Embed Size (px)
Citation preview
3. Column averaged concentrations
5. CH4 profiles
2. Method
Multi-layer Retrievals of Greenhouse Gases from a Combined Use of GOSAT TANSO-FTS SWIR and TIR
N. Kikuchi1＊, A. Kuze1, F. Kataoka2, K. Shiomi1 ,M. Hashimoto1, H. Suto1, R. Knuteson3, L. Iraci4, E. Yates4, W. Gore4, T. Tanaka4, and T. Yokota5
1. JAXA, 2. RESTEC, 3. University of Wisconsin, 4. NASA, 5. NIES ＊[email protected]
Acknowledgement Gas absorption tables for CO2 … H. Tran Solar spectral irradiance … G. C. Toon Aerosol global distribution … SPRINTARS developer team
Reference • De Lange, A., I. Aben, O. Hasekamp, A. Butz, J. Landgraf
(2016), Verification of CH4 profile retrievals from GOSAT thermal infrared measurements, IWGGMS-12 presentation
• Drouin, B. J. et al. (2017), Multispectrum analysis of the oxygen A band, JQSRT, 186, 118-138
• Herbin, H., L. C. -Labonnote, and P. Dubuisson (2013), Multispectral information from TANSO-FTS instrument-Part 1: Application to greenhouse gases (CO2 ad CH4) in clear sky conditions, Atmos. Meas. Tech., 6, 3301-3311
• Kikuchi, N., Y. Yoshida, O. Uchino, I. Morino, and Y. Yokota (2016), An advanced retrieval algorithm for greenhouse gases using polarization information measured by GOSAT TANSO-FTS SWIR I: Simulation study, J. Geophys. Res, in press
• Lamouroux, J., H. Tran, A. L. Laraia, R. R. Gamache, L. S. Rothman, I. E. Gordon, j.-M. Hartmann (2010), Update database plus software for line-mixing in CO2 infrared spectra and their test using laboratory spectra in the 1.5-2.3 μm region, JQSRT, 111, 2321-2331
• Rahman, H., B. Pinty, M. M. Verstraete (1993), Coupled Surface-Atmosphere Reflectance (CSAR) Mode. 2. Semiempirical surface model usable with NOAA Advanced Very High Resolution Radiometer data, J. Geophys. Res., 98, D11, 20791-20801
• Rothman, L. S. et al. (2009), The HITRAN2008 molecular spectroscopic database, JQSRT, 110, 533-572
Summary and Conclusions • From GOSAT TANSO-FTS measurements, we obtained CO2 and CH4 concentrations over
Railroad Valley Playa in Nevada during the vicarious calibration campaigns conducted every year around the summer solstice.
• We retrieved both column amounts and vertical profiles of CO2 and CH4 by using SWIR and TIR spectra simultaneously, and compared them with The Alpha Jet Atmospheric eXperiment (AJAX) measurements.
• To reduce the influence of a priori, forward model errors and/or possible L1B spectral biases, the gas concentrations were retrieved using a small number of vertical layers: 2 layers for CO2 and 4 layers for CH4.
• In some cases, the average CO2 profile of 3-4 GOSAT measurements captures the AJAX profile. For CH4, the GOSAT retrieval has a marginal sensitivity to the AJAX profile.
1. Introduction
4. CO2 profiles
Since the launch in January 2009, TANSO-FTS onboard GOSAT accumulates high resolution spectra in SWIR and TIR. TANSO-FTS is the only sensor that measures SWIR and TIR spectra at the same time and footprint, providing a unique opportunity to retrieve CO2 and CH4 concentrations from space (e.g., Herbin et al. 2013). We have started development of a retrieval algorithm which uses SWIR and TIR simultaneously, and analyzed GOSAT measurements over Railroad Valley Playa in Nevada during the vicarious calibration campaigns in 2009-2015.
Data From 2009 to 2015, there are 30 overpasses and 86 GOSAT measurements. Cloud contamination was seen in 3 overpasses (11 measurements) and these data were excluded. We analyzed 75 GOSAT measurements. The AJAX measurements are available for 10 overpasses (36 measurements) and were compared with the GOSAT retrievals. TANSO-FTS L1B V203 (precursor to the next version) was used.
BRDF model (SWIR) RPV model by Rahman et al. (1993) was adopted, and the model parameters were derived from MODIS MCD43C1.5.
Retrieved variables • Gas volume mixing ratios • Temperature profile • Surface temperature • Surface emissivity in TIR • BRDF parameter (1 of 3) in SWIR • Surface pressure (for dry air column) • Aerosol vertical profiles (fine and
coarse, logarithm of mass mixing ratio) • Wavenumber shift • Zero level offset
Window Wavenumber [cm-1] Absorbing gases 1 12950-13200 O2, O3
2 6180-6280 CO2, H2O, CH4
3 5900-6150 CH4, H2O, CO2
4 4800-4900 CO2, H2O 5 700-775 CO2, H2O, O3, HNO3
6 970-990 CO2, H2O, O3
7 1070-1110 CO2, H2O, O3
8 1200-1300 CH4, H2O, CO2, O3, N2O, HDO
Windows
Maximum A Posteriori Retrieval Given the measurement vector 𝒚𝒚, the optimal solution of the state 𝒙𝒙, which includes gas concentrations, is obtained by minimizing the cost function 𝑱𝑱 𝒙𝒙 = 𝒚𝒚 − 𝑭𝑭 𝒙𝒙 𝑻𝑻𝑺𝑺𝜺𝜺−𝟏𝟏 𝒚𝒚 − 𝑭𝑭 𝒙𝒙 + 𝒙𝒙 − 𝒙𝒙𝒂𝒂 𝑻𝑻𝑺𝑺𝒂𝒂−𝟏𝟏 𝒙𝒙 − 𝒙𝒙𝒂𝒂
𝑭𝑭 𝒙𝒙 : forward model 𝑺𝑺𝜺𝜺 : covariance of measurement noise 𝒙𝒙𝒂𝒂 : prior estimate 𝑺𝑺𝒂𝒂 : covariance of prior estimate
Spectroscopy O2 A band … Drouin et al. (2017) CO2 ... Lamouroux et al. (2010) Other molecules ... HITRAN2008 Continuum ... MT_CKD2.8 Solar lines ... G. C. Toon
Retrieval Algorithm Developed by adding a TIR module to the existing SWIR algorithm of Kikuchi et al. (2016).
3-4 measurements /overpass
Railroad Valley Playa
Meteorological data Atmospheric temperature, humidity, and surface pressure were measured by radiosonde, which were used as a priori in the retrieval.
Surface emissivity (TIR) Measured by a ground-based FTS (S-AERI), and was used in Window 6.
A priori profiles of gases CO2 … a flat profile of 380 ppm CH4 ... mid-latitude summer
SWIR only SWIR+TIR
SWIR only SWIR+TIR
A priori (380 ppm) and its variance (5 % = 19 ppm)
Average CO2 profile and the retrieval error of single measurement
AJAX measurements
CO2 profiles (3-4 measurements/overpass)
A priori (mid-latitude summer) and its variance (20 %)
AJAX measurements
Average CH4 profile and the retrieval
error of single measurement
CH4 profiles (3-4 measurements/overpass)
Bias correction scheme To reduce the bias of CH4 observed in the troposphere, spectral residuals are re-constructed through SVD, and a linear combination of the first 4 right singular vectors are added to 𝑭𝑭(𝒙𝒙) (cf. De Lange et al. 2016).
Bias correction The same correction scheme as CH4 was applied, but the effect was less marked.
𝑨𝑨 = 𝑼𝑼𝚲𝚲𝑽𝑽𝑇𝑇 Singular Value Decomposition
