GRWG UV Sub-Group Web Meeting 2014-12-16
Hyperspectral SNO comparisons of VIS channels
- Rose Munro (Chair UV Sub-Group) - Introduction
- Xu Na (CMA) Retrieving the SRF of VIS channels using hyperspectral instruments
- Dave Doelling (NASA) Introduction of the SCIAMACHY SBAF web tool
- Dave Doelling (NASA) Inter-calibration of both GEO (Met-9) and LEO (MODIS) against SCIAMACHY
- Yves Goevarts (Rayference) - Hyperspectral characterisation of Libya-4 site using GOME2 data and RTMs
- Ruediger Lang (EUMETSAT) GOME2 as hyperspectral reference for MSG & AVHRR
- All - Discussion
Guest Chair: Rose Munro
CMA: Xu Na, Lin Chen
CNES: Bertrand Fougnie
DLR: Thomas Walzel
EUMETSAT: Rose Munro, Marcel Dobber, Ruediger Lang, Tim Hewison, Sebastien Wagner, Rob Roebeling, Viju John
ESA: Berit Ahlers
JMA: Masaya Takahashi, Arata Okuyama, Ryoku Yoshino, Keita Hosaka
NASA: Dave Doelling, Amit Angal, Ben Scarino, David Flittner
NOAA: Fangfang Yu,
NPL: Nigel Fox, Emma Woolliams, Javier Gorrono and Tracy Scanlon
Rayference: Yves Govaerts
SRON: Ralph Snel
Other: Oksana Kushnyr
Xu Na presented her variational method of retrieving the SRF of IR channels in absorbing spectral bands (e.g. water vapour) using comparisons with hyperspectral sounders. Applicability to VIS/NIR? - Yes, but limited to spectral errors, rather than radiometric. The technique will also only work well for channels in absorbing bands - not for window channels.
Uncertainty in retrieved SRF? - Yes it is possible, but Xu has not looked at it yet - although she hopes to do so by the next GRWG meeting.
Dave presented NASA's new online tool for calculating the Spectral Band Adjustment Factor, which is based on SCIAMACHY observations over different surfaces. He also showed how limitng the range of conditions (e.g. Water Vapour column amount) can improve the uncertainty on the SBAFs. In response to questions, Dave confirmed this tool will include GOES-R and provide uncertainties and that it uses the same SRFs as used in GSICS. Sebastien asked whether the SCIAMACHY lunar observations could be used by this tool. Ralph Snel hopes to look at this.Ruediger pointed out the problem with analysing GOME-2 lunar observations, which would require a full
Dave also provided a refresher on the work his group has performed using hyperspectral sensors as a reference for the inter-calibration of GEO imagers. In this case, the calibration of SCIAMACHY was first checked to MODIS using polar SNOs and found to be <0.2%/decade. This was then compared with the VIS0.6 channel of Meteosat-9/SEVIRI, based on approximately 40 ray-matched observations per month, which have a standard error ~4-5%. He showed this method gave comparable results to other inter-calibration methods (e.g. DCC).
Yves Govaerts described comparisons of RTM simulations of TOA radiances spectra over the Libya-4 site with MODIS, MERIS and GOME-2. Yves concluded that he considered LibRadTran
V2.0b to be the most accurate model for this application. He believes the models are sufficiently independent of the BRDF model if the solar zenith angle is limited to 30°.
Rüdiger Lang provided an overview of inter-comparisons that have been conducted at EUMETSAT using GOME-2 as a hyperspectral reference instrument. This has been applied to the visible channels of Meteosat/SEVIRI and Metop/AVHRR. He also addressed the potential impact of these results to aerosol retrievals, for which inter-calibration will need to be peformed routinely and automatically. Rüdiger pointed out the impact of the differential degradation of the short-wave GOME-2 channels. A model has been developed to correct for this change (up to 3%). When this correction is applied to the AVHRR comparisons, the results suggest the AVHRR calbration is stable, with a constant bias. His initial comparisons with the geostationary Meteosat/SEVIRI are limited to the sub-satellite point, which still includes a 15% scatter, but shows a clear dark bias of ~8%, which show some variations with time, which need further investigation.