Written by:  

Guy Ariëns

Data-driven sustainability: minimizing fertilizer usage in sugarcane mills

Challenge

Colombian sugarcane mills are facing a challenge that is making it hard to increase profits. Given that each mill oversees more than 5,000 fields spread across 25,000 hectares, they are dealing with an enormous amount of land that is nearly impossible to monitor closely. 

And the crop does need monitoring – as it provides insights into which areas in the fields are lacking growth and require additional fertilization. It is crucial to have these insights in an early phase of the growth cycle since, after the fifth month, the density and size of the sugarcane crop limits fertilization options.

Fertilizer is not only costly, but its overuse can also harm the environment. Therefore, it is paramount for the mills to accurately discern where and how much fertilizer to apply, a practice known as “precision fertilization”.

Together with eLEAF, we’ve been looking for smart solutions to enable these sugarcane mills to do precision fertilization.

Technology

In the first months of the growth cycle, we use optical and radar satellite images, supplemented with weather data, to determine the quantity of sugarcane (in kilograms) per field. This information is used to make a forecast of the expected yield after a 12-month period.

Each field has a unique yield target (measured in tons/ha), which is influenced by factors such as soil quality and location. We compare these yield targets to the yield forecasts to provide sugarcane mills with insights on where to apply additional fertilizer. And where to reduce regular (scheduled) fertilization. This is a cost saving, but more importantly: has a positive sustainability impact.

Our (ViewJS) application enables the sugar cane mill manager to set own notification rules. Such as “notify me when the crop growth in month two is more than 5% lower than expected for a specific field.” 

Over time, we will be able to give smart recommendations in the app using a machine learning (ML) model. The model collects and stores data such as yield numbers, app-generated rules, and mill managers' feedback in a database. Once enough data has been compiled, the ML model can provide tailored notification rules to mill managers, to further increase yield and profit.

Impact

With precision fertilization, the mills do not have to overuse fertilizer anymore to achieve optimal yields; a significantly positive impact on the environment. 

Furthermore, sugarcane mills have the potential to increase yield in the areas that were lacking growth – resulting in a 6% to 14% increase in productivity per hectare. Given an average yield of 100 tons/ha (€3500,-), this translates to an additional revenue of €200,- to €500,- per hectare. Most of this revenue is direct profit, as the mill manager does not need to use much extra fertilizer – just redistribute it to areas in need.

Consider how this seemingly small profit increase could add up for a sugarcane mill operating on 25,000 hectares. 
And think how a seemingly small reduction of fertilizer per hectare will result in a large positive environmental impact.