For the convenience of research, we unify the uncertainty forms of isotope ratios and ages into standard deviations. Various forms of uncertainty have been disclosed in the original literature. The specific processing procedures for each form are as follows. The most ideal form of uncertainty is the standard deviation (such as 1 SD and 1 σ), which we directly collect from literature. If the uncertainty is in the form of 2 standard deviations (such as 2 SD and 2 σ), we divide the original uncertainty by 2 and then input the new uncertainty into the database. Uncertainty is displayed in percentage form, representing relative error. Firstly, we multiply this percentage by the corresponding isotope ratio or age to obtain the absolute error. Then, we repeat program (1) or (2) to normalize the absolute error to standard deviation.

    We divide the lithology of the parent rock into three categories, namely sedimentary rock, igneous rock, and metamorphic rock. In the initial collection process, we directly inputted the lithology published together with geological age data from the original literature. However, there are over 4000 phrases used in the literature to describe lithology, which poses great difficulties for statistical analysis of data grouped by parent rock lithology. In this case, we referred to the national standards "Classification and Naming Scheme for Igneous Rocks" (GB/T 17412.1-1998) (Qin et al., 1998), "Classification and Naming Scheme for Sedimentary Rocks" (GB-T17412.2-1998) (Yang et al., 1998), and "Classification and Naming Scheme for Metamorphic Rocks" (GB/T 17412.3-1998) (Zhuang et al., 1998), extracted the original detailed lithology, and classified it into three major types. If the original literature does not provide any lithological information, we will leave the "lithology" column blank. The null percentage is 9.70%.