Dr. Frederic Rees1, Dr. Bruno Andrieu1, Dr. Céline Richard-Molard1, Pr. Claire Chenu1
1UMR ECOSYS, INRA, AgroParisTech, Université Paris-Saclay, Thiverval Grignon, France
Plants represent the main source of organic carbon in soils. While inputs of carbon from the aerial parts are easy to measure, the difficulty in quantifying inputs from roots has prevented us to estimate the actual potential of soil carbon sequestration associated to each type of organic matters released by the roots. Besides the decay of root tissues, root systems have been shown to release organic carbon by various mechanisms, e.g. exudation of soluble compounds, mucilage secretion, and cells sloughing. Such rhizodeposition processes may consume 5% to 15% of the total amount of carbon photosynthetically fixed by the plant, and can generate an input of carbon to the soil ranging from 0.5 to 5 tC ha-1 yr-1. Because of this large range and the uncertainties associated to rhizodeposition mechanisms, current models of soil organic matter dynamics poorly assess the actual sequestration potential offered by plants. We performed a meta-analysis of the literature data generated over the last 60 years in order to assess which rhizodeposition processes quantitatively prevail, and what is the maximal amount of carbon inputs that can be expected from an entire growing root system, depending on plant species, growth stage, soil properties, and environmental conditions. According to our current dataset, the exudation of soluble sugars represents the major flow of organic carbon into the rhizosphere in most conditions, but other rhizodeposition processes such as mucilage secretion and cells sloughing can locally become as important in terms of carbon release. Following the presentation of these results, the consequences of these emissions on the actual soil carbon sequestration potential will be discussed based on our current understanding of carbon use efficiency and priming effect.