Skip to main content

Fertilizer recommendation technology: What is success?

Fertilizer recommendation technology

By: Brad Carlson, Extension educator

One of the hottest areas in precision agriculture the past few years is technology designed to make fertilizer recommendations. The technology can be used on a whole field basis, but typically is designed to make variable rate recommendations within a field. There are several very different technologies currently being used, including crop color sensors, drone imagery, in-season soil and tissue testing, and crop models. In general, most of these technologies focus on nitrogen, but some will make recommendations for other nutrients too. Despite the differences, there are some overarching concepts that need to be understood if you are going to use them.

To begin with, let’s look at traditional fertility research. University of Minnesota researchers conduct field trials across the state every year, with a variety of rates applied – from too little to too much. Crop response is measured in yield. From the data, a “response curve” is developed, where the maximum profit (income from yield vs. expense for fertilizer and application) is calculated. By combining many trials from many years, general recommendations are developed. It is important to understand that the recommendations are not an average rate where yield is maximized, but fit to a numeric model where most of the situations fall within the recommendation. There are situations where more than the recommended rate of fertilizer is necessary, but the recommended rate is appropriate for 80-90% of situations. It is reasonable to expect the same response pattern to occur within a field as it does when comparing one field to another.

There have been many field trials with variable rate N technology conducted by the University of Minnesota and others in the past several years. Some of the technologies work better than others, but that is a different topic for a different day. One thing that most of the research has confirmed is that most of the technologies (when properly calibrated and interpreted) will recommend lower rates of N in large parts of the field than would have otherwise been applied as a flat rate. It is therefore worth stressing an important point: These technologies are designed to make an accurate fertilizer recommendation, not to increase yield.

Many producers have inaccurately declared that the technologies do not work because they have not seen increased yields. When you think about it, though, when the technology recommends applying less N, you shouldn’t see a yield increase. If the technology works, it made an accurate rate recommendation and you will see the same yield with less fertilizer applied. The main take-home point is that, under these circumstances, you will see increased profits only via decreased fertilizer expense.

Many more trails need to be conducted over a wider variety of conditions before we are ready to declare which of the crop models and other technologies work, and understanding how to evaluate success is essential. As a final note, it is worth mentioning that simply taking a yield map and calculating crop removal of a nutrient and then variable rate applying those rates is not supported by research and is unlikely to result in increased yields or profits in most circumstances.


For the latest nutrient management information, subscribe to Minnesota Crop News, like UMN Extension Nutrient Management on Facebook, follow us on Twitter, and visit our website.

Support for this project was provided in part by the Agricultural Fertilizer Research & Education Council (AFREC).
Print Friendly and PDF