Look at a grassland and you are seeing about a third of it.
Across the world's grasslands, roughly two-thirds of plant biomass sits below the surface — root mass, crowns, and the soil those roots built. In forests the ratio inverts: about 22% belowground. The prairie you drive past is mostly the part you cannot see, and it took decades to put there.
That single fact explains most of what goes wrong with grasslands. A grassland does not read as an asset from the road. It reads as space — space between towns, space before the mountains, space that looks like it is waiting for something to happen to it.
what a grassland actually is
The working definition is plain. The Natural Resources Conservation Service calls grassland "land on which the vegetation is dominated by grasses, grass-like plants, and/or forbs." No trees required. No trees expected.
Cover is the operative word, not grass. A functioning grassland is a stack of living layers on the same acre. Three of them are always there. The fourth depends on where you are standing:
- Perennial grasses — bunchgrasses and sod-formers that green up, cure, and hold their root mass through drought years instead of starting over each spring.
- Forbs — the broadleaf plants and wildflowers carrying most of the insect and pollinator base, and much of the protein in a spring diet.
- Shrubs, where they belong — in shrub-steppe country, big sagebrush and its relatives are the winter cover and thermal shelter of the Intermountain West. On tallgrass prairie there is no shrub layer to speak of, and a woody stem is usually a warning. There, sage is not encroachment. It is structure.
- A living soil surface — litter, root crowns, and in dry country a biological soil crust of cyanobacteria, lichens, and mosses that holds the top inch against wind and lets water in instead of off.
Remove a layer that belongs there and the acre still looks like a grassland from the highway. It stops behaving like one. That is the whole problem in a sentence: a grassland is a cover system, and it fails one layer at a time, invisibly, from a moving car.
empty land, a crop, a tree farm, or a grassland
Four things can occupy the same flat acre. They are not substitutes for each other.
| what is actually there | who manages it | how it comes back | what it produces | |
|---|---|---|---|---|
| empty land | a label on a map — usually a grassland nobody has priced | someone, always — the label just says nobody | not applicable; the label is the problem | whatever the next use decides |
| crop | annual plants, tilled ground, agronomic inputs | a farmer, on a season clock | next spring, by design | food and fiber, deliberately |
| tree farm | planted woody stems, often one or two species | a forester or a project developer | decades, as a different biome | timber, or a carbon receipt |
| living grassland | perennial grasses, forbs, shrubs, root mass, soil crust | ecological management — grazing, fire, rest | years to decades; some layers far longer | forage, infiltration, erosion control, habitat, a longer fire interval |
The column that gets skipped is the last one. A grassland's output is mostly services performed in place: water that soaks in, soil that stays put, forage that shows up without being planted. Some of that does get priced — forage trades by the animal unit month in grazing leases across the West. What carries no price is the condition of the standing cover that keeps producing it, year after year, whether or not anyone leases it. That is why "empty" survives as a description of ground that is doing work even when it looks idle.
prairie, steppe, and sagebrush country
Three names for regional versions of the same thing.
Prairie is the tallgrass, mixed-grass, and shortgrass belt of the Great Plains, from Manitoba to Texas — grass-dominated, historically grazed by bison, and burned on a short interval. The fire had two sources: lightning, and Indigenous peoples who set it deliberately. Cultural burning shaped this cover for millennia before any of the mapping existed, and the nations who carry that practice are the ones to describe it.
Steppe is semi-arid grassland with shorter cover and a longer dry season. Much of the Columbia Plateau is steppe.
Sagebrush country is shrub-steppe: a sagebrush overstory with a perennial bunchgrass and forb understory, across the Great Basin and the wider Intermountain West. This is where the definition gets argued honestly. In the 15-stock accounting we use, Grasslands and Shrublands are separate stocks, and sagebrush steppe genuinely straddles them. It behaves like a grassland — the understory drives the production and most of the fire behavior — carrying a shrub layer that takes far longer to replace than anything under it. That clock gets its own post: sagebrush takes decades.
why grasslands stop being grasslands
Four mechanisms take the cover, and only one of them looks like destruction.
- Conversion to cropland. Sod broken for annual crops. Fast, legal, often rational for the individual operator, and effectively permanent on any operator's horizon — a seeding puts grass back; native diversity, root mass, and soil structure take decades to follow, if they do.
- Woody encroachment. Juniper, mesquite, and eastern redcedar moving in where fire used to hold them off. The acre gets greener on satellite imagery and stops being a grassland.
- Invasive annual grass. Perennial cover replaced by annuals that cure early and carry fire sooner. Cheatgrass is a fire-interval problem, not a weed problem, and it has its own post: cheatgrass is a fire cycle.
- Fragmentation. Subdivision, roads, energy and transmission footprints. Every piece still has grass on it. The system stops working at the scale herds and birds actually use.
The National Fish and Wildlife Foundation, announcing Great Plains grassland grants in June 2026, put the aggregate plainly: more than half of the region's historic native grassland footprint has been lost to land-use conversion, woody encroachment, invasive species, and fragmentation. Sharper "percent remaining" figures circulate widely and frequently mix denominators — tallgrass versus all grasslands, United States versus North America, one vintage of mapping versus another. The direction is not in dispute. The specific fraction depends on which grassland you mean, and it is worth asking anyone who quotes one.
The cleanest measurement of grassland condition is not acres. It is birds. Rosenberg and colleagues, in Science in 2019, found grassland birds down 53% since 1970 — more than 700 million breeding individuals across 31 species, with 74% of grassland species in decline. That is the steepest proportional loss of any breeding biome in North America. It is a bird count, not an acreage count, and it should be cited as one. It is also the closest thing we have to a running condition reading on the cover itself.
what is not on that list
Livestock. Most North American grasslands evolved under large grazing animals, and a well-managed grazing operation is one of the more reliable ways an acre stays in grass rather than becoming a cornfield or a subdivision. The failure mode is pattern, not presence — continuous season-long use with no rest, on the wrong soil, in a dry year.
Ranchers are the largest group of grassland managers on the continent. Tribal nations are among the most consequential. Reservation and trust lands across the Northern Plains carry large blocks of native cover, and tribal bison programs are managing grassland at scale under their own authority and their own objectives — which are theirs to state, not ours to summarize. Any grassland strategy that treats either group as the problem has misread its own map. Any strategy that speaks for them instead of asking has made a different mistake.
what the cover pays for
In range dialect first, accounting second:
- Forage that appears without being planted. The forage has a price — leases trade in dollars per animal unit month, and everyone in the business knows the number. The cover's capacity to keep producing it does not, until it fails and the herd gets destocked.
- Water abundance and clean water — infiltration instead of runoff, which becomes a downstream reservoir's problem the day it stops.
- Erosion control — root mass and soil crust holding the top few inches against wind and summer storms. Some of that crust does not come back on any human schedule: the crust that holds the desert together.
- Risk resilience — perennial cover carries a longer fire return interval than annual cover on the same ground.
- Habitat — grassland birds, pronghorn, sage-grouse, pollinators, and the insect base beneath all of them.
- Carbon that stays put. Because a grassland banks most of its carbon belowground, fire takes the standing crop and leaves the stock. Modeling work on California found grasslands to be the more reliable sink through 2101 under most emissions scenarios, while fire- and drought-stressed forests flipped to sources. That is an argument for grasslands as grasslands — not a reason to plant trees on one. The honest version of that question is post three: a tree on a prairie is not a grassland.
In our stocks and flows accounting, Grasslands is one of 15 ecosystem stocks and those services are named flows off it. Flows are what a beneficiary is actually funding when they fund condition — not a view, not a vibe, not a tonne of accounting.
the grassland is the asset
The grassland exists whether or not anyone buys a certificate. It was there before the accounting and it will outlast the accounting. It does not need us to be real. What it needs is for someone to be paying for the cover to stay standing while it is still standing.
That is the only job the instrument has. The people already absorbing the cost of failure — ranchers who destock, counties fighting rangeland fire, beef supply chains that need a forage base, insurers repricing exposure, bird groups watching a 53% decline — can fund the living condition of a named place now, instead of paying for the loss later. Mechanically, a certificate funds one named place's condition directly and a coin funds protection across the protocol indirectly; both route proceeds to an account that represents a place or a purpose — here, the grassland stock itself. That is the entire technical content of this post.
Plenty of good money is already moving. In September 2026 Ducks Unlimited took a $5.79 million grant through NFWF's Grassland Resilience and Conservation Initiative — funded by McDonald's USA, several of its beef and beverage suppliers, and NRCS — to replant more than 10,500 acres to native grass and improve 4,000 acres of existing grazing land across North Dakota, South Dakota, Montana, Minnesota, and Iowa, with ranchers as the delivery mechanism. A beef supply chain voluntarily paying to keep its own forage base intact is the right instinct, and it is a cousin of this work rather than a competitor.
The difference is what the money holds when the work is finished. A conservation easement holds a removed development right. A practice payment holds a completed action — seeding, fence, water development, a grazing plan. A carbon credit holds a tonne-year of accounting. All three are real and all three are useful. None of them is ongoing, priced funding of the cover's condition on a named place, which is the gap ensurance exists to fill. Post six works through who writes which check: who pays to keep grass standing.
Our own stage, stated honestly: grasslands.ensurance is live and volumes are small. The mechanism works and the market for it is early. That is a different sentence from "this is solved."
And the price is a bridge, never the worth. A grassland is not worth a forage lease, a carbon quote, or a fire-loss avoided. Those numbers exist so capital can see the cover at all.
frequently asked questions
what are grasslands?
Grasslands are ecosystems where vegetation is dominated by grasses, grass-like plants, and forbs rather than trees. In practice a functioning grassland is a layered cover system: perennial grasses and forbs above ground, a much larger root mass below, a shrub layer in shrub-steppe country, and a living soil surface that controls whether water infiltrates or runs off. Prairie, steppe, and sagebrush country are regional types of grassland, and roughly two-thirds of a grassland's plant biomass is belowground.
why do grasslands keep disappearing?
Four mechanisms, in rough order of permanence: conversion to cropland, woody encroachment where fire has been removed, replacement of perennial cover by invasive annual grasses that shorten the fire interval, and fragmentation by roads, subdivision, and energy infrastructure. Grasslands are unusually exposed because they are cheap to convert, easy to mistake for empty, and produce almost nothing that gets invoiced — infiltration, erosion control, forage, and habitat have no line item until they stop.
are prairies the same as rangeland?
No, and the two words are not even the same kind of category. Grassland is a cover type — what is growing there. Rangeland is a management category: land whose native or potential vegetation is predominantly grasses, grass-like plants, forbs, or shrubs, managed by ecological rather than agronomic methods. That definition, from the Society for Range Management, pulls in shrublands, savannas, many deserts, tundra, and wet meadows alongside grasslands. So most prairie is rangeland, but plenty of rangeland is not prairie. SRM puts United States rangeland at roughly 770 million acres — about 40% of the contiguous land area.
what to read next
Two spokes take the clocks that decide whether cover stays cover:
- cheatgrass is a fire cycle — how an annual grass shortens the fire interval, and why herbicide is a tool rather than a grassland.
- a tree on a prairie is not a grassland — the honest version of the tree-planting question.
Two things to look at:
- explore — places, agents, and what is currently funded.
grasslands.ensurance— the grassland door as it actually stands today.
And one sibling argument from a different biome with the same structure: you don't have a bear problem. you have a food-year problem. A bear in town is a receipt for a failed food year. A grassland is the biome that grows the forage in the first place.