Abstract
Understanding the processes that control deep soil carbon (C) dynamics and accumulation is of key importance, given the relevance of soil organic matter ( SOM ) as a vast C pool and climate change buffer. Methodological constraints of measuring SOM decomposition in the field prevent the addressing of real‐time rhizosphere effects that regulate nutrient cycling and SOM decomposition. An invasive lineage of Phragmites australis roots deeper than native vegetation ( Schoenoplectus americanus and Spartina patens ) in coastal marshes of North America and has potential to dramatically alter C cycling and accumulation in these ecosystems. To evaluate the effect of deep rooting on SOM decomposition we designed a mesocosm experiment that differentiates between plant‐derived, surface SOM ‐derived (0–40 cm, active root zone of native marsh vegetation), and deep SOM ‐derived mineralization (40–80 cm, below active root zone of native vegetation). We found invasive P. australis allocated the highest proportion of roots in deeper soils, differing significantly from the native vegetation in root : shoot ratio and belowground biomass allocation. About half of the CO 2 produced came from plant tissue mineralization in invasive and native communities; the rest of the CO 2 was produced from SOM mineralization (priming). Under P. australis, 35% of the CO 2 was produced from deep SOM priming and 9% from surface SOM . In the native community, 9% was produced from deep SOM priming and 44% from surface SOM . SOM priming in the native community was proportional to belowground biomass, while P. australis showed much higher priming with less belowground biomass. If P. australis deep rooting favors the decomposition of deep‐buried SOM accumulated under native vegetation, P. australis invasion into a wetland could fundamentally change SOM dynamics and lead to the loss of the C pool that was previously sequestered at depth under the native vegetation, thereby altering the function of a wetland as a long‐term C sink.