%
40%
60%
80%
100%
-90 -60 -30 0 30 60 90
Latitude
Tot
al C
loud
Am
ount
(%
)
GGEO DayGGEO NightVIRS DayVIRS NightISCCP Day+Night
NASA Langley Research Center / Atmospheric Sciences
Instantaneous VIRS-GOES-9 ComparisonCloud Percentage
VIRS: 71.1% GOES-9: 71.7% Mean Difference: 0.6% RMS:14.1%
0
500
1000
1500
2000
2500
3000
-1 -0.6 -0.2 0.2 0.6 1
Bin
Freq
uenc
y
NASA Langley Research Center / Atmospheric Sciences
Instantaneous VIRS-GOES-9 ComparisonCloud Temperature and Optical Depth
VIRS: 268.9K GOES-9: 269.2K VIRS: 7.8 GOES-9: 6.0
NASA Langley Research Center / Atmospheric Sciences
Summary of GGEO-VIRS InstantaneousCloud Fraction Comparison
0.600.690.460.46GMS-5
0.520.520.430.53METEOSAT-6
0.670.590.470.47GOES-9
0.590.720.670.68GOES-9
VIRSGGEOVIRSGGEO
LandOcean
NASA Langley Research Center / Atmospheric Sciences
Summary of GGEO-VIRS InstantaneousCloud Temperature (K) Comparison
252.9256.1270.8274.6GMS-5
261.0265.5272.6273.2METEOSAT-5
259.7260.0270.2270.4GOES-8
258.1255.2268.9269.2GOES-9
VIRSGGEOVIRSGGEO
LandOcean
NASA Langley Research Center / Atmospheric Sciences
Summary of GGEO-VIRS InstantaneousCloud Optical Depth Comparison
8.67.18.56.7GMS-5
9.710.66.84.6METEOSAT-5
14.010.810.48.8GOES-8
10.013.77.86.0GOES-9
VIRSGGEOVIRSGGEO
LandOcean
NASA Langley Research Center / Atmospheric Sciences
Future GGEO Cloud Property Validation
• Look at all 8 months
• Viewing and Solar Zenith Dependence
• Deep Convective Cloud Albedos
• Finalize Calibration– Smooth trend
– Eliminate issues from poor sampling
• Intercomparisons– Overlapped GGEO Satellites
– ARM Satellite & Surface Observations
NASA Langley Research Center / Atmospheric Sciences
Flux Comparisons
• ERBE-like– 2.5° Grid– VZ < 70 °– Scene Identification by MLE– Old ERBE ADM
• SS10 CERES Monthly Means– 1.0° Grid– VZ < 45 °– Scene Identification from VIRS– New CERES ADM
• Comparison Method– Scatter Plots of SS10 Data Matched to ERBE-like 2.5° Data– Compare Histograms of Fluxes
NASA Langley Research Center / Atmospheric Sciences
Monthly Mean Total-sky LW Flux(non-GGEO Method)
NASA Langley Research Center / Atmospheric Sciences
Comparison of ERBE-like and SS10 non-GGEOTotal Sky LW Fluxes
Mean Difference = 0.6 W/m2
Sigma = 6.2 W/m2
NASA Langley Research Center / Atmospheric Sciences
ES4 ERBE-like and SRBAVG Flux Summary
9.018.220.7Sigma
-0.949.949.0MeanClear-Sky
SW Flux
5.814.016.6Sigma
1.2290.7291.9MeanClear-Sky
LW Flux
12.029.233.2Sigma
Mean
Sigma
Mean
-2.796.193.4Total-Sky
SW Flux
6.228.729.1
-0.6261.3260.7Total-Sky
LW Flux
ES4 -SRBAVG
SRBAVGnonGGEO
ERBE-like
(ES-4)
40°N - 40°S
W/m2
NASA Langley Research Center / Atmospheric Sciences
Monthly Mean Total-sky LW Flux(GGEO Method)
NASA Langley Research Center / Atmospheric Sciences
Monthly Mean Total-sky LW Flux Difference(non-GGEO Method - GGEO Method)
NASA Langley Research Center / Atmospheric Sciences
Histogram of ERBE-like and SS10 Total-SkyLW Fluxes
0
0.5
1
1.5
2
2.5
3
3.5
4
150 200 250 300 350
Total-sky LW Flux (W/m2)
Fre
quen
cy (
%)
ERBE-likeSS10 non-GGEOSS10 GGEO
NASA Langley Research Center / Atmospheric Sciences
Monthly Mean Clear-sky LW Flux Difference(non-GGEO Method - GGEO Method)
NASA Langley Research Center / Atmospheric Sciences
Histogram of ERBE-like and SS10 Clear-SkyLW Fluxes
0
1
2
3
4
5
6
7
8
150 200 250 300 350
Clear-sky LW Flux (W/m2)
Fre
quen
cy (
%)
ERBE-likeSS10 non-GGEOSS10 GGEO
NASA Langley Research Center / Atmospheric Sciences
nonGGEO - GGEO Total-Sky LW FluxDifferences (W/m2)
0
2
4
6
8
10
12
14
16
18
-25 -15 -5 5 15 25
nonGGEO - GGEO Total-Sky LW Flux Difference (W/m2)
Fre
quen
cy (
%)
NASA Langley Research Center / Atmospheric Sciences
Histogram of ERBE-like and SS10 Total-SkySW Fluxes
0
0.5
1
1.5
2
2.5
3
0 50 100 150 200 250
Total-sky SW Flux (W/m2)
Fre
quen
cy (
%)
ERBE-likeSS10 non-ERBESS10 GGEO
NASA Langley Research Center / Atmospheric Sciences
Histogram of ERBE-like and SS10 Clear-SkySW Fluxes
0
2
4
6
8
10
12
14
16
0 50 100 150 200 250
Clear-sky SW Flux (W/m2)
Fre
quen
cy (
%)
ERBE-likeSS10 non-ERBESS10 GGEO
NASA Langley Research Center / Atmospheric Sciences
SRBAVG ERBE and GGEO Clear-Sky SWFlux Comparison
Mean Difference = -0.9 W/m2
Sigma = 5.0 W/m2
NASA Langley Research Center / Atmospheric Sciences
SRBAVG nonGGEO and GGEO Flux Summary
5.020.020.7Sigma
-0.951.550.6MeanClear-Sky
SW Flux
5.015.716.7Sigma
-1.9284.7282.8MeanClear-Sky
LW Flux
9.431.433.3Sigma
Mean
Sigma
Mean
1.399.3100.6Total-Sky
SW Flux
4.528.829.1
0.2257.6257.8Total-Sky
LW Flux
nonGGEO- GGEO
SRBAVGGGEO
SRBAVGnonGGEO
40°N - 40°S
W/m2
NASA Langley Research Center / Atmospheric Sciences
Tropical Mean Comparison
Total-Sky SW
Clear-Sky SW
NASA Langley Research Center / Atmospheric Sciences
Calibration Sensitivity Test
• Test effect of imager calibration on monthly meanfluxes
• Test by varying imager gain by ±5%
• Affects both radiances and cloud retrievals
NASA Langley Research Center / Atmospheric Sciences
Effect of Calibration Change (IR - 5%)
NASA Langley Research Center / Atmospheric Sciences
Calibration Sensitivity Summary
Vis - 5%Vis + 5%IR - 5%IR + 5%
-0.22
(0.94)
0.21
(1.11)
0.05
(4.20)
0.01
(1.33)51.5
Clear-sky SW
-0.02
(0.26)
0.01
(0.27)
0.30
(0.92)
-0.29
(0.69)284.7
Clear-sky LW
-0.94
(1.31)
0.94
(1.31)
0.54
(3.10)
-0.04
(1.35)99.3
Total-skySW
0.00
(0.00)
0.00
(0.00)
-0.01
(0.08)
0.01
(0.08)257.6
Total-skyLW
Mean & rms Flux Difference (W/m2)MeanFlux
NASA Langley Research Center / Atmospheric Sciences
Using Cloud Albedo Models for Imager TimeSeries
• Directional Models (DRM) areconstructed empirically fromCERES data
• Use imager-derived cloudproperties to construct total albedoat interpolation times
• Models based on phase and opticaldepth
• Only use overcast DRM’s• Albedo time series normalized to
CERES observations
• More details provided in Co-I talkby Jeff Boghosian
Albedo
Solar Zenith Angle
NASA Langley Research Center / Atmospheric Sciences
Difference in Monthly Men SW Flux Using GGEO-radiance and GGEO-Cloud Albedo Methods
NASA Langley Research Center / Atmospheric Sciences
Calibration issues GGEO-Cloud AlbedoMethod
NASA Langley Research Center / Atmospheric Sciences
Imager Calibration Effect on Monthly MeanTotal-sky SW
Vis - 5%Vis + 5%IR - 5%IR + 5%
-2.47
(2.97)
2.55
(3.02)
4.13
(5.82)
-0.44
(1.14)98.6
Cloud
Albedo
Method
-0.94
(1.31)
0.94
(1.31)
0.54
(3.10)
-0.04
(1.35)99.3
Radiance
Method
Mean & rms Flux Difference (W/m2)MeanFlux
NASA Langley Research Center / Atmospheric Sciences
Comparison of SS10 and Surface Fluxes
• Use GGEO-interpolated 1°x1° Gridded Fluxes
• TOA Flux interpolated to all hours of Month
• Surface Fluxes Computed Using CERES TOA-Surface Algorithms– Downwelling Clear-sky SW
– Downwelling Total-Sky LW
– Downwelling Clear-Sky LW
• Match with 30-minute Averaged Data (Centered onLocal Half-Hour) From ARM Central Facility
• Compare Bias and rms of Interpolated Comparisonwith Instantaneous Results of Kratz et al.
NASA Langley Research Center / Atmospheric Sciences
Downwelling Total Sky LW Time SeriesARM SGP February 1998
NASA Langley Research Center / Atmospheric Sciences
Surface Downwelling Clear-sky SW Time SeriesARM Central Facility February 7, 1998
CERES Observation Times
NASA Langley Research Center / Atmospheric Sciences
Instantaneous & Interpolated DownwellingSurface Total-Sky SW Flux ARM Central Facility February 1998
CERES Footprint vs. Surface
Bias = -8.0 W/m2
RMS = 58.8 W/m2
Interpolated vs. Surface
Bias = -1.4 W/m2
RMS = 90.1 W/m2
NASA Langley Research Center / Atmospheric Sciences
Instantaneous & Interpolated DownwellingSurface Clear-Sky SW Flux (Model A)
ARM Central Facility February 1998
CERES Footprint vs. Surface
Bias = 27.4 W/m2
RMS = 46.1 W/m2
Interpolated vs. Surface
Bias = -32.5 W/m2
RMS = 55.9 W/m2
NASA Langley Research Center / Atmospheric Sciences
Instantaneous & Interpolated DownwellingSurface Total-Sky LW Flux (Model B)
ARM Central Facility February 1998
CERES Footprint vs. Surface
Bias = -0.4 W/m2
RMS = 19.5 W/m2
Interpolated vs. Surface
Bias = 1.8 W/m2
RMS = 28.2 W/m2
NASA Langley Research Center / Atmospheric Sciences
Interpolated Downwelling SurfaceClear-Sky LW Flux (Model B)
ARM Central Facility
July 1998
Bias = 0.5 W/m2
RMS = 12.7 W/m2
February 1998
Bias = -10.5 W/m2
RMS = 27.3 W/m2
Nighttime points
NASA Langley Research Center / Atmospheric Sciences
Future Plans
• Finalize Algorithms– Final DRM
– Fine-tune Clear-sky Algorithms
– Finalize Total-Sky SW Algorithm
• Continued Validation– Finalize Geostationary Calibrations
– Nine Months of TOA and Surface Flux Comparisons
– Study Cloud Normalization and Night Clouds
– Compare Monthly Mean and Aggregate Clouds
• Scheduled Archival in May 2002