Week 7

Fencing, signage and wildlife crossings

The Best Result In The Literature Is Also The Hardest One To Fund

Four mechanisms

  1. exclude — fencing
  2. redirect — a safer crossing (underpass, culvert)
  3. change driver behaviour — signage, dynamic warnings
  4. change road-use conditions — sightlines, lighting, speed

Most measures you’ll compare today combine 1 and 2.

Fencing works until it doesn’t

A fence that ends doesn’t remove animals from the area — it relocates where they cross: to the end of the fence.

This is a real displacement effect, and a source of measurement noise in the studies you’re about to read.

Usage ≠ collisions

A well-used underpass is evidence the structure works as a crossing.

It is not, on its own, evidence that road collisions went down.

Two different measurements. Don’t recommend on the wrong one.

What the evidence says

83% reduction≈1% reductionFencing + crossing structuresRoadside reflectors

Rytwinski et al. (2016), PLoS ONE — a meta-analysis, not a single study.

Comparing candidates

MeasureEvidenceMaintenanceApproval
Fencing + crossingStrongHighHigh
Crossing aloneEvidence-limitedModerateModerate
Static signageWeakLowLow
Dynamic warningUntested hereModerateModerate
ReflectorsWeakLowLow

Evidence strength and approval complexity run in opposite directions.

Illustrative costs — fictional teaching figures

MeasureUpfront (AUD/km)Annual maintenance
Fencing + crossing$180,000$6,000
Crossing alone$140,000$4,000
Static signage$2,000$200
Dynamic warning$25,000$3,000
Reflectors$4,000$500

Rounded, illustrative, not a real quote.

Why “install everything” isn’t a recommendation

  • competes for the same limited budget
  • competes for the same finite approval bandwidth
  • a list of every measure is a refusal to choose

One primary measure, argued for one site. That’s the shape Week 12 will ask for again, at the scale of the whole proposal.

Three sites today

Bunjalung Ridge Road — 15 km, low traffic, small council budget.

Coorabin Road — 38 km, moderate traffic, two known clusters.

Merricourt Highway — 51 km, high traffic, slow state approval pipeline.

Same evidence table. Different answer at each site.

Next week

Whatever gets recommended, someone has to approve paying for it — and if a hazard was known and nothing was done, someone has to answer for that too.

Week 8 starts exactly where a sign, a fence, or the absence of either, turns into a question of who is liable.