Outsource your agronomist?

Nov 8, 2013
by DTN

—Climate Corp. adds consulting services

Farmers could outsource their agronomists in the near future, if weather insurer Climate Corporation succeeds in its latest product launch. Unlike traditional on-site crop advisers and field scouts, these agronomic experts include a team of PhDs with degrees in statistics, physics and applied math.

Climate Corporation is best known for applying Silicon Valley-style algorithms to the stodgy U.S.

crop insurance industry, then selling policies insuring against perils like drought, excess rain or heat at pollination. That was a departure from traditional government-subsidized crop insurance based on actual yields and futures prices to determine insurance claims. Climate Corporation’s innovations in math and climate research attracted a $930 million purchase by Monsanto that was finalized Nov. 1.

On Monday, Climate Corporation announced the launch of a fee-based weather and agronomic advisory service available in 2014. The service, Climate Pro, will use the company’s hightech weather simulations to provide field-level advisories. Service will include guidance on relative maturity and planting dates that will optimize yield; where, when, and how much to sidedress nitrogen; where pest management practices will lessen yield loss; and what harvest date would optimize dry-down costs, grain moisture and yield.

Unlike land grant university or previous seed company crop recommendations based on performance under historical average weather, Climate Corporation’s advice will be based on the computer models that will churn out the odds of weather conditions occurring from planting through harvest for the current growing season.

Such site-specific advice for corn will cost $15 per acre, but the company estimates it can return $100 per acre per year for a typical corn customer. Soybean advice could increase profits $50 per acre, at a cost of $7.50 per acre for a subscription. A “lite” version of the online service, Climate Basic, will provide site specific weather forecasts and crop tracking tools for free, but will not include site-specific agronomic advice.

Climate Pro will be sold through the company’s network of crop insurance agents nationwide. The agents already handle Climate Corporation’s unique weather-event-based policies and are selling traditional multi-peril insurance under the company’s name for the first time this crop year.

Ten years ago, relying on complex weather models to predict crop yields down to the field level wouldn’t even have been possible, Climate Corporation CEO David Friedberg told DTN. But the dramatic drop in the price of sensor technology has reduced the cost of weather monitoring and forecasts to a fraction of their pre-2000 levels. That doesn’t mean the company can predict weather, but it can predict the probability of weather throughout the growing season with a higher degree of precision, he said.

“A decade ago it cost $1 billion to develop a satellite that could take multi-spectral imagery of the earth, but just a few weeks ago a Silicon Valley firm launched two satellites for less than $7 million,” Friedberg said. “That’s almost nothing, really. The pixels in a cell phone today are better than the sensors on the original satellites.”

The National Oceanographic and Atmospheric Administration (NOAA) didn’t always give away all its data for free, but it does now, he added. “So someone here with a laptop can access radar sites and see how much rain is falling realtime in every field in the U.S.”

Simultaneously, the cost to store and transmit volumes of data—including the nation’s soil maps, historical weather, yield records and federal crop insurance field maps—has plummeted. “It cost $500 or more to store a gigabyte of data 10 years ago, but only 2 cents to 3 cents now,” he said, so companies need to add value to public data if they charge for service.

That ability to store and process immense amounts of crop growth and weather data will allow the company to forecast yields at a field level. Today it personalizes weather forecasts for rainfall, temperature and soil moisture on 0.6-mile grids for Climate Pro customers and 2.5-mile grids for its insurance customers. Each forecast uses tens of thousands of simulations.

Predicting yields based on weather events is the Holy Grail of crop forecasting, and an effort that has frustrated economists and USDA statisticians for decades. Some land grant universities contend they haven’t been able to find much correlation between the two, noting so many variables affect yield beyond pure weather. Others believe the science needs further refinement, but technology will eventually make it possible.

Friedberg admits the yield prediction model isn’t perfect. The company had to revise models to account for tiled fields or long distances from weather stations. It’s spent $100 million to improve basis risk, the difference between measured weather events at sites like weather stations and the actual events that take place on a farmer’s land. Access to Monsanto’s seed trial data will be another benefit.

“Over time we’re getting better at it,” he said. “It’s a lot like the algorithms that Google uses in its search engines. When you’re looking for something online, you want the first thing that pops up to be what you’re looking for.” It took a few years for the statistics experts to work out those kinks. — Marcia Taylor, DTN