Desktop Screening vs. Field Assessment for Data Center Sites

A desktop screen tests a parcel against mapped data without anyone visiting it: parcel boundaries, aerial imagery, flood zones, wetland inventories, soils, elevation, transmission lines and substations, fiber routes, zoning, airports and protected lands. It is fast and cheap enough to run on dozens or hundreds of parcels, which is why it is the first stage of almost every portfolio screen.

Its job is narrow. A good screen removes sites that clearly fail, ranks the survivors, and lists the specific questions each survivor still has to answer. It cannot tell you that a site works. Most of the layers it relies on were built for regional planning or inventory, and their publishers say so in the metadata.¹ Our guide to GIS data sources for site selection covers where each layer comes from.

Some questions are answered well from a desk, at least to the level a screen needs.

• Gross size, shape and adjacency. Parcel data and imagery show whether a tract is large and regular enough for a campus, and what sits next to it. Exact boundaries still need a survey.
• Terrain, where good lidar exists. The USGS 3D Elevation Program tracks where quality level 2 or better lidar is available, in progress or planned, and publishes an annual status map.⁶ Where it exists, lidar supports early grading and drainage judgments.
• Regional flood exposure. Flood Insurance Rate Maps show whether a parcel is in or near a mapped special flood hazard area, which is enough to flag it.
• Airspace triggers. Under the FAA’s notice rules, any structure more than 200 feet above ground level at its site requires notice, and closer to airports lower structures can too.⁷ Coordinates and heights are enough to know whether a Form 7460-1 filing is likely.⁸
• Context. Distance to highways, rail, towns, schools and existing industrial uses, and the zoning district on the county map.

The common thread is that these are questions of location and rough magnitude. Desktop data handles “where” and “roughly how much” well. It handles “exactly where,” “how much is really available” and “will the regulator agree” poorly.

Every major screening layer has a documented blind spot. Knowing them is what separates a useful screen from a false sense of certainty.

Wetlands are the clearest example. The Fish and Wildlife Service describes the NWI as reconnaissance-level information on the location, type and size of wetlands, and states it is not intended as a precise locator of wetland boundaries or for regulatory purposes.¹ Maps based on older black-and-white photography tend to omit forested wetlands and wet meadows, and imagery captured in a dry year can understate wetlands too.¹ A parcel with no NWI polygons can still hold regulated wetlands; our wetlands guide covers what that means for buildable land.

Flood data has a similar problem with age. A 2017 audit by the Department of Homeland Security Inspector General found that, as of December 2016, only 42% of the flood map miles in FEMA’s inventory were updated and valid.³ A parcel shown outside the mapped floodplain may simply sit on an old panel, which is why floodplain screening should look at elevation and drainage, not only the zone label.

Power data is the most dangerous gap. National transmission layers are compiled from many sources and generalized for scale.⁴ One county copy of a major power line layer was updated to include segments approved by the state commission but not yet built.⁴ Archived copies of the federal HIFLD transmission layer also circulate with different download dates, so the vintage of the file matters.⁹ No public map shows available capacity at a substation. That answer comes only from the utility, which is why substation capacity is always an open item after a desktop screen.

Some findings do not exist until someone collects them on site. These are also the findings that permits, lenders and purchase contracts depend on.

Wetland boundaries and jurisdiction

Federal wetland boundaries are set by a delineation done under the 1987 Corps of Engineers Wetland Delineation Manual and the applicable regional supplement, then reviewed by the Corps.² Most delineations are prepared by consultants and verified by the Corps.¹⁰ Districts publish their own submittal standards and the regional supplements they apply.¹¹ A preliminary jurisdictional determination is advisory and treats all waters in the review area as jurisdictional for permitting; an approved determination is the Corps’ official finding, can be relied on for five years and can be appealed.²

Environmental condition

A Phase I environmental site assessment includes a site reconnaissance visit by definition. Under the federal All Appropriate Inquiries rule, the site visit, interviews, lien search and records review must be completed no more than 180 days before acquisition to support liability protections.⁵ A desktop records search is useful in screening, but it is not a Phase I. See our Phase I ESA guide.

Ground, boundaries and access

• Subsurface conditions: rock, karst, shallow groundwater and bearing capacity come from borings, not soil maps. See slope, soils and geotechnical review.
• Boundaries and encumbrances: easements, encroachments and the exact property line come from an ALTA/NSPS survey and title work.
• Physical reality: drainage paths, haul road condition, culverts, neighbors, noise receptors, dumping and anything that changed since the imagery date.

The decision to send people to a site is a spending decision. The right moment is just before the cost of being wrong rises sharply. In practice, that happens at a few predictable gates.

1. 01Before signing an option or LOI with a short diligence period, walk the site and resolve any constraint that could cut buildable acreage materially.
2. 02Before paying for rezoning, engineering or a utility load study, confirm the constraints that would change the layout: wetlands, flood elevations and slope.
3. 03Before closing, complete the items lenders and statutes expect: Phase I, survey, title and any wetland verification.
4. 04Any time a desktop layer conflicts with imagery or local knowledge, such as a mapped wetland on what looks like a cornfield, or standing water in an area mapped as upland.

Field work is not one task. It is a set of studies with very different costs, lead times and seasonal limits. The cheapest way to use it is to fund first whatever could remove the most acreage or kill the deal, and to hold the rest until the site has earned it.

Every finding in a screen should carry its basis. “No wetlands mapped (NWI, desktop)” is a different statement from “no jurisdictional waters (approved JD).” When a report blurs those, readers act on more certainty than exists. Our guide to reading a feasibility report explains how to check for that.

Season matters too. Wetland hydrology and vegetation indicators are easier to read at some times of year than others, and access can be limited by crops, weather or the owner. Build that into the due diligence period rather than discovering it after signing.

The two methods are not rivals. A desktop screen without field follow-up overstates certainty, and field work without a screen spends money on sites that a map would have ruled out in an hour. The sound approach is to let the desktop pass narrow the list and define the questions, then send the right specialist to answer each question before it starts to drive money or time.

• Record the source and date of every layer used in the screen.
• Mark each finding as mapped, verified informally, or confirmed in the field.
• Get the utility’s read on capacity before paying for field studies on a site that may not have power.
• Walk the site early, even if formal studies come later.

BlackForge screens sites on the desktop and flags exactly which findings need field confirmation. If you have a parcel in mind, you can get a site reviewed. Our methodology explains how we grade confidence.