GIS Data Sources for Data Center Site Selection: What Each Layer Shows

A desktop screen stacks mapped layers over a parcel and asks what each one implies: power nearby, flood zone, wetlands, streams, slope, soils, and connectivity. The federal layers below are the usual starting point because they are free, national and consistent from county to county. Our guide to desktop screening vs. field assessment covers when to stop relying on them; this guide covers what each one is.

Two habits make any layer more useful. First, record the publisher, the download date and the data’s own effective or publication date, because copies of the same layer circulate with different vintages.⁹ Second, read the metadata’s stated purpose. A layer built for inventory or general planning will say so, and that sentence tells you how far to trust it at the scale of one parcel.³⁴

The Homeland Infrastructure Foundation-Level Data (HIFLD) program has been the standard public source for electric transmission lines and substations. A poster from Oak Ridge, Los Alamos and Idaho National Laboratories describes how that foundation electric data was built from open sources and updated annually from the latest available inputs.¹¹ Treat it as a map of where lines and substations probably are, with voltage classes where known.

Check where your copy came from. The Data Rescue Project archives a copy of the HIFLD Open transmission lines layer, listing a download date of August 26, 2025.⁹ Before relying on any version, confirm where the current federal release is hosted and when it was last updated, and compare it with recent imagery.

The U.S. Energy Information Administration’s U.S. Energy Atlas, launched in January 2021, offered 84 map layers, 60 of them based on EIA surveys, covering power plants, pipelines, storage, processing plants, refineries and other energy facilities.¹² Power plant locations come from the annual Form EIA-860, with latitude and longitude also published monthly in EIA-860M.¹ EIA states that it is not the source of transmission line data and directs users to HIFLD.¹

FEMA National Flood Hazard Layer

The NFHL combines effective Flood Insurance Rate Map databases with the Letters of Map Revision issued against them.² It includes only data FEMA considers final; preliminary maps shown at community meetings are excluded, and records enter the layer when they reach their effective date and become valid for regulatory use. It shows the 1-percent-annual-chance and 0.2-percent-annual-chance flood zones and areas of minimal risk.² It covers only parts of the country where maps have been modernized, so a blank area may mean no digital data rather than no hazard.⁵ Our floodplain guide explains what the zones mean for a campus.

National Wetlands Inventory

The Fish and Wildlife Service maps wetlands to provide reconnaissance-level information on their location, type and size, mostly from high-altitude imagery.³ The mapper uses a biological wetland definition and warns that it should not be read as showing the presence, absence or extent of wetlands regulated under federal, state, Tribal or local law; the Service recommends contacting the local Corps of Engineers office.³ Its metadata adds that it makes no attempt to define regulatory jurisdiction and that certain farmed wetlands are excluded by policy.¹³ See wetlands and data center development.

NHD and the 3D Hydrography Program

The USGS retired the National Hydrography Dataset on October 1, 2023. Its data remain available but are no longer maintained.⁶ Its successor, the 3D Hydrography Program (3DHP), derives streams, lakes and drainage features from 3DEP lidar and aims to align hydrography with elevation data; where no 3DHP data exist yet, the NHD is still used.⁷ For a site screen, that means stream lines from the older dataset may be offset from where water actually runs on lidar terrain.

The USGS 3D Elevation Program (3DEP) is the national lidar program. Its baseline is quality level 2 or better lidar (IfSAR in Alaska) collected within the last eight years, and its annual status map, built from the U.S. Interagency Elevation Inventory with NOAA, shows existing, in-progress and planned coverage.¹⁰ USGS reported that 3DEP data were available or in progress for 98.3% of the nation at the end of fiscal year 2024.¹⁴ Where quality data exist, lidar supports early slope, grading and drainage judgments; our guide to topographic surveys and lidar covers what it can replace and what it cannot.

SSURGO is the most detailed soil geographic database from the National Cooperative Soil Survey, designed mainly for farm and ranch, landowner, township and county natural resource planning.⁴ Because of mapping scale, features too small to delineate are shown separately as special point and line features rather than as polygons.¹⁵ Use SSURGO to flag shallow bedrock, hydric soils, high shrink-swell clays and poor drainage, then confirm with borings. See slope, soils and geotechnical review.

The FCC National Broadband Map is built from availability data that providers report for individual broadband serviceable locations.⁸ It shows availability, not performance: the FCC’s starter kit explains that challenges address whether a service is offered at a location, and a speed test showing underdelivered speeds is not a basis for a challenge.⁸ The kit lists grounds such as a provider being unable to connect a location within 10 business days of a request or charging more than its standard installation fee.⁸ Individuals file through the map, with details of their request and the provider’s response.¹⁶

For a data center, the map answers a narrow question: which wireline and wireless providers report retail service at nearby locations. That suggests which carriers have plant in the area. It does not show long-haul routes, dark fiber, route diversity, or whether any carrier can deliver the capacity a campus needs. Those come from carrier route maps, network consultants and direct carrier conversations; see our guides to fiber connectivity and route diversity and latency.

The pattern across the table is consistent. Federal layers are good at telling you that something is probably present and roughly where. They are weak on exact boundaries, regulatory status and capacity, which are the facts that decide whether a site can be built.

1. 01Pull each layer from its publisher or an identified archive, and record the download date and the data’s own effective date.
2. 02Add county parcel, zoning and utility service territory data, which federal layers do not provide.
3. 03Check every power feature against current imagery, and ask the utility what is energized, planned and available.
4. 04Read flood and wetland layers as flags for a floodplain review and a delineation, not as findings.
5. 05Use 3DEP terrain and SSURGO soils to plan where field work goes first.
6. 06Treat the FCC map as a list of carriers to call, not as proof of data center connectivity.

Layers feed directly into a scoring model, so their errors become scoring errors. Label each input with its source and confidence, as our methodology describes. BlackForge builds these stacks for candidate parcels and flags what needs field confirmation; you can get a site reviewed.