Neil Maitland grows sugarcane near Aloomba outside Cairns. For him, he wants to really know what’s going on with his crops and how they’re growing, and have real-time, accurate information so over time he can make better decisions.
“What it does by having it in real-time means we can isolate rain events, know whether we’re getting runoff, and maybe we can come up with some better practices that reduce the impact,” Neil said.
“I think the information is the most important thing first so we can build up some knowledge of what’s really happening.”
Nitrogen (N) losses into the environment from intensive crop production in wet tropical catchments of north Queensland are a major threat to the health of Great Barrier Reef ecosystems, and fertilisers applied to crops are an important source of that N. Considerable effort is being put into optimising N fertiliser management in the region, but more change is needed to meet water quality targets. This situation begs the question of how could change be accelerated?
Let’s face it: change is generally hard. For farmers, there are three main obstacles to tackle here. First, farmers are unsure about the link between N that appears in the environment and N fertiliser management on their farm. Second, there’s uncertainty about the effectiveness of new farm management practices and, third, it’s difficult for famers to assess the success of these new practices. These barriers are exacerbated in sugarcane production in north Queensland because farmers lack easy access to information on the impact of sugarcane production on N losses to catchments, extreme variability in climate and thus production, and the high variability in rainfall over very small distances.
We’re developing a suite of web apps, called 1622™, to deliver information services to farmers that address their concerns, make for faster change in farm management practices and help farmers reduce impacts of cropping on the Great Barrier Reef. Our 1622™ apps are at different stages of product development, with our first, 1622™WQ (‘water quality’), available at 1622.farm, released in January 2020.
1622™WQ: real-time water quality information
Using water quality data from high frequency automatic sensors deployed in coastal catchments for research projects and the Great Barrier Reef Catchment Loads Monitoring Program, we’re importing data streams and displaying them in the 1622™WQ app. 1622™WQ has been specifically co-designed with farmers to meet their needs. The app also features advanced data analytics and state-of-the-art deep learning architecture to automatically curate the data streams, which is not available in other data delivery platforms.
This app displays nitrogen losses at multiple sites, allowing farmers to see, for example, the influence of recent rainfall on water quality, how water quality differs between locations, or whether management actions such as recent fertilising has affected nitrogen losses.
We are also looking at new ways to estimate missing data and predict water quality in the days or weeks ahead based on novel applications of artificial intelligence.
1622™WQ also displays data from both Bureau of Meteorology weather stations and local on-farm weather stations to provide farmers a more accurate picture on rainfall variability.
1622TMWhatIf: a risk-based approach to optimising N fertiliser management
The second app in the suite of 1622™ products is 1622TMWhatIf. A prototype of this app is being tested with farmers and advisors, and we anticipate the first version will be released later in 2020. The Australian sugarcane industry has well developed recommendations for N fertiliser management that have been evaluated in around 30 experiments or demonstration trials in the wet tropics. That sounds a lot of studies, especially given the relatively small area - 176,000 hectares - of sugarcane crops in the region, but the wet tropics is a very heterogeneous region with a wide diversity of soils and climates. In the Tully region for example, recommendations have been extensively tested in only three main experiments, located on soils which cover only 26% of the region.
Further, there are two distinct sub-climates in the region and all experiments were located in the southern one. The other factor is the limited time (several years) in which the trials were conducted, and therefore capturing only a small fraction of the annual climate variations the region experiences. This lack of trial information creates uncertainty about the optimum N fertiliser management in farmers’ minds.
Cropping systems modelling is a way to extrapolate this limited experimental experience to different soils, climates and across years, and we have developed 1622™WhatIf to give farmers site- and time-specific information on the effects of N fertiliser rate on crop performance. For example, what if I change my fertiliser rate, harvest date and/or fertilising date, how would that affect my crop yields and nitrogen losses? Importantly, yield outputs from the app are expressed in terms of likelihood of yield loss, to both convey the uncertainty inherent in predicting the future behaviour of cropping systems and to allow farmers to integrate crop performance predictions into their own risk management preferences.
As well as yield, outputs include predictions of N losses to the environment through different pathways. This information will help farmers to participate in emerging markets for abating both greenhouse gas (i.e. nitrous oxide emissions from soils) and water-borne nitrogen discharges (https://www.reefcredit.org/) from these catchments.
1622™Crop: novel remote sensing of crops
The third product we are developing in the 1622™ app suite is 1622™Crop. This app is currently under construction but watch this space for a release in the not too distant future.
Persistent cloud cover in wet tropical catchments limits image acquisition from traditional satellites such as LANDSAT. Hence developing new and novel techniques for more timely monitoring of crop performance from satellite- and drone-based sensors is critical for farmers to better evaluate the effects of changed N management, particularly early in a crop’s life.
We are examining newer satellites, for example Sentinel-1 and -2, and developing new analytics to enhance the quality of their images. We’re also using LiDAR and multispectral sensors mounted on small rotorcraft drones to detect fine-scale variations in sugarcane as it grows to improve the efficiency of fertiliser inputs and maximise yields. We’re also exploring “fusing” multispectral and LiDAR data from drones, and using drone-acquired images to better calibrate satellite data.
We’ve tracked the performance of the crop from establishment to harvest, and can easily identify fertiliser-deficient sugarcane, which usually appears shorter and sparser as early as 10 weeks from planting.
Some of our research to date shows sugarcane farmers can cut their fertiliser use without losing profit.
In conclusion
The imperative to reduce N losses from sugarcane farming in the wet tropics means that the status quo in N fertiliser management is no longer tenable. It needs to be “disrupted”. It comes at a time when digital agriculture is developing rapidly, and harnessing these developments is a valuable opportunity to chart the future course of optimising N management.
As Neil said, “We have to know precisely that we’re not going to lose any production by cutting back on nitrogen. Production is the main thing but if there’s a better way of doing it, we need to know it.”
Check out 1622™WQ for yourself at 1622.farm or follow our developments on Twitter using #1622app.