What should a forestry database hold, and where?

A forestry database should hold dated measurements and dated events as well as the current total for each stand. Every cruise, harvest, load, contract and payment gets recorded with its date, its source document, and the same stand and contract identifiers the map and the ERP use. The stand shapes belong in the GIS, the invoices in the ERP, and the forestry database holds what links them over time. Of the two requirements, we find the shared identifiers are the harder one. Dated records usually exist somewhere, in cruise reports, scale tickets and settlements. Joining them is where it breaks.

What the usual answer says

The top results for this question are public databases describing their contents: growing stock, biomass, carbon, species, growth and mortality. A public subcompartment dataset calls itself an authoritative source for recording and reporting. Those descriptions are right for what they are. They say little about what a forest owner, manager or harvesting business should keep about its own land and operations, or how. A forestry consultancy's description of its own client geodatabase is closer, and typical: "Examples of attributes we track include forest type, recommended harvest, date a management plan was written, and more." Those are attributes of each property as it stands now.

What a public inventory does with time

The Forest Inventory and Analysis programme describes its nationwide inventory as "a national network of permanent plots remeasured every five to ten years depending on location." Its programme page shows the two design choices that matter for a business database.

The first is that each visit is its own record. A system that keeps only the latest cruise for a stand can report today's volume. It cannot report growth, or recompute past figures when a volume table or a cruise method changes. Keeping each measurement as a dated record, with its method, costs little and keeps both options open.

The second is that the units the data is reported by can move. The same page carries a user bulletin warning that some area-based estimates "require careful interpretation", because local boundaries "have changed multiple times since the inception of the FIA annualized inventory (~2004)." A forest business has the same problem at stand level. Stands are split, merged and renumbered after harvests, sales and replanning, and a figure keyed to an old stand number no longer lands anywhere on the current map.

Measured and modelled records

A second reason to keep method with each record is that more inventory figures are now modelled between visits. A 2024 study in Forestry by Mulverhill and colleagues, read from its abstract, notes that laser-scanning inventories "represent a static point in time". It updated one for a management unit of about 690,000 ha at a two-week interval from satellite data. The models' relative error ranged "from 11.47% (canopy cover) to 31.82% (stem volume)". A modelled update is useful, but it is a different kind of figure from a cruise. A database that stores both in the same field, without saying which is which, loses that difference for good.

What to hold

Hold measurements as dated records, each with the method used. Hold operations as events: each harvest, each load with its scale record, each treatment. Hold commercial records as events too: each contract and its terms, each settlement, each stumpage or levy payment. Give every record the identifiers of the stand and the contract it belongs to, using the same identifiers as the GIS, and keep a table of how stand identifiers have changed over time.

Hold the source document with the figure, or a link to it. A ledger total without its document can't be checked, and when the ledger and the documents disagree, only the documents can settle it.

Where each part belongs

Shapes and spatial analysis belong in the GIS, which is built for them. Invoices, payables and the general ledger belong in the ERP. Scanned contracts and tickets belong in a document store the database can link to. The forestry database, or a data platform acting as one, holds the dated measurements and events and the keys that connect all three. It should point to the stand shapes and the ledger rather than copy them, so that a change in one place does not leave a stale copy in another.

Whether that sits on a server in the office or in the cloud matters less than two things. It should be backed up, and it should keep its history, and its identifier table, when systems are replaced.

When it doesn't apply

A small woodland with one inventory and an occasional timber sale can keep its records in a well-organised folder and a spreadsheet. The discipline of dated records and linked documents still helps there. A business that only buys delivered wood and owns no land has little use for stand measurements. Its contracts and loads still benefit from the same event structure.

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.