Validating a biomass map against legacy plots, no new field work

A registry reviewer asking for a biomass map validation doesn't want a new inventory. They want proof the number on the map agrees with ground truth that's already on file. If your project area has legacy plots from the baseline survey or a prior monitoring period, you likely don't need to walk the stand again to answer that question.
Why legacy plots still work for validation
A plot measured three or five years ago isn't useless just because the forest has grown. Above-ground biomass changes on a predictable curve for a given stand type and age class, and most allometric models already account for that. What a reviewer actually wants to see is whether the mapped values track the plots in relative terms: do high-biomass plots land in high-biomass pixels, do low-biomass plots land low, and is the spread consistent across strata.
That means the validation question isn't "has the forest changed since the plot was measured." It's "does the regression between canopy structure and spectral response produce a biomass surface that's internally consistent with what the plots recorded, once you account for the time gap." Growth increment tables, species-specific yield curves, or even a simple age-adjustment factor pulled from the original inventory report usually get you there without sending anyone back into the field.
Running the residual check
The actual check is a standard regression diagnostic, not a new survey. Pull the plot coordinates, extract the mapped AGB value at each location (buffered to the plot radius, not a single pixel), adjust the plot-measured value for growth since the measurement date, and plot predicted against observed. A tight scatter along the 1:1 line is what you want. Systematic bias, where the map runs consistently high in dense stands or low in sparse ones, shows up immediately as a slope that drifts off 1:1 rather than scatter around it.
Report the residuals by stratum, not just as one pooled number. A registry reviewer evaluating a REDD+ or improved forest management project is going to ask how the map performs in the stratum that carries the most credits, and a single aggregate RMSE hides exactly that. If your legacy plots are concentrated in one forest type and thin in another, say so in the monitoring report rather than letting the aggregate number paper over it. Reviewers have seen that move before.
One thing worth flagging early: if the legacy plots predate a significant disturbance event, drop them from the validation set rather than forcing an adjustment that isn't defensible. A validation built on plots that no longer represent the stand is worse than no validation at all, and most verification bodies will catch the mismatch anyway.
What goes in the monitoring report
Three things, at minimum: the plot-to-pixel matching method (buffer radius, resampling if any), the growth adjustment applied and its source, and the residual plot by stratum with sample size per stratum called out. Skip the adjustment methodology and a technical reviewer's first question will be why the map and the plots were compared as if no time had passed.
If the number of legacy plots per stratum is thin (fewer than ten is a common soft threshold reviewers flag), say that plainly in the report rather than presenting the residual check as more conclusive than the sample supports. A reviewer who sees an honest small-sample caveat moves faster than one who has to go find the gap themselves.
None of this requires a fresh plot campaign, and it's the same logic you'd apply whether the AGB layer came from a one-off aerial LiDAR pass or from an annually refreshed biomass layer built for repeat reporting cycles, as long as the underlying method is documented. What changes with a repeat-cadence product is that you run this same residual check each crediting period, against the same stratification, so drift over time becomes visible instead of assumed.
If a fresh layer and a workable validation path for the next reporting cycle are what you're after, that's the problem Biomass Mapping is built to solve.