Water Changes: Why, How Much, and How Often?

From The Aquarium Wiki
Jump to: navigation, search

By Robert T. Ricketts, a.k.a. RTR


Once upon a time, not really so long ago, there lived a lovely theory of the "Balanced Aquarium". You had this glass box, with stainless steel edges, a slate bottom, a stainless steel hood with a couple of incandescent light bulbs. If you were going first class, the light bulbs were tubular "showcase" bulbs. If you were travelling economy, the frame was painted metal, not stainless, and so was the hood. The assembly sat on a wrought iron frame. In this tank you had some sand, some plants, and a pretty good selection of fish. The first-class tanks included a box filter loaded with come charcoal topped with "angel hair" (spun glass).

Or for the truly advanced hobbyist, an internal hang-on box filter was used with layered gravel or sand, charcoal, and spun glass. The air pump that provided water movement for this included a metal flywheel with one or two cylindrical pistons which were astoundingly noisy. It made slightly more noise than a heart-lung machine and had a tendency to walk across the bottom of the aquarium stand and fall off with a crash. These air pumps weighed about the same as modern full canister filter. The crash only occurred between two and four AM. The crash would also disconnect the airline tubing, so when you came in to correct the situation, the puddle into which you stepped brought you fully awake. Your valuable water was lost in the carpet. Change water? Why? It was conditioned, aged, mature water. My original LFS actually sold the stuff- clear amber in color; you really needed to get (read as: buy) some to start your tank.

The esteemed William T. Innes was one of the radical aquarium keepers in the USA, he actually suggested changing part of the water on a regular basis- 10% per week in winter, 25% per week in summer. Shocking. There was lots of controversy on this when I was young, even though Innes had proposed such before even I was born. There were numerous very fancy (read as: expensive) hand-blown glass devices that served as dip tubes for extracting debris from the bottom of the tank. It was suggest to me that I dump this collected material into a pitcher, allow the gunk to settle, and pour as much water as possible back into the tank. That way you only had to add the minimum amount of "raw" water. I am serious. This was standard tank care.

Well, eventually Dr. Innes' position won out, but I understand that there are still holdouts around who do not advocate water changes beyond some absolute minimum perhaps below Dr. Innes' suggested levels. I'm sorry, but to me this like telling people not to vacuum conventional under gravel filters. Where do they think the fish poop, uneaten food, plant debris, etcetera, is going to go? We do add food to our tanks- though in the mythical 'Balanced Aquarium' that was frowned upon also. We also add water to compensate for evaporation. Debris and solid wastes do not evaporate. Even the dissolved materials in the water remain behind when the water evaporates as water vapour, concentrating everything left behind.


Water is wonderful stuff, the universal solvent, responsible for life on earth as we orbit the sun in the wonderful region where water is liquid, over much of the planet, though that is hard to believe looking out my window right now. Fish live in it. They eat in it. They poop in it. Skipping over the actions of nitrification and heterotrophic bacteria, all the food we add to the tank ends as gunk on the bottom of the tank, or incorporated into bacterial or algal films on various surfaces (or even suspended in the water column), or dissolved in the water if it was not incorporated into living flesh. Lots of things are dissolved or suspended in the water. In a mature tank, we most often measure one: the nitrate concentration. We do not measure the dissolved organic compounds (DOC), rarely the rest of the mineral profile of the water, the suspended bacteria or algae. Sometime we will re-visit DOC for a good horror story. It includes anti-metabolites, toxins, colorants, and many other unseen undesirables.


Nitrate concentrations are not just used for themselves. Nitrate is much less toxic short term than its precursors in nitrification, ammonia and nitrite, close to two orders of magnitude less toxic. Long term is somewhat less comforting. Perhaps one order of magnitude (10-20 fold rather than 100-fold) if you are generous.

Part of this borders on the "survive" versus "thrive" arguments. Can we say that I am of the "attempt to have my fish thrive" group? When I advocate a particular water change policy, I use nitrate as the measure or baseline. The water change however does not know that. The siphon removes all of the DOC and other undesirables in that same volume of water along with the nitrate without discriminating. My hope is the same relative change in these unmeasured nasties will be as beneficial as the measurable change in the nitrate concentration. Does that make sense? Nitrate alone is far from the whole issue in water changes, it is just an accessible handle that we can grab and measure.

Calculating Volume

Okay, that has to be longest introduction to a simple technique ever given. How much water do we change, how often do we do that, and how can we tell we have done it right? We have this imaginary tank in our living room. It is a nominal 110 Litres (29 US G.) tank. My tank measures 30 ⅛" side-to side (called length), 12 ½" front-to-back (called depth), and 18 ⅝" tall (called height), maximum outside dimensions. By those outside dimensions, it would calculate to 115 Litres (30.36 US G.) (*), so a manufacturer's nominal volume of 29 gallons does not sound bad.

Ha! The glass is not anywhere near that size. And the glass has thickness too. The inside of the glass of tank is something like 29 ⅜ x 11 ⅞ x maybe 17 ½" effective depth to a practical water level. That is 100 Litres (26.4 US G.). Then I have a thick substrate, which displaces a significant amount of water. The actual fill measured volume of the tank is about 83 Litres (22 US G.).

But, I cheat. The tank manufacturers assign these imaginary numbers to the gallonage of our tanks. I use those numbers for water changes. If I post that I do a 20% change on that tank every week, I change 29 x 0.20 = 22 Litres (5.8 US G.), almost 6 gallons. Granted that as I am slightly neurotic about a few little things, I have 20-25 -33 -50% of nominal volume marked on the sides of most of my tanks. I vacuum the substrate by Python; drawing the water down approximately to my target mark, then refill the tank. It this particular tank, that 5.8 gallons is about 26% of the actual volume. It is 20% of the nominal volume. If the maker can lie to me, I can lie to myself. I will admit that when I post such information I try to say 20 - ~25%, so I am covered whichever way you measure. Most Web forum discussions are based on nominal volumes, so water partials in reality are larger than stated when measured in actual volumes. That to me is a very good thing. It is one case where you don't know doesn't hurt you, and it is good for the tank.

How Often?

How often do I do this change? If you ask I will calmly respond with "every week". If you pin me to the wall on the topic, I will confess that it is really more like every week to 10 days. Even the anal need some breaks. Sometimes I hold the schedule well, and am much happier with the world. But if a tank should get only three changes this month, the world will not end, the fish will not die, and I will be only slightly guilty.

Next month I will try to do better, and usually I do. I favor frequent changes because they do the best job of maintaining steady water parameters. The pH, hardness, nitrate, DOC, clarity, will not have shifted that much since the last change. If you wait two weeks, those things will have changed up to twice as much. Somewhere along the line you are increasing the stress on the fish and on the system itself. The more frequently the water is refreshed, the lower the stresses of the system and its inhabitants. Most fish "like" water changes; they seem energized by them, so long as they do not involve so great a change in water parameters that the change itself is inducing stress or even shock.

How can we tell that we have done it right? Easy. We test for nitrate again. You pick a level of nitrate that is comfortable for you. Many people select 20-30 ppm (mg/L). I aim for lower concentrations, 5-10 ppm. I do that in part because I have a lot of plants and am not a big fan of algae, I have enough of it now, thank you very much. While you are getting your feet wet (bad joke) with monitored water changes, keep some sort of record of your nitrate readings. If you start out at a level too high for comfort and it is going down gradually, you are winning. If it is rising (increasing over time), you need to re-evaluate. More volume per change or change more often, your choice, so long as do not shock the fish with too great a water parameter change at one time. Are you overfeeding? Overcrowded? If you are maintaining or decreasing the pre-water change nitrate total very slightly, you did not need to read this article, you are doing it right already.

Easiest Thing

Water changes are one of the easiest things you can do for your tanks and fish, and will do more to support their health and longevity than the most hi-tech filtration and control systems can do without water changes. It is a very good habit to develop. Don't really siphon into your shoes (yes, I have), but do get your feet wet.

Note: All of the above involving nitrate as a general pollution indicator applies only to tanks without plants. In a planted tank, and especially in a heavily planted tank, the plants use nitrate as a nitrogen source, as they also use ammonia. Plants do not take up nitrite; it is too toxic for easy uptake and internal transport. Under these circumstances, we do water changes arbitrarily. I do 50% weekly as I do add some supplements routinely, and partials reset those without my having to measure multiple items in multiple tanks - in other words, it is easier, faster, and cheaper. The KISS principle is good to you, and for you, so keep it simple, Sam.

  • To calculate the actual USA volume of a rectangular tank, measure (in inches) the inside length, width, and height (to a realistic water level, not to the rim). Multiply these three numbers together to get the volume in cubic inches. Divide that product by 231 (cubic inches per US** gallon) to get the actual volume without gravel, rocks, etc. If you want the actual working or fill volume - allowing for substrate and décor, you must measure the water the first time you fill the tank (the only time I ever carry buckets).

By Robert T. Ricketts, a.k.a. RTR