Space Sherpa COLA

What is SpaceSherpa Cola?

SpaceSherpa Cola is a decision cockpit for orbital conjunction assessment that extends CelesTrak's trusted data infrastructure into behavioral analytics, econometric insight, cross-source comparison, and evidence-based policy support.

Unlike tools that rely on a single collision probability (Pc) metric, COLA V4.1 uses the EVDT multi-factor triage engine to score conjunction events across four dimensions: Environment, Vulnerability, Decision urgency, and Technology confidence. This approach, grounded in Es haghi et al. (2024) and validated against six systematic literature reviews, catches high-consequence events that Pc-only screening would deprioritize.

Getting Started

Audience Modes

The header contains three audience buttons that control what you see:

Mode	Designed For	What Changes
Executive	Leadership, sponsors, decision authorities	Reduces card density, hides technical detail panels, shows only summary metrics and top-level risk
Analyst	SSA operators, conjunction analysts, engineers	Full detail: all cards, pipeline health, operator scorecards, model assumptions, EVDT factor breakdowns
Policy	Regulatory analysts, governance researchers	Shows policy levers with burden/benefit analysis, hides low-level pipeline data, emphasizes intervention scenarios

Keyboard Shortcuts

Key	Action
1	Jump to Monitor view
2	Jump to Assess view
3	Jump to Decide view
4	Jump to Explain view
5	Jump to Cross-Source view
6	Jump to Behavioral view
7	Jump to Globe view
8	Jump to SpaceMapper Intel view
Esc	Close any open panel (provenance drawer, this guide)

The EVDT Triage Engine

The core innovation of V4.1. Instead of ranking conjunctions by Pc alone, every event receives a composite score from four weighted dimensions:

Dimension	Weight	What It Measures	Key Inputs
Environment	0.35	How threatening is the physical conjunction geometry?	Collision probability (Pc), miss distance, relative velocity, covariance realism, screening volume overlap
Vulnerability	0.25	How consequential is a collision for the primary asset?	Crew presence, asset replaceability, constellation criticality, debris-generating potential, orbital regime congestion
Decision	0.25	How urgent is the decision timeline?	Time to TCA, maneuver window remaining, operator responsiveness history, coordination complexity
Technology	0.15	How confident are we in the underlying data?	Covariance age, propagation model quality, number of corroborating sources, atmospheric uncertainty

EVDT composite thresholds are configurable in Settings:

• CRITICAL score >= 0.80
• HIGH score >= 0.60
• MODERATE score >= 0.40
• LOW score < 0.40

Each conjunction in the Monitor table displays a mini radar chart showing the four EVDT dimensions, plus horizontal factor bars for quick visual comparison.

View-by-View Guide

📡 Monitor

Your primary operational view. Shows the current conjunction environment ranked by EVDT composite score (not Pc alone).

• Base-Rate Context Banner (top): Combats base-rate neglect (Bruno 2025) by showing the ratio of CRITICAL/HIGH events to total CDMs processed. Includes the historical false-alarm rate at the current threshold so operators calibrate their response appropriately.
• Top Risk Card: The highest EVDT-scored conjunction with full radar chart, factor bars, epistemic uncertainty flag, Pc, miss distance, TCA, and regime. Click to open the provenance drawer.
• Conjunction Queue: All active events sorted by EVDT composite. Columns include EVDT score badge, triage class, primary/secondary objects, TCA, miss distance, Pc, regime, epistemic uncertainty level, and responsible operator. Click "Sources" to jump to the Cross-Source view for any event.
• Data Pipeline Health (Analyst only): Real-time status of data providers (CelesTrak, Space-Track, LeoLabs, EU SST, TraCSS) with catalog sizes, latency, and last-update timestamps.
• Maneuver Decisions Pending: Queued maneuver recommendations with delta-V, fuel cost, deadline, and decision status.

📊 Assess

Economic exposure and environmental trend analysis.

• Economic Exposure Cards: Direct conjunction response cost, deferred congestion NPV, mission-life reduction cost, GDP exposure, and decision-velocity dividend. Each card is clickable for full provenance (model, assumptions, caveats, confidence level).
• Orbital Carrying Capacity: Utilization bars by shell (Dense LEO, Constellation Belt, SSO, MEO/GEO) with saturation year projections and sustainability indices.
• Operator Compliance Scorecards (Analyst only): CDM acknowledgment rates, average response times, data quality scores, and letter grades per operator.
• Normative Influence Metrics (Analyst only): Global averages for CDM acknowledgment, response time, voluntary sharing, and coordination effectiveness.

⚖️ Decide

Policy intervention testing and scenario comparison. Grounded in Rao et al. (2020) orbital-use fees, Muller (2022) Pigouvian debris tax, and Guyot-Rouillon (2023) optimal management frameworks.

• Policy Levers: Four instruments (mandatory sharing, orbital-use fee, debris tax, cap-and-trade) with expected conjunction reduction, implementation cost, timeline, operator burden, political feasibility, and burden/benefit distribution.
• Scenario Comparison: Three scenarios (Status Quo, Mandatory Sharing + Debris Tax, Ostrom Polycentric Framework) with projected conjunction reduction, cost, risk trajectory, and time horizon.

Each lever and scenario has an epistemic state chip indicating the confidence level. Click "View provenance" for the full assumption chain.

🔍 Explain

Full transparency layer. Every recommendation, finding, and model output is inspectable here.

• AI Findings: System-generated findings with structured fields: Why It Matters, Evidence Basis, Alternatives Considered, Recommended Action, and Expected Effect. Each links to its source module.
• Stated Assumptions (V4.1 new): Six explicit assumptions flagged from the literature review synthesis. Each shows status (UNTESTED, PARTIALLY_VALIDATED, TRANSFER_ASSUMPTION, ACKNOWLEDGED) and the architecture layer it belongs to. These must be validated empirically before the system treats them as established.
• Novel Research Contributions (V4.1 new): Five areas where SpaceSherpa Cola produces genuinely novel empirical baselines for the community, with current status tracking (IN_PROGRESS, PROTOTYPE, PLANNED).
• Model Assumptions Register (Analyst only): Key assumptions for each computational model (EVDT Triage Engine, Behavioral Proxy Model, Economic Exposure Model).
• Data Mode Legend: Definitions for all epistemic state labels (OBSERVED, DERIVED, FORECAST, RECOMMENDATION, SIMULATED, PROXY_OBSERVABLE, PROXY_INFERRED).

🔎 Cross-Source Comparison (V4.1 New)

When two providers disagree on a conjunction assessment, this view explains why and what to do about it. Based on BACARDI/W3C PROV provenance patterns.

• Source Calibration Table: Five providers (US 18th SDS, LeoLabs, EU SST, Russian ASPOS, Chinese SpaceMapper) with catalog sizes, update cadence, accuracy estimates, and operational status.
• Agreement Dashboard: Overall cross-provider agreement rate, the Weigel (2024) finding that 20-30% of disagreement is reducible via consider-covariance optimization, and count of active disagreement events.
• Disagreement Spotlight: For each materially disagreeing event, shows both assessments side by side, the root cause of the disagreement, and an actionable recommendation for which assessment to prefer and why.

🧠 Behavioral Analytics (V4.1 New)

Operator behavioral proxies derived from publicly observable data. This layer would produce the first structured behavioral dataset for orbital safety (no published precedent exists).

• Behavioral Proxy Indicators: Four indicators (Ephemeris Sharing Frequency, Conjunction Response Timing, Maneuver Compliance Rate, Post-Event Miss Improvement) with measurement class tags and literature references.
• Operator Behavioral Profiles: Table with sparkline trend charts showing 12-period rolling trends for sharing frequency and response timing per operator. Includes maneuver compliance rates, miss improvement metrics, and letter grades.

🌍 Globe

3D orbital visualization placeholder for CesiumJS integration. Requires a Cesium ION token (configure in Settings). When connected, renders real-time orbital tracks, conjunction vectors, covariance ellipsoids, and ride-along camera modes.

🛡️ SpaceMapper Intelligence

Analysis of China's commercial space surveillance platform (85,151 records, OSINT extraction). Shows surveillance asymmetry, catalog parity tracking, Chinese constellation mapping, the 14+ platform ecosystem, and strategic assessment of where the US leads, where China is close, and where China is ahead.

⚙️ Settings

System configuration including EVDT weights and triage thresholds, legacy Pc thresholds, display preferences (epistemic chips, provenance drawers, animations, color blind mode), data mode, refresh intervals, and Cesium ION token configuration.

Understanding Epistemic State Labels

Every data element in COLA carries an epistemic state chip showing how the data was produced:

Label	Meaning	Confidence Implication
OBSERVED	Direct measurement from sensor data or official catalog	Highest confidence
DERIVED	Computed from observed data using established models	Confidence depends on model quality
FORECAST	Projected using simulation or statistical models	Lower confidence; stated assumptions apply
RECOMMENDATION	System-generated action suggestion	Always requires human decision authority
SIMULATED	Generated for demonstration	Not based on live operational data
EPISTEMIC	Uncertainty from incomplete knowledge (reducible with more data)	Indicates where additional observations would improve the assessment

Additionally, V4.1 introduces PROXY_OBSERVABLE and PROXY_INFERRED labels for behavioral indicators, distinguishing data directly observable from public sources from data inferred through indirect evidence.

Provenance Drawers

Click any card with a provenance link (or any card marked "clickable") to open the provenance drawer on the right side. The drawer shows:

• Category: The epistemic state of the data
• Source Type: Where the data originated
• Model: Which computational model produced the output
• Confidence: HIGH, MEDIUM, or LOW
• Epistemic Uncertainty: Level of knowledge-gap uncertainty
• Aleatory Estimate: Irreducible randomness component
• Assumptions: What the model assumes
• Caveats: Known limitations
• EVDT Score: Full four-dimension breakdown (for conjunction events)

The separation of epistemic (reducible with more data) vs. aleatory (irreducible randomness) uncertainty follows Sanchez and Vasile (2022) Dempster-Shafer evidence theory.

Data Architecture

COLA V4.1 uses a JSON-driven data architecture. The application loads data from an external spacesherpa-cola-v4.1-data.json file. If the file is not found, embedded demo data provides a complete simulation for demonstration purposes.

The embedded data includes 8 conjunction events, 6 operator behavioral profiles, 5 source providers, 4 policy levers, 3 scenarios, 3 AI findings, 6 stated assumptions, and 5 novel contribution trackers.

The underlying architecture follows a six-layer design:

1. Layer 0: Source Adapters - Ingest from CelesTrak, Space-Track, LeoLabs, EU SST, TraCSS, commercial/foreign sources
2. Layer 1: Canonical Schema - Normalized data model (SpaceObject, Operator, Ephemeris, ConjunctionEvent, SourceRecord, BehaviorIndicator)
3. Layer 2: Decision Logic - EVDT triage engine, uncertainty separation, triage classification
4. Layer 3: Analytics - Behavioral proxy computation, econometric models, cross-source comparison
5. Layer 4: Policy Reasoning - Evidence-to-policy pipeline with traceable assumption chains
6. Layer 5: Presentation - This cockpit interface with audience-adaptive output

Literature Foundation

V4.1 integrates findings from six systematic literature reviews conducted April 2026:

#	Topic	Key Finding for COLA
1	Multi-Factor Conjunction Triage	EVDT framework (Es haghi et al. 2024) validates multi-factor approach. Alfano/Oltrogge/Arona (2022) quantifies operator heterogeneity in Go/No-Go criteria.
2	Behavioral Determinants of Orbital Safety	No published studies directly measure operator response timing. Behavioral analytics layer produces genuinely novel empirical baselines.
3	Econometric Methods for Orbital Safety	Full applied econometrics toolkit (DiD, IV, panel FE, RD) has never been applied to orbital safety outcomes. COLA's econometric layer is the first structured application.
4	Multi-Source SSA Comparison	Cross-source reconciliation methods are ad hoc. No standardized disagreement interpretation framework exists. Weigel (2024) shows 20-30% disagreement is reconcilable.
5	Evidence-Based Space Policy	Policy proposals lack empirical grounding. No quasi-experimental evaluation of a space safety policy has been published.
6	Uncertainty Communication	Base-rate neglect, probability weighting, and threshold effects are documented in cognitive science. Maritime/aviation design patterns are transferable.

Current Limitations

• TraCSS API access is pending; the TraCSS adapter is not operational
• CesiumJS globe requires ION token configuration
• Behavioral elasticities are assumed, not empirically validated
• Chinese SSA capabilities are estimated from OSINT; accuracy assumptions are not possible
• Policy behavioral response estimates use transfer assumptions from analogous domains (environmental economics, aviation safety)

Version History

Version	Date	Changes
V1.0	Feb 2026	Initial Vercel deployment with basic conjunction display
V2.0	Mar 2026	Local deployment, enhanced data model, CesiumJS placeholder
V3.0	Mar 2026	Multi-view architecture, audience modes, epistemic chips
V4.0	Apr 2026	Decision cockpit with Assess/Decide/Explain workflow, SpaceMapper Intelligence, provenance drawers
V4.1	Apr 2026	EVDT triage engine, behavioral analytics, cross-source comparison, stated assumptions register, novel contributions tracker, base-rate context banner, lit-review integration

Credits

Author: Dr. Nathaniel Dailey, FRAeS

Organization: MITRE Center for Government Effectiveness and Modernization (MPS/CGEM/Commerce)

Data Foundation: CelesTrak (T.S. Kelso), Space-Track.org, LeoLabs, EU SST

Literature Basis: Six systematic reviews via AnswerThis.io, synthesized April 2026

© 2026 The MITRE Corporation. All rights reserved.