Carbonate rhizoliths in loess and their implications for paleoenvironmental reconstruction revealed by isotopic composition: δ13C, 14C

Loess–paleosol sequences are important terrestrial archives for studying Quaternary climate changes. They often contain secondary carbonates including e.g. rhizoliths (calcified roots). These secondary carbonates are precipitated in isotopic equilibrium with root-derived CO2 and are therefore used to reconstruct the vegetation present during their formation based on stable carbon isotopic composition (δ13C). Usually, the chronological context of secondary carbonates in general is not mentioned, because it is assumed that they are formed synsedimentary with loess deposition. The loess–paleosol sequence at Nussloch, SW Germany, contains in its youngest part (Upper Würmian) large carbonate rhizoliths with diameters of up to 5cm and lengths of up to 1m and more, which have not been described in this profile so far. We investigated rhizoliths as well as loess adjacent to and distant from rhizoliths for carbonatic carbon (Ccarb) and organic carbon (Corg) contents, as well as their isotopic composition (δ13C, radiocarbon dating), to identify the rhizolith origin and the time frame of their formation. Considering the 13C fractionation by carbonate precipitation, the δ13Ccarb values (−10.9±0.1‰) revealed C3 plant origin of the rhizolith carbonate and the absence of large amounts of occluded primary loess carbonate. Similar 14C ages of rhizolith Ccarb and Corg (3788±59years BP and 3150±59years BP, respectively) argued for the absence of postsegregational alteration. Therefore they are suitable for the reconstruction of paleoenvironmental conditions after loess sedimentation. The 14C ages clearly indicate that rhizoliths did not form synsedimentary. Roots penetrated the loess at Nussloch after the deposition had ceased at ~15ka BP. Even in the loess adjacent to the rhizoliths (up to a distance of 5cm), δ13Ccarb values indicate the presence of secondary carbonate deriving from postsedimentary organic matter of origin other than that of the reference loess material. Hence, this postsedimentary input of younger root biomass might have masked the initial plant signal in loess–paleosol sequences, which could cause uncertainties for paleoenvironmental reconstructions based solely on loess organic matter.