If you're new to precision farming and can't figure out why yields vary so much within fields, perhaps a computer can help.
Using mathematical correlations, software can determine the relationship between yield and various measurable factors that affect yield, such as pH, P and K, or soil type.
"For example, we might want to start by correlating yield to pH because the correct pH level is so vital for optimizing yields," says Stephanie Dickson.
She's managing partner of Geo-Ag Tech, a Springfield, IL, firm that does computer analysis of yield maps.
"Let's say a field is soil sampled on 2.5-acre grids. We can determine the relationship between pH level and yield in each grid. We can, for instance, see how a pH of 5.7 affects yield compared to a pH of 6.4," Dickson claims.
A logical place to start using variable-rate application of a production input would be a liming program that brings an entire field to the most productive pH level.
"After correcting pH, we can act on other factors, depending on how closely they relate to yield," she says. "That probably will be field-specific."
Although computer analysis of yield maps shows promise, it's still in its early stages of development, she says. Also, the more sophisticated software can run up to $10,000 - quite expensive for an individual farmer. Commercial firms like Geo-Ag Tech can spread that cost over many farmer-clients.
The first level in yield map analysis doesn't involve sophisticated computer analysis, says Dickson. It calls for the farmer to gather basic production information on each field, such as fertility, pH, soil type, drainage, weeds and weather data.
Perhaps with the help of a field scout, he can use that information - plus his historical knowledge of the field - to study yield maps and determine the reasons for yield variability.
To do a computer analysis, which is the next level, Geo-Ag Tech takes that same information and feeds it into a software program. The software must be able to accept all the data and convert it to a form usable by an analysis program. The program manipulates the data into a form needed for correlations between yield and various factors affecting yield.
For example, the correlation between yield and plant population may be very high, while the correlation between yield and hybrid or variety may be lower. The farmer would then know how much priority to put on each.
"It likely will take several years of data on a field to see long-term patterns and get really meaningful analysis on most factors," says Dickson.