Advances in remote sensing promise spectacular views of crops while giving agribusiness considerable new marketing opportunities. Making the data meaningful, however, remains industry's big challenge.

As one commercial satellite after another rockets into orbit, new applications prepare to serve agriculture in ways unimagined a decade ago. Beyond closer aerial views, new technologies promise crop mapping with improved color schemes, varying swaths, more frequent flybys and even three-dimensional fly-throughs. Some improvements could magnify vegetative changes that otherwise go unnoticed, aiding earlier detection of crop stresses caused by pest or disease infestations.

Current remote sensing maps, for example, provide 256 shades of gray. A new satellite set for launch in 1999, QuickBird I from EarthWatch, records 2,011 shades of gray. Although the human eye can't see such differences, computers can read the unique numeric values assigned to each shade and foretell plant problems before they erupt in fields.

"Remote sensing technologies provide amazing and unfamiliar views of places both new and familiar. But novelty alone is not enough to justify the higher costs of obtaining such views - we must put them to work," says Dr. Bill Bland, Extension agricultural climatologist, University of Wisconsin-Madison. Bland believes the greatest potential for remote sensing data is as inputs to computer simulations of agricultural processes.

Data for computer models. "Remote sensing images are just snapshots of one place in time, and the crop has a long history," Bland explains. "Computer models can mimic how crop processes are going. They make more of an attempt to integrate situations the way the crop actually does."

Bland and his colleagues at the university currently are testing three mathematical computer models using remote sensing. One model estimates water used by irrigated potato crops throughout the season. It can pay for itself by scheduling irrigation to maximize yields without wasting water or the energy needed to pump it. Another potato program measures leaf wetness, using computer models that combine weather observations made at airports, radar readings, satellite images and irrigation records. Potato growers can use this measurement in models of disease outbreaks to time fungicide sprays for optimal pest control. Another model provides early frost warnings to Wisconsin cranberry growers and could replace a National Weather Service program that was cut several years ago.

Because remote sensing data won't stand on its own, Bland cautions farmers to be wary of "smooth guys trying to sell remote sensing data the way some people sell cars. They may forget to mention that the engine costs extra."

Commercial satellites. The market for selling remote sensing data and value-added services could well reach several billion dollars within the next 10 years as commercial interests vie for their share of agriculture's revenue. Several major communications companies have plans to launch high-resolution satellite systems, made possible by the relaxation of military intelligence constraints.

Since 1990 42 satellites have been in operation around the planet. On December 24 of last year, EarthWatch, of Longmont, CO, launched EarlyBird I from the Svobodny Cosmodrome in eastern Russia. Preflight publicity hailed EarlyBird I as the highest resolution satellite commercially available. They expected 15-meter color resolution and 3-meter black-and-white mapping. At 3-meter resolutions, buildings, roads and bridges are visible.

Four days into flight, however, EarlyBird I lost communication, and a month later the company dismissed 30% of its workforce. At press time, a company representative expressed little hope of recovering the satellite. Instead, EarthWatch is focusing on its next-generation satellite, QuickBird I, headed for space in 1999. This satellite offers 0.82-meter black-and-white resolution and 3.28 multispectral resolutions. It also adds blue to its color spectrum, along with red, green and near-infrared, as well as 11-bit quantization or pixel depth. The blue channel allows true color imagery from space. The added pixel depth more clearly defines objects.

Other pending commercial launches include Ikonos by Space Imaging/Eosat, set for later this year; OrbView-3 by Orbimage, expected in the fall of 1999; and a Resource21 satellite with a projected launch of 2001. Other planned advances in remote sensing include shorter revisit periods, advanced radar sensor systems and other features allowing earlier and more thorough detection of crop problems.

Key to product performance, according to Kevin Little, Resource21 director of agricultural sales, are highly calibrated sensors, techniques that compensate for atmospheric conditions, and more frequent, broad-based, overhead coverage. Based in Englewood, CO, Resource21 is already flying planes ahead of its expected satellite launch. The company has run remote sensing tests via airplane in more than 1,000 commercial fields growing corn, soybeans, cotton, wheat and potatoes.

A quantum leap. Using the wealth of remote sensing data for agriculture, however, will require significant advances in computer processing power and basic changes in data interpretation.

"The jump to 1-meter resolution for some applications is a quantum jump in the magnitude of data that they generate," says Dr. Thomas Lillisand, director of the Environmental Remote Sensing Center, University of Wisconsin-Madison. Founded in 1970, this center supports one of the oldest and largest remote sensing research facilities in the country.

"At 1-meter resolution, a 40-acre field encompasses approximately 162,000 picture elements, or pixels," Lillisand says. "This data volume necessitates tens of gigabytes of computer storage capacity and tremendous processing power to acquire the data and much, much more once you start to manipulate the data."

Even with computer advances, agricultural users will be forced into making tradeoffs in spacial resolution, visual range, swath width and data delivery, Lillisand says. He notes that other problems with satellite imagery could arise from the satellite's time schedule and interference from clouds of the view from space.

"Determining the optimal mix of space, airborne and ground-based data in the context of a variety of agricultural applications in near-real time will be a great challenge," Lillisand says. "Farmers will be forced to face numerous tradeoffs...especially within the cost structure and contractual agreement for the data."

Intuitive reasoning needed. Even bigger changes are necessary in the way agriculture interprets these data, says USDA soil scientist James Schepers, Lincoln, NE. He suggests that optimizing remote sensing data requires a paradigm shift toward more intuitive reasoning.

Current variable-rate technology (VRT) is based on numeric information, Schepers says, but the information is no better than the understanding that goes into its collection and interpretation. In a project testing remote sensing as a nutrient management tool, Schepers and research partner Daniel Hagopian concluded that positive results are possible. They also noted, however, their results were highly dependent upon the data analysis techniques.

"We're used to locking ourselves into numbers. If you torture data long enough, it will confess to anything," Schepers said. "With remote sensing, there are things you'll see in the field that only you may understand. Measuring these relationships requires more intuitive and instinctive feelings than traditional quantitative analysis."

Updated information on satellite launchings and remote sensing applications are available at various Internet sites: Visit www.ersc.wisc.edu/ for the University of Wisconsin Environmental Remote Sensing Center; www.digitalglobe.com for EarthWatch; www.orbimage.com for Orbimage/Orbital Sciences Corp.; www.spaceimage.com for Space Imaging/Eosat; www.asprs.org/asprs for the American Society for Photogrammetry and Remote Sensing; www.soils.wisc.edu/wimnext for Wisconsin/Minnesota Cooperative Extension Agricultural Weather and Timely Satellite Data for Agricultural Management program.