Computing the Uncertain: Environmental Data, Hazard Standardization, and the Foundations of Risk Assessment

During the 2025 hurricane season, two back-to-back hurricanes in the Atlantic Basin stressed the already ongoing concern about natural disasters and the impact of climate-risk-related environmental threats on the well-being of local populations. In recent years, environmental data streams have gained critical importance as they feed granular risk assessments and value calculations, which are increasingly relevant in decision-making processes like infrastructure spending, insurance premiums, and housing markets. The "vast machine" (Edwards 2010) appears to have transcended atmospheric politics and gone steadily into the ground. A deep entanglement of distributed sensing capabilities, decoupled from its exclusive human orientation, has assembled a new "world to be acted upon" (Gabrys 2019). Still, conceptualizations of disasters from science and technology studies have stressed the lack of straightforward confinement of their outcomes and the limits of standard operating procedures to address them (Kim Fortun et al., 2017). Uses of environmental data to minimize impact, standardize hazards, and project certainty over climate threats appear to be directed to avoid the disruption of specific sociotechnical systems. This paper is set to illuminate the infrastructural and epistemic processes that allow this by looking at the emergence and consolidation of risk assessment research since the 2010s. Using text-processing techniques on article metadata and close-up analysis of data-sharing statements, it identifies environmental data infrastructures critical in sustaining risk assessment efforts. Moreover, by showing the presence of specific entities and components of the sociotechnical systems as objects of disaster research, it describes the priorities, biases, and blind spots on what counts today as a risk, where it is perceived, and the prescribed ways to address it.