- Email: [email protected]
Origin of particulate organic carbon in the upper St. Lawrence: isotopic constraints
ELSEVIER
Earth and Planetary Science Letters 162 (1998) 111–121
Origin of particulate organic carbon in the upper St. Lawrence: isotopic constraints Johannes A.C. Barth a,Ł , Ja´n Veizer a,b , Bernhard Mayer b,1 a Ottawa-Carleton b
Geoscience Centre, University of Ottawa, Ottawa, Ontario K1N 6N5, Canada Institut fu¨r Geologie, Ruhr-Universita¨t Bochum, D-44780 Bochum, Germany
Received 17 November 1997; revised version received 10 July 1998; accepted 10 July 1998
Abstract Seven sampling locations in the upper St. Lawrence River near the city of Cornwall (Ontario, Canada), including the main river and six near-shore ecosystems (a creek, embayments and a wetland) were studied in order to determine the origin of particulate organic carbon. Parameters studied included chlorophyll-a (chl-a), particulate organic carbon (POC) and dissolved inorganic carbon (DIC), as well as the isotopic compositions of the latter two (δ13 CPOC , δ13 CDIC ). The results show that in situ photosynthesis and detrital inputs are both significant contributors to the POC pool in the isolated embayments. The former dominates during warm seasons, with POC concentrations up to 2663 µg=l and chl-a concentrations up to 26.1 µg=l. Near-shore ecosystems have a wide range of δ13 CPOC values ( 31.5 to 16.3‰), but this variability is not reflected in the ‘Main Channel’. There, the δ13 CPOC signal is uniformly close to 27‰, in accord with estimates from earlier studies on the river’s estuary. This suggests that the POC contribution from near-shore ecosystems is minor. Although the ‘Main Channel’ has low chl-a concentrations, model calculations suggest that most of its POC originates from photosynthetic activity, probably within the Great Lakes. 1998 Elsevier Science B.V. All rights reserved. Keywords: St. Lawrence River; organic carbon; stable isotopes
1. Introduction The study of organic carbon in rivers provides important information on the recycling of continental carbon [1]. Riverine particulate organic carbon (POC) consists partly of detrital compounds, but can also be a product of in situ phytoplanktonic activity [2]. Understanding the extent of in situ production Ł Corresponding
author. Fax: C1 613 562 5192; E-mail: [email protected] 1 Present address: Departments of Physics=Astronomy and Geology=Geophysics, University of Calgary, Calgary, Alberta T2N 1N4, Canada.
of POC is essential, because it influences the aqueous carbon cycle by consuming dissolved inorganic carbon (DIC). For the St. Lawrence River, the particulate organic carbon (POC) has been studied in detail, but only in the estuary [3–7]. In these studies, the nature of the upstream riverine POC was conjectured to be mainly of terrestrial origin and to have isotopic compositions between 24.1 and 27‰. As yet no measurements on the isotopic composition of the upper St. Lawrence POC pool have been available. The present study provides the first overview of this parameter from a monitoring of seven aquatic ecosystems near the city of Cornwall (Ontario, Canada), ¾150
0012-821X/98/$19.00 1998 Elsevier Science B.V. All rights reserved. PII S 0 0 1 2 - 8 2 1 X ( 9 8 ) 0 0 1 6 0 - 5
112
J.A.C. Barth et al. / Earth and Planetary Science Letters 162 (1998) 111–121
Fig. 1. Location map of the study area. Circles are proportional to the annual average concentrations of particulate organic carbon in the ecosystems studied.
km upstream of Montre´al (Fig. 1). The ecosystems studied include the center of the St. Lawrence River (‘Main Channel’), several of its embayments (‘Bay at Cooper Marsh’, ‘Long Sault Island’, ‘Whale Island’ and ‘Hoople Bay’), a creek that drains into the St. Lawrence (‘Hoople Creek’) and a wetland (‘Cooper Marsh’). The objectives of this research were: (1) to evaluate the relative importance of phytoplanktonic POC production versus detrital influx, seasonally and spatially, among the studied ecosystems; and (2) to test the proposition, based on estuarine studies, that the St. Lawrence is characterized by detrital POC. The St. Lawrence is particularly suitable for this type of study, because its average discharge of 4ð106 tonnes of suspended sediments per year ranks lowest among the world’s major rivers [8]. For comparison, the Mississippi has an average suspended load of 210 ð 106 t=yr [8]. The low suspended sediment load of the St. Lawrence River implies small POC
concentrations, minimizing the masking effect of the usually large allochthonous component on the overall balance of the riverine POC. Our study area is typical of the upstream parts of the St. Lawrence River, because at this point the water originates directly from Lake Ontario without any influence from tributaries. As a result, the ‘Main Channel’ sampling station is characterized by well mixed POC from the Great Lakes and can thus serve as a baseline for comparison with the other ecosystems studied.
2. Material and methods Seven stations in the vicinity of Cornwall (Fig. 1) were sampled repeatedly between June, 1994, and April, 1996. Samples were collected 2 m below the water surface with a 2-l narrow-mouth bottle attached to a metal weight. The sampling stations ‘Hoople Creek’, ‘Long Sault Island’ and ‘Cooper Marsh’ had
J.A.C. Barth et al. / Earth and Planetary Science Letters 162 (1998) 111–121
depths less than 2 m, allowing sample collection only at 0.5 to 0.75 m below the water surface. Water temperature, pH, and alkalinity were measured in situ in order to calculate the speciation of the DIC. The pH was determined with a Hach pH meter which was calibrated before every measurement with two standard buffer solutions. The 2¦ precision of š0.03 pH units was determined from multiple measurements of standards in the field and laboratory. The alkalinity titrations were performed with a Hach digital titrator on 100 ml of sample using 1.6 N sulfuric acid, the results being expressed in µM HCO3 . Repeat measurements of standards at various concentrations yielded 2σ values of š24.6 µM HCO3 , matching the recommended precision of the Compilation of EPA’s Sampling and Analysis Methods [9]. Together with routinely determined major anion and cation concentrations (Ca2C , Mg2C , NaC , KC , Cl , SO24 ) [10], the field parameters were then entered into the program Hydrowin [11] for calculation of CO2 (aq) activities (Appendix A). The samples for the δ13 CDIC measurements were filtered through 0.45-µm cellulose acetate filter papers, preserved with HgCl2 and stored at 4ºC until analysis. At the G.G. Hatch Isotope Laboratories, University of Ottawa (Canada), the DIC was converted to CO2 by acidification with 85% H3 PO4 and then purified under vacuum. The isotopic composition of the extracted CO2 was then determined on a triple collector VG SIRA 12 mass spectrometer. The results are expressed with reference to the Pee Dee Belemnite standard (VPDB), with an overall analytical precision (2¦ ) of š0.3‰. Samples for POC were collected on Whatman GF=C 2000 glass microfibre filter papers which had been prepared by overnight heating at 500ºC in order to remove organic carbon contaminations. After filtration, the samples were frozen at 20ºC until analysis. Prior to analysis, the filtered material was acidified with 10% HCl in order to eliminate any particulate inorganic carbon, then dried at 70ºC overnight. The suspended material and the filter paper could not be separated and therefore were ground up and analyzed together. The POC content and δ13 CPOC measurements were performed at the Isotope Laboratory of the Ruhr-Universita¨t, Bochum (Germany) by continuous flow isotope ratio mass spectrometry (CF-IRMS), using a Carlo Erba ele-
113
mental analyzer (CE 1110) connected to a Finnigan MAT mass spectrometer (delta C) via a ConFlo II interface. The results obtained reflected filter paper and sample material, and the POC contents were therefore calculated by subtracting average readings from multiple weighings of blank filter papers. The δ13 CPOC was then determined by mass and isotope balance calculations. The precisions for the C-content and its δ13 C value were determined from duplicate and triplicate measurements of samples as well as from multiple measurements of blanks. The 2σ precisions were š20 µg=l and š0.5‰ VPDB for the POC and the δ13 CPOC measurements, respectively. The suspended material for the chl-a analyses was collected on Whatman GF=C 2000 glass microfibre filter papers and frozen immediately after filtration at 20ºC. In the laboratory (Department of Biology, University of Ottawa), chl-a was extracted from the filter paper by leaching with concentrated dimethyl sulfoxide (DMSO) at 60ºC, and rinsing with 90% acetone [12]. Subsequently, the chl-a contents of the leachates were determined with a Pye Unicam double beam spectrophotometer (model SP8-100), following the guidelines of Jeffrey and Humphrey [13]. The chl-a content was then calculated utilizing the formula in Wetzel and Likens [14]. The precision of the chl-a measurements (2¦ ) was š0.18 µg=l, as determined from multiple measurements of standards.
3. Results The measured POC and chl-a concentrations range from 64 to 2663 and 0.3 to 26.1 µg=l, respectively (Table 1). Samples with high chl-a concentrations also have high POC contents, with the ‘Main Channel’ waters representing the lowest endmember and the ‘Hoople Bay’ and ‘Long Sault Island’ embayments having values up to an order of magnitude higher (Fig. 2). The waters of ‘Hoople Creek’, ‘Whale Island’, ‘Bay at Cooper Marsh’ and ‘Cooper Marsh’ are intermediate between the ‘Main Channel’ and the embayments in terms of their POC and chl-a levels. The isotopic compositions of the POC (δ13 CPOC ) range from 31.5 to 16.3‰, while those of the dissolved inorganic carbon (δ13 CDIC ) range between
114
J.A.C. Barth et al. / Earth and Planetary Science Letters 162 (1998) 111–121
Table 1 List of parameters studied Date
Chl-a (µg=l)
POC (µg=l)
POC=chl-a –
δ13 CPOC (‰)
δ13 CDIC (‰)
aHCO3 (µM)
aCO3 (µM)
aCO2 (aq) (µM)
DIC (µM)
Bay at Cooper Marsh 6=12=95 3.0 8=1=95 1.3 10=10=95 0.8 4=19=96 1.5
277 105 102 91
91.6 81.3 122.6 62.3
26.5 21.8 25.9 25.3
2.0 0.6 0.1 0.9
1711 1576 1688 1690
32 39 18 10
9 7 15 25
1752 1622 1721 1726
Cooper Marsh 8=14=95 10=18=95 12=6=95 4=19=96
4.3 8.6 4.8 2.3
391 711 964 361
90.7 82.6 202.7 159.6
25.7 31.4 27.0 30.9
5.6 9.0 10.6 11.3
1587 1587 2090 1955
9 2 1 3
30 156 369 115
1625 1745 2460 2073
10.7 11.2 9.3 23.3 6.5 8.9
1095 832 1578 1056 415 2107
102.7 74.6 170.0 45.3 63.7 236.2
28.2 30.0 27.3 24.7 27.3 28.5
5.8 1.4 1.2 1.1 10.3 11.0
2028 2386 1804 1713 2921 2430
13 21 19 30 6 12
31 27 18 9 136 47
2072 2434 1841 1752 3063 2489
4.0 2.6 1.6 1.4 1.2 3.4
211 747 173 162 112 1087
52.7 283.3 108.3 112.9 95.5 320.1
31.5 27.4 29.8 28.5 31.3 24.7
13.3 12.0 10.5 10.5 13.7 12.0
2622 3303 2329 2298 3358 2583
13 51 18 12 3 5
52 22 32 43 385 116
2688 3376 2378 2353 3745 2705
Long Sault Island 8=1=95 26.1 10=10=95 4.3
2663 429
102.0 100.3
25.4 21.4
1.3 1.2
1806 1882
18 7
19 52
1843 1940
Hoople Bay 5=30=95 6=13=95 8=2=95 10=17=95 12=5=95 4=21=96 Hoople Creek 5=30=95 6=14=95 8=2=95 10=11=95 12=4=95 4=21=96
Main Channel 7=14=94 11=3=94 5=29=95 10=17=95 12=4=95
0.7 0.3 1.1 1.1 0.7
102 95 64 76 99
136.6 334.7 59.1 66.2 141.7
27.4 26.7 30.0 28.5 26.3
1.3 1.8 1.0 0.3 0.1
1500 1654 1559 1689 1634
13 13 8 16 16
18 20 30 17 15
1531 1687 1597 1722 1665
Whale Island 5=29=95 6=13=95 8=1=95 10=10=95 12=5=95 4=20=96
2.0 1.2 0.8 2.1 1.0 2.8
264 162 274 226 264 245
135.4 136.6 340.0 110.1 261.1 87.2
25.7 27.0 26.3 22.4 16.3 27.0
1.3 2.6 0.4 0.2 0.8 4.8
1699 1712 1749 1579 1724 1762
7 30 14 20 9 19
42 10 23 12 29 15
1747 1752 1786 1611 1762 1796
13.7 and C0.3‰ (Table 1). Although these two parameters yield only a somewhat diffuse spatial pattern, the samples from the ‘Hoople Creek’ and ‘Cooper Marsh’ ecosystems show the most negative δ13 CPOC and δ13 CDIC values, indicating that these parameters are somewhat correlated (Table 1).
The CO2 (aq) activities of most water samples scatter around a mean of 20 µM, reflecting the fact that the ‘Main Channel’ provides the bulk of the water to most of the studied ecosystems. Only the near-shore ecosystems ‘Hoople Creek’, ‘Hoople Bay’ and ‘Cooper Marsh’ deviated from this mean,
J.A.C. Barth et al. / Earth and Planetary Science Letters 162 (1998) 111–121
115
Fig. 2. Temporal distributions of chlorophyll-a contents and particulate organic carbon concentrations in the seven ecosystems studied.
reaching peak CO2 (aq) activities of 385 µM (Table 1), an observation that will be of importance in the discussion of isotope discrimination between CO2 (aq) and phytoplanktonic POC.
4. Discussion 4.1. Sources of particulate organic carbon Particulate organic carbon represents a variable mixture of living and dead phytoplankton and other
components, such as detritus, bacteria and zooplankton [15,16]. Among all these components the major sources of riverine POC are (1) detrital matter, and (2) in situ produced phytoplankton. The latter changes the aqueous CO2 content and the 13 C=12 C ratio of dissolved inorganic carbon and can therefore significantly influence the riverine carbon cycle. Although the proportion of phytoplankton in the POC pool is not easily quantified, it can be approximated by chl-a measurements on the suspended matter. Seasonal variations of chl-a and POC contents that are specific to each ecosystem (Fig. 2) can yield
116
J.A.C. Barth et al. / Earth and Planetary Science Letters 162 (1998) 111–121
insight into POC generation, because phytoplanktonic production should result in high POC contents during warm seasons (May to October). In contrast, cold season POC (November to April) should be dominated by allochthonous detrital material. We shall discuss the studied ecosystems in a traverse across the river from ‘Hoople Creek’ to ‘Hoople Bay’, ‘Whale Island’, ‘Main Channel’ and finally the embayment at ‘Long Sault Island’ (Fig. 1). The ecosystems ‘Bay at Cooper Marsh’ and ‘Cooper Marsh’ are described separately, since they are located 20 km further downstream. The ‘Hoople Creek’ ecosystem shows two peak POC values, in June 1995 and April 1996 (Fig. 2). These may be a consequence of either higher terrestrial runoff or higher pelagic phytoplanktonic activity. The latter derives some support from concomitant high chl-a values, but both processes may have been involved, because the creek also receives input of terrestrial material, such as plant litter. The isolated embayments of ‘Hoople Bay’ and ‘Long Sault Island’ derive the bulk of their elevated POC contents from pelagic algal activities, as indicated by their high chl-a concentrations (Fig. 2). Nevertheless, a subordinate terrestrial input cannot be entirely ruled out, because of their proximity to land. Furthermore, re-suspension of loosely layered, organic rich bottom sediments, particularly during storm events, may be responsible for some of the POC peaks not supported by chl-a. Higher flow velocities combined with low nutrient levels, as well as a greater distance from the shoreline, may cause the generally low POC and chl-a levels in the ‘Main Channel’ (Fig. 2). Its POC concentrations in November 1994 and December 1995, when active photosynthesis must have been minimal, do not differ significantly from the warm-season values. This means either that most of this POC is of allochthonous origin, or that there is a constant low supply of phytoplanktonic POC from the Great Lakes. The latter scenario implies that, during the cold season, chl-a and POC are residual from the summer and arrive from the Great Lakes with some time delay. Very similar waters also dominate the ‘Whale Island’ and ‘Bay at Cooper Marsh’ ecosystems and, not surprisingly, their POC and chl-a characteristics are somewhat comparable to those of the ‘Main Channel’.
The ‘Cooper Marsh’ has higher POC concentrations, probably due to terrestrial influences, particularly during cold seasons. During the warm season a portion of the POC is likely produced by in situ photosynthetic activity, because POC and chl-a values move in parallel. More frequent sampling may be required in order to decipher clearer seasonal patterns of chl-a and POC distributions. Nonetheless, their relative concentrations and temporal trends indicate that terrestrial organic matter dominates in near-shore ecosystems during cold seasons and storm surges. The in situ phytoplankton productivity is likely an important source for POC in the ‘Hoople Bay’, ‘Long Sault Island’ and ‘Cooper Marsh’ ecosystems, at least during the warm seasons. These tentative conclusions can be further tested by isotopic techniques. 4.2. Isotopic constraints Most δ13 CPOC values plot within the range of C3 plants ( 22 to 34‰; [17]), thus indicating that this material contributes significant portions of the POC pool (Fig. 3). Both allochthonous as well as in situ produced material can have this C3 -δ13 CPOC composition [16,17], thus making a distinction between the two sources difficult. Deviations to more
Fig. 3. POC=chl-a ratio vs. δ13 CPOC . The samples underlain by grey are from the cold season (November to April), while the samples underlain by white are from the warm season (May to October).
J.A.C. Barth et al. / Earth and Planetary Science Letters 162 (1998) 111–121
positive δ13 CPOC values (e.g. ‘Whale Island’, ‘Long Sault Island’, ‘Bay at Cooper Marsh’) indicate terrestrial influences, such as decomposition of lignin (δ13 CPOC ¾ 13 to 14‰; [18]), or influx of detrital material from C4 plants ( 10 to 16‰; [17]). The latter could originate from corn grown in the Cornwall area. Although the δ13 CPOC compositions show no clear seasonal pattern, the POC=chl-a ratios of most samples are higher during the cold season (Fig. 3), thus indicating dominance of terrestrial material. Lower POC=chl-a ratios, typical for the warm season, on the other hand, are more indicative of photosynthetic activity when lower than 100 [19,20]. Nonetheless, exceptions to this pattern (i.e. warm season samples in the cold season field and vice versa) indicate that the POC budget in the study area is more complex. For example, the relatively low POC=chl-a ratios of most warm season ‘Main Channel’ data may hint at algal production as the source of POC, but higher ratios can be caused by infrequent flushing of detrital POC from the near-shore ecosystems that contain little chl-a. It is more difficult to explain the cold season samples with a low POC=chl-a ratio as perhaps containing residual amounts of chl-a or reflecting some contemporaneous algal activity. Note that, particularly in shallow water near-shore ecosystems, lower POC=chl-a ratios do not necessarily indicate phytoplanktonic POC generation, as chl-a-containing plant debris from macrophytes may have lowered this ratio. The isotopic discrimination between POC and DIC may help to further elucidate the origin of the POC. During photosynthesis, aquatic plants utilize aqueous CO2 , a process that preferentially removes 12 C from the dissolved inorganic carbon pool. This mechanism explains the isotopic discrimination between DIC and phytoplankton [16]. The isotopic composition of POC (δ13 CPOC ), if resulting from algal photosynthetic activity, should therefore reflect that of DIC (δ13 CDIC ), but at a more negative level. Based on this reasoning, the δ13 CDIC and δ13 CPOC values should track each other. In general, the temporal δ13 CPOC and δ13 CDIC curves for the seven ecosystems move in parallel, indicating some degree of participation of algal photosynthetic activity in the POC production (Fig. 4). Some δ13 CPOC peaks that do not have δ13 CDIC counterparts are likely due to the
117
influx of isotopically variable material, thus blurring the covariance. The degree of isotopic discrimination between DIC and POC may therefore be useful in constraining the relative proportion of POC that originates from in situ photosynthetic activity as opposed to allochthonous detrital material in a given ecosystem and season. Theoretical considerations of carbon isotope fractionation by phytoplankton may further help with this task. 4.3. Theoretical fractionation by algae Theoretically, carbon isotope fractionation by algae can be divided into two principal components: (1) aqueous diffusion of CO2 , and (2) enzymatic carboxylation reaction during carbon fixation [17]. The diffusion of CO2 within water .žd / results in a fractionation of 0.7‰ [21], whereas the enzymatic fractionation .žf / is caused by a catalyst known as ribulose 1,5-bisphosphate carboxylase=oxygenase (RuBisCO) and has been ascribed values between 20 and 29‰ [22–24]. For the subsequent modeling approach, we select an intermediate value of 25‰ for the latter process. Thus the algal matter is more negative than the ambient CO2 by 0.7 to 25‰. The 0.7‰ fractionation occurs when low CO2 concentrations are rate limiting during photosynthesis, causing utilization of all available CO2 (12 CO2 and 13 CO2 ) and turning the carbon fixation by RuBisCO into a non-selective process. In contrast, at higher CO2 concentrations, the isotopic fractionation is dominated by RuBisCO carbon fixation and approaches 25‰. One can therefore expect the phytoplanktonic carbon to be more negative (by 0.7 to 25‰) than CO2 (aq), depending on the activity of the latter. With respect to the phytoplanktonic growth rate, ¼, this fractionation can be modeled utilizing the approach of Rau et al. [25]. Note that this model was developed for marine systems, but its application to freshwater systems is possible by modifying the relevant equation (Appendix A). In order to use the model, the isotopic enrichment factor, ž, between the two reservoirs has to be defined [26]: ð Ł žA B D ÞA B 1 ð 1000 1000 C ŽA 1 ð 1000 ³ ŽA ŽB D 1000 C ŽB
118
J.A.C. Barth et al. / Earth and Planetary Science Letters 162 (1998) 111–121
Fig. 4. Temporal variations of the δ13 CDIC (upper curves) and δ13 CPOC (lower curves) in the seven ecosystems studied.
where ÞA B is the equilibrium isotopic fractionation factor and the subscripts A and B correspond to the isotopic composition of CO2 (aq) and the phytoplanktonic POC, respectively. The isotopic enrichment factor between CO2 (aq) and phytoplanktonic POC .žC O2.aq/ P OC / varies with the growth rate, ¼, over the range of the cell external CO2 (aq) concentration, Ce (Fig. 5). Most samples plot in the theoretical growth rate range of 0.5 to 2, with some samples form near shore ecosystems (‘Hoople Creek’, ‘Cooper Marsh’, ‘Hoople Bay’, ‘Whale Island’ and ‘Long Sault Island’) plotting
beyond 2. The latter are samples from ecosystems with elevated POC and variable chl-a concentrations (Fig. 2) and can plot into both, the photosynthetic and detrital fields in Fig. 3 For samples that plot in the photosynthetic field in Fig. 3, but to the right of the ¼ D 2 line in Fig. 5 (i.e. ‘Hoople Creek’ and ‘Cooper Marsh’), two explanations may be given: (1) they derive their high chl-a contents from non-phytoplanktonic material (i.e. macrophytes) that does not fractionate according to the model; (2) phytoplanktonic cells from these samples have a larger radius. For example, increasing the radius from 10 µm to
J.A.C. Barth et al. / Earth and Planetary Science Letters 162 (1998) 111–121
119
ecosystems (‘Hoople Creek’, ‘Cooper Marsh’ ‘Long Sault Island’). If present, chl-a-containing diatoms would cause the samples to plot in the photosynthetic field in Fig. 3, while making them deviate from the growth rates in Fig. 5, because they do not conform to the model predictions.
5. Conclusions
Fig. 5. Theoretical and measured fractionations, ž, between Ce (DCO2 (aq)) and phytoplanktonic POC with varying growth rates, ¼, following the model of Rau et al. [25] (Appendix A).
50 µm, a still realistic proposition, would shift the isolines for the growth rate, ¼, further to the right, causing all but six samples to conform to the model. It should be noted that the isotopic fractionation by diatoms exhibits a large scatter when plotted against the concentration of CO2 (aq) [27]. This may weaken the validity of the model, depending on the proportion of diatoms in the phytoplankton pool. However, only insignificant amounts of diatoms were detected in the phytoplankton community of the ‘Main Channel’ in the Cornwall area [28]. Phytoplanktonic activity in the ‘Main Channel’ and related ecosystems (‘Whale Island’, ‘Bay at Cooper Marsh’, ‘Long Sault Island’) therefore likely conforms to the model by Rau et al. [25]. No estimates of diatom proportions are available for the near-shore
The study of the isotopic composition of POC in representative ecosystems of the upper St. Lawrence has shown that its near-shore ecosystems have a wide range of δ13 CPOC values. Nevertheless, their direct contribution to the POC flux in the main river channel is relatively minor and relegated mainly to high water flushing during storm events. The ‘Main Channel’ POC has δ13 C at 27 š 3‰, confirming the interpolations from earlier estuarine studies, with most of this POC originating from phytoplanktonic activity, likely in the Great Lakes. Our findings indicate that elevated phytoplanktonic activity and detrital sources are both contributors of POC in the near-shore sampling stations (‘Hoople Creek’, ‘Cooper Marsh’, ‘Hoople Bay’, ‘Long Sault Island’ and ‘Bay at Cooper Marsh’). Detrital sources are important during cold seasons and storm surges, while phytoplankton dominates during warm seasons. Some warm season samples with low POC=chl-a ratios do not conform to model predictions and may be influenced by macrophytes and=or diatoms, or simply contain algae with a larger cell radius. In contrast, the POC in the Great Lakes dominated ecosystems (‘Main Channel’, ‘Whale Island’, ‘Bay at Cooper Marsh’) appears to have been derived largely from in situ photosynthetic production in the Great Lakes.
Acknowledgements This research was supported by a grant from the Institute for Research on Environment and Economy (IREE) as well as NSERC research funds. We thank Lee Willard for his assistance with the sampling work. For advice with the δ13 CDIC analyses, many thanks to Gilles St.-Jean, Natalie Morisset and Wendy Abdi from the G.G. Hatch Isotope Labora-
120
J.A.C. Barth et al. / Earth and Planetary Science Letters 162 (1998) 111–121
tories at the University of Ottawa. For support with the chlorophyll-a analyses, we thank Frances Pick from the Department of Biology at the University of Ottawa. [FA]
Appendix A A.1. Determination of the theoretical carbon isotope fractionation by algae The isotopic composition of plant material, δ13 CPhyto , can be written as [29]: δ13 Cphyto D δ13 CCe
žd
δ13 CCe
.žf
where is the of the ambient CO2 (aq) (DCe ), žd is the isotopic fractionation associated with diffusive transport of CO2 (aq) in water (D0.7‰; [21]), žf is the isotope fractionation associated with enzymatic, intracellular carbon fixation (D25‰ in this model; [25]). While most of these independent variables are known, the cell internal CO2 concentration, Ci , is unknown. It can however be expressed as [25]:
Ci D Ce
B Qs ð B @
1 DTŁ
r 1C C C r PA C rk
with: Q s D CO2 (aq) uptake rate per unit cell surface area D
c ð 2¼=.4³r 2 / [mol C m 2 s 1 ] with c D carbon content per cell (D1:76 ð 10 11 [mol C]); µ D specific growth rate (varies between 0 and 2 [d 1 ] and is always <2.3 [25]); r D average cell radius (assumed to be 10 µm D 10 5 [m]); DT D Temperature sensitive diffusivity of CO2 (aq) in freshwater (D1:543 ð 10 9 [m2 s 1 ] at 17ºC); rk D Reacto diffusive length (D2:06 ð 10 4 [m]); and P D Cell wall permeability to CO2 (aq) [m s 1 ]For a detailed derivation of these parameters refer to Rau et al. [25]. By substituting various values for ¼ (D0, 0.5, 1 and 2), the equation can be solved explicitly for Ce and the isotopic fractionation between CO2 (aq) and phytoplankton, ž, with: .žf Ce D
½
c ð 2¼ r 1 ð C 4³r 2 .DT ð .1 C r=rk / P δ13 CPhyto δ13 CCe C žd C .žd žf /
žd / ð
The δ13 CCO2 .aq/ values were evaluated by using the temperature dependent isotopic discrimination between HCO3 and CO2 (aq). This isotopic discrimination changes linearly from 12‰ VPDB at 0ºC to 8.4‰ at 30ºC [30]. The δ13 C value of the CO2 (aq) can be calculated by setting the δ13 CDIC equal to the isotopic composition of the HCO3 .δ13 CHCO /. It is safe 3 to assume that isotopic compositions of DIC and HCO3 are identical, as long as the dissolved HCO3 is the major species of the DIC, which is true for the pH conditions found in this study. The activities of the aqueous CO2 (DCe ) for the samples displayed in Fig. 5 were calculated from pH and alkalinity by using the program Hydrowin [11].
žd / ð Ci =Ce
δ13 C
0
A.2. Determination of the δ13 CCO2 .aq/
Temperature and pH influence the parameters DT and rk , but the model results proved to be essentially the same even under dramatic changes of these two parameters (i.e. change of temperature š10ºC and change of š1 pH unit). The model is, however, very sensitive to the cell radius, r, because it also changes cell carbon content, c . Increasing r causes the isolines for ¼ to plot further to the right in Fig. 5, thus integrating more samples into the model. However an average cell radius of 10 µm seems a reasonable assumption for the phytoplankton community in the ‘Main Channel’ at Cornwall [28].
References [1] W. Ludwig, J.L. Probst, S. Kempe, Predicting the oceanic input of organic carbon by continental erosion, Global Biogeochem. Cycles 10 (1996) 23–41. [2] W.G. Mook, F.C. Tan, Stable carbon isotopes in rivers and estuaries, in: E.T. Degens, S. Kempe, J.E. Richey (Eds.), Biogeochemistry of Major World Rivers, Scope 42, Wiley, Chichester, 1991, pp. 245–264. [3] R. Pocklington, The contribution of organic matter by the St. Lawrence River to the Gulf of the St. Lawrence, 1981– 1983, in: E.T. Degens, S. Kempe, R. Herrera (Eds.), Transport of Carbon and Minerals in Major World Rivers, Mitt. Geol.-Pala¨ont. Inst. Univ. Hamburg, 1985, pp. 323–328. [4] R. Pocklington, F.C. Tan, Seasonal and annual variations in the organic matter contributed by the St. Lawrence River to the Gulf of St. Lawrence, Geochim. Cosmochim. Acta. 51 (1987) 2579–2586. [5] F.C. Tan, P.M. Strain, Sources, sinks and distribution of organic carbon in the St. Lawrence Estuary, Canada, Geochim. Cosmochim. Acta. 47 (1983) 123–125. [6] M. Lucotte, Organic carbon isotope ratios and implications for the maximum turbidity zone of the St. Lawrence upper estuary, Estuar. Coastal Shelf Sci. 29 (1989) 293–304. [7] M. Lucotte, C. Hillaire-Marcel, P. Louchouarn, First-order organic carbon budget in the St. Lawrence lower estuary from 13 C data, Estuar. Coastal Shelf Sci. 32 (1991) 297– 312. [8] P.H. Gleick, Water in Crisis. A Guide to the World’s Fresh Water Resources, Oxford University Press, New York, 473 pp. [9] Keith, L.H., Compilation of EPA’s Sampling and Analysis Methods, Lewis, Boca Raton, FL, 1996, 1696 pp. [10] J.A.C. Barth, J. Veizer, G.E.M. Hall, Geochemical monitoring of the St. Lawrence River in the Cornwall area. Ecosystem recovery on the St. Lawrence (1996), IREE Research Paper, 23, Institute for Research on Environment and Economy, University of Ottawa, Canada, 13 pp. [11] Calmbach, L., Hydrowin (Computer Program), version 3.0, Institut de Mineralogie, Lausanne, 1995.
J.A.C. Barth et al. / Earth and Planetary Science Letters 162 (1998) 111–121 [12] B.K. Burnison, Modified dimethyl sulfoxide (DMSO) extraction for chlorophyll analysis of phytoplankton, Can. J. Fish. Aquat. Sci. 37 (1980) 729–733. [13] S.W. Jeffrey, G.F. Humphrey, New spectrometric equations for determining chlorophylls a, b, c1 and c2 in equal amounts in higher plants, algae and natural phytoplankton, Biochem. Physiol. Pflanz. (BPP) 167 (1975) 191–194. [14] R.G. Wetzel, G.E. Likens, Limnological Analyses, Springer, New York, 1991, 391 pp. [15] M.L. Fogel, L.A. Cifuentes, Isotope Fractionation during Primary Production, in: M.H. Engel, S.A. Macko (Eds.), Organic Geochemistry: Principles and Applications, Plenum Press, New York, 1993, pp. 73–98. [16] B. Fry, E.B. Sherr, δ13 C measurements as indicators of carbon flow in marine and freshwater ecosystems, Contr. Mar. Sci. 27 (1984) 13–47. [17] J.C. Vogel, Variability of carbon isotope fractionation during photosynthesis, in: J.R. Ehleringer, A.H. Hall, G.D. Farquhar (Eds.), Stable Isotopes and Plant Carbon–Water relations, Academic Press, San Diego, 1993, pp. 29–45. [18] R. Benner, M.L. Fogel, E.K. Sprague, R.E. Hodson, Depletion of 13 C in lignin and its implications for stable carbon isotope studies, Nature 329 (1987) 708–710. [19] B.J.F. Biggs, Biomonitoring of organic pollution using periphyton, South Branch, Canterbury, New Zealand, NZ J. Mar. Freshw. Res. 23, 263–274. [20] D.M. SØballe, R.W. Bachman, Influence of reservoir transit on riverine algal transport and abundance, Can. J. Fish. Aquat. Sci. 41 (1984) 1803–1813. [21] M.H. O’Leary, Measurement of the isotopic fractionation associated with diffusion of carbon dioxide in aqueous solution, J. Phys. Chem. 88 (1984) 823–825.
121
[22] R. Goericke, J.P. Montoya, B. Fry, Physiology of isotope fractionation in algae and cyanobacteria, in: K. Lajtha, B. Michener (Eds.), Stable Isotopes in Ecology, Blackwell, Boston, 1994, pp. 187–221. [23] W.W. Wong, C.R. Benedict, R.J. Kohel, Enzymatic fractionation of the stable carbon isotopes of carbon dioxide by ribulose-1.5-bisposphate carboxylase, Plant. Physiol. 63 (1979) 852–856. [24] C.A. Roeske, M.H. O’Leary, Carbon isotope effects on the enzyme-catalyzed carboxylation of ribulose bisphosphate, Biochemistry 23 (1984) 6275–6284. [25] G.H. Rau, U.H. Riebesell, D. Wolf-Gladrow, A model of photosynthetic 13 C fractionation by marine phytoplankton based on diffusive molecular CO2 uptake, Mar. Ecol. Prog. Ser. 133 (1996) 275–285. [26] I.D. Clark, P. Fritz, Environmental Isotopes in Hydrogeology, Lewis, Boca Raton, FL, 1997, 328 pp. [27] R.D. Pancost, K.H. Freeman, S.G. Wakeham, C.Y. Robertson, Controls on carbon isotope fractionation by diatoms in the Peru upwelling region, Geochim. Cosmochim. Acta 61 (1997) 4983–4991. [28] C. Hudon, S. Paquet, V. Jarry, Downstream variations of phytoplankton in the St. Lawrence River (Que´bec, Canada), Hydrobiology 337 (1996) 11–26. [29] G.D. Farquhar, M.H. O’Leary, J.H. Berry, On the relationship between carbon isotope discrimination and the intercellular carbon dioxide concentration in leaves, Aust. J. Plant. Physiol. 9 (1982) 121–137. [30] W.G. Mook, J.C. Bommerson, W.H. Staverman, Carbon isotope fractionation between dissolved bicarbonate and gaseous carbon dioxide, Earth Planet. Sci. Lett. 22 (1974) 169–176.
Earth and Planetary Science Letters 162 (1998) 111–121
Origin of particulate organic carbon in the upper St. Lawrence: isotopic constraints Johannes A.C. Barth a,Ł , Ja´n Veizer a,b , Bernhard Mayer b,1 a Ottawa-Carleton b
Geoscience Centre, University of Ottawa, Ottawa, Ontario K1N 6N5, Canada Institut fu¨r Geologie, Ruhr-Universita¨t Bochum, D-44780 Bochum, Germany
Received 17 November 1997; revised version received 10 July 1998; accepted 10 July 1998
Abstract Seven sampling locations in the upper St. Lawrence River near the city of Cornwall (Ontario, Canada), including the main river and six near-shore ecosystems (a creek, embayments and a wetland) were studied in order to determine the origin of particulate organic carbon. Parameters studied included chlorophyll-a (chl-a), particulate organic carbon (POC) and dissolved inorganic carbon (DIC), as well as the isotopic compositions of the latter two (δ13 CPOC , δ13 CDIC ). The results show that in situ photosynthesis and detrital inputs are both significant contributors to the POC pool in the isolated embayments. The former dominates during warm seasons, with POC concentrations up to 2663 µg=l and chl-a concentrations up to 26.1 µg=l. Near-shore ecosystems have a wide range of δ13 CPOC values ( 31.5 to 16.3‰), but this variability is not reflected in the ‘Main Channel’. There, the δ13 CPOC signal is uniformly close to 27‰, in accord with estimates from earlier studies on the river’s estuary. This suggests that the POC contribution from near-shore ecosystems is minor. Although the ‘Main Channel’ has low chl-a concentrations, model calculations suggest that most of its POC originates from photosynthetic activity, probably within the Great Lakes. 1998 Elsevier Science B.V. All rights reserved. Keywords: St. Lawrence River; organic carbon; stable isotopes
1. Introduction The study of organic carbon in rivers provides important information on the recycling of continental carbon [1]. Riverine particulate organic carbon (POC) consists partly of detrital compounds, but can also be a product of in situ phytoplanktonic activity [2]. Understanding the extent of in situ production Ł Corresponding
author. Fax: C1 613 562 5192; E-mail: [email protected] 1 Present address: Departments of Physics=Astronomy and Geology=Geophysics, University of Calgary, Calgary, Alberta T2N 1N4, Canada.
of POC is essential, because it influences the aqueous carbon cycle by consuming dissolved inorganic carbon (DIC). For the St. Lawrence River, the particulate organic carbon (POC) has been studied in detail, but only in the estuary [3–7]. In these studies, the nature of the upstream riverine POC was conjectured to be mainly of terrestrial origin and to have isotopic compositions between 24.1 and 27‰. As yet no measurements on the isotopic composition of the upper St. Lawrence POC pool have been available. The present study provides the first overview of this parameter from a monitoring of seven aquatic ecosystems near the city of Cornwall (Ontario, Canada), ¾150
0012-821X/98/$19.00 1998 Elsevier Science B.V. All rights reserved. PII S 0 0 1 2 - 8 2 1 X ( 9 8 ) 0 0 1 6 0 - 5
112
J.A.C. Barth et al. / Earth and Planetary Science Letters 162 (1998) 111–121
Fig. 1. Location map of the study area. Circles are proportional to the annual average concentrations of particulate organic carbon in the ecosystems studied.
km upstream of Montre´al (Fig. 1). The ecosystems studied include the center of the St. Lawrence River (‘Main Channel’), several of its embayments (‘Bay at Cooper Marsh’, ‘Long Sault Island’, ‘Whale Island’ and ‘Hoople Bay’), a creek that drains into the St. Lawrence (‘Hoople Creek’) and a wetland (‘Cooper Marsh’). The objectives of this research were: (1) to evaluate the relative importance of phytoplanktonic POC production versus detrital influx, seasonally and spatially, among the studied ecosystems; and (2) to test the proposition, based on estuarine studies, that the St. Lawrence is characterized by detrital POC. The St. Lawrence is particularly suitable for this type of study, because its average discharge of 4ð106 tonnes of suspended sediments per year ranks lowest among the world’s major rivers [8]. For comparison, the Mississippi has an average suspended load of 210 ð 106 t=yr [8]. The low suspended sediment load of the St. Lawrence River implies small POC
concentrations, minimizing the masking effect of the usually large allochthonous component on the overall balance of the riverine POC. Our study area is typical of the upstream parts of the St. Lawrence River, because at this point the water originates directly from Lake Ontario without any influence from tributaries. As a result, the ‘Main Channel’ sampling station is characterized by well mixed POC from the Great Lakes and can thus serve as a baseline for comparison with the other ecosystems studied.
2. Material and methods Seven stations in the vicinity of Cornwall (Fig. 1) were sampled repeatedly between June, 1994, and April, 1996. Samples were collected 2 m below the water surface with a 2-l narrow-mouth bottle attached to a metal weight. The sampling stations ‘Hoople Creek’, ‘Long Sault Island’ and ‘Cooper Marsh’ had
J.A.C. Barth et al. / Earth and Planetary Science Letters 162 (1998) 111–121
depths less than 2 m, allowing sample collection only at 0.5 to 0.75 m below the water surface. Water temperature, pH, and alkalinity were measured in situ in order to calculate the speciation of the DIC. The pH was determined with a Hach pH meter which was calibrated before every measurement with two standard buffer solutions. The 2¦ precision of š0.03 pH units was determined from multiple measurements of standards in the field and laboratory. The alkalinity titrations were performed with a Hach digital titrator on 100 ml of sample using 1.6 N sulfuric acid, the results being expressed in µM HCO3 . Repeat measurements of standards at various concentrations yielded 2σ values of š24.6 µM HCO3 , matching the recommended precision of the Compilation of EPA’s Sampling and Analysis Methods [9]. Together with routinely determined major anion and cation concentrations (Ca2C , Mg2C , NaC , KC , Cl , SO24 ) [10], the field parameters were then entered into the program Hydrowin [11] for calculation of CO2 (aq) activities (Appendix A). The samples for the δ13 CDIC measurements were filtered through 0.45-µm cellulose acetate filter papers, preserved with HgCl2 and stored at 4ºC until analysis. At the G.G. Hatch Isotope Laboratories, University of Ottawa (Canada), the DIC was converted to CO2 by acidification with 85% H3 PO4 and then purified under vacuum. The isotopic composition of the extracted CO2 was then determined on a triple collector VG SIRA 12 mass spectrometer. The results are expressed with reference to the Pee Dee Belemnite standard (VPDB), with an overall analytical precision (2¦ ) of š0.3‰. Samples for POC were collected on Whatman GF=C 2000 glass microfibre filter papers which had been prepared by overnight heating at 500ºC in order to remove organic carbon contaminations. After filtration, the samples were frozen at 20ºC until analysis. Prior to analysis, the filtered material was acidified with 10% HCl in order to eliminate any particulate inorganic carbon, then dried at 70ºC overnight. The suspended material and the filter paper could not be separated and therefore were ground up and analyzed together. The POC content and δ13 CPOC measurements were performed at the Isotope Laboratory of the Ruhr-Universita¨t, Bochum (Germany) by continuous flow isotope ratio mass spectrometry (CF-IRMS), using a Carlo Erba ele-
113
mental analyzer (CE 1110) connected to a Finnigan MAT mass spectrometer (delta C) via a ConFlo II interface. The results obtained reflected filter paper and sample material, and the POC contents were therefore calculated by subtracting average readings from multiple weighings of blank filter papers. The δ13 CPOC was then determined by mass and isotope balance calculations. The precisions for the C-content and its δ13 C value were determined from duplicate and triplicate measurements of samples as well as from multiple measurements of blanks. The 2σ precisions were š20 µg=l and š0.5‰ VPDB for the POC and the δ13 CPOC measurements, respectively. The suspended material for the chl-a analyses was collected on Whatman GF=C 2000 glass microfibre filter papers and frozen immediately after filtration at 20ºC. In the laboratory (Department of Biology, University of Ottawa), chl-a was extracted from the filter paper by leaching with concentrated dimethyl sulfoxide (DMSO) at 60ºC, and rinsing with 90% acetone [12]. Subsequently, the chl-a contents of the leachates were determined with a Pye Unicam double beam spectrophotometer (model SP8-100), following the guidelines of Jeffrey and Humphrey [13]. The chl-a content was then calculated utilizing the formula in Wetzel and Likens [14]. The precision of the chl-a measurements (2¦ ) was š0.18 µg=l, as determined from multiple measurements of standards.
3. Results The measured POC and chl-a concentrations range from 64 to 2663 and 0.3 to 26.1 µg=l, respectively (Table 1). Samples with high chl-a concentrations also have high POC contents, with the ‘Main Channel’ waters representing the lowest endmember and the ‘Hoople Bay’ and ‘Long Sault Island’ embayments having values up to an order of magnitude higher (Fig. 2). The waters of ‘Hoople Creek’, ‘Whale Island’, ‘Bay at Cooper Marsh’ and ‘Cooper Marsh’ are intermediate between the ‘Main Channel’ and the embayments in terms of their POC and chl-a levels. The isotopic compositions of the POC (δ13 CPOC ) range from 31.5 to 16.3‰, while those of the dissolved inorganic carbon (δ13 CDIC ) range between
114
J.A.C. Barth et al. / Earth and Planetary Science Letters 162 (1998) 111–121
Table 1 List of parameters studied Date
Chl-a (µg=l)
POC (µg=l)
POC=chl-a –
δ13 CPOC (‰)
δ13 CDIC (‰)
aHCO3 (µM)
aCO3 (µM)
aCO2 (aq) (µM)
DIC (µM)
Bay at Cooper Marsh 6=12=95 3.0 8=1=95 1.3 10=10=95 0.8 4=19=96 1.5
277 105 102 91
91.6 81.3 122.6 62.3
26.5 21.8 25.9 25.3
2.0 0.6 0.1 0.9
1711 1576 1688 1690
32 39 18 10
9 7 15 25
1752 1622 1721 1726
Cooper Marsh 8=14=95 10=18=95 12=6=95 4=19=96
4.3 8.6 4.8 2.3
391 711 964 361
90.7 82.6 202.7 159.6
25.7 31.4 27.0 30.9
5.6 9.0 10.6 11.3
1587 1587 2090 1955
9 2 1 3
30 156 369 115
1625 1745 2460 2073
10.7 11.2 9.3 23.3 6.5 8.9
1095 832 1578 1056 415 2107
102.7 74.6 170.0 45.3 63.7 236.2
28.2 30.0 27.3 24.7 27.3 28.5
5.8 1.4 1.2 1.1 10.3 11.0
2028 2386 1804 1713 2921 2430
13 21 19 30 6 12
31 27 18 9 136 47
2072 2434 1841 1752 3063 2489
4.0 2.6 1.6 1.4 1.2 3.4
211 747 173 162 112 1087
52.7 283.3 108.3 112.9 95.5 320.1
31.5 27.4 29.8 28.5 31.3 24.7
13.3 12.0 10.5 10.5 13.7 12.0
2622 3303 2329 2298 3358 2583
13 51 18 12 3 5
52 22 32 43 385 116
2688 3376 2378 2353 3745 2705
Long Sault Island 8=1=95 26.1 10=10=95 4.3
2663 429
102.0 100.3
25.4 21.4
1.3 1.2
1806 1882
18 7
19 52
1843 1940
Hoople Bay 5=30=95 6=13=95 8=2=95 10=17=95 12=5=95 4=21=96 Hoople Creek 5=30=95 6=14=95 8=2=95 10=11=95 12=4=95 4=21=96
Main Channel 7=14=94 11=3=94 5=29=95 10=17=95 12=4=95
0.7 0.3 1.1 1.1 0.7
102 95 64 76 99
136.6 334.7 59.1 66.2 141.7
27.4 26.7 30.0 28.5 26.3
1.3 1.8 1.0 0.3 0.1
1500 1654 1559 1689 1634
13 13 8 16 16
18 20 30 17 15
1531 1687 1597 1722 1665
Whale Island 5=29=95 6=13=95 8=1=95 10=10=95 12=5=95 4=20=96
2.0 1.2 0.8 2.1 1.0 2.8
264 162 274 226 264 245
135.4 136.6 340.0 110.1 261.1 87.2
25.7 27.0 26.3 22.4 16.3 27.0
1.3 2.6 0.4 0.2 0.8 4.8
1699 1712 1749 1579 1724 1762
7 30 14 20 9 19
42 10 23 12 29 15
1747 1752 1786 1611 1762 1796
13.7 and C0.3‰ (Table 1). Although these two parameters yield only a somewhat diffuse spatial pattern, the samples from the ‘Hoople Creek’ and ‘Cooper Marsh’ ecosystems show the most negative δ13 CPOC and δ13 CDIC values, indicating that these parameters are somewhat correlated (Table 1).
The CO2 (aq) activities of most water samples scatter around a mean of 20 µM, reflecting the fact that the ‘Main Channel’ provides the bulk of the water to most of the studied ecosystems. Only the near-shore ecosystems ‘Hoople Creek’, ‘Hoople Bay’ and ‘Cooper Marsh’ deviated from this mean,
J.A.C. Barth et al. / Earth and Planetary Science Letters 162 (1998) 111–121
115
Fig. 2. Temporal distributions of chlorophyll-a contents and particulate organic carbon concentrations in the seven ecosystems studied.
reaching peak CO2 (aq) activities of 385 µM (Table 1), an observation that will be of importance in the discussion of isotope discrimination between CO2 (aq) and phytoplanktonic POC.
4. Discussion 4.1. Sources of particulate organic carbon Particulate organic carbon represents a variable mixture of living and dead phytoplankton and other
components, such as detritus, bacteria and zooplankton [15,16]. Among all these components the major sources of riverine POC are (1) detrital matter, and (2) in situ produced phytoplankton. The latter changes the aqueous CO2 content and the 13 C=12 C ratio of dissolved inorganic carbon and can therefore significantly influence the riverine carbon cycle. Although the proportion of phytoplankton in the POC pool is not easily quantified, it can be approximated by chl-a measurements on the suspended matter. Seasonal variations of chl-a and POC contents that are specific to each ecosystem (Fig. 2) can yield
116
J.A.C. Barth et al. / Earth and Planetary Science Letters 162 (1998) 111–121
insight into POC generation, because phytoplanktonic production should result in high POC contents during warm seasons (May to October). In contrast, cold season POC (November to April) should be dominated by allochthonous detrital material. We shall discuss the studied ecosystems in a traverse across the river from ‘Hoople Creek’ to ‘Hoople Bay’, ‘Whale Island’, ‘Main Channel’ and finally the embayment at ‘Long Sault Island’ (Fig. 1). The ecosystems ‘Bay at Cooper Marsh’ and ‘Cooper Marsh’ are described separately, since they are located 20 km further downstream. The ‘Hoople Creek’ ecosystem shows two peak POC values, in June 1995 and April 1996 (Fig. 2). These may be a consequence of either higher terrestrial runoff or higher pelagic phytoplanktonic activity. The latter derives some support from concomitant high chl-a values, but both processes may have been involved, because the creek also receives input of terrestrial material, such as plant litter. The isolated embayments of ‘Hoople Bay’ and ‘Long Sault Island’ derive the bulk of their elevated POC contents from pelagic algal activities, as indicated by their high chl-a concentrations (Fig. 2). Nevertheless, a subordinate terrestrial input cannot be entirely ruled out, because of their proximity to land. Furthermore, re-suspension of loosely layered, organic rich bottom sediments, particularly during storm events, may be responsible for some of the POC peaks not supported by chl-a. Higher flow velocities combined with low nutrient levels, as well as a greater distance from the shoreline, may cause the generally low POC and chl-a levels in the ‘Main Channel’ (Fig. 2). Its POC concentrations in November 1994 and December 1995, when active photosynthesis must have been minimal, do not differ significantly from the warm-season values. This means either that most of this POC is of allochthonous origin, or that there is a constant low supply of phytoplanktonic POC from the Great Lakes. The latter scenario implies that, during the cold season, chl-a and POC are residual from the summer and arrive from the Great Lakes with some time delay. Very similar waters also dominate the ‘Whale Island’ and ‘Bay at Cooper Marsh’ ecosystems and, not surprisingly, their POC and chl-a characteristics are somewhat comparable to those of the ‘Main Channel’.
The ‘Cooper Marsh’ has higher POC concentrations, probably due to terrestrial influences, particularly during cold seasons. During the warm season a portion of the POC is likely produced by in situ photosynthetic activity, because POC and chl-a values move in parallel. More frequent sampling may be required in order to decipher clearer seasonal patterns of chl-a and POC distributions. Nonetheless, their relative concentrations and temporal trends indicate that terrestrial organic matter dominates in near-shore ecosystems during cold seasons and storm surges. The in situ phytoplankton productivity is likely an important source for POC in the ‘Hoople Bay’, ‘Long Sault Island’ and ‘Cooper Marsh’ ecosystems, at least during the warm seasons. These tentative conclusions can be further tested by isotopic techniques. 4.2. Isotopic constraints Most δ13 CPOC values plot within the range of C3 plants ( 22 to 34‰; [17]), thus indicating that this material contributes significant portions of the POC pool (Fig. 3). Both allochthonous as well as in situ produced material can have this C3 -δ13 CPOC composition [16,17], thus making a distinction between the two sources difficult. Deviations to more
Fig. 3. POC=chl-a ratio vs. δ13 CPOC . The samples underlain by grey are from the cold season (November to April), while the samples underlain by white are from the warm season (May to October).
J.A.C. Barth et al. / Earth and Planetary Science Letters 162 (1998) 111–121
positive δ13 CPOC values (e.g. ‘Whale Island’, ‘Long Sault Island’, ‘Bay at Cooper Marsh’) indicate terrestrial influences, such as decomposition of lignin (δ13 CPOC ¾ 13 to 14‰; [18]), or influx of detrital material from C4 plants ( 10 to 16‰; [17]). The latter could originate from corn grown in the Cornwall area. Although the δ13 CPOC compositions show no clear seasonal pattern, the POC=chl-a ratios of most samples are higher during the cold season (Fig. 3), thus indicating dominance of terrestrial material. Lower POC=chl-a ratios, typical for the warm season, on the other hand, are more indicative of photosynthetic activity when lower than 100 [19,20]. Nonetheless, exceptions to this pattern (i.e. warm season samples in the cold season field and vice versa) indicate that the POC budget in the study area is more complex. For example, the relatively low POC=chl-a ratios of most warm season ‘Main Channel’ data may hint at algal production as the source of POC, but higher ratios can be caused by infrequent flushing of detrital POC from the near-shore ecosystems that contain little chl-a. It is more difficult to explain the cold season samples with a low POC=chl-a ratio as perhaps containing residual amounts of chl-a or reflecting some contemporaneous algal activity. Note that, particularly in shallow water near-shore ecosystems, lower POC=chl-a ratios do not necessarily indicate phytoplanktonic POC generation, as chl-a-containing plant debris from macrophytes may have lowered this ratio. The isotopic discrimination between POC and DIC may help to further elucidate the origin of the POC. During photosynthesis, aquatic plants utilize aqueous CO2 , a process that preferentially removes 12 C from the dissolved inorganic carbon pool. This mechanism explains the isotopic discrimination between DIC and phytoplankton [16]. The isotopic composition of POC (δ13 CPOC ), if resulting from algal photosynthetic activity, should therefore reflect that of DIC (δ13 CDIC ), but at a more negative level. Based on this reasoning, the δ13 CDIC and δ13 CPOC values should track each other. In general, the temporal δ13 CPOC and δ13 CDIC curves for the seven ecosystems move in parallel, indicating some degree of participation of algal photosynthetic activity in the POC production (Fig. 4). Some δ13 CPOC peaks that do not have δ13 CDIC counterparts are likely due to the
117
influx of isotopically variable material, thus blurring the covariance. The degree of isotopic discrimination between DIC and POC may therefore be useful in constraining the relative proportion of POC that originates from in situ photosynthetic activity as opposed to allochthonous detrital material in a given ecosystem and season. Theoretical considerations of carbon isotope fractionation by phytoplankton may further help with this task. 4.3. Theoretical fractionation by algae Theoretically, carbon isotope fractionation by algae can be divided into two principal components: (1) aqueous diffusion of CO2 , and (2) enzymatic carboxylation reaction during carbon fixation [17]. The diffusion of CO2 within water .žd / results in a fractionation of 0.7‰ [21], whereas the enzymatic fractionation .žf / is caused by a catalyst known as ribulose 1,5-bisphosphate carboxylase=oxygenase (RuBisCO) and has been ascribed values between 20 and 29‰ [22–24]. For the subsequent modeling approach, we select an intermediate value of 25‰ for the latter process. Thus the algal matter is more negative than the ambient CO2 by 0.7 to 25‰. The 0.7‰ fractionation occurs when low CO2 concentrations are rate limiting during photosynthesis, causing utilization of all available CO2 (12 CO2 and 13 CO2 ) and turning the carbon fixation by RuBisCO into a non-selective process. In contrast, at higher CO2 concentrations, the isotopic fractionation is dominated by RuBisCO carbon fixation and approaches 25‰. One can therefore expect the phytoplanktonic carbon to be more negative (by 0.7 to 25‰) than CO2 (aq), depending on the activity of the latter. With respect to the phytoplanktonic growth rate, ¼, this fractionation can be modeled utilizing the approach of Rau et al. [25]. Note that this model was developed for marine systems, but its application to freshwater systems is possible by modifying the relevant equation (Appendix A). In order to use the model, the isotopic enrichment factor, ž, between the two reservoirs has to be defined [26]: ð Ł žA B D ÞA B 1 ð 1000 1000 C ŽA 1 ð 1000 ³ ŽA ŽB D 1000 C ŽB
118
J.A.C. Barth et al. / Earth and Planetary Science Letters 162 (1998) 111–121
Fig. 4. Temporal variations of the δ13 CDIC (upper curves) and δ13 CPOC (lower curves) in the seven ecosystems studied.
where ÞA B is the equilibrium isotopic fractionation factor and the subscripts A and B correspond to the isotopic composition of CO2 (aq) and the phytoplanktonic POC, respectively. The isotopic enrichment factor between CO2 (aq) and phytoplanktonic POC .žC O2.aq/ P OC / varies with the growth rate, ¼, over the range of the cell external CO2 (aq) concentration, Ce (Fig. 5). Most samples plot in the theoretical growth rate range of 0.5 to 2, with some samples form near shore ecosystems (‘Hoople Creek’, ‘Cooper Marsh’, ‘Hoople Bay’, ‘Whale Island’ and ‘Long Sault Island’) plotting
beyond 2. The latter are samples from ecosystems with elevated POC and variable chl-a concentrations (Fig. 2) and can plot into both, the photosynthetic and detrital fields in Fig. 3 For samples that plot in the photosynthetic field in Fig. 3, but to the right of the ¼ D 2 line in Fig. 5 (i.e. ‘Hoople Creek’ and ‘Cooper Marsh’), two explanations may be given: (1) they derive their high chl-a contents from non-phytoplanktonic material (i.e. macrophytes) that does not fractionate according to the model; (2) phytoplanktonic cells from these samples have a larger radius. For example, increasing the radius from 10 µm to
J.A.C. Barth et al. / Earth and Planetary Science Letters 162 (1998) 111–121
119
ecosystems (‘Hoople Creek’, ‘Cooper Marsh’ ‘Long Sault Island’). If present, chl-a-containing diatoms would cause the samples to plot in the photosynthetic field in Fig. 3, while making them deviate from the growth rates in Fig. 5, because they do not conform to the model predictions.
5. Conclusions
Fig. 5. Theoretical and measured fractionations, ž, between Ce (DCO2 (aq)) and phytoplanktonic POC with varying growth rates, ¼, following the model of Rau et al. [25] (Appendix A).
50 µm, a still realistic proposition, would shift the isolines for the growth rate, ¼, further to the right, causing all but six samples to conform to the model. It should be noted that the isotopic fractionation by diatoms exhibits a large scatter when plotted against the concentration of CO2 (aq) [27]. This may weaken the validity of the model, depending on the proportion of diatoms in the phytoplankton pool. However, only insignificant amounts of diatoms were detected in the phytoplankton community of the ‘Main Channel’ in the Cornwall area [28]. Phytoplanktonic activity in the ‘Main Channel’ and related ecosystems (‘Whale Island’, ‘Bay at Cooper Marsh’, ‘Long Sault Island’) therefore likely conforms to the model by Rau et al. [25]. No estimates of diatom proportions are available for the near-shore
The study of the isotopic composition of POC in representative ecosystems of the upper St. Lawrence has shown that its near-shore ecosystems have a wide range of δ13 CPOC values. Nevertheless, their direct contribution to the POC flux in the main river channel is relatively minor and relegated mainly to high water flushing during storm events. The ‘Main Channel’ POC has δ13 C at 27 š 3‰, confirming the interpolations from earlier estuarine studies, with most of this POC originating from phytoplanktonic activity, likely in the Great Lakes. Our findings indicate that elevated phytoplanktonic activity and detrital sources are both contributors of POC in the near-shore sampling stations (‘Hoople Creek’, ‘Cooper Marsh’, ‘Hoople Bay’, ‘Long Sault Island’ and ‘Bay at Cooper Marsh’). Detrital sources are important during cold seasons and storm surges, while phytoplankton dominates during warm seasons. Some warm season samples with low POC=chl-a ratios do not conform to model predictions and may be influenced by macrophytes and=or diatoms, or simply contain algae with a larger cell radius. In contrast, the POC in the Great Lakes dominated ecosystems (‘Main Channel’, ‘Whale Island’, ‘Bay at Cooper Marsh’) appears to have been derived largely from in situ photosynthetic production in the Great Lakes.
Acknowledgements This research was supported by a grant from the Institute for Research on Environment and Economy (IREE) as well as NSERC research funds. We thank Lee Willard for his assistance with the sampling work. For advice with the δ13 CDIC analyses, many thanks to Gilles St.-Jean, Natalie Morisset and Wendy Abdi from the G.G. Hatch Isotope Labora-
120
J.A.C. Barth et al. / Earth and Planetary Science Letters 162 (1998) 111–121
tories at the University of Ottawa. For support with the chlorophyll-a analyses, we thank Frances Pick from the Department of Biology at the University of Ottawa. [FA]
Appendix A A.1. Determination of the theoretical carbon isotope fractionation by algae The isotopic composition of plant material, δ13 CPhyto , can be written as [29]: δ13 Cphyto D δ13 CCe
žd
δ13 CCe
.žf
where is the of the ambient CO2 (aq) (DCe ), žd is the isotopic fractionation associated with diffusive transport of CO2 (aq) in water (D0.7‰; [21]), žf is the isotope fractionation associated with enzymatic, intracellular carbon fixation (D25‰ in this model; [25]). While most of these independent variables are known, the cell internal CO2 concentration, Ci , is unknown. It can however be expressed as [25]:
Ci D Ce
B Qs ð B @
1 DTŁ
r 1C C C r PA C rk
with: Q s D CO2 (aq) uptake rate per unit cell surface area D
c ð 2¼=.4³r 2 / [mol C m 2 s 1 ] with c D carbon content per cell (D1:76 ð 10 11 [mol C]); µ D specific growth rate (varies between 0 and 2 [d 1 ] and is always <2.3 [25]); r D average cell radius (assumed to be 10 µm D 10 5 [m]); DT D Temperature sensitive diffusivity of CO2 (aq) in freshwater (D1:543 ð 10 9 [m2 s 1 ] at 17ºC); rk D Reacto diffusive length (D2:06 ð 10 4 [m]); and P D Cell wall permeability to CO2 (aq) [m s 1 ]For a detailed derivation of these parameters refer to Rau et al. [25]. By substituting various values for ¼ (D0, 0.5, 1 and 2), the equation can be solved explicitly for Ce and the isotopic fractionation between CO2 (aq) and phytoplankton, ž, with: .žf Ce D
½
c ð 2¼ r 1 ð C 4³r 2 .DT ð .1 C r=rk / P δ13 CPhyto δ13 CCe C žd C .žd žf /
žd / ð
The δ13 CCO2 .aq/ values were evaluated by using the temperature dependent isotopic discrimination between HCO3 and CO2 (aq). This isotopic discrimination changes linearly from 12‰ VPDB at 0ºC to 8.4‰ at 30ºC [30]. The δ13 C value of the CO2 (aq) can be calculated by setting the δ13 CDIC equal to the isotopic composition of the HCO3 .δ13 CHCO /. It is safe 3 to assume that isotopic compositions of DIC and HCO3 are identical, as long as the dissolved HCO3 is the major species of the DIC, which is true for the pH conditions found in this study. The activities of the aqueous CO2 (DCe ) for the samples displayed in Fig. 5 were calculated from pH and alkalinity by using the program Hydrowin [11].
žd / ð Ci =Ce
δ13 C
0
A.2. Determination of the δ13 CCO2 .aq/
Temperature and pH influence the parameters DT and rk , but the model results proved to be essentially the same even under dramatic changes of these two parameters (i.e. change of temperature š10ºC and change of š1 pH unit). The model is, however, very sensitive to the cell radius, r, because it also changes cell carbon content, c . Increasing r causes the isolines for ¼ to plot further to the right in Fig. 5, thus integrating more samples into the model. However an average cell radius of 10 µm seems a reasonable assumption for the phytoplankton community in the ‘Main Channel’ at Cornwall [28].
References [1] W. Ludwig, J.L. Probst, S. Kempe, Predicting the oceanic input of organic carbon by continental erosion, Global Biogeochem. Cycles 10 (1996) 23–41. [2] W.G. Mook, F.C. Tan, Stable carbon isotopes in rivers and estuaries, in: E.T. Degens, S. Kempe, J.E. Richey (Eds.), Biogeochemistry of Major World Rivers, Scope 42, Wiley, Chichester, 1991, pp. 245–264. [3] R. Pocklington, The contribution of organic matter by the St. Lawrence River to the Gulf of the St. Lawrence, 1981– 1983, in: E.T. Degens, S. Kempe, R. Herrera (Eds.), Transport of Carbon and Minerals in Major World Rivers, Mitt. Geol.-Pala¨ont. Inst. Univ. Hamburg, 1985, pp. 323–328. [4] R. Pocklington, F.C. Tan, Seasonal and annual variations in the organic matter contributed by the St. Lawrence River to the Gulf of St. Lawrence, Geochim. Cosmochim. Acta. 51 (1987) 2579–2586. [5] F.C. Tan, P.M. Strain, Sources, sinks and distribution of organic carbon in the St. Lawrence Estuary, Canada, Geochim. Cosmochim. Acta. 47 (1983) 123–125. [6] M. Lucotte, Organic carbon isotope ratios and implications for the maximum turbidity zone of the St. Lawrence upper estuary, Estuar. Coastal Shelf Sci. 29 (1989) 293–304. [7] M. Lucotte, C. Hillaire-Marcel, P. Louchouarn, First-order organic carbon budget in the St. Lawrence lower estuary from 13 C data, Estuar. Coastal Shelf Sci. 32 (1991) 297– 312. [8] P.H. Gleick, Water in Crisis. A Guide to the World’s Fresh Water Resources, Oxford University Press, New York, 473 pp. [9] Keith, L.H., Compilation of EPA’s Sampling and Analysis Methods, Lewis, Boca Raton, FL, 1996, 1696 pp. [10] J.A.C. Barth, J. Veizer, G.E.M. Hall, Geochemical monitoring of the St. Lawrence River in the Cornwall area. Ecosystem recovery on the St. Lawrence (1996), IREE Research Paper, 23, Institute for Research on Environment and Economy, University of Ottawa, Canada, 13 pp. [11] Calmbach, L., Hydrowin (Computer Program), version 3.0, Institut de Mineralogie, Lausanne, 1995.
J.A.C. Barth et al. / Earth and Planetary Science Letters 162 (1998) 111–121 [12] B.K. Burnison, Modified dimethyl sulfoxide (DMSO) extraction for chlorophyll analysis of phytoplankton, Can. J. Fish. Aquat. Sci. 37 (1980) 729–733. [13] S.W. Jeffrey, G.F. Humphrey, New spectrometric equations for determining chlorophylls a, b, c1 and c2 in equal amounts in higher plants, algae and natural phytoplankton, Biochem. Physiol. Pflanz. (BPP) 167 (1975) 191–194. [14] R.G. Wetzel, G.E. Likens, Limnological Analyses, Springer, New York, 1991, 391 pp. [15] M.L. Fogel, L.A. Cifuentes, Isotope Fractionation during Primary Production, in: M.H. Engel, S.A. Macko (Eds.), Organic Geochemistry: Principles and Applications, Plenum Press, New York, 1993, pp. 73–98. [16] B. Fry, E.B. Sherr, δ13 C measurements as indicators of carbon flow in marine and freshwater ecosystems, Contr. Mar. Sci. 27 (1984) 13–47. [17] J.C. Vogel, Variability of carbon isotope fractionation during photosynthesis, in: J.R. Ehleringer, A.H. Hall, G.D. Farquhar (Eds.), Stable Isotopes and Plant Carbon–Water relations, Academic Press, San Diego, 1993, pp. 29–45. [18] R. Benner, M.L. Fogel, E.K. Sprague, R.E. Hodson, Depletion of 13 C in lignin and its implications for stable carbon isotope studies, Nature 329 (1987) 708–710. [19] B.J.F. Biggs, Biomonitoring of organic pollution using periphyton, South Branch, Canterbury, New Zealand, NZ J. Mar. Freshw. Res. 23, 263–274. [20] D.M. SØballe, R.W. Bachman, Influence of reservoir transit on riverine algal transport and abundance, Can. J. Fish. Aquat. Sci. 41 (1984) 1803–1813. [21] M.H. O’Leary, Measurement of the isotopic fractionation associated with diffusion of carbon dioxide in aqueous solution, J. Phys. Chem. 88 (1984) 823–825.
121
[22] R. Goericke, J.P. Montoya, B. Fry, Physiology of isotope fractionation in algae and cyanobacteria, in: K. Lajtha, B. Michener (Eds.), Stable Isotopes in Ecology, Blackwell, Boston, 1994, pp. 187–221. [23] W.W. Wong, C.R. Benedict, R.J. Kohel, Enzymatic fractionation of the stable carbon isotopes of carbon dioxide by ribulose-1.5-bisposphate carboxylase, Plant. Physiol. 63 (1979) 852–856. [24] C.A. Roeske, M.H. O’Leary, Carbon isotope effects on the enzyme-catalyzed carboxylation of ribulose bisphosphate, Biochemistry 23 (1984) 6275–6284. [25] G.H. Rau, U.H. Riebesell, D. Wolf-Gladrow, A model of photosynthetic 13 C fractionation by marine phytoplankton based on diffusive molecular CO2 uptake, Mar. Ecol. Prog. Ser. 133 (1996) 275–285. [26] I.D. Clark, P. Fritz, Environmental Isotopes in Hydrogeology, Lewis, Boca Raton, FL, 1997, 328 pp. [27] R.D. Pancost, K.H. Freeman, S.G. Wakeham, C.Y. Robertson, Controls on carbon isotope fractionation by diatoms in the Peru upwelling region, Geochim. Cosmochim. Acta 61 (1997) 4983–4991. [28] C. Hudon, S. Paquet, V. Jarry, Downstream variations of phytoplankton in the St. Lawrence River (Que´bec, Canada), Hydrobiology 337 (1996) 11–26. [29] G.D. Farquhar, M.H. O’Leary, J.H. Berry, On the relationship between carbon isotope discrimination and the intercellular carbon dioxide concentration in leaves, Aust. J. Plant. Physiol. 9 (1982) 121–137. [30] W.G. Mook, J.C. Bommerson, W.H. Staverman, Carbon isotope fractionation between dissolved bicarbonate and gaseous carbon dioxide, Earth Planet. Sci. Lett. 22 (1974) 169–176.