Hierarchical software quality assurance approaches to improve cybersecurity

MSU Cyber HSQA Leadership Team

Track ongoing cybersecurity research on the Software Engineering and Cybersecurity Laboratory webpage. Current students advised under this project include Aidan Keefe (BS), Brittany Boles (MS), Redempta Manzi Muneza (PhD), and Thomas McElroy (PhD).

Improving hazard preparedness with natural language processing and narrative science messages

Figure from Reinhold et al. (2023) depicting mixed-methods procedure for narrative message construction

Regardless of hazard type (from cyber hazards to hurricanes), effective risk communication is critical to circumvent a hazard-to-disaster trajectory. Unfortunately, conventional risk messaging is generally ineffective. However, a growing body of research indicates that hazard preparedness is improved by communicating risk information in story or narrative form. I have become passionate about applying natural language processing (NLP) approaches to improve the study of an innovative narrative-based risk communication framework that finds that effective communication employs story structure to engage the target audience. My initial work in risk communication was was focused on flood risk. Recently, my interests have expanded to cybersecurity risk communication in recognition of the fact that even the most robust technical cybersecurity defenses are ineffective if humans do not implement them. Initial work in the cyber domain was supported by a grant from DHS, but I am actively seeking funding to enrich it.

Collaborators: Liz Shanahan, Jamie McEvoy, Eric Raile, Clem Izurieta, Henry King, Geoff Poole, Richard Ready, Nicolas Bergmann; Barry Ezell, Ross Gore, Chris Lynch, Madie Munro

Funding: National Science Foundation grant #CMMI-1635885; United States Department of Homeland Security Research Appropriations Grant (Cyber QR Ops)

Expansive reactive transport from soils to streams: operationalizing theory with simulation science

Presentation from the 2021 Society for Freshwater Science Annual Meeting

Because clean water is a keystone of resilient coupled natural-human systems, understanding how solutes are transported and processed is a critical scientific objective. Water transport is a primary control on water quality as many solutes are transported and thereby redistributed by water. Across spatial scales from soil columns to catchments, water moves via a suite of flow paths with varying residence times; some water moves "fast" and other water moves "slow." The accumulation of these residence times constrains the biogeochemical potential of a system. I wrote the source code for the System for Environmental Reactor Simulation (SystERS) model and R package, developed to investigate how nutrient concentrations in surface waters are constrained by coupled reaction and transport dynamics occurring in distinct geomorphic process domains (e.g., soil, groundwater, stream).

Publications: Reinhold et al. (In Review)

Software: Reinhold et al. (2022) systERS: System for Environmental Reactor Simulation

Collaborators: Stephanie Ewing, Rob Payn, Maury Valett, Geoff Poole, Yvette Hastings, Stephan Warnat

Funding: Consortium for Research on Environmental Water Systems (CREWS) National Science Foundation EPSCoR Cooperative Agreement #OIA-1757351, NSF SitS Award #2034430

The nexus of parsimony and complexity: simulating whole-system biogeochemistry using first principles

Graphical abstract from Reinhold et al. (2019)

Understanding the linkages amongst biogeochemical cycles is a well-recognized, critical scientific objective. However, progress has been hampered by an inability to simulate several elemental cycles contemporaneously. Thus, to assist thinking in terms of biogeochemical systems rather than individual elemental cycles, we developed the "Generalized Algorithm for Nutrient, Growth, Stoichiometric and Thermodynamic Analysis" (GANGSTA) that automates the creation of user defined, constraint-based biogeochemical models with any number of elemental cycles, microbe types, and microbial pathways. Such models are founded in thermodynamic theory and simulate microbial metabolism, growth, and linked elemental cycling in user-specified in silico biogeochemical systems subject to stoichiometric constraints.

In our 2019 paper in Ecological Informatics, we present a series of GANGSTA-derived models that simulate linked carbon, nitrogen, oxygen, and sulfur cycling and reproduce realistic biogeochemical patterns. Among the most important implications of our modelling exercise is a clear demonstration of how ecosystem models that focus only on carbon and nitrogen cannot adequately account for the full energy and chemical budgets of ecosystems. We hope that the GANGSTA will inspire biogeochemists, systems ecologists, and computational biologists alike because it efficiently instantiates conceptual models via a computational framework, facilitating rapid hypothesis creation, testing, and validation. We don't yet have GANGSTA up on CRAN, but the 'gangsta' R package can be downloaded from GitHub here.

Publications: Reinhold et al. (2019)

Software: Poole and Reinhold (2019) GANGSTA: Generalized Algorithm for Nutrient, Growth, Stoichiometric and Thermodynamic Analysis

Collaborators: Geoff Poole, Clem Izurieta, Ashley Helton, Rob Payn, and Emily Bernhardt

Funding: National Science Foundation grant #DEB-1021001; National Science Foundation EPSCoR Cooperative Agreement #EPS-1101342

Quantifying hydrogeomorphic controls on invasions of Elaeagnus angustifolia (Russian olive) on riverine floodplains

Russian olive (red) in flood inundation zones on the Yellowstone River Floodplain from West et al. (2020)

Elaeagnus angustifolia (Russian Olive) invasions threaten native plant communities and are commonplace in many riverine corridors in western North America. Depth to water table, hydrochory, and seed deposition are all potentially important drivers of Russian Olive distributions in floodplains. Each of these mechanisms is governed by fluvial hydrogeomorphology; however, hydrogeomorphic controls on Russian Olive distributions within floodplains are poorly understood. My colleagues and I are working to determine how Russian Olive invasions are correlated with patterns in flood-inundation frequency and hydrogeomorphic legacy to begin to tease out the potential for hydrochory to accelerate invasion rates and to understand the floodplain habitats most vulnerable to invasion.

Publications: West et al. (2020)

Collaborators: Natalie West, Geoff Poole, John Gaskin, and Erin Espeland

Funding: United States Department of Agriculture Agricultural Research Service and Department of Interior Bureau of Land Management

Quantifying the importance of side channels for riverine biota

Yellowstone River fish habitat use of side and main channels according to hydroperiod; Reinhold et al. (2016)

The focus of my Ph.D. research was to understand how alterations to fluvial processes caused by manmade structures (e.g., bank stabilization, rip rap, dikes) influenced fish assemblages and fish habitats in a large, unimpounded alluvial river. I published three papers summarizing this work, which addressed (1) quantifying the changes in side channel areas over a 50 year period and relating these changes to the density of manmade structures that "plug" side channels; (2) comparing the habitat use of side channels to main channels by small fish during the late-spring/early-summer freshet; and (3) quantifying how bank stabilization and side channels influenced main-channel fish assemblages during base flow. Since completing my Ph.D., I've continued this vein of research in a collaboration led by Brian Tornabene, focused on understanding the movement and habitat selection patterns for a species of riverine turtle (Apalone spinifera).

Publications: Reinhold et al. (2016); Reinhold et al. (2017); Reinhold et al. (2018); Tornabene et al. (2019)

Collaborators: Al Zale, Bob Bramblett, Geoff Poole, Dave Roberts, Brian Tornabene, Mike Duncan, and Matt Jaeger

Funding: United States Army Corps of Engineers; Montana Cooperative Fishery Research Unit