What is an NFI plot, and how does it differ from a species map?

A National Forest Inventory plot is a fixed, permanently monumented sample point, usually laid out on a grid across a state, region, or national land base, where crews measure trees on a set schedule to track forest condition over time. In the US, the Forest Inventory and Analysis program runs these on roughly a 1-plot-per-6,000-acre grid, remeasured on a rotating cycle. Each plot is a cluster of subplots, typically a 24-foot-radius main plot with smaller microplots nested inside, where field crews record species, diameter at breast height, height, crown class, and condition for every tally tree that falls within the plot boundary.
An NFI plot measures real trees on real ground: species, diameter, height, condition, all recorded by a crew standing in that 24-foot circle. But it's one point among thousands. Everything between plots gets filled in statistically, using expansion factors and stratification by forest type, so the agency can report a volume-by-species estimate for a county or a national forest as a whole. The plot grid is built for a regional question: how much timber, by species, stands across that reporting unit. A compartment-level question, like what's growing in stand 14 this year, gets answered by whatever forest-type stratum that compartment was assigned to when the map was last drawn.
What the plot grid is good at, and where it stops
For state-level timber supply studies, FIA-type data is the backbone of American forestry. It is statistically defensible, it has decades of remeasurement history behind it, and the sampling design has been peer-reviewed more times than most of us want to count. If you need a defensible species mix for a 2-million-acre reporting unit, this is still the right tool, and nothing in a species map replaces it.
But a compartment manager planning a harvest block or a replant prescription isn't asking a regional question. They're asking what's growing on this 40-acre stand, right now, species by species. A plot grid that samples one point per several thousand acres was never built to answer that. Between plots, the inventory relies on stratification and extrapolation: this stand "should" look like the forest type it was classified into when the stratum was drawn. A hardwood pocket inside a planted pine stand, a beech component creeping into a sugar maple stand, a compartment where the actual conifer-broadleaf split has shifted since the last cruise: none of that shows up unless a plot happens to land there, or someone walks the ground and notices.
Foresters have always known this. It's why a timber cruise still happens before a sale even when FIA data exists for the county. The plot network gives you a trend line for the region; it was never meant to substitute for knowing your own compartments.
Wall-to-wall classification vs sample plots
The difference is in how each method covers ground. A sample-plot inventory measures points and extrapolates between them. A wall-to-wall species map classifies every pixel across the holding, so instead of inferring what's between plot 112 and plot 113, you get a species call for that ground too.
Satellite-based species classification works from narrow-band reflectance and seasonal phenology, the way different species leaf out, change color, and senesce on different calendars across spring, summer, and fall. A hyperspectral pass at 0.5 to 2 m resolution, run across multiple seasons, can pick up the spectral signature that separates conifer from broadleaf, and in many cases species from species within those groups, continuously across a compartment rather than at scattered sample points.
The plot network still earns its keep: it's the calibration and trend-tracking tool for regional reporting, built on decades of remeasurement that a satellite pass can't replicate. A wall-to-wall layer does a different job. It gives the compartment-level manager a species map for the ground they have to prescribe treatments on, this year, without waiting for a plot grid to intersect their stand. For habitat reporting, replanting stock orders, or harvest-mix planning across an entire holding, a species layer that covers every acre lets you plan off what's standing instead of an estimate stretched across the gaps between plots. Tree Species Mapping builds that kind of annual, compartment-by-compartment species layer from satellite reflectance and phenology, so the ground between plots stops being guesswork.
If your compartment records still lean on plot extrapolation for species mix, it might be worth seeing what a full species layer across the holding would show instead.