ZIP-0020: Impact Metric Oracle

Abstract

This proposal defines an on-chain oracle system for ingesting, verifying, and publishing conservation impact metrics on Zoo Network. The oracle aggregates data from field sensors, satellite imagery, partner NGOs, and citizen science platforms, producing standardized impact scores that can be consumed by smart contracts for reward distribution, grant evaluation, and donor reporting.

Motivation

Conservation funding is plagued by difficulty in measuring outcomes:

Accountability: Donors and governance need verifiable proof that funds create real impact
Incentive alignment: Impact-based reward distribution creates feedback loops that optimize for outcomes
Grant evaluation: On-chain impact data enables objective evaluation of grant-funded projects
Regulatory compliance: 501(c)(3) organizations must demonstrate charitable purpose; quantified impact satisfies this
Composability: Other protocols and DApps can build on standardized impact data

Specification

Oracle Architecture

┌─────────────────────────────────────────────────┐
│                Data Sources                      │
├──────────┬────────────┬────────────┬────────────┤
│ Satellite│ IoT Sensors│ NGO Reports│ Citizen Sci│
│ (MODIS,  │ (camera    │ (IUCN,     │ (iNaturali-│
│  Landsat)│  traps,    │  WWF,      │  st, eBird)│
│          │  acoustic) │  WCS)      │            │
└────┬─────┴─────┬──────┴──────┬─────┴──────┬─────┘
     └───────────┴──────┬──────┴────────────┘
                        ▼
              ┌─────────────────┐
              │  Oracle Network  │
              │  (5+ operators)  │
              ├─────────────────┤
              │  Aggregation    │
              │  Verification   │
              │  Consensus      │
              └────────┬────────┘
                       ▼
              ┌─────────────────┐
              │  Zoo Network    │
              │  Impact Registry│
              └─────────────────┘

Impact Metric Types

Metric ID	Name	Unit	Source	Update Frequency
IMP-001	Species Population	Count estimate	Camera traps, surveys	Monthly
IMP-002	Habitat Area Protected	Hectares	Satellite imagery	Quarterly
IMP-003	Deforestation Prevented	Hectares	MODIS/Landsat diff	Monthly
IMP-004	Species Sightings	Count	Citizen science	Weekly
IMP-005	Water Quality Index	0-100 score	IoT sensors	Daily
IMP-006	Carbon Sequestered	Tonnes CO2e	Remote sensing + models	Quarterly
IMP-007	Anti-Poaching Patrols	Km covered	GPS trackers	Weekly
IMP-008	Community Engagement	Participants	Platform analytics	Monthly

Oracle Operator Requirements

operator:
  stake: 50000 ZOO          # Minimum stake to operate
  slashing:
    deviation_threshold: 20%  # Max deviation from median
    penalty: 10%              # Of staked amount
    cooldown: 30 days         # Before re-entry after slash
  rewards:
    per_report: 100 ZOO      # Per accepted data submission
    bonus_accuracy: 1.5x     # For consistently accurate reporters
  minimum_operators: 5        # Quorum for data acceptance
  consensus: median           # Aggregation method

Impact Registry Contract

// SPDX-License-Identifier: MIT
pragma solidity ^0.8.20;

contract ImpactOracle {
    struct ImpactReport {
        bytes32 metricId;
        uint256 value;
        uint256 timestamp;
        bytes32 evidenceHash;  // IPFS hash of supporting data
        uint8 confidence;      // 0-100 confidence score
    }

    mapping(bytes32 => ImpactReport) public latestMetrics;
    mapping(bytes32 => ImpactReport[]) public metricHistory;

    event ImpactRecorded(
        bytes32 indexed metricId,
        uint256 value,
        uint8 confidence,
        bytes32 evidenceHash
    );

    function submitReport(
        bytes32 metricId,
        uint256 value,
        bytes32 evidenceHash,
        uint8 confidence
    ) external onlyOperator {
        // Aggregate with other operator submissions
        // Accept when quorum reached
    }

    function getImpactScore(bytes32 projectId) external view returns (uint256) {
        // Compute weighted composite score across all metrics for a project
    }
}

Evidence Layer

All raw evidence is stored on IPFS/Arweave with content-addressed hashes recorded on-chain:

Satellite images: GeoTIFF files with metadata
Sensor readings: JSON time-series data
Survey reports: PDF documents signed by field researchers
Photo evidence: Geotagged camera trap images

Composite Impact Score

Projects receive a composite score (0-1000) calculated as:

score = sum(metric_weight[i] * normalized_value[i] * confidence[i]) / sum(metric_weight[i])

Default weights: Species Population (30%), Habitat Protected (25%), Deforestation Prevented (20%), Community Engagement (15%), Carbon (10%).

Rationale

A median-based consensus mechanism is chosen over mean to resist outlier manipulation. The 5-operator minimum quorum ensures sufficient redundancy while keeping operational costs manageable. Slashing for deviation exceeding 20% from median prevents lazy or malicious operators from corrupting data.

The evidence layer on IPFS provides verifiable proof while keeping on-chain storage costs minimal. Only hashes and aggregated metrics live on-chain; full data is retrievable off-chain.

Security Considerations

Data manipulation: Multiple independent operators with different data sources reduce the risk of coordinated manipulation
Stale data: Metrics have defined update frequencies; consumers should check timestamps before relying on values
Sybil resistance: Operator stake requirement (50,000 ZOO) makes Sybil attacks expensive
Oracle front-running: Report submissions are committed in a commit-reveal scheme to prevent copying
Data source compromise: If a primary data source is compromised, governance can update source weights and add alternative sources
Privacy: GPS coordinates of endangered species are aggregated to grid cells (10km resolution) to prevent poaching

Impact Metric Oracle