Midnight sun: Contributing to climate change?

midnight sun

Editor's Introduction

Sunlight controls water column processing of carbon in arctic fresh waters

annotated by

Arctic soil that has been frozen for many years is beginning to thaw.  Organic carbon that has been trapped in the frozen ground is now able to move in the environment. Much of the newly released organic carbon ends up dissolved in the water flowing into arctic streams, rivers, and lakes. In this paper, the authors study the importance of sunlight in breaking down dissolved organic carbon (DOC) into carbon dioxide (CO2) in Arctic surface waters.  DOC degradation is important to understand because accurate knowledge of how much CO2 is released to the atmosphere is necessary to predict future climate change.

Paper Details

Original title
Sunlight controls water column processing of carbon in arctic fresh waters
Original publication date
Vol. 345, Issue 6199, pp. 925-928
Issue name


Carbon in thawing permafrost soils may have global impacts on climate change; however, the factors that control its processing and fate are poorly understood. The dominant fate of dissolved organic carbon (DOC) released from soils to inland waters is either complete oxidation to CO2 or partial oxidation and river export to oceans. Although both processes are most often attributed to bacterial respiration, we found that photochemical oxidation exceeds rates of respiration and accounts for 70 to 95% of total DOC processed in the water column of arctic lakes and rivers. At the basin scale, photochemical processing of DOC is about one-third of the total CO2 released from surface waters and is thus an important component of the arctic carbon budget.


Carbon dioxide emissions from inland surface waters to the atmosphere are as large as the net carbon transfer from the atmosphere to Earth’s surface (~2 Pg C year-1; 1 Pg = 1015 g) (15). This large flux is affected by the movement of dissolved organic carbon (DOC) from land (12) and its subsequent oxidation to CO2 in fresh waters (35). The remaining DOC may be unprocessed, flocculated, or partially oxidized and exported in rivers to oceans, where it has important consequences for coastal ecosystems (6). Although the processing and fate of DOC is thus a critical aspect of C cycling, the factors controlling oxidation of DOC in fresh waters are poorly understood (56).

DOC oxidation in inland waters is of special concern for the Arctic because permafrost soils contain twice as much C as there is in the atmosphere (7), and they are thawing (8) and releasing C to surface waters and the atmosphere. In turn, surface waters mediate C flux to the ocean and atmosphere that could represent up to 40% of the net land-atmosphere carbon exchange (9) (maximum freshwater flux of ~0.16 Pg C year-1 and a net arctic terrestrial sink of 0.4 ± 0.4 Pg C year-1). Because some fraction of newly thawed C is labile to biological and photochemical degradation to CO2 (1011), there is active debate on how and when the arctic C cycle will amplify global climate change (1213).

The fate of DOC released to inland waters is controlled by its coupled photochemical and biological oxidation. The current understanding of controls on DOC oxidation is that photochemical mineralization to CO2 constitutes only about 10% of bacterial respiration (i.e., biological mineralization) in inland or coastal waters (1419). However, no study has quantified concurrently all the dominant biological and photodegradation pathways of DOC, which include bacterial respiration to CO2, photomineralization to CO2, partial photo-oxidation of DOC (which is likely less labile to bacterial action and has a longer residence time in the DOC pool; DOCoxFig. 1), and photostimulated bacterial respiration of partially photo-oxidized DOC that is easily respired by bacteria to CO2. There is ample evidence that partial photo-oxidation can alter DOC quality and stimulate bacterial respiration (2022), and in arctic rivers this photostimulated bacterial respiration can be a substantial portion of dark bacterial respiration of DOC (11).

Fig. 1.  Conceptual model and pathways of DOC processing and fate.  Dissolved organic carbon (DOC) released from soils (left) can be completely oxidized to CO2 (dark bacterial respiration or photomineralization) or partially oxidized (partial photo-oxidation, DOCox) and remain in the DOC pool to be transported in rivers to oceans (right). DOC in headwater streams fed directly by soil waters has low prior light exposure and is more labile to photochemical mineralization to CO2 (e.g., red section of pie for Imnavait Creek) relative to Sagavanirktok River water, where partial photo-oxidation to DOCox dominates. Mean processing rates of C (mmol C m−2 day−1) are shown for three sites representative of the landscape continuum, from low history of light exposure leaving soils (left) to high history of light exposure after a longer time spent in surface waters (right). The C processed by sunlight ranged from 70 to 95% of the total C processed.

This figure is a flow chart showing how dissolved organic carbon (DOC) is processed as it moves from the soil, to small streams, and then into downstream lakes and rivers.

The pie charts show the relative importance of different DOC degradation pathways for the different field sites.

For new DOC in streams, photomineralization to carbon dioxide is most important, but partial photodegradation is more important for older DOC.

DOC chemistry


DOC and sunlight


Field site locations


Although no data exist that directly quantify both photochemical and bacterial degradation of DOC in arctic inland waters, these waters are characteristically rich in light-absorbing chromophoric dissolved organic matter (CDOM) labile to photodegradation (22), and they tend to be shallow (mean depth <2 m) (2324); hence, the entire water column may be exposed to ultraviolet (UV) light (11). We quantified the importance of photodegradation in DOC processing by directly measuring dark bacterial respiration and each of the DOC photoproducts listed above. Rates of photomineralization to CO2 and partial photo-oxidation of DOC were greater than rates of dark respiration by bacteria in the water column of arctic streams and lakes by a factor of 3 to 19, which suggests that sunlight controls the fate of DOC in arctic surface waters.

DOC processing was measured for streams and lakes in and near the Kuparuk River basin on the North Slope of Alaska during the ice-free summers of 2011 to 2013 (24) (fig. S1). We quantified dark bacterial respiration (N = 81) and measured the wavelength-specific apparent quantum yields of (i) photomineralization of DOC (N = 97), (ii) partial photo-oxidation of DOC (N = 97), and (iii) photostimulated bacterial respiration (N = 147) as measured by O2 consumption and CO2 production (24). Areal rates of bacterial respiration and each photochemical process were calculated by integrating apparent quantum yields with incoming UV radiation and light absorption by DOC over mean water column depth (24) in 73 different surface waters. Areal rates were then scaled to the entire Kuparuk basin, using optical properties from 135 different lakes and 73 rivers as well as 2153 measures of CDOM from the foothills to the coastal plains in the Alaskan Arctic.

Average areal water column rates of DOC photodegradation exceeded dark bacterial respiration of DOC and accounted for 70 to 95% of total C processed (Table 1). Photomineralization rates alone exceeded dark bacterial respiration rates by a factor of nearly 5 on average; in Imnavait Creek, a representative shallow-headwater stream characterized by high CDOM and low pH, photomineralization exceeded dark bacterial respiration by a factor of more than 10 (table S1). Even in the deepest site measured, Toolik Lake (mean depth 7 m), areal photomineralization was more than half the dark bacterial respiration in the water column (Table 1) and was a similar or greater percentage than comparable measurements made outside the Arctic (1417) or in coastal waters (18). Areal rates of partial photo-oxidation of DOC also exceeded dark bacterial respiration at every site except for Imnavait Creek, by a factor of 1.7 in Toolik Lake and by a factor of nearly 15 in the low-DOC, glacial-fed Sagavanirktok River (Table 1). Areal rates of photostimulated bacterial respiration were similar to water column rates of dark bacterial respiration in streams but were lower than the dark bacterial respiration rates in the deeper water columns of lakes studied by a factor of 2 to 7 (Table 1) and were similar to or higher than comparable measurements outside the Arctic (17).

Table 1.  Areal rates of water column processing of DOC by sunlight and bacteria in experiments at near-ambient temperature and light conditions.  Data are means ± SE averaged over 2011–2013 (24). Dark bacterial respiration and photomineralization both oxidize DOC to CO2. Photo-oxidation includes all photochemical processes (complete or partial oxidation of DOC) that consume O2. Partial photo-oxidation [calculated as (mol photochemical O2 consumption) – (mol photochemical CO2 production) × 0.5, assuming 0.5 mol of O2 consumed to 1 mol DOC oxidized (2437)] represents only the amount of O2 consumed that does not directly yield CO2 (i.e., DOC to DOCox in Fig. 1). Some DOCox produced by partial photo-oxidation is subsequently oxidized completely by bacteria to CO2(i.e., photostimulated bacterial respiration); however, in all but Imnavait Creek, the dominant partial photo-oxidation product was DOCox remaining in the dissolved C pool. Total C processed is the sum of dark bacterial respiration, photomineralization, and the larger of partial photo-oxidation or photostimulated bacterial respiration. All terms were converted to C basis on the assumption that 1 mol O2 is consumed per mol CO2 respired for dark or light bacterial respiration (measured as O2 consumption relative to killed controls) (24). Percentage of C processed by sunlight was calculated as [(photomineralization) + (the larger of partial photo-oxidation or photostimulated bacterial respiration)]/(total C processed). Note that partial photo-oxidation was greater than photostimulated bacterial respiration at all sites except Imnavait Creek. Ratio of photochemical to biological mineralization was calculated as (photomineralization)/(dark bacterial respiration).

At all of the sampling sites, at least 70% of DOC processing involves light.

Toolik Lake has the highest percentage of microbially processed DOC.  This is likely because this is the deepest water body in Table 1, and there is less light farther underwater.

At all sites except Imnavait Creek, more partially degraded DOC was produced than CO2.

Information about how the authors reached these numbers is provided in the following panels.

Areal measurements


Determining DOC degradation


O2, CO2, and DOC degradation


The dominance of photochemical over biological C processing in the water column of arctic surface waters, relative to that of lower-latitude waters, is likely due to three reasons. First, the measured volumetric rates of dark bacterial respiration tend toward the low end of the range reported for fresh waters (172527), thereby elevating the relative importance of sunlight over bacteria in this study. Second, our apparent quantum yields were on the high end of the range measured in waters outside the Arctic (162829) (fig. S4), which suggests that DOC in arctic surface waters is more labile to photodegradation relative to DOC in other fresh waters; this is consistent with prior evidence for high reactivity of arctic DOC to photodegradation (11223031) (discussed below). Third and perhaps most important, compared to previous, similar measurements in relatively deep lakes of forested regions (141729), the shallow and unshaded surface waters in the Arctic essentially confine photoprocessing of DOC to a thin boundary layer that maximizes DOC exposure to sunlight and facilitates photochemical degradation (table S2 and fig. S3). Given these conditions, it is likely that photochemical DOC degradation is substantial in any shallow and unshaded inland surface water and plays an important role in freshwater carbon cycling. For example, a recent study suggested that photomineralization alone provides ~10% of the CO2 emitted to the atmosphere from inland surface waters globally (29), and in our study partial photo-oxidation plus photostimulated bacterial respiration processed twice as much DOC as did photomineralization (Table 2).

Table 2.  Estimates of average annual photochemical and bacterial processing of DOC in the Kuparuk River basin, 2011–2013.  Data are means ± SE. Photochemical reaction rates were calculated as defined in Table 1. Estimates are based on 2011–2013 UV and photosynthetically active radiation data at Toolik Field Station, using an ice-out date of 24 May for streams and rivers and 16 June for all lakes; bacterial respiration is calculated for the annual range of temperatures in lakes that do not freeze to the bottom. The surface area of first- to third-order streams, fourth-order and larger rivers, and lakes was 5.937 km2, 12.26 km2, and 323.8 km2, respectively.

The total amount of DOC processed is highest in lakes and lowest in streams.

A much higher percentage of DOC processing is driven by sunlight in rivers and streams than in lakes.

With all of the measurements combined, approximately 55% of CO2 released from lakes, rivers, and streams was processed by light.

Table 1 -vs- Table 2

In Table 1, the DOC degradation rates were reported as rates per square meter (areal rates). This is useful for directly comparing the importance of degradation processes at different sites.

In Table 2, the reported numbers are total degradation rates per year for each type of surface water. Because lakes have by far the largest surface area, all of the types of degradation have the highest rate per year in lakes, even though sunlight-driven degradation is relatively more important in streams and rivers than in lakes.

Stream order


We scaled these measures of DOC processing to the open-water period in three water types—small streams, larger rivers, and lakes—using (i) mean biodegradation rates and mean apparent quantum yields for photoreactions (Table 2 and fig. S4), (ii) measured and modeled UV radiation in the atmosphere, and (iii) underwater light absorption (11). As expected, measured light attenuation coefficients Kd,λ were strongly related to underwater light absorption by CDOM (aCDOM,λ) (R2 = 0.94, N = 100), and we used our measured aCDOM,λ (N = 2153) to predict Kd,λ when it was not measured directly (fig. S2). Results for the open-water period highlight the day-to-day variability in available UV radiation due to atmospheric conditions and the seasonal decrease in zenith angle after summer solstice (Fig. 2). Although the overall dominance of photochemical processing in streams and lakes remained, the fate of DOC varied consistently by water type. In small streams, DOC was mainly mineralized by sunlight to CO2, whereas in lakes the main fate of DOC was partial photo-oxidation (Table 1 and Fig. 2). Large rivers were intermediate between these end members, and photomineralization to CO2 was about equal to or less than partial photo-oxidation (Table 1 and Fig. 2). We suggest that this pattern is a result of light exposure history; DOC leached from soils into headwater streams has little prior light exposure and is labile to complete photo-oxidation, but as light exposure increases in water moving downstream and into lakes with longer residence times, the DOC photolability declines (11223234). Thus, as easily photomineralized moieties are removed, DOC fate shifts toward partial photo-oxidation and downstream export in rivers and lakes.

Fig. 2.  Seasonal carbon processing in surface waters by bacteria and photochemistry. Daily areal water column rates of dark bacterial respiration, photomineralization, photostimulated bacterial respiration, and partial photo-oxidation of DOC in (A) first- to third-order streams, (B) fourth-order and larger rivers, and (C) lakes in 2012 (2011 and 2013 show similar patterns). Solid lines show the average rate; shaded areas in similar colors denote the upper and lower 95% confidence intervals of the apparent quantum yield and aCDOM,λ used to calculate photochemical rates, or the 95% confidence intervals on mean bacterial respiration, for each water type (24).

These graphs are easiest to read if you only look at the lines—they represent the average amount of DOC degraded by each process.

The shading around the lines represents the amount of uncertainty in the authors’ calculations based on their measurements of the water samples.


How is it different seeing these data in Figure 2 versus reading them in Table 2?

Why is there less of the sunlight-driven processes in September than in June?

What are some sources of day-to-day variation in these data?

Carbon processing at the scale of the Kuparuk River basin (~8000 km2) was estimated by summing areal rates over time (Fig. 2) for the period 2011–2013 and multiplying by the surface area of each water type (Table 2). Year-to-year variation was small because of similar degradation rates and light availability across years (Fig. 2), and photochemical reactions accounted for 75% of the total of 6.11 Gg DOC-C year-1 processed in surface waters, of which 3.34 Gg C year-1 was released to the atmosphere as CO2 (~55% from photoreactions; Table 2). Converting to a rate per square meter of land in the Kuparuk Basin, ~0.4 g C m−2 year-1 was mineralized to CO2in the water column, which is 25 to 40% of the estimated CO2 released to the atmosphere from arctic fresh waters (49). The remaining CO2 released to the atmosphere is likely generated in soil waters and transferred directly to surface waters (34) or is respired by bacteria and autotrophs in aquatic sediments (3536). For example, net benthic respiration of 6.6 mmol CO2 m−2 day−1 in a tundra pond (35) is about two-thirds of the photomineralization we estimate for coastal ponds (Table 1), and limited data for the Kuparuk River (36) show that net benthic respiration could exceed photomineralization (Table 1).

At present, our data show that for the entire Kuparuk basin 55% of the DOC processed in the water column of aquatic ecosystems is completely oxidized to CO2, and 45% is partially oxidized and transported downstream in the DOC pool (Table 2). This balance of DOC fate may change, as thawing soils will release DOC with little prior light exposure and greater potential for photomineralization to CO2, while earlier ice-out on lakes could increase UV and photoreactions substantially because solar radiation is highest and clear-sky days are more common in May and June relative to the July–October period (Fig. 2 and fig. S6). Understanding the controls on these pathways of DOC fate, and incorporating photochemical processing into models, will improve predictions of how the arctic C cycle will respond to and perhaps amplify climate change.

Supplementary Materials


Materials and Methods
Figs. S1 to S9
Tables S1 and S2
References (3854)

References and Notes

  1. T. J. Battin, S. Luyssaert, L. A. Kaplan, A. K. Aufdenkampe, A. Richter, L. J. Tranvik, The boundless carbon cycle. Nat. Geosci. 2, 598–600 (2009).

  2. K. Aufdenkampe, E. Mayorga, P. A. Raymond, J. M. Melack, S. C. Doney, S. R. Alin, R. E. Aalto, K. Yoo, Riverine coupling of biogeochemical cycles between land, oceans, and atmosphere. Front. Ecol. Environ 9, 53–60 (2011).

  3. J. J. Cole, N. F. Caraco, G. W. Kling, T. K. Kratz, Carbon dioxide supersaturation in the surface waters of lakes. Science 265, 1568–1570 (1994).

  4. G. W. Kling, G. W. Kipphut, M. C. Miller, Arctic lakes and streams as gas conduits to the atmosphere: Implications for tundra carbon budgets. Science 251, 298–301 (1991).

  5. P. A. Raymond, J. Hartmann, R. Lauerwald, S. Sobek, C. McDonald, M. Hoover, D. Butman, R. Striegl, E. Mayorga, C. Humborg, P. Kortelainen, H. Dürr, M. Meybeck, P. Ciais, P. Guth, Global carbon dioxide emissions from inland waters. Nature 503, 355–359 (2013).

  6. J. E. Vonk, Ö. Gustafsson, Permafrost-carbon complexities. Nat. Geosci. 6, 675–676 (2013).

  7. C. Tarnocai, J. G. Canadell, E. A. G. Schuur, P. Kuhry, G. Mazhitova, S. Zimov, Soil organic carbon pools in the northern circumpolar permafrost region. Global Biogeochem. Cycles 23, GB2023 (2009).

  8. Y. Zhang, W. Chen, D. W. Riseborough, Temporal and spatial changes of permafrost in Canada since the end of the Little Ice Age. J. Geophys. Res. 111, D22103 (2006).

  9. D. McGuire, L. G. Anderson, T. R. Christensen, S. Dallimore, L. Guo, D. J. Hayes, M. Heimann, T. D. Lorenson, R. W. Macdonald, N. Roulet, Sensitivity of the carbon cycle in the Arctic to climate change. Ecol. Monogr. 79, 523–555 (2009).

  10. E. A. G. Schuur, J. G. Vogel, K. G. Crummer, H. Lee, J. O. Sickman, T. E. Osterkamp, The effect of permafrost thaw on old carbon release and net carbon exchange from tundra. Nature 459, 556–559 (2009).

  11. R. M. Cory, B. C. Crump, J. A. Dobkowski, G. W. Kling, Surface exposure to sunlight stimulates CO2 release from permafrost soil carbon in the Arctic. Proc. Natl. Acad. Sci. U.S.A. 110, 3429–3434 (2013).

  12. E. A. G. Schuur, J. Bockheim, J. G. Canadell, E. Euskirchen, C. B. Field, S. V. Goryachkin, S. Hagemann, P. Kuhry, P. M. Lafleur, H. Lee, G. Mazhitova, F. E. Nelson, A. Rinke, V. E. Romanovsky, N. Shiklomanov, C. Tarnocai, S. Venevsky, J. G. Vogel, S. A. Zimov, Vulnerability of permafrost carbon to climate change: Implications for the global carbon cycle. Bioscience 58, 701–714 (2008).

  13. H. MacDougall, C. A. Avis, A. J. Weaver, Significant contribution to climate warming from the permafrost carbon feedback. Nat. Geosci. 5, 719–721 (2012).

  14. R. M. W. Amon, R. Benner, Photochemical and microbial consumption of dissolved organic carbon and dissolved oxygen in the Amazon River system. Geochim. Cosmochim. Acta 60, 1783–1792 (1996).

  15. W. Granéli, M. Lindell, L. J. Tranvik, Photo-oxidative production of dissolved inorganic carbon in lakes of different humic content. Limnol. Oceanogr. 41, 698–706 (1996).

  16. V. Vähätalo, M. Salkinoja-Salonen, P. Taalas, K. Salonen, Spectrum of the quantum yield for photochemical mineralization of dissolved organic carbon in a humic lake. Limnol. Oceanogr. 45, 664–676 (2000).

  17. V. Vähätalo, K. Salonen, U. Münster, M. Järvinen, R. G. Wetzel, Photochemical transformation of allochthonous organic matter provides bioavailable nutrients in a humic lake. Arch. Hydrobiol. 156, 287–314 (2003).

  18. S. Bélanger, H. Xie, N. Krotkov, P. Larouche, W. F. Vincent, M. Babin, Photomineralization of terrigenous dissolved organic matter in Arctic coastal waters from 1979 to 2003: Interannual variability and implications of climate change. Global Biogeochem. Cycles 20, GB4005 (2006).

  19. L. J. Tranvik, J. Downing, J. Cotner, S. A. Loiselle, R. G. Striegl, T. J. Ballatore, P. Dillon, K. Finlay, K. Fortino, L. B. Knoll, Lakes and reservoirs as regulators of carbon cycling and climate. Limnol. Oceanogr. 54, 2298–2314 (2009).

  20. L. J. Tranvik, S. Bertilsson, Contrasting effects of solar UV radiation on dissolved organic sources for bacterial growth. Ecol. Lett. 4, 458–463 (2001).

  21. K. E. Judd, B. C. Crump, G. W. Kling, Bacterial responses in activity and community composition to photo-oxidation of dissolved organic matter from soil and surface waters. Aquat. Sci. 69, 96–107 (2007).

  22. R. M. Cory, D. M. McKnight, Y.-P. Chin, P. Miller, C. L. Jaros, Chemical characteristics of fulvic acids from Arctic surface waters: Microbial contributions and photochemical transformations. J. Geophys. Res. 112, G04S51 (2007).

  23. J. E. Hobbie, Limnology of Tundra Ponds, Barrow, Alaska (Dowden, Hutchinson and Ross, Stroudsburg, PA, 1980).

  24. See supplementary materials on Science Online.

  25. R. Carignan, D. Planas, C. Vis, Planktonic production and respiration in oligotrophic Shield lakes. Limnol. Oceanogr. 45, 189–199 (2000).

  26. M. Pace, Y. Prairie, Respiration in Lakes (Oxford Univ. Press, New York, 2005), pp. 103–121.

  27. M. Berggren, J.-F. Lapierre, P. A. del Giorgio, Magnitude and regulation of bacterioplankton respiratory quotient across freshwater environmental gradients. ISME J. 6, 984–993 (2012).

  28. E. M. White, D. J. Kieber, J. Sherrard, W. L. Miller, K. Mopper, Carbon dioxide and carbon monoxide photoproduction quantum yields in the Delaware Estuary. Mar. Chem. 118, 11–21 (2010).

  29. Koehler, T. Landelius, G. A. Weyhenmeyer, N. Machida, L. J. Tranvik, Sunlight-induced carbon dioxide emissions from inland waters. Global Biogeochem. Cycles (2014).

  30. R. M. W. Amon, B. Meon, The biogeochemistry of dissolved organic matter and nutrients in two large Arctic estuaries and potential implications for our understanding of the Arctic Ocean system. Mar. Chem. 92, 311–330 (2004).

  31. P. J. Mann, A. Davydova, N. Zimov, R. G. M. Spencer, S. Davydov, E. Bulygina, S. Zimov, R. M. Holmes, Controls on the composition and lability of dissolved organic matter in Siberia’s Kolyma River basin. J. Geophys. Res. 117, G01028 (2012).

  32. H. E. Reader, W. L. Miller, The efficiency and spectral photon dose dependence of photochemically induced changes to the bioavailability of dissolved organic carbon. Limnol. Oceanogr. 59, 182–194 (2014).

  33. L. Molot, P. Dillon, Photolytic regulation of dissolved organic carbon in northern lakes. Global Biogeochem. Cycles 11, 357–365 (1997).

  34. Vallières, L. Retamal, P. Ramlal, C. L. Osburn, W. F. Vincent, Bacterial production and microbial food web structure in a large arctic river and the coastal Arctic Ocean. J. Mar. Syst. 74, 756–773 (2008).

  35. P. S. Ramlal, R. H. Hesslein, R. E. Hecky, E. J. Fee, J. W. M. Rudd, S. J. Guildford, The organic carbon budget of a shallow arctic tundra lake on the Tuktoyaktuk Peninsula, N.W.T., Canada. Biogeochemistry 24, 145–172 (1994).

  36. J. E. Hobbie, G. W. Kling, Eds., A Changing Arctic: Ecological Consequences for Tundra, Streams, and Lakes (Oxford Univ. Press, New York, 2014).

  37. S. Andrews, S. Caron, O. Zafiriou, Photochemical oxygen consumption in marine waters: A major sink for colored dissolved organic matter? Limnol. Oceanogr. 45, 267–277 (2000).

  38. G. W. Kling, G. W. Kipphut, M. M. Miller, J. W. O’Brien, Integration of lakes and streams in a landscape perspective: The importance of material processing on spatial patterns and temporal coherence. Freshw. Biol. 43, 477–497 (2000).

  39. R. M. Cory, K. McNeill, J. P. Cotner, A. Amado, J. M. Purcell, A. G. Marshall, Singlet oxygen in the coupled photochemical and biochemical oxidation of dissolved organic matter. Environ. Sci. Technol. 44, 3683–3689 (2010).

  40. B. C. Crump, G. W. Kling, M. Bahr, J. E. Hobbie, Bacterioplankton community shifts in an arctic lake correlate with seasonal changes in organic matter source. Appl. Environ. Microbiol. 69, 2253–2268 (2003).

  41. H. E. Adams, B. C. Crump, G. W. Kling, Temperature controls on aquatic bacterial production and community dynamics in arctic lakes and streams. Environ. Microbiol. 12, 1319–1333 (2010).

  42. H. E. Adams, B. C. Crump, G. W. Kling, Metacommunity dynamics of bacteria in an arctic lake: The impact of species sorting and mass effects on bacterial production and biogeography. Front. Microbiol. 5, 82 (2014).

  43. H. Xie, O. C. Zafiriou, W.-J. Cai, R. G. Zepp, Y. Wang, Photooxidation and its effects on the carboxyl content of dissolved organic matter in two coastal rivers in the southeastern United States. Environ. Sci. Technol. 38, 4113–4119 (2004).

  44. P. Ward, R. L. Sleighter, P. G. Hatcher, R. M. Cory, Insights into the complete and partial photooxidation of black carbon in surface waters. Environ. Sci. Process. Impacts 16, 721–731 (2014).

  45. V. Vahatalo, Light, Photolytic Reactivity and Chemical Products (Elsevier, Amsterdam, 2009), pp. 761–773.

  46. P. Morris, H. Zagarese, C. Williamson, E. G. Balseiro, B. R. Hargreaves, B. Modenutti, R. Moeller, C. Queimalinos, The attenuation of solar UV radiation in lakes and the role of dissolved organic carbon. Limnol. Oceanogr. 40, 1381–1391 (1995).

  47. J. A. L. Gareis, L. F. W. Lesack, M. L. Bothwell, Attenuation of in situ UV radiation in Mackenzie Delta lakes with varying dissolved organic matter compositions. Water Resour. Res. 46, W09516 (2010).

  48. S. C. Johannessen, W. L. Miller, Quantum yield for the photochemical production of dissolved inorganic carbon in seawater. Mar. Chem. 76, 271–283 (2001).

  49. W. Granéli, M. Lindell, B. De Faria, F. de Assis Esteves, Photoproduction of dissolved inorganic carbon in temperate and tropical lakes—dependence on wavelength band and dissolved organic carbon concentration. Biogeochemistry 43, 175–195 (1998).

  50. C. J. T. Spitters, H. A. J. M. Toussaint, J. Goudriaan, Separating the diffuse and direct component of global radiation and its implications for modeling canopy photosynthesis, part I. Components of incoming radiation. Agric. For. Meteorol. 38, 217–229 (1986).

  51. R. E. Bird, C. Riordan, Simple solar spectral model for direct and diffuse irradiance on horizontal and tilted planes at the Earth’s surface for cloudless atmospheres. J. Clim. Appl. Meteorol. 25, 87–97 (1986).

  52. S. Bertilsson, A. Burgin, C. C. Carey, S. B. Fey, H.-P. Grossart, L. M. Grubisic, I. D. Jones, G. Kirillin, J. T. Lennon, A. Shade, R. L. Smyth, The under-ice microbiome of seasonally frozen lakes. Limnol. Oceanogr. 58, 1998–2012 (2013).

  53. L. Vörös, A. Mózes, B. Somogyi, A five-year study of autotrophic winter picoplankton in Lake Balaton, Hungary. Aquat. Ecol. 43, 727–734 (2009).

  54. Environmental Data Center Team, Toolik Field Station, Institute of Arctic Biology, University of Alaska, Fairbanks (2012); http://toolik.alaska.edu/edc/abiotic_monitoring/data_query.php.

  55. Acknowledgments: We thank K. Harrold, J. Dobkowski, S. Fortin, S. Michael, J. Olsen, B. Papworth, A. Clinger, A. Rocha, W. Eugster, and researchers and technicians of the Toolik Lake Arctic LTER and Toolik Lake Field Station (GIS, R. Fulweber, J. Stuckey) for assistance. Supported by NSF grants OPP 1023270/1022876, PLR 1107593, and DEB-1147378/1347042. The data set provided by the Arctic Long Term Ecological Research Program (table S1) was supported by NSF grant DEB-1026843. The data set provided by the Toolik Field Station Environmental Data Center (table S1) was supported by NSF grants 455541 and 1048361. Other experimental data are presented in the supplementary materials.