Third-Party Trust
Carbon markets work on trust. Buyers pay for VCUs because an independent, accredited auditor has confirmed that emission reductions are real, measurable, and additional. Understanding verification helps project developers prepare correctly and avoid common pitfalls that delay credit issuance.
Analogy: Financial Audit
Verification is like a company's annual financial audit. The company keeps its own books (monitoring data), but an independent CPA firm reviews them against source documents. In carbon, the project proponent keeps monitoring records, the VVB audits them, and Verra registers the result.
📍 Real Verification: Kenya KACP Project
The Kenya Agricultural Carbon Project (KACP), one of the most cited agricultural soil carbon projects in sub-Saharan Africa, completed its first verification cycle using Bureau Veritas as the VVB. The verification team spent two weeks in the field, visiting 60 of the 60,000 enrolled farmer plots. They checked GPS-recorded field boundaries, asked farmers to demonstrate their understanding of conservation agriculture practices, and cross-checked activity records against satellite imagery (Sentinel-2 bare soil index). Corrective action requests were issued for 4 farmers whose boundary coordinates had shifted, requiring GPS re-measurement before VCUs for those plots could be issued.
Validation vs. Verification, A Critical Distinction
✅ Validation (happens BEFORE the project starts)
A VVB reviews the Project Description Document (PDD) to confirm that the project design is credible, the methodology is correctly applied, the additionality case is sound, and the monitoring plan is robust.
Output: Validation Report, project is registered in the Verra registry and can begin.
Timing: Once, at project start (before crediting period begins).
Think of it as: "Is this a credible plan?"
📋 Verification (happens AFTER each monitoring period)
A VVB reviews the Monitoring Report for a completed monitoring period to confirm that actual field data supports the claimed GHG reductions. The VVB audits the numbers, visits the site, and certifies the net ERR.
Output: Verification Report + VCU issuance by Verra.
Timing: Periodically (often every 5 years); can overlap with the next monitoring period.
Think of it as: "Did the plan actually deliver?"
Can the same VVB do both?
VCS rules require independence: the VVB that validates the project cannot verify the first monitoring period for that same project. This two-body rule prevents conflicts of interest (the same auditor shouldn't approve their own prior work). After the first verification, the same or a different VVB may do subsequent verifications.
The 5-Step Verification Process
Step 1: Contract and Scoping
Project developer contracts an accredited VVB. The VVB reviews project documents to understand scope, methodology version, and key risks. A verification plan is agreed upon.
Step 2: Desktop Review
VVB reviews all submitted documents: Monitoring Report, data files, sampling records, lab reports, activity logs, and model outputs. Clarification requests (CRs) and corrective action requests (CARs) are issued for gaps.
Step 3: Site Visit
VVB auditors visit project fields to confirm boundaries, interview farmers about practices, observe monitoring procedures, and cross-check activity data against farm records.
Step 4: Verification Report and Findings
VVB issues a Verification Report stating verified net ERRs (tCO2e), any deductions applied, and conclusion. Unresolved CARs must be addressed before credits can be issued.
Step 5: Verra Registration and VCU Issuance
Verra reviews the Verification Report and project documentation. If satisfied, Verra issues VCUs to the project's registry account. VCUs are now tradeable on the voluntary carbon market.
Common Verification Challenges
| Challenge | Why It Happens | Prevention |
|---|---|---|
| Missing activity records | Farmers don't maintain logs; records lost | Digital record-keeping, regular field officer visits |
| Sampling protocol deviations | Field teams not trained; equipment failures | Written SOPs, training refreshers, backup equipment |
| Lab data inconsistencies | QC failures, transcription errors | Accredited labs, electronic data transfer, blind duplicates |
| Boundary disputes | GPS data imprecise; field boundaries changed | Annual GPS confirmation, signed farmer agreements |
| Model input errors | Wrong default factors; calculation mistakes | Independent internal review before submission |
Mock Verification Scenario:
A VVB auditor arrives at a rice paddy project in Vietnam. She asks to see activity records for 30 randomly selected farmers. The field officer retrieves mobile app logs, but 6 farmers' records are missing because they switched phones. The auditor issues a CAR: "Provide or reconstruct records for 6 missing farmers, or exclude those fields." The developer reconstructs 4 from paper receipts and excludes 2 fields. The VVB re-reviews and closes the CAR. VCUs are issued for the reduced area.
Key Takeaways, Lesson 5.2
- ✓ Validation (before project start) confirms the plan is credible; verification (after monitoring) confirms it delivered
- ✓ VCS requires a different VVB for validation vs. first verification (independence rule)
- ✓ Verification is a 5-step process: contract, desktop review, site visit, report, registration
- ✓ The VVB is independent, they confirm data, not collect it
- ✓ CARs must be resolved before VCUs are issued
- ✓ Prevention is cheaper than correction, good record-keeping from day 1 saves money
Key Takeaways
- 1Validation (before project start) confirms the plan is credible; verification (after monitoring) confirms it actually delivered results
- 2VCS requires a different VVB for validation versus first verification to ensure independence - the same auditor cannot approve their own prior work
- 3Verification follows 5 steps: contract/scoping, desktop review, site visit, verification report, and Verra registration with VCU issuance
- 4Corrective Action Requests (CARs) must be fully resolved before any VCUs can be issued - unresolved CARs block credit issuance
- 5Digital record-keeping from day 1 is far cheaper than reconstructing missing records at verification - prevention beats correction every time