Carbon metabolism in a humic lake: Pool sizes and cycling through zooplankton

Limnol. Oceanogr., 35(l), 1990, 84-99 0 1990, by the American Society of Limnology and Oceanography, Inc. Carbon metabolism in a humic lake: Pool siz...
7 downloads 0 Views 2MB Size
Limnol. Oceanogr., 35(l), 1990, 84-99 0 1990, by the American Society of Limnology and Oceanography, Inc.

Carbon metabolism in a humic lake: Pool sizes and cycling through zooplankton Dug 0. Hessen Division of Zoology, Department of Biology, University of Oslo, N-03 16 Oslo 3, Norway

Tom A ndersen Division of Marine Botany, Department of Biology, University of Oslo

Anne Lyehe Division of Limnology, Department of Biology, University of Oslo Abstract To characterize the major carbon pathways in a humic lake, we determined carbon pool sizes and main pathways by long-term tracer studies in enclosures. Dissolved organic carbon (DOC) was by far the largest pool and constituted 80-85% of total carbon. In the water column particulate organic carbon was partitioned between detritus, zooplankton, bacteria, and phytoplankton at ratios of 22 : 4 : 3 : 1. Phytoplankton and bacterioplankton production averaged 24 and 32 pg C liter-’ d-l, while crustacean zooplankton production was very low (~5 pg C liter-’ d-l) during the experiment. Total pelagic community respiration was high, giving a net CO, flu K to the atmosphere of 44 lg C liter-’ d-l, while losses by sedimentation were negligible. Most of the particulate carbon available for zooplankton was highly recycled detritus of low nutritional value. The loop of ingestion and defecation of detrital particles was a major carbon pathway, giving detrital particles a turnover rate of 0.39 d-l. Detritus was found to support 4682% ofbody carbon in the surveyed species, with Acanthudiaptomus as the upper extreme. Bacterial carbon accounted for 1l-42% of body carbon and phytoplankton for 6-19% in the surveyed species.

plankton (bacteria and zooplankton) are often much greater than that of phytoplankton, indicating different pathways of organic carbon or energy flow (Salonen 198 1; Johansson 1983; Hessen 1985a; Salonen and Hammar 1986). Several studies have found zooplankton production higher than what could be supported by algal production alone. To explain this deficiency, workers have proposed the utilization of detrital particles (Naumann 19 18; Nauwerck 1963; Langeland and Reinertsen 1982; Forsyth and James 1984) or dissolved organic matter (Salonen and Hammar 1986). The ability to utilize detrital carbon has been demonstrated for several zooplankton species (see Chervin 1978; Melack 1985), although to our knowledge, no attempt has yet been made to calculate the quantitative imporAcknowledgments This study was supported by grants from the research tance of detritus in the carbon metabolism program on eutrophication of inland waters financed of zooplankton. In humic lakes, utilization by the Norwegian Council for Scientific and Industrial of detrital carbon would be of special inResearch (NTNP). We thank John Korstad, Yngvar Olsen, Olav Vadstein, and two anonymous referees for terest because detritus often constitutes almost 90% of the POC pool. As most of this comments and suggestions on the manuscript. 84

The flow of organic carbon in the pelagic food web is one of the main discriminators of lake types. The general picture from clearwater lakes is of a phosphorus-controlled energy flow, where primary production is a major determinant of production at higher trophic levels. Usually heterotrophic biomass ,and secondary production is strongly correlated with phytoplankton biomass and production, with dissolved organic carbon (DOC, pg C liter-l) released from phytoplankton as the main energy source for heterotrophic bacteria, and phytoplankton particulate organic carbon (POC, pg C liter-‘) as the main energy source for macrozooplankton (Riemann and Sandergaard 198 6 and references therein). In humic lakes, biomass and production of the heterotrophic

Carbon cycling in humic lakes

85

corresponds to 0.14 pg C liter-l or 2 X 1Om3 % of background DOC. Previous laboratory tests (Hessen et al. 1989) showed this amino acid mixture to be almost exclusively incorporated into particles of bacterial size (< 1 pm). To a second bag (CO, bag) we added 14C02(33 kBq liter-l initial activity) to obtain specific labeling of algae. A total of 20 ml of 1 mM C032- was added. The 14C02 was redistilled immediately before addition to remove nonvolatile contaminants from the stock solution. Samples for analysis of isotope distribution were taken eight times during the first 30 h after isotope addition. Then daily samples were taken around noon until the end of the experiment after 11 d. Simple sedimentation traps (made from glass reagent tubes, 15 cm long, 2-cm i.d.) were suspended at 2.5-m depth and harvested on days 5 and 10 of the experiment. Quantitative samples for biomass determination of bacteria-, phyto-, and zooplankton were taken daily. Zooplankton samples for isotopic analysis were taken by vertical net hauls (45~pm mesh size), and quantitative zooplankton samples were taken with a 3-liter Ruttner sampler at 0-, l-, Material and methods 2-, and 2.5-m depths and pooled. All other The experiments were performed during measurements were performed on integratJuly 1985 in Kjelsasputten, a small seepage ed samples taken with a 3-m plastic tube lake located in a coniferous forest area near sampler (3-cm i.d.). The integrated samples Oslo, Norway. The lake is surrounded by were filtered to exclude macrozooplankton bogs, has a mean depth of 8.5 m, and a (>45 pm) before subsampling. The phytosurface area of 0.8 ha. The water is acidic and zooplankton samples were preserved (pH range, 4.5-4.7) and strongly colored by with acid Lugol’s solution; bacterioplankhumic substances (So-150 mg Pt liter-l). ton samples were fixed with Formalin (2% During summer the lake is stratified with a final concn). pronounced thermocline at 2-3-m depth and Supplementary experiments-Additional anoxia below 3-4 m. short-term experiments were done at the Enclosure experiments -Two transparend of the experimental period to suppleent polyethylene bags (l-m diameter and ment the information gained from the long2.7 m deep) were anchored in the pelagic term labeling experiments in the bags. zone and filled with surface water. To comBacterial growth was measured in two sepensate for losses resulting from intensive ries of 250-ml Pyrex flasks, filled and insampling during the first day of the exper- cubated at 0-, l-, 2-, and 3-m depth. The iment, we enriched each bag with lake zoo- first series was screened through a 45-pm plankton from three net hauls (45-pm mesh plankton net to remove macrozooplankton. size, 3 m to surface). Isotope was added 3 The second series was filtered through 3-pm d after the bags were filled. One bag (AA Nuclepore membranes to remove microbag) was labeled with [U-14C] protein hy- zooplankton also. Subsamples for counts of drolysate (Amersham CFB.25), giving an bacteria were taken five times throughout initial 14Cactivity of 22 kBq liter-l. This an 18-h period. As bacterial growth was ex-

detritus is probably allochthonous, it is expected to be more refractory than autochthonously produced detritus. In such detritus-based ecosystems, detritus may support production of higher trophic levels either directly by assimilation of ingested detrital particles or indirectly through a bacterial food web. If detritus is poorly assimilated, and the main path of energy transfer is through bacteria, substantial respiratory loss of energy may occur toward higher trophic levels (Moran et al. 1988). The main purpose of this study was to characterize the major carbon pathways in a humic lake, with special reference to the role of zooplankton. Varying abilities to feed on bacteria have been demonstrated among species of crustacean zooplankton (Gophen et al. 1974; Pace et al. 1983; Forsyth and James 1984; DeMott 1985; Hessen 19856; Hessen et al. 1989). A second aim was to calculate the amount of phytoplankton vs. bacterial carbon incorporated in zooplankton of different species. A third aim of our study was to attempt to quantify the amount of detritus incorporated into zooplankton body carbon.

86

Hessen et al.

ponential in all bottles, specific growth rates were estimated by linear regression of logtransformed bacterial counts against incubation time. Primary production was measured by standard 14Cmethods (Vollenweider 1969). Light and dark bottles were incubated for 6 h (from 1000 hours) both in the lake (eight depths) and the CO,-labeled bag (six depths). High-specific-activity H14C03 was used to mask the background activity in the 14Clabeled bag while not perturbing the total CO, in the bottles. Zooplankton grazing experiments were performed in situ with labeled particles in both bags and in the laboratory with monodisperse latex beads and double-labeled (3H and 14C)natural particles. Diurnal grazing rates by different species and stages of zooplankton were measured. Details in experimental setup, analytical methods, and results of the grazing experiments will be given elsewhere (Hessen et al. 1989). Isotope loss was measured on zooplankton from the bags at isotopic equilibrium. Animals were transferred to cages covered with 45-pm plankton net, allowing them to graze on unlabeled lake seston. Samples were taken initially and after 24 and 48 h and processed the same way as animals from the grazing experiments. An,alytical methods -All radioactivity measurements were done on a Philips PW 4540 liquid scintillation counter with the preprogramed window settings for 14C. Counting efficiency for each sample type was calibrated with commercial 14C standards (Packard Instand). Quenching, which was generally low and stable within each sample type, was corrected with the external standard channels ratio method. All organic carbon samples were analyzed on a Carlo-Erba CHN 1106 elemental analyzer.. CO, in the exhaust from the CHN analyzer was trapped directly in vials containing a CO,-absorbing scintillation fluor (Carbomax; Lumac). The absorbent was carefully tested with [ 14C]sucrosestandards and found to capture >98% of evolved 14C02. This high extraction efficiency permitted simultaneous measurement of 14C and 12Con the sample and gave high precision in the calculation of specific 14Cac-

tivity. Specific activity is reported as the atomic ratio of 14Cto 12C[8.67 x 1O-8 x

dpn-@gWI.

Samples for determination of dissolved inorganic carbon (DIC, pg C liter-l) were prepared in the field in loo-ml serum vials. A 20-ml volume was removed and one drop of 6 N H2SO4 (giving pH= 2) added immediately before sealing. Although the CO, content in 20 ml of outdoor air is negligible in this context, care was taken to not contaminate the headspace with exhaled air. Samples were purged with CO,-free N2, and the evolved CO, directed through an Uras infrared CO2 analyzer connected to a stripchart recorder and an integrator. The outlet from the CO2 analyzer was directed through the same type of CO, trap as used with the CHN analyzer, so that 14Cactivity could be measured on the same sample. The DIC analysis was calibrated with Na,CO, standards, and the exact volume of the sample was found by weighing the vial after analysis. Samples for determination of DOC were screened through Whatman GF/F-filters, acidified to pH ~2, and purged with air for > 1 h to drive off all Dl C. A 2-ml subsample was evaporated on a strip of preignited glassfiber filter and total DOC determined by CHN analysis. Radioactivity in the same fraction was measured directly on a lo-ml subsample mixed with an equal volume of gel-forming, hydrophilic scintillation fluor (Lumagel; Lumac). POC was collected on washed, preignited Whatman GF/F filters and determined by CHN analysis. For a consecutive determination of isotope distribution in the particulate fractions, total particulate 14Cactivity was measured on 0.45-pm membrane filters (Gelman GN-6) while the activity retained by 3-pm polycarbonate membranes (Nuclepore) was taken as a bacteria-free fraction. All filters were counted in SafeFluor (Lumat). Zooplankton samples from the grazing and isotope loss experiments were dissolved in Lumasolve (Lumac) and counted in a lipophilic scintillation fluor (Lipoluma; Lumac). Bacterioplankton samples were stained and counted according to the AODC method (Hobbie et al. 1977). Cell volumes were

87

Carbon cycling in humic lakes

determined on 100 randomly chosen cells on each filter. The volume estimates were calibrated with monodisperse fluorescent beads of known dimensions. Mean volumes were converted to carbon assuming 12% of wet weight (wt/wt) is carbon (Watson et al. 1977). Phytoplankton were identified, counted, and measured in an inverted microscope. Biovolumes were calculated from geometrical models and converted to carbon assuming 10% of wet wt is C (Vollenweider 1969). Zooplankton were identified and counted under a binocular dissecting microscope. Abundance data were converted to carbon biomasses from carbon measurements on unfixed animals (CHN analysis). For determinations of individual biomasses and specific activities, zooplankton were transported live to the laboratory and anesthetized in carbonated water. From a counting tray under a binocular dissecting microscope, animals were transferred directly into tin capsules with an Irwin loop. Each sample consisted of up to 40 individuals of each species and stage (juveniles or adults). All samples were air-dried for 24 h (30°C) before CHN analysis. The high sampling frequency on the first day made processing of live animals impossible. The values from day 1 are therefore based on specimens fixed in Lugol’s solution. Fixed animals were carefully washed in tapwater before analysis and otherwise processed identically to fresh animals. The relation between specific activity (14C: 12C x 1O6atomic ratio) in fresh and . fixed zooplankton samples from the whole experimental period was not significantly different from a 1 : 1 ratio (Cfresh= - 1.2 + 1.04 Gixecl,r = 0.98, n = 49). Results Carbon pools-An

overview of the different carbon pools is given in Fig. 1, where relative pool sizes are based on averages from both bags. Of the three major fractions, DOC constituted 80-85% of the total carbon [DOC: 7,500+200 (SD) pg C liter-l]. Mean DIC was 570*60 pg liter-l, and total POC (~45 pm) was close to 350 pg C liter-l in both bags. The different pools remained fairly stable (Table l), and a mean

Table 1. Mean and standard error of mean (n = 12) of the particulate pools in the two bags given as pg C liter-l. SEM

CO, bag Detritus Zooplankton Bacteria Algae AA bag Detritus Zooplankton Bacteria Algae

354.4 61.7 45.1 16.1

20.2 6.7 2.0 1.5

353.5 62.0 53.9 12.9

28.8 8.5 3.5 0.9

sestonic atomic C:N ratio of 15.5 + 1.O was found in both bags. Although mean phytoplankton biomass in the lake was close to 30 pg C liter-l at the start of the experiment, biomasses in the two bags varied between 10 and 25 pg C liter-l, with mean values of 15.7 and 12.9 pg C liter-l. This disparity was probably due to the slight zooplankton enrichment in the bags compared to the lake. Bacterial biomass was close to 100 pg C liter-’ in the upper 1.5 m of the lake and < 30 in the 23-m layer. Bacterial biomass in the bags was in the range of 40-60 pg C liter-’ during the experimental period. Microscopic examination of stained bacteria and organic matter on 0.45-pm filters revealed that ~5% of the bacterial biomass was attached to particles. Crustacean zooplankton biomass was measured in the lake both initially and on day 7 during the diurnal grazing experiment (Hessen et al. 1989). In the lake, mean biomass increased from 40 to 55 pg C liter-l during this week. Zooplankton biomass in the bags remained fairly stable and oscillated around 60 (range 48-80) pg C liter-l. Neither of the analyzed carbon pools was significantly different between the two bags. Time-course of isotope labeling- In the AA bag, labile DOC (as added [ 14C]proteinhydrolysate) was depleted from the initial 1,300 dpm ml-l to a stable level at -200 dpm ml-’ within 24 h.The major activity was incorporated into the particulate fraction within the first few hours (Fig. 2A). Almost the entire uptake in this bag was within the bacterial size fraction (Fig.

88

Hessen et al.

DIC \

I

I CrustaceanPOC Daphnia

IAlgal POC

Bosmina

Fig. 1. Relative pool sizes of the different epilimnetic carbon fractions, and the species composition within the phytoplankton and zooplankton communities after zooplankton enrichment. The algal groups are Chrysophyceae, Dinophyceae, and Chlorophyceae. Species are Acanthodiaptomus denticornis, Holopedium gibber-urn, Diaphanosoma brachyurum, Daphnia longispina, and Bosmina longispina.

2B). Corresponding to the depletion in the DOC fraction, maximal activity in the particulate fraction was obtained within 30 h. At this time, 69% of the initial activity remained in the water, but ~40% (500 dpm ml-l) remained at the end of the experiment. DIC oscillated around 200 dpm ml-l t.hroughout the experimental period. The major fraction was probably lost through respiration (leaving as DIC). Adsorption to the walls of the bag could be another source of isotope loss as could sedimentation. Activity in the DOC pool remained stable from

-30 h onward. This level might be a real steady state but could also result from contamination by unavailable components in the added isotope. In the CO2 bag, there was a corresponding 52% reduction from the initial activity within 8-9 d, indicating a rapid efflux of CO, from the water (Fig. 3A). The major pool, DIC, decreased from almost 2,000 to 840 dpm on the last date. Only a small fraction of added isotope was incorporated into the particulate fraction. The total particulate fraction (0.45 pm < POC < 45 pm) exhib-

89

Carbon cycling in humic lakes

loo0

500

POC z-3vm o-l.,

0 0

2

4

6

8

10

1:

0

I. 2

I. 4

I. 6

I. 8

4. 10

2

Time (days)

Time (days)

Fig. 2. Isotope distribution in the dissolved (A) and particulate (B) compartments of the [14C]amino-acidlabeled bag.

ited rapid isotope incorporation within the first 2 d (Fig. 3B) and then reached an approximate steady state at 40-60 dpm ml-l. This level corresponds to 4-6% of total activity added. The slight increase could be accounted for by a slight increase in algal biomass and labeled particulate feces. Since a considerable part of the algae passed the 3-pm filters, the > 3-pm fraction constituted ~25% of the total particulate activity but experienced the same pattern of isotope incorporation. In both bags, there was good agreement (> 88%) between measured total activity in water and the sum of activity in the DOC, DIC, and POC pools. Radioactivity in the particulate fraction remained considerably lower in the CO1 bag than in the AA bag, although the COZ bag received the highest total activity. This difference was partly caused by the different sizes of the labeled precursor pools. The pool of free amino acids was presumably small

compared to the large CO2 pool. The high bacterial biomass compared to that of algae would also contribute to the difference. Flows-Major loss rates (sedimentation, community respiration, and grazing) and production rates (bacteria, phytoplankton, and zooplankton) were calculated from isotope distributions. Specific loss rates of [14C]POC, as measured in the sediment traps, was 0.030 d-l in the CO,-labeled bag and 0.0 17 d-l in the amino-acid-labeled bag. If we assume all isotope to be located in either algae or bacteria, this sedimentation loss corresponds to rates of 0.5 pg C liter-l d-l for algal POC and 0.9 for bacterial POC. A redistribution of isotope activity over other particulate pools would make these numbers maximal estimates. There was no visible growth of attached algae or bacteria on the bag walls during the experiment. The DIC concentrations were stable, both in the bags and in the lake, and correspond60

04. 0

, . , 2

4

, . , 6 The (days)

8

B

, . 10

1

4 Tii

Fig. 3. As Fig. 2, but of the 14C0,-labeled bag.

6 (days)

8

10

12

90

Hessen et al. pg C iii er-’ d-’ x(t) = 428 exp(-1.061 t) + 1,645 exp(-0.063 t) 0

0

10

20

30

40

50

1 33 b 2 750

1

5004 0

: 2

4

l Primary production 0 Bacterial production

3

v---v 6

8

10

12

Time (days)

Fig. 4. Double exponential fit to DIC activity in the 14CC),-labeledbag.

ed to 1’400% supersaturation of CO,. This level indicates a major efflux of CO, to the atmosphere, balanced by internal production In the CO,-labeled bag, for example [ 14C]DIC activity was reduced by more than 40% during the first 6 d of the experiment (Fig. 3A). This loss was obviously not a result of incorporation into particles, as > 90% of the 14C activity was in the DIC pool throughout the experiment. If we assume that the other losses (sedimentation and wall growth) are negligible compared to CO2 efflux, the tracer dynamics can be described as a single-exit compartmental (SEC) system (Anderson 1983). In an SEC system without traps, where the exit compartment is also the initially labeled one (DIC in our case), the specific efllux rate can be calculated from Hamilton’s formula (Rubinow 1975). If x(t) is the isotope activity in the DIC pool, this formula can be written as specific efflux rate = x(0) [la

Fig. 5. Depth distribution of primary production (from the CO, bag) and bacterial production (from the AA bag).

face (20°C) production was 37 pg C liter-l d-l (Fig. 5). Below the thermocline (8” 1OOC) production was reduced to 18 and 16 pg C liter-’ d-l at depths of i! and 3 m. Integrated bacterial biomasses in the bags were slightly higher and fairly constant at 45 -+2 pg C liter-l. If steady state is assumed, a bacterial net production of 32-1.3 pg C liter-l d-l can be calculated, corresponding to an integrated production of > 80 mg C m-2 d-l. At all depths, there was a noticeable dif-ference in bacterial growth rate between bottles with 45- and 3-pm prefiltered water (Fig. 6). It can be interpreted consistently as due to the presence of bacterial predators in the size range 3-45 pm. As ciliates were scarce both in the ba.gs and in the lake, phagotrophic feeding by chrysophytes of the genera Chromulina and Monochrysis seems the most likely source of predation. The difference in the observed growth rate in the 3- and 45-pm prefiltered water indicates a

x(t) dt],l,

where x is in dpm. In order to calculate the integr.al in the denominator, we fitted a sum of two exponential functions to the observed time series of 14Cactivity in the DIC pool (Fig. 4). This fit gave a specific loss from the DIC pool of 0.078 d-l (asymptotic SD = 0.0 1) or an efflux of 44+6 pg C liter-l d-l to the atmosphere. Specific growth rates of bacteria averaged 0.70k0.06 d-l in the bottle incubations. From bacterial biomass in the bottles, sur-

0.0

0.2

0.4

0.6

0.8

1.0

3 lrn Filtered

Fig. 6. Specific growth rate of bacteria at different depths in bottles with lake water prefiltered through 4% and 3-pm filters. Error bars indicate the standard errors of the slopes of the regression lines.

91

Carbon cycling in humic lakes

grazing rate on bacteria of 11+ 5 pg C liter-’ d-l. Although fl agellates like Monochrysis and Chromulina probably would be inactivated or damaged by passing the 3-pm filter, some of them may have remained active in the filtrate. If so, the 11 pg of C represents an underestimate of true flagellate grazing. Primary production measurements from bottle incubations gave very similar results in the lake (22 pg C liter-’ d-l) and in the bags (25 pg C liter-* d-l). Depth profiles (Fig. 5) showed a rapid decline in primary production from almost 45 pg C liter-’ d-l in the surface layers (O-l m) to < 10 below 1.5 m. Below 3 m, primary production was negligible. An independent estimate can be made from initial isotope uptake in the 14C0,-labeled bag (Fig. 3B). The linear increasesin particulate isotope activity during the first and second day of the experiment correspond to primary production rates of 2 1 and 27 pg C liter-l d-l. The close agreement between the two methods suggeststhat the containment effects from bottle incubations were negligible. The average of all four estimates is 24+ 3 pg C liter-l d-l, or 65+8 mg C m-2 d-l. With a mean algal biomass of 16+ 2 pg C liter-l it corresponds to a specific growth rate 5 0.9 d-l, with strict equality attained in the casewhere the measured primary production equals net production. Zooplankton egg production was negligible, and, as the biomass of Daphnia and Bosmina remained stable, net production in these species was very low. From the quantitative zooplankton data, mean daily biomass increase in Holopedium, Acanthodiaptomus, and Diaphanosoma together was estimated to be 1.7& 1.O pg C liter-’ d-l. Nonpredatory natural mortality must have been the main loss in the zooplankton community, as the only zooplankton predator in this system, Heterocope saliens, was virtually absent in the bags. From the small amounts of labeled organic matter caught by the sedimentation traps, the loss by natural mortality could maximally be of the same magnitude as the bacterial and algal sedimentation losses. Zooplankton net production could thus at maximum amount to 3 pg C liter-l d-l, corresponding to a specific growth rate SO.05 d-l.

Zooplankton gross production can be estimated from the isotope loss experiments. Assimilation and respiration rates can be assumed to be unchanged when labeled animals are transferred to cageswith unlabeled food at the same density. The specific activity (SA) of the animals will decrease as a result of isotope dilution by the unlabeled food, while respiration has no impact on SA. If u(t) denotes the specific activity of a given species, the isotope dilution process can be described by du -z----u

dt

P B 0

(2)

where P : B is the gross production per unit biomass and u the 14C: 12Cratio of a compartment. Solving this equation gives the following estimator for the gross P : B ratio P E

0

In u2 - In u, =-

t2

- t1

(3)

where Ui is the specific activity at time ti. Only Acanthodiaptomus, Holopedium, and Daphnia were present in sufficient quantities in the isotope loss experiments to permit this analysis. The biomassweighted, average gross P: B ratio in these three species was 0.30+0.02 d-l. Subtracting the upper limit of our specific growth rate estimate gives a minimum specific respiration rate of -0.25 d-l, comparable to measurements reviewed by Lampert (1984). If we assume negligible growth rate and that these rates apply to Bosmina and Diaphanosoma also, the maximal respiratory carbon loss rate from the zooplankton community would be 18+3 pg C liter-l d-l. Zooplankton

utilization

of carbon pools -

Differences in the utilization of the labeled carbon pools in the two bags (algae and bacteria) are reflected by the time-courses of the specific activities in the different zooplankton species. In the 14C02-labeled bag (Fig. 7A), Holopedium obtained the highest SA, clearly exceeding that of Diaphanosoma and Daphnia, while the lowest specific activities were found in Acanthodiaptomus and especially Bosmina. In the [14C]amino-acidlabeled bag (Fig. 7B), the SA was highest among Daphnia and Diaphanosoma (both juveniles and adults), while Holopedium had

93

Carbon cycling in humic lakes the trapezoidal rule). The regressions were taken over the time interval where the relationship was linear (l-3 d) as judged from graphical displays. The estimated net clearance rates imply that algae and bacteria together can supply only from 17 (in Acanthodiaptomus) to 47% (in Daphnia) of a respiratory carbon demand of 0.3 d-l (Fig. 8A). The rest of the respirational energy requirement must be acquired from other sources, most probably detrital carbon. As the calculation of net clearance rates does not take feedback into account, it will somewhat underestimate the true rates. This consideration implies that the calculated contributions from detrital carbon by this procedure will probably be maximal estimates. Although this first calculation does not rely directly on specific activity of the food pools (Eq. 6), a second independent, but less accurate, calculation can be obtained from the SA of algae, bacteria, and zooplankton. From an average specific gross production of N 0.3 d-l, the zooplankton should have reached nearly 95% of isotopic equilibrium with their food after 10 d. This close approach holds even for the slowest growing species, Acanthodiaptomus, which had a specific gross production close to 0.3 d-l. At isotopic equilibrium, the SA of a given species should be proportional to the specific activities of the food items and their contributions to the animals’ body carbon. If we assume that all particulate isotope activity is located in either algae or bacteria in the two bags respectively, the ratios of SA in a zooplankton species compared to algal or bacteiial SA will be estimates of the contributions of algae and bacteria in the diet of that species. As 12C in the different food pools could not be separated from each other on the filters (contrary to the SA calculations for zooplankton), 12C of algae and bacteria was based on quantitative counts. SA was calculated as the atomic ratio of 14C : 12C[8.67 x lo-* x dpm(pg C)-l]. Although SA of algae remained constant (N 250), SA of the bacteria decreased almost 50% (560-350) during the experimental period. The calculated contribution of bacterial C is thus a tentative estimate.

A

Initial

B

Steady

uptake

state

H.g. D.b.

4 0

:

: 20

Bacteria

!

: 40

:

Algae

: 60

:

0

: 80

:

1 100 K

Detritus

Fig. 8. Relative contributions of bacterial, algal, and detrital carbon sources incorporated as zooplankton body carbon. A. Calculated from initial labeling with net clearance rates, assuming 30% respiration and zero production. B. Calculated from specific activity in sestonic fractions and animals at isotopic equilibria. (See text for explanation.) Abbreviations as in Fig. 7.

Based on this assumption, we estimated that 6-l 9% of body carbon was of algal origin and 1 l-42% of bacterial origin in the different zooplankton species (Fig. 8B). In accordance with Fig. 7, the contributions from algal and bacterial carbon were lower in Acanthodiaptomus than in the cladocerans. The remaining 46-82% of body carbon would be expected to be of detrital origin. The persistent and gradual increase in SA in some species is probably caused by gradual labeling of the detritus pool by zooplankton feces of algal or bacterial origin. Hence, the detritus pool cannot be considered totally unlabeled, although the SA should be significantly lower than that of the primary labeled compartments (algae or bacteria). This labeling would give an overestimate of the contributions from algal and

94

Hessen et al.

Table 2. Gross and net specific clearance rates, ml (pg C)--I d-l, on bacteria and algae in the dominant species. Gross clearance rates (GCR) taken from Hessen et al. (1989). Net clearance rates (NCR), estimated by calculations described in the text. Assimilation efficiency (AE) is the ratio between NCR and GCIR. ----~~

-A

Z

Bacleria GCR

Acanthodiaptomus denticornis Hol’opedium gibberum Diaphanosoma brachyurum Daphnia longispina

p---p

1.0 5.5 5.5 5.8

bacterial carbon, and thus Fig. 8B presents minimal estimates of the amount of body carbon arising from detrital particles. The values as given in Fig. 8, although in basic agreement, represent maximal and minimal estimates of the contributions from detrital carbon. The ratio of net clearance rate to gross clearance rate, as measured in the shortterm grazing experiments (Hessen et al. 1989) gives an estimate of the assimilation efficiency on a given food source. Estimated assim:ilation cficiencies (Table 2) were in most cases lower for bacteria than for algae, with bacteria ranging from 10 to 32% and algae from 15 to 70%. Diaphanosoma and Daphnia had the highest assimilation efficiencies, both on algae and bacteria. Since - 50%) of total POC was < 3 pm, one would expect roughly half of the detritus to be in a bacterial size range and the gross clearance rates on detritus to be intermediate between those obtained on algae and bacteria by Hessen et al. (1989). This crude approximation leads to estimates of assimilation efficiency on detritus of from 6 to 14%. Although assimilation efficiencies in Daphnia and Diaphanosoma fit well with previous estimates (Hessen et al. 1989), the very low assimilation efficiencies on algae found for Acanthodiaptomus and Holopedium are somewhat surprising. If the algal assimilation is underestimated in these species, the contribution of algal carbon in Fig. 8A is underestimated. Discussion The presented results and discussion are based on measurements in enclosed water volumes. Each of the two treatments is unreplicated. The justification for discussing them is the consistency of results within and

NCR

0.2 0.6 1.8 1.7

Abe AE (46)

24 10 32 30

---

GCR

NCR

9.6 14.6 4.1 6.4

1.4 3.4 2.7 4.5

GE (%,

15 23 65 70

‘between treatments and between the enclosures and the lake. We think the enclosures are fairly representative for the epilimnetic ecosystem of this locality. The lake exhibited a strong thermocline between 2 and 2.5 m and was anoxic below 4 m. Oxygen content decreased rapidly below 3 m, which was close to the lower depth of zooplankton distribution. In general the abiotic and biotic parameters in the bags closely resembled those of the lake epilimnion. Moreover, no growth of epiphytes was observed on the bag surface during the experiment. Due to the low pH in the la:ke, which may slow down biological processes, the general inferences from this studly should be confined to acidic, humic lakes, Pool sizes-In Kjelsasputten, DOC was by far the largest carbon pool and exceeded DIC and POC by a factor ~8. The DOC concentration in the epilimnion (7.5 mg C liter-l) exceeds reported values from both autochthonously and allochthonously dominated systems (Jordan and Likens 1975; Ward and Wetzel 1984), but is comparable to the high concentrations reported from small bog pools and highly humic systems (Thurman 1985; Satoh et al. 1987). The POC concentration (range, 0.5-0.7 mg C liter-‘) is comparable to studies in other humic or allochthonously dominated lakes (Johansson et al. 1976; Hart 1986), but clearly exceeds those of clear-water, oligotrophic lakes (Saunders 1971; Jordan and Likens 1975; Persson 1985). A main characteristic of our study lake is a POC pool made up by > 7 5% detritus. The biomass of crustacean filter feeders equals bacterial biomass and exceeds algal biomass by a factor of 3-6 in the bags and a factor of 2 in the lake. Large bacterial biomasses exceeding those of phytoplankton have also

95

Carbon cycling in humic lakes 5 g

Dissolved Organic Carbon

g

Detrital Particulate Organic Carbon

44 (+6)

ag 4

7500 (Lmo)

350 (f20) A

Dissolved Inorganic Carbon 140 (f40)

r1401

32 (k3) +[24 - X] 570 (MO) L 7

1261

24 (13)

7 Crustacean Particu$a;bo($ganic

Bacterial Particulate Organic Carbon 45 w>

w

610% 7 Algal Particulate Organic Carbon 16 (f2)

8

P8 g

1 pg C liter-l or 3% of the measured bacterial production. With a reasonable choice: of bacterial growth efficiency, it is likely that < 10% of the bacterial production is directly supported by algal primary production. The main sources of bacterial carbon are therefore either allochthonously produced DOC or detrital POC, although our data give no simple ways to determine the relative importance of these two pools. For cl.arity, we have chosen the simplest solution in the flow-chart in Fig. 9 and assumed that DOC is the only substrate for bat terial growth. The measured primary production (24 hg C liter-l d-l) is comparable to that of other humic localities (Salonen 198 1; Johansson 19 8 3; Arvola 19 84). Primary production normalized to the standing crop of phytoplankton was high, as is to be expected for a system with high grazing pressure. Although bacterial production in clear-water lakes has bee-n reported to equal 18-63% of primary production (annual basis) (Jordan and Likens 1980; Love11 and Konopka 1985; Scavia and Laird 1987), bacterial production in the present study averaged 133% of primary production. Primary production is clearly exceeded by the CO2 production of the system (44 pg C liter-l d-l). Salonen et al. (1984) reported a similar result in two strongly humic lakes where a total plankton respiration of lOO200 ELgC liter-l d-l greatly exceeded primary production. This consistent finding indic,ates the important role of allochthonously produced DOC as an energy source in these lakes. In the isotope dilution estimate of total CO2 efflux, we cannot distinguish between biological CO2 production (respiration), and photooxidation or other chemical pathways. At steady state, chemical and biological CO2 production must sum up to the CO2 efflux. With phytoplankton net respiration loss equaling 2 pg C liter-’ d-l and a zooplankton respiration of 18 pg C liter-l d-l, the sum of chemical and bacterial CO, production must be 24 pg C liter-’ d-l. This corresponds to specific bacterial respiration of 10.53 d-l, where the maximal estimate is attained in the case of no photooxidation

of humic compounds. The possible flow due to photooxidation is indicated as “X” in Fig. 9, although its malgnitude is small and it has no important implications to overall carbon cycling. Community clearance rates-With a mean zooplankton biomass of 60 pg C liter-r, the bag would be cleared (Table 2) at a rate of 0.57 d-l for particles of algal size and 0.21 d-l for particles of bacterial size. About 50% of detrital carbon was found to be ~3 pm, and thus expected to be cleared at the same rate as bacteria. On the basis of this assumption, detritus would be cleared at a rate of 0.39 d-l. The total epilimnetic macrozooplankton community would on average remove 9 pg of algal C and 10 pg of bacterial C liter-’ d-l. With an estimated zooplankton respiration loss of 18 pg C liter-l, gross daily community intake of algae and bacteria would balance respiratory losses provided 100% assimilation. Although zooplankton production was low, it is evident from the obtained assimilation rates that algae and bacteria alone are insufficient even for respiratory demands in zooplankton. From the bacterial production estimates, heterotrophic eucaryotes < 45 pm seemed responsible for daily removal of 11 pg C liter-l, corresponding to the recorded bacterial grazing by macrozooplankton. As microscopic examination only occasionally revealed free-living ciliates, we suspect facultative phagotrophic activity by algae (cf. Bird and Kalff 1987) to be responsible for most of this removal of bacterial cells (Fig. 9). A discrepancy of - 11 pg C was found between daily net production of bacteria and removal by grazing and sedimentation. As bacterial biomass remained stable throughout the experimental period, this discrepancy suggests either underestimated grazing pressure or overestimated production rates. As some of the phagotrophic flagellates may have paseed the 3-pm filter, true rates of flagellate grazing may be underestimated. It seems likely that algal phagotrophic activity is an important pathway of energy in such nutrient-deficient and lightlimited systems with high bacterial biomass. The utilization of DOC and POC by zoo-

Carbon cycling in humic lakes plankton - Salonen and Hammar ( 19 8 6) suggested that a major fraction of incorporated carbon in zooplankton from humic lakes originated directly from DOC. Although our calculations are based on the assumption that zooplankton body carbon primarily originates from particulate sources, we cannot reject the possibility of some incorporation of DOC. Molecular weights in aquatic humic substances range from 500 to > 100,000 Daltons, of which the larger makes a significant contribution to the POC pool (Thurman 1985; Satoh et al. 1987). The amorphous nature of fluorescent humic detritus can easily be revealed on a Nuclepore membrane in the bacterial counts, and obviously the humic matter also contributes to the particulate fraction (>0.45 pm). The amorphous and collodial nature of humic compounds thus makes the distinction between the “dissolved” and “particulate” fractions arbitrary (cf. Ward and Wetzel 1984). In our study lake, the rapid cycling of humic detritus and autochthonous detritus further obscures these borders. Salonen and Hammar (1986) found the amount of carbon from primary production incorporated by zooplankton to be negligible. We found strongly species-dependent incorporation of algal carbon, although it never exceeded 30% of body carbon in any species. From Fig. 8, the importance of detritus as a source of carbon is evident and also strongly species-dependent. As previously stated, since detritus is calculated as unlabeled carbon, gradual labeling of this pool will cause underestimation of the actual relative contribution of detritus in Fig. 8A. Correct measurement of respiration rate is a prerequisite for Fig. 8B and might bias the results, as it is pooled for the whole zooplankton community. Still, these minimal and maximal estimates are quite consistent. Two possible sources of error should be mentioned. First, toward the end of the experiment, gradual cross-labeling may occur, i.e. respired label from bacteria may be incorporated by algae in the AA bag, and labeled algal exudates may be incorporated in bacteria in the CO, bag. This process would not alter Fig. 8A, which is based on the initial labeling, but could influence the rel-

97

ative contribution of bacteria and algae in Fig. 8B. Second, our 12% conversion factor from bacterial wet weight to carbon (cf., Watson et al. 1977; Vadstein et al. 1988) may underestimate true biomass (cf. Kogure and Koike 1987; Lee and Fuhrman 1987). If so, the relative contribution of bacterial carbon to overall energy flux and to zooplankton body carbon may be even larger than suggested by our data. Acanthodiaptomus and probably Bosmina obtain most of their body carbon from detrital carbon. This conclusion is consistent with the findings that these species sustain high population densities in the lake for several weeks after the disappearance of the other cladocerans (Hessen pers. obs.) and that they seem well adapted to otherwise poor food conditions. Laboratory experiments showed that Bosmina was able to reproduce and survive in lake water at 20°C in complete darkness for several weeks, while the other cladocerans survived the same treatment for only - 1 week (Hessen unpubl. data). These observations thus give support to the early suggestions of Naumann (1918) concerning the direct utilization of humic compounds by zooplankton. The results also support the common observation of a deficiency between available phytoplankton carbon and zooplankton carbon requirements even in nonhumic lakes, suggesting that bacteria and detritus must act as important additional carbon sources (Nauwerck 1963; Langeland and Reinertsen 1982; Forsyth and James 1984). On the basis of the community clearance rates, 39% of the detritus fraction was ingested each day, suggesting a turnover time of 3 d. It implies pronounced recycling of particulate matter and may explain the steady increase in specific activity of the zooplankton in the CO,-labeled bag, even after expected isotopic equilibrium. Olsen et al. (1986) found that particulate carbon released from Daphnia pulex could, in extreme cases, equal up to 80% of ingested algal carbon. Zooplankton would then mainly graze on detrital particles of fecal origin supplemented with bacterial and algal carbon. The lake is normally loaded with “fresh” allochthonous POC and DOC during vernal and autumnal floods, as reflected

98

Hessen et al.

in higher egg production in the zooplankton during these periods (Hessen 1989). As the season progresses, the pools of al1ochthLonous and autochthonous matter become mixed. ‘The present summer situation is characterized by a large pool of recycled detritus and low zooplankton production. Our findings support those of Salonen and Hamrnar (1986), who also found detrital carbo:n to be essential in zooplankton nutrition. It apparently contradicts, however, the data of Moran et al. (198 8) who found that

Suggest Documents