Share
Share
Share
Share
Most organisations that own land also own trees. Very few manage them the way they manage a boiler or a roof, with an inspection interval, a condition grade, a replacement cost and a named owner of the decision.
That gap was never carelessness. It was a measurement problem, and closing it has taken most of two decades of quiet technical work. How it closed is instructive for anyone trying to instrument an industry that has always run on experience.
A trade that ran on judgement because judgement was cheaper
For most of its history, arboriculture operated on visual assessment. A trained practitioner walked the site, read the crown, the trunk, the root plate and the ground around it, and formed a view. The output was a professional opinion, recorded in prose.
This was not primitive. Visual assessment is genuinely diagnostic. Internal decay expresses itself externally through growth patterns, through the reaction wood a tree lays down to compensate for weakness, through fungal fruiting bodies at the base. A trained eye reads real information from those signals.
The limitation was that the information stayed in one head, at one moment. It could not be aggregated across a portfolio, compared year on year, or handed to a finance director in a form anyone could act on. An authority with thousands of street trees held thousands of opinions and no register.
What the instruments actually added
Three distinct layers of technology arrived, and each solved a different part of the problem.
Geometry at scale came from aerial LiDAR and drone photogrammetry. Canopy extent, height, position and year-on-year change can now be captured across an entire borough in a single survey. What that gives is coverage. What it cannot give is condition — a structurally compromised tree can look entirely intact from above.
Internal condition came from sonic tomography and resistance drilling. Tomography times sound pulses between sensors placed around a trunk; sound moves more slowly through degraded wood, so the timings resolve into a cross-sectional map. Resistance drilling pushes a fine probe through the stem and records how hard the wood pushes back. Both convert a suspicion into a recorded measurement.
Structure came last, in the form of GIS inventories and formal risk frameworks such as QTRA and TRAQ, which oblige an assessor to separate target occupancy, likelihood of failure and consequence, then record each one. Modelling suites including i-Tree go further, converting canopy data into estimates of stormwater interception, air quality effect and stored carbon.
The last mile did not centralise
Every one of those tools produces an input rather than a decision. A tomogram shows a pattern of sound velocity. It does not say whether the tree stays.
The instruments centralised readily — a borough commissions a canopy survey once and holds the result as a dataset. Interpretation and remedial work never centralised at all. A Greater Manchester authority holding a completed inventory still depends on the kind of independent tree surgeon Stockport and its neighbouring districts have used for decades to turn a flagged record into a site visit, a judgement and, where the judgement goes that way, a climb.
That division is durable. Reading a measurement against species, site history, soil, target and season is contextual work, and the context is local.
Where the data layer outruns the decision layer
The honest position is that instrumentation has been oversold, and in a specific way worth naming.
Inventories are routinely commissioned and then never operationalised. A survey is a discrete capital project with a procurement route. Acting on it is recurring revenue expenditure with no natural sponsor. Those two things sit in different budgets and often answer to different people, which is why datasets age quietly.
Decay detection is over-claimed too. A tomogram is a single cross-section at a single height, inferred indirectly from sound behaviour. It can misread compartmentalised decay, radial cracks and moisture variation. It narrows uncertainty; it does not eliminate it, and a report presented as definitive is a report that has stopped being useful.
Modelled valuation has the opposite problem. Ecosystem-service figures are defensible as advocacy and awkward as accounting. No budget holder can spend an air-quality estimate, so the number persuades without ever quite deciding anything.
Taken together, the return on instrumenting this asset class is genuinely harder to demonstrate than vendors suggest. The survey cost is immediate and visible. The avoided failure is diffuse, delayed and invisible by definition.
What generalises
The sequence is not specific to trees. Any sector that ran on tacit expertise meets the same pattern: instruments arrive, produce far more information than the existing decision process was built to absorb, and the constraint migrates from knowing to acting.
The organisations that extract value from a canopy survey are not the ones with the best sensors. They are the ones that settled, before commissioning anything, who would read the output, what condition grade would trigger a visit, and which budget the visit would come from. That decision costs nothing. It is made far less often than the survey.
