"Understanding how that source of carbon dioxide will respond to future changes in temperature is really important for future prediction," Gabriel Yvon-Durocher of Queen Mary University of London and University of Exeter, UK, told environmentalresearchweb. "Previous studies focused on specific ecosystems such as lakes and rivers. We wanted to put it all together."

For individual species respiration depends exponentially on temperature, but at the ecosystem level factors such as the numbers and biomass of different species can come into play.

The team discovered that the respiration of ecosystems responds to seasonal temperature variations with remarkable similarity for a range of habitats, from lakes, rivers and the ocean to forests and grasslands. But over a full annual cycle the respiration of aquatic environments showed a greater sensitivity to temperature (average activation energy of around 0.65 eV) than terrestrial ones (average activation energy around 0.32 eV).

"On land, in the long-term, the oxidation of organic carbon must be limited by the amount of carbon fixed by plants," said Yvon-Durocher. "Because photosynthesis has a weaker temperature dependence than respiration, this constrains the long-term temperature sensitivity of respiration on land."

Aquatic environments, on the other hand, can receive organic carbon when it rains and the material is washed in, explained Yvon-Durocher. This removes the constraint of photosynthesis on the amount that the ecosystem can respire.

Over the short-term, the respiration of different ecosystems shows similar temperature sensitivity because of the fundamental biochemistry involved. "The temperature dependence of around 0.65 eV is identical to the activation energy of an individual mitochondrion," said Yvon-Durocher. "It is scale-invariant from an individual microbe to the entire ecosystem."

According to Yvon-Durocher, the study suggests climate change could potentially cause aquatic systems to become even stronger sources of carbon dioxide over a timescale of decades. "We cannot say for certain there will be a positive feedback without analysing climate change's effects on photosynthesis," he said. "But it should be considered in GCMs [global climate models]. At the moment many include a standard Q10 [a measure of the respiration increase for a 10° temperature rise] of around 2 but do not take account of this long-term feedback." (For an activation energy of 0.65 eV the Q10 factor is roughly 2.5 at 15°C.)

The team said that the results also uphold many of the predictions of the metabolic theory of ecology, for example that the size structure of the ecosystem determines its rate of respiration. "This gives us a theoretical framework to understand how biological communities are related to biogeochemical cycles at the global scale," said Yvon-Durocher.

The study incorporated measurements from a number of examples of nine ecosystem types worldwide – forests, non-forested terrestrial, soils, estuarine pelagic (open water), estuarine benthic (sea floor), lake pelagic, lake benthic, rivers, and microbial communities in the open ocean.

Yvon-Durocher believes that the most important next step is to characterize the temperature dependence of primary production and see if it differs between terrestrial and aquatic environments. Knowing the temperature dependence of both carbon fixation and respiration should help understand how climate change will affect the Earth's overall carbon sequestration capability.

The researchers reported their work in Nature.