Scientific Background

Wreck Environmental Risk Prioritisation (WERP): Applying Threat Intelligence Principles to Potentially Polluting Wrecks

Document ID: Deeptrek-WP-2025-01 • Date: October 23, 2025 • Version: 1.3

Back to Publications

Abstract

The world's oceans contain over 8,500 potentially polluting wrecks (PPWs), posing a significant and accelerating environmental threat. This paper introduces the Wreck Environmental Risk Prioritisation (WERP) protocol, a novel methodology applying established threat intelligence principles to systematically identify, assess, and prioritise these wrecks for monitoring and potential remediation. WERP leverages a multi-phase approach, beginning with rapid OSINT (Open-Source Intelligence) assessment accelerated by Large Language Models (LLMs), followed by expert review and optional in-situ data integration. It quantifies risk using scientifically grounded metrics: Wreck Condition Score (WCS), Pollutant Hazard Score (PHS), Environmental Sensitivity Index (ESI), and a dynamic Release Probability Modifier (RPM). Continuous monitoring for seismic and storm-related events further refines risk assessment over time. WERP provides a scalable, consistent, and scientifically defensible framework to address the global challenge of PPWs, enabling an intelligence-led scientific response to a critical environmental problem.

1. Introduction: The Unseen Environmental Threat

Millions of tons of oil and hazardous materials lie within thousands of sunken vessels worldwide, primarily casualties of 20th-century conflicts (Monfils et al., 2022; IUCN, 2025). Estimates suggest over 8,500 wrecks pose a significant pollution potential (IUCN, 2025). As these wrecks corrode, their structural integrity degrades, exacerbated by environmental factors like warming waters, increased storm frequency, and ocean acidification (Landquist et al., 2013; Van Landuyt et al., 2022). The potential for catastrophic release of pollutants presents a severe threat to marine ecosystems, coastal economies, and cultural heritage (ICOMOS & IUCN, 2024).

The sheer scale of the problem necessitates a structured, prioritised approach. Traditional methods often rely on reactive responses after a leak occurs or slow, resource-intensive surveys. This paper proposes the Wreck Environmental Risk Prioritisation (WERP) protocol, adapting the cyclical, iterative principles of threat intelligence (CIA, n.d.) to create a proactive, intelligence-led scientific framework for managing PPWs. Just as threat intelligence helps organizations anticipate and mitigate cyber or security threats based on incomplete data, WERP enables prioritisation and action on PPWs using available information, creating momentum for targeted scientific investigation and data refinement.

2. The WERP Protocol: An Intelligence-Led Scientific Framework

The WERP protocol mirrors the classic intelligence cycle (Direction, Collection, Processing, Analysis, Dissemination, Feedback) to manage the assessment of PPWs. It is designed to handle large volumes of data, make informed inferences where data is incomplete, and iteratively refine assessments as new information becomes available.

The protocol consists of distinct phases:

2.1. Phase 1: Initial OSINT Assessment & Triage

This phase leverages Open-Source Intelligence (OSINT) gathering, significantly accelerated by Large Language Models (LLMs), to conduct rapid initial assessments. Historical records, maritime databases, academic papers, and news archives are synthesized to estimate the core WERP scores (WCS, PHS, ESI, RPM) based on the WERP Phase 1 Assessment Definition (Deeptrek Ltd., 2025a). Key parameters like structural integrity, precise coordinates, and environmental specifics are often inferred based on established rules (e.g., inferring collapse based on age and sinking trauma). An expert maritime researcher reviews the LLM-generated assessment, validating sources, checking score calculations, and applying critical judgment before finalizing the Phase 1 score. This phase allows for rapid global screening, identifying high-priority candidates for further investigation from thousands of potential wrecks.

2.2. Phase 2: Expert-Driven In-Situ Reassessment

For high-priority wrecks identified in Phase 1, or when new survey data becomes available, a Phase 2 Reassessment is conducted according to the WERP Phase 2 Assessment Definition (Deeptrek Ltd., 2025b). This phase integrates high-fidelity, ground-truth data obtained from direct surveys (e.g., ROV, AUV, diver reports), including precise coordinates, depth, observed structural condition, seabed type, and potentially water temperature and pH readings. This in-situ data supersedes Phase 1 inferences, leading to a more accurate recalculation of WERP scores, particularly the WCS and RPM. The provenance and confidence levels are upgraded to reflect the higher data quality. The Amakasu Maru No. 1 case study provides an example of this process (Deeptrek Ltd., 2025d).

2.3. Phase 3: Continuous Monitoring & Dynamic Reassessment

WERP incorporates a dynamic monitoring component. Wreck locations are continuously cross-referenced with real-time environmental data feeds:

  • Seismic Activity: USGS Earthquake Catalog data is monitored for significant events (Magnitude 4.5+). Peak Ground Acceleration (PGA) is estimated at each wreck location. Events exceeding PGA > 0.05g trigger high-priority reassessment alerts, while PGA > 0.02g logs low-priority events. This assesses the potential for sudden structural damage (World Nuclear Association, 2021).
  • Tropical Storms: Active cyclone tracks are monitored. If a storm passes within 200 nm, local wave forecasts (e.g., WW3 data) are used to calculate the Bottom Orbital Velocity (Ub) at the wreck's depth. Ub exceeding seabed-specific scouring thresholds (e.g., 0.15–0.50 m/s) generates a low-priority alert. Ub exceeding structure-specific integrity thresholds (e.g., 0.50–0.80 m/s based on Phase 2 condition) triggers a high-priority reassessment alert (Liu et al., 2022).

These alerts feed back into the cycle, potentially triggering a need for a new Phase 2 survey or adjusting the RPM score based on recent stress events.

3. WERP Scoring Metrics: Quantifying Risk

WERP employs four core metrics, calculated using specific parameters and weightings derived from scientific principles and expert judgment.

3.1. Wreck Condition Score (WCS)

Assesses the physical state and potential for pollutant retention (Scale: 0-20).

  • Parameters (0-5 each):
    • Age: Time since sinking (proxy for corrosion duration).
    • Vessel Type/Size: Intrinsic capacity for pollutants.
    • Sinking Trauma: Likelihood of initial pollutant loss and structural damage.
    • Current Structural Integrity: Observed or inferred state (intact to collapsed). Phase 1 infers this based on Age and Trauma; Phase 2 uses direct observation.
  • Formula: WCStotal = ∑ WCSparam

3.2. Pollutant Hazard Score (PHS)

Quantifies the environmental threat posed by the wreck's contents (Scale: 0-10).

  • Parameters (Score 0-10, Weighted %): Assesses specific pollutant types (e.g., Fuel Oil, Munitions, Hazardous Cargo, Other Pollutants like heavy metals, POPs).
  • Weighting: The WERP protocol defines specific weight percentages based on pollutant type, prioritising persistent, high-volume pollutants (e.g., heavy fuel oil ~60-70%) over less persistent or lower volume ones (e.g., munitions ~10-20%, other ~10%). These weights reflect established environmental impact profiles (Landquist et al., 2013; Van Landuyt et al., 2022).
  • Formula: PHStotal = ∑ (PHSscore × PHSweight) / 100

3.3. Environmental Sensitivity Index (ESI)

Quantifies the vulnerability of the surrounding environment (Scale: 0-40). Based conceptually on NOAA's ESI methodology (Gundlach & Hayes, 1978; Scott et al., 2013).

  • Parameters (0-10 each):
    • Proximity to Sensitive Ecosystems: MPAs, coral reefs, mangroves, etc.
    • Proximity to Human Resources: Fisheries, coastlines, infrastructure.
    • Local Oceanography: Potential for pollutant transport by currents.
    • Baseline Biodiversity/Protected Species: Presence of vulnerable species/habitats.
  • Formula: ESItotal = ∑ ESIparam

3.4. Release Probability Modifier (RPM)

A dynamic multiplier reflecting environmental accelerants of degradation (Multiplier: >= 1.0).

  • Factors (Value 1.0 - 1.4+):
    • Thermal: Accelerated corrosion from ocean warming. Phase 2 uses in-situ temp (Cold water <5°C yields 1.0).
    • Physical: Stress from storms, seismicity, currents.
    • Chemical: Accelerated corrosion from ocean acidification. Phase 2 uses in-situ pH.
  • Formula: RPMfinal = 1.0 + (Thermal - 1.0) + (Physical - 1.0) + (Chemical - 1.0)

4. Final Severity Calculation

The overall risk is synthesized into a final severity score:

Severity = (WCStotal + PHStotal + (ESItotal / 3)) × RPMfinalMultiplier

This formula balances the wreck's condition, its hazardous contents, and the sensitivity of its location, while amplifying the score based on environmental stressors that increase the likelihood of release. The division of ESI by 3 provides a balanced weighting relative to WCS and PHS.

5. Conclusion: Towards Proactive Management

The WERP protocol offers a scientifically grounded, systematic, and scalable methodology for addressing the global threat of potentially polluting wrecks. By adapting proven threat intelligence principles, it enables:

  • Rapid Triage: Efficiently prioritising thousands of wrecks using OSINT and LLM acceleration.
  • Consistent Assessment: Applying standardized, scientifically-backed scoring metrics.
  • Iterative Refinement: Incorporating higher-fidelity in-situ data and dynamic monitoring.
  • Actionable Intelligence: Providing clear prioritisation for resource allocation towards targeted surveys, monitoring, or remediation.

WERP transforms the PPW challenge from an overwhelming historical burden into a manageable environmental intelligence problem. It provides the necessary framework for an intelligence-led scientific response, enabling proactive intervention before catastrophic environmental damage occurs. Further validation through broader application and integration with predictive corrosion modeling will continue to enhance the protocol's efficacy. The Project Guardian platform serves as an implementation of this protocol (Deeptrek Ltd., 2025c).

6. References