Remote Sensing of Spectral Aerosol Properties: A Classroom Experience

From my graduate school experience I find that the best way to learn and understand science is by getting your hands dirty with the relavant data when it comes to understanding remote sensing. Like remote sensing courses at many other universities in the United States, the University of Alabama in Huntsville offers two courses in Satellite Remote Sensing, ATS670 and ATS770. These courses are tailored in a manner to that allows students to get hands on experience with state-of-the art remote sensing datasets such as the MODIS . In ATS670, students select a MODIS image of their interest and perform a supervised and unsupervisd classification of the image to identify different features in the image such as land, water, clouds, aerosols, vegetation etc. The beauty of doing all this is that the student doesn't get to use any classification software! They write their own routines to perform all the required tasks including trivial tasks such as calculating minimum, maximum, mean, standard deviation to more sophisticated tasks such as histogram equalization, contrast streching, gray flipping, edge detection, fire detection, cloud detection in images to name a few. The journey begins with learning the basic principles of remote sensing and understanding the fundamentals behind seperating features in a remotely sensed image based on spectral signatures. Once the basics unfold, students write their own programs to read the MODIS image, perform a true color three band overlay, pick samples, perform image classification using several techniques such as the parellelopiped method, migrating means method, minimum distance, maximum likelihood methods and the mahalonobis classifier. The ingredients of this course a perfect blend of theoritical and practical classroom learning. ATS770 is more advanced and students use several radiative transfer models and other remote sensing tools to perform retrievals such as for cloud and aerosol properties.

Having taken these courses I have a great appreciation of hands-on experience in learning remote sensing and this is what attracted my attention to a paper by Robert Levy that appeared in the BAMS, 2007 (reference below). This paper talks about the challenge instructors face in bridging the gap between current research and the classroom and how the University of Maryland and NASA Goddard Space Flight Center teamed up to "design a graduate class project intended to provide a hands-on introduction to the physical basis for the retrieval of aerosol properties from state-of-the-art Moderate Resolution Imaging Spectroradiometer (MODIS) observations". "This paper reviews the basic physics of the remote sensing of aerosols and describes selected findings and lessons learned by the students." Students use both hand calculations based on given look-up tables of aerosol properties and the operational MODIS aerosol retrieval algorithm to carry out the class project. Aerosol retrievals are done over selected AERONET sites (shown in figure below) that aid validation of retrieved products.

Students investigated the reflectance v/s wavelength relations over these sites and they find a surface dependence. Figure alongside shows retrievals obtained by hand calculations by using selected fine and coarse mode geometries of aerosols and the fitting error was estimated to find the best fit. Overlaid is the MODIS retrieval in black. Best fit spectral AOT retrievals were then compared with AERONET AOTs . Details on the codes used and the instructions to perform the exercises is given at :

www.atmos.umd.edu/~levy/MODIS_Aerosol_Project/

This paper illustrates how " Projects such as this provide an opportunity for students and young scientists to become familiar with (and less apprehensive of) datasets of this magnitude". This paper is a must read for all those interested in learning aerosol retrieval techniques.

References :

Levy, R.C., and R.T. Pinker, 2007: Remote Sensing of Spectral Aerosol Properties: A Classroom Experience, Bull. Amer. Meteor. Soc., 88, 25–30

http://www.nsstc.uah.edu/~sundar/ats670.html

http://www.nsstc.uah.edu/~sundar/ats770.html