Parts | Days | Selection | Search | Downloads | Help
UP: Fachverband Umweltphysik
UP 17: Satellitengestützte Messmethoden
UP 17.6: Talk
Wednesday, March 28, 2007, 16:15–16:30, H48
Statistical Studies about Satellite Data of NO2 — •Michael Hayn1, Steffen Beirle1,2, Bjoern H. Menze3, Thomas Wagner1,2, Fred A. Hamprecht3, and Ulrich Platt1 — 1Institute of Environmental Physics, University Heidelberg, Germany — 2Max Planck Institute for Chemistry, Mainz, Germany — 3Interdisciplinary Center for Scientific Computing, University Heidelberg, Germany
Satellite measurements provide long time series of atmospheric trace gases. Many data about tracers like NO2 are obtained from GOME, SCIAMACHY and OMI. They allow detailed studies on their properties like coherences between different observations, which allow inferences on the physics and chemistry of tracers. Purpose of the talk is to present statistical methods which can be used to visualize relevant information contained in the large data sets of NO2 received by GOME and to discuss the corresponding results. Among them is the Fourier Analysis which provides information about sources. Principal component analysis (PCA) is a powerfull method to detect outliers. Extreme events contain many informations about sources, life time and burdens on the environment. Moreover, outliers can be caused by defects in the measurement process or by errors in the data processing algorithms. It is important to recognize them and to take them into account while interpreting results. PCA can also be used to replace missing data. Additionally, work on correlation with other datasets, like different meteorological parameters, using the generalized additive model will be included. These new results shall help to improve our knowledge about the sources of NOx as well as processes in atmosphere.