M. Zara1, K.F. Boersma1,6, I. De Smedt2 , A. Richter3, E. Peters3, J. van Geffen1, S. Beirle4, T. Wagner4, S. Marchenko5, N. Krotkov5 , H.J. Eskes1
OMI STM, September 2017
Quality Assurance for NASA, KNMI, BIRA-IASB and QA4ECV spectral fitting algorithms for NO2 and HCHO Slant Columns
from OMI and GOME-2A
1Koninklijk Nederlands Meteorologisch Instituut, KNMI, the Netherlands 2Institut royal d'Aéronomie Spatiale de Belgique, BIRA-IASB, Belgium 3Institut für Umweltphysik, IUP, Bremen 4Max-Planck-Institut für Chemie, MPI, Mainz 5Goddard Space Flight Center, NASA, USA 6 Wageningen University, WUR, Wageningen
Motivation
Recently, the Global Climate Observation System (GCOS) has identified NO2 and
HCHO as precursors to Essential Climate Variables (ECVs) because of their value in
detecting and attributing changes in ozone and aerosol distributions (GCOS-138, 2010).
Need for reliable and traceable quality information. The EU FP7-project Quality
Assurance for Essential Climate Variables (QA4ECV) (http://www.qa4ecv.eu/) is addressing
this need by providing a fully traceable quality assurance effort on all aspects of the NO2 and
HCHO retrieval algorithms.
Here, our topic is the quality assurance of state-of-science spectral fitting algorithms for
the NO2 and HCHO slant column densities (SCDs) from OMI and GOME-2A from
NASA, KNMI, BIRA-IASB and QA4ECV.
OMI NO2 SCD retrieval algorithms
(1) Bucsela et al., 2006; (2) Van Geffen et al., 2015; (3) QA4ECV Deliverable 4.2, 2016; (4) Marchenko et al., 2015
Reference
New
OMNO2A v1, OMNO2A v2, QA4ECV and NASA NO2 SCDs
SP v3.0 - online
SP v3.1
DOAS SCD uncertainty Statistical SCD uncertainty
Comparable SCDs: their spatial variability as proxy for the SCD uncertainty
We evaluate how good the uncertainty provided by the fitting procedure is.
DOAS: a (non) linear least-square fit between observed and modelled spectra
SCD uncertainty driven by: • Reflectance uncertainty (I, I0) • Wavelength calibration • Accuracy of fitting model to mimic
observations • Stripes
+ Possible systematic effects!
Random!
OMI NO2 Statistical & DOAS SCD uncertainty
OMNO2A v1 OMINO2-QA4ECV
DOAS = random + systematic
Statistical = random
SP v3.0 SP v3.1
OMINO2-NASA
2005-2015: OMI NO2 SCD uncertainty
OMINO2-QA4ECV
The effect of stripiness on the OMI NO2 SCD uncertainty
OMI NO2: 2005-mean solar irradiance
The effect of stripiness on the OMI NO2 SCD uncertainty
OMI NO2: 2005-mean solar irradiance
OMI HCHO: daily radiance spectra vs
50 × 1015 molec. cm−2
2 × 1015 molec. cm−2
Stripiness for HCHO before and after daily radiance as reference spectra:
OMNO2A v1 (01/01/2012)
1. Reduction of NO2 SCD DOAS uncertainty globally.
2. Removal of extreme (red) uncertainties from the swath edges.
QA4ECV (01/01/2012)
Clouds affect the SCD uncertainty brighter scenes.
The SCD uncertainty is mostly driven by the SNR (TOA-reflectance) rather than the signal strength (SCD).
OMINO2-QA4ECV DOAS unc. (01/01/2015)
www.temis.nl
GOME-2A NO2 SCDs from BIRA-IASB and QA4ECV
Similar to OMI
2007-2015: GOME-2A NO2 SCD uncertainty from BIRA-IASB and QA4ECV
September 2009
Conclusions
This research was funded by the FP7 project QA4ECV grant No. 607405