Bacterial metabolism in small temperate streams under contemporary and future climates

1. We examined the detailed temperature dependence (0–40 °C) of bacterial metabolism associated with fine sediment particles from three Danish lowland streams to test if temperature dependence varied between sites, seasons and quality of organic matter and to evaluate possible consequences of global warming. 2. A modified Arrhenius model with reversible denaturation at high temperatures could account for the temperature dependence of bacterial metabolism and the beginning of saturation above 35 °C and it was superior to the unmodified Arrhenius model. Both models overestimated respiration rates at very low temperatures (<5 °C), whereas Ratkowsky's model – the square root of respiration – provided an excellent linear fit between 0 and 30 °C. 3. There were no indications of differences in temperature dependence among samples dominated by slowly or easily degradable organic substrates. Optimum temperature, apparent minimum temperature, Q₁₀‐values for 0–40 °C and activation energies of bacterial respiration were independent of season, stream site and degradability of organic matter. 4. Q₁₀‐values of bacterial respiration declined significantly with temperature (e.g. 3.31 for 5–15 °C and 1.43 for 25–35 °C) and were independent of site and season. Q₁₀‐values of bacterial production behaved similarly, but were significantly lower than Q₁₀‐values of respiration implying that bacterial growth efficiency declined with temperature. 5. A regional warming scenario for 2071–2100 (IPCC A2) predicted that mean annual temperatures will increase by 3.5 °C in the air and 2.2–4.3 °C in the streams compared with the control scenario for 1961–1990. Temperature is expected to rise more in cool groundwater‐fed forest springs than in open, summer‐warm streams. Mean annual bacterial respiration is estimated to increase by 26–63% and production by 18–41% among streams assuming that established metabolism–temperature relationships and organic substrate availability remain the same. To improve predictions of future ecosystem behaviour, we further require coupled models of temperature, hydrology, organic production and decomposition.