Appendix A — Appendix A — Model Permit-Condition Language

Draft regulatory text for a causal volume condition on saltwater-disposal permits. Written for adaptation by Commission staff or legislative counsel; bracketed terms are policy choices, not analytical outputs.

A.1 A.1 Definitions

Dose-response curve. For a disposal well, the estimated relationship between 365-day cumulative injection volume and expected seismicity within [7–19] km, produced by a targeted-learning estimator meeting the standards of §A.4, with 95% confidence bands, evaluated only within the supported data region (volumes at or below the 99th percentile of observed well-volumes).

Volume threshold. The lowest 365-day cumulative volume at which the lower bound of the well’s dose-response confidence band exceeds [the Commission-adopted risk tolerance].

Attributable-risk ranking. For a catalogued seismic event, the ranking of wells within [19] km by estimated causal contribution, with confidence intervals.

A.2 A.2 Permit condition (new and amended permits)

  1. For any disposal permit within [a Seismic Response Area or a Commission-designated monitoring area], the permitted 365-day cumulative volume shall not exceed the well’s volume threshold as most recently computed under §A.4.
  2. Where the well’s dose-response curve does not resolve a threshold within the supported data region, the permit shall instead carry [the area’s standard volumetric condition], and the well shall be re-evaluated at each annual refresh.
  3. A permittee may at any time submit volumes below the threshold as evidence of proactive seismic-risk management in any Commission proceeding.

A.3 A.3 Event response (complement to existing SRA protocol)

  1. Upon a catalogued event of [M3.0]+ within a monitored area, Commission staff shall obtain the attributable-risk ranking for the event.
  2. Wells whose contribution interval excludes zero shall receive priority review. Curtailment of wells whose interval includes zero may proceed under existing protocol but shall be flagged for threshold re-evaluation rather than presumptive contribution.

A.4 A.4 Evidence standards

A dose-response curve or attribution is cognizable under this section only if produced by a method that:

  1. Adjusts for confounding between injection volume and location (minimally: fault distance, formation depth, well age, and neighboring injection) using a doubly-robust estimator with published convergence guarantees;
  2. Publishes its uncertainty — confidence intervals constructed with cluster-robust variance at the well level;
  3. Declares its supported region and renders no operative estimate by extrapolation;
  4. Is reproducible — code and processed data sufficient to regenerate every operative number are available to the Commission and, [save for trade-secret well identifiers], to the public;
  5. Maintains a revision record documenting every material change in estimates across data refreshes, including artifacts detected and repaired.

Note: standard 5 is deliberate. The program maintaining this manual has itself detected and repaired a cross-validation artifact (see the Evidence Scoreboard); a method without a revision record has not been tested hard enough to be load-bearing in a permit.

A.5 A.5 Annual refresh

Thresholds and rankings shall be recomputed against refreshed TexNet and H-10 data at least annually. A permittee whose threshold rises may apply the increase immediately; a threshold reduction takes effect at the next permit anniversary, with [180 days] notice.