Annual NLCD land cover across the seven-island archipelago, 2021 and 2025

Summary

Annual NLCD supplies the first consistent current land-cover frame across all seven islands of the archipelago: Beaver, Garden, High, Hog, Whiskey, North Fox, and South Fox. The official time-enabled land-cover service, at 30 m nominal resolution, provides both 2021 and 2025 snapshots. The results are especially useful for the Fox Islands, where the MNFI 2017 natural-community report has no occurrence records and the DNR public stand-inventory service returned no polygons.

Annual NLCD is broad remote-sensing land cover. Its forest classes distinguish deciduous, evergreen, and mixed forest, but not individual tree species such as cedar, pine, birch, or oak. Its wetland classes distinguish woody and emergent wetlands, but do not establish legal wetland boundaries. It is therefore a regional context layer that complements the MNFI observations, SSURGO soils, NWI wetlands, circa-1800 reconstruction, and DNR stand comments.

Seven-island 2025 frame

Areas below are summed from classified 30 m pixels whose centers fall inside the held island polygons. Forest is NLCD classes 41–43; wetlands are classes 90 and 95; developed is classes 21–24; barren is class 31. They are not parcel, survey, or natural-community acreage.

The broad pattern is useful but should be read cautiously. Beaver is the largest and most developed frame, with substantial forest and woody-wetland coverage. High, Hog, and South Fox retain substantial forest pixels. Garden is dominated by the NLCD woody-wetland class, while North Fox and South Fox show very different proportions of forest, barren land, and wetland. These patterns are hypotheses for connecting to soils, shoreline geology, historical land cover, and field records—not substitutes for those layers.

Forest and wetland detail

The 2025 forest totals preserve the NLCD split needed for comparison with the archipelago’s tree-history material:

These are land-cover classes, not species inventories. For example, NLCD evergreen forest does not establish that a pixel is cedar, spruce, fir, or pine. The species-level evidence remains in the MNFI site summaries, DNR comments, historical sources, and future field or inventory records.

Change signal, 2021–2025

The extractor compares aligned normalized rasters and preserves every changed class transition. Changed pixels are not automatically interpreted as real ecological change: shoreline water, classification updates, mixed pixels, and 30 m boundary effects can all contribute.

The index contains the full from-class/to-class transition list. Treat these as screening signals for follow-up against aerial imagery, MNFI monitoring, NWI updates, forest inventories, fire or storm records, and local knowledge.

Connectivity to the existing ecology corpus

This layer closes a specific scale and coverage gap:

• MNFI Report 2017-11 provides 43 detailed natural-community occurrences on Beaver, Garden, High, and Hog, but no North or South Fox occurrence records.
• The DNR public forest-inventory stand layer provides detailed stand and comment data for Beaver but no returned stand polygons on the other six island frames.
• Annual NLCD now gives all seven islands the same 2021/2025 land-cover vocabulary, including forest, shrub, barren, developed, water, and wetland classes.
• The existing SSURGO, NWI, EGLE hydric-soil, geology, and circa-1800 layers remain separate evidence layers. Their categories should be compared rather than collapsed into NLCD classes.

The practical next step is to use the NLCD raster as the common outer-island frame, then add Fox-Island-specific historical, GLO, NWI, and biological records without implying that MNFI’s Beaver/Garden/High/Hog survey scope extends to Fox.

The machine-readable ecology crosswalk makes that relationship explicit: it contains seven NLCD island records, four islands with MNFI occurrences, seven with SSURGO context, one with returned DNR public-forest stands, and the held environmental-layer summaries where those layers were queried. Missing MNFI or environmental-context entries remain null or zero with a source-scope note; they are not converted into ecological absence claims.

Evidence limits

The raw WMS GeoTIFFs and normalized class rasters are retained for every island and year. The normalized rasters are derived from the official rendered RGB palette; they are not presented as a separate USGS distribution. WMS pixels outside the island mask are set to zero in normalized rasters. The data do not provide ownership, parcel boundaries, species-level tree identification, natural-community element-occurrence geometry, legal wetland determinations, or proof that a 2021–2025 class transition represents an on-the-ground event.

Added 3 August 2026. Recent changes