According to this analysis, these theoretical models have different applicability for different $a\u2032$ ranges. For a more detailed study of the applicability of these models, 19 sets of optical properties were selected, which all have the same $\mu s\u2032$ ($1\u2009\u2009mm\u22121$) but different $\mu a$. For $\mu a$ equal to 0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, and $1\u2009\u2009mm\u22121$ respectively, we can obtain the reduced albedos ranging from 0.5 to 0.99. By comparing the derivation error, the best model to derive $\mu a$ and $\mu s\u2032$ at different ranges of $a\u2032$ using the six theoretical models is summarized in Table 1. First, this means that the best forward model of $\mu a$ is used to derive $\mu a$ and $\mu s\u2032$ simultaneously and then the derivation result of $\mu a$ is used to be the best result of $\mu a$; secondly, the best forward model of $\mu s\u2032$ is used to derive $\mu a$ and $\mu s\u2032$ simultaneously, and then the derivation result of $\mu s\u2032$ is used to be the best result of $\mu s\u2032$.