This update shares a practical rebuild of a Gbiota bed after I noticed uneven water spread between rows. The core lesson is simple: the liner shape matters more than expected. A “saucer” profile gives too little contact area for wicking when water is only trickling through the pipe. A flatter base with a small lip dramatically improves lateral wetting. I also share two clear checks you can do before building: use stable compost in the base and test your slope using water, not your eye.
Why I’m rebuilding
I had some problems with my Gbiota beds and decided to rebuild one line so I can compare designs under similar conditions. I drafted an earlier version of this update to warn members about poor water movement across rows. Since then, a small change to the liner profile has made a substantial improvement, and it is important you hear this before you start digging.
The good news is the soil biology looks strong. I have never seen so many worms in a garden bed, which suggests we are heading in the right direction biologically. Some plants have flourished—especially the more aggressive, water-hunting types—while others have struggled.
In my beds, plants like Purple Amaranth, Kang Kong, Spinach, Okra, and Comfrey have done very well. Others, like lettuce and radish, went to seed quickly, and Chinese cabbages were decimated by insects. These can be normal Queensland summer issues, but the symptom that really worried me was poor germination and weak growth between rows. That pointed to a basic engineering problem: poor water distribution across the bed width.
What I was trying to achieve
Wicking beds are excellent for small areas. They are a reliable, well-tested system for home production. But the scale of chronic disease, especially diabetes, is enormous. The intention here is to produce food that is rich in minerals and phytonutrients, with active plant-associated biology that can help improve human gut biology. To do that at scale, the growing system must be low cost, simple, and able to operate with automation.
The idea was to use multiple rows of modified open wicking beds, potentially up to 100 metres long, supplied by an external reservoir. Water delivery would be automatic using a pulsed cycle, and the same loop would allow compost tea (with mineral additions) to be introduced to the root zone.
In theory the system is straightforward: create a gentle slope, dig a narrow channel, line it with plastic film, lay in a perforated pipe, and connect it to a reservoir and return path. Add a pump, a manifold, and a drainage system so excess water returns to the reservoir. Then provide a method to introduce compost tea into that loop.
Practical realities: slope, soil, and water flow
For pumping, I wanted to use a pond pump because they are cheap and have an open impeller, so they can tolerate a fair amount of dirt. They do not have much head—about a metre or so—but that is adequate for this type of distribution. The most convenient placement was near my shed where power is available. A commercial system would likely use solar, which is generally sufficient.
In my case, the shed location forced me to reverse the slope direction compared to what I first wanted. I do not have hard experimental data on the ideal gradient, but from flood irrigation experience I guessed a slope of about 1 in 100 would be fine. In a 7 metre bed, that is around 70 mm of fall.
I also needed soil to create that gradient. My existing soil is poor—essentially a silt layer sitting above clay—and I wanted to inoculate the system with a very active biology. That led me to bring in biologically active material for earthworks and soil building.
Two essential build lessons
After watching what happened over time, two practical lessons became very clear:
- Use well-composted material in the base. Fresh, labile compost settles, and its open structure does not wick well. Wicking depends on a finer, compacted structure with small pores.
- Measure your slope with water, not your eye. If you do not have laser levelling, lay out plastic and do a water-flow test. A small gradient is difficult to judge visually, but water will tell you immediately whether the slope is uniform.
The world’s most effective soil moisture monitor
In a previous life I ran a company making electronic soil moisture monitors. These days I use a simple wood auger I bought at the market for $8 because it gives me the information I actually need.
I could see the surface soil between rows was dry, but the auger showed moisture underneath. To test properly, I planted radish seeds across the rows, gave one watering, and waited. The result was unambiguous: the radish grew well over and near the pipes, but between the pipes there was almost nothing. The problem was not “maybe.” It was water distribution.
I explored the obvious suspects. I wondered if the soil was not wicking laterally, so I trialled a strip of rotted grass clippings in one place and a cotton cloth in another to improve capillary continuity. Those spots looked moist, but overall the improvement was not enough to solve the pattern.
I also tried a second pipe above the ag pipe, like a homemade dripper tube. Water use increased, but there was no strong sign it improved lateral wetting across the bed. The patches were starting to look like “patch after patch,” and that is often a warning sign: the basic geometry might be wrong.
Finding the real cause
I eventually accepted I needed to pull the bed apart and watch what was actually happening. Replacing liners is hard work—shovelling soil out and back in again—so I do not say this lightly. But it was monsoon season, and digging in wet conditions made it easier to see how water behaved.
Once I exposed the pipe and watched the flow, the mistake became obvious. Water flowing down the pipe was only a trickle—about a litre per minute—so the water depth in the pipe was only around a millimetre. Wicking is not a strong force. It needs the right conditions, and it needs contact area.
I had shaped the liner channel like a saucer and placed the ag pipe at the bottom. That meant there was only a tiny contact area between the soil and the water. With such a small contact strip, not enough water could move out of the pipe and into the surrounding soil to wet laterally.
The fix: a flatter base and a small lip
The solution was to change the channel profile. Instead of a saucer, I formed a flatter base with only a shallow lip at the edges. That creates a meaningful flat area where water can spread, increasing the soil-to-water contact zone. The difference was dramatic.
This change worked so well that I was tempted to trial a simple flat sheet on another bed. It felt like a silly idea at first, but sometimes a “simple” idea is exactly what the physics needs. In this case, it worked.
From here, the next refinement is to reduce the lip as much as possible while still preventing leakage. The goal is enough containment to guide flow, but not so much shaping that you lose effective contact area.
Following the water: a quick system walk-through
It helps to follow the water path as a complete loop. Water starts in the main tank, which catches rainwater from the shed. In Bundaberg, we either have too much rain or too little, so I use a float valve to top up when needed. That also gives me a simple way to estimate how much water I am using.
From there, water goes into a feeder tank (in my case about 40 litres). It is small, but it does the job. The pond pump sits in this feeder tank. Water then travels through a pipe to the distribution point at the top of the block, where a tap adjusts flow.
The flow is set so the header tank does not empty during a 10-minute irrigation cycle. That gives roughly 40 litres, plus whatever comes in from the main tank during the run. In my layout, that is just over 1 mm of water per cycle—if the soil is already wet, the excess simply returns to the header tank.
I irrigate on a pulse cycle (every two hours in this setup), which gives capacity for around 12 mm per day. Evaporation can be around 10 mm per day in hot conditions. Vegetable crop factors are generally under 1, but I also have trees pulling extra water, so overall the net demand can be near that 1:1 range. In practice, this level of delivery should be adequate, provided the bed actually distributes the water laterally—hence the importance of the liner geometry.
Key dimensions and planting test
In the revised bed, each lined channel is around 200 mm wide, and the plastic liner width is about 270 mm. The base of the channel sits roughly 150 mm below the soil surface. I would prefer deeper pipes, but depth is constrained by the header channel level.
To confirm wetting, I plant across the rows. This is a simple visual test: if seedlings are uniform across the bed width, water distribution is working. If growth repeats in stripes, distribution is still failing somewhere.
Compost tea flushing and next improvements
I am still using a regular compost bin for household compost and flushing manually at the moment. The intention is to automate compost tea flushing so the biology and nutrients can be delivered consistently through the root zone. Automation is the goal, but it needs to be built on top of a distribution system that works reliably first.
If you are about to build beds, especially in wet season, watch for updates and do not hesitate to check your slope and liner profile before you commit to filling and planting. A small geometry change now can save a huge amount of labour later.
