Allochthonous organic carbon decreases pelagic energy mobilization in lakes

Over the past decade, it has been shown that unproductive lakes worldwide are net heterotrophic because bacterial respiration of allochthonous organic carbon (AOC) makes community respiration exceed primary production. Net heterotrophy means that aquatic systems are net sources of CO2 to the atmosphere but also that bacterial utilization of AOC increases bacterioplankton production (BP) and bacterial uptake of limiting inorganic nutrients at the expense of phytoplankton production (PP). We studied 15 unproductive lakes in northern Sweden with dissolved organic carbon concentrations between 3 and 22 mg L21. We found a highly significant negative relationship between the degree of heterotrophy and total pelagic energy mobilization (PP 1 BP based on AOC) per unit of limiting nutrient. We suggest that this is because the high cell phosphorous (P) requirement of bacteria makes energy mobilization per P unit considerably lower in bacterioplankton than in phytoplankton. We also suggest that the productivity of the entire pelagic ecosystem is determined by the availability of inorganic nutrients and AOC and by whether nutrients are allocated to BP or PP. Pelagic production depends on biological energy mobilization from external energy sources. Photosynthesis by phytoplankton is here an important process. In addition to autotrophic production, there is also mobilization of energy by bacteria utilizing allochthonous organic carbon (AOC) as an energy source (Jones 1992). Bacterial utilization of AOC must be separated from the bacterial secondary production in the microbial loop (Azam et al. 1983) and should be regarded as mobilization of energy from an external source by analogy with photosynthesis (Jones 1992; Jansson et al. 2000). Consequently, the energy mobilization in pelagic food webs is based on both light energy and imported chemically bound energy in the form of AOC. The importance of heterotrophic energy mobilization relative to autotrophic energy mobilization increases with increasing input of AOC, and bacterial energy mobilization is often entirely dominant in brownwater lakes (Hessen 1998; Jansson et al. 2000) and can exceed phytoplankton primary production (PP) even in ultraoligotrophic clearwater lakes (Karlsson et al. 2002). Therefore, energy consumption in pelagic food chains in a large variety of unproductive systems might depend on energy mobilized by bacteria from C sources other than PP. It can be hypothesized that pelagic systems dominated by bacteria should be less efficient in mobilization of energy and production of biomass than autotrophic systems. Bacteria can generally out compete phytoplankton for low concentrations of inorganic P (Vadstein 2000). Bacteria also have a very high P content (median, lower, and upper quartiles, respectively, from the review by Vadstein 2000: 32, 15, 55 mg P [mg C]21) compared to phytoplankton (3.8, 2.5, 5.2 mg P [mg C]21) and typically contain about 10 times (by weight) more P per C unit than phytoplankton (Vadstein 2000; Wetzel 2001). These differences are critical when P is a limiting inorganic nutrient for production of bacteria and phytoplankton, as it often is in lakes (Schindler 1977; Vadstein 2000), and imply that bacteria mobilize considerably less biomass C per P unit than phytoplankton. However, it is an open question whether, and to what extent, pelagic energy mobilization is lower with a heterotrophic than with an autotrophic base for the total production. Therefore, we tested the hypothesis that pelagic energy mobilization per unit of limiting nutrient is lower when dominated by heterotrophic bacterioplankton than by autotrophic phytoplankton. We used data from 15 unproductive small (0.02–0.27 km2) lakes in northern Sweden, studied with identical methods. The lakes have low concentrations of N and P and concentrations of dissolved organic carbon (DOC) between 3 and 22 mg L21 (Table 1). Thus, the lakes represent a gradient from ultraoligotrophic clearwater lakes to highly stained brownwater lakes (i.e., the lake types that dominate the world population of unproductive lakes). Lakes 1–4 in Table 1 are located close to Örträsket (648109N, 188559E) in the temperate region of northern Sweden and lakes 5–15 are located close to Abisko (688219N, 188499E) in the subarctic region of northern Sweden. Sampling and analytical procedures have been reported in detail for lakes 1–4 by Jansson et al. (2001) and for lakes 5–15 by Karlsson et al. (2002). All lakes were sampled during the period June–September at intervals of every second to every fourth week. Photosynthetically active radiation (PAR) was measured at every meter of the water column using a Windaus Luxmeter (Windaus Labortechnik, Clausthal-Zellerfield) in lakes 1–4 and an IL-1400 Radiometer (International Light) in lakes 5–15. Attenuation coefficients and daily irradiation data from the Umeå Centre for Marine Sciences (lakes 1–4) and Abisko Scientific Research Station (lakes 5–15) were used to calculate daily effective light climate (mean light intensity in the mixed layer) for PAR in the lakes according to Blomqvist et al. (1981). Water temperature was measured at every meter of the water column with a Ruttner sampler thermometer (lakes 1–4) and a WTW multiline P4 instrument (lakes 5–15). Water for analyses of water chemistry, bacterial biomass, bacterioplankton production (BP), and phytoplankton species composition and biomass was gathered from composite samples collected with a tube sampler (1 or 2 m long, diameter 3.4 cm). Each layer of the investigated water columns (epilimnion in stratified lakes and the whole water column in nonstrati-