What Actually Makes a Hill Burn
Essay · Land use
The wildfire argument has become a fight about two destinations. The hills themselves are somewhere else entirely.
By Caroline Grindrod · 18 min read · August 2026
For most of my working life I have had a slightly unfashionable habit. I read the old accounts of how these hills used to be worked, and I ask old shepherds what the ground looked like when they were young.
Frank Fraser Darling’s West Highland Survey, and the practical crofting manual he wrote before it. I F Grant on Highland folk ways, recording the shieling summers before the last of the people who had lived them were gone. The parish accounts. And, closer to home, years of conversations with older fell farmers in the Lakes, who described fells clad in deep heather and bilberry, black grouse still flying in the 1940s, hay meadows full of herbs, hardy cattle and ponies grazing alongside the Herdwicks, and shepherds who slept in huts on the hill to keep their hefted flocks moving so that parts of the fell could freshen up.
I started reading all this as agricultural history and cultural heritage. That is how it is usually filed.
This summer I have been rereading it as fire history.
There is a detail about Fraser Darling that matters here. He gave us the line that every rewilding campaign now quotes — that the Highlands and Islands are largely a devastated terrain, and that no policy ignoring that can hope to rehabilitate them. He was right, and the Department of Agriculture disliked it enough to sit on his survey for years, because the criticism implied was of their own policies.
But Fraser Darling was not arguing for emptiness. He was a farmer. He had reclaimed derelict ground himself on Tanera Mòr, and he wrote a manual to help crofters farm better rather than less. His argument was for rehabilitation through husbandry.
Both camps in the current argument quote him. Only one of them has read him.
I have spent the best part of thirty years working in upland systems, and I have watched this argument come round three or four times now. It arrives with the smoke and it leaves with the rain, and each time it is conducted in almost exactly the same words.
One side says the hills burned because we stopped managing them. Grazing pulled back, burning restricted, fuel building up on ground that used to be worked. The other says the hills burned because we drained and burned and grazed them into a dry, degraded, flammable state, and the answer is to restore them.
I have a good deal of sympathy for both, and I want to say so before anything else, because the people making these arguments are not fools and they are not lying. Hill farmers watching bracken and rank vegetation take ground they used to be able to manage are describing something real. Ecologists watching drained peat go up in smoke and release a century of carbon are describing something real. Fire crews asking for tools to reduce fuel are not a lobby; they are the people who have to walk into it.
What neither side is describing is the hill as it is today.
Because both arguments are about a destination. Keep the system we have, or restore the system we lost. And almost no ground in Britain is in either of those states. Most of it is somewhere in the middle, halfway through a change nobody designed, and the middle behaves differently from either end.
That is where the nuance has gone missing. It is also where the fires are actually happening.
So let me set the ideology aside and describe what determines whether a hillside carries fire. Not what it ought to be. What it does.
There are four processes, and every one of them is something we can influence.
The first is water — not how much falls, but where it sits.
A functioning upland holds water near the surface. Sphagnum and the moss layer act like a sponge; the water table sits high; the peat below stays saturated. Rain arrives and is received rather than shed.
We spent much of the last century cutting drains through that. Hundreds of thousands of miles of grips, much of it publicly funded, to make wet ground carry more stock. A drained peat stops building and starts oxidising, and the moss layer is the first casualty. Once the sponge is gone the ground has no reserve. It dries from the top down, and it dries fast.
This is the deepest lever we have. It is also the slowest; water tables take years to decades to come back up.
Of the four processes, this is the one with the strongest evidence behind it. The comparative work on how differently wet and dry ground burns is about as settled as anything in this field gets, and we have collected it in the companion paper for anyone who wants to check it.
The second is age — specifically, how much of the plant is already dead.
Heather has a life cycle. It comes away young and dense, matures, and then somewhere around thirty or forty years it begins to break up in the middle: central branches die back, gaps open, and the proportion of dead woody material in the canopy climbs steadily.
That dead fraction is what ignites. Not the green growing tips — the grey brittle stuff in the middle of the bush and the dry papery litter underneath it. Put your hand into almost any moor in August and you will find it.
Both current systems produce a problem here, from opposite directions. Rotational burning holds everything young, which reduces the dead fraction, but only by resetting a whole block to the same age so that it ages back up together. Stopping management altogether lets everything go degenerate at once. Either way you end up with one age class across a wide area.
Anyone about to tell me that rotational burning is therefore the answer should read the Scottish regulator’s own review of the muirburn evidence base, which found rather less under it than either side of this argument assumes. It is in the paper, along with the strongest arguments on the other side, which are better than they are usually given credit for.
The third is diversity — because different plants dry at different rates.
This is the one people find least intuitive, and it may be the most important.
Bilberry, cotton grass, sphagnum, heather, birch and the various dwarf shrubs root at different depths, hold water differently, and give it up at different speeds. A hill made of one plant dries as one thing. A hill made of eight plants dries unevenly — parts of it still holding moisture when the rest has gone. That unevenness is not decoration. It is the difference between a fuel bed that is uniformly available and one that is only patchily available.
I should be straight about this one. There is good measured work showing that variation in fuel moisture across a landscape creates real discontinuity in what a fire can burn, and that it matters most in spring, which is our worst season by area burned. There is not yet a study directly comparing dead fuel loads in diverse versus even-aged heath, which is the specific test this argument deserves. The paper says which is which. I would rather show you the edge of what we know than pretend there isn’t one.
The fells the old shepherds described to me — deep heather and bilberry, buzzing with life — were hills that dried unevenly.
And this is the process we have most thoroughly dismantled, by two methods working together.
Fire arrests it directly. That is what a burn rotation is for: to hold the vegetation at an early stage and prevent it developing into anything more complicated. Nothing woody gets through, the moss layer never rebuilds, and the sward stays at whatever age the rotation dictates.
Sheep arrest it more quietly, and only when they are set-stocked without proper shepherding. A sheep is a selective grazer with a small mouth and a preference. Left on the same ground all season, every season, it will find the palatable, nutritious plants, eat them, find the regrowth, and eat that too. Year after year the plants that taste good are weakened until they disappear, and what thrives is what nothing wants — bracken, rush, purple moor-grass, mat grass — growing ever more rank.
That is not what sheep inherently do, and it is emphatically not an argument that there are no shepherds any more. There are, and they are as skilled as they have ever been. Flocks are still hefted. Gatherings still happen, and anyone who imagines that is easy work has never tried it.
What has gone is narrower and more specific than that, and it is three things at once.
The first is the routine moving of stock around the hill through the season, so ground could be rested before it was grazed again. That was the practice that stopped any one species being grazed out, and it is the part that no longer pays for itself. Not because anyone stopped knowing how, but because headage payments rewarded animals rather than management, and the margins hill farming now runs on will not fund someone to spend a season shifting flocks across a mountain.
The second is that there are more sheep on the hill than the old system carried.
And the third is that there are far fewer cattle and ponies alongside them — the less selective animals that will take the rank stuff sheep leave behind, and the heavy ones that break the surface and press dead material down into it.
It is the combination that does the damage. One selective grazer, at higher numbers, with almost nothing eating what it will not, and no rest between visits. Any one of those on its own is manageable. Together they take the complexity out.
The fourth is continuity — because a fire needs somewhere to go next.
A fire that cannot spread is an incident. A fire that can spread freely is a catastrophe. Everything in between is a question of whether the front, arriving at the edge of a patch, finds something it can move into.
Interruption is the whole game. Wet flushes, bog pools, woodland in the folds, vegetation of a different age or a different species, tracks, breaks. Every one of those is a place where a fire has to work harder, or stop.
Now look at those four together, because the important thing is that they are not four separate problems.
They are all outputs of one thing: whether the system is being held simple or allowed to become complex.
Simplify a hill and all four go the wrong way at once. The water leaves, because there is nothing holding it at the surface. The age classes collapse into one. The species collapse into one or two. And the continuity becomes total, because nothing is left that is different from anything else.
Complexity does the opposite, on all four, at the same time. This is not a shopping list of interventions. It is one condition with four symptoms.
Which brings me to the history, and I want to handle it fairly, because both camps invoke tradition and neither has much claim on it.
The Highland hills once ran on a seasonal system. Stock went up around Beltane and came down around Samhain; people went with them and lived up there through the summer at the shielings, making butter and cheese. The animals were mixed — cattle mostly, but sheep and goats and horses too. Groups grazed an area for weeks and were then brought down while another went out. The stocking was low. The hill was in use for a few months and recovering for the rest of the year.
In the Lakes the same intelligence took a different form: the hefted flocks moved around so the fell could freshen up, cattle and ponies alongside the sheep, and down in the valley the rule that you would not let stock hear the church bell strike twice in the same field.
Both systems were then replaced, twice, in two directions.
From the 1790s the Highlands were converted to commercial sheep, because wool prices were rising and a sheep run paid several times what small tenancies did. The people went, and the cattle went with them, and the hill stopped being empty for most of the year. Then when wool collapsed later that century, driven grouse shooting arrived — the breech-loader, the railways, Balmoral — and estates began burning heather in short rotations to produce a particular age structure for a particular bird.
Modern hill sheep systems are around two hundred and thirty years old. Driven grouse management is around a hundred and seventy. The intensity of the burning is younger still; it stepped up sharply in this century, within the working lives of people currently on the hill.
Every upland has its own accent on this. In the Welsh mountains the story is mostly sheep, and the signature is Molinia and bracken taking the ffridd and the open hill. In the Peak District, the South Pennines and the North York Moors, grouse is the dominant hand — and in the South Pennines the moss layer had already been stripped by industrial-era pollution before anything else happened to it, so the ground started further back. Dartmoor and the Lakes carry the commons and the ponies and a different governance problem. Different histories, different vegetation, different politics.
Same three moves. Drain it so it carries more. Burn it so it stays young. Graze it continuously with one animal so it stays open.
Two opposing land uses, two centuries apart, several regions, arriving at more or less the same landscape. Which is why it matters rather less, for fire, which of them you happen to be defending.
And I would say here what I have said about the Lakes for years. This is not a story about anybody’s negligence. Farmers and estates were responding to the prices and the policies of their day, and those prices and policies were set by all of us, because we wanted cheap food and we wanted the countryside to look a particular way. The responsibility for the state of our uplands sits squarely with all of us.
Now the middle. This is the part nobody is talking about and the part I spend most of my working week in.
When someone decides — entirely reasonably — that this has to stop, what usually happens is that the sheep come off or the numbers fall hard, the burning stops or is restricted, and then nothing else happens. Because blocking grips costs money that most upland ground will never see.
Stopping is not the same as starting.
Fuel begins accumulating the season you stop. The water table takes years to decades to recover and the moss layer longer still. So there is a window — a decade, maybe two — in which the ground carries more standing dead material than it did under management and is still as dry underneath as it ever was. Three of the four processes are unchanged and one of them has got worse.
And into that window, on the wetter ground, comes purple moor-grass. It builds a thatch of its own dead leaves that stands up rather than lying down, so it dries in the wind and stays dry. It burns fiercely. Its growing points sit inside the tussock where fire cannot reach them, so it comes back afterwards with more vigour than before. It is a plant that is good at being burned, and it arrives because the animal that was holding it in check has gone and nothing has replaced it.
None of this is a surprise to anyone who has looked. That reducing stocking on already-degraded heath produces purple moor-grass rather than heather was published in 2008, and it has been sitting in the literature ever since, largely unread by the people it is most relevant to. It is in the paper.
This is what most of the ground everybody is pointing at is actually doing. And both camps are reading it as a verdict on the destination.
The farming side sees fuel piling up on destocked hills and concludes that removing management causes fires. The restoration side sees drained, burned, degraded moorland going up in flames and concludes that management causes fires. Both are looking at half-finished ground and calling it a finished argument.
I have sympathy with the frustration on the farming side here, and it is worth naming plainly. A tool has been restricted and the mechanism meant to replace it does not yet function. The licensing system in England has approved almost nothing; Scotland’s new muirburn regime lands next month after two delays. If you are the person standing in front of that hill in April, that is not an abstraction.
But the answer to a badly designed middle is to design the middle. It is not to go back.
Because here is what neither system has, and it is the process we have not talked about yet.
Before we simplified any of this, these hills had wild herbivores in them, and they did not behave like livestock. They moved in herds, because a herd is how you survive predators. They grazed hard and fast, taking what was in front of them rather than picking through it, and moved on. They trampled a great deal of what they did not eat, pressing standing dead material down onto the soil where it could rot in rather than stand up and dry. They dunged as they went, and their feet broke the surface enough to make places for seed to land.
And then they left. Not for a fortnight — for months, because the herd had gone and their own fouling kept them from coming back too soon.
Dense impact, then long absence. A pulse, not a hum. That is what built the mixed-age, mixed-species, moisture-holding vegetation we are now trying to reconstruct from fragments. And the shieling summers, and hill shepherding, were human versions of the same thing: animals in the right place at the right time, moved before they did damage, then gone long enough for the ground to come back.
Neither of our two modern options reproduces it.
Set-stocked sheep are a constant, low-grade signal. The animal is always there and never impactful, and it never allows full recovery. And the lower the numbers go without changing the pattern, the worse it gets, because the grazing stays just as selective and there is even less impact on the things nothing eats.
Removing the herbivores altogether is also a constant signal. It is just a constant of zero. No grazing, no trampling, no dung, no disturbance, no germination sites. Whichever plant is most competitive in the absence of anything that eats it simply takes the ground.
I understand the instinct behind the second one. When you have watched a landscape damaged by animals for two centuries, taking the animals off feels like the obvious repair. But there is nothing natural about a hill with no herbivores on it. Every temperate system on this island evolved with them moving through it. Take them out entirely and you have not restored the system — you have removed one of its central processes and left the rest to sort itself out.
Set-stocking is a constant signal. Removing the animals is a constant of zero. The system evolved with neither.
And I would say the same, more gently, about most conservation grazing. Putting a few animals on a large block at low density is closer in intention, and it is not nothing. But it is not mimicking the natural process either, because what made wild herbivory work was predation — the bunching, the urgency, the moving on, the long absence afterwards. A handful of cattle drifting across a big unit for six months are still a constant signal. They still select. They are simply selecting a little more slowly and over a wider area. That gets you a slightly better outcome than no stock at all, and nothing like the outcome the process is capable of.
The animal is not the point. What the animal does, and when it stops, is the point.
So what we work towards is neither camp’s version, and it is not a reinstatement of the shieling system either. You cannot rebuild a system that depended on a population of people who are no longer there.
But you can put back what it did.
Restore the water first and hardest. Deep peat protected and rehydrated — grips blocked, water table up, sphagnum back. On the much larger area of ground that restoration funding will never reach, the grazing design has to do the work the money is not doing.
Use cattle and ponies as the primary herbivores rather than sheep. They are less selective, they will take purple moor-grass in summer when sheep will not, and they are heavy enough to trample. And use them the way the system evolved with them: concentrated, brief, then gone long enough for full recovery. Virtual fencing has made that genuinely possible on open hill for the first time — not a replacement for a shepherd but a way of giving one the reach that used to take several, at a cost a hill business can actually carry. Sheep stay in the system wherever their timing and duration can be properly controlled. This is not an argument against sheep. It is an argument against sheep at large on an open block for the whole season.
Let succession run instead of arresting it. Mixed ages, mixed species, bilberry and the whole dwarf shrub complement rather than heather alone, moss back in the ground layer, birch and willow coming through where they will, and real closed woodland in the folds that can carry it.
And place firebreaks deliberately — on the access routes, at the ignition hotspots, along the edges of what you most need to protect. Tracks used as breaks and as stock corridors. Wet features and ponds as landscape-scale interruptions. This belongs in the design, and fire crews are right to ask for it.
Read that list back against the four processes and you will see it is one intervention rather than five. Water back. Age classes spread. Species diversified. Continuity broken. All four move together, because they were only ever one thing.
What you end up with is a hill you cannot describe in a single sentence. Heather at four or five different heights. Birch and willow in the shelter. Bog pools and wet flushes. Moss in the hollows. Several dwarf shrubs rooting at several depths. Woodland in the folds and open ground on the tops.
Something, in other words, remarkably close to what the old accounts describe.
And I want to be clear about what that hill is, because this is where I lose people on both sides. It is not a landscape with the farming taken out of it. It is a working hill, managed deliberately and intensively — intensively in attention rather than in inputs — and it is producing food.
Managed that way, over time, it will carry more livestock than the same ground set-stocked, and considerably more than the same ground under low-density conservation grazing. That is not a paradox. You are growing more forage, of more kinds, over a longer season, on ground that holds its water instead of shedding it — and the animals are healthier on it, because a diverse sward gives them something like the range of nutrients they actually need. A declining sward rationed out among fewer and fewer animals is not a lighter touch. It is a slow liquidation, and it produces less every year.
Better outcomes for the hill and more production from it are not in tension here. They are the same thing, arriving together. That is the part both camps find hardest to believe, and it is the part I have watched happen most often.
That hill still burns under a severe enough drought — wet ground burns too when conditions are extreme enough, and anybody promising fireproof is selling something. But it does not burn as one event the size of the hill, because a fire arriving at its edge does not find a mile of the same dry thing waiting.
I have kept the citations out of this piece deliberately. The companion paper carries them — the sources, the caveats, the strongest version of each camp’s case, and an honest list of the things nobody has measured yet. If you want to argue with me, argue with that.
But I will say this, having watched the argument come round for three decades. Both camps are holding one true thing and using it to avoid the other’s true thing. Meanwhile the hills are neither of the two landscapes being argued over, and the middle years — which is where every estate and every hill farm in this country is actually standing — have no plan attached to them at all.
We are not trying to build a living museum. Nobody is going back up to the shielings for the summer. But the people who kept those summers, and the shepherds who slept in huts on the fells, understood something we have spent two hundred years unlearning twice over.
A hill is not improved by being asked to carry the same thing all the time.
It needs the water back in it. It needs the succession allowed to run. And it needs the animal returned to doing what it always did — hard, briefly, and then gone.
Written in dialogue with Claude (Anthropic). Ideas, arguments and experience are mine; AI helped research, structure and articulate them.
Caroline Grindrod has spent nearly thirty years in environmental land management and regenerative agriculture, from lowland production beef enterprises to Cumbrian fell farms and Highland estates.
Wilderculture CIC · info@wilderculture.com · wilderculture.com
No Comments