Intact Forest Landscape¶
An operational forest-conservation class for a contiguous forest-zone mosaic that clears mapped human-disturbance exclusions and fixed minimum area, width, and corridor thresholds.
Core Idea¶
An Intact Forest Landscape (IFL) is an operational map class for a large, contiguous mosaic of forest and naturally treeless ecosystems within the current forest zone that shows no remotely detected sign of specified significant human activity. In the global IFL method, a surviving polygon must have at least 500 km² (50,000 ha), admit an inscribed circle 10 km in diameter, and retain corridors or appendages at least 2 km wide. These fixed spatial gates follow removal of mapped settlements, industrial infrastructure, agriculture, logging, mining, and other disqualifying disturbance, with specified buffers.
Scope of Application¶
The IFL class was designed for global and regional forest monitoring, conservation prioritization, carbon and biodiversity analysis, forest certification, and change detection. Potapov and colleagues used a consistent remote-sensing workflow to map the world's large unfragmented forest landscapes, while later editions compare the fate of the baseline polygons through time.
The scope includes boreal, temperate, tropical, and subtropical forest zones. Associated natural nonforest areas are retained when they form part of the same landscape mosaic; this avoids cutting wetlands, grasslands, lakes, and high-elevation openings out of an ecologically continuous system merely because they lack tree cover.
Clarity¶
A four-stage diagnostic keeps the identity clear.
First, subtract disqualifying influence: apply the declared disturbance categories and their buffers. Second, form residual connected components: determine what remains spatially unbroken after subtraction. Third, test geometry: check area, inscribed width, and corridor/appendage width. Fourth, version the result: attach a reference date, source imagery, and method edition.
Manages Complexity¶
Forest intactness varies by canopy condition, species composition, hunting pressure, fire regime, roads, settlement, patch size, connectivity, and observation quality. Modeling all dimensions continuously at global scale is difficult and makes cross-region comparison sensitive to weighting. IFL compresses the problem into an interpretable screen: remove specified industrial footprints, retain large and wide residual mosaics, and compare like-class polygons over time.
Abstract Reasoning¶
The classification licenses several deductions.
Monotonic loss under a fixed baseline. If qualifying land is removed by newly mapped disqualifying disturbance and no restoration/reclassification rule adds land, IFL extent can stay equal or decrease. This makes the baseline suitable for loss accounting.
Threshold discontinuity. A small physical intervention can cause a large categorical change. A narrow road plus buffer may split one qualifying component into two subthreshold components, so the membership loss exceeds the disturbed area.
Knowledge Transfer¶
The full identity transfers literally among forest-monitoring regions because the same roles recur: forest-zone baseline, disturbance mask, buffers, connected components, geometry gates, and versioned outputs. It also transfers from mapping into certification and policy when those institutions explicitly import the IFL definition.
The method teaches broader spatial-analysis lessons—make the observation window explicit, separate evidence subtraction from geometric testing, disclose detection limits, and version classifications. Those lessons route to Classification, Threshold, Boundary, Spatial Indexing, and Monitoring.
Relationships to Other Abstractions¶
Current abstraction Intact Forest Landscape Domain-specific
Parents (1) — more general patterns this builds on
-
Intact Forest Landscape is a kind of Classification Prime
The minimal prospective parent is
prime:classificationthrough strict subsumption.
Hierarchy path (1) — routes to 1 parentless root
- Intact Forest Landscape → Classification
Neighborhood in Abstraction Space¶
Intact Forest Landscape sits in a sparse region of the domain-specific corpus (88th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.
Family — Unclustered & Miscellaneous (1565 abstractions)
Nearest neighbors
- Potential natural vegetation — 0.81
- Native Species — 0.80
- Mesohabitat Simulation Model — 0.80
- Intermediate disturbance hypothesis — 0.79
- Habitat Fragmentation — 0.78
Computed from structural-signature embeddings · 2026-09-08