July 28, 2009 | Observing System Simulation Experiments, known as OSSEs, are studies that quantify just how useful a particular observation is for answering a given science question. Scientists have used this technique for some time to test if incorporating new measurements improves weather forecasts. A new study, led by NCAR scientist David Edwards and highlighted in Journal of Geophysical Research on July 23, is the first to apply the concept of OSSEs to chemical weather (predicting pollution events and variability in the atmosphere’s chemical composition).
The research will help NASA choose the satellite instrumentation and observing strategies that will be most useful and cost effective for predicting chemical weather as the agency plans its next generation of Earth science satellite missions. Edwards and colleagues are investigators for GEO-CAPE (Geostationary Coastal and Air Pollution Events), a satellite that will observe air pollution over North and South America. The mission will help scientists identify different sources of aerosols and ozone precursors, track air pollution transport, and study the dynamics of coastal ecosystems, river plumes, and tidal fronts.
The team needed a way to determine whether the proposed new instruments for GEO-CAPE would be capable of making the measurements necessary to fulfill the satellite’s science objectives. To that end, the researchers developed a software instrument simulator testbed that allows them to “fly” conceptual instruments through a model of the atmosphere (CAM, or Community Atmosphere Model) to see what the instruments observe. Using data assimilation tools (NCAR/CISL’s Data Assimilation Research Testbed, or DART) they can then evaluate how well the simulated measurements represent the real atmosphere and, by extension, determine if the proposed instruments would be useful for GEO-CAPE.
"It's very important to establish exactly how the science questions we want to answer determine where, when, and how we make our measurements,” Edwards says. "Using simulated observations, the testbed allows us to mimic the way future real observations might be used. Because we can test the performance of several measurement scenarios, we can hopefully make better decisions about the mission design and how we can most effectively meet our science goals."