FIA Forest Type Transitions

Explore how forest types and forest regeneration shifted across Connecticut, Maine, Massachusetts, New Hampshire, New York, Rhode Island, and Vermont, 2005-2024, from remeasured USDA Forest Service inventory plots.

Four data layers tell the story from canopy to forest floor: established forest types, and the dominant tree, sapling, and seedling species per plot. Ribbons trace every plot's path between five-year measurement intervals, so type conversions and regeneration shifts are visible as flows rather than snapshots.

Below the chart, four hexagon maps, one per interval, show where those plots sit on the landscape, so what changed and where it happened read together.

A Forest Ecosystem Monitoring Cooperative project using USDA Forest Service FIA (FIADB) data.

Tabs above the chart switch the data layer: forest types, or the dominant tree, sapling, and seedling species.
Filters on the right set the detail level, starting category, resampling filter, and geography (state or ecoregion section).
Hover a band to trace its flows across the intervals; click it to pin the highlight. A pinned type also filters the maps, showing where it sat in each interval.
The maps under the chart shade each hexagon by how many plots of the current selection were measured in that interval; outlined hexagons hold none.
The menu in the chart corner offers full screen, image downloads, and the data table.
The address bar tracks your view; copy the URL to share exactly what you see.

Data come from remeasured USDA Forest Service Forest Inventory and Analysis (FIADB) plots in the seven-state region. Forest type follows the dominant condition by condition proportion; dominant species are the greatest live basal area per plot and interval (trees 5.0 in DIA and larger; saplings 1.0-4.9 in) or the greatest expanded seedling count (seedling layer).

FIA remeasures each plot on a roughly five-year cycle, so each measurement interval is a lustrum (a five-year span) and every ribbon is a transition between successive lustra. Plots appear when measured in at least two lustra. The resampling filter (on by default) keeps only plots measured in every interval, so every ribbon is a real plot-to-plot transition. Types under 1.5% of an axis are lumped into "Other (small types)". Plot coordinates are the public fuzzed values.

Percentages are shares of plots measured in that interval, not of all plots on the axis. The "Not measured" band is therefore shown as a plot count rather than a percentage, and the remaining bands add to 100% among themselves. Earlier versions of this page divided by the padded total, which understated every band wherever plots were missed - New York's maple / beech / birch share on the last interval, for example, reads about 52% here where the earlier page showed about 33%. Figures cited or captured before August 2026 will not match.

Nonstocked plots are still forest land. FIA assigns this label when a stand temporarily carries too few live trees to type - most often just after a harvest, or while a young stand regenerates - so what changed is the tree cover, not the land use. It is not the same as a plot converting to nonforest, a true land-use change, which the underlying data tracks separately. Nonstocked is ordered by size alongside the forest types rather than pinned to the bottom of the axis, so flows into and out of it trace stands through harvest and back into a type.

The maps bin plots into the standard 64,000-acre FIA display hexagons using the public fuzzed coordinates (displaced up to about a mile), so no exact plot location is shown. One color scale spans all four maps, letting shading compare across intervals.

About the Forest Ecosystem Monitoring Cooperative