Farming through the lens of a satellite

Views

Managing on-farm inputs has always been one of the keys to a successful farm business, and in recent times this has become even more pivotal with high prices for urea, chemical and even some seed. So, in a world where we more frequently question just exactly how much product we should be putting out, it is no surprise that many are turning to technology to help with the more strategic placement of these inputs.

One such technology aiding in this process is satellite imagery. Yes, I know, this isn’t particularly new technology, in fact we have been using it for years—GPS, Google Earth, NDVI imagery, etc. However, a lot of untapped potential remains.

COALA

The COALA project (which stands for ‘COpernicus Applications and services for Low impact agriculture in Australia’), is looking into expanding the application of satellite imagery to support decision making in agriculture, using technologies proven and operational in Europe.  The COALA project is funded by the European Union. They are introducing satellite imagery-based services around paddock zoning, crop yield potentials, in season nutrient status and others more suited to irrigation in the Australian agricultural environment after success with European farmers.

The satellite imagery is provided by the Sentinel-2 satellites. Each satellite in this constellation orbits the earth every five days, with a spatial resolution of 10m, 20m and 60 m. Satellite data is compilations of millions of pixels that relate to a distance on Earth. Sometimes, a pixel can represent the equivalent of one square metre on the ground; in the case of the Sentinel-2 satellites a pixel represents an area of 10 x 10 m on the ground (or 20×20 m if we use images that capture light in the infrared part of the spectrum).

BCG is one of eleven partners from across Europe and Australia that make up the project consortium. Over the last two years BCG has been working closely with fellow project partners AgriSat Iberia, S.L., a Spanish agronomic consultancy company along with growers from the Wimmera and Mallee to trial the following services:

Some of the features of each product are outlined in Table 1.

Table 1: The services being trialled by BCG and their potential application in broadacre cropping.

COALA Service How they are created Potential application Suitable crop types
MZMs From layers of satellite imagery measuring crop biomass

 

Characterising the paddock in zones to aid in variable rate applications of seed, nutrients, soil sampling etc. Cereals, canola and pulses

 

NNI Maps A combination of using Sentinel 2 satellite red edge bands using the MTCI (MERIS terrestrial chlorophyll index) and crop biomass Aid with nitrogen decisions around when and how much should be applied during the season. Wheat and barley (and under the project canola is being trialled)
Yield forecast A model called MYRS uses a time series of NDVI images and selected meteorology parameters to model crop biomass and yield Aid with logistics management at harvest and selling grain. Wheat and barley

Interim results and grower feedback

The first two years of results (2020 and 2021) has shown some promising alignment with the satellite yield forecasts and the actual paddock yields (Figure 1 and 2). There were a few outliers but upon further interrogation we have been able to explain the gaps, i.e. longer season varieties weren’t as well predicted due to the timing of the satellite images used, frost events and herbicide damge were just some of the contributing factors.

Cereal yield estimated (MYRS) vs. measured for 3 growers’ paddocks in 2020

Figure 1: Cereal yield forecasts (estimations) in comparison to the actual paddock yields in 2020 across the Wimmera and Mallee. Source: AgriSat Iberia. S.L.

Cereal yield estimated (MYRS) vs. measured for three growers’ paddocks in 2021

Figure 2: Cereal yield forecasts (estimations) in comparison to the actual paddock yields in 2021 across the Wimmer and Mallee. Source: AgriSat Iberia. S.L.

The MZMs have generated some alternative imagery for paddock zoning which has built upon existing methods by the growers trialling the services. For some of the paddocks where the growers have still been tweaking the zoning the MZMs have anecdotally been helpful. Figure 3 is an example of a MZM generated for one of the growers paddocks and Figure 4 shows how this info can then be used to generate a variable rate (VR) prescription map.

 

Figure 3: An example of an MZM map. The ‘cool’ colours are areas below the average paddock yield and the ‘warm’ colours are above Source: AgriSat Iberia. S.L.

Figure 4: An example of a VR prescription map generated off the zones from the MZM. Source: AgriSat Iberia. S.L.

The NNI maps have shown some interesting spatial variability of nitrogen during the season and now need further ground truthing.

What’s ahead?

This season we will continue to work with the three growers already engaged with the project to continue to validate the services and ground truth the satellite imagery as well as provide assistance with VR appliations based off the satellite imagery. A selection of this imagery i.e. NDVI, RGB, NNI, MZM will also be available for Yield Prophet users to view (but not Yield Prophet Lite).

If you are interested in trialling the services or to hear more about the project get in touch with Kate at [email protected] or 0448 823 353.

Upcoming Events

Become a BCG Member

BCG exists for its members. Research and extension activities are designed to provide members with information and resources that will help them improve the productivity, profitability and sustainability of their farm businesses.

Improve your profitability

Receive the latest research, extension and event news direct to your inbox! For a limited time, receive a free technical bulletin when you subscribe.