Advancing Technology in Weather Data & Forecasting

The drive to gather valid weather data and best react to the quirks and vagaries of Mother Nature has been a preoccupation for as long as farming has been a profession. Every year, farmers and their crop production partners risk everything against climatic forces that can swing from idyllic to completely unmanageable from week to week, or even from hour to hour.

With so much on the line, it’s no wonder that the business of weather monitoring and data collection is among the most competitive and complex segments of the agriculture industry. Providers of weather data and weather monitoring tools and systems agree that we’ve come a long way, and farmers have more and better information at their disposal.

We’ve had weather technology innovation and investment for years courtesy of the federal government. But private investment in systems and products for ag weather have skyrocketed in recent years, led by Bayer/Monsanto’s investment in Climate Corp., and a big push from organizations outside of traditional agriculture like IBM to bring its technologies and capabilities to the industry.

The result has been a lot of excitement, along with a bit of confusion about what weather systems actually bring to the table as far as actionable information. We spoke with some of the leading organizations and experts to get some sense as to the state of weather data and systems, and where technology and product offerings are headed in the future.

Forecasting Success

Everybody has taken a turn beating up the meteorologist from time to time, but experts agree (and statistics concur) that forecasting overall has improved significantly over the past four decades. The combination of fully maintained, on-the-ground stations along with satellite sensing capabilities, coupled with the proliferation of advanced super-computing power possessed by organizations beyond the federal government is creating forecasts of unprecedented accuracy.

“In particular with five-day forecasting, we’re seeing accuracy in the area of 90%, where 30 years ago that accuracy was closer to 60% to 70%,” says Jim Pollack, a consultant with the Denver-based Prassack Advisers. Pollack spent several years with the weather data company aWhere prior to his current role. Longer range forecasts are still a relative coin flip at 50%, in particular when trying to pinpoint precipitation, but are still very useful for agriculture’s purposes.

Recent years have seen a spate of consolidation in the marketplace, with companies like DTN and IBM scooping up forecasting assets in a race to have a forecasting quality advantage, says Tim Marquis, a meteorologist who’s now working in the equipment logistics efforts for technology company Uptake.

“We’re seeing a lot of these organizations for the first time be able to run global weather models that they are aiming to have meet or exceed what either of the two primary models — the United States and Europe — are capable of delivering, and which offer at no charge,” Marquis explains.

Increasing resolution is also the reason for the push by private weather organizations for additional ground stations across farm country. Theoretically, more surface data ingested would lead to a higher surface analysis accuracy that could be monetized in the agriculture market.

“They would not offer it for free, unlike government sources,” says Marquis. “They would argue that they have the best weather model and are improving on it to meet the specific business targets of agriculture.”

The question, “How much resolution do we need?” is a point of debate, and competition, in the market place. Nate Taylor, Sales Director, Agriculture for Iteris, Inc., sees the primary value from weather forecasting is incorporating it with other data toward the application of a specific task. “The discussion should be around the question, ‘do you have the right resolution to solve the problem you are trying to solve?’”

Taylor points out that the many top weather forecasting systems are audited through a service called ForecastWatch, which ensures a base level of forecast validity. From this base data, Iteris builds models for ag tasks such as spray application, fertilizer, and seeding to ensure optimum timing of field operations. Beyond tools for the production cycle, Iteris develops custom analytics for a variety of companies in the agriculture value chain.

Taylor is not a big believer in the benefit of a wide proliferation of weather stations for additional data collection, asserting that models can be constructed from existing satellite and ground systems that will provide enough resolution to extrapolate accurate data for rural users.

“When a company is looking to use environmental data, we ask what problems they want to solve that weather data-infused modeling can deliver,” says Taylor. “It’s not reporting on a computer dashboard that it’s cold or windy outside, it’s ‘what is the data telling me I can do today?’”

On the other end of the spectrum, DTN is leaning on the proliferation of local weather stations in an effort to improve its forecast resolution, in particular for wind speed and precipitation, and where the probability of weather variance is especially high across distance.

“Having local observation available allows a company like DTN to create a forecast for a given location, even in remote locations,” says Jim Foerster, Certified Consulting Meteorologist with the company. “Farmers are making decisions along a wide spectrum of work, and we see providing that information in real time as critical.”

Weather stations in and of themselves are not a perfect answer, as conditions can change significantly in relatively short distance from the station location. But as more stations are added, and then combined with high resolution radar and satellite information, the forecast accuracy improves dramatically, says Foerster.

“For example, if it’s always windier at your location than the airport that’s closest to you, the forecast will be able to learn that and the data from your farm will reflect that,” says Foerster. This is especially important when making difficult and potentially expensive decisions around irrigation and chemical applications.

What model delivers answers at an appropriate, dependable resolution that’s cost-effective for service providers and farmers will depend on the crops grown, the variability of field conditions, and the goals of the end-user. The best advice: Ask questions and do your homework.