 | 
| |
GreenSeeker Sensor and NDVI images for Sugar Beet Nitrogen Management
Gary Wagner, President
A.W.G. Farms
Crookston, Minnesota
www.awgfarms.com
January 2004
When our farm was first introduced to the GreenSeeker, we immediately saw a personal remote sensing device that could give us real time, high resolution, remote sensing data.
Our farm is located in Northwest Minnesota; our main cash crop is sugar beets. In 1997 local researchers discovered that a tremendous amount of nitrogen fertilizer is stored in the tops of sugar beets, which is shredded and returned back to the soil. Amounts of N could range from 50 to 400#'s, depending on the amount of bio-mass. It was also determined, that 70 to 80% of this potential Nitrogen is available in time for the next year's crop. (The image to the below/left shows an interpolated Landsat 7 NIR image, and 42 hand gathered sugar beet tops points analyzed for Nitrogen content.) Armed with this knowledge we variable rate applied nitrogen based on the reflectance values. (Shown on the image below/right ). Saving on average $12/ acres, in fertilizer cost, over 1000 acres of beets was very beneficial.
Also realizing excess nitrogen is the main contributor to poor quality, by correcting these high Nitrogen regions we are able to increase crop quality in subsequent sugar beet crops. (The 2nd set of numbers reflect the sugar content of the beet tops in 1997 (lower number) and 2000 (top number). In 1997 the sugar content varied about 4%, in 2000, after 4 years of VRT, the content varies only 1%)
American Crystal Sugar, our parent Cooperative, in 1999 started purchasing Landsat Imagery, and giving these images back to the beet farmers to start adopting VRT fertilizer applications. As of today, approximately 150,000 acres to 200,000 acres (of Crystal Sugar's 500,000 acres) are treated in this manner.
Here is where the GreenSeeker will play a very important role in the Red River Valley of Minnesota and North Dakota. Landsat imagery has very course resolution, but is very inexpensive. The timing of the image (once every 16 days), clouds, and other factors, sometimes forces an image to be acquired 4 to 6 weeks before harvest, which at best gives us approximate regions of high and low bio-mass. This last fall we modified a hand held GreenSeeker, and using a iPaq handheld computer, and FarmWorks software we mapped our beet tops at the time of harvest.
The results at best can be describes as "WOW". Higher resolution, less atmospheric interference, the list can go on and on, we are starting to generate more precise data on the bio-mass of the beet tops.
The image on the right compares the output of the GreenSeeker, the image we get from American Crystal (top image), and the fertilizer application map (bottom image) shows pattern not consistent. I need to be mindful that the GreenSeeker recorded an NDVI reading and the Landsat image is NIR, but even compared to a calculated Landsat NDVI map the relationship is not even close.
The GreenSeeker, in our estimation, will become more valuable as a remote sensing device (or at least as valuable), as an application device. It is our intention to have several sensors on our ground sprayer and every time we cover our crop (in beets 8 times) we will map crop differences throughout the growing season, generating difference maps, for scouting tool, to determine areas of better on poorer growth. We are also working with Dr. Dave Frazen (NDSU), on adding a canopy height sensor besides the reflectance reading from the GreenSeeker, to generate a more precise fertilizer application map.
|
|
|
|
