This summer, Jackson State Community College will help several Mid-South cotton producers make more money — and spend less.
The farmers will be participating in a variable-rate application study on 4,000 acres that could reduce their inputs by at least $63 per acre while bumping yield by 60 pounds.
Late last summer, the producers acquired aerial images of several of their cotton fields. The images, shot with an airplane-mounted camera in green, red and near infrared, were loaded into a JSCC computer and geo-referenced. Then a software function called normalized differential vegetative index (NDVI) classed each field into three productivity zones — low, medium and high. These zones are highly correlated to eventual yield.
JSCC then developed a “computerized mission plan” for VR applications in the fields. This coming season, when it's time to apply inputs, the growers will simply load a chip in the controller on a VR sprayer to variably apply lime, preplant fertilizer, seedling rate, in-furrow fungicide, in-furrow insecticide, Pix, top-dress nitrogen and in-season insecticides. Inputs will be varied according to which zone the equipment is passing through.
The VR application itself is easy and the cost of modifying equipment for VR relatively inexpensive — about $18,000 for a 1,000-acre operation.
In 2002, JSCC conducted a similar VR experiment on 1,000 acres in west Tennessee. It indicated that growers could pay for the technology in one year, even adding the cost of gathering and analyzing the data.
According to Tim Sharp, JSCC professor, three west Tennessee farms were used for the 2002 study. In 2001, an airplane shot imagery of the fields late in the season and JSCC computers generated the NDVI images of the field.
Fields across the three farms were paired — one of each pair was managed using conventional application techniques and the other with VR application.
Data was collected in the zones at various times. Information included stand count, first position retention, plant height, total nodes, identification of first fruiting branch, elongation at the fourth internode, yield maps, total final plant maps and soil grid sample data.
For the study, the variable-rate fertilizer application was grid-sample driven and applied on all fields, even those conventionally managed — to remove fertility as a variable in the test results.
Seeds were planted at a rate of two seeds per foot in low zones, three in medium zones and four in high zones. Most of the VR technology used on the sprayer rig was also used on a planter.
In-furrow fungicides and Temik were applied in the low zones at 50 percent of what was applied in the high zones, and at 75 percent in the medium zones.
One application of Bidrin was made over the study at 60 percent, 80 percent and 100 percent in low, medium, and high zones. Pix was applied at 0, 0, and 100 percent, respectively.
Plant mapping indicated the most significant improvement in the retention of first position bolls, nodes and total bolls in the low zones with variable-rate application versus conventional, according to Sharp.
Overall, yield in variable-rate fields was 63 pounds higher than in conventional fields, “plus we spent about $60 less in reduced input costs from variable-rate application. So we're looking at $100 less per acre in the variable-rate fields.”
Sharp noted that there was no difference in yield between conventionally managed and variable-rate managed fields on first pick. However, the second pick was where growers picked up the extra yield — 33 pounds more in the high zones, 73 pounds more in the low zones and 83 pounds more in the medium zones.
Sharp also observed some other characteristics of the productivity zones. For example, medium zones can act like high zones if they're managed correctly and like low zones if they're not. In many cases where a grower has unexpectedly high yields, it's often the medium zone that has done the work.
In addition, in irrigated cotton, where farmers often push the low and medium zones for more cotton, they can inadvertently “blow up” their high zones — create boll rot or allow excessive growth. In that situation, it may actually be better to reduce inputs such as nitrogen in the high zone, and/or increase Pix. VR can be used to pinpoint those areas.
Before growers undertake a VR program, they need a base image of a field indicating their low, medium and high productivity areas. That image will be difficult to obtain after VR has begun.
Sharp explained that the effect of variably applying inputs over the field tends to remove variability as seen by cameras, sensors or yield monitors. Therefore, in most cases, subsequent yield or image maps will no longer reveal the low, medium and high zones. “VR simply flattens out variability,” Sharp noted.
Of course, there are definitely more than three levels of variability in a given field and certainly no distinct lines of demarcation between zones, noted Greg Evans, JSCC laboratory assistant. “We're not always doing exactly what is called for, but what we're doing is certainly better than what we were doing before.”
“Three management zones are okay for seeding rate, but for in-season PGR applications there is a level of variability that we are not capturing,” added Sharp. “We need sensors for the next level of technology for in-season application.”
Such sensors are on loan from Oklahoma State University and are a part of the study, too. The sensors use a light source to gather plant reflectance data, which provides information for on-the-go applications of PGRs.
In addition, “zones can swap on you during the season,” Sharp said. “About mid-season, those sandy loams are going to start senescing and the clays are going to start growing. When that occurs you don't know, but the relative attractiveness to insects within that field is going to change. We can't grab that with aircraft images. We can with sensors.”
VR is a practical, grower-ready technology “that any farmer can do with some consulting help,” said Sharp.
Finding a qualified consultant and a computer processing center for VR will be more difficult, according to Sharp, noting that the manipulation of huge imagery files requires powerful computers and sophisticated software “which is not something the average farmer is going to have on his farm.”
That side of the VR equation will require significant investment by private industry or public institutions like JSCC.
Still, Sharp figures the farmer would pay around $20 an acre for VR computer services and consulting. With potential savings and increased yield amounting to $100 an acre, it would be a good buy for producers.