What it is
Site protection is the line between your seedlings and every mouth that wants to eat them.
Deer, rabbit, goat, cow, sheep. To them, a soft nursery-grown sapling is the best meal on a degraded hillside.
One bite to the growing tip costs a full year of growth. Two seasons of browsing kills the tree or locks it into a shrubby crouch it never escapes. On heavy-pressure sites, herbivory mortality runs above 70 percent.
That number ends most reforestation projects before they start.
Why it matters
The arithmetic is brutal. You can spend thousands on prep, nursery stock, planting crews, and aftercare. A flock of sheep on a spring hillside erases the lot in a week.
Deer are worse. They are selective. They walk past the grass and pick out the broadleaf hardwoods, which are the species you actually planted.
This is the single most common cause of restoration failure worldwide.
Tony Rinaudo's farmer-managed natural regeneration worked in the Sahel for one reason: farmers protected the regrowing stumps from livestock and fire first. Decades of tree-planting in the same region had failed because seedlings were browsed flat in their first dry season. The trees were never the problem.
Build the fence
Permanent post-and-wire is the gold standard for large herbivores. For deer, run it 1.8 m high, or 2.0 m where pressure is heavy. Stock fence at 1.2 m stops cattle and sheep, nothing else.
Steel post-and-wire lasts 20 to 30 years. Expensive on rocky or steep ground. Wooden post-and-rail is heavier work and costs more again.
Electric is the flexible option. A live wire at the right voltage turns deer, livestock, and some rabbits. You can install it fast, shift it as the planting expands, pull it out once the canopy is safe.
The trade is maintenance. Walk the line, clear vegetation off the wires, keep the energiser fed. Solar units are cheap enough now to make remote sites viable.
Individual guards suit small plantings. Spiral guards stop rabbit bark stripping. Tubular shelters 1.2 to 1.8 m tall block deer and warm the seedling into a faster start.
Per tree, guards are expensive. Below a few hundred trees they often beat perimeter fencing. Above that, fence the block.
Tend it
A fence with one gap is no fence. Deer find the hole in days and teach the rest of the herd.
Inspect after every storm. Repair the moment you see damage. Where deer numbers are extreme, stack defences inside the fence: scent deterrents, sacrificial planting of preferred species to pull pressure off your targets, night surveillance during the spring flush.
Rabbits need a different fence. Mesh of 30 mm or less, buried 150 mm down, turned outward at the foot so they cannot dig under. Where that is impractical, every seedling gets a spiral guard. Where rabbit numbers are very high, you cull before you plant.
When fencing meets people
In landscapes where grazing is the local economy, a fence is a political object.
Gates get left open. Wires get cut. Community engagement from day one is the only way through. Bring herders into the design. Offer alternative grazing. Improve pasture on unfenced land. Build controlled grazing into the management plan as a grass-suppression tool.
The fence holds when the neighbours want it to hold.
Letting the animals back in
Permanent exclusion is not the goal. Grazing is a natural process and most ecosystems need it.
The question is when, and how much.
Reintroduce grazing once trees are out of reach of the target herbivore. For cattle and sheep, that means a stem of 8 to 10 cm and a height above 2 m. Expect 5 to 10 years depending on species and site. For deer, taller and sturdier still.
Start low. Short pulses, light stocking, watch survival and regeneration before you push the numbers up.
Conservation grazing is a tool, not a default. Low-density cattle in a young woodland keep glades open, hold back grass competition, and feed pollinator habitat. Year-round grazing undoes the work. Seasonal pulses, short and moderate, then long rest, mimic wild herds and give the cleanest results.