What does the research say about decision support in forest planning?

For operational harvest planning, that the tools are strong on calculation and thin on reported proof. A 2025 systematic review of 23 studies of harvest planning tools, published from 2005 to 2024, found none that explicitly reported independent validation and none with open code. At least 19 depended on parameters specific to one site. A March 2026 conference abstract, from a survey of 42 tool managers, found tools strong on timber yield, stand development and carbon, and weak on insects and pathogens. It also found the tools have "advanced considerably".

What the usual answer says

Standard accounts say decision support systems improve harvest scheduling, sustainability and multi-objective trade-offs through optimisation, simulation and mapping, with adoption held back by complexity and data. The research agrees on the capabilities. It is sharper about what the published work does and doesn't show.

What the 2025 review found

Jaffray, Coupland and Paradis posted their review as a preprint on EarthArXiv on 15 October 2025. They searched two major databases and kept 23 peer-reviewed studies of operational harvest planning tools, covering machine location, road design and harvest scheduling. Eleven used mathematical programming, four simulation, three spatial systems and three AI or machine learning. Most, 17, were developed on real-world case studies. Cost was the most common objective, in 10.

The reproducibility counts are low. "Independent validation (separate test data or third-party replication): 0 explicitly reported in the extracted set." Open code: 0 of 23. Public datasets: 1 of 23. A prototype interface for users was mentioned in 2. Site-specific parameters were required in "at least 19/23".

Two caveats apply. The counts are of what the papers reported, not of what their authors did. And the review team used an AI assistant "to assist with coding during data processing and analysis", which they disclose.

Their summary: the models "provide valuable insights and demonstrate technical expertise", but "they are often hard-coded to specific sites, lack reproducibility and are rarely open-source". They don't treat site-specific inputs as a flaw to remove. They argue for frameworks "that embed site-specificity as a structural element rather than a limitation".

What the other studies add

A March 2026 conference abstract by Mazziotta, Kurttila and Vacik, not peer reviewed, reports a survey of 42 decision support tool managers, scored against 40 variables. It finds that tools "have advanced considerably, enabling multi-objective analyses and holistic assessments that were unattainable a generation ago", and "increasingly integrate ecosystem services and climate-related risks". They do well "in traditional forestry domains, like estimating timber yield, stand development metrics, and carbon accounting". But "few address biotic threats (insects, pathogens)", and "many tools remain tailored to scientific rather than operational contexts".

A September 2026 preprint by Labarre and colleagues, on plantation forests, found that adding fire simulation to a growth model changed the optimal plan, "favoring diversified portfolios that balance high-yield but fire-prone regimes with lowerrisk alternatives". In our reading, a tool that leaves a real risk out will tend to show a plan as better than it is.

Reading it through a data lens

The review measured validation in the model-building sense: testing on separate data, or replication by others. An operation cares about something the review didn't count, which is how a plan compared with what followed. In our view, that is mainly a records question. It needs the plan as made, kept beside volumes by block, costs by operation and dates as they happened.

With those records, a buyer can put concrete questions to any tool. Did its volume estimate for a block match what the scale tickets later showed? Did its cost model predict what contractors invoiced? Did blocks scheduled for summer get cut in summer? A provider that can answer from a client's history is offering more evidence than the published literature does.

The strongest objection

Commercial planning tools are checked in use every day, and those checks are rarely written up. The 23 papers are mostly method demonstrations on case studies, whose job is to show a formulation can be solved. So the review describes the literature, and says little about any particular tool an operation might buy.

When it doesn't apply

Small ownerships planned by rule of thumb gain little from formal tools. Strategic plans running decades ahead can be checked only against intermediate outcomes. And the 2025 review covers operational harvest planning, a narrower field than forest planning as a whole.

Quarri for forest management is built around how a forest operation runs, from the cruise to the settled account.

Quarri is an AI-native data platform for the timber supply chain. It connects buying, production, sales and inventory for forest management, sawmill, wood products and pulp, paper and packaging operators.