Traditional fingerprinting methods are limited in their ability to identify soil erosion sources where geologic

variations are small or where different land uses span geological boundaries. In this study, a new biomarker

for fingerprinting, specifically, n-alkanes, was used in a small catchment to identify sediment sources. The nalkanes

were based on land uses, could provide vegetation information, and were relatively resistant to diagenetic

modifications and decomposition. This study used a composite fingerprinting method that was based on two types

of fingerprint factors (27 biomarker properties and 45 geochemical properties) with 60 source samples (i.e., gully,

grassland, forest, and cropland) and nine soil profiles. Genetic algorithm (GA) optimization has been deployed to

find the optimal source contribution to sediments. The biomarker results demonstrated that young forest is the

main sediment source in this catchment, contributing 50.5%, whereas cropland, grassland and gully contributed

25.6%, 14.4% and 9.5%, respectively; the geochemistry results were similar to the biomarkers. The forest and grassland

contributions gradually increased from upstream to downstream, and the sediment contributions of cropland

gradually decreased in the direction of the runoff pathway at the check dam. In a comparison of biomarker and geochemical

fingerprinting data, the lattermay have overestimated the forest inputs to the catchment sediment yields

because of a mixed land use history (i.e., forest and grassland). The geochemical fingerprint approach limits its ability

to fully discriminate sources based on land management regimes, but the biomarker (individual n-alkanes)