Publication details
Series: NILU OR 53/2014
Publisher: NILU
Year: 2014
File:
NILU OR 53/2014 (pdf)
Summary: NILU has a mandate to monitor air quality and particularly its changes over time, both nationally through Miljødirektoratet (MD) and internationally through the European Monitoring and Evaluation Programme (EMEP). Satellite data related to atmospheric composition are increasingly used for monitoring as they provide long time series of spatially continuous observations. It is therefore essential for NILU to begin preparing for the upcoming Copernicus missions. Here, we evaluate methane products from AIRS, TES, TANSO-FTS and SCIAMACHY as added value for GHG monitoring in Norway and Svalbard. As expected, due to the low sensitivity of the sensors to ground-level Artic large deviations are seen when comparing to in situ data from Birkenes and Ny-Ålesund. Higher level products (L4), combining satellite and ground-based information, seem more appropriate for future reporting purposes. Further, we investigated the usability of the current set of long-term operational ground-based MAX-DOAS stations worldwide for inter-comparing their NO2 observations to those of satellite-based instruments, in particular OMI and GOME-2A. The two data sources agree very well for sites located in rural, non-polluted regions. For sites located in polluted areas we found strong systematic biases, large random errors, or slightly shifting systematic biases. The systematic biases can be explained primarily by the strong spatial gradients in NO2 levels in urban areas in conjunction with the large differences in the spatial representativity of the measurements. We evaluated the possibility to use the now relatively long time series of MAX-DOAS observations to fit a statistical trend model and to directly compare the resulting trends to those obtained for the satellite-based time series for the same area and time period. It was found that the sites with approximately 50 months of valid data for both data sources showed quite similar long-term trends and that sites with fewer than 30 months of valid data exhibited significant discrepancies in the resulting trends.