Power density and how much land each energy source needs

Originally published August 9, 2013. Updated September 8, 2026.

Power density is how much electricity a source produces per square meter of land, and the gap between sources is enormous. A median natural gas plant produces about 482 watts per square meter, while wind comes in under 2. Nuclear runs about 241 watts per square meter. Utility solar comes in around 6.6 once you count the land under the panels. That gap is why a wind or solar buildout needs so much more land than a gas plant delivering the same power.

What power density measures

Robert Wilson, a mathematical ecology researcher, raised the term on this site in a 2013 piece called “Why Power Density Matters.” His point wasn’t about cost or capacity factor, the share of a plant’s rated output it delivers over a year. Power density measures something narrower, how many watts come out per square meter of land the plant or farm occupies.

That’s also a different measure from energy density, which packs energy into a fuel’s weight, in megajoules per kilogram, and matters more for questions like why jet fuel is hard to replace with batteries. Power density is about the footprint a source needs on the ground to make that power.

Wilson’s 2013 estimates put wind at 2 to 3 watts per square meter and solar at 5 to 20 depending on climate. Biofuels came in at 0.5 to 2, and fossil fuel or nuclear plants ran above 100. Those were rough, self-sourced figures. The peer-reviewed data since then is sharper.

How many watts each energy source packs into a square meter

A 2018 meta-analysis by van Zalk and Behrens, published in Energy Policy, pulled power-density figures from real US project data across nine energy sources. Their medians:

Source: van Zalk and Behrens, Energy Policy (2018).

Biomass sits at the bottom of that table even though it’s often marketed as a low-carbon fuel. Is biomass carbon neutral? covers why that pitch is more complicated than it sounds.

Vaclav Smil, who wrote the book on this exact subject in Power Density (2015, MIT Press), reports broadly similar ranges from his own research. His published figures, as reported by MasterResource, put solar PV at 4 to 9 watts per square meter and wind at 0.5 to 1.5, depending on the specific site. Natural gas, in his numbers, runs far higher, from 200 to 2,000.

What it takes to power a city-sized load, source by source

Put those medians to work on a fixed job, supplying 1,000 megawatts, roughly what a large power plant delivers, around the clock. At natural gas’s median density, that takes about 2.1 square kilometers, about four-fifths of a square mile. Nuclear needs about 4.2 square kilometers (1.6 square miles), and coal needs about 7.4 (2.9 square miles).

Solar and wind change the math by an order of magnitude. At solar’s median density, the same 1,000 megawatts needs about 151 square kilometers of panels, 58 square miles, close to 37,000 acres. Wind’s footprint, spacing included, runs to about 544 square kilometers, 210 square miles. That’s more than triple solar’s footprint and roughly 260 times gas’s.

Those wind and solar figures assume the plant runs at that output non-stop, which real wind and solar farms don’t do without storage or backup on top. Even before capacity factor enters the picture, wind’s raw land footprint is the largest of the five. That’s the number a planning board sees when a developer files for two hundred square miles of leased land, and it’s also the number that gets misread, because leased and paved aren’t the same thing.

Most of a wind farm’s footprint stays in crops or pasture

That leased-versus-paved distinction is what a 2024 USDA Economic Research Service study measured directly. Looking at wind and solar projects built between 2012 and 2020, USDA found that more than 99% of the agricultural land inside wind project boundaries stayed in farming or ranching after construction. Only about 85% of agricultural land near solar farms did the same.

NREL’s land-use numbers point to why the two differ. A summary of NREL’s utility-scale solar data, published by YSG Solar, puts the land panels directly occupy at a capacity-weighted average of 7.3 acres per megawatt, rising to 8.9 acres per megawatt once support structures and setbacks are counted. The gap between those two figures is small. Solar’s leased land and its occupied land are close to the same thing.

Wind doesn’t work that way. Hannah Ritchie’s analysis for Our World in Data, looking at 23 major US wind farms, found land-use estimates ranging from 8 to 184 square meter-years per megawatt-hour, depending on whether the count includes the whole leased area between turbines or only the turbine pads and access roads. The ground in that gap keeps its old job. That’s what the USDA figure above is measuring.

Why low density hits renewables harder than fossil fuels

Wilson’s 2013 argument was that this gap matters most where people are packed in tight. He pointed to Manhattan, where electricity use worked out to more than 100 watts per square meter, something like two orders of magnitude above what a wind farm produces per square meter of its own footprint. His conclusion was that a city that dense can’t generate enough power inside its own borders from wind or solar alone. It has to import power from somewhere with more room, or from a source with far higher density.

The World Economic Forum, citing a 2021 UN Economic Commission for Europe life-cycle study, put a number on the compact end of that same tradeoff. Nuclear needs 27 times less land than coal and 34 times less than solar PV for the same output. That’s the density gap Wilson described, viewed from the small-footprint side instead of the sprawling one.

None of this makes wind and solar a bad bet. It does mean the land has to come from somewhere with room to spare, often a long way from the city that’s using the power.

Wind and solar’s numbers are sharper than they were in 2013

The peer-reviewed data since 2013 has sharpened Wilson’s picture more than it’s overturned it. Van Zalk and Behrens’ 2018 medians put wind and solar even lower, at 1.84 and 6.63 watts per square meter, than the low end of Wilson’s own estimates once real project spacing and siting losses get counted in. Fossil fuels and nuclear are now pinned to precise medians, 135 to 482 watts per square meter, instead of Wilson’s rougher “above 100.”

The new finding is USDA’s, and it didn’t exist in 2013. The 2024 study measured what happens to the farmland around a wind project once it’s built. Over 99% of it stays in agriculture. We think that’s the more useful figure for anyone deciding whether to lease land to a wind developer, since it describes what happens to the ground afterward.

That USDA dataset only covers projects built between 2012 and 2020. Whether the same farming-and-turbines coexistence holds as more wind and solar gets sited on cropland over the next decade is the number worth checking next.