Agriculture has been told a simple story for years. The world is getting warmer, so the main risk to food production must be heat. Prepare for drought. Prepare for heat stress. Prepare for water scarcity. Prepare for crops moving north. There is truth in that story, but I do not think it is complete enough. In some places, it is dangerously incomplete.
The problem is that crops do not grow inside climate models. They grow in fields, orchards, vineyards, plantations, valleys, hillsides, and frost pockets. They grow through awkward seasons, false springs, cold nights, wet soils, late snow, fungal outbreaks, and short windows where one bad week can destroy the year. Averages do not harvest grapes. Averages do not protect coffee trees. Averages do not stop blossom from freezing at 3am.
That is where the climate conversation often becomes too narrow. Agricultural companies, policymakers, insurers, and investors have been encouraged to think in terms of warming. But many of the most damaging recent agricultural shocks have come from volatility. Warmth arrives too early. Buds break. Plants lose dormancy. Then cold returns, and the damage is worse because the crop has been fooled into vulnerability. That is not the simple heat story. That is a volatility story.
One of the most dangerous patterns in modern agriculture is the false spring. A warm spell arrives early. Crops respond as nature designed them to respond. Vines bud. Fruit trees flower. Coffee plants move into a sensitive stage. Then a cold outbreak arrives late enough to strike the crop when it is most exposed.
This is exactly why frost can be so destructive in a warming world. It is not always the absolute cold that matters. It is the timing. A vineyard in full dormancy can tolerate cold that would destroy it weeks later. A fruit blossom can be killed by a short freeze that would have meant little before bud break. Coffee trees can suffer structural damage from frost that affects not only one harvest, but future production.
France’s 2021 wine frost is a useful example because it was so visually dramatic. Fires and candles burned through the night in vineyards as growers tried to protect young shoots from freezing temperatures. Government estimates at the time suggested French wine production could fall by 24 percent to 30 percent from the previous year after spring frost and later summer weather damage. That is not a marginal inconvenience. That is an agricultural shock with regional, commercial, and cultural consequences. The key point is not just that it was cold. The key point is that warmth came first, and frost came after. That sequence is what made the damage so severe. 
I think this is where many climate discussions fail farmers. They talk about warming as though it moves in one clean direction. Farmers know better. They live in the messy middle, where a warm March can make an April freeze more destructive, not less.
Brazil’s 2021 coffee frost showed the same pattern in a different agricultural system. Coffee is often discussed in terms of heat, drought, and shifting suitability zones, which are all real issues. But frost remains one of the most damaging risks to coffee production because it can harm the tree itself, not just the fruit.
In July 2021, unusually cold weather hit Brazil’s coffee belt, including key areas of Minas Gerais and São Paulo. Reuters reported that preliminary Brazilian government estimates suggested frosts had affected 150,000 to 200,000 hectares, around 11 percent of the country’s total arabica crop area. Another estimate cited a possible 4 percent loss to Brazil’s next coffee crop, while some regional assessments suggested heavier localised damage. 
The numbers matter, but the geography matters more. Frost does not spread evenly across a coffee region. It collects in low-lying areas. It follows elevation, slope, exposure, air drainage, and local topography. Two farms close to each other can experience very different outcomes because cold air behaves like water. It flows downhill and settles. If a plantation is in the wrong pocket, it can be hit hard while another area escapes.
That is why I think plantation risk needs to be mapped at a much finer scale than many broad climate narratives allow. Saying a region is getting warmer does not tell a coffee company whether a particular valley remains exposed to damaging frost. Saying climate change will move coffee uphill does not tell a grower which hillside is safer, which road can support expansion, or which existing fields should be replanted with more frost-resistant varieties.
The map is not a decoration. It is the risk model.
Cold damage is not only about frost. In many agricultural systems, cool and damp conditions create a second wave of losses through disease. Wet leaves, poor sunlight, delayed flowering, waterlogged soils, and reduced field access can turn a mediocre season into a disastrous one.
French wine production in 2024 offers a useful reminder. France cut its wine output estimate after wet weather, with production projected around 22 percent below the previous year and 15 percent below the five-year average. Reports pointed to rain, frost, mildew, hail, and poor flowering across wine regions. That is the real agricultural world. Not one clean variable. A chain of stress. 
This matters because carbon-focused climate policy can push attention toward the wrong dashboard. Emissions matter. Heat matters. Drought matters. But disease pressure, field access, cold snaps, and damp growing conditions matter too. If advisory systems become too focused on warming averages, they may underprepare producers for the practical hazards that actually destroy yield.
A potato farmer dealing with ground frost and rot does not care whether the annual average temperature is up. A grape grower watching mildew move through a vineyard after wet weather does not care that the long-term climate trend is warmer. A citrus grower looking at freeze-damaged trees does not care that the policy paper said heat was the main risk.
This is not an argument against climate planning. It is an argument against lazy climate planning.
The 2021 Texas freeze is another useful case because it struck a region not popularly associated with severe cold. Winter Storm Uri brought freezing temperatures deep into Texas and caused huge disruption across power, water, transport, and agriculture. In the Rio Grande Valley, citrus growers suffered major damage. Industry reporting later cited estimates that around 10 percent of the state’s citrus trees were killed or damaged, with roughly 24,000 acres remaining in production compared with about 27,000 acres at the time of the storm. 
Citrus is a good example of how agriculture depends on thresholds. A few degrees can separate manageable stress from serious damage. A few hours can matter. A crop can survive a brief dip and fail under prolonged freeze. Irrigation systems, wind machines, grove location, tree age, and local cold pooling all influence the outcome.
What strikes me about these events is how often they expose hidden fragility. A system can look adapted to its normal climate and still be poorly prepared for volatility. A region can specialise successfully for decades, then discover that its infrastructure, insurance, and emergency planning are not designed for the wrong kind of weather.
That phrase matters: the wrong kind of weather. Not because it is unexpected in some absolute sense, but because it sits outside the institutional story people have been told to prepare for.
The Maunder Minimum is worth introducing carefully because it is often dragged into climate debates in a sloppy way. It refers to a period from roughly 1645 to 1715 when sunspots were exceptionally rare. This period overlapped with some of the coldest decades of the Little Ice Age in parts of Europe and North America. It was associated in popular historical memory with harsh winters, shorter growing seasons, frozen rivers, and crop stress.
It would be careless to claim that a new Maunder Minimum is definitely coming or that solar cycles alone explain present agricultural risk. That is not my argument. The point is more sober. The history of the Maunder Minimum reminds us that agriculture has always been vulnerable to cold, shortened seasons, and climatic instability. It reminds us that food security can be damaged by cooling shocks as well as heat shocks. It reminds us that societies often underestimate cold risk when they become comfortable with a dominant climate narrative.
The lesson is not that the world is about to repeat the seventeenth century. The lesson is that agricultural resilience must include the possibility of colder disruptions, not only warmer averages. A farm system that prepares for drought but ignores frost is not resilient. A plantation strategy that models long-term warming but ignores cold-air drainage is not complete. A policy regime that encourages decarbonisation but neglects food system volatility is solving one problem while leaving another exposed.
I think the Maunder Minimum matters because it breaks the arrogance of the present. It reminds us that climate is not a straight line drawn through a policy document. It is a complex system that can hurt food production in more than one direction.
Modern agricultural policy is increasingly shaped by carbon. Land use emissions. Fertiliser emissions. Methane. Soil carbon. Carbon reporting. Carbon markets. Carbon border rules. Supply chain disclosure. Some of this is valuable. Some of it is inevitable. But it can also distort priorities.
A producer under pressure to report carbon exposure may spend more time measuring emissions than mapping operational risk. A food company may build a sustainability framework that looks impressive in a boardroom while its growers remain exposed to frost pockets, snow cover, fungal pressure, or cold-related planting delays. An investor may reward a company for climate pledges while ignoring whether its plantations are geographically resilient.
This is where I think the climate conversation has become too bureaucratic. It often rewards the appearance of climate awareness rather than the reality of climate preparedness.
Real preparedness is not a glossy emissions statement. It is knowing which farms are vulnerable to late frost. It is knowing which roads become impassable after snow or heavy rain. It is knowing where disease risk rises after damp conditions. It is knowing which crops are too concentrated in one climatic zone. It is knowing whether insurance models reflect volatility rather than historical averages.
The most dangerous risks are often the ones that do not fit the reporting template.
If I were advising an agricultural company, I would start with exposure rather than ideology. Where are the crops. What are the thresholds. What weather events can destroy them. What is the timing sensitivity. Where has volatility increased. Which fields, orchards, or plantations sit in frost-prone terrain. Which areas suffer from poor drainage. Which transport routes fail under snow, flood, or cold.
Then I would map the operational chain. Seed, fertiliser, labour, irrigation, harvesting, storage, processing, transport, export. Cold weather does not need to destroy the crop directly to cause losses. It can delay planting. It can stop field access. It can damage roads. It can reduce labour availability. It can disrupt processing energy supply. It can increase disease pressure after the event.
This is where GIS becomes essential. Not as a map-making exercise, but as a way of seeing risk before it appears in the financial statements. Satellite data, elevation modelling, soil moisture analysis, weather station records, crop calendars, disease models, and historical yield data should be brought together into one spatial framework. The goal is not to prove a political point. The goal is to protect production.
Agriculture needs less narrative and more exposure mapping.
The title of this briefing is deliberately blunt because the mistake is expensive. If agriculture is told to fear only heat, then it will prepare for only part of the threat. It may invest in drought measures while neglecting frost protection. It may relocate crops based on average warming while missing local cold-air dynamics. It may comply with carbon regulation while remaining vulnerable to the next cold, damp, disease-heavy season.
The correct frame is not warming versus cooling. That is too crude. The correct frame is volatility. More unstable seasons. More damaging timing. More compound events. More situations where the crop is pushed out of rhythm and then punished.
That is why frost matters. That is why snow cover matters. That is why damp conditions matter. That is why fungal disease matters. That is why the Maunder Minimum still matters as historical memory, even if it should not be abused as a simplistic prediction.
Agriculture is not threatened by one neat climate story. It is threatened by the failure to understand the full range of climatic stress.
The practical conclusion is simple. Food producers should prepare for heat, but they should not be hypnotised by it. They should prepare for drought, but not forget frost. They should reduce emissions, but not confuse compliance with resilience. They should listen to climate science, but not reduce it to averages.
The crop does not fail in an average year. It fails in a particular place, in a particular week, under a particular combination of conditions.
That is the geography of agricultural risk.
And it is the part of the story that too many people still prefer not to see.