Site location and descriptions

Incubation experiments were conducted in 11 cryoconite holes on three valley glaciers, Midtre Lovenbreen (ML), Vestre Brøggerbreen (VB) and Austre Brøggerbreen (AB), situated in the northwest of the Svalbard archipelago, close to the scientific base at Ny-Alesund (78°55’48’’ N, 11°56’59’’E; Fig. 1). All incubations took place in July 2009. All the cryoconite holes studied were open, i.e. did not have ice lids. Owing to the relatively late start of the melt season on the glaciers in 2009, six of the 11 cryoconite holes were still covered in slush.

Fig. 1. Location of cryoconite holes where in situ rates of photosynthesis and respiration were measured. AB = Austre Brøggerbreen, ML=Midtre Lovénbreen, VB=Vestre Brøggerbreen. Note that x and y labels are in Universal Tranverse Mercator (UTM) coordinates.

Cryoconite hole diameters, heights, water depths, slush depths and sediment depths were measured using a 30 cm ruler, to allow estimates of relative sediment-to-water volumes.

O₂ , temperature and dissolved inorganic carbon (DIC) measurements of water

Oxygen and temperature measurements were taken with a Hach HQ30d meter with luminescent dissolved oxygen probe following the method of Reference HodsonHodson and others (2010b). The relative standard deviation (RSD) of the luminescent oxygen probe was 1.5%. The precision of the temperature sensor is quoted by the manufacturer as ±0.1 °C. Measurements of DIC were made using a PP systems EGM-4 infrared CO2 meter, following the method of Reference HodsonHodson and others (2010b). The RSD was 0.98%.

Total organic carbon (TOC) content of cryoconite sediment

Cryoconite was first dried at 70°C for 2 days then analysed for total carbon (TC) on a Eurovector EA3000 Elemental analyser. Inorganic carbon (IC) was determined using a Coulomat 720 analyser. TOC was defined as the difference between TC and IC. The detection limit was 100| μgCg⁻¹ dry sediment.

Batch incubations for estimating rates of photosynthesis and respiration

The ΔO₂ and ΔCO₂ methods

At each site, sediment was collected from the hole and added to six glass bottles with tapered glass stoppers (60 mL BOD bottles, Wheaton) to give a sediment depth of ∼2–4mm. All bottles were filled completely with in situ supraglacial water to give approximate sediment/water ratios of 1 :60. At each site, three of the bottles were chosen randomly and covered with aluminium foil (dark bottles). The remaining three bottles (light bottles) were illuminated by the natural levels found in situ. At three sites (one on each glacier), an additional two bottles (one light, one dark) were filled completely with supraglacial water only. All bottles were submerged completely in water within cryoconite holes and incubated for 24 ± 2 hours. If present, slush was scooped back over the bottles in the holes to give the same approximate depth of slush as was present initially.

Starting oxygen concentrations and temperature were measured on each individual bottle prior to incubation. Starting values for DIC were established by filling three identical BOD bottles with supraglacial water and measuring immediately for DIC, as described above. Oxygen, temperature and DIC were also measured in the bottles at the end of the incubation period. All measurements were made in situ on the glaciers immediately after removing the bottles from the cryoconite holes. Corrections for sulphide oxidation (for oxygen measurements, Equation (1)) and carbonate dissolution (for DIC measurements, Equation (2)) during incubation were made following the procedure of Reference HodsonHodson and others (2010b).

(1)

(2)

Start and end incubation concentrations for Ca²⁺, Mg^{2 +} and SO42– were measured on a Dionex 4000i Ion Chromatography System at the University of Bristol. The acidified samples were stored refrigerated for up to 4 months prior to analysis. The RSD for Ca²⁺, Mg^{2 +} and SO₄ ²– determinations was 8.7%, 5.6% and 7.3%, with corresponding detection limits of 2.0, 0.93 and 3.6 μM, respectively (based on mean + (3 × SD) of eight blanks).

Total respiration and net community production (NCP) were determined from the O² and DIC measurements in dark and light bottles, respectively. NCP is defined here as the net balance of gross photosynthesis and respiration for the entire microbial community. Gross photosynthesis rates were calculated by subtracting the O2 and DIC concentrations of dark bottles from light bottles. In all cases, rates were normalized for the different dry weights of sediment in the bottles, determined by drying and weighing the sediment.

The detection limits for the ΔCO₂ method were 18 μgC L⁻¹ d⁻¹ (1.2 μgCg⁻¹ d⁻¹) for rates of NCP and respiration, and 26 μgC L⁻¹ d⁻¹ (1.7μgCg⁻¹ d⁻¹) for gross photosynthesis. The detection limits for the ΔO₂ method were 96 μgCL⁻¹ d⁻¹ (6.2μgCg⁻¹ d⁻¹) for rates of NCP and respiration and 136 μgCL⁻¹ d⁻¹ (8.8μgCg⁻¹ d⁻¹) for gross photosynthesis. The sediment weight normalized detection limits assume 1 g of dry sediment, which was typical for the incubations of the present study.

The ¹⁴C method

In parallel to the ΔO2 and ΔCO2 method incubations, sediment was added to six 7 mL transparent polystyrene vials with screw-cap lids to give sediment/volume ratios of approximately 1 :50. All vials were glued horizontally to glass slides so they sank to the bottom of cryoconite holes, so that they received the same ambient light levels as the glass BOD bottle incubations used for the ΔCO₂ and ΔO₂ methods. Supraglacial water was then added to fill the vials completely, leaving no headspace. At three sites (one on each glacier), an additional two vials (one light, one dark) were filled completely with supraglacial water only. Aliquots (7 μL) of a 37MBq stock of ¹⁴C-NaHCO₃ (Perkin-Elmer) were added via an automatic pipette below the level of the water and the vials capped immediately. At each site, three of the six vials were covered with foil (dark controls) and the tubes left to incubate for 24 ± 2 hours. After incubation, samples were filtered immediately onto 0.2 μm Nuclepore polycarbonate filters, the filtrates collected into 20 mL scintillation vials and 300 μL of 3 M HCl added to acidify the samples and remove DIC. The filters were then transported back to the laboratory within 4 hours and left under an upturned beaker in a fume cupboard for 1 hour with small open beakers of 25% glutaraldehyde and concentrated hydrochloric acid to kill microbial activity and remove DIC. The filters were next placed in pre-weighed scintillation vials, allowed to dry overnight at 70°C and reweighed. The acidified filtrates were shaken and left to degas overnight in the fume cupboard. Samples were transported in 20 mL scintillation vials back to the UK, then 5 mL aliquots of scintillation cocktail (Ultima Gold^TM, PerkinElmer) were added to 1 mL aliquots of filtrates, and 10 mL of scintillation cocktail added to filters. The vials were then left in the dark for 24 hours prior to counting on a PerkinElmer Tri-Carb 2810 TR scintillation counter, using external standards to correct for quenching.

DPM values were converted to carbon uptake values by:

(3)

where ¹²C uptake is the amount of carbon taken up into cells (μg C L⁻¹ d⁻¹), ¹⁴C sample is the activity of the sample (sum of filtrate and filter DPMs), C is the concentration of DIC in the incubation water (μg C L⁻¹), 1.06 is the isotope constant effect between ¹²C and ¹ C, ⁴C added is the activity of NaH¹⁴CO₃ added to each vial (DPM) and t is the duration of incubation (days).

Rates of photosynthesis were calculated by subtracting dark control from light bottle incubations. The percentage of ¹⁴C incorporated into extracellular dissolved organic material (¹⁴C-DOC), rather than into microbial cells or particulate matter, was defined as the percentage of the ¹⁴C measured in the filtrates compared with that on the filters.

Statistical analysis

Preliminary statistical tests (for normality and variance) were carried out using the SPSS statistical software package. These tests demonstrated that the ¹⁴C, ΔCO₂ and ΔO₂ datasets had approximately normal distributions; however, there were differences in variance between methods. Consequently, comparisons between methods were carried out using two-tailed t tests (p<0.05) assuming non-equal variances. Correlations were carried out using Pearson’s product-moment correlation method.

Temperature, O₂ saturation and DIC of in situ cryoconite water

The temperature of the water in all 11 cryoconite holes used for incubation experiments was 0.1 ±0.1°C. O₂ saturation levels were 100 ± 4 . 6% and DIC concentrations ranged from 29.8 to 44.9 μM (Table 1).

Table 1. Physical and chemical characteristics of cryoconite holes

Temperature, calcium, magnesium and sulphate changes in incubation water

There was no systematic difference in temperature between light and dark incubations of the ΔO₂ and ΔCO₂ methods, with a mean difference of 0.0±0.7°C (1 × SD). The mean temperatures measured at the start and end of incubations were 2.1°C ( ± 1.1°C, 1 × SD) and 1.3°C (± 0.7°C, 1 × SD), respectively, slightly higher than the mean in situ temperature of 0.1°C. However, temperatures in incubations were measured after bottles had been removed from the cryoconite holes and oxygen measurements taken. The warmer temperatures could therefore be an artefact of short-term surface warming during measurements rather than representative of incubation temperatures.

The maximum changes in Ca²⁺ and Mg^{2 +} during incubations were <6 and <2 μM, respectively, with the exception of AB1 incubations where the Ca²⁺ and Mg^{2 +} changes measured were up to 31 and 18μM, respectively. No changes in SO₄ ^{2 -} concentration during incubations were detected.

TOC content of cryoconite

The TOC content of cryoconite ranged from 14 to 32 mg Cg⁻¹ dry sediment, with an average of 23 mgCg⁻¹ dry sediment d⁻¹ (Table 1).

Comparison of A O₂ and ΔCO₂ methods for respiration and gross photosynthesis and NCP

Rates of gross photosynthesis and respiration determined by the ΔCO₂ method in cryoconite-amended incubations were of similar magnitude (6.5–32.2 μgCg⁻¹ d⁻¹ for gross photosynthesis, 4.3–37.1 μgCg^{- 1} d⁻¹ for respiration; Table 2), resulting in relatively small changes in NCP (–11.9 to 8.0 μgCg⁻¹ d⁻¹; Table 2). A comparison of paired ΔO₂ and ΔCO₂ rates of gross photosynthesis, respiration and NCP is shown in Figure 2. For all paired measurements above the detection limit (Fig. 2), the mean photosynthetic quotient (PQ, the molar ratio of O₂ produced to CO₂ consumed) was 1.24 ± 0.20 (1 × SD). The mean respiratory quotient (RQ, the molar ratio of CO₂ produced to O₂ consumed) was 0.80 ± 0 . 1 7 (1 × SD). The mean NCP quotient (defined here as the molar ratio of O₂ produced to CO₂ consumed) was 2.34 ± 1 . 9 0 (1 × SD). There were significant differences between paired ΔO₂ and ΔCO₂ measurements for photosynthesis (p<0.05, two-tailed t test) and gross photosynthesis (p<0.05, two-tailed t test), but not for NCP (p = 0.27, two-tailed t test).

Table 2. Rates of gross photosynthesis, respiration, NCP (all from ΔCO₂ method) and ¹⁴C photosynthesis in cryoconite sediment and water incubations. Mean rates are shown with minimum and maximum rates in parentheses. Negative numbers for NCP indicate net heterotrophy; positive numbers indicate net autotrophy

Fig. 2. Comparison of paired measurements of rates by CO₂ and O₂ methods. (a) Gross photosynthesis, (b) respiration and (c) NCP. Dotted lines show typical ranges of photosynthetic and respiratory quotients (PQ and RQ) from other environments: PQ range of 1.0–1.4 (Reference Williams and RobertsonWilliams and Robertson, 1991); RQ range of 0.7–1.0 (Gnaiger, 1983). Shaded areas are below the detection limit.

Rates of gross photosynthesis and respiration were substantially lower in the water-only incubations compared with cryoconite-amended incubations (Table 2). Mean rates of respiration and gross photosynthesis in water-only incubations were 7.39 and 85.6 μgCL⁻¹ d⁻¹, respectively (ΔCO₂ method) (Table 2). In situ ratios of cryoconite:water in the cryoconite holes (mean 1 :34, range from 1 :6 to 1 : 75; Table 1) were generally similar to, or lower than, that of the incubation bottles (1 :60). Rates of gross photosynthesis and respiration measured in bottle incubations should therefore approximate in situ rates of cryoconite holes.

The daily rates of gross photosynthesis were sufficient to utilize 41–291% (mean 146%) of the starting DIC pool in incubation bottles (Table 2). Since rates of NCP demonstrate only relatively small changes in DIC over the course of the incubations (Table 2; Fig. 3), this indicates that the DIC pool in incubation bottles is typically recycled completely in 24 hours. In contrast, daily rates of gross photosynthesis in water-only incubations were only capable of utilizing 0.5–1.9% (mean 1.2%) of the starting DIC pool (Table 2).

Fig. 3. Comparison of rates of photosynthesis by ¹⁴C method with (a) rates of gross photosynthesis by the ΔO₂ and ΔCO₂ methods and (b) rates of NCP by the ΔO₂ and ΔCO₂ methods. Error bars are 1 SD. Note the logarithmic scale in (a).

Comparison of the ^{1 4}C method with estimates of NCP and gross photosynthesis by the ΔCO₂ method

Measurements of photosynthesis by the ¹⁴C method ranged from 0.3 to 3.9 μgCg⁻¹ d⁻¹ (Table 2). These are closer to rates of NCP rather than gross photosynthesis as determined by the ΔCO₂ method (Fig. 3). There were significant differences between mean rates of ¹⁴C photosynthesis and gross photosynthesis (p<0.05, two-tailed t test), but not between ¹⁴C photosynthesis and NCP (p = 0.186, two-tailed t test). Paired rates of ¹⁴C photosynthesis with NCP and paired rates of ¹⁴C photosynthesis with gross photosynthesis were poorly positively correlated (R=0.57 and 0.53, respectively), with neither correlation significantly different from zero (p<0.05) (Fig. 4).

Fig. 4. Scatter plots of rates of photosynthesis measured by the ¹⁴C method vs (a) gross photosynthesis (ΔCO₂ method) and (b) NCP (ΔCO₂ method). Errors bars are 1 SD of triplicate measurements. Pearson’s product-moment correlation indicated no significant relationships (p<0.05) in either (a) or (b).

Dark bottle uptake of ¹⁴C and percentage of ¹⁴C released as dissolved organic matter

There was a significant uptake of ¹⁴C in dark bottles, with higher rates in water-only incubations (11–33%, mean 22%; Table 2) compared with sediment-amended incubations (2.6–16.0%, mean 8.0%; Table 2). The percentage of ¹⁴C released as ¹⁴C-DOC into the filtrate fraction of sediment-amended incubations was low (0.2–1.4%; Table 2). A larger percentage of ¹⁴C-DOC was measured in the filtrate fraction of water-only incubations (13.0–37.3%).

Respiratory and photosynthetic quotients (RQ and PQ) in Arctic cryoconite holes

The sediment fraction of cryoconite holes in the present study dominates both rates of gross photosynthesis and respiration (Table 2). This is consistent with previous studies on Arctic cryoconite holes (Reference Säwström, Mumford, Marshall, Hodson and Laybourn-ParrySawstrom and others, 2002; Reference Anesio, Hodson, Fritz, Psenner and SattlerAnesio and others, 2009; Reference HodsonHodson and others, 2010b). Previous studies have also assumed a 1 :1 stoichiometry between ΔO2 and ΔCO2 in Arctic cryoconite holes (Reference HodsonHodson and others, 2007; Reference Anesio, Hodson, Fritz, Psenner and SattlerAnesio and others, 2009). However, the present study demonstrated that there were significant differences (p<0.05) between the ΔO₂ and ΔCO2 methods for measuring rates of respiration and gross photosynthesis, resulting in a mean RQ of 0.88 ±0.17 (1 × SD) and a mean PQ of 1.24 ± 0.2 (1 × SD) (Fig. 2). The mean RQ is within the range of RQ values expected from variations in the composition of biological organic matter (0.7–1.0; Reference Gnaiger, Gnaiger and ForstnerGnaiger, 1983). Similarly, the mean PQ value of the present study is within the typical PQ range for algae (1.0–1.4; Reference Williams and RobertsonWilliams and Robertson, 1991), with PQ varying with the source of nitrogen (e.g. ammonia or nitrate) used by the photosynthetic cells (Reference Williams and RobertsonWilliams and Robertson, 1991).

Rates of gross photosynthesis in cryoconite holes

The rates of gross photosynthesis in cryoconite-amended incubations (mean of 17.3 μgCg⁻¹ d⁻¹ by the ΔCO₂ method) are similar to previous estimates of gross photosynthesis in Arctic cryoconite holes in Svalbard and Greenland made using the same method (mean between 15.7 and 23 μg C g⁻¹ d⁻¹, respectively; Reference HodsonHodson and others 2010a,Reference Hodsonb; Table 3). The factors limiting rates of gross photosynthesis in cryoconite holes have yet to be established, although phosphorus has been suggested as a likely limiting factor on rates of overall microbial growth (Reference Säwström, Laybourn-Parry, Granéli and AnesioSäwström and others, 2007a; Reference Stibal, Tranter, Telling and BenningStibal and others, 2008b).

Table 3. Summary of published rates of gross photosynthesis, NCP and ¹⁴C photosynthesis in cryoconite holes

The average rates of gross photosynthesis from the only available data of Antarctic cryoconite holes are an order of magnitude lower than Arctic cryoconite holes at 2.4 μg C g⁻¹ d ^{− 1} (Reference HodsonHodson and others, 2010b; Table 3). This could reflect the closed more hydrologically isolated nature of Antarctic cryoconite holes due to the thick ice lids that commonly form above them (Reference TranterTranter and others, 2004). Greater isolation both from the atmosphere and from hydrological flow would reduce the flux of both DIC and nutrients (Reference TranterTranter and others, 2004), while the thicker ice lids may lower the levels of light available to microorganisms.

The potential problem of bottle effects

This study assumes that rates of photosynthesis and respiration measured within incubation bottles accurately reflect those of the in situ cryoconite holes. However, there remains the possibility that nutrient concentrations in the bottles could change over 24 hours from in situ concentrations. This could potentially alter both rates of microbial processes and community structure in the bottles (i.e. ‘bottle effects’; Reference Li and DickieLi and Dickie, 1991). Ideally, future studies of cryoconite holes should compare shorter-term measurements of photosynthesis and respiration with 24 hour measurements to check for any bottle effects.

A further assumption of the light–dark bottle technique for estimating rates of gross photosynthesis is that rates of respiration are the same in the dark bottles as in the light bottles. Previous studies on other aquatic environments have shown that, while often true, this is not always the case and that respiration rates in the light can be higher than in the dark, leading to an underestimation of gross rates of photosynthesis (Reference YallopYallop, 1982; Reference BenderBender and others, 1987; Reference KanaKana, 1990). Processes that can cause increased respiration in light bottles include photorespiration and the Mehler reaction involving uptake of oxygen by photosystem 1 (Reference BurrisBurris, 1977; Reference YallopYallop, 1982; Reference Lancelot and MathotLancelot and Mathot, 1985). Conversely, respiration rates in the light could be lower than in the dark, either due to differences in temperature within dark and light bottles or inhibition of microorganisms by ultraviolet (UV) radiation (Reference Thomson, Dyke and WorrestThomson and others, 1980; Reference Herndl, Müller-Niklas and FrickHerndl and others, 1993). The former can be discounted, as there were no systematic temperature differences between paired light and dark bottles in the present study. Inhibition by UV radiation is also unlikely to reduce rates of respiration, since the glass walls of the incubation bottles used for the ΔCO2 and ΔO₂ methods will cut out the majority of harmful UV-B radiation (Reference Ferreyra, Demers, del Giorgio and ChanutFerreyra and others, 1997).

However, it is possible that the elimination of UV-B radiation in glass bottle incubations may have artificially increased rates of photosynthesis and respiration in the ΔCO₂ and ΔO₂ methods relative to in situ rates (Reference Cullen, Neale and LesserCullen and others, 1992; Reference Herndl, Müller-Niklas and FrickHerndl and others, 1993). In contrast to the glass bottles, the polystyrene vials used in the ¹⁴C method will have transmitted a significant fraction of the UV-B spectrum (Reference Furgal and SmithFurgal and Smith, 1997).

Uncertainties in interpreting the ¹⁴C results

Importantly, this study has shown that ¹⁴C measurements do not accurately measure rates of either gross photosynthesis or NCP in cryoconite holes (Fig. 4). However, rates of ¹⁴C photosynthesis are closer to those of NCP than gross photosynthesis (Fig. 3). The majority of aqueous ecological studies that have used the ¹⁴C method over 24 hours have concluded that it measures a rate closer to net photosynthesis rather than gross photosynthesis (Reference PetersonPeterson, 1980; Reference Dring and JewsonDring and Jewson, 1982; Reference Behrenfeld and FalkowskiBehrenfeld and Falkowski, 1997). Net photosynthesis is defined here as gross photosynthesis minus photoautotrophic respiration, in contrast to NCP which subtracts total community respiration (Reference PetersonPeterson, 1980). The model most commonly used to explain why the ¹⁴C method can measure a rate closer to net photosynthesis is that of Reference Hobson, Morris and PirquetHobson and others (1976). In this model, photosynthetically fixed ¹⁴C-organic matter is respired back to ¹⁴CO₂ and exuded outside the cell. The initial rate of photosynthesis will be closer to gross rates of photosynthesis because the initial organic matter within the cell will be low in ¹⁴C, so respiration will initially remove little ¹⁴C. As incubation time increases, the ¹⁴C content of the cell organic matter increases, so that the relative loss of ¹⁴C back to ¹⁴CO₂ will also increase, until rates eventually approach that of net photosynthesis (Reference Hobson, Morris and PirquetHobson and others, 1976; Reference Williams and LefèvreWilliams and Lefèvre, 2008).

The model of Reference Hobson, Morris and PirquetHobson and others (1976) assumes that photoautotrophic respiration is responsible for all the respiration of photosynthetically fixed ¹⁴C-organic matter. However, if heterotrophic microorganisms respire a significant fraction of the fixed ¹⁴C-organic matter, then the ¹⁴C method can potentially measure a rate lower than net photosynthesis and approach NCP (Reference Gieskes, Kraay and BaarsGieskes and others, 1979). In net heterotrophic ecosystems, however, the ¹⁴C method can only approach zero rather than NCP, as an intrinsic drawback of the ¹⁴C method is that it can only measure positive (i.e. net autotrophic) values (Reference PetersonPeterson, 1980).

There was significant heterotrophic activity within the incubations of the present study. Indeed, many of the incubations were net heterotrophic (Fig. 3), although the proportion of heterotrophic to photoautotrophic respiration in the cryoconite is not known (Reference AnesioAnesio and others, 2010). Furthermore, the dark ¹⁴C uptake results of the present study were consistent with substantial heterotrophic activity. The mean dark ¹⁴C uptakes of 8.0% and 21.9% (for sediment-amended and water-only incubations, respectively; Table 2) were of a similar magnitude to those reported previously in Svalbard cryoconite (6.4–16.0%; Reference Stibal and TranterStibal and Tranter, 2007) and higher than the 1–2% typically found in surface ocean water (Reference Li, Irwin and DickieLi and others, 1993). Such relatively high dark uptake of ¹⁴C has previously been interpreted as the results of a high proportion of heterotrophic to photoautotrophic respiration (Reference Gieskes, Kraay and BaarsGieskes and others, 1979). However, other autotrophic bacteria (in particular, ammonia-oxidizing bacteria) could also be responsible for part of the dark ¹⁴C uptake (Reference BillenBillen, 1976).

Heterotrophic microorganisms need to gain access to, and respire, the photosynthetically fixed ¹⁴C-organic matter in order to directly affect the results of the ¹⁴C method. The principal mechanisms for this are cell lysis (and promotion of cell lysis by viral infection), grazing by predators (Reference Karl, Hebel, Björkman and LetelierKarl and others, 1998) and the release of organic matter from photosynthetic cells by passive or active exudation (Reference FoggFogg, 1983). While rates of cell lysis in cryoconite holes are unknown, the virus-induced lysis of cells in cryoconite holes may be an important source of organic carbon for heterotrophic microorganisms (Reference Anesio, Mindl, Laybourn-Parry, Hodson and SattlerAnesio and others, 2007; Reference Sä, wström, Granéli, Laybourn-Parry and AnesioSäwström and others, 2007b). A high incidence of microbial infection by phage cells has also been reported in cryoconite waters on ML (Reference Sä, wström, Granéli, Laybourn-Parry and AnesioSäwström and others, 2007b) and high ratios of viruses to bacteria have been documented in cryoconite sediments and waters on AB and ML (Reference Anesio, Mindl, Laybourn-Parry, Hodson and SattlerAnesio and others, 2007). In addition, eukaryotic predatory microorganisms, including tardigrades, rotifers and ciliates, have been documented in cryoconite holes (Reference De Smet and van RompuDe Smet and Van Rompu, 1994; Reference Porazinska, Fountain, Nylen, Tranter, Virginia and WallPorazinska and others, 2004), although the ecological importance of grazing in cryoconite holes has yet to be established. However, over the timescale of incubations in our study, the exudation of dissolved or colloidal organic matter by living photosynthetic algae and cyanobacteria is likely to be more important as a source of organic matter for heterotrophic microorganisms than cell lysis or grazing (Reference Karl, Hebel, Björkman and LetelierKarl and others, 1998).

The percentages of ¹⁴C exuded as DOC in water-only incubations of the present study (13.0–37.3%; Table 2) were similar to those in other aquatic environments (typically 5–40%; Reference FoggFogg, 1983; Reference Baines and PaceBaines and Pace, 1991; Reference Morán, Estrada, Gasol and Pedrόs-AliόMorán and others, 2002), whereas cryoconite-amended incubations had relatively low percentages of ¹⁴C-DOC (0.2–1.4%; Table 2). However, in microbial mats in some benthic environments, as much as 70% of total primary production may be exuded as colloidal organic matter in the form of extracellular polysaccharides (EPS) (Stal, 1993; Reference Goto, Kawamura, Mitamura and TeraiGoto and others, 1999). The relatively low percentages of ¹⁴C-DOC in sediment incubations of the present study could therefore be explained by the exudation of colloidal ¹⁴C-organic matter. Certainly, the production of colloidal EPS is consistent with a study indicating that cyanobacteria in cryoconite bind granules together with EPS (Reference HodsonHodson and others, 2010b). Alternatively, however, the relatively low ¹⁴C-DOC percentages could reflect either an overall low rate of organic matter exudation, or an efficient uptake of exuded ¹⁴C-DOC by heterotrophic bacteria in the cryoconite-amended incubations (Reference Karl, Hebel, Björkman and LetelierKarl and others, 1998). Further research is needed to distinguish between these hypotheses.

Interpreting the ecological significance of the ¹⁴C method in the present study is therefore problematic. If the bulk of photosynthetically fixed ¹⁴C-organic matter was respired by the photosynthetic microorganisms rather than heterotrophic microorganisms, then the 24 hour ¹⁴C method could have measured a rate close to that of net photosynthesis. If, however, heterotrophic recycling was substantial, then the ¹⁴C method may have measured a rate in between net photosynthesis and zero, and of potentially little value in ecological interpretation.

NCP in Arctic cryoconite holes

Measurements of NCP by the ΔCO₂ method should be robust, as it relies simply on changes in the chemistry of one bottle (the light bottle). The relatively small rates of NCP (mean of –0.3 μgCg⁻¹ d⁻¹, with a range of –11.9 to 8.0 μg Cg⁻¹ d⁻¹, ΔCO2 method; Table 2) reflect the close balance between rates of respiration and gross photosynthesis within incubation bottles. Many of the NCP measurements were close to the detection limit (Fig. 2), probably accounting for part of the large variation in NCP quotients (1.98 ± 1.76).

Previous estimates of NCP in Arctic cryoconite holes have been similarly low to those of the present study (–0.6 to –1.8μgCg⁻¹ d⁻¹; Table 3). Previous mean measurements of photosynthesis using the ¹⁴C method have all been <1.3 μgCg⁻¹ d⁻¹ (Table 3). As discussed above, the ecological significance of the 24hour ¹⁴C data is uncertain; however, the low ¹⁴C rates are consistent with a closely balanced or net heterotrophic ecosystem. This study therefore supports the conclusion that photosynthesis in cryoconite holes has at most a small, and more likely a negligible, impact on the net creation of organic matter, suggesting instead that organic matter may be derived largely from allochthonous sources (Reference Stibal, Tranter, Benning and RehákStibal and others, 2008a; Reference HodsonHodson and others, 2010b). The major impact microbes can have in affecting the albedo of glaciers is then through the transformation and cementing together of allochthonous organic matter, rather than through the net creation of new dark organic matter (Reference HodsonHodson and others, 2010b).