How to read a utility GIS extract for substation capacity

A GIS extract from a DNO or TO lands in your inbox as a zip file full of shapefiles, and somewhere in there is the substation layer you care about. Before you build a shortlist on it, it helps to know what each field is telling you and what it's quietly not telling you.

What's in the extract

Most extracts give you a point layer for substations and a line layer for the circuits connecting them, each with an attribute table. The fields worth reading closely:

• Voltage level. Usually the cleanest field, and the one that sorts your candidate sites fastest. A 33 kV bus and a 132 kV bus are not interchangeable for a given MW ask.
• Capacity or headroom. Sometimes expressed as firm capacity, sometimes as a non-firm figure that assumes curtailment agreements. Read the metadata before you compare two numbers across different utilities, because "capacity" isn't defined the same way in every extract.
• Feeder or bay count. Tells you how many connection bays exist or are spare, which matters more than the headroom number once you're past screening and into real queue conversations.
• Last updated date. The field everyone skips and shouldn't. Extracts get refreshed on the utility's schedule, not yours, and a 2022 snapshot sitting in a 2026 queue review is a liability, not a convenience.

The hosting capacity map that sits on top of this data is a visualization, not a new dataset. It's coloring in the same headroom field across a region so you can eyeball where the green zones are. Useful for a first pass across a portfolio, not something to cite in a connection application.

Where the extract stops being useful

The substation point in the GIS layer tells you a capacity number. It doesn't tell you how a cable would get from your site to that bus. Two things trip up analysts who treat the extract as the whole picture.

The extract reflects the network as registered, not as built. Reconductoring, new bays, and local reinforcement work often lag the published layer by a year or more. If your shortlist hinges on a headroom figure from an extract that predates a known upgrade, you're scoring sites against stale numbers.

Distance in the attribute table, if it's there at all, is usually straight-line. A ruler measurement from site centroid to substation point ignores the river crossing, the road corridor, the third-party landowner you'd need wayleaves from, and the existing transmission corridor you'd realistically want to follow rather than cut a fresh route through. Two sites with identical "distance to substation" fields in the GIS extract can carry very different real connection costs once you account for terrain and rights-of-way.

That gap, between what the attribute table says and what a route on the ground would need to cross, is exactly where early-stage site screening goes wrong. It's also where a distance-and-route layer built from current satellite imagery earns its keep: instead of a straight-line guess sitting on top of a utility snapshot, you get a plausible corridor and a connection distance that accounts for what's between your site and the substation.

Reading the extract without over-trusting it

None of this means the GIS extract is wrong to use. It's the only source that gives you a capacity number tied to a specific bus, and nothing replaces that for the queue application itself. The read is: use the extract for what it's authoritative on (voltage, nominal headroom, bay count) and treat anything spatial in it, distance, routing, corridor availability, as a starting point that needs checking against what's on the ground before it goes into a site ranking.

If your shortlist spans a dozen candidate sites and three different utility territories, that checking adds up fast. Worth putting a real route and a current connection distance on each site before the capacity numbers decide the order.