Farming through the storms

There is a stretch of lane on this farm where the season arrives underfoot before it arrives anywhere else. In a wet June the gravel goes soft and takes the print of a tire and keeps it. In a dry June the same lane rings with every step, and the dust comes up pale and fine and settles along the fencerow like ash. By the middle of that month you know which year you are standing in, and you knew it before you looked at a field.

A farm learns weather the way it learns everything else, which is by repetition, from somebody who learned it the same way. Which week the frost lifts. Which fortnight the flies come. How late in the season you can push hay before rain starts living in the forecast. That knowledge was never written down anywhere, because it did not need to be. It was the inheritance, and it was handed on the way a good fence is handed on, which is by walking it with the person who built it.

The inheritance worked because the pattern held long enough to be worth learning. That is the part that has changed.

The same province, in the same month

Start with one month, in one province, because a decade of climate data is easy to argue with and a single summer is not.

In July 2026, Edmonton recorded 324 per cent of its normal rainfall. Regions south of Edmonton received less than 40 per cent of normal, in the same province, in the same four weeks. Several counties in central Alberta declared an agricultural disaster on account of flooding and localized crop damage, while the south of the province dried out. The federal government's own agroclimate report for mid-August put it in one line, which is that recent flooding in Ontario and early-season flooding in Manitoba were concerns at the same time as drought conditions in British Columbia were the most significant climate-related agricultural risk in the country. By the end of July, 23 per cent of Canada's agricultural landscape was abnormally dry or worse, up from 13 per cent in June.

That is the thing a farm cannot prepare for by choosing a side. A farm used to be able to ask which kind of year it was going to have, and the answer stayed the same for a season. Too dry or too wet was a different set of decisions, and you made one set. What the record now shows is both, arriving in neighbouring counties and sometimes in the same one: the mechanism, in the federal climate science, is that precipitation is concentrating into intense and infrequent downpours separated by longer dry spells. It rains harder, less often, and not where you needed it.

The old preparation was for a shape of year. The new problem is that there is no shape.

Two years on two farms

Simon Ellis farms near Wawanesa, Manitoba. He is a fourth-generation farmer, and the 2025 season on his place ran flooding first and then a long drought. When he drove the combine into that crop he expected, in his word, "catastrophic failure." What came out of it instead were plump seeds of wheat, oats and soybeans.

He does not describe this as a triumph or a technology. He describes it as maintenance. The investments that carried him through were minimum and zero tillage to keep the soil where it belongs, tile drainage, which is an underground system that moves water off a field it would otherwise sit on, slow-release fertilizer, and an agronomist's advice on weedkillers. "We are constantly making little tweaks," he said. "That's how we're going to be able to keep fighting the changing climate."

Now put the same season under a different farm. Quinton Jacksteit farms in southwest Saskatchewan, near Golden Prairie, and is the reeve of the Rural Municipality of Big Stick. His 2025 was the ninth year in a row of drought. His crops were extremely short and he planned to salvage most of them as animal feed. "They're not going to be able to make much of a yield," he said.

Two farmers, one country, one season, and the difference between them is not skill or effort. It is position, and underneath the position it is water.

Little tweaks, not miracles

The word that keeps turning up in the accounts of adaptation is a small one. Ellis says tweaks. Almost nothing in the literature of farm adaptation is a transformation, and almost all of it is maintenance that compounds.

South of Prince Albert there is a dairy farm outside Birch Hills, Saskatchewan, where the difference was visible from the road. Paul Kernaleguen's fields stood waist-high in the 2021 drought beside neighbours' fields that had gone yellow and short, and the reason was that he had stopped growing one crop in a field at a time. His fields are now a mix of cereals, pulses, clovers and turnips, chosen because their roots sit at different depths: fibrous shallow roots that grab rain quickly, deep tap roots that reach subsoil water when the top is dry. He reports having doubled the organic matter in his soil, and organic matter behaves, as researchers put it, like a bigger bucket. "It's these type of years that you can see the drastic difference," he said.

None of that requires a sensor or a subscription. It requires having made the change some years earlier, in a normal year, when it looked like an unnecessary expense.

That is the arithmetic that is hardest to see from outside a farm. Adaptation is not a purchase you make in the year you need it. It is a series of decisions made in years you did not need it, which is exactly the kind of decision a farm under financial pressure stops making first.

An average crop is not enough

Here is what the changing normal does to the money.

Stephen Vandervalk is a fourth-generation wheat grower in southern Alberta and vice-president of the Wheat Growers Association. His 2026 crop was not bad, and it was not good. He described the year's crop as "by far, with no comparison ever in history, the most expensive crop that farmers have put in the ground," which is a careful way of saying the inputs cost more than the crop was worth.

His own summary was blunter. "You get an average crop, you are losing money hand over fist. It's a pretty scary thought when your average isn't nearly good enough."

An average crop used to mean you continued. It now means you probably did not, which is the plainest description available of what adaptation as survival means: the bar has moved past the middle. Rob Saik, an agronomist who has advised governments on several continents, put the same shift in bushels. "Back in the day, 30, 35 bushels an acre (for wheat) would have been a bumper crop," he said. "Now it's an abject failure."

The cost side of this has a number attached. Crop insurance payments in Canada went from 890 million dollars in 2018 to 4.9 billion dollars in 2022, with drought among the significant drivers. That is a public record of private losses, and it is the mechanism by which a bad decade on the land turns into a line item in somebody else's budget.

And then there is the year that no amount of preparation can absorb. On the fourth of July, 2026, a hailstorm crossed three-quarters of John Kowalchuk's wheat, barley and canola acres in central Alberta and probably halved the yield on those fields. "It was looking really good until the hail," he said. None of that would have changed the outcome for him, and a book about tools should say so.

Where the water rule is written

If a farm cannot control the sky, it can at least understand who controls the water, because those are two different questions and only the second one is written down.

Alberta has 2 per cent of Canada's water supply, and in March 2026 the province proclaimed the first substantive update to its Water Act in more than 25 years, framed as making more water available to farmers and ranchers while keeping environmental protections in place. Buried in the framework, and reported on the government's own water management pages, is the principle that governs everything under it: licences issued earlier have priority over licences issued later, regardless of what the water is being used for. First in time, first in right.

That is a queue ordered by the date somebody applied. It means the farm with the oldest licence has the strongest claim on a short river, no matter how good a case a newer neighbour has, and no matter whether the newer neighbour is growing food that people need more. The queue is the allocation.

The water rule on this continent is not open source and not closed source. It is administrative, it is old, and it is made of dates. You can read it, and you can find out where your farm sits in it, and that is a more useful piece of literacy than most dashboards.

What a farm can still run itself

There is a real open layer in water, and it is narrower than the enthusiasm around it.

At the valve there is OpenSprinkler, an open hardware and firmware controller under a public licence, in production across a lot of horticulture and smallholding. Its specific virtue is that weather-adjusted irrigation scheduling, the kind that decides whether to water based on what the sky is about to do, does not require a company's servers. The comparison is not abstract: Hunter, Rain Bird and Rachio sell the same function with the reasoning happening somewhere you do not control. A farm that has watched a vendor turn off a service knows the difference already.

In the model there is AquaCrop-OSPy, an open implementation of the UN Food and Agriculture Organization's crop-water model. This matters more than it sounds. When a dry year makes the local rule of thumb wrong, the useful thing is not a number but a way of reasoning about water, and an open model is a piece of reasoning the farm can inspect, modify, and keep.

And then the edge, stated plainly: above the scale of a garden or a small holding it stops. Broad-acre canal automation and centre-pivot telemetry remain closed industrial systems, and the search for a verified open-source large-scale pivot control stack came up empty. Irrigation districts in the American west now use open evapotranspiration data to make allocation decisions, and there is no Canadian equivalent. The open layer in water is real at the valve and at the model, and absent exactly where the biggest water decisions are made.

The place that already built for this

There is one part of this country where the resilience problem was never theoretical, and it shows what the answer looks like when nobody can pretend otherwise.

In Nunavut, intermittent connectivity and fragile supply chains are the ordinary condition rather than a risk to be modelled. So the systems that work there are built the other way around: they store data locally and forward it when a connection appears, they run on community machines, and they keep working when nothing outside the community answers. Paper fallbacks are not a degradation, they are a layer.

The finding from the research is blunt and it belongs in a chapter about weather: the northern food system did not choose that architecture out of ideology. It was forced into the design the southern open movement spends its time advocating. Which suggests the useful question for a farm in Lanark County or the Okanagan is not whether to build for interruption. It is how many bad seasons you get to have before the interruption arrives on its own.

Staying with the trouble, together

Donna Haraway, the feminist theorist and historian of science whose book Staying with the Trouble sets the terms for a lot of the thinking on climate and care today, makes one demand this chapter has been circling. Face the trouble without either of the two exits people take from it, which are the doom and the denial. The doom says the farm is finished. The denial says the next version of the tools will fix it. Both are ways of not having to decide anything on Monday.

Her book earns its title the long way, and the ideas in it are worth more to a farm than the phrase alone lets on. Haraway replaces the word making with one she writes sympoiesis, making-with. Nothing on a farm is made alone. A crop is a joint project between seed, soil life, weather, and the person who decides what to plant and when. A farm that survives a hard year is a project between neighbours, a veterinarian, a supplier, and a lender, whether or not any of them would put it that way. The trouble is not solved by the lone operator gritting through it, and it is not solved by a single machine doing the same. It is carried, or not, by everything the farm does with.

The second idea follows from the first, and it has the chapter's question inside it. Haraway writes the word responsibility as response-ability, the capacity you build to answer what is actually in front of you. It is closer to competence than to guilt. It says that what you can respond to is not fixed, that it grows with knowledge and with practice, and that it grows fastest when it is made together. That is the chapter's answer to the reader who finishes a description of volatile weather and asks what to do about it. You do not wait for a plan that works on its own. You extend the range of things you are able to respond to, in the company of people doing the same, and you make it more literate and more shared every year.

Begin with the register itself. Rob Stone, who farms in central Saskatchewan, described the actual situation better than any report I read, and he did it with two images. "We're used to the blast furnace. We aren't used to a (steamy) sauna." The old enemy was heat and dryness, and a farm knew how to stand in it. The new conditions are humid heat, which stores energy, which is why the same warmth that helps a crop also produces tornadoes and hail that take it away. Naming that is not doom. It is the first response.

Virginia Maier, the reeve of Enterprise, Saskatchewan, is standing in a drought bad enough that she said, plainly, "Everything is short, everything is dying," and then, about how long it had been going on, "When (drought) goes on for so many years, it's starting to get depressing." And then she said the thing I keep coming back to, which is that she still expects to be all right, and that the real question on the table belongs to her neighbours, some of whom are asking "Do I just get out and sell the land and rent it and just be done with it?"

That is the register. It is not optimism and it is not despair. It is a person with her hands in the work, telling you precisely how bad it is and what she intends to do next, and neither of those facts cancelling the other.

Haraway's word for what Maier is doing with her neighbours is kin. Making kin is the third idea, and it is the one this chapter has to name plainly because it carries the open-source argument. When the pattern stops holding, nobody adapts alone. The volatility is shared, so the way through it has to be shared too: tools, systems, and knowledge made in the open, so a fix built on one farm in a wet year is sitting where a farm in a dry county can find it the next. That is the ordinary sense of the word kin. The unusual sense is that Haraway means it across species, which a farmer already believes and the rest of us take a while to catch up with: the soil organisms, the bees, the seed, are partners in the response, not scenery for it.

Put the two ideas beside the chapter's earlier workings and the picture closes. OpenSprinkler is response-ability with the company left the field: a water decision that keeps working when nobody outside the farm answers. AquaCrop-OSPy is response-ability made literate, the reasoning about water kept in a form a farm can inspect, modify, and pass on. And the more volatile the weather, the more valuable it is to harvest the forces being unleashed rather than only absorb them. The clearest version of that turn is agrivoltaics, growing under panels that also make power, where shade and electricity become two products of the same acre instead of a competition for it. It is a young field and the economics are still settling, so this book treats it as a direction rather than a promise. The point is the posture. A farm that stops asking how to defend against the weather and starts asking what the weather and the sun can produce together has changed the question from survival to harvest.

That is what staying with the trouble means in the concrete: not hope, not despair, but a growing capacity to respond, made together, and made to harness what cannot be avoided. The register Maier and Stone already stand in is the only one that gets a farm through. This book's wager is that the tools of that register are the ones the earlier chapters have been laying out: literacy, agency, and the habit of building in the open with other people.

Open questions

If adaptation lives in the tweaks made during normal years, what does a farm do when it cannot afford a normal year?

Does a queue ordered by the date an application was filed count as a rule, or as the absence of one? First in time, first in right has the virtue of being clear and the defect of being decided by paperwork from decades ago.

What would open infrastructure for a centre pivot even look like? Nobody has built it, and the scale is the likely reason.

If the northern approach is the correct approach, why is it the exception rather than the design standard?

And the one the book cannot answer: how many farms does the country lose before the losses stop being a statistic and start being a shortage?

Moves you can make

Find out where your water comes from and who decides about it. Not the general answer. Your watershed, your municipality or your district, and the actual rule under which allocation happens. If you farm, find your licence or your place in the queue and write down the date. If you do not, find out which body governs the river or the aquifer your town draws from, and whether it publishes what it decides.

Then check one thing you own that touches water or weather and ask the interruption question. If the company that made it stopped answering tomorrow, what still works? A controller that schedules locally still works. A model you can run on your own machine still works. A cloud subscription does not, and the difference shows up in the season you need it.

The situation: five years of weather where you live

Write down, from memory, what the last five years were like where you live. Rough is fine. Hot summer, wet spring, the year the basement flooded, the year the garden died, the year nothing happened at all.

Then go and get the actual record and compare. Environment Canada publishes historical weather data for stations across the country, including the one nearest you, and the numbers are free. Put your memory beside the measurements and mark every place they disagree.

You will find two things, and both are useful. The first is a surprise at how much you did not notice, because memory is a bad instrument for slow change: the year you remember as normal usually was not. The second is that the disagreements run in a direction. What people misremember is not random. The pattern people carry is the older one.

Keep the page. Almost nobody has one, and it is the only kind of document that gets more useful with age. A record will not stop a hailstorm or bring rain. What it does is tell you which of your assumptions about this ground were ever true, and that is not something you can buy in the year you need it.

The next gate

Every chapter so far has taken a system the farm uses and asked who controls it. This one has done the opposite, which is to describe a force with no vendor and no platform and no licence to revoke, and the reason it belongs here is what it does to the older promise.

The book argues that anyone who wants to farm should get to. Quinton Jacksteit has a couple of boys who want to farm, and this is what he says about it, in a year when he was scraping through his ninth drought: "it's not something that I would recommend to them at this particular time."

That sentence does not refute the promise. It changes what the promise costs. Resilience is an argument for more farms, run in more ways, on cheaper terms, and it is not an argument that farming is getting easier. The next chapter is about what stands between a person and a farm, which is a list that has been getting longer.