Application Study of Principal Component Based Physical Retrieval Algorithm for Hyperspectral Infrared Sensors

摘要

An ultra-fast principal component based physical retrieval algorithm has been developed at NASA Langley research center. Works are under way to maximize the application potential of the algorithm in order to generate reliable products from hyper-spectral sensor data for climate studies. The algorithm has been tested using synthetic data for various infrared sensors. This paper describes in detail about retrieval sensitivity study carried out for several hyper-spectral infrared sensors using this physical algorithm. The retrieval accuracy obtained using the algorithm for the atmospheric parameters including trace gases of interests is discussed. Its dependence on the sensor system noise and spectral resolution is illustrated by comparing the retrieval performance achieved for different sensors. PCRTM has been demonstrated to be a reliable tool for end-to-end sensor performance simulations and has great potential for real-time trace gas retrieval applications.