Sand Backfill Over Clay: The Perched Water Table You Built Yourself

Ninety of 400 trees died in a retail car park, and every one of them was in a pit that had been built exactly to the landscape architect’s detail: a 900 mm augered hole in stiff clay, 300 mm of washed sand at the base for drainage, a 45-litre root ball, then sand backfill with 60 g of gel mixed through. The detail is common. It is also, hydraulically, a bathtub.

The reasoning behind the detail is that sand drains. It does, but only if the water has somewhere to go. Sand sitting on clay is not a drain; it is a sponge on a saucer, and adding gel to the sponge makes the saucer matter more.

What happens at the interface

Water entering sand moves down under gravity until it reaches the clay. Clay has a saturated hydraulic conductivity several orders of magnitude lower than sand, so the water stacks up. A 300 mm layer of sand above a clay base can hold roughly 90 to 110 mm of water at saturation before it starts to drain laterally, and in a pit with vertical clay walls there is no lateral path either. The result is a saturated sand layer 300 mm deep that persists for days after rain and, in a wet winter, for weeks.

Roots need oxygen. In saturated sand at 15 °C, oxygen is depleted within 24 to 48 hours in the root ball, and fine roots begin to die within three to five days. The tree does not drown in the sense of being underwater; it suffocates in a saturated, airless cylinder.

The same pits performed beautifully in the first summer, which was dry. The gel in the sand held water, the sand drained fast enough in the surface layer, and survival was 98 percent. The failures came in the second winter, when 140 mm of rain fell in nine days. That pattern is the reason this fault survives in specifications: it looks like a success for one full season.

There is a second, subtler effect. The gel swells at the sand-clay interface, where water accumulates, and blocks the remaining pore space. A sand that infiltrates at 200 mm per hour clean can drop below 20 mm per hour once the gel at the interface is fully swollen. The pit seals itself from the inside.

Four ways to build it properly

For most street and car park trees, the second option is the cheapest fix: stop building a sand layer in a clay pit. If the specification calls for drainage, the drainage has to lead somewhere, and a 150 mm slot of gravel to a subsoil drain is drainage while a 300 mm sand layer is storage.

The dose question nobody asks

On a site with a perched water table, the gel is not short of water; it is holding water in the wrong place. Reducing the dose is more useful than increasing it. On the retrofit of the car park above, the pits that were rebuilt with clay backfill and 30 g of gel had 94 percent survival over the following two winters, against 78 percent for the untreated survivors in the original detail.

The counterintuitive part is worth stating plainly: on this site, and on three others like it, gel made the wet-season mortality worse and the dry-season mortality better. The net was negative because the wet season killed in weeks and the dry season killed in months, and the irrigation programme covered the dry season anyway. If your site has any risk of a perched water table, the gel rate should be set by drainage, not by rainfall deficit.

A five-minute test before you plant

Dig a hole to the proposed pit depth, fill it with water, and measure the drop after 30 minutes and after 24 hours. If the water is still standing at 24 hours, the pit will perch, and no gel rate will fix it. If it drains within an hour, the sand layer is redundant. That test costs a bucket of water and it would have saved 90 trees.

Write the result into the planting method statement next to the dose, and photograph the hole with the water in it. The infiltration figures and mixing guidance for our soil-applied water retaining polymer assume a draining profile, and the technical sheet says so explicitly.