Funded by EPSRC, Innovate UK and Cytiva, the research could help growers work out the best time to use fertiliser on their crops and how much is needed based on factors including weather and soil condition.
This could reduce expensive and damaging effects of over-fertilising soil, which releases the greenhouse gas nitrous oxide and can pollute soil and waterways.
Over-fertilisation has so far rendered 12 per cent of once-arable land worldwide unusable and the use of nitrogen-based fertiliser has risen by 600 per cent in the last 50 years. However, it is difficult for crop growers to precisely tailor their fertiliser use: too much and they risk environmental damage and money wastage, too little and they risk poor crop yields.
Named ‘Chemically functionalised paper-based electrical gas sensor (chemPEGS)’, the team’s sensor measures levels of ammonium in soil, the compound that is converted to nitrites and nitrates by soil bacteria.
Using machine learning, it combines this with weather data, time since fertilisation, pH and soil conductivity measurements. It uses this to predict how much total nitrogen the soil has now and how much it will have up to 12 days in the future to predict optimum time for fertilisation.
Published in Nature Food, the researchers’ study identifies how the AI solution could help growers yield maximum crops with minimal fertilisation, particularly for fertiliser-thirsty crops like wheat. According to the team, the technology could simultaneously reduce growers’ expenses and environmental harm from nitrogen-based fertilisers, the most widely used fertiliser type.
“It’s difficult to overstate the problem of over-fertilisation both environmentally and economically,” said lead researcher Dr Max Grell, co-developer of the technology at Imperial’s Department of Bioengineering. “Yields and resulting income are down year by year, and growers don’t currently have the tools they need to combat this.
“Our technology could help to tackle this problem by empowering growers to know how much ammonia and nitrate are currently in soil, and to predict how much there will be in the future based on weather conditions. This could let them fine-tune fertilisation to the specific needs of the soil and crops.”
Nitrous oxide, released into the air by excessive nitrogen fertiliser, is 300 times more potent than carbon dioxide. Excess fertiliser can also be washed by rain into waterways where it deprives aquatic life of oxygen, leading to algal blooms and reduced biodiversity.
Current ways of measuring soil nitrogen involve sending soil samples to laboratories, a lengthy and expensive process with results of limited use by the time they reach the grower.
The team hopes its new low-cost approach could expedite the process of testing the soil. Researchers expect chemPEGS and associated AI technology, currently in prototype stage, to be available for commercialisation in three to five years with more testing and manufacturing standardisation.