A Sensor Web Simulator for Design of New Earth Science Observing Systems

A strategic goal of NASA\´s Science Mission Directorate is the development of highly advanced global Earth observing systems to help monitor and predict key elements of the Earth\´s ocean-atmosphere system. A new concept now being considered is the "sensor web", which would allow collaborative measurements to be made by linking together various spaced-based and in situ observations along with output from numerical models. A underlying sophisticated communications fabric would enable rapid, seamless interaction across observing platforms and between observations and Earth system models. For weather forecasting applications, we envision a future architecture in which Earth observing systems are coupled with many different atmospheric, chemical, and oceanographic models and data assimilation systems such that our ability to monitor, understand, and predict the evolution of rapidly evolving, transient, or variable features in the atmosphere are better understood. In theory, tasking the observing system to collect data at specific locations in space and time could produce notable gains in predictive skill. In 2006 our team was awarded funding from NASA\´s Earth Science Technology Office (ESTO) to design and build an end-to- end sensor web simulator (SWS) based upon the proposed architecture that would objectively assess the scientific value of a fully functional model-driven meteorological sensor web. The effort is based upon two ESTO-funded studies that have yielded a next-generation weather observing system architecture and a preliminary SWS software architecture developed in 2003. Simulation is essential: investing in the design and implementation of such a complex observing system could be potentially costly and almost certainly involve significant risk. The SWS will help provide information systems engineers and Earth scientists with the ability to define and model candidate designs, and to quantitatively measure predictive forecast skill improvements. We will repor- - t on our efforts to develop a prototype simulator for a weather forecasting application, and present preliminary results in which the steps of the simulator were manually executed.