CanopyProof

CanopyProof helps forest restoration teams estimate defensible carbon credits using IPCC science, live Earth-risk data, transparent deductions, and integrity evidence—before costly validation.

Comment

Inspiration

Nature-based carbon finance could fund restoration at enormous scale, yet the earliest project decisions are often made inside opaque spreadsheets.

A headline estimate may count all forest growth as additional while ignoring the baseline, leakage, uncertainty, reversal risk, community safeguards, or double counting. Small landowners, community forests, and local NGOs may spend significant resources on validation before discovering that the project has fundamental evidence gaps.

We built CanopyProof to make the first carbon assessment more honest and accessible: expose every assumption, deduct every material risk, and show exactly what still needs to be proven.

What it does

CanopyProof is a carbon MRV preflight for forest restoration and afforestation/reforestation projects.

A user selects any project location, then enters:

  • project area and crediting period;
  • an IPCC forest stratum;
  • baseline biomass growth;
  • annual project emissions;
  • leakage and uncertainty assumptions;
  • six evidence declarations covering baseline, tenure, community consent, monitoring, leakage, and registry commitment.

CanopyProof then:

  1. Estimates above- and below-ground biomass CO₂ removals using official IPCC defaults.
  2. Subtracts the counterfactual baseline and project emissions.
  3. Applies visible leakage and uncertainty deductions.
  4. Uses live climate and natural-event context to recommend a bounded permanence buffer.
  5. Screens additionality, robust quantification, permanence, no double counting, leakage, and safeguards.
  6. Uses live GBIF occurrences to show biodiversity context without presenting records as population estimates.
  7. Generates a SHA-256 fingerprint binding the normalized inputs, model version, source status, and retrieval time.
  8. Exports the complete assessment as a machine-readable evidence pack.

The output is deliberately labeled a pre-feasibility estimate, never an issued or verified carbon credit.

For the default Cát Tiên restoration scenario, our tested run estimated 87,826 retained credits over 20 years, after leakage, uncertainty, project emissions, and permanence deductions. The purpose is not the headline number—it is making every tonne traceable and challengeable.

How we built it

CanopyProof is built as a responsive React and TypeScript decision workspace with a same-origin edge data-fusion API.

The CP-MRV 1.0 engine uses the transparent equation:

Area × (project growth − baseline growth) × (1 + root:shoot) × 0.47 × 44/12 × years − emissions − leakage − uncertainty − permanence buffer

Above-ground growth factors and root-to-shoot ratios come from Tables 4.9 and 4.4 of the 2019 Refinement to the 2006 IPCC Guidelines, AFOLU Volume 4. The selected factor and source table remain visible in the interface. No AI model invents a biomass value.

The backend validates inputs and concurrently queries:

  • Open-Meteo for current and seven-day heat, rainfall, evapotranspiration, and gust context;
  • GBIF for biodiversity occurrence context within the project window;
  • NASA EONET for open wildfire and natural-event context;
  • OpenStreetMap for project-location context.

Optional source failures remain visible instead of being replaced by mock values.

The integrity screen is structured around six concepts found in the ICVCM Core Carbon Principles. A user declaration may improve preflight readiness, but CanopyProof never presents it as verified documentary evidence.

Challenges we ran into

Translating inventory science into a project workflow

IPCC Tier 1 defaults are designed for greenhouse-gas inventories, not automatic credit issuance. We had to make them useful for early project screening without overstating their precision or authority.

Keeping every deduction understandable

Leakage, uncertainty, project emissions, and buffer deductions must not disappear inside one final number. We implemented a sequential credit waterfall so users can see which tonnes were removed and why.

Using live risk data responsibly

A seven-day forecast cannot replace a registry non-permanence risk tool. We use live environmental data only as contextual screening and clearly label the resulting buffer as a CanopyProof recommendation.

Avoiding biodiversity overclaims

GBIF occurrence density reflects reporting activity as well as biodiversity. We display it as contextual evidence, never as a population count or verified co-benefit claim.

Designing graceful partial failure

Environmental APIs have different geographic coverage and availability. CanopyProof preserves missing states and source health instead of fabricating convenient demo results.

Accomplishments that we're proud of

  • A working global preflight for any searchable project location.
  • A traceable IPCC factor library with visible source tables.
  • Live Open-Meteo, GBIF, and NASA EONET integration.
  • No API keys, user account, personal data, or mocked environmental results.
  • A transparent credit waterfall and retained-removal projection.
  • Six-dimensional integrity screening with prioritized evidence gaps.
  • A climate-contextual permanence-buffer recommendation.
  • SHA-256 evidence fingerprints for reproducible assessment runs.
  • A downloadable JSON evidence pack.
  • Responsive, accessible design for desktop and mobile.
  • Clear boundaries preventing estimates from being mistaken for issued credits.

What we learned

Carbon credibility depends less on making a model appear sophisticated and more on making its boundaries visible.

Baseline, leakage, permanence, safeguards, and double counting are not footnotes. They are core product requirements.

We also learned that transparency can become an interaction model. A project developer can change one assumption, rerun the assessment, and immediately see which tonnes moved, why the integrity score changed, and which evidence should be collected next.

What's next for CanopyProof

  1. Add project polygons and satellite-derived land-cover history.
  2. Support evidence-document upload and verification instead of declarations alone.
  3. Match projects with registry-approved methodologies.
  4. Replace Tier 1 defaults with field-plot and remote-sensing measurements.
  5. Add monitoring-period comparison and versioned evidence packs.
  6. Support jurisdictional nesting and registry serialisation workflows.
  7. Add multilingual community review, FPIC, and benefit-sharing records.
  8. Pilot the workflow with community forestry organizations, project developers, and validators.

CanopyProof does not manufacture credits. It helps credible restoration projects make every tonne defensible.

Built With

Share this project:

Updates

Submission history