La Candelaria: Precision Agriculture

La Candelaria consists of 118 lots which are constantly being seeded, receiving chemical applications, being monitored for humidity, temperature variations, topographical changes, or being harvested. Understandably, with approximately 12,833 hectares being cultivated, there is always work to be done by the agricultural division.

Currently, the team is in the first steps of developing an online geographic information system (GIS)database to improve their knowledge of their fields and thereby improve yields. This was my first introduction to precision agriculture.

There are a number of variables that exist within fields. By doing landscape analyses of the harvest yields annually, slowly a landscape inventory of the variations in the fields' productivity is being developed. Currently the project is in its third introductory year. In 2010, the first harvesting machines with the ability to track and document the quantities being harvested at precise locations for each crop. The ability to track this information has been made possibly by using GPS technology amongst others. The following year, in 2011, pretty much all of the machines had the technology to capture the precise yield information more fully, though some holes still exist.

Here’s how it works. First the raw data is exported to the company's network in the office. The raw data then needs to be input manually into an agricultural software called SMS by Ag Leader. The information currently needs to be input manually for two reasons. 1) Unfortunately, the GIS software systems for each of the major agricultural mechanical manufacturers - John Deer, New Holland, Claas – are currently incompatible, and since contrators don't all use the same mechanical brands to harvest, the administration needs to assimilate the information themselves. 2) La Candelaria developed a lot (field) numbering system to organize the land being cultivated over many decades. This information isn't currently represented in the collection software by the mechanical harvesters. For this reason, the administration needs to identify and label each lot manually when inputting the data into the SMS system so the information can be appropriately referenced.

Once the information is in the SMS system, different variables can then be isolated and mapped for analysis. Though the software is complex enough to map very precise variations, currently the administration is creating maps with three colour variations for simplicity - green, yellow and red. Using the mapped out yields, it is then the task of the administration to determine if the lower yields are because of humidity, pests, disease or other factors.

About five years of information is needed on a field before any major decisions can be taken since weather changes from year to year and the overall climatic conditions of each year also need to be considered for yields.

Ultimately, the assimilated field variation information will be able to be transferred to the seeders, fertilizers, and fumigators to improve the efficiency and precision with which chemical applications are dispered or seed varieties are chosen to meet specific growth conditions. The goal with precision agriculture is to take advantage of modern agricultural technology to maximize production. For example, there are machines that currently use ‘green technology’ when fumigating, which means the machines have the ability to detect weeds and apply herbicides only in the locations where weeds are growing. I am very impressed to learn about such technologies, as I was under the impression that most often chemical applications were spread across the entire field regardless of need. Though I do not know the commonly practiced methods in Canada and the US,  this technology is certainly more cost effective and more environmentally friendly.

In these first years, a major challenge has been that not all contractors have the technology in their machines to map the fields. Also, some don’t have the proper training to make sure the information is being collected properly. This has resulted in discrepancies from year to year. One of the responsibilities of the agricultural division during harvest season is to be on top of which machines have the technology and to calibrate the machines, which takes about 20 minutes, once a week to ensure the information being collected is accurate. For the agricultural division too, learning to apply the technology takes time. Ag Leader provides a trainer that comes in to improve the capacity of the administrative team on its software. However, scheduling and finding time amongst the other responsibilities on the farm is often difficult.

To test the technology, the administration has conducted experiments on certain lots. For example, on one sunflower lot this year they applied fungicide in its usual dosage on half of the field and didn’t apply anything on the other half. Come harvest time, the sunflowers that had received the application yielded 400 kg more per hectare. The lot was 92 hectares which means about 18,000 kg of seed were lost on the 45 hectares that didn’t receive the fungicide which is certainly significant.

Last, the raw data that is being collected by the harvesters is currently being archived in case one day a program that can process and interpret all the mechanical manufacturers' systems is developed, La Candelaria has its history available and can be ahead of its competition.