Researchers and Energy Safety staff push for open data, shared modeling and a 'commons' to speed wildfire risk tools

Researchers from the University of California, San Diego's WIFIRE Lab and data staff from the Office of Energy Infrastructure Safety (OEIS) told the California Wildfire Safety Advisory Board on March 5 that making models and data more findable, interoperable and reusable will speed practical tools for mitigation and response.

University researcher Ilkay Altintas, director of the WIFIRE Lab, described a 'commons' approach that bundles a community of practice, a marketplace for reusable open models/data and a cloud-hosted environment where researchers and practitioners can run and compare models. Altintas urged the board to favor open, transparent datasets and modular models so utility and public-safety planners can test alternatives and scale successful tools.

Altintas said the nature of fire modeling requires a range of model types and resolutions. "One size does not fit all in fire modeling," she told the board, noting a tradeoff between fast 2D rate-of-spread models used for early incident awareness and deeper 3D fuels-and-behavior models used for fuel treatment and prescribed fire planning.

Why this matters: Board members and OEIS staff repeatedly raised resource and interoperability limits during the meeting. The WIFIRE Lab and OEIS presenters argued those limits make it harder for regulators and utilities to evaluate mitigation options consistently and that a commons and common interfaces can reduce the need for duplicated, closed analyses.

OEIS data and analytics staff described both their internal data work and a recent program-level example. Shafi Mohammed, chief data officer for OEIS, said the agency has expanded its data and analytics unit to support decision-focused analysis and to move from descriptive reporting (what happened) to predictive and prescriptive capabilities (what will likely happen; what actions to take).

Jenny Reid, research data supervisor, explained how OEIS frames complex problems: convert policy goals into research questions, translate those to tractable use cases, then design data products and architectures (data collection, analysis tools and visualizations) to answer those use cases. Reid described the approach as a repeated, collaborative cycle between policy experts and data teams so technical outputs map to policy needs.

OEIS engineer Stefan Shonczak walked the board through a recent example: the office's work to assess SB 884 undergrounding proposals. He said the statute asked OEIS to approve 10-year undergrounding plans only if they "substantially increase electrical reliability and substantially reduce the risk of wildfires," and to compare undergrounding to other mitigation strategies. OEIS developed a four-screen process that: (1) checks eligibility and location requirements; (2) tests financial feasibility; (3) does project-level probabilistic risk analysis (PRA); and (4) prioritizes projects. The process asks utilities to provide standardized data products and analysis so OEIS can stress-test model assumptions and compare outcomes across projects and corporations.

Shonczak emphasized the difference between probabilistic risk assessment and risk-based decision frameworks used for rate-setting. He urged a consistent separation: use PRA to estimate likelihoods and consequences, then use decision frameworks to incorporate policy choices such as risk tolerance.

Panel Q&A and board reaction Board members asked how to make utility risk models and AI-driven tools transparent without creating a flood of unusable technical detail. Panelists said standardization, documented provenance and clear metadata make disclosure useful; a commons-style approach would let utilities publish model inputs and outcomes in a discoverable format OEIS and others can query and analyze.

The board's members flagged two practical needs: (1) better bookkeeping of outage causes and equipment changes; and (2) accessible, interpretable presentations of model outputs for nontechnical decision makers. The OEIS team said the agency is piloting data catalogs and standards and that additional staff capacity will be needed to implement data-sharing at scale.

Ending: Panelists urged the board to support standardized, interoperable data and open-model practices that are combined with clear governance and provenance so regulators, utilities and communities can compare mitigation options and measure real-world outcomes more quickly.