Accelerated Discovery Cycles

Problem Statement

Despite challenges with significant implications for mankind, the considerable recent advances in earth system science have nonetheless failed to keep pace with the accelerating need to better understand .how the world works. in the face of rapid global change. The science of global change requires a new model for comprehensive study of complex systems that combines: (a) experimental infrastructure that encompasses realistically complex earth systems, but simultaneously allows for precise manipulation of that system, and (b) information integration and informatics capacity to both ingest the massive datasets needed to capture large-scale dynamic ecosystem complexity, and to instantly process and update it in order to test contrasting mechanistic models and drive the next set of experiments. Furthermore, to ensure that such an approach is accessible to the entire scientific and educational community, the data, models, and knowledge generated need to be easily accessible and visualized. Currently, no coupled physical and computational infrastructure exists to address this pressing need. A computational collaboratory that can support such integration would generate immediate transformative science and educational opportunities.

Project Goals

The overarching objective of this project is to enable Accelerated Discovery Cycles (ADCs) by coupling recent advances in computational models and cyber-systems with the unique experimental infrastructure of Biosphere 2 (B2), a large-scale earth system science facility now under management by the UofA. The hallmark of ADC is the ability to dynamically couple and autonomically configure, in real-time, complex and customized interaction between biophysical models, environmental sensors and effectors, and experimental protocols. This ability will enable a new era of global and collaborative investigations in earth system science to address critical issues underlying global changes.

We propose a service-oriented cyberinfrastructure for pervasive access to and coordinated sharing of geographically distributed hardware, software, and information resources, as well as interactions between computations, experiments, information/data and experts distributed across a Grid. This cyberstructure will be integrated with the Biosphere 2, experimental infrastructure, which is separately supported by private funds and is available at no cost to this grant. Together, these infrastructures can revolutionize how earth science is conducted, shared and disseminated. Our central and unique deliverable will be a fully controllable and observable experimental test bed, which will enable ADCs that will fundamentally advance understanding of complex earth ecosystems and processes. Such ADCs will result from computer/computational scientists (Hariri, Parashar, Flikkema, Lewis and Akoglu) working together with earth scientists/engineers and mathematicians (Huxman, Saleska, Zeng and Jenerette) to address the core challenges of heterogeneity, complexity and scale of the data, the computational models and the system infrastructure.
However, achieving this vision presents significant computing and computational challenges that exceed current practices, paradigms and infrastructures in earth sciences, which are based on static, preorchestrated and centrally mediated models and implementations. Security, privacy, ownership, and quality of data and service require a paradigm shift that will enable ecosystems of accelerated discovery cycles. The goal of this research is to design, deploy and evaluate a decentralized collaboratory that is based on an innovative design paradigm inspired by nature and biological systems (i.e., Autonomic Computing), and can support transformative earth science research. This effort will catalyze computational thinking, enabling scientists to collectively address biophysical responses to global changes. It will also produce the next generation of earth scientists and students who can effectively exploit the capabilities and services offered by B2 and the cyberinfrastructure. The innovative computer science and information science contributions of this effort will be an infrastructure for the synthesis, enactment and runtime management of dynamic and data driven workflows that enable
(1) robust data acquisition from a large network of heterogeneous sensors,
(2) autonomic decision-making to robotically reconfigure the distribution of sensors and steer data acquisition,
(3) the ability to implement competing biophysical models and assess uncertainty in data-model fusion,
(4) analysis, visualization, and interpretation of the results from both the experimental and computational ADCs, and
(5) development of mechanisms and material to disseminate the research results and knowledge of this study.

Solving these challenges will generate active workflows capable of learning about the environment, reconfiguring to address deficiencies in the sensing network, and assessing the potential knowledge gain from different experimental protocols.

Project Deliverables

The primary deliverable of this effort will be an integrated collaboratory composed of models, tools and a data repository aimed at accelerating research and discovery towards addressing complex earth science grand challenges. The deliverables will be associated with the key phases of the ACD, i.e., physical infrastructure consisting of autonomous sensors and robots for advanced dynamic environmental sensing, autonomic infrastructure for computation and experiment management, portals and interfaces for analysis and visualization, and educational and outreach.

Broader Impacts

The proposed collaboratory for ADC will integrate the research, educational, and teaching activities carried out by earth scientists, and will empower earth scientists, computer science and engineering scientists, and statistical and applied mathematicians to integrate their research and collaborate to address the earth science grand challenges. By integrating cyberinfrastructure resources and tools (high performance computing and storage resources, high speed networks, autonomic computing software and tools, sensors and robot technologies) with a grand scale B2 facility, accelerated discovery cycles can be developed to provide an integrative framework for bridging the gap between data collection and development of earth science models. These accelerated discovery cycles will reduce significantly the time to investigate grand earth science challenges and will lead to development of accurate models that can be validated rapidly using the B2 experimental test-beds. The collaboration tools will enable earth scientists to scale up these results to relevant spatial and temporal realms . an important grand challenge. Furthermore, the resulting knowledge environment will have broad applications for addressing challenging science goals in diverse domains. The proposed collaborative environment will also lead to rapid dissemination of knowledge and discovery results to the research and education community. Specifically, the collaboratory will spur the development of new multidisciplinary research teams to collaborate on accelerated discovery cycles in an array of earth science problems.