The levels of control
Development of modern agriculture is occurring at a time of massive adoption of information technologies. To increase the efficiency of resource utilization and environmental savings, the merging of information technologies and intensive crop production systems has resulted in the concept of precision agriculture. Ranging from automatic navigation systems to remote sensing, some elements of precision agriculture already have been well-integrated into modern crop production. However, many producers have expectations that cannot be satisfied through the technical capabilities of available equipment and services. The main problem is the mixing together of two different levels of control in crop production management.
How to define the economically optimal rate of seeding or the application of agrochemicals? Frequently it is expected that sensor technologies in combination with variable rate equipment can answer this question for every location across every field. In fact, it is not true. Equipment and software can handle massive data obtained at the farm or field level; however, this does not ensure a profitable response to applied chemicals. In general, once the economically optimal range of application rates is established, it is possible to expect the higher probability of increased profits. Therefore, before implementing variable rate technology, it is important to make sure that existing growing conditions require site-specific management of a given field or farm. There should be at least two specific locations where optimal application rates differ from each other. After that, redistribution of agricultural inputs according to prescription maps, or real-time sensors, becomes reasonable. This constitutes the second level of control.
The first level of control remains undefined. It should start by formulating the optimization objective function. The most common assumption is that the objective function means profitability. However, it is not clear whether short-term of long-term profitability is important and what is the acceptable risk tolerance. Production processes are normally quantified by a productive function, which describes change in yield in response to varying rates of seeding or fertilization. Key parameters of such a function depend on soil conditions, weather and past management. In addition, profitability depends on crop prices and the cost of agricultural inputs. There are many traditional and modern methods for determining the parameters of the yield response function and these methods extend beyond information technologies. It is important to relate to past experience, data from seed and fertilizer trials, a fundamental understanding of soil processes, plant physiology, and meteorology. Without a doubt, there is a need to integrate multiple databases containing relevant knowledge, conduct analysis of soil and plant tissue, etc. This is the first level of control.
An advantage of agricultural services promoted by AgriLab is the approach to optimize plant production using two levels of control. Comprehensive analysis of soil test results, meteorological data, crop scouting reports and monitoring the prices of agricultural commodities provide an opportunity to compare alternative management decisions in terms of discovering the least risky plan of action. Diverse scientific and commercial organizations strive to develop and improve existing prediction models and analytical methods to account for uncertainties that agricultural producers are required to deal with. Certainly, the modern quest to engage in deep learning that is based on the “big data” phenomenon will eventually help producers and their advisors make informed decisions to improve agricultural processes while relying on the two levels of control.
Viacheslav Adamchuk, PhD
Associate Professor in Bioresource Engineering