93.7 [w/m2/um/str ] ocean - jaxa · this presentation summarizes these comparisons of gosat and...

1
GOSAT has 3-days revisit cycle and 44 repeat orbits. The local equator crossing time in a descending node is 13:00+/- 15 minutes. OCO2 join the international Afternoon Constellation, or A-Train, of Earth-observing satellites. A-train has 16-days revisit cycle and 227 repeat orbits. The local equator crossing time in a ascending node is about 13:30. TANSO-FTS onboard GOSAT and the grating spectrometer on OCO-2 use different measurement techniques to measure molecular carbon dioxide (CO2) and molecular oxygen (O2). Both instruments observe sunlight reflected from the Earth’s surface in almost the same spectral range. As a first step in cross calibrating these two instruments, we compared spectral radiance observations within the three short wave infrared (SWIR) spectral bands centered on the O2 A-band (O2A), the weak CO2 band near 1.6 microns (Weak-CO2) and 2.06 microns (Strong-CO2) bands at temporally coincident and spatially collocated points. In this work, we reconciled the different size of the footprints and evaluated at various types of the surface targets such as ocean, desert and forest. For radiometric inter-comparisons, we consider long term instrument sensitivity degradation in orbit and differences in viewing geometry and associated differences in surface bidirectional reflectance distribution function (BRDF). Measured spectral radiances agree very well within 5% for all bands. This presentation summarizes these comparisons of GOSAT and OCO-2 spectral radiance observations and associated estimates of carbon dioxide and related parameters retrieved with the same algorithm at matchup points. We will also discuss instrument related uncertainties from various target observations A41F-0110: GOSAT and OCO-2 Inter-comparison on Measured Spectral Radiance and Retrieved Carbon Dioxide Fumie Kataoka (1) , Akihiko Kuze (2) , Kei Shiomi (2) ,Hiroshi Suto (2) , David Crisp (3) , Carol Bruegge (3) and Florian Schwandner (3) (1)Remote Sensing Technology Center of Japan, (2) Japan Aerospace Exploration Agency (3) NASA Jet Propulsion Laboratory Greenhouse gases Greenhouse gases Observing SATellite Observing SATellite © GOSAT GOSAT Greenhouse gases Greenhouse gases Observing SATellite Observing SATellite © GOSAT GOSAT Introduction Level2 : Carbon Dioxide Match - up criteria [Product version] ACOS FTS L2 B7.3 Lite (no-gainM data) OCO2_L2_ OCO2_L2_standard v7r [Data period] 2014/092016/02 ; Level2 matchup : 715 points Matchuped XCO2 data of GOSAT/ACOS and OCO2 are generally in good agreement with no obvious change in time. ΔXCO2 has no strong correlation with observation geometry, AOD and topography (Psurf). But some area (especially, central Asia and South of Africa) shows that GOSAT/CO2 is slightly higher than OCO2. This area corresponds to negative ΔPsurf (ACOS-OCO2) and high band1 albedo area. The negative ΔPsurf is thought to be related to high radiance level of GOSAT band1 spectra. We need to investigate the relationship between ΔPsurf and band1 radiance. Inter - comparison between GOSAT and OCO2 Level1 : Spectral Radiance Mauritania (Southwest area of Sahara desert) Gain : M GOSAT OCO2 GOSAT OCO2 (BRDF not applied) OCO2 (BRDF applied) band1 band2 band3 Center of GOSAT obs. point OCO2 obs. points within GOSAT-IFOV band1 band2 band3 BRDF plot [Product version] GOSAT FTS L2B V201 OCO2 L1B Science.7r [Data period] 2014/092016/09 ; Level1 matchup : 257 points Figure2. The map of GOSAT and OCO2 matchup from Sep2014 to Sep2016. The red plots shows the GOSAT gainM observation and bule plots are GOSAT gainH. Figure1. The schematic image of GOSAT and OCO2 match-up. . Figure3. The flow chart of GOSAT-OCO2 inter-comparison of spectral radiance. [Math-up criteria] Temporal : GOSAT OCO2 time difference : 1 hour Spatial : OCO2 footprint data within about 10.5 km-diameter GOSAT footprint XCO2 (ACOS-OCO2) error bar: XCO2 deviation within GOSAT-IFOV 6 XCO2 (ACOS-OCO2) AOD_total (OCO2) Albedo band1 (OCO2) Psurf (ACOS-OCO2) 8 Land Ocean (n=448) (n=267) Psurf vs XCO2 (ACOS-OCO2) ΔPsurf vs XCO2 (ACOS-OCO2) Psurf vs XCO2 (ACOS-OCO2) ΔPsurf vs XCO2 (ACOS-OCO2) (n=715) ACOS Albedo band1 vs XCO2 (ACOS-OCO2) OCO2_AOD_total (775nm) vsXCO2 (ACOS-OCO2) Land (n=448) Ocean (n=267) (n=715) Land & Ocean Land & Ocean Figure7. (Left) Time series plots of XCO2 (GOSAT OCO2) (Right) The scatter plots of XCO2 and related parameters. . Figure6. The map of GOSAT and OCO2 Level2 matchup parameter from Sep2014 to Feb2016. Figure4. Example of OCO2-GOSAT spectral comparison and BRDF plots. Figure5. (Upper) The scatter plot of GOSAT and GOSAT-OCO2 radiance. (Lower) The ratio of GOSAT/OCO2 spectra. Figure.5 shows the spectral comparison of OCO2 and GOSAT. This comparison includes not only clear sky data, but also homogeneous FOV of cloud and snow data. The ratio of GOSAT/OCO2 is within about +/- 4% except for band1. The ratio of band1 has wavelength dependency. Especially for shorter wavelength of band1, GOSAT spectra are quite higher than OCO2. This is thought to be causally related to the band1 non-linearity correction, but we need further investigation. MODIS BRDF (MCD43B1) Match-up temporally and spatially * Kuze et al, 2014 Select homogeneous site within FOV GOSAT (V201201) OCO2 (Science.7r) GOSAT Degradation Factor* w/ or w/o BRDF corr. GOSAT-OCO2 spectral comparison Wavenumber correction BRDF apply BRDF non apply Ocean Cloud Land Figure3 shows the flow chart of inter-comparison of spectral radiance. In this work, we applied GOSAT radiance degradation factor, described in Kuze et al (2014), wavenumber correction for GOSAT SWIR spectra and BRDF from the MODIS BRDF product (MCD43B1). 1.5929 [nm] 1.5968 [nm] 1.6056 [nm] 1.6161 [nm] 1.6195 [nm] 2.0492 [nm] 2.0528 [nm] 2.0618 [nm] 2.0723 [nm] 2.0795 [nm] gainM: n=16 gainH/Land : n=59 gainH/Ocean : n=89 band1 band2 band3 band1 band2 band3 gainM: n=94 gainH/Land : n=86 gainH/Ocean : n=77 gainM: n=94 gainH/Land : n=86 gainH/Ocean : n=77 Band1 saturation level 93.7 [W/m2/um/str] 0.7595 [nm] 0.7619 [nm] 0.7676 [nm] 0.7696 [nm] 0.7707 [nm] 1.5929 [nm] 1.5968 [nm] 1.6056 [nm] 1.6161 [nm] 1.6195 [nm] 2.0492 [nm] 2.0528 [nm] 2.0618 [nm] 2.0723 [nm] 2.0795 [nm]

Upload: others

Post on 11-Jun-2020

0 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: 93.7 [W/m2/um/str ] Ocean - JAXA · This presentation summarizes these comparisons of GOSAT and OCO-2 spectral radiance observations and associated estimates of carbon dioxide and

●GOSAT has 3-days revisit cycle and 44 repeat orbits.

The local equator crossing time in a descending node is

13:00+/- 15 minutes.

●OCO2 join the international Afternoon Constellation, or A-Train,

of Earth-observing satellites. A-train has 16-days revisit cycle

and 227 repeat orbits. The local equator crossing time in a

ascending node is about 13:30.

TANSO-FTS onboard GOSAT and the grating spectrometer on OCO-2 use different measurement techniques to measure molecular carbon dioxide (CO2) and molecular oxygen (O2). Both instruments observe sunlight reflected from the Earth’s surface in almost the same spectral range.As a first step in cross calibrating these two instruments, we compared spectral radiance observations within the three short wave infrared (SWIR) spectral bands centered on the O2 A-band (O2A), the weak CO2 band near 1.6 microns (Weak-CO2) and 2.06 microns (Strong-CO2) bands at temporally coincident and spatially collocated points. In this work, we reconciled the different size of the footprints and evaluated at various types of the surface targets such as ocean, desert and forest. For radiometric inter-comparisons, we consider long term instrument sensitivity degradation in orbit and differences in viewing geometry and associated differences in surface bidirectional reflectance distribution function (BRDF). Measured spectral radiances agree very well within 5% for all bands.

This presentation summarizes these comparisons of GOSAT and OCO-2 spectral radiance observations and associated estimates of carbon dioxide and related parameters retrieved with the same algorithm at matchup points. We will also discuss instrument related uncertainties from various target observations

A41F-0110:

GOSAT and OCO-2 Inter-comparison on Measured Spectral Radiance and Retrieved Carbon Dioxide

Fumie Kataoka(1)

, Akihiko Kuze(2)

, Kei Shiomi(2)

,Hiroshi Suto(2)

, David Crisp(3)

, Carol Bruegge(3)

and Florian Schwandner(3)

(1)Remote Sensing Technology Center of Japan, (2) Japan Aerospace Exploration Agency (3) NASA Jet Propulsion Laboratory

Greenhouse gases Greenhouse gases

Observing SATelliteObserving SATellite

©

GOSATGOSAT

Greenhouse gases Greenhouse gases

Observing SATelliteObserving SATellite

©

GOSATGOSAT

Introduction

Level2 : Carbon Dioxide

Match-up criteria

[Product version]

ACOS FTS L2 B7.3 Lite (no-gainM data)

OCO2_L2_ OCO2_L2_standard v7r

[Data period] 2014/09~2016/02 ; Level2 matchup : 715 points

Matchuped XCO2 data of GOSAT/ACOS and OCO2 are generally in good

agreement with no obvious change in time. ΔXCO2 has no strong correlation with

observation geometry, AOD and topography (Psurf). But some area (especially,

central Asia and South of Africa) shows that GOSAT/CO2 is slightly higher than

OCO2. This area corresponds to negative ΔPsurf (ACOS-OCO2) and high band1

albedo area. The negative ΔPsurf is thought to be related to high radiance level

of GOSAT band1 spectra. We need to investigate the relationship between

ΔPsurf and band1 radiance.

Inter-comparison between GOSAT and OCO2 Level1 : Spectral Radiance

Mauritania (Southwest area of Sahara desert)

Gain : M

GOSATOCO2

GOSATOCO2 (BRDF not applied)OCO2 (BRDF applied)

band1

band2

band3

Center of GOSAT obs. point

OCO2 obs. points within GOSAT-IFOV

band1 band2 band3

BRDF plot

[Product version]

GOSAT FTS L2B V201

OCO2 L1B Science.7r

[Data period] 2014/09~2016/09

; Level1 matchup : 257 points

Figure2. The map of GOSAT and OCO2 matchup from Sep2014 to Sep2016.

The red plots shows the GOSAT gainM observation and bule plots are GOSAT gainH.

Figure1. The schematic image of GOSAT and OCO2 match-up. .

Figure3. The flow chart of GOSAT-OCO2

inter-comparison of spectral radiance.

[Math-up criteria]

● Temporal : GOSAT –OCO2 time difference : 1 hour

● Spatial : OCO2 footprint data within about 10.5 km-diameter

GOSAT footprint⊿XCO2 (ACOS-OCO2)error bar: ⊿XCO2 deviation within GOSAT-IFOV

6

⊿XCO2 (ACOS-OCO2)

AOD_total(OCO2)

Albedo band1 (OCO2)

⊿Psurf(ACOS-OCO2)

8

Land

Ocean

(n=448)

(n=267)

Psurf vs⊿XCO2 (ACOS-OCO2) ΔPsurf vs ⊿XCO2 (ACOS-OCO2)

Psurf vs⊿XCO2 (ACOS-OCO2) ΔPsurf vs ⊿XCO2 (ACOS-OCO2)

(n=715)

ACOS Albedo band1 vs ⊿XCO2 (ACOS-OCO2)

OCO2_AOD_total (775nm)vs⊿XCO2 (ACOS-OCO2)

Land (n=448)

Ocean (n=267)

(n=715)Land & Ocean

Land & Ocean

Figure7. (Left) Time series plots of ⊿XCO2 (GOSAT – OCO2)

(Right) The scatter plots of ⊿XCO2 and related parameters.

.

Figure6. The map of GOSAT and OCO2 Level2 matchup parameter from Sep2014 to Feb2016.

Figure4. Example of OCO2-GOSAT

spectral comparison and BRDF plots.

Figure5. (Upper) The scatter plot of GOSAT and GOSAT-OCO2 radiance. (Lower) The ratio of GOSAT/OCO2 spectra.

Figure.5 shows the spectral comparison of

OCO2 and GOSAT. This comparison includes not

only clear sky data, but also homogeneous FOV

of cloud and snow data.

The ratio of GOSAT/OCO2 is within about +/- 4%

except for band1. The ratio of band1 has

wavelength dependency. Especially for shorter

wavelength of band1, GOSAT spectra are quite

higher than OCO2. This is thought to be causally

related to the band1 non-linearity correction, but

we need further investigation.

MODIS BRDF(MCD43B1)

Match-up temporally and spatially

* Kuze et al, 2014

Select homogeneous site within FOV

GOSAT

(V201201)

OCO2

(Science.7r)

GOSAT Degradation Factor*

w/ or w/o BRDF corr.

GOSAT-OCO2 spectral comparison

Wavenumber correctionBRDF apply BRDF non apply

Ocean

CloudLand

Figure3 shows the flow chart

of inter-comparison of spectral

radiance. In this work, we

applied GOSAT radiance

degradation factor, described

in Kuze et al (2014),

wavenumber correction for GOSAT SWIR

spectra and BRDF from the MODIS BRDF

product (MCD43B1).

1.5929 [nm]

1.5968 [nm]

1.6056 [nm]

1.6161 [nm]

1.6195 [nm]

2.0492 [nm]

2.0528 [nm]

2.0618 [nm]

2.0723 [nm]

2.0795 [nm]

gainM: n=16gainH/Land : n=59gainH/Ocean : n=89

band1 band2 band3

band1 band2 band3

gainM: n=94gainH/Land : n=86gainH/Ocean : n=77

gainM: n=94gainH/Land : n=86gainH/Ocean : n=77

Band1 saturation level93.7 [W/m2/um/str]

0.7595 [nm]0.7619 [nm]0.7676 [nm]0.7696 [nm]0.7707 [nm]

1.5929 [nm]1.5968 [nm]1.6056 [nm]1.6161 [nm]1.6195 [nm]

2.0492 [nm]2.0528 [nm]2.0618 [nm]2.0723 [nm]2.0795 [nm]