fun stuff
contact us
Time spent in the pursuit of Leisure
Priority solutions for Leisure-time!
Custom Search

Building a Stream using Liner and Clay (Bentonite)


For quite a while now I have been meaning to make this page to discuss an interesting technique used by someone who contacted me for my comments on filtering for his wet-look Zen-garden style of stream. His name is Steve and he lives in South Africa, and this page is composed from various fragments of our email conversation.

The stream that he built was just that; Simply a stream, no pond, and no fish!

What I found interesting though was Steve's approach to building his project, and I believe this could be applicable to many other folk interested in building a stream for their pond. Below is a photo of the finished Zen stream. The water appears muddy because the picture was taken very shortly after completing the construction. It's not the water-chute that I want you to focus on. Pay attention further "upstream" to the nice levels of the rocks and pools of the stream bed.

A Zen-style garden stream

Jims Pond Blog

To get the latest news on my ponding bio-filter and venturi experiments why not visit Jim's Pond Blog. You can subscribe to my blog's RSS Feed powered by Feedburner to ensure you get the latest updates. It will work with most Atom and RSS 2.0 compatible news reader software, such as Bloglines, Desktop Sidebar, NewsGator, MyYahoo, etc.


For great fish-keeping communities visit my favourite forums at:-

They are very friendly and knowledgeable groups of people who will make you feel very welcome. There is tons of discussion going on about fish of all kinds, problems whether relating to the health of your fish or the state of your pond, and advice on filters, pumps and anything else you can think of!

I highly recommend Bradshaws Direct (UK) for all your ponding supplies:-

Sick of SPAM? This is what I use every day - It's very good.....

Read my review about Mailwasher Pro Anti Spam Software

How To Build A Zen-style Garden Stream

The best thing is just to let you hear from Steve the ideas he had for making the stream, and how he did it, and since this was an email conversation over the course of a few weeks, my responses are included as well:-


I have just completed (2006) a "natural" stream in my garden here in Johannesburg, South Africa - natural in the sense that despite being 7m long and with a fall 0.75m, it has no concrete. Now that it is "working" after 7 months of on-off weekend work, I'm beginning the fine tuning phase.

The project is not yet complete and the plants are still very small. Much of the 'workings' are therefore still open.

I don't have a pond. Instead I have a closed weir / filter at the end of the stream into which the water flows (the black tank). I may install a pond later but ponds are a lot of extra work (as you know!!) and this is a new venture for me.

Weir Filter Fall
Early construction of the weir/water chute.

After the weir, the water overflows into a large green tank where the pump is situated.

Green Reservoir Tank
The Green Tank is the main water reservoir.

The design is more a dry zen garden look. Water flows for one minute every two hours and irrigates the fringe plants and fills an area (only 1.5 cm deep) that birds can drink from.

The pump can be turned on manually if you want to sit and enjoy the sound of the water from the deck.

The layers are as follows:-

  1. Plastic liner at the bottom
  2. A mixture of clay and small pebbles giving a smooth foundation that runs downhill without terraces
  3. A layer of stones, pebbles and rocks giving the completed look with levels and terraces.

The absence of concrete means that any element / level of the feature can be changed.

Stream Bed Layers and Pond Liner
Early layers of the stream bed

The issue now is this. Because of its construction my water is not yet clear - I don't want it to be totally clear BUT it needs to be clearer than it is now. I do have a "filter design" in the weir (black tank) but it is not working efficiently yet.

Because the stream bed is open to direct sun light and temperatures of 30 deg C I don't think I will get the same problems that a deep, dark, still pond will get. But I do need to filter out the very, very fine particles of light brown clay in the water.

Any suggestions?

Construction Phase
The yard before starting the project


Thanks for the photos Steve. That looks like a great project you've been working on. I really like the levels of the rocks (some nice sizes and proportions) and the intermediate pool.

When we first finished our own stream and pond, we also had problems with fine water borne clay, dust, silt and stone particles. This should just settle out after a week or two. One contributory factor is that as algae slowly forms a film on the stream bed and stonework (anywhere that water is in contact), this will gradually act as a natural sticky filter to trap all those very fine particles. Also a lot will settle out in the bottom of your chamber at the bottom of the stream.

It sounds like you already have some type of filter media in your weir/pump chamber? This will help to trap particles and clean the water. It will also provide a home for algae, and a place for bacteria to develop, although it will not have much work to do in clearing ammonia (nitrification cycle) because you do not have any fish to produce any excrement for bacteria to work on.

I am sure that your stream will quickly build up a layer of algae because of the high temperature and shallow water, and certainly that direct sun will contribute to this. However, you only notice "green water" when there is a degree of depth to look into. That is to say, that even if you had very green water, since your project is simply a pleasurable stream to look at and enjoy, there is no depth of water and you are unlikely to notice any water-borne algae (it should always look clear). Of course where fish ponds are concerned, it is this "green pea-soup" that greatly annoys fish keepers because they cannot see and admire their fish!

String algae will surely develop, but I think this will contribute to the natural look of the stream.

Why do you want the stream running for only 1 minute every hour? I presume for the Zen look? Which I suppose is a damp, almost mossy look, rather than a full blown stream. I just wonder whether turning the water on and off all the time will make it take longer to settle, mature and become "natural" looking? Starting and stopping it may cause the silt to take longer to clear. I don't know for sure in your special case, because I've never seen anything quite like your project, so just thinking out aloud! If you keep the water running all the time it will help oxygenate the water, which will help the natural cleaning cycle of the stream and weir chamber. It will also prevent any stagnant water build-up, or flocculation film on the water surface.

One thing we find is that birds love to come and wash in our stream here in the UK, particularly blackbirds and wagtails which nest in our garden, and thoroughly enjoy the running water. Again a good reason for leaving the water running all the time.

I also wonder what effect the constant dampening and drying out will have? Nature is tremendously adaptable, and you might find that it attracts unusual creatures to your stream.

Hmm, I'm also wondering whether mosquitoes and gnats might like a primarily static section of water!

I am sure that in a years seasoning your plants will have established themselves and matured and the stream will look lovely. And in a short time you will decide what is best. It's your stream, and how you make it work will be an interesting study. I suppose being Zen-like you will keep night-lighting very subtle and low-key?

I would love to see how it progresses and looks in time, and I would be interested to know if my comments have been helpful.

Construction Phase
Initial digging work


Thanks for your input which I am in the processing of carefully considering. I have never done this before and thus every decision is mulled over for a long time but despite this, trial and error seems to be unavoidable. I will send you some photos for comparision at the end of our growing season, i.e. April, for comparison. These will either support or refute the theories as discussed.

The internal design of the filter (which I did not provide) has been a complex process as I had to build it into an existing drum and thus every part had to be handmade (using mainly perspex - it does not rot or float). I currently have a problem with it making a ghastly sucking noise when the flow rate is set at a specific level (it runs fine on max and min, but not so midrange). This problem is as a result of a poor position choice for the exit outlet going into the main tank below.

Jim: I wonder if either a larger bore pipe will help? Or simply a piece of wire across the entrance of the outlet to disrupt the flow, therefore breaking up any vortex causing the noise. I expect it is a small whirlpool making the sound by drawing air down with the water.

Stream Construction Phase
Laying sand layer to protect the pond liner from punctures

Stream bed laid
Additional contouring of the stream bed, and rockery progressing nicely - complete with cat!



From the photos you can see that the entire area has become a natural catchment area. In Johannesburg we experience cloud burst thunderstorms that can cause sedate streams to turn it raging torrents that wash away cars in less than 30 minutes (that's what happenes when you cover thousands of hectares with concrete and asphalt). Thus an overflow system
for the tanks still remains on the to-do list. Going out into the lightening, thunder and rain to pump the main tank half empty (so that it can overflow again 10 mintues later) is no fun! (the large green tank is 500 liters).

Stream bed
Stream bed leading up to the weir/water chute into the filter

The on/off interval is definitely a matter for debate, and again trial and error based on comments like yours will have to be the order of the day. Yes, the Zen look is definitely dry, with the colours of the stones and sand being a very important element. Johannesburg is very dry 7 months of the year and thus a 'dry look' is more in keeping with our climate. On the
other end of the climate scale our summer offers very high evaporation rates.

Jim: Yes, I was wondering whether the choice of running the stream all the time, or on/off would also affect the rate of evaporation, i.e. if it were kept running all the time it may evaporate less quickly because the moving water maintains a more constant (cooler) temperature, whereas if the water stops frequently, where the sun is shining on it, the temperature would rise rapidly and make for faster evaporation?

Zen Garden style stream
View of the Terraced Stream Bed with Rocks and Pebbles

Path and Decking
View of lower end of stream, filter and decking area.

Decking Area
Looking down stream toward decking

Five Months Later.....

Some time afterwards I sent another email to Steve asking how his stream had matured......

  • How's it doing?
  • How was the clay bed structure in the end?
  • Has it stood the test of time?
  • Or have the storms washed everything away?
  • How long did it take the clay to clear?
  • What sort of problems have you had along the way?

At this time I do not have any more up to date photographs, as I am waiting for Steve to send me some new ones.


Jim, I am more than pleased with the outcome.

Water spilling from Stream into the Filters
Reinforced clay/gravel/cement outlet - Water flows into the Filter
(Notice how careful consideration for the levels of the rocks has created shallow pools further up stream)

The area is fast becoming lush and overgrown resulting in a profusion of wildlife, i.e. insects, birds, lizards, frogs (and the neighbour's cat who now lives under the wooden decking !). This is despite us living in the city, in a housing complex.

The pump only works one minute in every hour, and for 20 minutes once in the morning (there is a manual override however).

The type of clay we used was bentonite, which we bought from a chemical company.

The compacted clay and pebble stream bed structure is totally stable, with two exceptions ....

  • First, the 7 places where the water enters the stream. There I have to use a large flat rock, then large pebbles followed by small pebbles to resist the attrition from the water coming out of the pipe. This is not a problem because it looks nice anyway.
  • Also the 'waterfall/weir' area (the water drops straight down about 15 cm). The water falls onto a mixture of clay, small pebbles and cement (which I did not allow to dry). This made the clay harder in this area which resists the attrition of the water very successfully.

Storms have not been a problem despite us having a very wet summer. Even the biggest storm is 'small' compared to running the pump for 20 minutes. The stream bed shows no signs of erosion. This is borne out by the fact that the filter fills up with sand and clay particles very slowly.

Two of the things I really like about the clay bed is the fact that it irrigates the surrounding plants automatically. Also, seeds that land on the clay bed germinate by themselves. The river bed is fast becoming a natural mixture of deliberate and natural planting. This is also causing the stream to slow down - although the pump runs for 1 minute every hour,
the stream actually flows for a further 15 minutes after the pump has turned off.

The stream bed and puddle areas are not smelly, nor are they covered with excessive unsightly algae. Yes, they do have algae but just enough to look natural. Breeding mosquitoes is also not a problem.

We lose 50 litres of water per day to summer evaporation, and keeping the thick clay/soil base wet (about 15cm thick). This may sound a lot but bear in mind that the water is being used to irrigate as well. Fifty litres to keep a "swamp area" of 8m x 4m wet for 24 hours a day is not a lot. It will be considerably less during the winter.

Getting the water clear took about two weeks - but also took the creation of a homemade filter - see attached two files. I didn't give you a copy last time we chatted as I was not sure how well it was going to work. Well it works very well and the water is crystal clear, even after rain.

Fileter Design 1

The timer settings and the overall design of the water feature results in the water flowing very slowly through the filter 23 hours of the day. The agitated water is contained in the first compartment allowing maximum settling in the next two.

The 3-compartment manifold design serves to make the water meander through the filter, the openings being left and right, down and up. This helps the particles to settle out. Once settled, they stay settled, due mainly to the honeycomb element at the bottom of the middle compartment.

Filter Design 2
Filter Design (click for full-size image)

The stream does not feature an open pond. This is important as we are expecting a baby and an open pond is a worry (and a technical demand) we don't want at this stage. The 170 litre tank with lid is an off-the-shelf product here in South Africa. Items like stones, leaves, twigs, etc. don't even enter the filter as they are trapped among the pebbles lying on
the perforated filter cover. This means that keeping it clear and clean is very easy with a total clean out of the whole filter being necessary only every 4/5 months.

The final compartment is filled with Koier (spelling?) - this is a hard, dry mass of grass-like natural fibres that can be bought by weight. Water passes easily through it while acting as a basic pre-filter. The same can be said of the foam rubber element that stands in front of the opening to the third compartment.

Problems we encountered:-

  1. Water leaving via the downhill pipe into the final large reservoir caused a vacuum and an awful sucking noise in the filter. This was solved by drilling a small hole in the pipe allowing air to break the vacuum.
  2. Flow controls (ball valves) at the end of the 7 feeder pipes into the stream caused an awful hissing noise as the water was compressed on exit. They were moved further back and away from the openings.
  3. The water falling into the final tank was too noisy despite the tank being buried. I had to create a special drop-pipe that dissipated the pressure of the falling water so as not to disturb the stored, stagnant water as much.
  4. I buried my final tank too low. Thus I don't have a good overflow solution should it get too full. Probably never will without the addition of another pump.
  5. Submerged plastic tanks can crack if they are allowed to get too empty (due to lateral pressure from water-logged soil). This means that the tank must be kept at least 50% full at all times. This is easily managed by the pump's built-in float valve that turns the pump off when the level gets too low BUT . . . it means that you only have 50% of the capacity of the tank to work with (before you have to top it up again). At some point I am going to have to automate the topping up of the tank from the nearby tap.
  6. I could not get a tank with an opening big enough to take my chosen submersible pump. By cutting the top opening bigger I brought on all sorts of problems that I had to solve later on, e.g. the tank was no longer a sealed component which can potentially allow pebbles and sand to fall in.

Feel free to pass the info on (I made a lot of mistakes that were expensive to fix). Yes, anyone can email me. I have documented its creation which I can send on to interested parties.


I hope you will agree that Steve has produced a very interesting look to his stream. The primary consideration before applying this build-method to a proper pond environment, and most important is avoiding the use of cement at all within the clay bed, since lime in the cement would probably cause a toxic problem for any fish.

Remember that Steve's stream was not intended for fish, so using cement in one part of it did not matter.

However bear in mind that the only place cement is used, is in the weir/water chute, where the water is particularly rapid, and a light cement mix was required to make the surface hard enough to prevent erosion of the materials into the water filter. Regardless, with aging, this does not appear to have affected the wildlife that Steve reports making the stream their habitat.

Provided the stream is built without such a rapid flow area, or the use instead of slate or rocks at the edge of any waterfall feature, all the other constructional methods should make for a very practical and "living" stream, in the sense that it can be continually manipulated to improve pooling, terracing and interesting flow areas, and I am certain can be used safely with a pond with fish.

To see additional pictures of Steve's project now it has matured go to his website at Garden Stream Workshops:-


In particular see the "two years on" photo gallery at:-


Useful Links:

Wikipedia Bentonite reference

http://www.mkm.co.uk/ - MKM Clays & Absorbents Group is a leading supplier of absorbent clays to the UK and Europe.

http://www.akwaseal.com/ - Akwaseal® Pond Liner Roll is made with two water barriers: Bentonite Clay and a Flexible Plastic Membrane

If you have any comments or suggestions about this project please contact me:

And if you have found these web pages useful and interesting and would like to make a donation, you may do so using the Donations Page.

Terms of Use |  Privacy Policy
Copyright © 2016 Prior IT Solutions - All rights reserved