2.3.2. Spatial Autocorrelation Test

The first law of geography stipulates that the closer things are in space, the stronger their correlation, also referred to as "spatial autocorrelation" [27]. To elucidate the spatial distribution of O3, global Moran's index *I* was utilized according to Equation (1):

$$I = \frac{N\sum\_{i}\sum\_{j}W\_{ij}(X\_i - \overline{X})(X\_j - \overline{X})}{\left(\sum\_{i}\sum\_{j}W\_{ij}\right)\sum\_{i}(X\_i - \overline{X})^2} \tag{1}$$

where *N* represents the number of municipal administrative divisions, *Xi* and *Xj* represent the average value of O3 concentrations in administrative regions *i* and *j*, respectively; *X* represents the mean value of O3 concentrations in all administrative regions; and *Wij* represents the spatial weight matrix. The value range of *I* was considered [−1, 1]. Note that *I* < 0, *I* = 0, and *I* > 0 indicate a spatially negative correlation, the absence of correlation, and positive spatial correlation, respectively, whereas the closer *I* is to 1, the stronger the spatial correlation is.

To facilitate the interpretation, *I* is usually transformed into a standardized statistic, *Z(I)*, using Equation (2):

$$Z(I) = \frac{[I - E(I)]}{\sqrt{Var(I)}} \tag{2}$$

where *Z(I)* represents the significance level of the global Moran's index, *E(I)* represents the expected value, and *Var(I)* represents the variance. In particular, Z < −2.58 indicates that O3 concentrations have a negative spatial correlation, and −2.58 < Z < 2.58 indicates that the spatial correlation is not significant. Finally, Z > 2.58 indicates the positive spatial autocorrelation of O3 concentrations [28,29].

#### **3. Results**

#### *3.1. Characteristics of Ozone Time Variation*
