Title of Presentation: Sensor-web Operations Explorer (SOX) for Integrated Air Quality Campaign

Primary (Corresponding) Author: Meemong Lee

Organization of Primary Author: Jet Propulsion Lab

Co-Authors: Kevin Bowman, Adrian Sandu

Abstract:  A sensor-web in this context is defined to be an integrated observation infrastructure composed of multiple sensors (instruments) deployed on multiple orbital/sub-orbital platforms. The number and type of each sensor/platform combination defines a specific sensor-web measurement design strategy, and the optimal strategy varies with the science question(s) to be addressed and available resources. To explore operational scenarios of a complex sensor-web that answers science questions optimally requires an advanced design process that overcomes the limitations of the single-platform-oriented design paradigm.

The observation involves forward modeling and inverse modeling activities where the forward modeling establishes the transformation of an atmospheric state to measurements and the inverse modeling establishes the transformation of the measurements to atmospheric state. The platform and instrument concepts as well as operation scenario concepts can be evaluated in a virtual observation system by applying retrieval analysis methods and data assimilation algorithms to the simulated measurements and quantitatively analyze their impacts on science objectives.

The goal of the Sensor-web Operations Explorer (SOX) is to provide an advanced observation system simulation experiment (OSSE) capability for the global and regional Earth atmospheric science community. A three-step approach has been devised to achieve the goal. The first step is to develop a flexible concept design exploration space for sensor-web system architectures and operational scenarios. The second step is to develop a science impact metric that can be applied to quantitatively evaluate science-return from the explored system architectures and operational scenarios. The final step is to establish a process that coordinates interdisciplinary collaboration between scientists and engineers to develop optimal observation scenarios.

Development of a flexible concept design exploration space involves population of sensor-web architecture concepts and sensor-web operation concepts that can be explored for optimality. The optimality criteria may vary depending on the project phase as different challenges and constraints surface. During the early design phase, the design trade space needs to be populated by exploring the sensitivities of multiple design concepts with respect to the potential challenges and constraints. The populated design trade space can be used to progressively optimize the design as the challenges and constraints become more specific. Development of a science impact metric involves sensitivity analysis methods for each sensor-web architecture and operation scenario concept with respect to measurement requirement and science objective. The sensitivity analysis evaluates the effectiveness of the candidate sensor-web operation scenarios with respect to the retrieval accuracy and the prediction accuracy. The retrieval accuracy is measured by the estimation error of the atmospheric state from the simulated measurements. The prediction accuracy is measured by the assimilation error of the atmospheric state from the estimated atmospheric states. Establishment of the interdisciplinary collaboration process involves multi-faceted interface infrastructure and multi-staged integration infrastructure. The multi-faceted interface infrastructure addresses interactions with the concept designers, exploration processes, and computational resources. The multi-staged integration infrastructure addresses coordination of the discipline models, data product generation, and data/information management.