GOSAT CALIBRATION PLAN

Kei Shiomi(1), Shuji Kawakami(1), Tomoko Kina(1), Mayumi Yoshida(2), Nami Sekio(2), Yasushi Mitomi(2), and Fumie Kataoka(2)

(1) Japan Aerospace Exploration Agency, 2-1-1 Sengen, Tsukuba, Ibaraki 305-8505, Japan, e-mail: shiomi.kei@jaxa.jp

(2) Remote Sensing Technology Center of Japan, 1-6-1 Takezono, Tsukuba, Ibaraki 205-0032, Japan ABSTRACT

Greenhouse gases Observing SATellite (GOSAT) is a Japanese MOE/NIES/JAXA joint program to observe greenhouse gases, such as CO2 and CH4, from space. The GOSAT will be launched in 2008. The GOSAT carries a Fourier transform spectrometer and a push broom imager. The GOSAT development is going on in phase-C/D and characterized the sensor performance in laboratory. The post-launch calibration items are planned and the methods will be developed before the launch. In this study, post-launch calibration methods are investigated by analyzing the current MODIS data, which has similar wavelength bands to GOSAT. The calibration sites are selected on the basis of this study. 1. GOSAT SATELLITE AND SENSORS

Thermal And Near infrared Sensor for carbon Observation (TANSO) is carried on the GOSAT satellite for precise measurement of GHGs [1]. The TANSO is composed of 2 sensors, Fourier Transform Spectrometer (FTS) for greenhouse gases observing sensor, and Cloud and Aerosol Imager (CAI) for cloud detection and correction of optical thin cirrus and aerosol interference within the FTS instantaneous field of view (IFOV). Fig.1 shows the TANSO sensors on the GOSAT satellite.

The FTS observes at wide spectral range from shortwave infrared (SWIR) to thermal infrared (TIR) with the same optics. The incident lights are modulated by double cube corner mirrors. The modulated lights are separated into SWIR 3 bands with P/S polarization and TIR band by dichroic mirrors settled after optics. The FTS sensor characteristics are shown in Tab. 1. The SWIR bands are utilized for retrieval of CO2 and CH4 column density, on the other hand the TIR band for CO2 and CH4 concentration soundings. Both observations are efficient to improve the measurement accuracy of GHGs. Moreover, the FTS TIR band is utilized for retrieval of another trace gases, such as O3.

The CAI has continuous spatial coverage swath, and higher spatial resolution than the FTS in order to detect the aerosol spatial distribution and cloud coverage. The CAI sensor characteristics are shown in Tab. 2. Using the multi-spectral bands from UV to SWIR, the spectral characteristics of the aerosol scattering can be retrieved together with optical thickness.

Figure 1. GOSAT satellite and TANSO sensors

Figure 2. TANSO-FTS spectral coverage

Table 1. TANSO-FTS characteristics

Band Band1 Band2 Band3 Band4 Polarization P, S P, S P, S - Wavelength 0.75-0.78

µm 1.56-1.73

µm 1.92-2.09

µm 5.5-14.3 µm

Target Gases O2 A band CO2, CH4 CO2 CO2, CH4, O3

Spectral Res. 0.5 cm-1 0.2 cm-1 0.2 cm-1 0.2 cm-1 SNR > 300 > 300 > 300 > 300

Time Res. 4 sec / interferogram (5-point observation CT) Spatial Res. 10 km (nadir view at sub-satellite)

Swath Width 750 km (5-point observation cross track)

TANSO-FTS TANSO-CAI

GOSAT

_____________________________________________________

Proc. ‘Envisat Symposium 2007’, Montreux, Switzerland 23–27 April 2007 (ESA SP-636, July 2007)

Table 2. TANSO-CAI characteristics Band Band1 Band2 Band3 Band4

Center Wavelength

0.38 µm 0.67 µm 0.87 µm 1.60 µm

Bandwidth 20 nm 20 nm 20 nm 90 nm Spatial Res. 0.5 km 0.5 km 0.5 km 1.5 km

Num. of Pixels 2000 2000 2000 500 Swath Width 1000 km 1000 km 1000 km 750 km

SNR > 200 > 200 > 200 > 200

2. SENSOR OPERATION

The FTS observation is operated by pointing mirror as shown in Fig. 3. The FTS can observe over the land in lattice with its pointing mechanisms in every 3 days. The normal observation over the land is operated at 5 points cross tracking direction in 4 sec with image motion compensation (IMC) along tracking direction during obtaining an interferogram. The FTS can also observe over the oceanic sun glint with tracking by the pointing mechanisms. Similarly, particular observations for calibration and validation are operated by previous planning.

The collection of calibration datasets are planned in 2 ways. One is the calibration sites are observed as particular mode actively. The other is the data collection from normal observation mode because observation mode change has an impact in decreasing of observation grid points.

Figure 3. Operation of the TANSO-FTS observation

3. ON-BOARD CALIBRATION

The measurement spectra of TANSO-FTS are evaluated in orbit by onboard and vicarious calibration methods. The onboard method is implemented by using calibration mode data, on the other hand, the vicarious method by using observation mode data basically. The complementary evaluations are important for keeping sensor performance throughout the mission life.

The onboard calibrations of FTS are listed in Tab. 3. The solar irradiance data by onboard diffuser is obtained when the satellite passes over the northern polar region before starting SWIR observation. The SWIR spectra are

evaluated by solar irradiance and cold space view. The ILS calibration is operated only for Band 2 by using the 1.55µm diode laser after the reflection of solar diffuser. The TIR observation spectra are calibrated with the hot blackbody and cold space view. The blackbody is equipped without active heater control, but temperature of the whole sensor body is under control.

The onboard calibration of CAI is also listed in Tab. 4. The sensitivity cannot estimate continuously by onboard calibration except lunar observation once a year. The pre-launch calibration of CAI is most important activity because of less onboard calibration.

Table 3. TANSO-FTS Onboard Calibration FTS OBC B1 B2 B3 B4 Freq. Electric Noise O O O O TBD Deep Space O O O O Every pass Solar Irradiance O O O Every pass Blackbody O Every pass Lunar Observation O O O Once a year 1.55µm ILS Laser O TBD IFOV Camera O O O O TBD

Table 4. TANSO-CAI Onboard Calibration

CAI OBC B1 B2 B3 B4 Freq. Electric Noise O O O O TBD Pre-scan Pixels O O O Every scan Optical Black Pixels O Every scan Night Observation O O O O Every pass Lunar Observation O O O O Once a year

4. VICARIOUS CALIBRATION

After the operational on-board calibration, the verification of calculated L1B radiance and spectra are ensured by the other calibration method. The sensor characterization in orbit is also checked by using observation datasets. The data accumulation enables the estimation of SNR and sensitivity in orbit.

The vicarious calibration is implemented by cross calibration using other satellite sensor data, which has similar wavelength bands, and physical parameter datasets, which are generated from L2 data and global climatology. The comparable sensors are required to be well calibrated and located near local time of GOSAT orbit. The strategy of post-launch calibration is based upon match up datasets for direct comparison with other sensor radiance and data accumulation of limited conditions, such as clear-sky and stable surface albedo, for calculated spectra from physical parameters. 5. CALIBRATION SITES

Appropriate calibration sites are investigated by analyzing the available satellite sensor data, such as MODIS. The selected calibration sites are utilized for cross calibration with other well-calibrated sensors. The data accumulation enables the estimation of SNR and sensitivity in orbit.

88 – 800 km

GPS

GPS receiver

Pointing mechanism

GOSAT

Star tracker

Satellite direction(Along track)

5 cross track patterns 1, 3, 5, 7, 9 points/cross track scan

Cross track

The surface reflectance data, Aqua+Terra MODIS global reflectance of 16 days average data. After preparing 1-year data, 22 datasets are available in this study. The calculated indices are as follows. (1) Mean reflectance, (2) STDEV/Mean reflectance in a year, (3) STDEV/Mean reflectance in 5x5 pixels (approximately 25km area). The index (2) shows time stability in a year, and (3) shows spatial homogeneity in a large footprint of TANSO-FTS IFOV. The sites, where the above indices (2) and (3) have low value, are appropriate for calibration in terms of stability and homogeneity. Fig. 4 – 7 shows these 3 indices of each calibration site candidate. The numbers in these index figures show priorities at calibration site, which are selected from high homogeneity.

The desert has high reflectivity at SWIR bands. Several calibration sites are selected in Sahara (Fig. 4) and Rub Al Khali (Fig. 5) deserts using MODIS Band 6 (1.64µm) data. Desert has a very stable reflectance throughout a year, hence the desert sites are utilized for accumulation of datasets. In-flight calibration of POLDER on ADEOS-1/-2 was implemented by using the Sahara desert for performance characterization in orbit [2]. The POLDER site is near the site candidate of this study. On the other hand, forest and snow filed are available for calibration in limited season.

The Amazon has moderate reflectivity at SWIR for the forest moisture as shown in Fig. 6 of MODIS Band 6 data. The data from May to July will be available at Amazon site for dataset accumulation.

The Greenland as shown in Fig. 7 of MODIS Band 2 (0.86µm) has high reflectivity and homogeneity at UV-VIS bands of CAI. The data from Apr. to Aug. will be available at Greenland for dataset accumulation. However, the reflectivity at SWIR is small because of absorption by water content.

At the selected sites, the database of reflectance and BRDF will be prepared for calibration system, especially for desert, which is stable in time. The typical parameters are obtained by the current MODIS and SCIAMACHY data. The database will work as climatology of surface parameters if we cannot have enough match-up data with other sensors in operation.

Figure 4. Calibration site of Sahara desert (Investigation

for FTS and CAI SWIR with high reflectivity target)

Figure 5. Calibration site of Rub Al Khaki desert

(Investigation for FTS and CAI SWIR with high reflectivity target)

Figure 6. Calibration site of Amazon forest (Investigation

for FTS and CAI SWIR with moderate reflectivity target)

Figure 7. Calibration site of Greenland snow field

(Investigation for CAI VIS with high reflectivity target)

6. CALIBRATION METHOD The calibration datasets are accumulated at the

calibration sites with cloud free condition. The spectral mismatch between TANSO and other comparable sensors will be compensated by relations from current spectral observation such as SCIAMACHY. Fig. 8 shows the vicarious calibration flow for TANSO-FTS and CAI of UV-SWIR bands. Database of surface parameters, which works as climatology, is established previously.

Figure 8. Calibration method by using other satellite

sensor 7. SUMMARY

This study shows mainly the vicarious calibration plan and method for radiometry using MODIS data. The calibration dataset will be accumulated exactly at these sites by particular pointing mode or vicinities by normal observation mode. The observation plan and characteristics of calibration dataset will be reconsidered from the GOSAT orbit. The pre-launch calibration of EM and PFM is going on and the performance in orbit will be validated by the above plan and method. 8. REFERENCES 1. Kuze, A., Kondo, K., Kaneko, Y., & Hamazaki, T.

(2005), Greenhouse Gases Observation from the GOSAT Satellite, AGU 2005 Fall Meeting Abstracts, p. A14C-03.

2. Hagolle O., Goloub P., Deschamps P. Y., Cosnefroy H.,

Briottet X., Bailleul T., Nicolas J. M., Parol F., Lafrance B., & Herman M. (1999), Results of POLDER In-Flight Calibration, IEEE Trans. Geosci. Remote Sens., 37, 3, 1,550-1,566.