A new method to evaluate the accuracy of the sediment source mixing model

Fingerprint identification technology has been widely used to extract sediment source proportions, but the only indicator currently available to assess the accuracy and applicability of the results is good-fit (GOF). We have proposed a new method to evaluate sediment source results and quantitatively evaluate the applicability of sediment source mixing models. A typical check dam in the Loess Plateau was used to evaluate the new method by combining field sampling and numerical simulations. Collins (C) and modified Hughes mixing (M-H) models were used to quantitatively analyze the sediment sources of the check dams built and operated until July 2017. The results showed that the best combination for the fingerprint factors in the dam-controlled watershed, was Zn, Mg, Mn, and d (0,1), which had an 86.1% identification ability. With rainfall, the relative sediment contribution rates from gullies, sloping farmland, grasslands, and branch ditches were 54.22%, 23.56%, 15.54%, and 6.68%, respectively. The M-H mixing model had a higher comprehensive score (2.72) when compared with the C mixing model (2.54). The comprehensive evaluation method could provide an effective scientific theoretical basis for optimal allocations of water and soil conservation in small watersheds.