A simulation for UV-VIS VIS observations of tropospheric composition from a geostationary satellite over Asia

Transcription

1 A simulation for UV-VIS VIS observations of tropospheric composition from a geostationary satellite over Asia Hitoshi Irie 1, Hironobu Iwabuchi 2, Katsuyuki Noguchi 3 Yasuko Kasai 4, Kazuyuki Kita 5, and Hajime Akimoto 6 1 Japan Agency for Marine-Earth Science and Technology 2 Texas A&M University, 3 Nara Women s University 4 National Institute of Information and Communications Technology 5 Ibaraki University, 6 Asia Center for Air Pollution Research

2 Simulation is performed to address the following questions Do the measured BUV radiances penetrate deep inside the PBL? This is important, because observations from GEO are often performed not at nadir. Also, observations at large SZAs are needed to see diurnal variations. What is a relationship between SNR and measurement precision? This will translate scientific requirements into instrument specification requirements and vice versa.

3 Overview of simulation method

4 Geometry A GEO satellite is assumed to be located at an altitude of 36,000 km over the Equator at 120 E.

5 Radiative transfer model intercomparison (K. Noguchi, H. Irie, and H. Iwabuchi) Comparisons between radiances calculated by SCIATRAN and JACOSPAR Tokyo summer 12LT Tokyo summer 15LT Tokyo winter 12LT Tokyo winter 15UT Radiances agree to within 2% in UV and 1% in VIS. JACOSPAR is used to calculate AMFs useful to investigate the penetration depth of sunlight. Tokyo summer 12LT Tokyo summer 15LT Tokyo winter 12LT Tokyo winter 15LT

6 Calculated AMF profiles at different geometries White lines indicate AMF=1 or 2 SZA=12 SZA=40 SZA=59 SZA=73 As SZA increases, stratospheric info increases but PBL info decreases.

7 Sensitivity to near-surface pollutants Nadir Tokyo For Nadir simulation that covers SZA=2-49 deg, the sensitivity is almost unchanged, suggesting that the SZA dependence should be very small for Nadir observations, including existing LEOs. For Tokyo, the sensitivity is low at Geo#3 (SZA=73 deg, winter, 15LT). In general, however, elevated conc. of surface ozone is not anticipated under this condition. In contrast, for other cases showing AMFs similar to those of Nadir, it is suggested that a GEO satellite would enable observation with a surface sensitivity similar to that of existing LEOs.

8 Preparation of synthetic spectra to quantify the relationship between SNR and precision for each 0.01 nm

9 Levenberg-Marquardt method: Forward model: Fitting window: nm for UV O nm for VIS O nm for NO nm for HCHO DOAS analysis x = x + ( KTSK + γ D) 1KTS[ y F( x )] i+ 1 i i i n i i i ln I ( λ) = ln( I 0 ( λ) c( λ)) σ i ( λ) ΔSCD i p( λ) Degree of offset polynomial : 2 nd (i.e., c(λ) = a 0 + a 1 λ + a 2 λ 2 ) Degree of polynomial : 3 rd (i.e., p(λ) = b 0 + b 1 λ + b 2 λ 2 + b 3 λ 3 ) Precision estimate For each geometry and each SNR given, 100 synthetic spectra containing different random noises are analyzed by DOAS. The mean and its 1σ standard deviation for 100 SCDs retrieved are calculated. The 1σ standard deviation is regarded as the precision. i= 1

10 Our DOAS analysis tools have been validated CINDI Cabauw Intercomparison Campaign of Nitrogen Dioxide measuring Instruments Place: Cabauw, the Netherlands (52.0ºN, 4.9ºE) *A KNMI s meteorological observation site Period: June-July 2009 Agreement within 10% Main objective To intercompare the NO 2 measuring instruments that can be used for validation of tropospheric NO 2 from satellites. Roscoe et al. (2010)

11 Relationship between SNR and precision for O 3 O 3 (UV:330 nm) O 3 (VIS:500 nm) Tokyo Dec. 20, 15:00 Dec. 20, 12:00 June 20, 15:00 June 20, 12:00 FWHM = 0.6 nm Requirement

12 Relationship between SNR and precision for NO 2 and HCHO NO 2 (VIS:450 nm) HCHO (UV:330 nm) Tokyo Dec. 20, 15:00 Dec. 20, 12:00 June 20, 15:00 June 20, 12:00 FWHM = 0.6 nm

13 Required SNR Product Required precision(vcd) Required precision(scd) SNR O 3 (UV) O 3 (VIS) NO 2 (VIS) HCHO(UV) 2.5x10 17 cm 2 (50 ppb in PBL) 2.5x10 17 cm 2 (50 ppb in PBL) 2.0x10 15 cm 2 (0.4 ppb in PBL) 5.0x10 15 cm 2 (1 ppb in PBL) 2.5x10 17 cm x10 17 cm x10 15 cm x10 15 cm

14 Comparison with SNR expected at various temporal and spatial resolutions Species for UV/VIS Sensor Required Precision Slant Column (cm 2 ) Required SNR O 3 (UV) nm O 3 (VIS) > nm NO 2 (VIS) > nm HCHO (UV) > nm Instrumental SNR 10 km, 1h 20 km, 1h 20 km, 2h

15 Summary A feasibility simulation for UV-VIS observations dedicated to define the instrument concept for GMAP-Asia has been performed. Detailed analysis of AMFs suggests that a GEO satellite would enable observation with a surface sensitivity similar to that of existing LEOs, when elevated surface ozone concentration is generally anticipated. Analyzing RTM-calculated spectra by the DOAS method gives a clear relationship between SNR and measurement precision. Observations of lower-tropospheric ozone would be feasible using Chappuis bands, with an improved sensitivity over Huggins bands. Is this valid for various surface features? To confirm this, analysis of spectra actually taken from space would be important.

16 Relevant information I will be able to share synthetic spectra (AOD = 0.1, 0.2, and 0.4; Albedo = 0.05, 0.10, 0.20). It is a good idea? Please tell me, if you are interested. To promote Korean/Japanese geostationary activities, I guess that it would be a good idea to participate in the next DOAS workshop together. The next DOAS workshop is planned to be held at Mainz, Germany, on July 25-27, 2011.