Conditions for sink-to-source transitions and runaway feedbacks from the land carbon cycle

Abstract

The first GCM climate-carbon cycle simulation indicated that the land biosphere could provide a significant acceleration of 21st century climate change (Cox et al. 2000). In this numerical experiment the carbon storage was projected to decrease from about 2050 onwards as temperature-enhanced respiration overwhelmed CO2-enhanced photosynthesis. Subsequent climate-carbon cycle simulations also suggest that climate change will suppress land-carbon uptake, but typically do not predict that the land will become an overall source during the next 100 years (Friedlingstein et al., accepted). Here we use a simple land carbon balance model to analyse the conditions required for a land sink-to-source transition, and address the question; could the land carbon cycle lead to a runaway climate feedback? The simple land carbon balance model has effective parameters representing the sensitivities of climate and photosynthesis to CO2, and the sensitivities of soil respiration and photosynthesis to temperature. This model is used to show that (a) a carbon sink-to-source transition is inevitable beyond some finite critical CO2 concentration provided a few simple conditions are satisfied, (b) the value of the critical CO2 concentration is poorly known due to uncertainties in land carbon cycle parameters and especially in the climate sensitivity to CO2, and (c) that a true runaway land carbon-climate feedback (or linear instability) in the future is unlikely given that the land masses are currently acting as a carbon sink.