I had a QT tank set up and cycled, getting ready to order some angels, and then had to put the order on hold for a while. I had been keeping the cycle going by adding ammonia every day, but rather than tear down the tank I decided to see how nitrification affects the pH in RO water.

The purpose was to find a method of keeping a low pH tank using nitrification and timed water changes instead of peat, and what I have found is that this method can only work if your water has a KH as close to zero as possible. If your source water has an appreciable amount of alkalinity, only RO followed by mixed-bed deionization can produce such water. This is because by the time the pH falls the nitrate is too high. You need the pH to fall before the nitrate reaches 10-20 ppm. After the first month or two the pH is more likely to stay low, and water changes can be more frequent and nitrate kept under 10. I don't know why the ageing of the filter does this.

To do the test I started keeping a log of the amount of ammonia added and the time it was added, and logging the pH along the way. When the pH reached 4.8+/- and the filter stopped working, I did 100% water changes and started over.

The results are shown in this graph:

phdriftgraph0001.jpg

I ran the test 12 times, but only the last 4 runs are shown in the graph. I wasn't able to be here often enough to get closely timed data points until the last few weeks. In the earlier runs the pH behaved in a similar way but the curve was not as smooth.

Extracting useful information about pH drift from the graph is difficult unless you can get a sense of how quickly time flows along with the increase in nitrates. There are too many variables involved to make a blanket statement like, "one ppm of ammonia per day". Stocking level or bioload is by far the biggest variable, but feeding methods, tank cleanliness, temperature, plants... all these effect the rate of nitrification. If I had to take a wild guess I would say that 2 ppm per day of ammonia is a crowded tank, and that 0.5 ppm per day is a lightly stocked tank. Any other guesses or input on this would be appreciated.

One possible way to measure the rate of time along the graph is to have X fish in Y gallons, with no biofilter at all, and carefully measure how much ammonia accumulates in the water over time. This could be done without forcing a fish to stew in toxins as long as the pH was fairly low and water is changed every day after measuring total ammonia. I have done this with GBRs in hospital tanks but its hard to extrapolate to bigger fish.

Another method, which would only be valid without plants or algae, would be to measure nitrate before regular water changes in a functioning BB tank, and try to relate the number of discus per gallon to the build-up of nitrate, then do a little math. The drawback to this method is that the common nitrate tests have crappy accuracy.

Some observations.

1. In these tests the pH was brought below 5.0 in an average of about 4 days from the complete water changes. The nitrifying bacteria active in the filter are the species that thrive in the midrange of pH from 6.0 to 8.0, the ones that have been studied and given names, and since these species become inactive when the pH falls below 5, the biofilter was incapable of processing ammonia. This explains why the graph levels out.

2. If the pH is kept low and steady for several weeks, or lowered very gradually, a colony of low-pH nitrifiers will develop, allowing the cycle to continue. This is what happens with Heckel biotopes.

3. After each pH drop and water change, the nitrifying bacteria were able to function again without any apparent setback. If the legends about dead biofilters from pH crashes are true, the pH would need to fall lower than 4.7 somehow, or stay there long enough to become lethal.

Here is another interesting graph. It is easier to read.

graph-duck-rabbit.jpg