A paludarium land section rebuild: what to change the second time

Fourteen months after the first build, the land section had dropped 3 cm, the water line had crept up the glass, and the fern roots were sitting in a soggy mat that smelled of pond bottom. The tank had looked superb for eight months. The second build took a weekend and has now run for two years without the same failure.

Paludariums fail at the boundary. The water side is straightforward and the land side is straightforward, but the line between them is where water does what it wants, and it wants to move from wet to dry until everything is wet. Gel in the land section makes that movement slower, which is helpful right up to the point where it hides the problem.

What went wrong the first time

The land section was built as a mound of coir and bark resting on a false bottom, with the water level set 2 cm below the top of the mound. Water wicked upward through the substrate continuously. The gel held it, so the surface stayed damp and looked correct, while the base of the mound was permanently saturated.

That saturation did three things over a year. It collapsed the structure, so the mound sank. It killed the roots in the lower half, which then rotted and released nutrients. And it kept the water side at a stable high level, because the mound was acting as a wick feeding water back into the tank. The whole system was one continuous wet body with a dry crust on top.

The barrier is the whole design

The second build started with a physical separation, not a substrate decision. A sheet of corrugated plastic, cut to the shape of the land section and sealed at the glass with silicone, runs from the false bottom to 4 cm above the intended water line. The land section sits entirely behind it.

That single change means the land section can only get water from above. It gets misted, it drains down to the false bottom through a small weep hole at the lowest corner, and it never sits in the tank’s water. The weep hole is 6 mm, positioned at the base, and packed with coarse gravel so it does not clog with fines.

Gel placement on a slope

With the barrier in place, the gel has a clear job: hold moisture in the root zone and release it slowly upward. On a slope, that means dosing by height. The top of the mound dries fastest and gets 5 percent. The middle gets 4. The base, which stays damp from drainage, gets 2 or none at all.

Mixing one batch for the whole mound is what most people do, and it produces a wet base and a dry top within six weeks. Three small batches and a layered build take an extra forty minutes and solve it permanently.

• Base layer: coarse gravel and bark, no gel, drains freely
• Lower mound: 2 percent gel, packed moderately
• Upper mound: 5 percent gel, packed lightly so roots can move
• Surface: 2 cm leaf litter, kept dry, replaced twice a year

The rebuild checklist

Drain the tank fully and let the glass dry before you cut anything, because silicone will not bond to a damp surface. Test the barrier with a jug of water before the substrate goes in: pour 2 litres behind it and watch where it comes out. If water appears on the tank side anywhere other than the weep hole, the seal is wrong and you will find out again in eight months.

Plant before you fill. Roots placed into a finished mound are roots that get pushed into a hole that is too small, and those are the ones that rot. Build the mound around the root ball in stages, firming gently, and water each stage as you go so the gel swells in place rather than after the plant is settled.

The detail I would change a third time is the weep hole size. Six millimetres clogs with fines within a year in a coir mix. Twelve millimetres with a gravel collar would drain better and still keep the substrate out. If you are building now, go larger than you think you need.

For the humidity side of the same build, the notes on SAP powder grades explain the difference between a fast-swell and a slow-release product, and the slow-release grade is the better choice for a mound that you want to stay stable rather than cycle.

Before you rebuild, take the old land section apart slowly and photograph the layers as you go. The wet line in that cross-section tells you exactly where your barrier failed, and it is the only drawing you will ever need.