QPRC 2016

Wrangling Big Data to Herd Geoprocessing for Analysis of Engineered-Natural Systems


Kelly Rose and Jennifer Bauer

National Energy Technology Laboratory


Despite more than two centuries of exploration and rapid engineering of the Earth’s crust, resulting in more than six million deep wellbores with depths exceeding 40,000 feet in some parts of the world, our ability to constrain subsurface processes, properties, and interactions remains limited.  Subsurface properties vary and can be analyzed on a variety of spatial scales.  Characterization and prediction of subsurface properties, such as depth, thickness, porosity, permeability, pressure and temperature, are important for developing robust models and interpretations of subsurface engineered and natural systems.  These studies contribute to insights and understanding of the natural, subsurface system but are also used for predictions and assessments of subsurface resources (hydro, heat, hydrocarbon, mineral, storage capacity) and support environmental and geohazard assessments.  However, the availability of data used to characterize these systems as well as the techniques that utilize those data vary significantly.  There is a wealth of data and information in structured and unstructured datasets stemming from subsurface characterization and interpretation studies.  In addition, the geo-data science landscape is shifting, becoming more open, and affording opportunities to fill in knowledge gaps through continued analysis and characterization of subsurface systems as well as innovate how existing data and results are consumed and utilized to improve our understanding of the subsurface.  A significant shift in this landscape centers on the implementation of novel big data computational technologies and approaches in combination with data-science and geologic methods to better address some of the difficulties encountered when using subsurface data to characterize, analyze, and model engineered and natural systems. In this presentation, we will discuss the breadth of subsurface data, discuss challenges researchers often encountered when working with these data, and detail on-going big data computing approaches in development to integrate geoscience computing and geospatial analytical methods to improve data integration, processing, and analysis as well as reduce uncertainty pertaining to engineered-natural systems.