It is more than likely that if you are directly involved in the agricultural world at some point you’ve heard about precision agriculture. Just like the tablet computers today, or the Macarena in the 1990s, precision agriculture is that unavoidable thing that everyone is talking about and with good reason. With a promise of dramatically increasing productivity by integrating 21st century agriculture technology into the farm, one can’t help but to get excited about all the potential benefits. And since precision agriculture relies on advanced on-board computing systems you can expect tomorrow’s tractors to more closely resemble the car from Knight Rider than your dad’s John Deere.
The premise behind precision agriculture is simple; since terrains are not uniform in most of their dimensions such as soil composition, nutritional needs, crop yields, and pest/disease presence, then the traditional practice of treating them uniformly is hardly ideal. Precision agriculture intends to solve this problem by splitting each block into a grid of smaller plots of land and micromanaging them individually, or performing “site specific management” as we say in agriculture. Of course this is too big a task for people to do with since it requires collecting and analyzing millions of pieces of data. This is where technology comes into play with a variety of advanced technologies such as GPS systems, yield monitors, variable rate applicators, and geospatial statistical analysis software.
Soil composition and yield maps
As with any successful scientific endeavor, a good precision agriculture process requires accurate data. A good place to start is usually by generating soil composition and yield maps of the blocks you wish to analyze. These maps represent the large numbers of samples that will be the basic input for geospatial statistical analysis software that will produce recommendations for different processes such as fertilization, sowing density, and pest/disease control.
The problem with these types of maps is that they can be very difficult to build correctly. A cumulative set of errors coming from many different elements such as GPS system accuracy limitations, usage of multiple harvesting machines on the same fields (Multiple yield monitors), and time delays on data capture due to hardware capacity, can drive the margin of error exponentially even to the point of rendering it useless. This is why having a proper process for minimizing errors during data capture, and where the posterior filtering of suspiciously out of place data (Outliers) become crucial activities.
Variable rate applications
Currently there are two major schools of thought behind variable rate applications and potentially even precision agriculture in general; on one side we have those who wish to increase the fields to further heights than had been previously possible with traditional methods. Their rationale lies behind the fact that since we’re not using optimal agricultural methods for each specific site, then the maximum potential of fields has not been reached and large benefits can be achieved by trying to maximize production. On the other side of the fence we have experts who think that due to the fact that an extremely high level of production has already been reached on most crops, possible gains in production due to site specific management of fields are marginal and not worth pursuing. The focus of these experts instead is centered in cost-cutting by reducing waste and over application of agricultural supplies (Seeds, fertilizers, pesticides, etc.). What I’ve noticed is that actual precision agriculture savvy farmers in many cases have a position that sits half-way between both camps, and thus treat each field as a unique case that needs to be treated individually to determine the best approach.
Once you have decided which approach suits your current situation better, modern geospatial statistical analysis software tools will offer a myriad of mathematical methods for determining an optimal application pattern that takes into account the specific circumstances of each individual site. This application patterns will then be transferred to an on-board computer that can automatically control application rates of the tools attached to the tractors (Planters, sprayers, etc.) based on the GPS position of the vehicle. Here again ensuring the minimization of errors in data, such as ensuring an accurate GPS position, can have a big effect on the final results of the exercise.
Is it worth it?
I am sorry to say that the jury is still out on this question. So far you will get an even amount of both positive and negative stories about real-life applications of precision agriculture. This could be due to the fact that a clear definition of when and where this methodology makes sense has yet to be agreed. Expecting precision agriculture to be a magic solution for anyone with a plot of land is not realistic, and it’s even more likely that there will be plenty of cases where precision should be used and even more where it shouldn’t. Once we get to that point and past the hype we can gauge the real impact of precision agriculture in different situations, then we will find its real value. Thankfully there is an ever growing group of pioneers from all over the world that have taken the first step.