2.2.3. OC and EC Analysis

The concentrations of OC and EC were analyzed by DRI Model 2001 OC/EC Analyser, which was developed by the American Desert Institute (DRI). The main testing principle of this method is as follows: the sample is heated and converted into CO2 under different temperature gradients and gas environments. CO2 is reduced to CH4 by catalyzing MnO2 and is detected by using a flame ion detector (FID). Then, using a 633 nm helium/neon laser to detect the anti-light intensity of filter paper to detect the production of organic pyrolysis carbon (OPC), eight different carbon components (OC1, OC2, OC3, OC4, OPC, EC1, EC2, and EC2) were obtained. IMPROVE (Interagency Monitoring of Protected Visual Environments) defines OC as OC1+OC2+OC3+OC4+OCPyro and EC as EC1+EC2+EC3- OCPyro. The detection limits were 0.82 (OC), 0.19 (EC) and 0.93 (TC) μg/cm2, and the measuring range was 0.2~750 μg/cm2.

## 2.2.4. Principal Component Analysis (PCA) Modeling

PCA is an important multivariate statistical tool that can reduce the dimensionality of large datasets and extract the number of principal components needed to explain all the variance of such datasets, which is much less than the original number of variables [31,32]. PCA extracts new variables by the correlation between all variables, which contain most of the information about the data, called principal components. Each variable has the same significance, and each topic has the same weight. The first component extracted explains the maximum amount of data variance. The maximum amount of remaining data variance will be further explained by successive components [33–35]. This process sets up the orthogonal distribution of components to each other, and the result of the regression adjustment of factors is simple and stable, regardless of how large a dataset is and how many variables are included in the study [36].
