Introduction

In the earlier part of this article, I described the role of oxygen in seawater and its potential effects on marine species. I noted, particularly, that hypoxia is likely to occur in reef waters, within coral colonies, at the coral tissue's surface, and that hypoxia potentially affects species exposed to it. The scientific literature suggests that such conditions might be relatively common at night when respiration is high and primary oxygen production through photosynthesis is not occurring. On coral reefs, although oxygen levels are supersaturated near the surface during the day in shallow waters, they are often reduced by 40-90 percent or more at night. Primary factors affecting the oxygen content of reef waters at night include surface conditions and water mixing, as well as total community respiration. So how do aquariums compare to reefs?

In this article, I report the results of numerous tests of various water conditions in closed system aquaria. I utilized various ways to "oxygenate" water and compared their effectiveness. While the data I presented at the 2005 IMAC conference were limited to some degree, I believe that enough measurements have now been taken (including suggestions made by attendees during the question and answer session following my presentation of this material at the recent IMAC conference in Chicago) to present a somewhat comprehensive picture of oxygen dynamics in reef aquaria. The most notable additions presented here are the effects of oxygen production in tanks that are initially hypoxic. In the next article, I will continue to report my results and discuss more natural oxygen dynamics of several reef aquaria.

General Methods

I used a YSI Model 58 submersible oxygen field probe to record oxygen levels in all tests. This particular probe does not have a self-stirring BOD probe. Without the self stirring device the probe must be exposed to flowing water or be moved through the water at a rate fast enough to obtain accurate measurements. More information on this meter's specifications can be found here.

All readings were taken following the installation of a new membrane, and the meter was calibrated by containment in a 100% humidity environment adjusted for Hg pressure at altitude, according to instructions provided by the manufacturer. Another calibration for a zero oxygen value was performed with a solution of stirred water with a known volume of sodium thiosulfate, an oxygen scavenger.

During the installation of the membrane, it is important to avoid letting bubbles be trapped under the membrane or readings will vary significantly. It is also important that bubbles are not present in the water sample being recorded. One very small bubble was present under the membrane, which may account for some minimal variability in the results; although after taking hundreds of readings and analyzing standard deviations of the measurements, I can state that the readings are statistically accurate and that any variance in replicated readings is insignificant at a 95% confidence interval.

All readings were taken after a 10-minute stabilization period for the probe in any given solution. In all measurements, four readings were taken 15 seconds apart and the mean of the values was recorded.

Experimental tanks included controls with sterile containers and sterile seawater, either closed or open to the air, tanks containing freshly mixed seawater using various water flow or aeration devices but no organisms present and with or without light, and various "reef aquaria" containing a variety of organisms and apparatus. All tanks mentioned above contained water at ambient levels of oxygen, either following mixing in the case of the sterile controls, or those levels already present in the tanks that were exposed to air or enclosed. In addition, measurements were taken in tanks containing freshly mixed seawater that had been made hypoxic by the rapid addition of nitrogen gas to displace oxygen to a level below 20% of saturation for all the conditions tested above.

All tanks contained seawater at 35psu as measured by a Reichert salinity refractometer calibrated using double distilled water. Although some variance in tank temperature occurred in the tanks where lighting was a variable, the oxygen meter is temperature compensating. The tank temperatures without lighting, and some with lighting, were maintained at 26°C, with some of my home systems reaching 26.6°C during the late afternoon.

For informational purposes, oxygen saturation values at various salinities and temperatures are provided in Table 1.

Results

Control tank 1

Distilled water was autoclaved for one hour prior to use. The salt source was an unopened 50-gallon bag of Crystal Seas Bioassay Formula™ salt (Marine Enterprises, Inc.). Seawater was prepared in clean two liter beakers, covered, and stirred on a stir plate with a stir bar until completely dissolved. Six liters of seawater at 35psu were filtered to 0.1µ through graded Whatman papers and then through Millepore filters under a vacuum. Filtered seawater (1500ml) was poured into a desiccation bowl that was triple rinsed, then fitted with a stir bar and rubber stopper through which the oxygen probe was placed into the seawater solution. The bowl was placed onto a stir plate and its water was stirred at a minimal rate in excess of that required for probe measurements for a period of 240 minutes under ambient room light or in darkness by covering the desiccation container with a thick cardboard box. Results are shown in Table 2.

Experimental Tank 1

Ten gallons of seawater were mixed to 35psu using an unopened bag of Instant Ocean™ salt and distilled water poured into an acid-rinsed and neutralized 10-gallon glass aquarium.

As a first variable, the tank was either sealed with a sheet of acrylic using removable tack adhesive to prevent air from entering the aquarium, or it remained open to the air. A second variable was the presence or absence of water circulation by a single MaxiJet 1200™ (Aquarium Systems) placed at the center of the tank's short end so that the flow moved along the aquarium's long centerline. It was also placed just far enough below the surface so that no air was pulled into the pump. Visual inspection of circulation indicated a highly turbulent flow. The oxygen meter was placed on the aquarium's opposite end through a hole cut in the acrylic and sealed with parafilm to fill any gaps. For stagnant tests the meter was carefully rotated in a circle within the tank at a rate allowing readings to be recorded; the parafilm provided a flexible but airtight barrier to allow this movement. A third variable was light. Lighting was provided by a single 18" fluorescent fixture, or the room lights were turned off. While not completely dark, there was too little light in the unlit room to see the oxygen meter reading without a flashlight, and it was assumed that this latter condition sufficed as a "dark" condition. Results are shown in Table 3.

Experimental Tank 2

A 15-gallon (8" x 24" x 12") aquarium was filled with seawater at 35psu prepared with deionized water and Instant Ocean™ salt. Hypoxic water to below 20% saturation was achieved by bubbling nitrogen gas into a powerhead's intake to rapidly displace oxygen by saturating the tank's water with fine nitrogen bubbles. A ten minute wait then allowed all residual bubbles to rise to the surface and break so that no bubbles would affect test readings by the probe. Results are shown in Table 4. The conditions tested were as follows:

1. A single MaxiJet1200™ was placed at one end of the tank, creating highly turbulent flow throughout the water volume and visibly stirring the surface.
2. The tank was left open to the air without any water flow and the oxygen probe was slowly moved at a rate just fast enough to take a reading.
3. A single ceramic airstone was added at one end of the tank using a Reno 400 air pump. Air bubbles were coarse and did not reach the other end of the tank. The power head was turned on for 15 seconds to mix the water and a reading was taken by slowly moving the probe through the water where no bubbles would affect the reading.
4. Thirty-seven grams of live Chaetomorpha sp. algae were added to the tank and illuminated by two 15" 18-watt fluorescent daylight bulbs placed close to the water's surface. The powerhead was turned on for 15 seconds to mix the water and a reading was taken by slowly moving the probe through the water.
5. A four-pound piece of live rock approximately 40% covered with Protopalythoa sp. zoanthids and 60% with coralline algae was placed into the tank and illuminated by two 15" 18-watt fluorescent bulbs placed close to the water's surface. The power head was turned on for 15 seconds to mix the water and a reading was taken by slowly moving the probe through the water.

Experimental Tank 3

A 10-gallon reef aquarium established for three years in my lab is described as follows:

Substrates: 5 cm medium grain aragonite, 3 kg live rock
Fish: none
Corals: small colonies: Ricordea florida, Millepora sp., Porites cylindrica, Zoanthus sp.
                                       Rhodactis inchoata
             medium colonies: Porites cylindrica, Protopalythoa sp.
Invertebrates: five Trochus sp. snails
Water flow: a single MaxiJet 1200™ powerhead (Aquarium Systems, Inc.)
Other filtration: CPR BakPak™ skimmer
Water changes: none
Additions: kalkwasser as replacement water
Lighting: a single 18" fluorescent bulb (Lights of America, Inc.)

The oxygen probe was either stabilized with a ring stand so that it remained submerged in the powerhead's water flow (stirred conditions), or the meter was carefully rotated in a circle within the tank at a rate that would allow for readings to be taken. In the first set of experiments, nitrogen gas was bubbled into the powerhead's intake to rapidly drop oxygen levels until they reached approximately 20% saturation. Results are shown in Table 5. The conditions tested were as follows:

1. A single MaxiJet 1200™ placed in the center of the tank's short side and adjusted to create a lengthwise circular water flow in the tank. The tank was illuminated by a single 18" fluorescent daylight bulb (Lights of America, Inc.).
2. A single MaxiJet 1200™ placed in the center of the tank's short side and adjusted to create a lengthwise circular water flow in the tank. All lights, including ambient room lights, were turned off to simulate darkness.
3. A medium pore ceramic airstone connected to an air pump (Supra™, Tetra Inc.) was placed in the center of the tank. The powerhead was turned off. All lights, including ambient room lights, were turned off to simulate darkness. The meter was carefully rotated in a circle within the tank away from any bubbles at a rate that would allow for readings to be recorded.
4. A skimmer was placed on the tank (Remora hang-on™, Aqua C). No powerheads or airstones were in operation. All lights, including ambient room lights, were turned off to simulate darkness. The meter was carefully rotated in a circle within the tank away from any bubbles at a rate that would allow for readings to be recorded.

Experimental Tanks 3 and 4

In addition to the tank described above, a second 10-gallon reef aquarium established for three years in my lab is described as follows:

Substrates: 5 cm medium grain aragonite, 3 kg live rock
Fish: none
Corals: small colonies: Capnella sp. and Millepora sp.
             medium colonies: six Rhodactis inchoata
             medium to large colony: Zoanthus sociatus
Invertebrates: five Trochus sp. snails
Water flow: a single MaxiJet 1200™ powerhead (Aquarium Systems, Inc.)
Other filtration: none
Water changes: none
Additions: kalkwasser as replacement water
Lighting: a single 18" fluorescent bulb (Lights of America, Inc.)

In this series of tests, the relative effects of photosynthesis and respiration on the oxygen content of two 10-gallon reef tanks were measured in the absence of any water flow or other apparatus. Lights above the aquarium were either turned on, or all lights, including ambient room lights, were turned off to create darkness. Prior to beginning the experiment, all tanks were operating with all apparatus as described above. Results are shown in Table 6.

Experimental Tanks 5, 6 and 7

The following tanks are part of an interconnected six-tank propagation system. It has been established for three months and is described as follows:

Total system gallons: 600
Water flow: Ampmaster 2600™ centrifugal pump (Dolphin Aquarium and Pet Products, Inc.) fitted with eductors at three of the six outflows (one per tank); some tanks also have high volume powerheads for circulation (6060 Stream, Tunze, Inc.; Seio 1100; Rio pumps, Inc.)
Other filtration: MR-2™ protein skimmer (My Reef Creations, Inc.) powered by a Gen-X MAK 4™ centrifugal pump (Pacific Imports)
Water changes: none
Additions: kalkwasser as replacement water, calcium chloride, sodium carbonate, sodium bicarbonate
Lighting: various, depending on the species being cultured.

Tank 5

Size: 30-gallon
Substrates: 7.5 cm medium grain aragonite, small amount of live rock rubble (<0.5 kg)
Fish: none
Corals: small colonies: 63 Capnella sp.
             medium colonies: three Capnella sp.
             medium to large colony: one Capnella sp.
Invertebrates: five Trochus sp. snails
Water flow: slow flow from an Ampmaster 2600™ centrifugal pump (Dolphin Aquarium and Pet Products, Inc.)
Lighting: four 48" 110 watt VHO fluorescent bulbs (two actinic, two daylight - URI, Inc.)

Tank 6

Size: 30-gallon
Substrates: 7.5 cm medium grain aragonite, small amount of live rock rubble (>25 kg)
Fish: none
Corals: none
Invertebrates: five Trochus sp. snails
Water flow: slow flow from an Ampmaster 2600™ centrifugal pump (Dolphin Aquarium and Pet Products, Inc.)
Lighting: four 48" 110 watt VHO fluorescent bulbs (two actinic, two daylight - URI, Inc.)

Tank 7

Size: 30-gallon
Substrates: 7.5 cm medium grain aragonite, live rock (>10 kg)
Fish: Chaetodon rostratus
Corals: none
Invertebrates: five Trochus sp. snails, 50 Nassarius sp. snails
Water flow: slow flow from an Ampmaster 2600™ centrifugal pump (Dolphin Aquarium and Pet Products, Inc.)
Lighting: 65 watt power compact fluorescent 6500K (Lights of America, Inc.)

Discussion

The tanks tested in this article show some interesting and unexpected results. In the control tanks with sterile water, a small initial peak occurred in both samples, despite being sealed from outside air. This probably resulted from the equilibration of existing chamber air (21% oxygen) with the seawater since only half the desiccation chamber was actually filled with sample water. The later slight decline may be due to slight warming of the water sample from the motor of the stir plate. Differences in the curves are, as expected, statistically insignificant.

In the first experimental tank, the variations are slight and the curves are statistically insignificant (Borneman, in preparation). Any variations between the lines are likely due to the reasons mentioned above, to the order in which each condition occurred and to stochastic variations that are nearly impossible to adequately control in multiple tests involving numerous variables, non-sterile conditions, and environmental exposure in a working laboratory.

The second and third experiments produced surprising results, especially in terms of the variables' effects in non-hypoxic tanks (next article). This set of experiments was recently performed after Martin Moe, Julian Sprung and other attendants at IMAC suggested that water movement and airstones, in particular, might play a role in oxygenating water that was already reduced in that gas. My data at the time had explored the role of these variables only on various tanks or chambers with significantly higher "normoxic" water conditions. In the case of hypoxic water, such as might be found at night or in the event of power failures, the results clearly show that airstones, water pumps and skimmers are quite effective in rapidly raising the oxygen level of both plain seawater and small reef aquariums. Their effect on larger tanks, however, remains minimal and, again, will be shown next month.

One finding which conflicts with previously explored data presented at IMAC is the ability of photosynthesis to rapidly raise oxygen levels in hypoxic seawater only. I will present conflicting data and potential reasons for the discrepancy in next month's article. It is also notable that a small piece of coralline algae-covered live rock with a smallish colony of Protopalythoa produced as much oxygen as a medium sized tangle of Chaetomorpha sp. algae. In contrast, stagnant diffusion across the surface of the test tank's water was very ineffective in raising oxygen levels, and although the test was discontinued after one hour, I believe it unlikely that oxygen levels would approach the baseline levels for a very long time. Part of this result might be due to the test aquarium's surface area-to-volume ratio, which was much lower than that of other tanks I had tested and found to equilibrate much faster, even without circulation. Another factor that I will examine in more detail is the PAR level to which the algae were exposed (much lower in this experiment than the previous data presented at IMAC). These comparisons and statistics will also be provided in the next article.

In terms of the ten-gallon laboratory reef tanks, and despite their lack of fish, it is apparent that supersaturation of oxygen never occurs, although it is close to 100% in one case. I attribute this, as a contrast to other tanks that will be described next month, to the low irradiance provided by a single 18" fluorescent fixture not allowing for maximum photosynthesis, although a relatively high saturation in excess of 90% is testimony to the relatively low light levels capable of maintaining many corals and, together with a single water flow device, far outpacing the respiratory demands of a reasonable biomass of bacteria, algae and invertebrate species.

The fourth experiment examines the effects of light and darkness without the variable of water motion. In both tanks that are similar in many ways with some species differences and a lack of a skimmer on one tank, the effects are expected and the same in nature. Light alone in normoxic water provides enough oxygen for the community to maintain stable oxygen levels, while the community's respiration in darkness causes oxygen levels to drop. Tank 2 displayed a slight increase in oxygen in the lack of water motion, which corresponds with a similar finding under high illumination of Chaetomorpha in unstirred conditions that will be discussed next month.

The fifth experiment is similar in nature to the experiment above except that its communities are distinct, the tank volumes are tripled and the surface area-to-volume ratio is much greater. Furthermore, a less precipitous drop in terms of the shape of the curves, or even a leveling off in the case of the sand/live rock and the sand/soft coral tanks, is interesting. Because this system is exposed to ambient room light in my sunroom, despite this period being considered "night" as the system is set on a "reverse daylight" cycle, there may have been enough light present for some photosynthetically efficient organisms to produce some oxygen, leading only to a slightly lower P:R compensation point in these tanks than occurs when artificially illuminated. Alternatively, a continuing decline might have occurred had the test period been carried out over a longer duration.

Next month will conclude this series of articles, and I will show the results of other experimental tanks, shipping containers, and the dynamics of oxygen over normal and unmanipulated conditions. Finally, I will prepare graphs showing and explaining the various contributions of the variables tested in terms of their effects on the communities found in reef aquaria. I will produce several new sets of data, some of which were presented at IMAC, but will concentrate on how the major variables affecting oxygen in tanks play into an "average" aquarium running normally. I will also tease apart each variable and compare its relative importance under normoxic and hypoxic conditions. Finally, I will conclude next month's article with suggestions as to how to maintain higher oxygen levels in tanks with very low P:R ratios (i.e. heavily stocked tanks, tanks with high fish loads, tanks with low water flow or those with low surface area-to-volume ratios) and consider the implications of shipping livestock that must endure long travel times in small water volumes.