Earlier this year, we started a new series on our blog to detail the crop production lifecycle – or the key areas that are focused on helping to drive agronomy ahead and the role that technology plays. In our first post, we highlighted research and product development; and today, we’ll go into a deeper dive around plant breeding.
Experts estimate that plant breeding has existed about as long as agriculture itself; mostly because agriculture started with taming wild plants and starting to create hybrid seeds to improve growth and yield.
It hasn’t been until the last few decades where plant breeding started to get more attention. The practice changed significantly with advancements in technology, but it also started to get more traction because there were bigger trends at play.
Read any article about future food productivity and you’ll likely see a statistic from the United Nations Food and Agriculture Organization, which estimates that farmers must produce 70% more food by 2050 to feed the world’s population.
What does this mean? It means that agronomy needs to continue to advance. While there are products, planting strategies, and sustainability initiatives that can support maximizing yields, the biggest gains will come from the development of hybrid seeds through the practice of plant breeding.
Plant Breeding: What are the Challenges?
Like much of agriculture, one of the biggest challenges for plant breeding comes from being able to access the data – or measurements – to analyze and validate outcomes, helping to identify and inform the trait selection process.
There are a few key challenges that have traditionally caused challenges with getting this data.
Unlike most industries, the hyper-locality of agriculture puts different constraints on data for the industry. There isn’t a core location where activities occur; it can vary field by field – even if the fields are located within the same county.
This has posed challenges from a data collection perspective. With numerous fields that need to be audited throughout the season, how can you account for data variability? How can you ensure that you are collecting the swath of data points that you require to validate outcomes? And, how can you ensure accuracy?
But, like most things, the current processes may have challenges, but they have brought us to this point and advanced the practice of plant breeding to where it is today.
How Can Technology Advance the Practice of Plant Breeding?
The reality is that demands of our current world require a more advanced approach to plant breeding.
From uncovering how we can grow crops more sustainably to meeting the food productivity needs of the future, unlocking plant breeding methods that can move trait selection and development forward faster than we have in the past will be paramount to our success in feeding our future world.
Technology plays a role by helping to uncover the astute differences between different traits and plant lines to truly identify which one will produce the best outcomes. Using an ag drone system, for instance, can capture aerial imagery that depicts crop health throughout the season. These images can then be turned into detailed data sets via machine learning and artificial intelligence, which means that the data sets – or the numerical measurements – can be used.
Aerial imagery can serve as one type of data used in a more sophisticated agronomic model, which is designed to use a number of different variables within a machine learning engine. The result? A retrospective analysis using historical data, or real-time predictions or forecasts as new data becomes available around maturity date, yield and quality, and nutrient status.
It can open the door to more standardized data. Traditionally methods of having people in the field counting leaves, assessing plant health, and identifying maturity parameters can be open to subjectivity. It can also result in a higher error rate.
Secondly, using these manual data collection processes can limit the amount of data that you are able to capture. There are only so many hours in a day. And, there are only so many plots that humans can analyze in this period of time.
Drones can solve both issues by eliminating variability. As a machine, it provides objective data. Radiometric calibration can further support the quest for standardized data that can be compared to different measurements throughout the season to understand growth, maturity and vigor.
What’s important to note: It does not have to be one way or another. Using a combination of aerial imagery technology with manual methods can provide a robust solution; a solution that augments data collection and analysis while still capturing boots-on-the-ground qualitative analysis to further support data sets and measurements.
Where to Start?
Starting with new technology can be overwhelming; particularly when your goal may be running a successful plot trial program to understand the performance of a particular trait.
As we noted above, it doesn’t have to be an extreme shift. Starting with a trial to see how it works within your process and the data sets that are captured can give you enough insight to understand the ROI and what it can do for your plant breeding program.