The Fate of Organic Carbon

Abstract

In Lake Kinneret, the majority of photosyntetically produced organic carbon (OC) is cycled through the microbial loop. Taken together, bacterial production (BP) and bacterial respiration (BR), i.e., bacterial carbon demand (BCD), accounted for about 65 % of gross primary production (GPP), measured biweekly and averaging 2.3 g C m^–2 day^–1 during the last decade (2001–2011). Community respiration (CR) was 2.1 g C m–2 day–1. The major contributors to total CR were bacterial and phytoplankton respiration (~80%) while zooplankton respiration accounted for the reminder. Most (~ 83 %) of the OC input were eventually respired, ~3 % lost to outflows, while ~15 % of the total OC input were transferred annually to the sediments. Here oxic mineralization is gradually replaced by anoxic processes as a function of the availability of suitable electron acceptors. After the depletion of oxygen in the hypolimnion, sulfate (500 μM) becomes the dominant oxidant. Depending on the settling flux of OC sedimentary sulfate reduction (SR) rates were measured from 0.01 to 1.67 µmol cm–3 day–1 in December and July, respectively. SR is the dominant anaerobic terminal decomposition process in Lake Kinneret and is responsible for the accumulation of sulfide in the hypolimnion to concentrations up to 400 μM. Methanogenesis is restricted to those sediment layers that are depleted of sulfate (below 3–5 cm). Seasonal profiles and 13C signatures of dissolved methane in the sediment pore water of Lake Kinneret have indicated anaerobic methane oxidation in the deeper sediments (below 20 cm), with Fe(III) as electron acceptor. Lake Kinneret resembles the first aquatic ecosystem where the existence of this process could be verified. Changes in the watershed and lake environment are suggested as possible causes for the apparently significant declines in bacterial numbers, BP, and BCD that have taken place over the last decade in Lake Kinneret.

1 Heterotrophic Bacterial Production, Respiration, and Growth Efficiency

Tom Berman, Arkadi Parparov, and Yosef Z Yacobi

1.1 Heterotrophic Bacteria in Organic Carbon Processing

As in most lakes, so too in Lake Kinneret, the metabolic activities of aerobic heterotrophic bacteria are major drivers of organic carbon (OC) cycling (Cole et al. 1988; del Giorgio and Williams 2005a). In Sect. 15.1, the patterns of bacterial abundance (BA) based on 4′,6-diamidino-2-phenylindole (DAPI) counts and some phylogenic and morphological data relating to Lake Kinneret bacteria are presented. Here, we shall show data concerning various aspects of bacterial metabolic activity and their far-reaching impacts on OC cycling of the lake.

1.2 Bacterial Production

The first set of monthly bacterial production (BP) measurements were made using the ³H-thymidine method (Riemann et al. 1982) from June 1988 through May 1993 at five depths at Station A (location in Fig. 1.1 of Chap. 1). Routine measurements, using ¹⁴C-leucine (Kirchman et al. 1985) were resumed only in 2001 on samples collected at Sta. A from 1, 5, 10, 20, 30, and 40 m and mid-thermocline depth. (For the ³H-thymidine method, empirical conversion factors were used; literature values were used for the ¹⁴C-leucine method.) The annual means for the two data sets (1988–1993 and 2001–2011) are shown in Fig. 25.1. Although the mean annual BP values for both periods appear to be very similar, it should be noted that during the first period sampling was much less frequent or complete than in 2001–2011. The former data set was used for the first models of carbon cycling in Lake Kinneret that emphasized and quantified the central role of the bacteria in carbon flux in this ecosystem (Stone et al. 1993; Berman and Stone 1994; Hart et al. 2000).

Fig. 25.1

Annual means (mg C m⁻² d⁻¹) and standard deviations of bacterial production (BP) in the euphotic zone, 0–15 m at Sta. A. a From 1988 to 1993, measured with ³H-thymidine. b From 2001 to 2011 measured with ¹⁴C-leucine

From 2001 to 2011, the mean annual BP in the lake epilimnion was 558 ± 179 mg C m⁻² day⁻¹ with annual values ranging from 902 to 1,320 mg C m⁻² day⁻¹ in 2011 and 2001, respectively. Semiannual means of BP are given in Table 25.1. On a semiannual basis, BP ranged from 12 to 38 % (mean 25 %) and from 18 to 57 % (mean 38 %) of gross primary production (GPP) and net primary production (NPP), respectively. Until 2008, mean semiannual BP was always greater in January–June as compared to July–December, a pattern that also corresponded with PP levels. However, from 2009 through 2011, this pattern reversed, as has also been reported for primary production (see Chap. 24). Furthermore, as noted in Chap. 15, in the years from 2001 to 2011, there was a significant trend (r ² = 0.93, p <  0.0001, n = 11) to lower annual average BA that coincided with a significant decrease (r ² = 0.57, p = 0.004, n = 11) in BP; the annual means of BA and BP were also significantly correlated (r ² = 0.43, p = 0.028, n = 11). The marked drop in BP has had a significant impact on OC flows within the lake system (see below).

Table 25.1 Semiannual means of gross primary productivity (GPP, calculated as 1.5 *14C-measured primary productivity (see Chap. 24)), phytoplankton respiration (PR), community respiration (CR), bacterial productivity (BP), bacterial respiration (BR), and zooplankton respiration (ZR) from 2001 to 2011

The declines in the annual averages of both BA and BP in the epilimnic water do not appear to be correlated to any change in annual averages of phytoplankton biomass , primary production, or chlorophyll . Nevertheless, the apparently real and significant declining trend in BP and the changes in semiannual patterns of primary production and BP (Table 25.1; Fig. 25.1) appear to reflect a definite slowing down in the rate of OC cycling in the lake over the last decade. This trend may be a further indication of the profound ecosystem shift that has occurred (and is probably still ongoing) since the mid-1990s (Zohary and Ostrovsky 2011).

As previously reported for the period 2001–2007 (Berman et al. 2010), strong correlations were obtained from 2001 to 2010 between semiannual BP and total community respiration (CR; r ² = 0.67, p = < 0.0001, n = 22). A rather weaker, but still significant correlation was observed between semiannual BP and GPP (r ² = 0.22, p = 0.03, n = 22) from 2001 to 2011.

1.3 Bacterial Growth Rates and Turnover Times

The estimation of cell-specific C incorporation rates (expressed as fg C cell⁻¹ day⁻¹ and calculated as BP/BA from samples taken monthly at five depths in Sta. A) from 2001 through 2011 falls within the expected ranges. Overall mean cell-specific carbon uptake for epilimnic depths (0–15 m) was 18.0 ± 1.2 fg C cell⁻¹ day⁻¹. Assuming a mean bacterial cell content of 30 fg C bacterial cell⁻¹ (Stone et al. 1993; Hart et al. 2000), this would give a mean carbon turnover (doubling) time of 3.1 ± 0.4 days. Note that despite the large decreasing trends observed in both the annual mean BA and BP, the ratio BP:BA (i.e., cell-specific carbon uptake) showed no significant trend and remained more or less constant throughout the period 2001–2011.

For hypolimnic depths (samples taken close to 40 m), mean cell-specific carbon uptake was distinctly slower with much higher variability (13.4 ± 5.8 fg C cell⁻¹ day⁻¹), and with longer turnover times (7.3 ± 10.9 days). These values for samples from the anaerobic hypolimnion should be viewed with caution since the actual BP measurements were not made under strictly anaerobic conditions. Nevertheless, because metabolic efficiency would be expected to be lower in anaerobic than aerobic ­environments (Fenchel and Finlay 1995), it is reasonable that bacteria in the hypolimnion would have considerably slower C uptake rates and turnover times than epilimnic bacteria.

The above values for cell-specific carbon uptake and turnover times were based on the assumption that all DAPI-counted bacterial cells (Chap. 15) were ­metabolically active. Berman et al. (2001) used three different staining techniques, each of which detected different aspects of cellular state or metabolic activity to ­assess the proportion of active bacteria within the total counted population. Their data indicated that in Lake Kinneret, usually only a small fraction (probably < 20 %) of the entire bacterial assemblage was metabolically active at any given time. In this case, the active bacterial population would have cell-specific carbon uptake rates of ~ 90 fg C cell⁻¹ day⁻¹ and turnover times of about 14–15 h. Unfortunately, no studies have been made in Lake Kinneret of the impact on bacterial cell turnover of lysis by bacteriophage (Suttle 2007).

Berman et al. (2001) also observed that even apparently inactive bacteria could become highly active under appropriate conditions (Choi et al. 1999). This “switching on” might occur as a result of nutrient inputs caused by upwelling events, or on a more localized scale, because of inputs from protistan or zooplankton excretion or sloppy feeding (Jumars et al. 1989) or from algal cell lysis (Berman and Wynne 2005). Thus, the bacterial populations in the lake should be regarded as a dynamic mixture with various components functioning at different degrees of activity at any given moment, constantly modulating their physiological activities to changes in their localized environment.

1.4 Bacterial Respiration, a Major Fraction of Total Community Respiration

Bacterial respiration (BR) was estimated using the method suggested by Berman et al. (2004, 2010) as follows:

$$\text{BR }=\text{ CR }\text{ PR}\text{ ZR}$$

(1)

where CR is community respiration , PR is phytoplankton respiration, and ZR is zooplankton respiration. CR was directly measured by the ΔO₂ method (Sect. 25.2). PR was calculated as 0.3* GPP derived as 1.5* primary production (PP) determined by routine ¹⁴C measurements of PP (Sect. 24.1). Based on previous studies (Berman and Pollingher 1974; Berman et al. 1995), these PP values were assumed to be close to NPP.

Some explanation of the ZR term is required. Historically, crustaceans ­(cladocera and copepods) together with the rotifers were included in the category of “zooplankton” (Gophen 1978). Hambright et al. (2007) used the term “microzooplankton” to describe the assemblage of rotifers, numerous ciliated and flagellated protists, and nupliar stages of copepods. In their previous work, Berman et al. (2010) used ZR values based on routine monitoring measurements of the crustacean and rotifer biomasses multiplied by experimentally determined group-specific respiration rates (Gophen 1981) to obtain BR (Eq. 1 above); however, similar ZR values were calculated by assuming that ZR ranged from 0.33 to 0.5 BR (for details, see Berman et al. 2010). There are no data available on ciliate or flagellate respiration in the lake and no explicit mention of protist respiration has been made in previous studies. In the present study, ZR was taken to include both crustacean and microzooplankton respiration and assigned a value of 0.5 BR; therefore

$$\text{BR }=\text{ }\left( \text{CR-PR} \right)/1.5$$

(2)

This implies that the directly measured values for crustacean ZR in Berman et al. (2010) were overestimated and that the ZR values used in their study included respiration of the protists as well as crustaceans and rotifers. (Note: Fish respiration, based on biomass, has been estimated at between 3 and 5 % of CR (Stone et al. 1993), which is below the sensitivity of this method; therefore, this source of respiration has been omitted.)

The semiannual means of BR (Eq. 2) from 2001 to 2011 are given in Table 25.1. There was a strong, positive correlation between BR and BP (r ² = 0.41, p = 0.0018, n = 21), although the decrease in BR during this period was only weakly significant (r ² = 0.20, p = 0.0412, n = 21). Generally, with the exception of 2009 and 2010, BP levels were higher in January–June than in July–December as might be expected because of the higher levels of phytoplankton biomass and GPP at this season (see Sect. 24.1).

The relative proportions contributed annually by PR, BR, and ZR to total CR are shown in Fig. 25.2. The 10-year averages of PR, BR, and ZR as percentages of CR were 36, 43, and 21 %, respectively. With the exception of 2008 and 2009, BR was always the major component of CR. As previously observed (Hart et al. 2000; Berman et al. 2004, 2010), ZR was consistently the smallest contributor to CR ranging from 14 to 24 % on an annual basis and showing no clear trend during these years.

Fig. 25.2

Relative contribution to total community respiration (CR) by phytoplankton (PR), bacteria (BR), and zooplankton (ZR), based on annual means from 2001 to 2011 (Table 25.2), Sta. A. (Note: 2011 values are based on January–June data only)

1.5 Bacterial Carbon Demand and Bacterial Growth Efficiencies

Previous studies (Berman et al. 2004, 2010) had shown high levels of bacterial carbon demand (BCD; defined as BP + BR) in Lake Kinneret, consistent with the generally high levels of BP and BR. As shown in Table 25.1 and Fig. 25.3, a large proportion of GPP was cycled through the epilimnic bacterial community . These are high, but by no means impossible levels for BCD:GPP (Table 25.2). Even in the absence of external inputs, the total bacterial carbon uptake by heterotrophic bacteria and by secondary and tertiary consumers can sometimes exceed GPP because of recycling within the system (Strayer 1988), as was shown to be the case for Lake Michigan (Scavia 1988). For the years 2001–2007, Berman et al. (2010) reported annual mean BCD:GPP % of 75 ± 17 % for the entire period 2001–2010; however, BCD:GPP % decreased to 65 ± 13 % as a result of the sharp drop in BCD since 2007. There was an overall slight but significant decrease (r ² = 0.46, p = 0.029, n = 10) in BCD from 2001 to 2010.

Fig. 25.3

Annual means of bacterial carbon demand: gross primary production (BCD:GPP) and bacterial growth efficiency (BGE) from 2001 to 2011, Sta. A. (Note: 2011 values are based on January–June data only)

BCD levels tended to be somewhat higher during the first half of the year (average 66 ± 14 %) than in the second half (60 ± 10 %) with the notable exception of 2009 and 2010. In these years, BCD fluxes were only about half of those measured in peak years, although GPP levels did not show a similar drop. BCD reached a maximum of 84 % during the P. gatunense bloom of 2007 (Table 25.2). It is reasonable to assume that this seasonal pattern reflects the increased cycling by heterotrophic bacteria of dissolved organic carbon (DOC) derived directly or indirectly from the larger biomasses of phytoplankton usually present in winter–spring than in summer–fall. Sherr et al. (1982) demonstrated that bacterivorous flagellates increased the decomposition rate of the dinoflagellate carbohydrate thecae by increasing the rate of recycling of nutrients, which stimulated bacterial activity. Thus, the heterotrophic flagellates could be a driver for the increased bacterial (i.e., microbial loop) carbon demand during dinoflagellate blooms in the lake.

Over the period 2001–2010, there was a significant trend of decreasing BCD:GPP measured in winter–spring but not in summer–fall. The drop in both the absolute amounts of BCD and the ratio of BCD:GPP in the winter–spring season may have resulted from the trend to fewer Peridinium years. In any event, the decrease in the flux of OC via the microbial loop may be connected with lower levels of BP and BA (see above and Chap. 15).

Table 25.2 Semiannual means of bacterial carbon demand (BCD), bacterial growth efficiency (BGE), bacterial production: gross primary production (BP:GPP) and BCD:GPP, 2001–2011 at Sta. A

In contrast to BCD, annual mean bacterial growth efficiencies (BGE; defined as BP/(BP + BR)) showed no consistent, significant change over the period 2001–2011 (Fig. 25.3). The overall, average annual BGE (41 ± 6 %) was very similar to that reported for the years 2001–2007 (Berman et al. 2010). In general, estimates for BGE levels in this lake tend to be higher than in most reported freshwater systems. However, it is presently unclear whether these high BGE estimates are accurate or reflect some inherent difference between various methods used for estimating BGE. In the present study, we have assumed that the respiration of protists is included in the ZR term; thus, BR refers to bacterial respiration only. The indirect method of calculating BR (and hence BGE) used here has not been widely used. Most BR measurements reported in the literature to date have utilized experimental data from 1-μm filtered water to directly estimate BR; this approach also has some limitations (see Berman et al. 2010). Future research should compare results obtained with both the above approaches to estimate BR and BGE.

1.6 What Caused the Significant Drop in BP, BA, and BCD From 2001 Through 2010?

As documented elsewhere (Chaps. 10 and 11), the most visible evidence of ecosystem change in Lake Kinneret has been the disruption of the regular seasonal pattern of phytoplankton development (as typified by late winter–early spring blooms of P. gatunense with dominance of small chlorophytes in summer–fall) to highly erratic dinoflagellate blooms and the increasing dominance of N₂-fixing cyanobacteria especially in summer–fall. Here, we have shown that dramatic changes have also occurred both in the abundance (Chap. 15) and in the metabolic activity of the bacterial populations from 2001 to 2011.

Not much is known about limiting factors for the bacterial populations in this lake. Zohary et al. (2000) found that P addition enhanced bacterial degradation of dinoflagellate thecae in Lake Kinneret water. Pinhassi and Berman (2003), working with samples taken from near-surface lake waters, found that P appeared to be ­limiting BP and growth at most times. Additionally, similarly to previous research (Berman et al. 1993), they found that the availability of Fe or chelators could play an important role in regulating bacterial metabolism and growth even in Lake Kinneret where ambient concentrations of total Fe are relatively high. These observations might suggest two possible causes for the significant continuous decreases in BP, BA, and BCD levels from 2001 to 2011 documented here: (1) From the mid-1990s, major hydrological changes in the catchment area have altered the amounts and characteristics of organic chelators and/or Fe and other trace metal ions flowing into the lake. (2) Prolonged periods of low lake levels and low volume winter inflows may have been responsible for the increasing abundance of N₂-fixing cyanobacteria (Chap. 12). These cyanobacteria would be expected to compete more effectively with the heterotrophic bacteria for relatively scarce available nutrients, especially P and Fe, than the previously dominant chlorophytes and dinoflagellates. In any event, there appears to be little doubt that a profound shift in carbon cycling has taken place over the last decade in Lake Kinneret.

2 Community Respiration

Arkadi Parparov and Tom Berman

2.1 Background

Dark pelagic respiration (hereafter, CR) is a major flux in biogeochemical cycles, being the largest sink for organic matter and oxygen in freshwater lakes (del Giorgio and Williams 2005b; Pace and Prairie 2005). Traditionally, CR has been measured as oxygen depletion (ΔO₂) in dark bottles (“dark respiration”). A basic assumption of the oxygen method is that CR rates in the light and dark are similar. Although this assumption has been controversial (Pringault et al. 2007, 2009), in most ecological situations, differences between light and dark CR appear to be insignificant (Winberg 1960; Aristegui et al. 1996; Gazeau et al. 2007).

The first systematic determinations of CR in Lake Kinneret were made by Hepher and Langer (1969), who used the ΔO₂ method. Subsequently, lake CR has been estimated from the budget of OC (Serruya et al. 1980), from daily changes in the pool of dissolved oxygen (Nishri et al. 1998) and using ¹⁸O₂ (Luz et al. 2002). From 2001, CR was monitored biweekly at the central pelagic Sta. A (see location in Chap. 1, Fig. 1.1), using the ΔO₂ method. This involved in situ incubations of water from five depths (0, 1, 5, 10, and 15 m) in dark bottles over 24 h. A factor of 0.3 was used to convert mg O₂ L⁻¹ to mg C L⁻¹ (i.e., Respiration Quotient, RQ = 0.8) . In natural aquatic environments, an RQ of 0.8 may be more accurate than the often used RQ = 1 for CR because the respiration of organic matter includes proteins and lipids in addition to carbohydrates (Winberg 1960; Geider 1997; Robinson 2008).

2.1.1 Community Respiration in Lake Kinneret (1964–2010)

Estimates of CR in the lake, based on measurements made by a variety of methods, are shown in Table 25.3. The annual average value of CR for the period from 2001 to 2010, when the most detailed monitoring of this parameter was carried out, is surprisingly close to early estimates obtained in 1964–1965 (Hepher and Langer 1969) and in 1972 (Serruya et al. 1980). Although it is tempting to suggest that this similarity of CRs attests to a relative stability in the rates of aerobic decomposition of organic matter over the decades, we note that the early data are based on relatively few actual measurements. Much lower CR values (3.18 g O₂ m⁻² day⁻¹ or 0.95 g C m⁻² day⁻¹) were calculated by Nishri et al. (1998) for the period 1970–1990, based on oxygen mass balance calculations. It seems probable that these CR values were considerably underestimated because they are much lower than values for average daily primary production (1.54 g C m⁻² day⁻¹) measured by the ¹⁴C method for the same period (Berman et al. 1995; Yacobi 2006). Furthermore, it seems unlikely that lower CR would occur during a period of overall higher standing stock algal biomass. In contrast, extremely high CR values were reported by Luz et al. (2002) for the years 1996–1998 using the ¹⁸O₂ method (Table 25.3).

Table 25.3 Community respiration (CR) in Lake Kinneret measured at different periods

2.1.2 Community Respiration from 2001 to 2010

Berman et al. (2010) reported that from 2001 to 2007, annual CR averaged 7.28 g O₂ m⁻² day⁻¹. Community respiration was consistently higher in winter–spring than in summer–fall by an average factor of 1.48 ( ± 0.37). It was also correlated closely with BP as well as with parameters of phytoplankton biomass (Chl) and photosynthetic activity.

The average annual values for CR for the period 2001 to 2010 (Fig. 25.4a) averaged 6.86 ± 1.91 g O₂ m⁻² day⁻¹ and ranged from 4.93 ± 2.53 g O₂ m⁻² day⁻¹ in 2009 to 9.91 ± 6.18 g O₂ m⁻² day⁻¹ in 2007. Daily levels of CR varied very widely from 0.61 to 31.7 g O₂ m⁻² day⁻¹. The highest CR values (> 15 g O₂ m⁻² day⁻¹) were recorded in spring during intense dinoflagellate blooms (“Peridinium years,” when maximum phytoplankton biomass exceeded 100 g m⁻²; Zohary et al. 2012) that occurred in 2003, 2004, and 2007; in years without such blooms (“non-Peridinium years”) , the spring maxima of CR were smaller or even absent. The discrepancy between these two annual patterns of CR is most clearly shown by comparing the monthly CR recorded in 2004 and 2009 (Fig. 25.4b). Community respiration was much higher during the first half of 2004 than in 2009; subsequent monthly differences between these years were minimal. It seems reasonable to assume that the greater input of OC that occurred as a result of the Peridinium blooms provided the substrate for enhanced respiration rates recorded during these years. During dense dinoflagellate blooms, high CR values were measured even when the upper water layer (0–15 m) was supersaturated (up to 130 %) with oxygen. On these occasions, GPP should be higher than the corresponding measured CR values which ranged from 4.50 to 31.7 g O₂ m⁻² day⁻¹ (1.35–9.51 g C m⁻² day⁻¹) assuming no significant allochthonous inputs of OC . Such high values for GPP greatly exceed primary production rates measured with the ¹⁴C method during Peridinium blooms. In general, the ¹⁴C method as used in the long-term studies in Lake Kinneret gives values close to NPP (Berman and Pollingher 1974); during dense Peridinium blooms, the ratio of GPP:NPP was close to 2 (see Chap.  24.1).

Fig. 25.4

Lake Kinneret community respiration (CR, g O₂ m⁻² d⁻¹) recorded at Sta. A, 2001–2010. a Annual averages and overall average (Avg). b Monthly CR for a “Peridinium year” (2004) and “non-Peridinium year” (2009). The annual mean contributions of dinoflagellates to total phytoplankton biomass were 65.5 % in 2004 and 38.4 % in 2009. c Multiannually averaged monthly CR for the period 2001–2010. Whiskers in a and c indicate standard deviation (SD)

The mean monthly CR (averaged for the period 2001–2010) is shown in Fig. 25.4c. Despite the large variability shown in these measurements , there was nevertheless a clear trend of higher CR in the late winter–spring months (April, May). Overall, CR was consistently higher in winter–spring than in summer–fall by a factor of 1.44, consistent with the generally higher levels of phytoplankton biomass measured during the former season.

2.2 Contributions of Major Biological Groups to CR

Phytoplankton, heterotrophic bacteria , and zooplankton (including protista) all contribute to the total measured CR in the epilimnic waters of Lake Kinneret. Some attempts have been made to measure directly respiration rates for heterotrophic bacteria (Berman et al. 1979), phytoplankton (Berman and Gerber 1980; Berman and Kaplan 1984a, 1984b), and zooplankton (Gophen 1981). The first estimates for the specific contribution of the major planktonic biological groups to overall lake CR were derived from carbon mass balances measured from 1989 to 1992 (Hart et al. 2000). These indicated (Fig. 25.5) that heterotrophic bacterial and PF contributed about equally (~ 40 % each) to total CR, with only ~ 20 % contributed by ­zooplankton (cladocera, copepoda, rotifers) and protozoa (ciliates and flagellates). The contribution of fish respiration was at most only a few percent of total CR.

Fig. 25.5

The relative contributions of three groups of planktonic organisms to total community respiration (CR). DH (D. Hart): estimates based on Hart et al. (2000) for the years 1989–1992; TB (T. Berman): estimates based on Berman et al. (2010) for the years 2001–2010. Whiskers indicate standard deviation (SD)

The data on dry weight- specific respiration (mg O₂ mgTSS⁻¹ day⁻¹) associated with seston dominated by different algal groups are presented in Table 26.3 of Chap. 26.

Berman et al. (2004, 2010) introduced a method of estimating the relative amounts of respiration contributed annually by PR, BR, and ZR to total CR. Using this approach, the calculated, overall, 10-year (2001–2011) averages of PR, BR, and ZR of CR were 36, 43, and 21 % of CR, respectively (details in Fig. 25.2, Sect. 25.1). With the exception of 2008 and 2009, BR was always the major component of CR, slightly higher than PR. As previously estimated (Hart et al. 2000; Berman et al. 2004, 2010), ZR was consistently the smallest contributor to CR, ranging from 14 to 24 % on an annual basis and showing no clear trend during these years. Note, the contribution of fish respiration to total CR was excluded here because it was assumed to be too low to be within the resolution of this method.

3 Sulfate Reduction

Ora Hadas and Riki Pinkas

3.1 Introduction

Anaerobic mineralization of organic matter is mediated by a series of bacterial metabolic pathways such as nitrate reduction and manganese and iron reduction with sulfate reduction and methanogenesis being the terminal processes in the anaerobic decomposition of organic matter (Oremland and Polcin 1982; Ingvorsen and Jorgensen 1984; Capone and Kiene 1988). Sulfate-reducing bacteria (SRB) play a key role in the mineralization processes in marine sediments, but significant sulfate reduction rates have also been reported for freshwater sediments, despite low sulfate concentrations in the pore water (King and Klug 1982; Landers and Mitchell 1988). This finding suggests that SRB in freshwater have acquired high-affinity sulfate uptake systems to cope with the low sulfate concentrations (Smith and Klug 1981; Ingvorsen and Jorgensen 1984). High substrate affinities of SRBs for acetate, H₂, and other electron donors can account for the inhibition of methanogenesis in sulfate-rich environments (Lovley and Klug 1986; Skyring 1988), although syntrophism in methanogenic environments is not excluded (Plugge et al. 2011). Although until recently sulfate reduction was considered a strict anaerobic process (Postgate 1984), we know today that SRBs may thrive under oxidizing conditions in which they respire with nitrate or even oxygen (Canfield and Des Marais 1991; Frund and Cohen 1992; Cypionka 2000).

Sulfate reduction is one of the major driving forces in the maintenance of deep water anoxia in Lake Kinneret. As in other sites, sulfate reduction forms the basis of the biological sulfur cycle in this lake. The factors controlling the rate of sulfate reduction are: (1) availability of organic matter from decomposing material, mostly planktonic carbon in the water column; (2) ambient SO₄ ²⁻ concentration; (3) redox potential; (4) temperature; and (5) pH (Westrich and Brenner 1988).

The conditions in Lake Kinneret hypolimnion and sediments regarding those controlling factors were followed closely at Sta. A (station location map in Fig. 1.1 of Chap. 1) during 1988–1989. Two other stations Sta. G and the shallow Sta. S were studied during different seasons in 1991 and 1992. The findings are summarized below.

3.2 Sources of Organic Matter

The annual average of total phytoplankton biomass was similar in 1988 and 1989 (73 g wet wt m⁻² in 1988 and 74 g wet wt m⁻² in 1989). In 1988, the highest biomass (225 g wet wt m⁻²) was observed in February due to a bloom of the filamentous diatom Aulacoseira granulata (182 g wet wt m⁻²) whereas the Peridinium peak appeared in May (124.7 g wet wt m⁻²). In 1989, the Peridinium peak was in March–April (166, 184 g wet wt m⁻²), contributing more than 90 % of the total phytoplankton biomass; no significant biomass of Aulacoseira was observed in 1989 (Table 25.4) .

Table 25.4 Monthly mean total, Peridinium gatunense and Aulacoseira granulata biomass (g w.w. m⁻²) at the trophogenic layer at Sta. A, and sulfate reduction rates (mmol m⁻² month⁻¹)

The heavy filaments of A. granulata appear in the water column in times of high turbulence, and at the end of the turbulent period, they sink toward the sediments (mixed water column), supplying fresh organic material to the sediments. These blooms supplied the organic matter for sulfate reduction by SRBs in the hypolimnion and sediments, where anoxia existed, sulfate was available, and temperature and pH were also suitable, as discussed below.

3.3 Sulfate and Sulfide Concentrations

Hypolimnion: Strong seasonal variations in sulfate and sulfide concentrations in the hypolimnion of Lake Kinneret were found. Levels of sulfate ranged from 531 to 547 µM during holomixis and dropped to minima of 198–219 µM in December, at the end of the stratification period. An inverse pattern was observed for sulfide concentrations, 0 during the mixing period and reaching maxima of 309 and 366 in 1988 and 1989, respectively, just before overturn (Fig. 25.6). The decrease in sulfate concentrations corresponded stoichiometrically to the increase in that of sulfide, suggesting that intensive sulfate reduction was occurring in the hypolimnion of the lake, with sulfide accumulating as a result of sulfate reduction. The higher sulfide concentrations in 1989 compared with 1988 were probably due to a smaller hypolimnion volume, due to lower water levels (1,555 and 1,296 million m³ in 1988 and 1989, respectively).

Fig. 25.6

Monthly average concentrations (µM) of sulfate at 10-m depth (empty squares) and sulfate (full squares) and sulfide (circles) at 40-m depth at Sta. A in Lake Kinneret, 1988–1989. (Reproduced from Hadas and Pinkas 1995a with permission from Wiley)

Sediments: The highest sediment pore water sulfate concentrations were found during the mixing period in February, 438 µM at 0.3 cm depth in 1988 and 781 µM at 0.6 cm depth in 1989 due to turbulence, oxygen diffusion into the first few millimeters and bioturbation by transitory dwellers (e.g., Leydigia sp., Cladocera), and some copepods and chironomids. Minimum values were detected toward the end of the stratification period, 8–20 µM, of sulfate in November–December 1988 (Fig. 25.7). In some years, no sulfate is detected in December below 1.5 cm.

Fig. 25.7

Depth profiles of H₂S and SO₄ concentrations and pH and sulfate reduction rates in different seasons in Lake Kinneret sediments and overlying water. (Reproduced from Hadas and Pinkas 1995a with permission from Wiley)

3.4 Arylsulfatase

When sulfate concentrations are low, the enzymatic activity of arylsulfatases may supply part of the sulfate demands of SRB in the sediments. In Lake Kinneret, the activity of arylsulfatase varied with depth and season (Hadas and Pinkas 1992, 1997). Maximum activity was found in the sediments in July (670 nmol p-nitrophenol, PNP, g w.w.⁻¹ h⁻¹) and December, and minimum in February. The lower activity in February was correlated with high SO₄ ⁻² available. During the stratified period, intensive sulfate reduction resulted in decrease in SO₄ ⁻² concentrations and increase in arylsulfatase activity. At the decline of the Peridinium bloom, supplying organic matter and sulfate esters, high arylsulfatase activity was induced (Hadas and Pinkas 1992, 1997).

3.5 Redox, Temperature, and pH

The pH in the sediment cores decreased with depth and ranged between 7.45 and 7.05 with no seasonal variation (Fig. 25.7). The temperature of the hypolimnion at Sta. A is in the range of 14–16 °C and does not vary during the year, with interannual variations averaging 2 °C. There is a drop in redox potential in the sediment due to availability of organic matter, high microbial activity, and depletion of oxygen at the beginning of summer. These conditions enabled intense sulfate reduction activity (Figs. 25.7 and 25.8).

Fig. 25.8

Time–depth distribution of rates of sulfate reduction in Lake Kinneret sediments in 1988–1989. (Reproduced from Hadas and Pinkas 1995a with permission from Wiley)

3.6 Sulfate Reduction Rates

Sediment cores were taken with a Tessenow gravity sampler (Tessenow et al. 1977) at Sta. A (Central, 42 m depth) during 1988–1989 and at Sta. G (Northern, 22 m) influenced by the Jordan River, and Sta. S (shallow, 10 m) representing sediments covered with oxic water the whole year. Sampling was carried out at different seasons of the year, corresponding to the thermal stratification and Peridinium bloom: (1) in February during full mixing at the beginning of the Peridinium bloom, (2) in July after the crash of the bloom, and (3) in December during stable stratification. Sulfate reduction rates were measured by the method of Jorgensen (1978) using Na₂ ³⁵SO₄ as described previously (Hadas and Pinkas 1992, 1995a, 1995b).

3.6.1 Hypolimnion

According to the stoichiometric equation:

$$2{{\left( \text{C}{{\text{H}}_{\text{2}}}\text{O} \right)}_{n}}+\text{ S}{{\text{O}}_{4}}^{-2}\to \text{ }2\text{HC}{{\text{O}}_{3}}^{-}+\text{ }{{\text{H}}_{\text{2}}}\text{S}$$

(3)

2 mol of OC are oxidized per mole of sulfate reduced. The accumulation of H₂S in the hypolimnion was accompanied by a decrease in SO₄ ⁻² concentrations of 0.31 and 0.35 mM in 1988 and 1989, respectively (Fig. 25.6). The average hypolimnion volume of the lake was 1,555 and 1,296 million m³ in 1988 and 1989, respectively. Accordingly, 964 × 10⁶ and 907 × 10⁶ mol C could have been oxidized in 1988 and 1989 via sulfate reduction . The total accumulated yearly phytoplankton biomass was about 150,000 t wet weight. Since most of it was Peridinium with 20 % C content, about 36 and 39 % of the phytoplankton biomass could have been oxidized via sulfate reduction in the hypolimnion in 1989 and 1988, respectively. This is in agreement with the reports that 50 % of the primary production may be decomposed by SRB in the hypolimnion (Smith and Oremland 1987; Canfield 1989).

3.6.2 Sediments

The rate of sulfate reduction in Lake Kinneret sediments varied seasonally (Figs. 25.7, 25.3) and was dependent on both the availability of sulfate in the pore water and organic matter supplied from the epilimnion and reaching the sediments. The relative fraction of organic matter reaching the sediments depends on the timing of the algal bloom and the shape of the isotherms. An early bloom of Peridinium (or other species) and a long mixing period would usually increase the amount of organic matter reaching the sediments (Serruya et al. 1974). At the beginning of summer, the crash of the Peridinium bloom resulted in an increased input of fresh organic matter.

During thermal stratification, the phytoplankton is mostly decomposed in the epilimnion and the supply of organic matter reaching the sediments is low resulting in lower sulfate reduction rates in autumn (Fig. 25.3). During the mixing period, although sulfate concentrations are high, the limited supply of organic matter resulted in low sulfate reduction rates. An exception was February 1988, in which a second peak of sulfate reduction occurred due to the A. granulata bloom sinking toward the sediments (no thermal barrier) and providing fresh organic matter (Fig. 25.3).

High sulfate reduction rates were observed in July 1988 throughout the whole 2.4-cm depth sediment core, reaching a maximum of 1,699 nmol SO₄ ⁻² reduced cm⁻³ day⁻¹ at 0.3-cm depth (Figs. 25.7 and 25.8). These high rates were also observed in June and August at the upper layers of the sediments. In November and December, due to the strong thermal stratification, depletion of organic matter, and of sulfate, lower rates were measured (14 and 80 nmol SO₄ ⁻² reduced cm⁻³ day⁻¹). In 1989, the sulfate reduction peak occurred in July at 0.3-cm depth and, at least at this depth, was as high as in 1988 (Figs. 25.7 and 25.8). Lower rates were observed in December, just before the overturn (12 nmol SO₄ ⁻² reduced cm⁻³ day⁻¹ at 2.4-cm depth; Hadas and Pinkas 1995a).

The integrated values of sulfate reduction in the sediments in 1989 were only 63 % of those in 1988. The difference in the phytoplankton composition and the higher productivity during the first half of the year (January–June), 2,300 and 2,000 mg C m⁻² day⁻¹ in 1988 and 1989, respectively, could probably account for part of the difference in sulfate reduction rates between those years (Table 25.5). Furthermore, the mixing period of the lake was a month longer in 1988 than in 1989, resulting in higher amounts of organic matter reaching the sediments. Only 10 % of the carbon fixed by photosynthesis reaches the sediments (Serruya 1978). Based on this value, the phytoplankton primary production reaching the bottom could account for most or all the potential sulfate reduction in the sediments at Sta. A (Table 25.5). δ¹³C of the organic matter in the sediments have shown values lower by 4 ‰ as compared to the organic matter sinking from lake water indicating diagenetic processes in the sediments (Stiller and Magaritz 1974).

Table 25.5 Semiannual and annual monthly average carbon oxidized through sulfate reduction (mmol C m⁻² month⁻¹) and carbon supplied by primary production (mmol C m⁻² month⁻¹) in the sediments at Sta. A

3.7 Sulfate Reduction Rates at Stations G and S as Compared to Station A (Based on the Years 1990–1992)

Station G, located in the northern part of the lake, is mostly influenced by the inflow of the Jordan River . The Jordan and its tributaries bring relatively high amounts of particulate organic matter . The sediments at station G are rich in detrital elements (Fe, P, Mn) as the result of deposition of suspended matter associated with the early floods. High sulfate reduction rates were observed immediately after the first heavy rainfalls in December 1991. Another peak (1,350 nmol SO₄ ⁻² reduced cm⁻³ day⁻¹) was recorded in July at the crash of the Peridinium bloom when organic matter and sulfate were available. During drought years when the contribution of the Jordan River was small, low sulfate reduction rates were recorded (Hadas and Pinkas 1995b).

Station S (10 m depth, in the littoral at the western side of the lake) is exposed to an oxygenated water column during the whole year; however, the sediments at Station S are anoxic and oxygen penetrates to less than 0.6 mm. At low lake levels, if less organic matter was available, low sulfate reduction rates were recorded (6–48 nmol SO₄ ⁻² reduced cm⁻³ day⁻¹ in February, 1991). During periods of low water levels (e.g., 1990–1991) shoreline vegetation (Tamarix jordanis and Phragmites australis) developed on the exposed shores. This vegetation decomposed after being covered by the rising water (like in 1992), supplying fresh organic material that resulted in high sulfate reduction rates.

3.8 Prospective

The main trigger for the sulfate reduction process in Lake Kinneret was the decline and decomposition of the Peridinium bloom. In years of low water levels, the bloom started early (February), peaked in April, and declined in May–June supplying fresh organic material to the hypolimnion and sediments. At high water levels (usually correlated with cold and rainy winters), the bloom of A. granulata added to the organic fraction supplied to the sediments. During the past 2 decades, the characteristic annual bloom of Peridinium does not appear every year, and in some years other species (e.g., Mougeotia) are the main phytoplankton contributing to the primary production. Sulfate reduction rates calculated from sulfate concentration profiles using a numerical model and based on profiles in 2008 (“non-Peridinium year,” Adler et al. 2011) showed values in the range of the rates measured by Hadas and Pinkas (1995a) without the peaks measured in July after the degradation of the Peridinium bloom. The changes in the phytoplankton composition pattern of the lake and invasion of new species with different seasonal distribution patterns and bloom timing may have an impact on sulfate reduction in the hypolimnion and sediments and on the mineralization processes of organic matter in the lake.

4 Methanogenesis

Werner Eckert and Orit Sivan

Methanogenesis is a key terminal process in the anaerobic decomposition of organic matter that dominates in the absence of other electron acceptors such as nitrate, ferric iron, and sulfate. It is meditated by methanogenic bacteria (MB) , a diverse group of microorganisms, all belonging to the Archaea. They are responsible for ~ 70 % of the globally produced methane estimated at 500–600 Tg CH₄ per year (Conrad 2009). This biologically produced methane derives mainly from the reduction of CO₂ with hydrogen (hydrogenetic methanogenesis), or from the fermentation of acetate (acetoclastic methanogenesis), which is split into CH₄ and CO₂ and, to a lesser extent, from other C −1 compounds (Whitman et al. 1991). In aquatic environments, methanogenesis is primarily a sedimentary process controlled by the prevailing sulfate concentrations due to the competition between MB and SRB for common substrates. Even in freshwater sediments that are relatively poor in sulfate, SRB can outcompete MB (Lovley and Klug 1983) restricting the process of methanogenesis to the deeper sediment layers that are depleted of SO₄ ²⁻. Nevertheless, methanogenesis was repeatedly confirmed as the dominant process in the final mineralization of organic matter (Kuivila et al. 1989; Sinke et al. 1992). On a global scale, the contribution of aquatic ecosystems to the atmospheric methane budget is relatively small (3 %)—a fact that seems to disagree with the well-established importance of methanogenesis in carbon cycling (e.g., Rudd and Taylor 1980). This apparent contradiction is the result of the activity of methane-oxidizing prokaryotes, or methanothrophs, that in aquatic environments effectively attenuate the sedimentary CH₄ flux by oxidizing upward diffusing methane in the presence of suitable electron acceptors (Conrad 2009). For a long time, methanotrophy was regarded as a strictly aerobic process performed by gamma- or alpha-proteobacteria (e.g., Trotsenko and Murrell 2008). However, increasing evidence suggests the existence of a group of Archaea capable of anaerobic oxidation of methane (AOM) . AOM was estimated to consume the methane equivalent to 20 ± 5 % of the present atmospheric methane flux (Valentine and Reeburgh 2000).

In Lake Kinneret, both methanogenesis and methanotrophy were unexplored processes until 1998. At that time, a scientific program was launched aimed at learning about methane evolution in the water column, its role in the lake’s carbon cycle, and physiological aspects of methanogenesis . The study of hypolimnetic methane accumulation together with that of sulfide during several annual lake cycles showed that the biogeochemical processes in the lake are tightly linked to the physical forcing of the water column (Eckert et al. 2002). During the stratified period, strong daily westerly winds lead to the formation of a benthic boundary layer (BBL) that is characterized by enhanced microbial activity. As such, oxygen and nitrate are depleted first in this ~ 10 m thick BBL delineated by a sharp drop of the oxygen-reduction potential, as exemplified by the April profile in Fig. 25.9. Only when the upper hypolimnion, with its significantly lower microbial activity (e.g., rate of oxygen consumption was shown to be ca. four times lower in this layer; Eckert et al. 2002), has turned anoxic as well, does the chemocline rise to the depth of the thermocline (August profile in Fig. 25.9). In the absence of suitable electron acceptors, sedimentary release causes the continuous increase of CH₄ into the hypolimnion with near-bottom concentrations reaching up to 400 µmol L⁻¹ (Eckert and Conrad 2007). The peaks of both solutes in the upper hypolimnion are likely the result of BBL mixing due to internal wave braking followed by advection across the laminar hypolimnion (Eckert et al. 2002; Chap. 9, physics). Compared to other freshwater systems, this concentration is relatively high. Similar quantities measured in the hypolimnion of some temperate lakes in the northern USA (L. Paul: 250 µmol L⁻¹; L. Peter: 600 µmol L⁻¹, and L. Hummingbird: 300 µmol L⁻¹, Bastviken et al. 2008) could be attributed to low sulfate (< 60 µmol L⁻¹) and high DOC concentrations (> 500 µmol L⁻¹; Houser et al. 2003). But Lake Kinneret generates such concentrations at a tenfold higher sulfate and five times lower DOC. Conditions are similar for Lakes Konstanz and Lugano where CH₄ concentrations in the anaerobic zone remain significantly lower at 120 µmol L⁻¹ (Schmidt and Conrad 1993) and 80 µmol L⁻¹ (Liu et al. 1996), respectively. The only ­environmental ­parameter that may explain the relative high methane concentrations in Lake Kinneret is the elevated temperature in comparison to temperate lakes (Thebrath et al. 1993).

Fig. 25.9

Biological stratification in Lake Kinneret during summer in response to physical forcing. Microbial activity is highest in the benthic boundary layer and the metalimnion , and lowest in the laminar “upper hypolimnion.” The effect of the stratification on the evolution of biochemical processes in the water column is indicated in the right panel by the sulfide (S²⁻), CH₄, oxidation reduction potential (ORP), and temperature (T) profiles from April and August. (Reproduced from Eckert et al. 2002 with permission from Springer)

During the stratified period, the accumulation pattern of methane in the hypolimnion of Lake Kinneret follows closely that of sulfide (Fig. 25.10), with annual average concentrations around 3 and 6 mol m⁻², respectively. The relative impact of both processes on the mineralization of OC was elaborated further by Eckert and Conrad (2007), who established a carbon budget for Lake Kinneret. Sulfate reduction was identified as the dominant process in the decomposition of organic matter that seasonally settles into the hypolimnion (~ 20 mol C m⁻² year⁻¹), accounting for 60 % of the C input. Methane release represented 30–40 % including that of methane gas ebullition . The latter constituted approximately 25 % of the sedimentary methane flux (Eckert and Conrad 2007), a figure that was confirmed by echosounder measurements (Ostrovsky 2003; Ostrovsky et al. 2008; Ostrovsky and Tegowski 2010).

Fig. 25.10

Seasonal patterns of near-bottom dissolved methane and total sulfide concentrations in the hypolimnion of Lake Kinneret during the years 1998–2000. (Reproduced from Eckert and Conrad 2007 with permission from Springer)

Physiological studies targeting methanogenesis in the sediment of Lake Kinneret have indicated extreme substrate limitation, with methane production from acetate being partly due to syntrophic oxidation coupled with hydrogenotrophic methanogenesis rather than direct acetoclastic cleavage (Nüsslein et al. 2001, 2003; Schwarz et al. 2008). Molecular analysis of the microbial community structure in the upper 10 cm of profundal sediments revealed that the numbers of archaea and bacteria, quantified by real-time polymerase chain reaction (PCR), amounted to about 10⁸ and 10¹⁰ 16S rRNA gene copies cm⁻³ sediment, respectively, suggesting that Archaea account for only a minor fraction (approximately 1 %) of the total prokaryotic community (Schwarz et al. 2007a, 2007b). Since methanogenesis is carried out by Archaea only, this observation supports the overall assumption that methanogenesis plays only a minor role in the carbon cycle of the Lake Kinneret.

Seasonal profiles of dissolved methane in the pore water of intact sediment cores from the centrally located Sta. A (see location in Fig. 1.1a, Chap. 1) typically show a prominent increase in sediment depth with a maximum around 10 cm (Fig. 25.11a, redrawn from Adler et al. 2011). Below the peak zone, dissolved methane decreases again to concentrations similar to those measured in the upper sediment layers. Model results (Adler et al. 2011) as well as incubation experiments suggest that this decline is the result of a methane sink in the deeper lake sediments. Apparently, methanogenesis is not substrate limited in this zone, based on the DOC profile that is characterized by increasing concentrations with depth (Fig. 25.11a). Thus, the hypothesis evolved that methane production may be balanced by methane consumption.

Fig. 25.11

Sediment pore water depth profiles of a dissolved methane and dissolved organic C (DOC) and b δ¹³C–CH₄ signature measured in intact sediment cores sampled in August 2009 at Sta. A at the center of Lake Kinneret (location in Fig. 1.1 of Chap. 1). (The DOC profile represents unpublished data of Bar Or and Eckert; CH₄ and δ¹³C profiles are redrawn from Sivan et al. 2011—Copyright 2014 by the Association for the Sciences of Limnology and Oceanography, Inc.)

The assumption of AOM could be further manifested by means of the δ¹³C signatures measured within the zone of deep methane decrease (Fig. 25.11b, modified from Sivan et al. 2011). The δ¹³C-CH₄ profile shows a decrease from − 60 ‰ at 1-cm depth to about − 65 ‰ at 7-cm depth and then an increase to a maximum of − 53.5 ‰ at 24 cm, an increase that can only be explained by methanotrophy , in which the residual methane becomes isotopically heavier. Besides this geochemical evidence for AOM in deep Kinneret sediments, Sivan et al. (2011) suggested that AOM is likely driven by iron (Fe) reduction, a hypothesis that was verified by ­dissolved iron concentration and isotopic profiles measured in the pore water of intact sediment cores and by Fe(III)-amended mesocosm studies.

Analysis of seasonal sediment cores from Sta. A revealed that below the sediment water interface sulfate reduction is the dominant process throughout the year with rates peaking at around 1.5 × 10⁻¹² mol cm⁻³ s⁻¹ and declining exponentially with depth (Fig. 25.12). Methanogenesis starts at ~ 3 cm depth with maximum rates of 4 × 10⁻¹³ mol cm⁻³ s⁻¹ at 5–12 cm depth, followed by decreasing methane concentrations below the maximum methane production zone. Apparently, this decline is caused by AOM taking place at a rate of about 5 × 10⁻¹⁴ mol cm⁻³ s⁻¹ (Adler et al. 2011). To explain the average annual hypolimnetic CH₄ accumulation of 3 mol m⁻² (Fig. 25.10) with the reported rates of methanogenesis requires a 20–25 cm-thick sedimentary zone of methane production—a presumption that matches well our measured profiles (Fig. 25.12).

Fig. 25.12

Geochemical model output of the changes with sediment depth of rates of sulfate reduction (dotted line), methanogenesis (full line), and anaerobic methane oxidation (AOM, dashed line) in Lake Kinneret. (Drawn from data published by Adler et al. 2011 and Sivan et al. 2011)

In summary, the study of methane cycling in Lake Kinneret revealed some unique features in comparison to other freshwater lakes. Besides the relatively high concentrations of dissolved CH₄ in spite of high sulfate and low DOC concentrations, there is the important discovery of sedimentary AOM with ferrous iron, shown for the first time ever in a natural aquatic system.

5 Fluxes of Organic Carbon in the Epilimnion

Tom Berman, Arkadi Parparov, Yosef Z Yacobi, and Ilia Ostrovsky

5.1 Organic Carbon Sources and Sinks in Lake Kinneret

The first delineation of carbon flows in the Lake Kinneret system was by Serruya et al. (1980) who emphasized the different patterns observed during the “Peridinium season” and during the summer–fall. In winter–spring, the major OC input was from primary production during the regular annual dinoflagellate bloom by large Peridinium cells that were mostly not grazed by zooplankton but broken down in the water column by heterotrophic bacteria . By contrast, in summer–fall, primary production was carried out by smaller phytoplankton, mostly chlorophyta , that were readily grazed by the resident cladocerans, which in turn provided sustenance for copepods and fish. Later studies provided more detailed quantitative descriptions of carbon fluxes in the lake ecosystem, based on more extensive data including bacterial and protozoan (ciliates and flagellates) participation in OC cycling (Stone et al. 1993; Hart et al. 2000). Here, we present an overview of OC fluxes in the lake, based on data from 2001 to 2010.

In Table 25.6, we show the major, annual, averaged daily OC fluxes for the years 2001–2010; during this period, Peridinium blooms occurred only in 2003, 2004, and 2007. Table 25.6 shows only the major autochthonous OC input, GPP (calculated from routine measurements of ¹⁴C primary production, see Sect. 24.1), and sinks (CR , measured with ΔO₂, see Sect. 25.2 and sedimentation, SED, determined from direct measurements with sediment traps). These fluxes are also shown separately for the winter–spring and summer–fall seasons in Table 25.7.

Table 25.6 Lake Kinneret: Annual, averaged, daily organic carbon fluxes in the epilimnion (0–15 m) for the years 2001 through 2010

Table 25.7 Lake Kinneret: Semiannual, averaged, daily organic carbon fluxes (January–June and July–December) in the epilimnion (0–15 m) for the years 2001 through 2010

By far, the major input of OC into Lake Kinneret was from GPP by phytoplankton (Sect. 24.1), with GPP estimated as 1.5 × ¹⁴C-measured primary production, except for January–June 2003, 2004, and 2007, when, during Peridinium blooms, GPP was calculated as 2.0 × ¹⁴C-measured primary production. Minor additional inputs were from chemoautotrophic bacteria (Sect. 24.2), watershed inflows, ­direct ­rainfall, benthos, littoral vegetation, and dust . The major loss of OC derived from community respiration (Sect. 25.2), followed by losses from sedimentation (Fig. 25.13). Other losses through outflows via the National Water Carrier , local water consumption (including water to the Kingdom of Jordan), and to the southern Jordan River were very small.

Fig. 25.13

Organic carbon (OC) fluxes for the epilimnic waters of Lake Kinneret based on averaged annual rates from 2001 to 2010. Sources of OC are gross primary production (GPP), chemosynthesis (CH), and inputs (INF) from watershed inflows and dust. Sinks of OC are community respiration (CR), sedimentation (SED), and outflows (OUT) via the National Water Carrier, southern Jordan River, and pumping for local water consumption and to the Kingdom of Jordan. The OC pool in the lake is comprised of particulate and dissolved OC (POC and DOC). The numbers represent annual OC fluxes as tons C lake⁻¹× 1,000, based on lake surface area = 151 km² measured for an average lake depth of 24 m

Note that the data (given as mg C m⁻² day⁻¹) in Tables 25.6 and 25.7 were derived from measurements made in the epilimnion (here, taken as the water layer from 0 to 15 m) and show only the major OC fluxes in this system. Neither OC production nor OC losses have been quantified in the deeper waters of the lake (average lake depth ~ 24 m). In addition, we have not included in these tables ­autochthonous inputs of OC resulting from chemoautotrophic activity. Chemoautotrophic inputs may add somewhere between ~ 5 and 10 % OC on an annual basis, but these have not been consistently monitored and occur only in aerobic/anaerobic interfaces such as in the metalimnion and sediment–water interface (Sect. 24.2). Additional inputs of allochtonous OC into Lake Kinneret brought by river inflows from the watershed (Chap. 18), direct rainfall, benthos, littoral vegetation, and dust constitute a very small percentage (~ 1–2 %) of the GPP and have not been included in Tables 25.6 and 25.7. Other losses that occur but are not shown in these tables include OC removed in fish biomass by commercial fishing and birds, and in the water outflows to the National Water Carrier , to the southern Jordan River , to local water ­consumers, and to the Kingdom of Jordan (Chap. 31). In total, these OC losses are estimated to be not greater than ~ 1–5 % of GPP.

The major sink for OC was always CR; annual losses by SED averaged 16% varying from 12 to 22% of GPP (Table 25.6). Note that CR data in Table 25.6 were based on measurements that did not include respiration by fish, estimated to be &lt;3% of CR (Hart et al. 2000). Overall, the 10-year record of the difference “Inputs-Sinks” indicated that some further inputs and sinks not shown in Table 25.6 could be required to better balance the OC budget; however, these “missing” components were quantitatively minor. Also, inherent problems in the measurements of GPP, CR and SED add a measure of uncertainty to the data.

Clear differences in OC semiannual flux patterns over the period 2001–2010 were observed (Table 25.7). With few exceptions, the relative amounts of GPP, CR, and SED were usually greater in the first half of the year in comparison with the second half of the year, resulting from the higher levels of primary production and nutrient inputs in winter–spring (Table 25.8). This was particularly evident with respect to GPP and CR in 2003, 2004, and 2007 when dense blooms of Peridinium occurred. In winter–spring season of these years, measured rates of CR exceeded those of GPP, presumably due to the high intrinsic respiration of the dinoflagellate bloom and subsequently to elevated rates of bacterial respiration linked to the breakdown of the phytoplankton biomass.

Table 25.8 Ratios of winter– spring to summer–fall averages of gross primary production (GPP), community respiration (CR), sedimentation (SED), and bacterial C demand (BCD) measured in Lake Kinneret, 2001–2010, and the multiannual average and standard deviation (SD), (nd: not defined)

The ratio between GPP (winter–spring) and GPP (summer–fall) in any given year ranged from 0.9 to 1.7 (average 1.2 ± 0.2) whereas the ratio between SED (winter–spring) and SED (summer–fall) was always higher, ranging from 1.3 to 3.9 (average 2.0 ± 0.8). This suggests that the main reason for the lower average percentage of summer–fall SED:GPP (11 ± 2 %) compared to winter–spring SED:GPP (20 ± 6 %) was because more of the autochthonous particulate organic carbon (POC) generated by GPP was lost by respiration or transformation to DOC in the second half of the year. This observation implies that the POC formed in the latter half of the year was relatively more labile (i.e., less POC reached the lake bottom) than the POC formed in winter–spring. Another possible reason for the observed difference in SED: GPP is a more efficient recycling of POC within the epilimnion in summer and fall (see Chap. 27).

The annual averages of CR (Table 25.6) were weakly correlated to GPP (r ² = 0.47; p = 0.028; n = 10); a much stronger correlation (r ² = 0.67; p = 0.004; n = 10) was found between CR and GPP measured for the January–June season only (Table 25.7). No significant relationship was observed for CR and GPP in the latter half of the year. Curiously, no correlations were found between GPP and SED or, less surprisingly, between CR and SED.

5.2 Net Autotrophic and Heterotrophic States of the Lake

In view of the unexpectedly high potential of lakes to act as sources or sinks of greenhouse gasses and to function as significant regulators of climate change on a global scale (Williamson et al. 2009), considerable interest has focused on their overall net autotrophic or heterotrophic status. Given the low OC external inputs, it is not surprising that the euphotic zone of Lake Kinneret is generally net autotrophic. However, when examined on a semiannual basis, measured CR was higher than GPP in the winter–spring (January to June in 2003, 2004, 2005, and 2007), indicative of net heterotrophic conditions. As noted, 2003, 2004, and 2007 were years with large Peridinium blooms with intrinsically high levels of respiration. Additionally, after the rapid die-off of the bloom, the dense phytoplankton biomass would be expected to undergo extensive degradation by heterotrophic bacteria, giving rise to heightened CR. In spring 2005, there was an unusual bloom of the filamentous green alga Mougeotia sp . (Zohary et al. 2012) which may have resulted in a net heterotrophic situation where CR > GPP (Table 25.7). Overall, however, despite relatively short, occasional periods of net heterotrophy , Lake Kinneret is clearly net autotrophic (see below).

5.3 Organic Carbon Flux Through the Microbial Loop

In early studies of OC cycling in Lake Kinneret (Serruya et al. 1980), the role of microorganisms was scarcely considered. However, later modeling studies of carbon fluxes in the lake indicated the central function of bacteria and protista in cycling of OC (Stone et al. 1993; Hart et al. 2000); this was subsequently confirmed by experimental measurements (Berman et al. 2004, 2010). As can be seen in Table 25.7, BCD (calculated as the sum of BP plus BR), a measure of the OC flux through the heterotrophic bacteria, was always a relatively high proportion of the total autochthonous OC inputs, ranging from 45 to 91 % of GPP in winter–spring and 48 to 77 % of GPP in summer–fall (for more details on BCD, see Sect. 25.1).

5.4 Generalized Organic C Budget for Epilimnic Waters of Lake Kinneret

We show a generalized diagram quantifying annual OC fluxes for the epilimnic waters of Lake Kinneret based on averaged annual rates from 2001 to 2010 in Fig. 25.13. Here, we have included estimates for minor inputs such as chemosynthetic primary production, OC in river inflows from the watershed, direct rainfall on the lake, benthos, littoral vegetation and dust , as well as minor OC sinks such as OC in outflows to the National Water Carrier , local water consumers, southern Jordan River outflow, and water supply to the Kingdom of Jordan.

The decade-long summary shown in Fig. 25.13 would imply that over this period, ~ 94 % of OC input derived from autochthonous , photosynthetic carbon fixation by lake phytoplankton. Possibly, this underestimates the contribution of chemosynthetically fixed OC (Sect. 24.2), albeit most chemosynthetic activity would not occur within the epilimnic waters of the lake. Most (~ 83 %) of the OC input was eventually respired. A considerable portion of the OC pool was taken up, recycled, and respired through the microbial communities; BCD accounted for ~ 65 % of GPP annually (Table 25.7). Reflecting the net autotrophic status of Lake Kinneret, about ~ 20,000 t C (~15 % of the total OC input) was transferred to the sediments annually.