Which NLCD land cover classes get excluded when screening solar parcels

If you've pulled the National Land Cover Database into QGIS to start a regional screen, you already know the first problem: the legend has 20 classes and nobody on your team agrees on which ones are automatic no-gos versus which ones just need a closer look. Here's how most land acquisition teams actually sort it.

The classes that get excluded outright

Open Water (11). Self-explanatory, but worth flagging because NLCD's water classification picks up seasonal ponds and flooded ag fields that may not show up on a county flood map. If a parcel reads mostly class 11, don't waste a site visit confirming it.

Woody Wetlands (90) and Emergent Herbaceous Wetlands (95). These are the two classes that cause the most rework later in the process. A parcel can look clean on a plat map and still carry a wetland delineation requirement that adds months to permitting. Teams that skip this check at the screening stage end up paying for it during Section 404 review.

Developed, Medium and High Intensity (23, 24). Dense development. Not buildable for utility-scale ground mount, and usually not worth modeling even for a feasibility pass.

Perennial Ice/Snow (12). Rare outside specific latitudes, but it shows up in some western screens and is an automatic cut.

A few classes get excluded by some developers and kept by others, and this is where regional practice and interconnection economics start to matter more than the raster.

The classes that depend on who's screening

Deciduous, Evergreen, and Mixed Forest (41, 42, 43). Forest cover isn't a hard exclusion everywhere, but clearing cost, canopy removal permitting, and local ordinances (several counties in the Southeast and Mid-Atlantic now restrict large-scale tree clearing for solar) push most teams to flag these as a downgrade rather than a pass. If land is cheap enough and the interconnection queue position is good enough, some developers will still model through forest. Most won't lead with it.

Developed, Open Space (21). This one trips people up. NLCD puts golf courses, large-lot residential, and cemeteries in the same bucket as some undeveloped buffer land. It's worth a manual look rather than a blanket exclude.

Pasture/Hay (81) and Cultivated Crops (82). Generally preferred classes for solar, not excluded. Flat, already cleared, usually fewer wetland and forest permitting entanglements. The complication here is prime farmland soil designations layered on top of the land cover class, which is a separate dataset entirely.

Barren Land (31) and Shrub/Scrub (52). Usually fine, sometimes flagged in arid regions for dust control or revegetation bonding requirements, but not a land-cover exclusion in the way wetlands are.

Where the raster runs out of answers

The honest limitation here: NLCD runs at 30-meter resolution and updates on roughly a two-to-three-year cycle. A parcel that was cultivated cropland at the last NLCD release could be fallow, regrown to scrub, or partially cleared by now, and you won't know it from the raster. The class boundaries also don't track parcel lines, so a 40-acre field can show up as three or four different classes stitched together in a way that doesn't mean anything on the ground.

That's the gap between "ran an NLCD query" and "screened a region." NLCD tells you the land cover class as of its last update. It doesn't tell you current slope, current flood exposure, or how a parcel's cover interacts with its shape and size once you draw the actual array footprint. Most GIS analysts end up stacking NLCD against a slope raster, a FEMA flood layer, and a parcel boundary file by hand, parcel by parcel, which is the exact bottleneck that turns a regional screen into a multi-week project.

If you're running this process across a few hundred parcels at a time, you can see how an automated pass/fail screen built from current satellite imagery handles the land cover, slope, water, and flood exclusion together instead of as four separate layers someone has to reconcile manually. Worth a look before your next regional push.