Cooperative Institute for Research in the Atmosphere

How Do We Make And Improve Satellite Data Products That Enable Breakthrough Science? Journeys Through Solar Backscatter Observations

Over the past two decades, satellite solar backscatter instruments have been a cornerstone for monitoring atmospheric composition, including near surface pollutants and climate agents such as ozone, sulfur dioxide, and nitrogen dioxide. Some of these same sensors were also surprisingly able to measure a small signal arising from solar-induced fluorescence inside leaves, global measurements related to the total amount of carbon taken up by plants. All these satellite-derived products require complex physics-based retrieval algorithms that transform spectral measurements of backscattered sunlight into useful geophysical quantities that can be used in scientific studies. In this talk, I will present several examples that illustrate how important advances in retrieval algorithms occur unexpectedly or in a non-linear manner, and how retrieval improvements then lead to important new science. For example, machine learning is now being used not only to speed up processing of the massive amounts of available satellite data, but also to improve products by reducing noise and expanding coverage in cloudy areas. These improvements will continue to yield new science and applications.