Cotton Growth Management with Precision Agriculture

Satellite imagery cotton field-in-Brazil

From left: Satellite imagery one week before application, as applied rate, and satellite imagery one month after application of plant growth regulator in cotton fields in Mato Grosso, Midwest Brazil.


A cotton crop presents some important differences to other commodities that are widely produced in Brazil, such as soybean and corn. The first one concerns the importance of quality, since the main reason for growing cotton is the production of fibers that are used in the textile industry, which has high standards of fiber uniformity, staining, strength, length, and maturity.

The second difference is of physiological order, and is related to the origin of the cotton plant, which is perennial. When conditions are adequate, the plant prioritizes vegetative growth to the detriment of reproductive development, which is undesirable for obtaining high yields, requiring the control of plant growth through the use of plant growth regulators (PGRs) that change the hormonal balance and promote the reduction of growth and retention of vegetative structures.

Several factors can cause spatial variability in the growth and development of cotton crops. For example, differences in soil texture and fertility and occurrence of pests, diseases, and nematodes. Thus, the rates of growth regulator required to obtain a homogeneous crop and express the greatest yield potential may vary within a field.

An important feature of this variability is that, in general, it presents low temporal stability, except in areas where it is directly related to soil texture. This makes it necessary to identify the differences in season.


Some tools have been developed to measure the height of plants with high spatial resolution, initially with the use of ultrasonic sensors and more recently with the use of LIDAR technology. Although the assessment of the need for PGRs is usually performed by measuring plant height and number of nodes, this is directly related to plant vigor and has good correlations with vegetation indices obtained by active canopy sensors or satellite images. Satellite images have the advantage of low cost and high spatial resolution (3 – 5 m), but are dependent on the climatic conditions, being limited mainly by the existence of clouds. Active canopy sensors, on the other hand, have the advantage of being less dependent on climatic conditions, and can even operate at night. A comparison of these tools can be seen in TREVISAN et al. (2015).

The use of vegetation indexes presents advantages mainly in more advanced phases of crop development, since the effect of the plant growth regulator is dependent on accumulated biomass and vegetative vigor, which makes the vegetation index a better estimator of the required rate. The same is true in the case of the use of defoliants and boll openers, since the percentage of defoliation and open bolls is better correlated with the vegetation index than with the height of the plants. Once a spatial variability map of crop development has been obtained, a PGR prescription can be created using existing models or from field checks for dose determination at contrasting points and the use of regression.

The execution of the prescription depends on the availability of equipment for variable rate liquid applications, which usually presents some limitations. The most widespread form is the application with variable flow, in which a controller changes the pressure of the system, increasing or decreasing the flow and consequently the applied rate. The range of variation allowed in this system is limited because the variation in pressure may adversely affect the quality of the application. Another technology, which solves this problem, is the use of pulse width modulation (PWM) system to control flow at the nozzles. This allows wide variation of doses without changing the pressure of the system.

However, this technology is still little widespread and has a common limitation with the flow variation, which is related to the use of tank mixtures. Thus, when it is desired to apply a PGR and an insecticide in the same operation, it is not possible to only vary the dose of the PGR and keep the insecticide constant. An alternative to this is the use of direct injection systems, which inject concentrated chemicals and water into an in-line mixer on the sprayer. In general, most of these available systems have limited spatial resolution because the injection occurs on the main line and the applied rate is the same across the bar. The ideal system for this type of application could be the use of two independent spray systems on the same equipment or direct nozzle injection systems, but none of these technologies are commonly found.

It is important to match the resolution of variability mapping to that of the application. Thus, for an application where the control occurs only in the whole boom width, free satellite images such as Landsat 8 and Sentinel-2 may be sufficient, whereas for equipment with section control the use of images with better spatial resolution such as RapidEye and PlanetLabs can bring better results. With the use of PWM valves and individual control in each nozzle, the use of active sensors such as LIDAR and WeedIT can provide the best result, which also makes it possible to do a real-time application without the need of a previous map.

We have been testing, developing, and using these technologies for some years, and the tools for mapping spatial variability have evolved greatly and are no longer limiting. However, machine technologies for the application of the prescriptions, although they have existed for several years, are still not very accessible and limit the use of variable rates liquid application. In our case studies, we have observed savings from 15% to 40% in the volumes of products used, resulting in greater uniformity of areas, such as in the area of 450 ha shown in the figure above. For the creation of the prescription, the NDVI of a RapidEye image and a calibrated algorithm for these fields were used. The application was performed by a self-propelled sprayer with section control and flow variation. The prescribed application rate varied between 50 and 70 L/ha and in 18% of the area it was not necessary to apply the PGR.

The main expected results using the variable rate application of PGR, defoliants and boll openers are product savings, more uniform crop development, higher yields and better fiber quality. We are seeing these results even in areas with uniform soil characteristics, because too many factors can affect cotton development.

Leave a Reply