PTEM Labs

Product

PTEM structural intelligence

PTEM operates a governed structural analysis engine for tropical cyclone risk. Institutions use PTEM to quantify internal storm organization, identify high-readiness regimes, and add structural context alongside intensity and environmental models.

PTEM is not a forecast model. It is a structural instrument evaluated on a frozen, multi-decadal dataset and delivered as governed structural artifacts, with optional enterprise SDK access for reproducible integration.

What you get

PTEM consists of three governed components that map directly to the access tiers described on the Pricing page.

Frozen structural record (Atlantic v1.2, 1851–2024)

The Atlantic v1.2 frozen structural spine (1851–2024), 6-hour aligned and versioned for reproducible validation, research, and integration pilots. Includes:

  • Structural invariants (HC_ops, precursor flags, priming windows)
  • Readiness channels derived from structural organization metrics
  • Storm-level structural timelines and advisory-aligned metrics
  • Version-tagged scoring tied to PTEM model revisions

This dataset is frozen for auditability and forms the foundation for all evaluation and early-stage integrations.

Governed structural sequences

Historical structural outputs delivered through the SDK so teams can:

  • Retrieve structural snapshots for any advisory in the frozen dataset
  • Analyze readiness regimes and precursor patterns
  • Perform RI stress tests and scenario analysis

For pilot engagements, PTEM can optionally run structural snapshots on current storms by mutual agreement, without implying a full operational live feed.

Enterprise SDK

The SDK is a delivery surface, not the PTEM instrument itself.

The sole integration surface for PTEM signals, providing:

  • Authentication and organizational access control
  • Replay-safe retrieval of structural snapshots and channels
  • Convenient access to governed structural metrics and readiness channels (e.g., HC_ops, PL, composite indices)
  • Built-in provenance, versioning, and governance metadata

Example

from ptem import PTEMClient

client = PTEMClient.from_env()

# Historical structural snapshot
snapshot = client.snapshots.get(
    storm_id="2024181N09320",
    dtg="2024-06-29T00:00:00Z",
)

# Readiness channel
ppi = client.channels.get(
    storm_id="2024181N09320",
    channel="ppi",
    dtg="2024-06-29T00:00:00Z",
)

print(snapshot.summary)
print("PPI:", ppi)

What PTEM measures

PTEM quantifies structural organization — how internally coherent a storm is — rather than intensity alone. Structural readiness matters because storms with similar intensities often behave differently depending on their internal structure.

PTEM’s metrics provide:

  • Lead-time context before intensity changes materialize
  • Regime identification for operational triage and research
  • Structural baselines for scenario modeling and model comparison
  • Audit-ready signals tied to a frozen validation harness

These signals complement, rather than replace, existing hazard or physics models.

Validation & governance

All PTEM model releases are validated against a frozen multi-decadal Atlantic harness (1851–2024) with:

  • Forward-only scoring for reproducibility
  • Per-lead and per-regime diagnostics
  • Versioned validation notes for every engine revision
  • Governance gates and suppressions to prevent uncontrolled drift

Enterprise clients receive structured audit packs suitable for CROs, regulators, and modeling leads.

Who PTEM is for

PTEM supports institutions that require governed structural visibility into storm evolution:

  • Reinsurers & carriers — portfolio stress testing, event triage, scenario evaluation
  • Public programs — structural readiness insights for planning and communication
  • Modeling teams — add structural context to hazard or ML pipelines
  • Research groups — benchmark against frozen structural baselines

Access tiers

PTEM access is licensed through structured enterprise tiers. See the Pricing page for details on evaluation access (frozen dataset + SDK), internal-signal tiers, and governed operational feed options.