GOSAT Project is a joint project of MOE (Ministry of the Environment), JAXA (Japan Aerospace Exploration Agency) and NIES (National Institute for Environmental Studies). Data acquired by TANSO(Thermal And Near infrared Sensor for carbon Observation)-FTS (Fourier Transform Spectrometer) and TANSO-CAI (Cloud and Aerosol Imager) on GOSAT will be collected at Tsukuba Space Center, in JAXA. The level 1A and 1B products of FTS (interferogram and spectra, respectively) and the level 1A product of CAI (uncorrected data) will be generated at JAXA and will be transferred to GOSAT Data Handling Facility (DHF) at NIES for further processing. Radiometric and geometric correction will be applied to CAI L1A product to generate CAI L1B product. From CAI L1B product, cloud coverage and aerosol information (CAI Level 2 product) will be estimated. The FTS data that is recognized to have “low cloud coverage” by CAI will be processed to generate column amount of carbon dioxide CO2 and methane CH4 (FTS Level 2 product). Level 3 product will be “global map column amount” of green house gases averaged in time and space. Level 4 product will be global distribution of carbon source/sink as well as 3-dimensional atmospheric CO2 concentration derived from inverse and other models. Major data flow will be also described. The Critical Design Review of the DHF was completed in the end of July of 2007 to prepare the scheduled launch of GOSAT in December 2008. The data products can be searched and will be open to the public through GOSAT DHF after the data validation process.
Greenhouse Gases Observing Satellite (GOSAT) will observe infrared light reaching its sensors from the earth's surface and the atmosphere and give spectra which can be used to derive the column abundances of carbon dioxide and methane. The observation instrument onboard the satellite is called the Thermal And Near-infrared Sensor for carbon Observation (TANSO). TANSO is composed of two sensors : a Fourier Transform Spectrometer (FTS) and a Cloud & Aerosol Imager (CAI). This article describes characteristics of data from each band of TANSO-FTS, atmospheric radiative transfer of the input signal to the sensor, and outline of the data retrieval algorithms of the Short Wavelength Infra-Red (SWIR) bands of TANSO-FTS.
The Fourier Transform Spectrometer (FTS) is a powerful tool for measuring weak radiations with high spectral resolution, by virtue of its multiplex advantage. However, to fully appreciate this advantage, we have to be patient during a time for the acquisition of an interferogram that is to be converted to a spectrum of the band concerned. In case of the FTS of GOSAT (Greenhouse gases Observing Satellite), it takes four seconds for obtaining an interferogram of Earth-reflected solar radiations. To observe the reflected solar radiation from the Earth from moving satellite in space, we encounter a difficulty that during the acquisition of the interferogram, the optical characteristics of instantaneous filed of view (IFOV) could change. One of the causes of IFOV radiance fluctuation is the fluctuation of line-of-sight of the spectrometer, which is induced by the image motion compensation (IMC) of scanning mirror. Simulations showed that such disturbances could generate serious errors in CO2 retrieval. A method has been shown by Aoki et al. (2006)1) to correct this type of disturbances in the interferograms. The disturbance to IFOV radiance could also be induced by other causes. In this paper, we examine the effects of pulse-like glittering caused by the reflection from roof of houses, cars or others, and show the same correction method to the interferogram as that adopted in IMC correction well works.
A method of retrieving CO2 column amount has been developed for near infrared solar radiation measurement in clear sky condition. It is based on the maximum a posteriori retrieval method in which a vertical pressure grid is optimized in terms of uniform data density using the normalized Jacobian matrix. CO2 retrieval error due to instrumental noise, aerosol, and surface property has been evaluated for the CO2 1.6μm absorption band. The instrumental noise yields about 0.11% retrieval error in CO2 column amount in the case of surface albedo of 0.3. Although retrieval error due to aerosol depends on a composition and size of aerosol, it rises with increase in the aerosol optical thickness and tends to overestimate with increase in the ground surface albedo. When the aerosol distributes near surface and the optical thickness is smaller than 0.2 (at 0.55μm), CO2 retrieval error is almost within 1%. However, the aerosol floating in the upper troposphere and the stratosphere such as the yellow sand and volcanic ash significantly affects CO2 column amount in which retrieval error results in much larger than 1%. The wavenumber dependency of surface albedo is also an error factor that can not be neglected. But the linear approximation at wavenumber intervals of 25cm-1 can reduce the retrieval error due to the wavenumber dependency.
The Greenhouse Gases Observing Satellite (GOSAT) will be launched in 2008 for global observations of greenhouse gases such as CO2 and methane. This study examines the feasibility of retrieving CO2 vertical profiles from spectra from 700 to 800cm-1 (referred to as “CO2 15-μm band”) of the GOSAT/Thermal And Near infrared Sensor for carbon Observation (TANSO)-FTS band 4. Retrieval simulations in which the non-linear maximum a posteriori (MAP) method was applied to pseudo-spectra at CO2 15-μm band (signal to noise ratio of 300) showed that retrieved CO2 profiles agreed with true CO2 profiles to within ±1% above 800hPa without strongly depending on a priori when atmospheric conditions such as temperature were known. These simulation results confirm the validity of CO2 retrieval at CO2 15-μm band of TANSO-FTS. Differences between retrieved and true CO2 concentrations greatly increased if atmospheric data, especially temperature data, used in the retrieval included some bias and random errors; even 1K bias of temperature yielded as many as 9% bias in retrieved CO2 concentrations. However, treating such uncertainties in atmospheric conditions as part of measurement noise and selecting retrieval channels on the basis of the information contents of CO2 retrieval could reduce the discrepancies between retrieved and true CO2 concentrations; the effect of bias in atmospheric conditions decreased by almost half and that of the random error decreased to be negligible. Furthermore, channel selection could cut the computational cost for the retrieval depending on the number of the selected channels. It depends on seasons and regions how many channels should be selected in the retrieval; therefore, it should be needed to examine the method of the channel selection for each atmospheric condition.
The GOSAT satellite has been designed for retrieving the carbon dioxide amount in the atmosphere by a Fourier transform spectro-radiometer (TANSO-FTS). It is, however, known that FTS-observed radiances in the near-infrared spectral region are heavily contaminated by a solar radiation component scattered by atmospheric aerosols, which is known as “path radiance”. In order to correct the path radiance, a multispectral visible-near infrared imager (TANSO-CAI) will be on board the GOSAT satellite to acquire the aerosol and cloud information that is indispensable for correcting the atmospheric path radiance to improve the carbon dioxide remote sensing by FTS. In this article, we study aerosol retrieval algorithms and cloud screening algorithms using the four channel radiances of CAI. We also present the methodology for vicarious calibration and validation for the CAI remote sensing, intercomparison of CAI-retrieved aerosol products with those from other satellite-borne sensors, and the data fusion strategy of satellite-measured and model-simulated aerosol properties.
We studied the impact of the planned atmospheric CO2 observations by GOSAT (Greenhouse gases Observing SATellite) on the inverse model estimate of monthly and annual mean regional CO2 fluxes, using the inverse models of the CO2 fluxes for 22 large regions of a globe. The inverse model of the atmospheric transport uses monthly and annual means of simulated GOSAT CO2 observation data, which were aggregated to grid cells with the size of 7.5°×7.5° (horizontal), and surface-station observations as formulated in Transcom-3 inverse modelling inter-comparison. The observation frequency of GOSAT in our analysis was corrected for simulated probability of clear-sky conditions, assuming global mean clear sky probability of 11%. Our results demonstrate that the mean regional flux uncertainty can be reduced by about 50% by adding satellite observations with single shot precision of 5ppm and randomly-distributed retrieval bias of 0.5ppm.
Validation plan for GOSAT-TANSO products is described in this article. The validation will be started 3 months after launch. The present main ideas of acquiring validation data for TANSO-FTS SWIR L2 products of the column abundance of CO2 and CH4 are as follows. A ground-based high-resolution Fourier transform spectrometer will be installed and co-located with a lidar and a sky-radiometer at some stations for extensive validation data acquisition. In-situ measurements and/or sampling of CO2 and CH4 onboard commercial and charter aircrafts will be also employed. National Institute for Environmental Studies (NIES) GOSAT project team and as well as other organizations responsibly concerning with GOSAT project, research teams selected through Research Announcement (RA) will conduct the validation of products.