This article on the GreenReport, maintained by the Bio Survey's Kansas Applied Remote Sensing program, was put online in October 2019 (dated May 2019).
In the spring of 1996, the Kansas Applied Remote Sensing (KARS) program on the campus of Kansas University started chronicling the greening and browning of America.
Where was new plant life emerging from the fading winter? Where were the corn, the soybeans, and the wheat sprouting in the fields? As March turned to April, then May, how robust and vibrant did those greening fields become? And when summer arrived, how healthy did the plants remain in places where rain was sporadic, and where the sun baked the landscape?
All those questions KARS tried to answer—and does so yet today—in a series of maps called the GreenReport® that it puts out with assistance from the Earth Resources Observation and Science (EROS) Center.
To measure vegetation conditions across the United States—not only mapping the location and amount of plant biomass but comparing that information as well to the relative norm, to the previous week and year, or to a 30-year average—KARS relies on Normalized Difference Vegetation Index (NDVI) images compiled at EROS from Advanced Very High Resolution Radiometer (AVHRR) satellite imagery.
KARS puts its greenness and change maps on a website that provides weekly updates on vegetation health and plant stress and are available to anyone. It also uses the AVHRR images in concert with historic satellite and weather data for bi-weekly forecasts of district, state, and national-level crop yields for eight major crops in the U.S.
Those forecasts are distributed by KARS’ business partner, Berwyn, PA-based Planalytics, to help clients better understand how weather and crop yield forecasts can affect their bottom lines.
Planalytics’ expertise is business weather intelligence, says Jude Kastens, a research associate professor at KARS. The company works with hundreds of retail clients to assess and address how weather may impact their businesses so they can predict such things as when to order more umbrellas, or to have seed available to plant, or to stock enough tractors to sell for the coming growing season. Planalytics has a Life Sciences sector that works with people in agriculture as well, Kastens said, “and that’s where we come in with supplying information they can use.”
That KARS can do what it does was made possible when the National Oceanic and Atmospheric Administration (NOAA) decided in 1978 to launch the AVHRR sensor aboard the first in a series of polar-orbiting meteorological satellites. AVHRR’s mandate was to monitor clouds and to measure the thermal emission of the Earth. But it turns out the data are good for other things, too. With its red and near-infrared bands, AVHRR can monitor vegetation health and change on a daily basis.
Data Management and Information Distribution (DMID) at EROS has put together AVHRR composites since 1988. Using a strategy called Maximum Value NDVI, DMID staff can pull out the greenest pixels during a seven- or 14-day time period, said Research Geographer Jess Brown at the Center. Understanding that clouds decrease signals from the sensor, the greenest pixels are assumed to be the most cloud free. EROS then combines those greenest pixels into weekly and bi-weekly composites that clearly show vegetation health and stress.
“The value of long-term NDVI time series is just phenomenal,” Kastens said. “We can’t say enough about how interesting and useful those data have been to us that EROS distributes.”
Here’s where a slight hitch comes into the conversation now for KARS. It has relied almost entirely on AVHRR since it started its GreenReport® and crop yield forecasting. But the days of AVHRR data are growing short as the aging satellite platform carrying the current AVHRR nears its end. The Moderate Resolution Imaging Spectrometer (MODIS) sensors aboard NASA’s Terra and Aqua satellites have continued AVHRR’s land-imaging legacy, but those are both aging as well. And the next generation sensor—the Visible Infrared Imaging Radiometer Suite, or VIIRS, aboard the Suomi National Polar-orbiting Partnership and NOAA-20 weather satellites—is close but not exactly a carbon copy of AVHRR.
“There are differences,” Brown said. “If your history is based on one sensor, and then your new data for monitoring is based on another, that introduces some error, and scientists need to quantify that. We have started research to look at the relationship between MODIS and VIIRS, but bringing three sensors into that study would be quite difficult.”
She has suggested that the KARS folks switch to MODIS for now. A small team at EROS is now testing the ability to make regular greenness composite products from VIIRS, Brown said, adding, “because there’s no telling how much longer we will have these data, we are trying to prepare for the time when both AVHRR and MODIS are no longer available.” MODIS NDVI is also a current input to VegDRI and QuickDRI products developed at EROS.
Kastens said he understands all that. His office has used MODIS data when there were hiccups with AVHRR, such as when it was unavailable during the government shutdown in October 2013, or during transition periods when the primary AVHRR sensor fails and EROS has to switch to a different one. He did indicate that the KARS staff “has done some work cross-walking the two datasets ... and they map up sufficiently for what we do.”
Still, there are fundamental differences between MODIS and AVHRR, Kastens added.
“We do less gymnastics if we just maintain our AVHRR dataset because what we get now is still very consistent with what we’ve received in the past,” he said. “If we go fully on to MODIS, we just have to make sure that things are behaving well. It’s a little more of an effort because it’s a little different data stream. So really, the main reason we haven’t switched yet is because of the record. The historical record can’t be beat for AVHRR.”
They want to continue the GreenReport® and crop yield forecasting as long as they can, Kastens said. Since the two MODIS sensors have already exceeded their lifetimes, if AVHRR goes away and the data products they need become available on VIIRS, “I imagine we would make the jump to that sensor right off the bat,” Kastens said.
Of course, to do so likely would require some assistance from an old friend, he added.
“We’re really counting on EROS,” Kastens said, “for whatever direction we take in the future.”