Some information on the IUP-UB satellite based verification approach
Satellite data used
Tropospheric nitrogen dioxide (NO2) columns are computed using a three step approach: 1) Applying the DOAS retrieval on the measured spectra which yields the total slant column, 2) correcting the stratospheric contribution by applying the reference sector method and 3) converting to vertical columns by applying an air mass factor. Only data having a FRESCO+ cloud fraction smaller than 0.2 are used but no further cloud correction is applied.
The data product used here has some limitations, in particular because of the simplistic stratospheric correction used which can result in negative NO2 columns at high latitudes in spring. The air mass factors are based on a monthly climatology created from one year of MOZART model profiles and thus cannot reproduce short-term variations in meteorology or biomass burning activity.
However, for the purpose of verification, the most important point is to have a product which is completely independent from the models and data assimilated into the MACC / CAMS IFS / CIFS model runs. This is the case for the product provided here.
The data is retrieved from measurements of two satellite instruments, SCIAMACHY on ENVISAT and GOME-2 on MetOp-A. The transition between the two instruments took place in April 2012 as SCIAMACHY data are only available until April 2012.
Details on the NO2 satellite retrieval can be found in Richter et al. (2005) for SCIAMACHY and Richter et al. (2011) for GOME-2.
It is important to note that the the stratospheric columns computed from the SCIAMACHY and GOME-2 observations are in fact total columns derived using a stratospheric air mass factor. To minimize the impact of the troposphere, only data over the clean Pacific region are used (180°E - 220°E). Still, the amount considered here as stratospheric includes a weighted part of the tropospheric background NO2 over that region.
Tropospheric formaldehyde (HCHO) columns are computed under the assumption that stratospheric HCHO amounts are negligible. HCHO columns are derived using the DOAS method and the air mass factors are based on a static monthly climatology using land surface classification and biomass burning statistics to estimate the vertical HCHO profile. For SCIAMACHY data, no cloud screening has been performed while for GOME-2 data, the same 0.2 FRESCO+ threshold is applied as for tropospheric NO2. As both the SCIAMACHY and GOME-2 HCHO retrievals are subject to drifts and offsets, the values are normalised over the Pacific sector where the amount of tropospheric HCHO is assumed to be negligible.
Compared to NO2, the HCHO columns have much larger random and systematic uncertainties. Therefore, satellite data are smoothed to reduce the impact of noise.
Details on the HCHO satellite retrievals can be found in Wittrock et al., 2006.
Treatment of model data
The first step consists in selecting the model output corresponding to the satellite overpass time (approx. 10:00 solar local time (LT)).
The model results are then integrated in the vertical to derive tropospheric and stratospheric vertical columns, as only vertical columns are retrieved from the satellite measurements and not concentrations at specific heights. For the separation of the stratosphere and troposphere, a latitude dependent tropopause height is used from Santer et al. (2003).
The spatial resolutions of the models are coarser than that of the standard SCIAMACHY/GOME-2 product (0.125° x 0.125°). Thus, the daily satellite measurements are re-gridded to the correspondent model resolution: 1.875° x 1.875° for IFS-MOZART NRT-Analysis and NRT-Forecast, 2° x 3° for IFS-TM5 NRT-Analysis, 1.125° x 1.125° for IFS-MOZART Re-Analysis and 0.75° for CIFS runs. In this process, the average of all valid SCIAMACHY/GOME-2 grid boxes within one model grid box is taken without applying any area weighting, i.e., all satellite data are considered as long as part of it is located within the model box.
Finally, the model data is selected according to the existing satellite data, ensuring that both datasets consist of data for the same days at the same locations. This is important because for example SCIAMACHY data are not available at daily global coverage. In optimal conditions, the global coverage would be obtained every 6 days. Moreover, NO2 tropospheric columns are only determined for clear sky pixels, i.e., cloud fraction smaller than 20% according to the FRESCO+ data (Wang et al., 2008). For this reason, the daily model data previously selected for the overpass time are then matched to the available SCIAMACHY/GOME-2 data which is already converted to the model resolution.
In order to get sufficient signal to noise and statistical representativeness, all data are considered as monthly averages only.
Richter, A., Burrows, J. P., Nüß, H., Granier, C, Niemeier, U., Increase in tropospheric nitrogen dioxide over China observed from space, Nature, 437, 129-132, doi: 10.1038/nature04092, 2005
Richter, A., Begoin, M., Hilboll, A., and Burrows, J. P.: An improved NO2 retrieval for the GOME-2 satellite instrument, Atmos. Meas. Tech., 4, 1147-1159, doi:10.5194/amt-4-1147-2011, 2011
Santer, B.D., R. Sausen, T. M. L. Wigley, J. S. Boyle, K. AchutaRao, C. Doutriaux, J. E. Hansen, G. A. Meehl, E. Roeckner, R. Ruedy, G. Schmidt, and K. E. Taylor, Behavior of tropopause height and atmospheric temperature in models, reanalyses, and observations: Decadal changes, JGR, VOL. 108, NO. D1, 4002, doi:10.1029/2002JD002258, 2003
Wang, P., P. Stammes, R. van der A, G. Pinardi, M. van Roozendael, FRESCO+: an improved O2 A-band cloud retrieval algorithm for tropospheric trace gas retrievals, Atmospheric Chemistry and Physics, 8, 6565-6576, 2008
Wittrock, F., A. Richter, H. Oetjen, J. P. Burrows, M. Kanakidou, S. Myriokefalitakis, R. Volkamer, S. Beirle, U. Platt, and T. Wagner, Simultaneous global observations of glyoxal and formaldehyde from space, Geophys. Res. Lett., 33, L16804, doi:10.1029/2006GL026310, 2006