Accurately predicting sugarcane yield could be the key factor for a successful harvest season. Good historical data can be an optimal way to develop yield models that quantify the relationship between field actions and yield outcomes. But sometimes this data is not available or requires time and work to gather. Digital Harvest has developed a powerful tool to estimate yield using remote sensing and other external data sources starting with a simple field polygon.
Understanding and predicting yield is critical for the sugarcane industry because a lot of decision making depends on having accurate predictions, such as harvest timing, logistics and forecasting revenue and profits. Digital Harvest has developed a powerful tool to predict yield outcomes at field level with just your field polygon and dates of growth start and intended date of harvest. This solution scales from individual fields up to entire regions. These predictions can be made months in advance using satellite data, daily weather, and minimal cultivation details about the crop. Our tools and predictions don't end here because with the combination of our multi layered biophysical model and supervised machine learning approach we can get even more accurate field level yield predictions to get your organization to the optimization level.
The first layer of this predictive model approach is a sugarcane-specific biophysical simulation model. This model is the base for all our predictions. It is a process-based model that describes key ecophysiological processes to determine plant growth and development. It simulates four main processes that take place in the sugarcane: C4-photosynthesis, partitioning of carbohydrates and sucrose accumulation, respiration and senescence. These processes are simulated differently for green leaf blades, non-millable top, cane, and roots, as their behavior is different. Last, three developmental stages are defined according to GDD (Growing Degree Days): Emergence, Tillering, and Grand growth and Maturity. It’s important to mention that this model assumes that there are no nutrient or water stress conditions.
One of the main advantages of this model is that it requires few data from the customers: a polygon for each field and dates of growth start (or the date a field was most recently harvested) if you don’t have this data we can help you with our polygon drawing app and our proprietary Land Clearance Detector (LCD) to derive the last date of harvest. With only this, Digital Harvest can obtain valuable remote sensing data, such as daily maximum and minimum temperatures, daily leaf area index, or perform the calculations to obtain daily active solar radiation, and daily day length.
We want to invite all sugar cane mills and producers with more than 100 hectares to test our tools. Just click on the next link to go to our SIP section, and read all about it.