Oxygen saturation

BRT and oxygen saturation in phytoplankton growth experiments

The measured parameter that correlates best with BRT is oxygen saturation (Table 4.2), which is in turn linked to the rate of net phytoplankton production. For all phytoplankton growth experiments (Kiel Firth water as well as monoculture experiments), correlations between oxygen saturation and BRT were highly significant (Table 4.2). For those

experiments where BRT increased significantly with increasing chlorophyll concentration and thus with increasing oxygen saturation, the absolute saturation values reveal that it was necessary for the tank water to be sufficiently supersaturated with respect to oxygen for major changes in BRT to occur. This corresponds well with the findings of Keeling (1993), who discussed that the tendency of bubbles to grow or contract depends on the

concentration of dissolved oxygen in seawater. The results showed that a certain threshold in oxygen saturation needed to be reached, ranging from 110-140% for the different experiments, until significant increases in BRT became apparent (Figure 4.3). The variation of this threshold between different phytoplankton growth experiments may be explained by several factors. First, it is likely that among the different growth experiments, bubbles of slightly different sizes were created. From the findings of Wang and Monahan (1995) and Monahan (2001) the different bubble sizes may be attributed to differences in salinity for the various experiments (see Table 4.1). The amount and composition of organic surfactants produced by the different phytoplankton species may have changed the surface curvature of the bubbles and thus their sizes to different degrees, as has been

discussed by Woolf and Thorpe (1991). If smaller bubbles were created in the tank, they should have dissolved faster than larger bubbles as described by Harris and Detsch (1991).

Therefore, smaller bubbles would need different degrees of supersaturation in order to inhibit bubble dissolution and to initiate bubble growth. Thirdly, dissolved and colloidal organic substances produced by the different algal species may have inhibited the

exchange of gas from bubbles to the surrounding water at different degrees, depending on the effectiveness of the substances to form a more or less impermeable cover on the bubble surface (Woolf, personal communication, Krägel, personal communication). The

dependency of BRT on oxygen saturation was further concluded as the maximum values of BRT temporally occurred with the maximum oxygen saturation of the tank water. Declines in BRT during the senescent phases of the phytoplankton growth cycle corresponded to decreases in oxygen saturation. The dependency of BRT on the oxygen saturation of the water becomes apparent especially for experiments 1 and 6, where, additionally to changes in oxygen saturation and BRT with increasing phytoplankton biomass, diurnal changes in BRT during photosynthesis (i.e. when the tank system was illuminated and positive net production occurred) and during respiration (no illumination) covaried strongly with diurnal oxygen saturation as indicated by the distinct peaks in the Fourier analyses of BRT at a frequency of 1 (Figures 3.7 and 3.68). The two remaining Kiel Firth water experiments (experiments 2 and 3) only revealed these diurnal changes in BRT when oxygen

supersaturation did not exceed 150%. At oxygen supersaturations > 150%, this diurnal fluctuation in BRT was not apparent (Figure 4.3). This could have been due to the production of oxygen bubbles by degassing of photosynthetically produced oxygen,

resulting in higher BRT. Furthermore, the reduction in oxygen saturation for experiments 2 and 3 during respiration was not strong enough to result in lower BRT, thus it was most likely still leading to the formation of oxygen bubbles as a result of high supersaturation. In comparison to those experiments where the tank water was supersaturated with oxygen (experiments 1, 2, 3 and 6), saturation to ~ 100% as well as undersaturation of the tank water did not result in major changes of BRT, as shown by the results of experiments 4 and 5. This is the main difference compared to the remaining growth experiments, when the tank water was highly supersaturated with oxygen and BRT changed significantly with time. From these comparisons, it becomes evident that it is a precondition for the tank water to be sufficiently supersaturated with oxygen for major changes in BRT to occur.

This can be explained by the net diffusive flux of gases across the air bubble - water interface, which is proportional to the concentration gradient driving this flux, as explained

by Woolf and Thorpe (1991), trying to achieve equilibrium between the gas pressure inside an air bubble and the surrounding water. Bowyer (1992) described that for a certain degree of gas saturation, the initial radius of a bubble is important with respect to its lifetime. If this initial radius is below a certain threshold, the bubble will collapse immediately after formation as a result of the Laplace pressure. In order for bubbles not to dissolve

immediately after their formation but to grow, gas must diffuse into the bubble from the surrounding water (Bowyer, 1992). As the exchange of gas is always from high to low concentration, gas, respectively oxygen will diffuse into the bubble only if the water is supersaturated with oxygen. This will then result in stabilisation and growth of the bubble, enabling it to reside in the water for a longer period of time. However, as stated by Bowyer and Woolf (2004) the processes of bubble gas dynamics are non-linear, and

interdependencies exist between the exchange of gas across a bubble’s surface, its resulting size and further gas exchange. If the water is undersaturated with respect to oxygen, as was the case for experiment 4 (Chaetoceros muelleri) and for most of experiment 5

(Phaeocystis), higher oxygen concentrations of atmospheric level inside the bubble lead to an adjustment of the equilibrium by diffusion of oxygen into the surrounding,

undersaturated water. This in turn, will result in rapid dissolution of small bubbles and short BRT. Nevertheless, this process may be accelerated during respiration and may proceed slightly slower during net production, thus explaining why slight day-night fluctuations are still recognisable for experiments 4 and 5.

The regression models investigating the relationship between oxygen saturation and BRT show that for some experiments, the relationship is modelled more accurately by using a quadratic regression, while for other experiments, the relationship is best described by linear regression (Figure 4.2). This finding together with the differences of the regression line slopes as well as the differences in absolute oxygen saturation and BRT for the different experiments indicate that despite the importance of oxygen saturation, BRT in phytoplankton enriched seawater is not only a function of oxygen saturation. The presence of different types and concentrations of algae most likely play an important role with respect to BRT, even though the different degrees of supersaturation of seawater with oxygen are the major influence on BRT. This becomes apparent when referring back to experiment 4 with Chaetoceros muelleri, where despite very high chlorophyll

concentrations and thus a high number of particles as well as presumably organic exudates present in the water, no major changes in BRT occurred as a result of consistent

undersaturation of the seawater with oxygen. The undersaturation despite high net

productivity during this particular experiment resulted from the very low oxygen saturations at the start of the growth phase, which, in turn, originated from the long residence time of the seawater in the tank system before the addition of the algal culture, following enhanced bacterial activity and thus oxygen consumption.

BRT and oxygen saturation in gas saturation experiments

The first gas saturation experiment showed a very important result i.e. increasing the saturation of oxygen on its own by bubbling deionised water with pure oxygen has no effect on BRT. The bubbling of deionised water in the tank system with pure oxygen has been repeated on several occasions at other times to verify this result and has never caused a significant change in BRT. From this it follows that despite the strong correlations between oxygen saturation and BRT for the phytoplankton growth experiments, oxygen saturation on its own is not the only factor responsible for the observed increases in BRT.

It demonstrates that sufficient oxygen saturation is a prerequisite for increased BRT but that the increases in BRT are linked to the presence of phytoplankton cells and/or the production of organic exudates as well. Nitrogen saturation, however, seems to have a greater influence on BRT than oxygen saturation, even though the effect of increasing the nitrogen saturation on BRT is not comparable in order of magnitude to the increases

observed during most of the phytoplankton growth experiments. However, this implies that oxygen does not play an exceptional role with respect to BRT but that it is rather the

overall gas saturation that needs to be sufficient for bubbles to grow. This is also confirmed by the bubbling with air (i.e. bubbling with both nitrogen and oxygen), where increases in BRT were noticeable but small (between 30 and 100 seconds), depending on the saturation prior to bubbling. From the results of bubbling with pure oxygen and nitrogen, it follows that most likely nitrogen saturation was the dominant factor that caused the increases in BRT when bubbling with air.

The variation of gas saturation through temperature changes produced significant increases in BRT that showed strong correlation with oxygen saturation and good agreement of results between the two (second and third gas saturation experiment) experiments.

Nitrogen saturation should have increased by the same order of magnitude as oxygen saturation, given that the initial saturations were equal, as both gases have a saturation increase of ~ 2% per °C. However, as the absolute values of oxygen saturation and thus most likely nitrogen saturation resulting from temperature increase were only slightly higher compared to the oxygen saturation values determined after bubbling with air, the

large increase in BRT cannot be attributed to increasing gas saturation alone. It is more likely a result of varying two parameters (gas saturation and temperature) simultaneously.

The reference measurements described in section 3.1 have shown that mean BRT is slightly higher for 18°C reference measurements than for 12°C reference measurements.

Coupled with rapid changes in gas saturation, these temperature differences of BRT could be more pronounced, thus accounting for the high BRT observed during the second and third saturation experiment.

Figure 4.2 Oxygen saturation versus BRT for all phytoplankton growth experiments.

The dashed lines are lines of best fit for the respective experiments. For experiments 2-6, mfBRT values were used.

Oxygen % saturation

0 50 100 150 200 250

Bubble residence time [sec]

0 100 200 300 400 500 600 700

Exp. 1 Exp. 2 Exp. 3 Exp. 4 Exp. 5 Exp. 6

1 Chapter Four - Discussio Figure 4.3 Changes in BRT (Exp. 1)/mfBRT (Exp. 2-6), chlorophyll concentration and oxygen saturation with time for all phytoplankton growth experiments.

a) Exp. No. 1; b) Exp. No. 3; c)Exp. No. 4; d) Exp. No. 2; e) Exp. No. 5; f) Exp. No.6. Dashed black line: 100% oxygen saturation mark.

Figure Fehler! Kein Text mit angegebener Formatvorlage im Dokument..1 Changes in BRT(Exp. 1)/mfBRT(Exp. 2-6), chlorophyll concentration and oxygen saturation with time for all phytoplankton growth experiments.

a)Exp. No.1; b) Exp. No.2; c)Exp. No.3; d) Exp. No.4; e) Exp. No.5; f) Exp. No.6. Dashed black line: 100% oxygen saturation mark.

b

a c

d e f

Median filtered bubble residence time [sec] Median filtered bubble residence time [sec]

Median filtered bubble residence time [sec] Median filtered bubble residence time [sec] Median filtered bubble residence time [sec]

Figure 4.3 Changes in BRT(Exp. 1)/mfBRT(Exp. 2-6), chlorophyll concentration and oxygen saturation with time for all phytoplankton growth experiments.

a)Exp. No.1; b) Exp. No.2; c)Exp. No.3; d) Exp. No.4; e) Exp. No.5; f) Exp. No.6. Dashed black line: 100% oxygen saturation mark.