Dalhousie University study applies artificial intelligence to sustainable dairy farming

A researcher at Dalhousie University in Halifax is applying modern technology to ancient practice to identify trends in methane emissions from Canadian dairy farms.

Suresh Neethirajan is an associate professor and research chair in the faculties of computer science and agriculture at Dalhousie University. He led a study to benchmark methane emissions that could predict future trends in the Canadian dairy sector.

Neethirajan said benchmarking is a systemic process that measures and compares performances practices and outcomes against established standards.

“It's kind of having a reference point to evaluate how we are doing currently and identify how do we go about improving and putting in achievable goals as a country,” Neethritrajan said. “We had a 2030 net zero goal, and we moved the goal post to net zero 2050. How do we actually go about it?”

Researchers used satellite data from NASA and the European Space Agency to track weekly methane emissions and concentration data from 575 dairy farms and 380 dairy producers. They identified that emissions peak during autumn. They also found that Ontario has the highest methane concentration in the country.

Neethirajan said it’s like monitoring your energy output at home so you can implement strategies to cut costs.

“The research we have done establishes a baseline for methane emission levels across Canadian dairy farms and processors,” he said. “If we have to do a comparative analysis, we need a certain standardized technique.”

The Canadian dairy industry contributes to methane emissions through the digestive processes of cattle and manure management. These activities have led to a fourfold increase in emissions over the past 130 years said a news release from Dalhousie University.

Neethirajan said standardization allows him to study changes across regions or during anomalous events — such as the COVID-19 pandemic — to see what factors most affect emissions. He listed different weather patterns, feed storage and farm sizes as examples of factors worth studying. He said it’s all made possible by artificial intelligence.

“It's like the ultimate data detective,” Neethirajan said. “We have shifted through about 10 to 15 years of satellite data to understand and uncover hidden patterns in the methane emissions.” Neethirajan said the technology manages millions of data points and helps develop predictive analytic tools that show future emission trends.

“This predictive capability is crucial for developing proactive strategies and advising policy makers in the Government of Canada to be able to mitigate emissions before they escalate,” Neethirajan said. “It also helps to understand the historical changes, even when satellite data is not available.”

Neethirajan said the technology they’ve developed can be applied to other sectors, such as poultry, through comparative analyses.

“If the number of dairy cows in a farm is approximately 100 it typically emits about 11,500 kilograms of methane over 12 month period, which is roughly equal to 74 cars emitting carbon dioxide annually,” Neethirajan said. “So a similar basic framework can be easily developed for other species.”

Neethirajan said if we know how much farms are emitting we can develop thresholds and systems that help Canada become net zero by 2050. He said this creates a competitive advantage for low emission farms.

“These tools provide the ability to implement preemptive, prescriptive strategies, such as optimizing the feed regimens, the cleaning protocol, manure management in the dairy farm,” Neethirajan said. ”By using the tool we have developed the dairy industry has the potential to achieve a much more sustainable and economically viable future, because the environmentally responsibility and profitability go hand in hand.”

For more Nova Scotia news, visit our dedicated provincial page.