I have been writing increasingly long threads on Mastodon about my aquarium and the mathematics I have been using to make sense of its water tests. The latest ran to nine posts, with links back through several earlier threads. By that point, a blog seemed a more sensible place to put it all.
I am very much a nerd, so it was perhaps inevitable that keeping an aquarium would lead to a mathematical model. I wanted to understand what the water tests were telling me and how much confidence to put in the readings. I am not a probabilist in my day job, so building and testing the model has also been a useful excuse to revisit some Bayesian filtering.
For this first post, though, I thought I would start with the aquarium itself, where nitrate comes from and why I monitor it. The mathematics can follow in later posts.
The aquarium
My tank is a Fluval Roma 125, a freshwater aquarium with a nominal capacity of 125 litres. I upgraded from a smaller tank at the end of August. It has a dark sand substrate and a mixture of live plants, including cryptocorynes, Java fern and several stem plants.
The tank has an EcoFlow 150 internal filter and a Fluval AquaSky light. I aim to keep the water at about 22.9°C.
It is currently stocked with:
- Eight male guppies.
- Seven platies.
- Ten panda corydoras.
- Two honey gouramis.
There are also cherry shrimp and nerite snails.
I record water tests, temperature and maintenance in Home Assistant, the software I use for home automation. Having a record makes it easier to see what has changed, particularly after adding fish or plants or altering the feeding.
Where nitrate comes from
Nitrate is one of the forms in which nitrogen occurs in water. In an aquarium, much of that nitrogen arrives in fish food.
Fish food contains nitrogen, particularly in proteins. Fish use some of that nitrogen and excrete some as ammonia, mainly through their gills. Uneaten food, faeces and decaying plant material also release ammonia as they break down.
Microorganisms living on the filter media and other surfaces convert ammonia into nitrite, then nitrite into nitrate. This process, called nitrification, requires oxygen. It is an important part of biological filtration, because ammonia and nitrite are harmful to fish at much lower concentrations than nitrate.1
The nitrate can then accumulate as feeding continues to add nitrogen to the aquarium.
Some nitrate can also arrive in the tap water. It occurs in source waters through natural processes and human activities, including agriculture. Fresh water from the tap does not necessarily start at zero nitrate.2
Why monitor it?
Plants can use nitrate as a source of nitrogen for growth. As they take it up, nitrogen moves from the water into plant tissue; removing plant trimmings carries some of it out of the aquarium. The amount taken up will change as the plants grow and become established.3
Nitrate is generally less acutely toxic to fish than ammonia or nitrite, but sustained exposure to elevated concentrations can still be harmful. The effects depend on the species, life stage, water chemistry and length of exposure. Studies have reported effects on growth, development, behaviour and physiological performance. I would be cautious about treating one concentration as a universal limit for every aquarium fish.4
I therefore monitor nitrate alongside ammonia, nitrite and the other water parameters, and carry out regular partial water changes. These replace aquarium water with tap water, which may already contain nitrate. I test the tap water as well so that I know what I am putting back.
I am also interested in how the nitrate concentration changes between water changes. With an increasing number of plants, I would like to know whether their uptake is making an appreciable difference to the amount that accumulates each week.
A nitrate test cannot tell me separately how much was generated from fish waste and how much the plants removed. I am looking at their combined effect on the water.
My test involves comparing the colour of a treated water sample with a printed chart. That is useful for routine monitoring, but estimating a rate of change from a few readings is less straightforward. The colour comparison is uncertain, and I also need to account for the water changes.
In the next post I will explain how I combine the tests and water-change records to estimate nitrate concentration and how quickly it is accumulating, together with the uncertainty in both estimates.
Footnotes
-
Craig Watson, Ruth Francis-Floyd, Denise Petty and Deborah B. Pouder, Ammonia in Aquatic Systems, University of Florida IFAS Extension. ↩
-
Matthew Recsetar, Aquaponics Basics: How Fish, Plants, and Microbes Work Together, University of Arizona Cooperative Extension, July 2026. ↩
-
Camargo, J. A., Alonso, Á. and Salamanca, A. (2005), Nitrate toxicity to aquatic animals: a review with new data for freshwater invertebrates, Chemosphere 58(9), 1255–1267. ↩