Published in Groundwater Monitoring & Remediation, vol. 44, issue 3, Summer 2024
Co-Authors: Craig Divine, Everett Fortner III, Colleen O. Barton, Caitlin Cisco, Colin Hollister, David Profusek, and Matthew Spurlin
Introduction
Hydrogeologic-focused data (e.g., groundwater levels, precipitation, hydraulic conductivity values) collected from environmental, water resources, and remediation projects are essential building blocks for site characterization, evaluation, and design. While these data may be initially obtained for specific objectives, the assemblage of this information into big data sets to gain additional insights via data transformation or advanced analytical methods is often underutilized, if not completely overlooked. This is particularly true when data are collected solely for compliance purposes without consideration of using the data to further inform and refine the conceptual site model (CSM). Leveraging these big data sets to be “mineable” to develop more robust CSMs within a hydrostratigraphic flux framework (an approach where geologic information is integrated and interpreted primarily in the context of groundwater flow) is even more critical today to address the need for resilient and cost- effective remediation strategies that meet the challenges posed by emerging contaminants. To do this effectively groundwater practitioners must embrace data science principles, which will require involvement of specialized data analysts, application of enhanced visual presentation tools, and use of stakeholder- friendly data sharing platforms (Horst et al. 2022).
Robust CSMs, which ultimately lead to successful project outcomes, are data rich and adaptive, or “living,” and, as such, discourage risky shortcuts in contaminant characterization and remedial design (Payne et al. 2008). Furthermore, the common occurrence of sites or projects
changing ownership over time can easily lead to the loss or underutilization of valuable historical data. Integrating these data within an established, centralized platform and workflow enhances efficiency, standardizes data use, facilitates knowledge transfer, and streamlines data management efforts, thereby reducing the overall effort and redirecting time toward data interpretation.
The technical advancements in big data concepts (storing, accessing, transforming, visualizations, and artificial intelligence [AI]) should support hydrogeologic-focused data or other data types to better inform on site characterization without overwhelming practicality. Establishing
goals to leverage data sets for repurpose can provide further insights. For example, automating the analysis of existing multiple soil particle- size distribution data or low- flow groundwater sampling data through a common data platform can be used to quickly characterize and/or provide useful insights on permeability distribution, design criteria, and well performance. This data repurposing takes advantage of (historical, current, and future) data collection efforts and unlocks additional data potential.
The development and use of AI as a tool for managing “big data” is gaining momentum among groundwater practitioners to enhance the utilization of diverse data types. While the underlying conditions for effective data processing in the groundwater sciences are suitable for codification by AI (i.e., explicit knowledge, objective, logical, and easily transferable), tacit knowledge, rooted in personal experiences and observations, presents a unique storage and management challenge. Training AI tools may help bridge this gap by extracting new insights based on tacit knowledge from subject matter experts. This has the potential to enrich data- driven interpretations that include the widest possible background considerations, weighted for importance. The integration of AI tools by using this approach can reimagine the way we collect and analyze knowledge to unlock opportunities in efficiency and innovation.
By centralizing and streamlining data management efforts, organizations can reduce inefficiencies and redirect efforts toward high- value interpretations. Development and application of automated data processing using established protocols, without sacrificing data quality, further supports this expansion of knowledge use. By integrating AI tools and adopting methods that connect explicit and tacit knowledge, the application of big data analytics can vastly enhance our decision- making capabilities as groundwater practitioners. Through the themes of reduce, repurpose, and reimagine, organizations can more effectively unlock the full potential of hydrogeologic data to drive efficiency and foster innovation.