PCA-OS: A Planetary Climate Adaptation Operating System

Predictive climate machine learning is increasingly good at forecasting hazards, but hazard maps alone do not decide what to do, where, when, for whom, and under which futures. We argue that climate ML remains insufficient for adaptation unless interventions are treated as first-class, versioned, and auditable objects. This matters because many climate digital twins still prioritize state estimation and simulation, while adaptation requires intervention observability, counterfactual effect estimation, and constrained portfolio choice. We propose PCA-OS (Planetary Climate Adaptation Operating System), a decision-support operating abstraction that uses an intervention-aware global causal knowledge graph and standardizes object schemas, versioned updates, query primitives, and audit interfaces across three core system objects: (1) an Adaptation Intervention Ledger that records measurable interventions with provenance and uncertainty; (2) a Causal Effect Atlas that stores scenario-indexed, spillover-aware estimands, identification assumptions, diagnostics, and sensitivity bounds; and (3) a Robust Portfolio Decision Layer that optimizes intervention portfolios under budget, equity, and no-harm constraints. We argue that foundation models and intervention-aware world models should support, rather than replace, identification-aware causal analysis by surfacing candidate confounders, mechanisms, and spillover pathways for human review. Finally, we outline AdaptBench, an evaluation suite in which systems can fail for inequitable or maladaptive recommendations even when predictive accuracy is high. The result is a field-level provocation: move climate ML from read-only hazard intelligence to auditable decision support for adaptation.