The FieldApex 2018 results are out!
The 2019 season is approaching. Let’s see how FieldApex performed in light of the 2018 crop yields.
The particularly dry season of 2018 has been providential in demonstrating the value of using FieldApex year after year knowing that nitrogen needs can vary from up to 112 lbs/ac for the same plot, depending on whether the season is very dry or very wet around fertilization time.
High satisfaction with commercial validation
Based on the technical results we obtained from our main agronomic partners, on average 8 out of 10 FieldApex estimated rates were the closest to the most profitable rate.
This demonstrates the value of FieldApex in commercial conditions and makes it an essential tool for agronomists and producers willing to apply the principles of 4Rs: right rate, right source, right time, and right place.
To carry out their validations, our partners used their own protocols, such as test strips, or comparisons with other tools for assessing nitrogen requirements or other measurement methods of their choice.
Testimonials from participating agronomists
|Jacques Nault, Vice-President and Agronomist at Logiag, discusses how he introduced the use of the FieldApex algorithm to farmers, and the spectacular results they got in terms of yield and reaching the economically optimal nitrogen rate in 2018.|
|Dale Cowan, Senior Agronomist at AGRIS Cooperative discusses how he introduced the use of the FieldApex algorithm to corn producers, how nitrogen rates were predicted well in 2018, and how weather data is a critical input that adds accuracy to the prediction.|
Independent scientific results
The SCAN.AI algorithm has been considerably improved prior to the 2018 season. According to the latest calibration from Agriculture and Agri-Food Canada and Université de Sherbrooke team, the algorithm achieves a level of accuracy still unmatched with an RMSE of 8 lbs N / acre.
As shown in the figure below, economically optimal rates are very close to the estimates of our algorithm.
The accuracy of the estimates has been reached using the following methods, as shown in the following figure:
- field tests in Canada;
- data collection from Quebec and Ontario databases;
- a major North American database.
A simple approach with increasingly suitable results
Another distinguishing feature of FieldApex is its metadata-based approach. It minimizes logistical costs by estimating optimal nitrogen rates with no need for soil samples or laboratory analysis.
As shown in the figure below, the most commonly used tools in nitrogen management, i.e. Pre Side-dress Nitrate Test (PSNT) or Late Spring Nitrate Test (LSNT) are not able to help generate accurate estimates of nitrogen requirements.
These tools are designed to detect conditions that would enable a reduction in nitrogen rates as measured in a typical soil sample using a soil nitrate-nitrogen in parts per million (ppm) model. However, they don’t provide conclusive estimates, particularly when measurements are below 15-20 ppm —which turns out to be the vast majority of cases.
Broadly speaking, the relationship between the economically optimal rate and the estimates calculated by the different formulas used with these tests remains vague.
In comparison, FieldApex stands out for its ability to provide a much more accurate and balanced estimate in much larger rate intervals. Therefore, the solution is better suited to reflect the agronomic reality of each field, as evidenced by the validation below carried out with independent data covering the years 2014, 2015 and 2017 assembled from partners by Agriculture and Agri-Food Canada.
Source: Agriculture and Agri-Food Canada
Considering how close the economically optimal rates are to the FieldApex estimates and the magnitude of the latter’s intervals, FieldApex appears to be the market solution that currently offers the best mean to assess nitrogen requirements in corn crops.
Customizing for the specificities of each field with deep learning
Evaluating the specificities of each field is paramount, such as compaction zones and agronomic characteristics impacting soil health. Therefore, it is important to build the history of each field. FieldApex is an excellent place to start to support you in your analyzes for each growing season. The future development of autonomous learning capabilities through machine learning and deep learning is expected to enable further customization of FieldApex estimates while keeping the rate calculation process just as simple.
Where to find FieldApex
Discuss about FieldApex with a partnering advisor to see how it can contribute to increase your field profitability and protect your yields.