Background The fingerprinting method has become one of the main methods for identifying sediment sources due to its simplicity and directness. However, there are still deficiencies in the source of uncertainty and quantitative analysis of the discriminant results in its practical application, especially in wind and water interactive erosion, the relevant research is still blank. The Danghe River originates the western Qilian Mountains and flows through the southern Mingsha Mountain, which has suffered severe siltation with the current effective reservoir capacity less than 50.0%. In this study, this watershed was regarded as the study region to evaluate the resulted uncertainty under multiple composite fingerprint method.

Methods Based on the remote sensing and field investigation, and geomorphic types. Dune, gobi, and upstream mountain within the watershed were identified as the sediment source areas, and the Danghe Reservoir was identified as mixed sediment area. For the dune and gobi, 16 surfaces samples were collected. For the upstream mountain and mixed sediment area, and 9 samples were collected from upstream mountain and mixed sediment area during two major flood events and two minor flood events using sediment traps., respectively. Based on the first-order Gaussian approximation equation, the sediment source uncertainty was quantified and analyzed using the multiple composite fingerprints method.

Results 1) In the six particle size groups, the sediment of Danghe Reservoir is mainly from the upstream mountain, followed by the gobi, and finally the Dune. 2) An appropriate particle size range is a pivotal aspect in sediment source identification using the multiple composite fingerprinting method. Different particle size groups bear distinct geochemical signatures. Selecting the appropriate range ensures that the fingerprints are truly representative of the sediment from specific sources. Therefore, meticulous consideration and optimization of the particle size range are essential steps for enhancing the accuracy and effectiveness of the multiple composite fingerprinting method in sediment source discrimination. 3) When the multiple composite fingerprint method is utilized for sediment source discrimination, the uncertainty of the results from the mixed sediment area (sink), up to 48.0% of the total uncertainty. This phenomenon can be ascribed to the integration of the geochemical characteristics and their spatio-temporal variability of all provenance areas in the mixed sediment area. Among the diverse provenance areas, the upstream mountain area leads, contributing 29.4% to the total uncertainty. Following gobi, accounting for 18.5%, while the dune has the least impact, with only 3.1%. This divergence is likely closely linked to the sample quantity and representativeness. This enables the accurate reflection of the geochemical characteristics of each sediment source area. Therefore, a larger number of samples collected from various locations within a specific sediment source area can comprehensively reveal its chemical composition and temporal-spatial variations, thereby enhancing the reliability of the fingerprinting method results.

Conclusions Multiple composite fingerprints method is used to identify sediment sources, the mixed sediment area has the most significant uncertainty on the results. In the future, the sampling quantity is appropriately increased to accurately reflect the geochemical characteristics of each sediment source. In addition, reasonable selection of the appropriate particle size range is a key issue for the fingerprint method to identify sediment sources.