Numerous studies suggest that denitrification in riparian zones removes nitrogen from groundwater as it moves from terrestrial to aquatic ecosystems. However, removal rates vary widely among sites complicating the incorporation of riparian zones into models of nitrogen movement across landscapes. Because denitrification in the riparian subsurface is often limited by the supply of microbially-available carbon, explaining how and why carbon supply varies among riparian zones using mappable landscape attributes holds practical and theoretical appeal. First principles suggest three carbon sources for subsurface microbes: (1) dissolved organic carbon leached from surface soils; (2) deep plant roots; and (3) buried, carbon-rich soil horizons deposited long ago. Working in Rhode Island USA at riparian zones mapped as outwash and alluvium, I investigated the relative importance of different carbon sources to 3 meters depth.

Field and laboratory experiments showed that both roots and buried horizons can supply carbon in the shallow subsurface (40-75 cm), but that buried horizons dominate below 75 cm. Radiocarbon dates and results from ingrowth cores showed that roots 40-75 centimeters deep grow and decompose on decadal time scales and form patches of organic matter that may influence nitrogen removal from groundwater. However, in both alluvial and outwash profiles, most roots below 80 cm are relics (usually > 140 years old) and therefore do not act as direct carbon conduits between the surface and deep subsurface. Laboratory incubations of buried soils from many sites demonstrated that high rates of carbon mineralization associated with these soils are common. In-situ groundwater incubations and 14C dating demonstrated that metabolism of ancient carbon constitutes at least 31% of total carbon mineralization >2 meters below the surface at some sites.

My results suggest that: (1) the depth of the biologically active zone extends as deep as buried horizons; (2) on outwash and alluvium the riparian surface and subsurface are largely decoupled on time scales of months to years; (3) functional classifications of riparian zones intended to support management need to include buried horizons and recognize the limited influence of surface vegetation on subsurface biogeochemistry over short time frames.