Field note · 15 November 2026
Designing a survey grid before you buy a single lure
Most disappointing trap programmes were mis-designed before anything was ordered. Three decisions fix the grid, and all three come before the purchase order.
A pheromone trap does not sample a forest. It samples a small volume of air around one point, and everything you conclude afterwards rests on where that point was. The uncomfortable part is that grid design is usually decided by whoever is holding the map, in an afternoon, and then inherited for a decade.
Decide what the grid must detect
There are three different jobs, and they need different grids:
- Early detection — is the insect here at all? Sensitivity matters, coverage does not. Few traps, carefully placed on the highest-risk hosts and edges.
- Population trend — is it going up or down? Consistency matters. Same trees, same height, same spacing, same servicing interval, every year, without exception.
- Delimitation — where is the boundary? Density matters. Many traps, tight spacing, and a defined radius around a known find.
Most programmes try to do all three with one grid and end up doing none of them well. If the budget only supports one, choose trend: a comparable series is worth more than a wide scatter of unrelated points.
Fix the spatial unit before the trap count
Traps should map onto the unit you actually make decisions about — a stand, a block, a management compartment, a quarantine zone. Three traps in a stand give you a stand-level number; one trap in each of three stands gives you three numbers you cannot compare with confidence.
The practical rule used in a lot of forest programmes is three traps per stand, positioned to represent the stand rather than the road: one on the exposed edge, one inside, one on the downwind side. That is not statistics, it is insurance against a single trap failing or being vandalised, which happens more often than most protocols admit.
Spacing, height, orientation
Spacing is usually 100–200 m along a trap line for bark beetles and budworms; closer than that and neighbouring traps compete for the same beetles, which flattens the relationship between catch and population without ever announcing itself. If doubling trap density does not roughly double total catch, you were already sampling the same insects twice.
- Height: bark beetles 1.5–3 m at the host stem; budworm and tortricids 8–12 m in the live crown; clearwings 0.5–2 m at the trunk; longhorn beetles in the canopy of host trees.
- Orientation: on the sun-exposed edge for bark beetles; facing into the prevailing wind for most moths.
- Anchoring: a trap that swings in the wind releases differently from a rigid one. Hang from a rigid arm, not a branch tip.
A worked example
Four hundred hectares of spruce, split into eight stands averaging fifty hectares. Three traps per stand gives twenty-four traps. At one hundred and fifty metres along the edge, that is roughly one trap per kilometre of edge — comfortable. If the same twenty-four traps were spread evenly by area, each would cover sixteen hectares and sit nowhere in particular.
Now add the servicing arithmetic: twenty-four traps checked weekly for fourteen weeks is 336 visits. Two people, half a day per round, is 28 person-days. That number, not the lure price, is what kills programmes in September.
Where traps should not go
At the base of a slope where cold air pools, inside dense unthinned stands where flight is restricted, beside a road where dust coats the sticky surface, and anywhere you cannot reach in a wet week. A trapping point you skip twice in a season is a data gap, not a data point.
Write it down
The grid is an instrument, so it needs a specification: trap type, lure, lot, height, tree, coordinates, deployment date, servicing interval, and the rule for what happens when a trap is lost. Two seasons of that record outranks any single season’s peak count, and it is the only thing that makes a change of supplier or a change of staff survivable.