We have now been running our aquaponics system for a year. It was installed as part of a PhD project emerging from Sheffield Hallam University, exploring how a lay-person would take to soil-free gardening, and capturing lessons that may prove of use if the approach is to be taken up more widely.
Last September we posted our lessons from the initial design and cycling of the aquaponics system, and here is the experience gained from using it in earnest.
- Temperature: we are still waiting for the live temperature monitoring system planned by the University, but temperatures in the tank got down to 5C in this uncharacteristically long, cold winter. We keep carp in our system in part due to their ability to cope in a range of temperatures, and whilst their activity seemed to slow, there was no significant problem.
- Temperature time-lag: the salads planted in the spring were a lot faster growing in the soil than aquaponics. Without monitoring data it is hard to specify the cause but it is likely that problems with the auto-siphon (see 8) slower supply of feed and oxygen, and we suspect that the temperature of the water was slower to rise than that of the soil.
- Over-winter plants: our main problem over winter was vermin rather than temperatures. The kohlrabi were eaten in one night. But other plants thrived, and we had a supply of green leaves from Mizuna, chard and spinach plants through til Spring.
- Thermal mass: Another success were the tomatoes which survived 6 weeks longer than those a meter away that were planted in soil. This may be due to their later planting, but their close proximity to heat-storing thermal mass may also play a part. The mass is in the clay balls in which they are planted, the water, and the dense concrete blocks that support our growing beds – against which the tomatoes were trailed.
- Minerals: We share the produce grown between our five households, and concern was raised about whether they were as healthy as soil-grown plants due to a lack of minerals. This prompted some further reading but we were on the right track:
- Maxicrop to add micronutrients in our initial cycling
- Rocks from our carp pond helped introduce naturally-occuring bacteria
- To this we added, after three months, composting worms which seem to be surviving well, and there is no sign of mineral deficiency in the leaves or crops.
- Sowing medium: some seeds have self-sown (tomatoes), and some peas and beans have successfully grown from seed scattered in the beds, but vermin were a significant problem – perhaps due to the long winter. To get round this we started seedlings off in an Ikea hydroponics kit in a conservatory, made more sustainable by using our own sheeps’ wool to form the sowing medium, instead of mineral wool.
- Evaporation: again, we are yet to receive the monitoring equipment but it seems more water is evaporating than we would expect, with the system requiring approximately 300 litres top-up every 2-3 weeks in growing season.
- Cleaning: it took a long time to select fish food and it’s not clear we made the right choice. A lot of it is either not properly digested or dissolved and so occasionally blocks the pipes or is spewed out in to the last growing bed in the system. This bed is now protected to some degree by a sieve under the inlet pipe, but we would advise others to experiment with different feeds before buying a 25kg bag! The outcome of this was that the inlet water slowed to the point it was insufficient to trigger the auto-siphons, the growing beds did not drain once flooded (unless triggered manually) and so plants were starved of oxygen for much of the day.
The plants we are now testing for the first time (for us, at least) are French beans, squash, kale and spring onions. The French beans are densely planted, and are producing many more although much smaller, finer beans than those planted in neighbouring soil; the squash and courgette are slower to take off than those planted in soil, as is the kale, so it is still early days; but it is a delight to finally grow some spring onions which have never taken off in our soil beds.
You can follow more of our food-growing ventures on Instagram.
Last year we embarked on a trial of aquaponics in a domestic setting, as part of a PhD undertaken by John Grant of Sheffield Hallam University. Our related blogs are not a guide in themselves, but perhaps help fill some of the gaps we found in existing literature. If you want an introductory guide we recommend Sylvia Bernstein’s Aquaponic Gardening.
This blog follows up on our previous post with a review of our first 6 weeks of hosting fish, and the snags we had with our initial build. There was a gap of some months as having built the physical system we were nervous of the temperatures that could be reached in our polytunnel and so waited until the summer so we could monitor temperatures before introducing fish to the system.
The build
Our first visit to the polytunnel after the winter break was a bit of a shock, as the foundations for the water tank had subsided. This meant we had to empty the water tank and dig out the earth around the sump tank to the extent that we could embed boards to stop the earth from collapsing against the tank.
Lesson learnt: if placing the fish tank over an IBC sump tank, box in the sump tank – however firm you think the ground is. We’d also suggest that the top water tank is filled some time before adding fish, if you are not building on a very firm base. The fix for our problem would have been a lot harder if we had already put fish into the top tank, and very disheartening if cycling had already started.
Having plumbed the growing beds’ inlets and outlets we added water to check the autosipons worked. It is worth doing this without the clay balls in place. We also discovered a leak in one of the tanks at this point.
Lesson learnt: It is much easier to fix these things before the clay balls are added.
Down the track we have realised that it will be difficult for us to alter the height of the outlet pipe that sits up in the tank. Ours is a single piece of pipe down to the junction with the outlet pipe that returns to the sump tank.
Lesson learnt: Plumb the system so that the pipe within the tank can be removed.
We have also realised that whilst we take water from close to the top of the fish tank, this can still carry a fair amount of gunk (technical term). It is likely that this affects the water flow.
Lesson learnt: Install a settling tank in the plumbing immediately after the fish tank.
Which fish?
Originally we had hoped to stock Tilapia or trout, but were concerned about the need to heat or cool the tanks. Having reviewed various guides it became clear that carp were one of the most robust fish when it comes to temperature range and fluctuations. And, as we have a lake full of carp on our doorstep, carp it was. We don’t normally eat carp because it is very muddy when fished from the lake, but we realised that keeping it in the aquaponics fish tank offers a way of cleaning out their systems,and perhaps make them more appetising.
We spent a significant amount of time trying to source organic fish food but it seems to be available solely to large scale fish farms. The first batch was a floating feed from a local ‘World of Water‘, which we have followed up with a slow sinking feed for coarse fish from Skretting. We are also considering how we can farm duckfeed as a high-protein green for them.
Lesson learnt: We recommend a floating feed. Whilst carp are, naturally, bottom-feeders, they were happy eating from t
he surface after 1-2 days, and this gives you a chance to check on them.
Whilst considering the animal kingdom, it is worth noting that we found a frog in the sump tank one day. It is worth ensuring that the tank is sealed, and/or providing a route out for anything that decides to go for a swim.
Cycling the system
We used the Murray- Hallam cycling technique to start the system off. This means adding liquid seaweed and adding plants, and then waiting for 2 weeks before adding fish. There was concern about the pH of the water, which was higher than the ‘ideal’ range given in the literature. However, having tested the pH of the lake from which the fish would be taken, we realised they were the same, so no action was taken on its acidity. This has dropped over time.
We were also able to introduce some rocks from the lake into the growing beds. They carry ‘healthy bacteria’ that would help speed up the cycling process.
3 solar powered oxygenators were used over the summer, which are now powered by the mains as daylight hours start to reduce. We cannot currently monitor oxygen levels in the water, but hope to get a more efficient approach to the energy consumption in future.
The rationale for this approach was that the water would be prepared for the fish. However, 2 weeks into their residence we found a number dead. This hit us hard as the whole idea of this approach had been to avoid stressing the fish. It appears to have happened because the stocking levels were too high, and they received more food than necessary due to some enthusiastic helpers. We reduced the fish levels to a quarter of that suggested in the literature, and their feed is also given at a lower rate than suggested. A month on the pH, Ammonia, Nitrite and Nitrate levels remain spot on.
Lesson learnt: Use the Murray-Hallam method but introduce the fish very gradually, and give a single person responsibility for feeding. It is better to lose some early plants than put the fish through stress.
Data monitoring
We are yet to start the formal monitoring of the system but readings were taken daily during cycling and the first two weeks of the fish residency. Now the system is up to speed we are testing pH, ammonia, nitrite and nitrate levels on a weekly basis. We found (but don’t know if this is the norm) that the Nitrite level rose and fell very quickly, and the Nitrate level rose to 80ppm on the day after the fish were introduced and haven’t fallen much since.
We also visually check the water levels and flow rates in the tanks and from the outlets into the grow beds. We have found that we have to add around 300 litres of water on a fortnightly basis over the summer, suggesting that evaporation is more of an issue than we expected.
Our greatest concern is the temperature, as this is very difficult to control, and there is no data available on this for our particular site (a polytunnel in the UK). The good news is that the water appears to limit temperature swings with its highest temperatures on the hottest days reaching 5 degrees less than the outside air, and staying 5 degrees warmer than the outside air overnight. It is likely that the water will provide a similar cushioning effect over winter, and it is hoped we will have full temperature monitoring to provide more information to protect our fish, and help others with their aquaponics plans.
Output!
First plants go in, and visibly struggle to find nutrients
The planting of the grow beds was circumstantial rather than planned. A tomato cutting or two have taken hold surprisingly quickly and are now fruiting. Some supermarket coriander plants thrived for around 6 weeks but didn’t survive the first cold night of the autumn, whilst some strawberry plants surprisingly fruited within weeks. Spare pea seedlings grew, but failed to fruit. And of the greens, the lettuces have taken well but various brassicas have been eaten by unseen bugs (or slugs?).
At this point it seems that the output is different to that achieved in soil. Some leaves, like spinach and mizuna, are growing well and deliver on taste and texture; but other lettuces are just too limp. Next year we will need to more consciously trial what works and what doesn’t.
And finally, perhaps most exciting for a beginner, that thing where people just scatter seeds into the grow beds – it works! In early August, in a rush before going on holiday, I sprinkled a handful of dwarf peas into the bed and covered it with a single layer of clay balls. Those plants are now growing nicely and producing flowers. Light levels may now be too low to deliver many peas this year, but they may well survive the winter ready to grow on in the Spring.
Aquaponics /akwəˈpɒnɪks/: a system that combines aquaculture (the raising of aquatic animals) and hydroponics (cultivating plants in water) in a symbiotic environment.
Earlier this summer we were contacted by John Grant of Sheffield Hallam University to ask if we were interested in participating in his PhD research into the domestic application of aquaponics. We were, due to…
- Desire for water efficiency – aquaponics counter-intuitively is claimed to use a fraction of the water used in conventional food production
- Low maintenance – aquaponics offers automated water provision and no weeding
- Problems with slugs decimating conventional crops
- High yield and the potential to raise different fish
- Concern about soil degradation, and interest in alternatives
- Opportunity to reuse surplus food-grade IBCs
We are also able to test a system in a domestic setting, whilst also demonstrating it to visitors on our public and educational tours.
Site
John visited us in July 2016 to consider site options. The factors considered were:
- accessibility, as the system will need checking each day to ensure fish welfare
- orientation, to maximise light
- existing shelter, or potential cost of a new shelter
- access to water and power
- available IBCs, which had previously been used to store water on a smallholding.
The key factor was the availability of space in an existing polytunnel. This is about 200m from the homes and offers good light. Access to filtered water and to power had to be installed, and the challenge remains to stabilise temperatures to ensure fish welfare.
System design and build
![]()
Tank preparation
Two food grade IBC tanks were used:
- one for the sump tank (700l) and one for a grow bed (300l)
- the other provided two grow beds
Cutting IBCs with a thin metal angle grinder blade proved surprisingly efficient and produced a clean smooth edge on the plastic. The alternative using a jigsaw would have been more difficult to control on the bendy sides and produced a rougher cut edge. There seemed to be the possibility of fine plastic dust created by the grinding although mostly the plastic melted when cut and a dust mask was worn to protect against this, along with ear defenders and goggles. The containers were swept and washed to remove this “dust”.
Measuring to create an accurate cutting line was difficult with the slightly rounded shape of the IBC. Best estimates need to be used.
Metal work cut easily and it took about 5 hours for all cutting of the IBCs. Creating a lid for the sump tank out of spare bits took about 1 hour. This minimized waste materials from adapting the IBCs.
Sump tank
Fish health and welfare is a key requirement and has driven the design. To keep the water cool, the sump tank has been sunk into the earth floor of the polytunnel. Whilst the temperature of the soil surrounding the tank will vary with the seasons, it will remain cooler than the air temperature in summer and warmer in winter. The degree of difference will be monitored and recorded as part of this project. There is the additional benefit of keeping the sump tank dark to prevent (or at least slow) algae growth. It was for this reason that we chose a dark tank, when initally we thought a light tank could be useful to mark water levels.
The main challenge with burying the sump tank is the pressure of soil on the sides of the plastic part of the IBC. This is averted to some extent by the use of metal panels in the centre of each side.
An inverted growbed was used to protect the sump tank whilst the soil was tamped down, as it protected the water from soil. This was then replaced with a lid, made from a timber frame and the ‘waste’ from the IBC that had been cut to provide two growbeds.
Fish tank
The fish tank is located over the sump tank. The factors considered as part of this decision are access to the fish tank, the simplicity of plumbing, and temperatures. These two tanks are placed next to the doorway with an area in which to put a small table on which to keep records for research. This area is slightly cooler than the centre of the polytunnel, and it is (in broad terms) preferable to give warmer temperatures to the grow beds and a cooler temperature to the fish tank.
The fish tank is offset so that dense concrete blocks can take the weight of the tank on 3 sides, and to allow access to the sump tank (allowing for the potential need to remove and maintain/replace the pump). This access is on the grow bed side of the sump and fish tanks to reduce the complexity of plumbing. The tank is being filled gradually both to limit impact on the availability of water to HHP residents, and to monitor the stability and level of the fish tank.
The fish tank is dark as this is the preference for the fish under consideration, but there is the risk that this will increase the temperature of the water. Another way to manage the temperature is to insulate the tank, and we are considering the use of sheepswool from the HHP flock.
Grow beds
The next stage is to install the grow beds. This needs to take into account:
- Accessibility for planting, including the rinsing of roots, and harvesting
- Plumbing – length and complexity; any impact on accessibility
- Shading of other planting areas
The two options considered were linear and keyhole.
Keyhole: a permaculture approach which minimises path to bed ratio, reducing the distance travelled to work on a bed. This has a particular value with aquaponics as plant roots must be cleaned prior to planting. Plumbing would be to the rear of the beds, with ease of access dependent on the distance of the beds from the polytunnel wall. Planters can be placed at the end of the paths, and potentially incorporated into the aquaponics system at a later date. Additional beds would be added in a similar fashion.
Linear: A more efficient approach in terms of materials used as plumbed routes are marginally shorter (the above diagram is not to scale). The pipework from the fish tank to the growbeds, and from the growbeds to the fishtank each have one less joint. It also allows for planting on a 20cm strip on the south of the polytunnel. The access path will need to be wider than for the ‘keyholes’ as it will need to allow for a working area to clean roots. Additional beds could leave a break to allow easier access to the south edge, with plumbing bridging any gap.
The impact on ease of use will only be fully understood from practice.
