Scorecards
Versioned scorecards and per-storm summaries that partners use to validate and audit engine behavior.
Structural intelligence
Governed structural insight for hurricane risk
PTEM operates a governed structural analysis engine paired with a frozen, multi-season dataset. Risk teams use it to benchmark scenarios, evaluate readiness regimes, and add a validated structural view alongside traditional hazard models.
At a published reference operating point (p80) on the frozen Atlantic v1.2 harness, PTEM shows 36h median lead time on true positives (mean 50h), with 0.396 precision / 0.205 recall under forward-only evaluation.
Licensed only to institutional risk owners, reinsurers, and public programs that require governed audit trails.
Validation surface
Frozen dataset, governed delivery
Historical archives, structural snapshots, and audit notes ship with every tier so partners can reproduce PTEM scoring.
- Versioned validation pack with every agreement
- Structural signals at a 6-hour cadence
- SDK integration support for risk workflows
Why PTEM?
PTEM offers a rigorously validated structural view of storm evolution, drawn from a frozen, multi-decade dataset and a governed analytical engine. Risk teams use PTEM to benchmark scenarios, evaluate structural regimes, and integrate a trusted second perspective into existing models.
Canonical validation is run on the Atlantic v1.2 frozen record (1851–2024) across declared operating points, with the reference operating point recorded in run manifests and validation receipts. See /validation for the canonical run summary and illustrative sensitivity sweeps, which do not tune or select operating points.
Under the hood, PTEM tracks structural coherence, phase alignment, and priming indices over time and looks for sustained tightening patterns that have historically preceded major regime shifts.
The result is an orthogonal signal: a structural engine that can flag when a storm is entering a high-readiness regime before intensity-based thresholds are met.
Snapshot
A compact view of the PTEM hurricane engine
A frozen structural dataset, a reproducible engine, and a narrow surface area for integration.
Structural snapshots
Compact structural snapshots at a 6-hour cadence so teams can integrate PTEM signals without exposing internal model fits.
Governed access
Access tiers that separate evaluation/research from production-grade enterprise integrations with governance and audit support.
How PTEM works
A structural engine on a frozen, auditable dataset
PTEM runs on a frozen, multi-season tropical cyclone dataset. The engine produces governed structural signals that can be audited, compared against baselines, and blended into existing risk workflows.
Step 1
Frozen dataset
PTEM is anchored to a multi-season dataset so partners can reproduce scoring and compare model snapshots fairly.
Step 2
Structural engine & governance
PTEM extracts structural signals, applies governance layers, and delivers versioned outputs for audit and integration.
Step 3
Integration & validation
Partners start with evaluation materials and can progress to enterprise integrations supported by validation packs and governance processes.
SDK integration
PTEM is delivered through a governed SDK that fronts the structural engine. Risk teams pull structural metrics and readiness channels through that single surface and wire them into existing hazard workflows, internal tools, or research environments.
{
"storm_id": "al152025",
"dtg": "2025-09-29T18:00:00Z",
"hc_struct": 0.81,
"fme_struct": 0.72,
"pl_struct": 0.67,
"dc_struct": 0.03,
"ppi_live_norm": 0.92
}This example shows a single structural snapshot for a hypothetical storm: a normalized structural coherence score, energy-like change metrics, phase alignment, structural trend, and an aggregate readiness measure. The actual schema is versioned and documented in the SDK.
PTEM operates as a managed, trade-secret-protected system. The SDK carries version tags for both dataset and model revisions so downstream systems always know what they are consuming.
Key features
Structural Coherence Engine
Captures changes in a storm’s internal organization using a governed set of structural metrics. Designed to complement—not replace—existing hazard models.
Governed readiness channels
Composite structural readiness channels derived from storm organization, symmetry, and burst patterns. Highlight periods where storms may be structurally receptive to rapid changes.
Enterprise SDK
Python SDK client for seamless integration into insurance, reinsurance, or research modeling workflows.
Frozen Validation Harness
A forward-only, 6-hour aligned evaluation harness executed across the full frozen Atlantic structural spine (1851–2024), enabling reproducible scoring, sensitivity analysis, and audit-safe validation.
Who uses PTEM
Built for institutional risk owners
PTEM is designed for reinsurers, carriers, public programs, and modeling teams that need structural visibility into storm evolution.
Reinsurers & carriers
Use structural signals to stress-test portfolios, examine alternative scenarios, and compare PTEM behaviour against existing cat models.
Public agencies & programs
Support planning and communications with a structural view on storm organisation, without replacing official forecasts.
Research & modeling teams
Use the frozen dataset and structural outputs as a benchmark for new models, or as a complementary signal inside broader risk experiments.