A McGill University research team has built a geospatial model of food waste from Montreal businesses and institutions, finding that a small portion of the mapped city accounts for a large share of the estimated disposal.
The model estimates that industrial, commercial and institutional operations send 215.4 ± 7.8 kilotonnes of food waste to landfill each year. Six per cent of the mapped grid cells account for more than half that total, while food service and retail businesses contribute an estimated 83 per cent.
These are modelled estimates, not a direct weighing of waste from every building. That limitation is central to understanding what the study can—and cannot—tell municipal planners.
How the researchers built the map
The team identified 14,508 active businesses and institutions from municipal inspection records and grouped them into 14 subsectors. It then linked establishment locations to building-footprint data and combined those records with operational information drawn from 421 Canadian facility audits.
A Monte Carlo model was used to estimate both the amount of waste and the uncertainty around those estimates. The results were summarized on a grid covering Montreal, allowing the researchers to identify areas where estimated disposal is particularly concentrated.
That spatial detail is the study’s main contribution. A citywide total can show the scale of a problem, but it cannot tell a planner where collection routes, redistribution programs or processing capacity might have the greatest effect.
Why uncertainty matters
Food-waste estimates depend on imperfect information. Buildings house different activities, businesses change over time, and waste audits capture only a sample. The researchers explicitly modelled that uncertainty rather than presenting every location as equally well known.
The paper reports that areas with the greatest estimated waste intensity tended to have lower relative uncertainty. That makes the highest-volume zones potentially useful places to begin targeted policy work, while still leaving room for local audits before money is committed.
The available evidence does not show that the City of Montreal has adopted the model or changed collection policy because of it. Municipal implementation remains unresolved. The team says it is working with researchers in Halifax and Guelph to replicate the approach, which may reveal how well the method transfers to cities with different data and waste systems.
From map to intervention
A hotspot map does not determine the best response. Prevention, donation, animal feed, composting and anaerobic digestion have different costs, regulatory requirements and environmental effects. A dense restaurant district may need a different system from a hospital, school or food-processing site.
The model can narrow the search area. It cannot replace decisions about collection frequency, contamination, storage, transport distance or the most appropriate destination for the material.
The work also connects two Canadian methane stories. Food in landfill can produce methane as it decomposes, while the federal government is trying to build an export market for Canadian detection and reduction technology through AVOID-Methane. Better local waste data and better emissions measurement address different parts of the same climate problem.