Multi-Sensor Data Fusion Technique

We describe a data fusion technique for combining lidar measurements with correlative observations made by passive sensors. Simultaneous measurements obtained by the Cloud Physics Lidar (CPL)1 and the MODIS Airborne Simulator (MAS)2 serve as inputs to a Kohonen self-organizing map (SOM)3 algorithm, which in turn classifies the collocated MAS+CPL pixels according to scene type; i.e., according to the number, the type, and the vertical locations of the cloud and aerosol layers present. Tests conducted using the MAS data alone show that the SOM algorithm recognizes a much greater percentage of the pixels containing high, thin clouds than does the standard MAS cloud mask algorithm. Results obtained when using the combined measurements identify a greater number of distinct classes of data (i.e., scene types) within individual MAS pixels, and thus will allow the selection of more accurate physical models for the retrievals of radiatively significant properties. The wealth of information gained by including the lidar profile in the SOM algorithm clustering study is, however, limited by (to) the instruments nadir only footprint. To fully utilize this information a method for extending or transferring the classification to the full passive swath was developed. We will describe our technique for extending the classifications derived from the nadir track analysis, where we have coincident measurements from both instruments, to the full passive sensor swath, for which we have only MAS measurements. Preliminary validation studies show that we can expect a classification success rate of better than 70 percent when applying this method. The recent CALIPSO-CloudSat validation campaign will provide additional datasets to validate our technique. REFERENCES M. J. McGill, D. L. Hlavka, W. D. Hart, V. S. Scott, J. D. Spinhirne, and B. Schmid, "The cloud physics lidar: Instrument description and initial measurement results", Applied Optics, 41, pp. 3725 3734, 2002. King, M. D., Y. J. Kaufman, W. P. Menzel and D. Tanré, "Remote sensing of cloud, aerosol, and water vapor Properties from the Moderate Resolution Imaging Spectrometer (MODIS)", IEEE Trans. Geosci. Remote Sens., 30, pp. 2 27, 1992. T. Kohonen, Self-Organization and Associative Memory, 3nd Edition, Springer-Verlag, Berlin, 2000.