Ground Truth: DBH and Core Sampling as Calibration
No instrument reads a tree's age. An increment borer does. The relationship between what LiDAR measures and what a core reveals is the foundation of defensible forest science.
When you tell a hearing that a given tree is approximately 600 years old, you are making a claim that needs to be traceable to an observation. The point cloud doesn't contain that number. TreeLearn's model doesn't produce it. What produces it is a physical instrument — an increment borer — extracting a core from the tree and counting the rings. Everything else is inference from that.
DBH measurement works the same way. LiDAR-derived diameter — a cylinder fitted to points in the 1.3-metre-height band on the stem — is an estimate. It is highly accurate under good conditions: clean stem, sufficient point density, no obscuring understory. But "highly accurate" is not the same as "ground-truth." The ground truth is a diameter tape wrapped at breast height on the actual tree.
We measure both, always. The gap between them is calibration data.
The Calibration Sample
On each block scan, we select a stratified calibration sample: typically 80 to 150 individual trees spanning the diameter range and species composition of the block, identified from the preliminary TreeLearn output. For each calibration tree, a field team member takes three measurements: a tape DBH at breast height, a borer core if the tree is old enough to warrant aging, and a GPS fix precise to 30 centimetres.
The DBH data allows us to correct the LiDAR-derived diameter model. We compute a residual error by species — typically less than 3% across the diameter range on well-scanned stems — and apply that correction to the full-block TreeLearn estimates. The corrected estimates are what we report in deliverables. The raw calibration measurements are included in the data package.
The core data is different. We are not sampling every tree. We are sampling a subset of large-diameter individuals — those ≥ 120 cm DBH — that are candidates for prominence in advocacy and legal materials. A core from a 1,500-year western redcedar is not just a calibration measurement. It is a voucher specimen, a primary source, the only thing that makes "this tree established before the Norman Conquest" a citable fact rather than an estimate.
Age and Diameter: A Complicated Relationship
A common assumption is that bigger diameter means older tree. On BC's coast, this is broadly true but locally unreliable. Western redcedar grows faster on rich floodplain sites than on exposed ridgelines. A 200 cm cedar on a wet river terrace might be 600 years old. A 200 cm cedar on a dry rocky slope might be 1,200. A Sitka spruce of the same diameter might be 250 years on a nutrient-rich beach-fringe site.
TreeLearn uses a diameter-to-age relationship calibrated on correlated data — sites where we have both LiDAR-derived DBH and core-confirmed age for the same trees. The model is site-type sensitive. But it is still a model. The core is the measurement.
We are explicit with partners about what the numbers mean. When we say "this block contains an estimated 847 trees ≥ 100 cm DBH, of which an estimated 29% are over 500 years old," we specify the DBH values are LiDAR-derived with ±2.8% RMSE against tape measurements on a sample of 112 trees, and the age estimates are from a site-calibrated model with calibration cores from 31 large-diameter individuals in this block. That level of disclosure is not bureaucratic caution. It is what makes the number defensible when someone in the room questions it.
Why Coring Matters Beyond Calibration
An increment core is also a climate record. The annual rings in a 600-year Douglas fir contain a high-resolution record of wet years and drought years, insect defoliation events, fire events visible as narrow bands followed by release growth. Dendrochronologists can cross-date cores from a site, building a chronology that identifies specific historical events.
We do not produce full dendrochronological analyses on every scan — that is a specialist service beyond our scope. But we retain and archive all cores taken during a capture, and where partner organizations have academic relationships with dendrochronology labs, we facilitate transfers for analysis. The Elphinstone core that confirmed p'élán-ay's 1540 CE establishment date was analysed by a team at Snowline Ecological Research using exactly this pathway.
The core becomes the root of the evidentiary tree: a physical object that can be dated, cross-referenced, cited, and, if necessary, produced in a hearing room.