Blog·

Wall-to-wall biomass mapping vs sample-based forest inventory

A forested mountain ridge and valley stretching into the distance under a cloudy, stormy sky.

A project developer filing under a REDD+ methodology or a jurisdictional forest carbon program eventually has to answer the question a registry reviewer will ask: how was the biomass number built, and what happens when someone checks it five years from now.

Plot-based carbon inventory answers that with a sampling design. Strata get drawn from forest type and age class, plots get allocated by expected variance within each stratum, field crews measure diameter and height, allometric equations convert those measurements to above-ground biomass per hectare, and a stratified mean gets applied across the project area. It's the method most registries were built around, and for good reason. A well-designed plot network gives you a confidence interval you can defend in a validation review.

It also has a shape that doesn't match the landscape. Strata get drawn once, often from an older forest-type map, and biomass gets treated as uniform within each one after that. A stratum covering degraded secondary growth next to intact canopy gets a single mean applied to both. Move the strata boundary and the project-level number moves with it, before a single tree has grown or been cut.

What wall-to-wall coverage changes

A landscape-scale biomass estimate built from full coverage doesn't assume uniformity inside a stratum. Every pixel across the project area gets its own biomass value, regressed from canopy structure and spectral response rather than carried over from a handful of plots and stretched across the polygon. That shows up most at edges: forest-non-forest boundaries, degradation gradients, the irregular patches that stratified sampling tends to average away.

This doesn't replace the plot network. Field data is still what trains and checks the regression, and a registry reviewer is still going to want ground truth behind any remote-sensing layer. What changes is the unit of comparison between inventory cycles. Instead of re-flying strata boundaries and re-running plot allocation each time, a wall-to-wall layer gets refreshed on the next imagery pass and compared directly to the last one, pixel by pixel, without re-deriving a stratification scheme first.

Stratified sampling vs full coverage, in practice

A single stratified plot, measured well, gives a tighter local confidence interval than any single pixel in a biomass raster. What matters more to a registry reviewer is what happens at monitoring time.

Stratified sampling vs full coverage comes down to what gets re-run at each verification cycle. A plot-based system re-runs field logistics: access to the same plots or a resampled subset, measurement, QA, re-aggregation by stratum. A wall-to-wall system re-runs an image regression over the same footprint, so the year-two map and the year-five map sit on the same grid and can be differenced directly for degradation or regrowth, without reconciling two different stratification schemes first.

For a monitoring report that has to show change across a reporting period, that grid consistency often does more work than a tighter interval at any single plot. For a baseline that has to survive a validation body's review of the sampling design itself, the documented statistics behind a plot network still carry weight a raster can't replace by itself.

Biomass Mapping builds the wall-to-wall layer from wide-swath multispectral imagery at 10 to 30 meter resolution, refreshed annually, so each cycle's map sits on the same footprint as the last one. It's built to sit alongside a plot network, not instead of it: a landscape-scale biomass estimate to check strata assumptions against, and a consistent surface to monitor against between the field years when nobody's back out in the inventory plots.

If you're scoping how a wall-to-wall layer would slot into your next reporting cycle, Biomass Mapping lays out the specifics.