How AI can help farmers “climate-proof” their acres

ClimateAi uses a blend of proprietary machine learning-based forecasts/analyses and climate physics, along with public and private weather and agronomic data, to create accurate and actionable predictions and decision recommendations for growers.

Himanshu Gupta, founder and CEO of ClimateAi, says climate adaptation – growing the right crops in the right places and planting/applying inputs at the climate-optimal time – reduces water, fertilizer, and pesticide use, reduces GHG emissions and improves financial stability for local communities and agribusinesses.

So, how can growers climate-proof their acres, for this season and for the coming decade(s), and how can ClimateAi help?

“ClimateAi uses innovative climate forecasting technologies (one day – 40 years out) to help agribusinesses and farmers adapt to the changing climate, for the season and decades ahead. With our machine-learning approach to forecasting, we can find key climate risks for the next decade by specific 5-square kilometer locations.

The impacts of climate change – such as rising temperatures, changes in precipitation patterns, increases in extreme weather events, and reductions in water availability – will have unique but varied consequences for the agricultural sector in the coming years. The exact effects will vary based on crop type, variety, geographic region, and growing practice. As a result, understanding the specific key risks that will arise, dependent on each of these categories, is crucial to adapting in a climate-changing future.

For the coming decades, our tools don’t just give an average increase in temperature in a region, but instead forecast how temperature and precipitation evolve by month in a certain time frame. For example, our tools provide the different probabilities of if temperatures will hit over 35 degrees Celsius during July in say, Florida, when a tomato crop would be flowering and its quality would be vulnerable to high heat.

The forecast image shown below was delivered to an anonymous customer, and it visualizes how heat risk evolves over the next two decades for a selected specialty crop-location pair in the Northeastern U.S. Because the heat risk was too high, the customer canceled the investment, thus avoiding necessary and – without ClimateAi – unforeseen downside risk.

The customer instead used ClimateAi’s long-term strategic dashboard to identify and invest in crop-location pairs that are positioned to experience future climatic conditions that will boost yield, increasing upside potential.”

_{Text continues underneath image}

What are the crop- and location-specific risks facing an ag operation this year and over the next 10?

“Key risks are for example such as heat stress (extreme heat days and overall warming trends that affect growing degree days and nighttime temperatures), frost risk, extreme weather events (the 1-in-100-year storms or wildfires that are occurring at 1-in-10 year frequencies), drought risk, and soil/water quality degradation. Factors such as heat will directly affect quality of specialty crops such as strawberries in California. Extreme weather events such as storms in the Midwest will directly affect yield of crops such as sorghum and corn. Still, again, there are many intersecting factors that will determine the success or failure of crop production, because both plant physiology and climate change are complex.

Let’s look at one plant, the D’anjou pear, commonly grown in Washington state orchards — as an example of one crop that might find more success in a warmed world.

Similar to many fruit trees and certain flowers, this pear requires a certain number of chill hours to thrive (that is, the number of hours between about 32 degrees and 45 degrees Fahrenheit during the winter when the plant goes dormant, and after it reaches the required number, a biological alarm clock rings and the plant wakes up and blossoms or sets fruit). On the flip side, temperatures below 32 degrees are “frost hours,” and too many inhibit blossoming and fruiting.

But with global temperatures rising, winters are not as cold on average, so chill hours outnumber frost hours. We found that with only 500 chill hours in Washington, on average, pear yields go down 50%. With more chill hours and less frost risk, yields are expected to go up.

Still, despite less frost risk in a climate-changing world, there is still a strong risk of freak storms and cold snaps because of the increase in extreme weather events – which would be extremely damaging to these sensitive fruit trees. These risks, as well as chill hours, are huge issues for almonds, pistachios, and stone fruit in places like California’s Central Valley.”

How can your company help growers mitigate effects of climate change on their crops?

“Our platform offers an analytical decision-making layer that can help farmers understand how to respond to the risks, as they all depend on many factors. So, for pre- and in-season decisions, the platform can help optimise planting, input timing, harvest timing, pricing decisions, and contracting decisions with growers. In addition, it can recommend ramping up or down production in different areas including counter-season production to get the production needed (if it expects less yield in one place, it can double down in a place that is expecting to do better).

It also helps manage equipment logistics – for example, one of our co-op customers shares equipment across its farmer-members, and knowing when and where farmers across the country will need access based on climate-smart yield/harvest timing forecasts (delivered on a rolling basis, months ahead of time) can keep costs down and manage relationships.

For example, if we see a high likelihood of a heat stress event coming during flowering (temperatures over 35 degrees C, the threshold for X crop variety), we can advise farmers to move the planting date by two weeks to avoid the heatwave and still have sufficient growing degree days (GDD) to hit yield/quality.

On the mitigation-or-no-mitigation front, we are able to show the direct impact of future climate conditions on crops. There’s no question that overall it will get hotter within the next decade, but crops can stay on-location if the heat is concentrated in a month that doesn’t actually matter to the crop cycle, such as if the crop is not in the ground or if the crop is not sensitive at that stage. In addition, as mentioned in the pear example above, sometimes a risk decreases in select locations like extreme cold becoming less likely for crops that are at risk of frost (presenting an opportunity).

If the risk is increasing or already too high, options include moving to a new location where conditions are suitable. Our climate analog feature is able to find new suitable locations that have come online or will become optimal in the next five years – essentially a “copy paste” function for climates across time and space. It can get down to the microclimate level and look at climate, soil, water availability, and consider other variables like pest and disease as well, either honing in on a specific region of interest or searching the whole world. When we worked with a major seed company on a blind trial, we were able to identify three new unsuitable regions for a crop within minutes – ones that the company had identified itself after three years of visiting and testing all around the country.

Long-term forecasts (decadal by region) can show future climate conditions, so seed processors can breed for traits that are resilient to these risks. We also size the analysis according to their target markets i.e. what is the opportunity if we do breed or the risk if we don’t in terms of profit. One seed company customer, Advanta Seeds (UPL subsidiary) in Australia, was able to realize a 5-10% increase in sales by using our precipitation forecast (we predicted a precipitation event in a key sorghum growing region weeks ahead of anyone else) to inform the decision to move inventory into this region and capture the market.

Another facet of the long-term tool is the “Year of Departure” component – this calculates when such key risks will start to significantly affect companies’ bottom lines. For example, it can answer questions along the lines of, how long do we have before we need to take serious action before it’s too late? Sometimes, the risk is already too high, or it will be in five years, or down the road in 20-30 years. Different for each location, it equips companies with more intelligence for strategic decision-making.”

What risk tolerances need to be prioritized when breeding climate-optimized hybrids and varieties for target markets, 10-20 years in the future? Where are the new locations (microclimate-level) that can optimally grow planned focus crops?

“A cross made today must suit a market in 2030 — what will the climate look like in 2030, what type of attributes does the product need, where can we test today which mirrors these future environments? These are the questions we are working to address with ClimateAi. When breeding climate-optimized hybrids and varieties for target markets, 10-20 years in the future, the specific traits depend on the region.

For locations like California’s Central Valley, a key production region for cereal grains, hay, cotton, tomatoes, vegetables, citrus, tree fruits, nuts, table grapes, and wine grapes that faces more frequent and severe droughts, drought- and pest-resistant crops are highly relevant.

In addition, crops that thrive in highly specific microclimates are at risk. Take coffee: The most popular type of coffee, the arabica variety, thrives at temperatures between 64° and 70°F, on shaded mountainside farms ranging from 1,800 to 6,300 feet above sea level – basically a sliver of global arable land. A great deal is grown in Latin America’s so-called coffee belt.

But according to a study in the journal Climatic Change, nearly half of current global high-quality coffee-growing land could be unproductive by 2050. Heat-sensitive coffee production would suffer in a hotter world where heat reaches higher latitudes and altitudes, with Brazil and Vietnam most at-risk. Meanwhile, some areas that are not currently ideal for growing coffee may become hospital to the crop. That could happen in Nicaragua, where analysts say the optimal altitude for coffee growing will rise from 1,200 meters above sea level to 1,600 meters by mid-century.

Another example is a shift in wine regions, because fine wine grapes are one of the most sensitive crops to variations in temperature and precipitation. Three conditions must be met to grow them: warm temperatures, low risk of frost damage, and no extreme heat. Climate change affects all of these factors in traditional growing regions, from warming France to dry and wildfire-struck California.

Some wine grape farmers in Sicily have switched to avocados and other tropical fruits to keep up with the climate shifts already taking place. Winemakers and retailers are also seizing new opportunities in farmland that had previously been unsuitable for grape growing — such as in Belgium, England, China, Japan, and Poland.”