Sequestr - Nigerian Mangrove Project

Methods & Validation

Scientific approach and validation framework for the Research-Based Monte Carlo model.

Chapman-Richards Growth Model

Asymptotic growth trajectory

Growth Equation
y(t) = initial + (asym − initial) · (1 − exp(−k·t))c
Parameter Uncertainty

Asymptote values sampled from priors with species-specific distributions

Structural Uncertainty

Growth rate constants (k, c) vary by species and site conditions

Asymptote uncertainty is applied via clipped z-scores to create parallel growth trajectories. This captures both parameter uncertainty (asymptote values) and structural uncertainty (growth rate constants).

Self-Thinning Density Dynamics

Size-density feedback

Self-Thinning Equation
Nmax = α · Dq−β
α (Alpha)

Calibrated to anchor mixed stands to sampled dy40 target

β (Beta)

Default 1.6 controls self-thinning strength

Dq

Quadratic mean diameter drives density limits

α is calibrated so mixed stands anchor to the sampled dy40 target while respecting species-specific density limits. The beta parameter (default 1.6) controls the strength of self-thinning.

Allometry Weighting

Structural equation uncertainty

45%
WA-Mama2023
West African
45%
GAB-Gabon
Regional
10%
GLOBAL-Komiyama
Global

Primary equations: West African allometry (Zanvo et al. 2023, Benin) calibrated to Niger Delta conditions. Weighted sampling captures structural uncertainty across regional equation sets, yielding approximately 50% lower (more conservative) biomass estimates than global IPCC equations.

Dependence Structure

Shared latent site quality factor

Site Quality zsite
correlates with
Growth (ρ = 0.70)
SOC (ρ = 0.45)
Density (ρ = 0.20)

A shared latent factor induces realistic correlation across growth, SOC, and density. This ensures biological coherence: high-quality sites have both better growth and higher soil carbon accumulation.

Model Validation

Conservative vs field measurements

Model vs Literature

Growth model projections at 30 years are 1–9% lower than weighted-average values from literature, validating realistic size trajectories.

Model vs Nigerian Field Data

Model projections are 29–68% more conservative than actual measured growth from 3 Nigerian pilot sites (1.5–2 year monitoring). This conservatism is appropriate for ex-ante projections submitted for certification.

Validation approach: Model conservatism is validated against both regional literature benchmarks and direct field measurements from pilot restoration sites in the Niger Delta. The conservative bias ensures credible ex-ante carbon estimates suitable for carbon credit certification.