Multi-Layer Plot Sampling: How Nested Plots Capture Every Carbon Pool

Multi-layer plot sampling uses concentric plots of different sizes at the same location, each sized to the carbon pool it measures. Large stems are counted over a wide radius, saplings over a smaller one, litter and soil in small quadrats. The design exists because the pools have wildly different densities, and sampling them all at one scale wastes effort on some and misses others entirely.

A hectare of tropical forest might hold a few hundred large stems, several thousand saplings, tens of thousands of seedlings, and a continuous litter and soil layer.

Sample all of that with one plot size and you get a bad trade in both directions. A plot large enough to capture enough big trees makes counting seedlings impossibly laborious. A plot small enough to make seedling counts tractable will contain zero or one large stem, and large stems hold most of the above-ground carbon, so their variance dominates the estimate.

Nesting resolves it. Each pool is measured at the scale where it is both statistically adequate and practically countable.

Exact dimensions are set by the methodology and the ecosystem as mangrove, peat and dryland protocols differ substantially. Take radii from the applicable methodology rather than from a generic template.

Above-ground biomass is usually the largest pool in forest systems and the one most sensitive to large-stem sampling. Because biomass scales roughly with the square of diameter, a handful of very large trees can carry a disproportionate share, which is exactly why plot size and count for this layer drive the confidence interval.

Below-ground biomass is almost never measured directly. It is derived from above-ground biomass using root-to-shoot ratios, which means its uncertainty inherits from the above-ground estimate and adds model error on top.

Deadwood is highly variable and easily missed. It also responds sharply to disturbance, which makes it informative as well as material.

Litter is small per unit area but continuous, so it is cheap to sample and contributes steady rather than volatile mass.

Soil dominates in mangrove and peat systems, frequently exceeding all living biomass combined, and it is the pool where decisions regarding sampling depth matter most. In these ecosystems, treating soil as a minor pool is the single largest potential source of error.

Mangrove and tidal wetland systems are where multi-layer design stops being an optimization and becomes a necessity. Published protocols, such as the widely used CIFOR reference work by Kauffman and Donato, set out nested plot layouts specifically because soil carbon extends to depths that make accounting based only on above-ground biomass badly wrong.

Projects that apply a generic upland forest protocol to mangrove routinely misestimate, and the direction of the error depends on which pools were skipped. This is a documented failure mode, not a theoretical one.

Three decisions determine whether a nested design works.

Which pools to include. Methodologies specify required and optional pools. Optional pools may be excluded conservatively, but only if excluding them understates the result, and that has to be demonstrated rather than asserted.

Plot sizes per layer. Driven by the density and variance of each pool. Pilot data beats convention.

Plot count and allocation. Derived from within-stratum variance, allocated across strata by area and variability together.

All three interact with stratification. A nested design inside poorly drawn strata still carries high within-stratum variance, and no amount of nesting fixes that.

The recurring questions are consistent: are the pools included appropriate to the ecosystem and the methodology; are plot dimensions and layouts documented and consistently applied; is deadwood and litter sampling actually performed rather than assumed; for soil, are depth intervals, bulk density and carbon fraction measured and recorded; and does the field data trace back from the headline tonnage without gaps.

A nested design is straightforward to defend when it was planned. It is very hard to reconstruct afterwards.