The primary goal of the present study was to gain quantitative insights into how SOC and hypoxia interact with each other, and how seasonal bottom hypoxia would affect the bacterial community richness, diversity and composition in the surface sediment of Omura Bay, a typical enclosed bay in Japan. In a previous study, Wada et al. (2012) revealed that the SOC as measured by the INT reduction method, and the bacterial community structure as revealed by ARISA at the center of Omura Bay under persistent summer hypoxia were distinct from those in the reference site (the south fringe of the bay). The results suggest the interaction of the microbes with different DO regimes would change significantly. Considering the spatio-temporal extent of hypoxia development and maintenance in Omura Bay vary within and between years (Nogami et al. 2000; Suzaki et al. 2013), it is necessary to conduct multi-year observations of hypoxia in the bay to gain detailed understanding on variations of SOC and bacterial community under bottom hypoxia. In order to achieve this goal, I modified the INT reduction method used in Wada et al. (2012) to distinguish BOC and COC relative to WSOC, and determined the R/ INT-F ratio of the surface sediment of Omura Bay. With this new technique and two different DNA finger printing methods (ARISA and dsrA-TRFLP), I examined systematically the samples collected over multiple years to gain comprehensive understanding of the
127
dynamics of sediment bacterial community structure, diversity and composition in relation to seasonal hypoxia. Here, I discuss further the findings in each chapter.
Does INT reduction method give realistic estimates for SOC in hypoxic condition?
When the R/ INT-F ratio of WSOC and WIR derived from the present study was applied to estimate SOC in Omura Bay, it fell within a range of SOC that had been reported in other coastal area (Wada et al. 2012). Therefore, it is very likely that INT reduction with sediment sample was dependent on SOC and INT reduction method gives realistic estimates for potential SOC in hypoxic condition. The INT reduction method also made it possible for the first time to demonstrate COC, CIR and sulfide concentration correlate with each other. However, the R/ INT-F ratio of COC and CIR was not directly comparable with that of SOC and WIR, because the two ratios were derived from different experimental conditions. Another drawback of the present study is the addition of unbuffered formalin into the sediment sample when measuring COC: it could have caused pH drop that may lead to an increase in COC. In the future, buffered formalin or other fixatives that would not change pH of the sediment sample should be used to estimate COC. Negative values of BOCINT found in the present study might have been ascribed to the pH effect caused by unbuffered formalin. Besides, it should be noted that INT
128
reduction may not necessary be associated with DO consumption because some reduced compounds such as glutathione, cysteine and ascorbic acid are known to reduce INT without appreciable DO consumption (Maldonado et al. 2012). For these reasons, the relative contribution of BOCINT and of COCINT to WSOCINT remains to be carefully re-examined in the future study.
How does SOC change in seasonally hypoxic coastal area?
We found that WSOCINT, COCINT and BOCINT in the center region of Omura Bay increased and peaked during hypoxia. This finding is consistent with the report by Murrel and Lehrter (2010). Accumulation of reduced compounds such as sulfides in sediment should be responsible for the increased potential SOC under hypoxia. This has implications in DO dynamics of bottom water. In Omura Bay, there have been times when strong wind caused transient re-oxygenation in the bottom water in the middle of hypoxic period (usually during July through August). However, elevated DO would soon be consumed and hence bottom-water hypoxia would be often re-formed (Nguyen et al.
2018). Similar variation of DO level was reported in other coastal environments (e.g.
Rabalais et al. 2007). Increased potential of not only COC but also BOC, can contribute to the re-formation of bottom-water hypoxia at least partly.
129
As controlling factors for SOC, DO and temperature of bottom-water can affect WSOCINT and COCINT, presumably through attenuating the activity of SRB, while DO alone can strongly affect BOCINT. Although details on how DO affected BOCINT under hypoxia are yet to be examined, increase in the metabolic activities and/or abundance of facultative anaerobic bacteria may have contributed to the increase in BOCINT.
How do BCC and diversity change in seasonally hypoxic coastal sediment?
The results of microbial community analysis with ARISA clearly demonstrate that DO availability can be a good predictor of the bacterial community diversity and composition in the surface sediment of Omura Bay. There are two major findings: (1) unimodal relationships between sediment bacterial diversity indices and DO availability, and (2) a seemingly inverse relationship between the relative abundance of Gammaproteobacteria and Deltaproteobacteria along a DO gradient, both of which have
potential implications for detection and diagnosis of changes in the sediment microbial ecosystem affected by DO decline.
In support of the findings, massive parallel sequencing of the bacterial 16SrRNA genes of the uppermost sediment samples in 2011 showed Gammaproteobacteria was less dominant in suboxic condition, while Deltaproteobacteria became dominant under
130
decreasing DO conditions. Facultative chemolithoautotorophic Woeseiaceae was the most dominant family within the Gammaproteobacterial group, which can potentially oxidize sulfides (Dyksma et al. 2016). As the bacterial OTUs tended to decrease with DO decline, they may serve as an indicator of low DO condition. Furthermore, considering the aerobic nature of these predominant sulfur oxidizing microorganisms, they can make important contribution to the potential BOC regardless of the season (Fig. 20).
How do SRB change in response to temporal variation of DO in bottom water?
SRB communities as revealed by dsrA-TRFs were significantly different between the uppermost and subsurface sediment layers in Omura Bay, regardless of the changing concentration of bottom-water DO. Nearly 70% of the total OTUs were shared between the two layers, with a few predominant in both. Based on amplicon sequencing of the bacterial 16SrDNA, members of Desulfococcus (Desulfobacteraceae) represented the dominant population of SRB in surface sediment and were likely to most strongly contribute to the stability of the bacterial community. However, a more substantial shift in the bacterial community structure is likely to be seen if the severity of hypoxia is exacerbated under an increase in water temperature associated with climate change. This would also compromise the stability of the sediment SRB community under seasonal
131 hypoxia.
Future research should involve validation of all the above-mentioned findings in different coastal areas, and integration of shifts in sediment microbial activities (respiration) and bacterial community composition into ecosystem modeling under varying DO concentrations in the water overlying the sediment in order to infer the possible consequences imposed by global trends in ocean deoxygenation.
132
Fig. 20 A conceptual diagram of sediment oxygen consumption process
133