Why Water Changes During Cycling Are Good

FishForums.net Pet of the Month
🐶 POTM Poll is Open! 🦎 Click here to Vote! 🐰
I am so new to keeping fish its painful, I freaking LOVE your posts.  Thank you so much for taking the time to write them!  You have answered so many of the questions I have had.  I also know what I will do different with my next tank (shhh don't tell my husband)!
Another article that should be deleted. From the outset it starts with an argument that can't hold water. At a level of 5ppm as nitrogen, both ammonia and nitrite will begin to stall a cycle and rising higher they kill the bacteria we want and encourage the bacteria we don't. On the API kits this is 6.4 ppm for ammonia and 16.4 for nitrite.
Next, it totally ignores the NH3/NH4 issues involved. And this means all the fancy spread sheet work is worthless in this article. Where is the pH and temperature data that is relevant. What about carbonate and oxygen levels. What about different reproduction rates for bacteria depending on temperature. All of this has been ignored.
Then it ignores the biology of the the bacteria. They do not multiply until there is sufficient excess food to make that possible. And the thing is a Nitrosomonas type AOB can increase how much ammonia it converts before it reaches the point of dividing.
I would say the urban myth here is that many water changes are good during cycling. The science doesn't support this. I would say that in fish in cycling the goal is to allow both ammonia (esp NH3) must be allowed to be as high as reasonable as long as the fish show no signs of distress and the N3 levels are held under .05 ppm.
During any cycle, with or without fish, the goal is not to do multiple water changes, it really is just the opposite. The point at which they might be necessary depends on if fish are involved or not. The biggest issue in fish in cycles isn't the ammonia so much as the number and types of fish used. If one uses the right fish in the right number, ammonia is much easier to manage and fewer water changes should be needed.
As for the whole snapshot argument it is silly. The spreadsheet is a series of snapshots. The way we know how a cycle is progressing is by having a series of those "useless" snapshots as they will tell us clearly what is going on during a cycle.
But here is the real nail in the coffin of the whole idea of this article. There is a simple way that we all know already why what is stated in here is simply not true. And that proof is seen by anybody who has done a fishless cycle dosing to 2 or 3 ppm of ammonia. We are controlling ammonia levels closely when we fishless cycle. We can hold them so they never go above that 2 or 3 ppm. These are the same levels the author suggests would speed up a fish in cycle . But if this were the case why do these same ammonia levels not speed up a fishless cycle? If the logic of this article held water, the average time for fishless cycling using 2 -3 ppm levels of ammonia should be fewer than 3 weeks tops. Case closed.
ps- During a fish in cycle one rarely should allow ammonia levels to rise much above 2 ppm for any length of time even if NH3 is below .05 ppm. But there is a huge difference between doing water changes because they are needed and just doing them because somebody thinks they have shown they cycle faster when they haven't.
For a long time, I've seen the "common knowledge" that while a tank is cycling, you don't want to do any water changes. (Obviously, this is cycling with fish, not fish-less cycling). One of the reasons given is that you'll take away the bacterial colony's food and the colony won't grow. Or that the colony won't grow large enough to handle the fish's bio-load if you do water changes during the process. Another reason given is that by doing water changes, you'll lengthen the cycling process, and that since the fish in the tank are being exposed to ammonia anyway, getting it done as soon as possible is desirable.

There are many, many websites out there that repeat this same information, Google brings up many examples.

It's time to show everyone that this information, this "common knowledge" is just plain wrong.

First, let me explain the theory behind cycling. The steady state of the situation is well known to us. The size of the bacterial colony is large enough that it consumes all the ammonia the fish produce immediately. Worded another way, the rate of consumption (by the bacteria) is equal to the rate of production (by the fish). Or, to write it mathematically,

r_p = r_c

r_p stands for rate of production
r_c stands for rate of consumption

Let me re-write this equation as (the reason why will become apparent soon)

0 = r_p - r_c

However, while the tank is cycling, the rate of production and the rate of consumption are not equal. Specifically, the rate of production is greater than the rate of consumption (mathematically r_p > r_c). The bacterial colony hasn't grown large enough to consume a large amount of ammonia yet. That leads to an accumulation.

a = r_p - r_c

where a stands for the accumulation of ammonia. (for the real math nerds out there a = dc/dt where c is the concentration of ammonia and t is time)

The real power of this equation and the above one is that it is not the amount of ammonia that is in the tank that is important. It is rates of production and consumption that are important. When the rates are equal, no ammonia will accumulate in the tank, and that is what we read when we take the sample. A reading of zero indicates that the rate of consumption by the current bacterial colony is equal to or greater than the rate of production by the fish.

However, a reading of some ammonia, tells you virtually nothing about the rates in the tank. The ammonia reading is like a snapshot of a moving target. You get a picture of the tank at that exact moment, but it doesn't tell you anything about the dynamics of the situation at that time.

Also, it is very important to note that the rates themselves are independent of the ammonia reading at any time. This can be taken advantage of, because, if you do waterchanges while the cycling process is going on, you can dilute the poison the fish are living in, and not disturb the cycling process. The bacteria can only eat so much in a given day, and until the very end of the cycling process, the bacteria aren't going to consume as much as the fish produce in a single day. For example, if on a certain day, the bacteria can consume 0.1 ppm of ammonia, they don't care if there is 1 ppm of ammonia or 5 ppm of ammonia in the tank. Just so long as there is at least 0.1 ppm of ammonia in the tank, the colony gets as much food as it can consume, it will grow and consume a certain amount the next day. The bacteria only stop growing if there is insufficient food. So, since the bacteria don't care if there is 0.1 ppm, 1 ppm, or 5 ppm, there is no reason not to do a water change to dilute that down -- just so long as you don't dilute it down below 0.1 ppm. Because, the fish very much care if they are living in 0.1 ppm, 1 ppm, or 5 ppm. That this can be done is what I am going to show below.

I set up an excel sheet to simulate this equation. One where I simulate straight cycling without water changes, and one where I simulate doing a 10% water change at the end of every day. I picked r_p = 1 ppm of ammonia per day, and r_c = 0.000014 ppm of ammonia per day for the first day. Then, everyday after after that r_c doubles (approximating that the filter bacteria double in number roughly once every 24 hours). That is, on day 2, r_c = 0.000028, on day 3 r_c = 0.000056 and so on. This choice of r_c on the first day will become apparent below.

day....ammonia reading at end day (ppm)

The tank took 21 days to cycle, and the ammonia got up to 15.16 ppm as a maximum concentration. This is pretty high, and might have needed emergency action at that level. (If I may not too immodest, see my other article to compute when emergency action is needed: http://www.fishforums.net/content/forum/15...Ammonia-Charts/ ) Most test kits don't go up to that high anyway.

Now, let's look at a tank that gets a 10% water change at the end of each day:


This tank cycled one day faster. And the maximum level of ammonia that built up, 7.52 ppm, was less than half of the maximum of the non-water changed maximum of 15.16 ppm. 7.52 is still pretty high, but probably not nearly as bad a 15 ppm. The filter bacteria do not get starved, there is plenty of ammonia to eat every day until the last one when the tank is cycled. In fact, at that point, the bacteria colony has grown large enough to consume 3.67 ppm of ammonia per day, more than 3 times as much as needed -- recall that the fish are assumed to excrete 1 ppm of ammonia per day. And, of course, the objection that cycling takes longer is smashed, because this tank actually cycled one day sooner.

Normally, this is where my argument would end. Personally, I think the math is convincing enough, but I also know the math really well. I know that most others don't know the math as well as I do, or quite get the math. However, Over the last month, I've been performing an experiment. I took my quarantine tanks, two bare 10 gallon tanks, and two extra filters I have and set the tanks up. I performed the experiment that the math above describes.

Here's the details. I did fishless cycles, because then I could control exactly how much ammonia went into each tank. Both tanks were set up the same, and I took the readings at the same time each night. I added the ammonia at the same time each day. I added enough ammonia to be equal to 1 ppm per day. The test kit I used was brand new, and measured from 0 to 7 ppm of ammonia. The gradations were 0, 0.25, 0.5, 1, 2, 4, and 7+ ppm. I had to guess some of the intermediate values, so, I only report whole numbers (no decimals) because they are just guesses. Anything at 7 or above is just recorded as 7, because that is the limit of the tank.

Here's my computational results side-by-side with the experimental results.

Now you see why I used r_c = 0.000014 as the first day's consumption value. It is the value that best fit my data. It is actually a pretty good fit between the experiment and data. The tank cycled on the same day, and the decrease at the end is missed just a little. I got a value of 4 ppm for the 20th day's experimental value, when the computer predicts 5.3 ppm. But, I think that is pretty good. Most likely, the simple doubling rule isn't quite right when the food is starting to become scarcer. Nevertheless, it shows that model used is pretty darn good, really.

Here's the tank with the 10% water changes done every day (I did the water change everyday immediately after doing taking the ammonia readings).


Again, very good agreement between simulation and experiment. The tank did indeed cycle one day earlier than the non-waterchanged tank. The readings taken during the beginning and the simulation agree very well, though again at the end of the cycling the experiment and the computer had some small disagreement. The disagreement is not large, and I think that the general trend is captured really well.

Since I feel the model is pretty well verified now, I ran a few more simulations. A tank that undergoes 25% waterchanges everyday reaches a maximum of 3.82 ppm of ammonia, and is cycled by the end of day 19. A tank that undergoes a 50% waterchange every day reaches a maximum of 1.99 ppm of ammonia and is cycled by the end of day 18.

I am fairly confident of the 25% water changed tank, but not sure of the 50% tank. The levels are kept very low in that tank, and since the bacteria don't grow exponentially (doubling every day) when the food levels are lower, it may not be exactly right Nevertheless, the idea is very similar, you can keep the ammonia levels low without sacrificing speed of the cycling process. Keeping the ammonia levels low keeps the fish in much better health than otherwise.

For example, see "Low levels of environmental ammonia increase susceptibility to disease in Chinook salmon smolts " by Ackerman PA, Wicks BJ, Iwama GK, Randall DJ in PHYSIOLOGICAL AND BIOCHEMICAL ZOOLOGY volume 79 JUL-AUG 2006) which showed that fish exposed to low levels of ammonia were more susceptible to disease later on in life. Basically, exposure to any levels of ammonia leads to greater health problems for the rest of the fish's lives. But, the lower that exposure, the less susceptible the fish will be later in its life.


In conclusion, I have given both experimental and computational evidence why it is not bad to perform waterchanges while cycling, and indeed actually beneficial. The experimental evidence I think is particularly compelling, it is no longer just a theory or a mathematical construct that I came up with. The theory matches the experiment pretty well. The experimentally waterchanged tank cycled one day earlier than the non-waterchanged one. Unfortunately the test kit couldn't measure high enough to measure the peak concentration each tank go up to, but I think that the very good matching of the curve up and then the pretty good matching of the curve down is excellent.

So, everyone, please do waterchanges even when cycling. Especially at the beginning when the bacterial colony is just starting to grow and can only consume a very tiny percentage of the fish waste every day. At that stage you can do very large water changes to keep the ammonia level down then and later. It doesn't matter because the bacterial colony will grow until the rate of consumption is equal to the rate of production. And, so long as the temperature, pH, and hardness of the replacement water is the same as the tank water, there is no harm in doing large water changes, 50%, 70% even 90% if you have the capability.
Wonderful, ty so much

Most reactions