01
Water quality, which is the entire game
Nearly every failure in small-scale aquaculture is a water quality failure, and the fish are usually the last thing to show it.
Dissolved oxygen is the one that kills. It falls at night, when plants respire instead of photosynthesising, and it falls in warm water because warm water holds less. The dangerous combination is warm, still, heavily stocked and heavily fed, at dawn, which is when fish kills happen. Movement at the surface is the cheapest insurance there is.
Ammonia is the second. Fish excrete it, and it's toxic. In a healthy system bacteria convert it through nitrite to nitrate, which is far less harmful and is what plants use. Establishing that bacterial population takes weeks, which is why a new system must not be stocked immediately, and why overfeeding a young one is the classic mistake.
Temperature governs everything else: how fast fish grow, how much they eat, how much oxygen the water holds, and which species will live there at all. And pH shifts through the day with photosynthesis, more sharply in a small water body than a large one.
The practical upshot is to watch the water in place of the fish. By the time behaviour changes, the problem has been present for a while.
02
Building a pond that works
Siting and construction decide how much management the system needs afterwards.
Water supply first: a reliable inflow, or enough catchment and storage to hold a level through the dry period. A pond that drops sharply concentrates everything in it, which is a water quality problem arriving on a schedule.
Depth matters more than area for stability. Deeper water holds temperature steadier, gives fish somewhere to escape heat and cold, and is less prone to swings. A shallow pond is a warm pond, and a warm pond is an oxygen problem.
The ground has to hold water, which means clay or a liner. A pond that leaks is a permanent cost and a permanent worry, and the traditional method of puddling clay to seal a base is still sound.
Build in the ability to drain and to control flow. Being able to lower a pond makes harvest, cleaning and repair possible, and an inlet and outlet you can regulate is what lets you manage quality and not watch it. Screen both, or you stock the whole catchment.
03
Choosing and stocking
Species choice is climate first, then feed, then what you're allowed to keep.
Water temperature is the hard constraint, and species differ enormously in what they tolerate and where they grow best. Beyond that, the biggest practical division is between species that feed low on the food chain, eating plants, algae and detritus, and those that need concentrated protein. The former are far cheaper and simpler to keep, and traditional systems are built around them for that reason.
Stocking density is the lever you actually control. Understocking wastes the system; overstocking is behind almost every water quality failure. Start well below what looks possible, because a lightly stocked pond is forgiving and a heavily stocked one has no margin at all.
Polyculture, keeping several species that feed at different levels, is the traditional answer and it works. Species that eat plants, species that eat what settles, and species that eat what the others miss, together use far more of what the pond produces than any single one.
⚠ Stocking and moving fish is regulated in most places, and introducing a species where it doesn't belong can be seriously damaging as well as illegal. This is a check-locally matter before it is anything else.
04
Feeding, and letting the pond do it
How you feed determines both cost and water quality, and the two are the same decision.
A fertile pond grows its own food. Algae and the tiny animals that eat it are the base of the food chain, and modest fertilisation encourages that natural production, which then feeds the fish. This is the oldest method, the cheapest, and the most stable, and for lightly stocked systems it can be nearly all of the feeding.
Supplementary feed lets you carry more fish and increase growth, and every uneaten pellet becomes a water quality problem. Feed to what's actually consumed within a few minutes, feed less in cold water when fish eat little, and stop entirely if the water looks stressed.
Household and farm by-products, spoiled grain, brewing residue, vegetable waste, insects, worms, have fed fish for a very long time and are the reason integrated systems are efficient. Turning waste into protein is the point.
05
Integrating with everything else
Aquaculture is at its best when it isn't a standalone enterprise, and traditional systems almost never treated it as one.
Fish in rice paddies eat pests and weeds, fertilise the crop and produce protein from ground already growing food. Ducks over ponds contribute manure that feeds the base of the food chain. Pond water and sediment are among the best fertilisers on a holding, and irrigating from a pond delivers both water and nutrient.
Recirculating systems that grow plants on fish-rich water take the same principle indoors: the plants use the nitrate the bacteria have made from the fish waste, and in doing so they clean the water. It's elegant, and it depends on pumps and therefore on power, which is a real dependency worth being clear-eyed about.
The unifying idea is that the waste of one part is the input of another, which is the same logic that makes a mixed farm work on land.
06
Harvest and keeping the system going
Harvesting is straightforward and the handling around it decides quality.
Partial harvesting, taking the largest fish and leaving the rest, keeps a system producing continuously and reduces the density pressure as fish grow. Full harvest by draining is simpler and resets the pond, which is a chance to clean, repair and check the base.
Fish spoil fast, faster than most meat, so the plan for what happens after the harvest matters as much as the harvest. Cooling immediately, and having the preserving or the eating organised in advance, is what stops a good crop being wasted.
Breeding your own stock closes the last loop and removes a dependency, and how difficult it is varies enormously by species. Some breed readily in a pond with the right conditions; others need specific triggers and are far harder.