For example, farmers may decide they need to buy supplemental feed from outside the drought-affected area to support their livestock, or they may need to adjust contract or insurance decisions.

The U.S. Drought Monitor already uses a combination of indices, such as rainfall, to describe drought conditions each week. The monitor currently does not include plant stress, but the potential is being explored. "Plant stress is one representation of drought impacts, and the drought monitoring community agrees that you can't do this with just one tool – you need a lot of different tools," Anderson says. 

Plant stress information has the potential to improve the skill of existing forecasts that predict drought out to weeks or months. Also, because the plant stress information is derived from satellites, it can describe drought conditions in areas where rain gauge and radar networks are sparse -- and it can do so at the scale of individual fields. 

To produce the maps of plant stress, scientists start with the Moderate Resolution Imaging Spectroradiometer (MODIS) on NASA's Terra and Aqua satellites. Images are processed to distinguish between land surfaces covered by soil and surfaces covered with vegetation.

Narrowing their focus to vegetated areas, scientists set out to measure moisture availability. Plants cool themselves by sweating water extracted from the soil by their roots. When access to water is limited, plants lessen their consumption and reduce evapotranspiration from leaf surfaces. As a result, leaves heat up and produce an elevated leaf or canopy temperature, which can be detected by thermal sensors on NOAA's geostationary weather satellites. Hotter plants imply limited water in the soil. 

"This is not a drought forecast. It's a map of what's going on right now," Anderson says. "Is there more or less water than usual?" 

What is "usual" or normal, however, can depend on the season or even the year. Scientists currently define normal by calculating and mapping plant stress averaged over periods of 1-3 months, from the start of MODIS data collection in 2000 to present. The mean of these historic maps is considered normal. Compare a current map with the longer-term "normal" map, and scientists get a picture of the magnitude by which current conditions deviate from normal. 

"What was normal back in 1920 is not what's normal now, so the more years we have under the belt the better we can define normal," Anderson says. "But this year is so far out of line with respect to previous years, it is unusual regardless of the period of record used as the baseline."