Interactive comment on “ Net community production and stoichiometry of nutrient consumption in a pelagic ecosystem of a northern high latitude fjord : mesocosm CO 2 perturbation study ”

This study reports the results of a large-scale mesocosm experiment on ocean acidification carried out in Kongsfjorden, West Spitsbergen, presenting the effects of CO2 perturbations on net community production (NCP) and stoichiometry of nutrient consumption. This is a valuable study providing rare information at the community level of a high latitude ecosystem in response to CO2 increases. The manuscript is well organized and written, however, it does not provide an integration of the information about the response of this system to CO2 perturbations.


Conclusions References
Tables Figures

Introduction
The Arctic Ocean is a key player in the global cycling of carbon (e.g. Bates et al., 2009), and the Arctic shelves are currently amongst the most productive areas in the Introduction

Conclusions References
Tables Figures
The largest ocean acidification signal in the world oceans is projected to occur in Arctic surface waters (Steinacher et al., 2009).Undersaturation with respect to aragonite is found already in the surface waters of the Canada basin (Yamamoto-Kawai, 2009).Model studies show that the Arctic Ocean may become entirely undersaturated with respect to aragonite already by 2050 (Andersson et al., 2010).Experiments performed on different marine ecosystems with different CO 2 gradients (and thus pH and saturation states gradients) indicate potentially deleterious effects of ocean acidification on marine calcifying organisms (e.g.coccolithophores; Riebesell et al., 2000;Orr et al., 2005;Ridgewell et al., 2009), and on organisms at higher trophic levels (e.g.Comeau et al., 2009;Lischka et al., 2011;Frommel et al., 2011).
Increasing inorganic carbon concentrations have also been shown to promote primary production and carbon assimilation in marine phytoplankton (Hein and Sand-Jensen, 1997;Engel et al., 2008;Tortell et al., 2008).Increasing carbon uptake could cause a shift towards ecosystems with higher carbon-to-nutrient utilization ratios (Riebesell et al., 2008;Bellerby et al., 2008).Model studies show that by consuming more carbon in the surface layer, marine phytoplankton may potentially reduce atmospheric pCO 2 globally by 20 % (Schneider et al., 2004).However, the ecosystem of the Arctic Ocean is dynamic in terms of production and respiration of organic matter.Therefore pelagic system will not necessarily act as sink of atmospheric CO 2 (Borges et al., 2005;Thomas et al., 2005).Introduction

Conclusions References
Tables Figures

Back Close
Full Despite the hypothesis that Arctic marine ecosystem may experience the greatest changes under ocean acidification (e.g.Merico et al., 2006), no mesocosm studies testing that hypothesis have, to the best of our knowledge, been conducted in the northern high latitudes.This paper presents results from the first large-scale mesocosm experiment on ocean acidification conducted in the Arctic.Here we present the net community production (NCP) estimated from a net biological uptake of inorganic carbon and the stoichiometry of inorganic carbon to nutrient consumption in an ecosystem of a high latitude fjord under a range of pCO 2 levels from 175 to 1085 µatm.The effect of CO 2 perturbation on NCP and stoichiometry of element utilization varied through the course of the experiment, thus our results emphasize the variability in the coupling between autotrophic and heterotrophic processes.

Study area
The experiment took place in Kongsfjorden (78 • 56.2 N, 11 • 53.6 E, Fig. 1), West Spitsbergen, Svalbard archipelago.The fjord is 26 km long and its width varies from 4 to 10 km.The water in Kongsfjorden is characterized by a mixture of Arctic water masses (which are transported by the coastal current flowing from the Barents Sea over the West Spitsbergen Shelf), Atlantic water masses (West Spitsbergen current), and freshwater input from calving and melting glaciers as well as precipitation (Svendsen, 2002;Hop et al., 2006).In winter the hydrography is dominated by Arctic water masses and in summer it is under Atlantic dominance (Svendsen, 2002).The dominant water mass may determine the structure of the microbial community, thus influencing the biogeochemical cycling in the fjord.Heterotrophic bacteria, picoplankton and nanoflagellates contribute to ecosystem structure and functioning in all seasons.Therefore, in nutrientlimited post-bloom conditions, there is an efficient microbial loop (Rokkan Iversen and Introduction

Conclusions References
Tables Figures

Back Close
Full Seuthe, 2011) that provides inorganic nutrients to phytoplankton and bacteria through rapid organic matter remineralization.

Experimental set up
Nine mesocosm bags two m in diameter and 17 m long were deployed in Kongsfjorden in late May of 2010 (for details see Riebesell et al., 2012).Briefly, bags attached to hard floating frames, were made of TPU (thermoplastic polyurethane), which transmits ∼ 95 % of the incoming solar radiation.The mesocosms enclosed ∼ 45 m 3 of fjord water with its natural phyto-ans bacteriaplankton assemblages.After closing the mesocosms at the bottom, there was no exchange with the ambient fjord water.The bottom plate was made of polycarbonate.A hood made of transparent PVC (polyvinyl chloride) plastic was mounted on the top of the floating rafts to minimize precipitation and birds' excrement input into the systems.
The experiment lasted 31 days, from 7 June (day t0) to 7 July (day t30).A CO 2 addition was implemented in four steps (Schulz et al., 2012).Filtered seawater, enriched with CO 2 was injected into the mesocosms and evenly distributed throughout the water column.Exchange of CO 2 -enriched water with unperturbed water in a "dead" volume underneath the sediment traps caused an initial decline in pCO 2 levels until day t8.At this stage, pCO 2 in the mesocosms covered a range from 185 to 1420 µatm (Bellerby et al., 2012).Table 1 shows mean pCO 2 and pH values in seven perturbed (M1, M2, M4, M5, M6, M8, M9) and two control mesocosms (M3, M7) for different periods of the experiment that followed peaks of biomass growth: Phase I (t4-t13), Phase II (t14-t21), Phase III (t22-t27), post-nutrient period t14-t27, and total period of the experiment t8-t27.
Nutrients were added to mesocosms on experimental day t13 to stimulate a phytoplankton bloom.Water samples were collected daily using a 5 l depth-integrated sampler lowered down to 12 m.A detailed description of the experimental set up can be found in Riebesell et al. (2012), Czerny et al. (2012) and Schulz et al. (2012).Introduction

Conclusions References
Tables Figures

Back Close
Full

Data
To estimate net community production (NCP) and the stoichiometric rates of carbon to nutrient utilization, we used measurements of total inorganic carbon concentration (CT), determined by coulometry; total alkalinity (AT), determined by Gran titration (Gran, 1952;Bellerby et al., 2012); inorganic nutrient concentrations (phosphate - (Schulz et al., 2012); air/sea gas exchange (CO 2(ex.) ) (estimated by measured loss of N 2 O added to the mesocosms as a deliberate tracer) (Czerny et al., 2012).We also show temporal evolution of chlorophyll a concentration, measured fluorometrically according to Welschmeyer (1994) and Schulz et al. (2012).

Net community production derived from changes in CT concentration
To estimate the net effect of CT utilization by phytoplankton during photosynthesis and CT release due to bacterial degradation of organic matter, we calculated NCP with the method previously employed in the PeECE mesocosm studies (Delille et al., 2005;Bellerby et al., 2008).
The change in CT concentration was corrected for the air/sea gas exchange (Eq.2).
CT corrected = CT measured − CO 2(ex.)Corrected AT and CT concentrations were normalized to salinity to account for evaporation from the first day of each phase (Eqs.3 and 4) (Schulz et al., 2012).
where S is salinity, x n and x 1 are corresponding to day n and day 1 respectively, of the time period for which AT and CT are normalized.
The net community calcification NCC was estimated as cumulative change in AT norm.(Eq.5).
Calcification was insignificant during the experiment, therefore calculated NCC expresses the precision of AT measurements, which was < 2 µmol kg −1 (Bellerby et al., 2012).Net community production is computed as a cumulative change in CT norm., accounting for cumulative change in AT norm.(Eq. 6).

Statistical analysis
To test the difference in NCP between CO 2 scenarios, values of a cumulative NCP calculated for every time period, were plotted against mean pCO 2 values for t8-t27 period.Additionally we plotted the ratio of values of a cumulative NCP to a cumulative Introduction

Conclusions References
Tables Figures

Back Close
Full difference in nitrogen and phosphate, calculated for every time period, against mean pCO 2 values for t8-t27 period to test the difference in C : N and C : P utilization ratios among pCO 2 scenarios.Linear regression analysis (F-test) was applied to identify the effect of pCO 2 treatment on NCP, C : N and C : P utilization ratios.Regression coefficients, slopes, R 2 and p-values derived from this analysis are shown in Table 2.
To calculate C : N and C : P utilization ratios we plotted cumulative NCP against a cumulative difference in nitrogen and phosphate uptakes for every period in every mesocosm.Linear regression analysis (F-test) was applied to estimate a ratio of carbon to nitrogen and phosphate consumption.Regression coefficients, R 2 , and p-values derived from this analysis are shown in Table 3 for C : N uptake ratio and in Table 4 for C : P uptake ratio.Analyses were performed with a Statistics toolbox in Matlab.

Net community production
The initial pCO 2 of the ambient water in the fjord was ∼ 170 µatm, corresponding to a pH of ∼ 8.3 (Bellerby et al., 2012).Concentrations of nitrate and phosphate in the water were close to detection limit at the beginning of the experiment (0.11 µmol kg −1 for nitrate, 0.13 µmol kg −1 for phosphate).Concentration of ammonia was 0.7 µmol kg −1 (Schulz et al., 2012).Additionally, there were 5.5 µmol kg −1 of dissolved organic nitrogen, 0.20 µmol kg −1 of dissolved organic phosphorus (Schulz et al., 2012) and 75.0 µmol kg −1 of dissolved organic carbon (Engel et al., 2012).Reduced pCO 2 and inorganic nutrient concentrations, and increased concentrations of organic carbon, nitrogen and phosphorus indicated a post-bloom situation in the fjord at the start of the experiment.
The concentration of chlorophyll a increased three times during the course of experiment in the form of three bloom events (Fig. 2).Peak concentrations were observed on days t6, t19 and t27 (Riebesell et al., 2012).According to observed peaks, the results Introduction

Conclusions References
Tables Figures

Back Close
Full of the experiment are presented for three phases: Phase I from t4 to t13; Phase II from t14 to t21; Phase III from t22 to t27.The CO 2 added between t1 and t4 was found to have equilibrated with the water in the dead volume below the sediment traps by t8.Therefore we discuss the NCP of Phase I only from t8 to t13 (Fig. 3a and b).
Additionally we show some results for the total period of the experiment t8-t27, and for the post-nutrient period t14-t27.
Our approach implies evaluating the net utilization of total carbon (CT) accounting for the net change in total alkalinity and CO 2 gas exchange through the air/water interface.Net community production (NCP) is the result of community production and respiration.We observed NCP variability among three phases of the experiment and between the nine CO 2 scenarios.Overall cumulative NCP, calculated for the period t8-t27 showed similar values in all mesocosms -∼ 50 ± 5 µmol kg −1 (Fig. 3a).However in the different phases NCP varied among different pCO 2 scenarios (Figs.3b and 4).In Phase I NCP was positive in mesocosms with high and intermediate CO 2 levels (cumulative NCP on t13 in high ∼ 6.1 ± 1.5 µmol kg −1 and intermediate ∼ 2.8 ± 1.4 µmol kg −1 ; Fig. 4), indicative for an autotrophic system.Cumulative NCP in mesocosms with low CO 2 treatment was close to zero by t13, indicating that autotrophic and heterotrophic processes were evenly active (∼ −0.2±0.9 µmol kg −1 ).In Phase II all mesocosms were autotrophic and had positive values of NCP.The highest mean cumulative NCP by t21 was in mesocosms with high CO 2 treatment -∼ 13.9 ± 4.3 µmol kg −1 .In mesocosms with intermediate and low CO 2 treatments mean cumulative NCP was lower: in intermediate -∼ 10.3 ± 3.9 µmol kg −1 , in low -∼ 8.9 ± 0.9 µmol kg −1 .
In Phase III all mesocosms were autotrophic with the highest cumulative NCP rates during the experiment.The highest NCP in Phase III was in mesocosms with low (34.4 ± 1.7 µmol kg −1 ) and intermediate CO 2 levels (31.4 ± 6.2 µmol kg −1 ).In mesocosms with high CO 2 level cumulative NCP was only ∼ 19.2 ± 3.2 µmol kg −1 .
Regression analysis (F-test) of a cumulative NCP calculated for different time periods against mean pCO 2 values calculated for t8-t27 periods showed a positive response of NCP to pCO 2 in Phase I (p = 0.000) and in Phase II (p = 0.076) (Table 2, Fig. 5).A Introduction

Conclusions References
Tables Figures

Back Close
Full negative response of NCP to pCO 2 was observed in Phase III (p = 0.000) and for the total period t8-t27 (p = 0.094).

Stoichiometry of carbon-to-nutrients consumption
Since the concentrations of inorganic nutrients in Phase I (prior nutrients addition) were low, we evaluated the stoichiometry of carbon uptake (in this study shown as NCP) to inorganic nitrogen uptake (presented as sum of nitrate, nitrite and ammonium) (hereafter C : N uptake ratio) and inorganic phosphate uptake (hereafter C : P uptake ratio) only for Phase II, Phase III (Figs. 6 and 7) and additionally for the post nutrient period t14-t27.C : N and C : P uptake ratios in period t14-t27 were close to Redfield values, and ratios varied among pCO 2 treatments.A mean C : N uptake ratio in mesocosms with low CO 2 level was ∼ 8.9, while it was ∼ 8.7 in intermediate CO 2 level and ∼ 6.6 in high CO 2 level (Table 3).Thus the negative trend of C : N uptake ratio with increasing pCO 2 was pronounced in t14-t27 period (Fig. 8a).Similar to C : N uptake ratio, C : P uptake ratio was close to Redfield, and varied amongst CO 2 treatments.Mean C : P uptake ratio was ∼ 136.3 in mesocosms with low pCO 2 level, ∼ 127.3 in mesocosms with intermediate pCO 2 level and ∼ 93.0 in mesocosms with high CO 2 level (Table 4).
A negative response of C : P uptake ratio to CO 2 treatment was observed in the post nutrient period t14-t27 (Fig. 8b).
Values of C : N uptake ratio in Phase II were similar and lower than Redfield ratio (6.6) in all mesocosms (Fig. 6a-c).C : N uptake ratio was on average 4.01 in mesocosms with low, 4.0 in mesocosms with intermediate and 4.2 in mesocosms with high CO 2 treatment (Table 3).C : P uptake ratio was two times lower than Redfield value (106) in Phase II (Fig. 7a-c).Mean C : P uptake ratio in Phase II did not show a clear response to CO 2 treatment, and was 62.0 in mesocosms with low, 54.6 in mesocosms with intermediate and 55.3 in mesocosms with high CO 2 treatment (Table 4, Fig. 8b).
In contrast to Phase II C : N uptake ratio in Phase III was slightly higher than the Redfield value (Fig. 6d-f).The mean C : N uptake ratio was 8.8 in mesocosms with Introduction

Conclusions References
Tables Figures

Back Close
Full low and intermediate CO 2 treatments, but only 6.7 in mesocosms with high CO 2 treatment (Table 3).Thus the C : N uptake ratio had a slight negative trend with increasing pCO 2 (Table 2).In Phase III C : P uptake ratio was higher than Redfield (Fig. 7d-f) and showed a negative response to CO 2 perturbation (Fig. 8b).C : P uptake ratio decreased from 136.0 in low, to 127.0 in intermediate and 93.0 in high CO 2 treatment (Table 4).

Summary
Cumulative NCP was similar in all CO 2 scenarios by t27, but varied among CO 2 scenarios in different experimental phases.For the major part of experiment the pelagic system in the mesocosms was autotrophic, but in Phase I mesocosms with low CO 2 levels were close to heterotrophic.The highest cumulative NCP rates were observed in Phase III.The cumulative NCP was increasing with higher CO 2 in Phases I and II.The mean cumulative NCP was decreasing with higher CO 2 in Phase III, which caused negative response of NCP to CO 2 in total experimental period t8-t27.C : N and C : P uptake ratios were lower than Redfield in Phase II, higher than Redfield in Phase III, and close to Redfield in post-nutrient t14-t27 period.The mean C : N and C : P uptake ratios were decreasing with higher CO 2 treatment in post-nutrient t14-t27 period.In Phase II the mean C : N and C : P uptake ratios were slightly increasing with higher CO 2 .In Phase III, C : N and C : P uptake ratios were decreasing with higher CO 2 .

Discussion
The overall effect of CO 2 treatment on net community production and net stoichiometry of nutrient and CT utilization during the experiment on ocean acidification in the high northern latitude fjord was not clear.A distinct succession in phytoplankton groups occurred over the experimental period (Schulz et al., 2012;Brussaard et al., 2012).The Introduction

Conclusions References
Tables Figures

Back Close
Full rates of net community production and net stoichiometry varied over the experimental period and most likely reflected the specific sensitivities of the dominant phytoplankton groups to CO 2 perturbation.The experiment in Kongsfjorden started after the spring phytoplankton bloom.Dissolved organic nutrients and carbon were the substrate for a recycling system in the fjord.Due to an efficient microbial loop, inorganic nutrients and CT were resupplied to the water column by heterotrophs.In Phase I, concentrations of mineral nutrients were low, suggesting a tight coupling between the release of nutrients from organic matter and their rapid utilization in primary production.Therefore, the first chlorophyll a peak occurring in Phase I was most likely fueled by a phosphate remineralized from organic matter and most importantly ammonium as N-source, remaining after the spring bloom.Peaks of phytoplankton growth similar to the one observed in Phase I have been described as common events in the fjord during the summer (Hop et al., 2006).In Phase I after t8 the concentration of chlorophyll a was already declining, which was probably caused by grazing (Niehoff et al., 2012) or viral infection (Brussaard et al., 2012).However, we observed net autotrophy in the mesocosms with high CO 2 treatment, while mesocosms with intermediate and low CO 2 treatment were net heterotrophic.Production rates, which were higher than respiration rates in mesocosms with high CO 2 treatment, were stimulated by CO 2 perturbation (Engel et al., 2012).
Nutrient addition on t13 stimulated biomass growth in all mesocosms from t15 peaking at t19.In Phase II that followed nutrient addition, we observed net autotrophy in all mesocosms.NCP was increasing with higher CO 2 treatment.Nutrients were consumed at higher rates with increasing CO 2 (Schulz et al., 2012).Both increased NCP and nutrient uptake rates in mesocosms with higher CO 2 treatment resulted in similar C : N and C : P ratios among pCO 2 treatments.Nevertheless, stoichiometric uptake ratios were lower than Redfield in Phase II.This may be the result of luxury consumption of inorganic nutrient that followed nutrient addition on t13.It is also possible that through the microbial loop N and P were utilized, but CT was released to the system due to organic carbon respiration (Thingstad et al., 2008).This result shows that along with variable Introduction

Conclusions References
Tables Figures

Back Close
Full elemental stoichiometry (e.g.Sterner and Elser 2002) net uptake stoichiometry of carbon and nutrients during photosynthetic fixation and bacterial respiration varies in time and regionally.The decline of phytoplankton production after t19 was probably caused by viral infection (Brussaard et al., 2012).The decline was the end of Phase II.On t22 chlorophyll a concentration increased for the third time during the experiment, peaking on t27 (Schulz et al., 2012).Although the majority of inorganic nutrients were consumed during Phase II, we observed the highest rates of NCP in Phase III after t22.C : N and C : P uptake ratios were higher than Redfield in Phase III, because of high NCP rates and low nutrients concentrations towards the end of experiment.
Despite variable NCP response to CO 2 treatment in different phases, cumulative carbon uptake was similar in all mesocosms by the last day of the experiment.C : N and C : P uptake ratios during the post-nutrient period (Phase II + Phase III) were close to Redfieldian, and decreased from low to high pCO 2 levels.
Overall trends of net community production estimated in this study are in good agreement with a temporal evolution of chlorophyll a (Schulz et al., 2012).Net community production estimated from CT uptake, and its negative response to pCO 2 treatment in Phase III are consistent with NCP estimated by oxygen consumption and release (Tanaka et al., 2012), and with NCP estimated by net 13 C-POC production (de Kluijver et al., 2012).Despite the similarity, the results should be taken with caution, because in low and intermediate CO 2 treatments the bloom was peaking on t27 with the following decline.In high CO 2 treatment, however, the bloom was still developing on t27, allowing for different carbon-to-nutrient coupling and, hence, different NCP than it was in low and intermediate CO 2 treatments.Engel et al. ( 2012) observed higher production and release of dissolved organic carbon in mesocosms with high CO 2 treatment.Thus there was a stimulated bacterial growth, which was leading the system of high CO 2 mesocosms to a heterotrophic state.Thus, NCP dynamics in Phase III was based on biogeochemical conditions resulted from Phase II.Figures

Back Close
Full

Conclusions
We studied net community production and the stoichiometry of carbon and nutrient consumption during a mesocosm experiment to test the effect of CO 2 perturbation on the marine pelagic ecosystem in a high latitude Arctic fjord.This study emphasizes the importance to account for the presence of a strong microbial loop in the Arctic waters, which leads to a high plasticity in production and stoichiometry of uptake ratios.Thus the observed NCP, C : N and C : P uptake ratios varied between different phases of the experiment indicating species-specific responses to pCO 2 treatment.Overall cumulative net community production, however, was similar in all mesocosms by the last day of experiment.Mean C : N and C : P uptake ratios in the post nutrient period were close to Redfield ratio although were different from it when considering different phases.Biogeochemical conditions established in the fjord in the beginning of the experiment played an important role in the community response to experimental treatment.Similarly to that, biogeochemical conditions resulted in a community response in a certain phase, affected the community response in the following phase.This observation may lead to a conclusion, that microbial community is adapting to a treatment to a certain degree, affecting following bloom stages.The evidence of a strong microbial loop in the Arctic Ocean in this study strengthens the recommendation to implement the microbial loop into regional biogeochemical models for more reliable projections of the future of a polar marine ecosystem and its feedback to a changing climate and ocean acidification.Introduction

Conclusions References
Tables Figures
Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper |

Fig. 1 .Fig. 2 .
Fig. 1.Map of the Arctic Ocean with the Svalbard archipelago highlighted in red, and enlarged map of the latter with a red square indicating the location of Kongsfjorden.

Fig. 3 .Fig. 4 .Fig. 5 .Fig. 6 .Fig. 7 .Fig. 8 .
Fig. 3. (A) Cumulative net community production for the total period of the experiment; (B) cumulative net community production in every phase of the experiment.Horizontal dashed line on both figures shows the border between heterotrophic (below 0) and autotrophic (above 0) systems.Line colors and numbers in a legend are as described for Fig. 2.

Table 2 .
Dependence of NCP, C : N and C : P uptake ratios on pCO 2 (F-test).

Table 3 .
Dependence of NCP on N consumption (C : N uptake ratio, F-test).