May 20, 2026
Reading time: 6 minutes
Canada holds an outsized share of the planet’s natural wealth. One-fifth of the world’s fresh water moves through our rivers, lakes, and wetlands, and the boreal forests and temperate woodlands shelter an estimated 80,000 species. This ecological abundance is a responsibility, and that responsibility now carries a deadline.
Under the Kunming-Montreal Global Biodiversity Framework, Canada has committed to protecting 30 per cent of lands and waters by 2030. The country currently protects 14 per cent. Closing that gap over the next four years, across 9.9 million square kilometres of varied and often remote terrain, demands more than political will. It demands a method.
Our new peer-reviewed study published in the journal Earth’s Future offers one. Led by Camilo Alejo and co-authored by Amy Luers, Andréa Ventimiglia, María-Isabel Arce-Plata, and H. Damon Matthews, “Maximizing nature-based solutions using artificial intelligence to align global biodiversity, climate, and water targets” applies an AI-driven framework to the full sweep of the Canadian landscape, identifying where conservation and restoration actions would generate the greatest simultaneous gains across four ecological dimensions: threatened species coverage, ecological integrity, carbon storage, and surface water stability.
The study arrives alongside the federal government’s own roadmap. In 2026, Canada released A Force of Nature: Canada’s Strategy to Protect Nature, committing $3.8 billion to conservation and naming AI, biodiversity mapping, and carbon data as core instruments for delivering on the 30×30 pledge. Where the federal strategy provides the policy architecture, the Al-driven study provides the scientific foundation beneath it.
A framework built for multiple goals
Conservation planning has long struggled with structural tension. Protecting one ecological value often diminishes another. A wetland managed for water stability may rank poorly for carbon storage. A forest prioritized for biodiversity may conflict with Indigenous land use or extractive industry. Existing frameworks tend to optimize for a single goal, leaving others to be addressed in separate, disconnected processes.
The Al-driven study breaks from that approach. At its centre is a Reinforcement Learning agent: a form of artificial intelligence that learns not from static rules but from experience, refining its decisions across thousands of simulated scenarios until it identifies choices that consistently produce the best outcomes. Applied to Canadian landscape data spanning habitat change from 1992 to 2020, carbon stocks, land cover transitions, and satellite measurements of surface water, the AI agent evaluated conservation and restoration options across the country’s full ecological and geographic complexity.
The result is six priority scenarios stress-tested against real conditions: three for conservation, three for restoration.
Where conservation priorities align
Across all three conservation scenarios, one variable surfaced as the clearest guide to high-value land: irrecoverable carbon. This refers to carbon stored in soils and vegetation that cannot recover by 2050 if it’s released. When the three scenarios were overlaid spatially, the areas appearing in two or all three were strongly associated with irrecoverable carbon stocks concentrated in forested ecosystems.
The conservation scenario pairing threatened species protection with irrecoverable carbon in forest-dominated zones produced the strongest outcomes across all four measured dimensions simultaneously. Relative to Canada’s existing Protected Areas network, this approach improved coverage for threatened species, species richness overall, ecological integrity, and surface water stability. The convergence of biodiversity and carbon priorities is not incidental. Canada’s boreal forests contribute an estimated $703 billion annually in ecosystem services. Its wetlands provide roughly $225 billion per year in water and climate benefits. The AI framework identifies where those values are most concentrated and most exposed to loss.
Two pathways to restoration
Restoration priorities follow a more divided geography. Our study identifies two distinct pathways, each with its own timeline, constraints, and sectoral implications.
The first targets disturbed natural lands, primarily forests in Boreal ecozones and the Montane Cordillera, where habitat quality and carbon stocks have measurably declined. Improved forest management and targeted reforestation in these areas could produce relatively rapid environmental returns. Many of these restoration zones sit adjacent to conservation priority areas, making coordinated management across both objectives spatially feasible.
The second pathway is more demanding. It targets converted lands: agricultural fields, grasslands, and urban zones across the Prairies, Mixedwood Plains, and Atlantic Maritimes. Applicable interventions include tree intercropping in agricultural zones, riparian planting along croplands and waterways, and urban greening measures such as green roofs and bioretention systems. Large-scale land-use shifts in agricultural regions carry real economic costs, and the research acknowledges those constraints directly. Reaching the 30 per cent restoration target will require drawing on both pathways simultaneously.
The pressures on priority land
The AI framework maps the human pressures working against conservation and restoration priorities, across Canadian industry sectors from 1992 to 2020. At present, forestry dominates as the primary pressure on conservation priority areas. Agriculture and urban development exert the greatest strain on water-focused restoration zones. But the picture shifts over the coming decade. Mining emerges as the single largest growing threat to both conservation and restoration priorities. Oil sands extraction in northeastern Alberta, mining along the Ontario-Quebec border, potash extraction in the Prairies, and pipeline development across western ecozones all appear as significant and expanding risks to high-value ecological land. This sector-level spatial analysis provides precisely the kind of evidence base that science-grounded permitting and regional assessments require.
Canada’s nature strategy requires coordination among different sectors and government levels, asserting Indigenous land governance, as Indigenous-led Nature-based Solutions have shown to play a key role in Canada’s environmental targets.
Investing in nature
Canada’s nature strategy also aims to attract private capital and build public-private partnerships by improving how the value of nature is measured and incorporated into financial decision-making. According to the strategy: “the annual funding gap globally, from all sources, to meet the commitments in the Kunming-Montreal Global Biodiversity Framework has grown to over $1 trillion USD.”
Our research emphasizes the importance of a multi-dimensional approach to investing in Nature-based Solutions. Identifying the best combinations of conservation and restoration actions that maximize both biodiversity protection and other environmental co-benefits can help direct funding. Decision-makers need to know not only where ecologically valuable land exists, but which specific combination of actions will produce the strongest simultaneous return across biodiversity, climate, and water outcomes.
Science in service of strategy
Our Al-driven study has limitations worth naming. The framework does not fully account for land conversion costs or local economic trade-offs. It does not yet integrate climate change projections directly, a gap the researchers flag as a priority for future work, and high-resolution scenarios tailored to specific communities or species would sharpen the picture considerably.
However, what our study does accomplish is substantive. It demonstrates that Reinforcement Learning can hold biodiversity, carbon, and water outcomes in simultaneous consideration across a continental landscape. It identifies where Canada’s conservation and restoration investments would generate the greatest co-benefits. It maps, with geographic precision, the industrial pressures most likely to undermine those investments in the decade ahead.
A Force of Nature: Canada’s Strategy to Protect Nature articulates where the country needs to go. Our AI-driven framework developed by Alejo and colleagues maps a rigorous, replicable, and spatially grounded path for getting there. The science and the strategy are oriented toward the same landscape.
References
Alejo, C., Luers, A., Ventimiglia, A., Arce-Plata, M.-I., & Matthews, H. D. (2026). Maximizing nature-based solutions using artificial intelligence to align global biodiversity, climate, and water targets. Earth’s Future, 14, e2025EF007560. https://doi.org/10.1029/2025EF007560
Government of Canada. (2026). A Force of Nature: Canada’s Strategy to Protect Nature. Environment and Climate Change Canada.