**2. Materials and Methods**

### *2.1. Study Area Overview*

The total management area of Dongshan Forest Farm in Dali city, Yunnan Province, China, is 98,300 hm2, including 58,100 hm2 of forest land (44,200 hm<sup>2</sup> of natural and 13,900 hm2 of artificial forest). The topography of the forest farm is high in the east and low in the west. The highest altitude is 2046 m, the lowest altitude is 1416 m, and the average altitude is 1600 m. The annual sunshine in the forest farm is more than 2700 h, the frost-free period is 125–150 days, and the annual average precipitation is 500–550 mm. The soil type in this region is mainly neutral brown soil, with a small amount of cinnamon soil and mountain

meadow soil. In 2016, according to the distribution of typical forest types of the study area, six forest stands were constructed and set as study sites; I: *Pinus yunnanensis* × *Alnus nepalensis* (2:1); II: *Pinus yunnanensis* × *Alnus nepalensis* (3:1); III: *P. yunnanensis* × *Quercus acutissima* (2:1); IV: *P. yunnanensis* × *Quercus acutissima* (3:1); V: *P. yunnanensis* × *Celtis tetrandra* (2:1); VI: *P. yunnanensis* pure forest. These forest types belonged to undeveloped forest land before the construction of mixed forests, and the plants on the land were mainly *Heteropogon contortus* and *Dodonaea viscosa*. The six forest types are 100 m apart and do not interfere with each other. The tree species arrangement of the six forest types is shown in Figure 1. The soil's properties before the tree planting were very similar. We assumed that the local soil's properties were largely a consequence of plant growth and soil protection of forest types, and the initial conditions or management for the sites were similar. To reduce the effects of slope and elevation on the soil's properties, all the selected plots were taken at a similar slope (around 4) and elevation (around 1880 m) (Table 1).

**Figure 1.** Area settings. Forest I: *Pinus yunnanensis* × *Alnus nepalensis* (2:1); Forest II: *Pinus yunnanensis* × *Alnus nepalensis* (3:1); Forest III: *Pinus yunnanensis* × *Quercus acutissima* (2:1); Forest IV: *Pinus yunnanensis* × *Quercus acutissima* (3:1); Forest V: *Pinus yunnanensis* × *Celtis tetrandra* (2:1); Forest VI: *Pinus yunnanensis* pure forest. -: *Pinus yunnanensis*; ♦: *Alnus nepalensis*; : *Quercus acutissima*; -: *Celtis tetrandra*, hereinafter the same.

**Table 1.** Basic information of experimental plots.


Notes: Forest I: *Pinus yunnanensis* × *Alnus nepalensis* (2:1); Forest II: *Pinus yunnanensis* × *Alnus nepalensis* (3:1); Forest III: *Pinus yunnanensis* × *Quercus acutissima* (2:1); Forest IV: *Pinus yunnanensis* × *Quercus acutissima* (3:1); Forest V: *Pinus yunnanensis* × *Celtis tetrandra* (2:1); Forest VI: *Pinus yunnanensis* pure forest.

### *2.2. Sample Plot Selection and Soil Sample Collection*

In March 2016, we selected *P. yunnanensis* seedlings with consistent and healthy growth cultivated in seedling base to be transplanted to Dongshan Forest Farm in Midu County, Dali, Yunnan Province. The tree species used to set the same kind of mixed forest with different mixed proportions were also seedlings with consistent growth and no diseases and pests in the plantation. The soil samples were collected in mid-June 2016 and mid-June 2018 in this experiment.

Each plot of each forest type sample plot is more than 100 m apart, and the slope and planting management measures of the plantations are basically the same. In each forest type, three plots on the diagonal were taken, and the area of each plot is 625 m2, i.e., 25 m × 25 m (Figure 2). Thus, a total of 18 standard plots were set up in our study sites. In a 625 m2 plot, five large sample plots of size 5 m × 5 m were set in the four corners and center of each forest type plot to measure the ground diameter and tree height of *P. yunnanensis*. In each 625 m2 plot, five small sample plots of size 2 m × 2 m were selected within the five large sample plots, and soil samples were divided into soil layers with a diameter of 4 cm. During sampling, litter on the soil surface was removed, and the sampling depth was 60 cm (divided into three layers of 0–20, 20–40, and 40–60 cm). Soil samples collected from five 2 m × 2 m plots of 625 m2 were layered and mixed as one replicate, and a total of three 625 m2 areas were used as three replicates. Samples were sealed in soil bags and taken back to the laboratory for testing. Each layer of soil samples in each repetition was divided into two parts. One part was sieved and stored in a refrigerator at 4 ◦C to determine the enzyme activities, and the other was dried naturally at room temperature to remove roots, stones, and other impurities in the soil sample and then screened for 2 mm to determine the physicochemical properties of the soil.

**Figure 2.** The diagram for the positions to collect soil samples in each plot.

### *2.3. Measurement Items and Methods*

A set of three physical variables of the soil was evaluated: bulk density (BD), soil moisture content (MC), and total porosity (TOP). In the laboratory, undisturbed samples were weighed, dried in a forced-air oven at 105 ◦C for 48 h, and weighed again. The bulk density (BD, g/cm3) and total porosity (T porosity, %) were determined by the cutting-ring method [23]. A total of 20 g of fresh soil was dried in an oven at 105 ◦C to a constant weight, and the soil's moisture content was measured by gravimetric analysis [24].

Four chemical variables of the soil were evaluated, pH, total nitrogen (TN), available phosphorus (AP), and total phosphorus content (TP). Soil pH was determined using a pH meter with a soil/water ratio of 1:2.5 [25]. The total nitrogen (TN) was determined by the automatic Kjeldahl method [26]. The total phosphorus (TP) was determined via the acid solution–molybdenum–antimony anti-colorimetric method [27]. The available phosphorus (AP) was extracted using 0.5 M NaHCO3 (Olsen) and measured by a spectrophotometer [28].

Three enzyme activities of soil samples, which were stored in a refrigerator at 4 ◦C, were measured: urease activity (Ure), sucrase activity (Suc), and catalase activity (Cat). Soil urease activity (Ure) was determined using the phenol-sodium hypochlorite colorimetric method, sucrase activity (Suc) was determined using the 3,5-dinitrosalicylic acid colorimetric method, and catalase (Cat) activity was determined using the potassium permanganate titration method, respectively [29]. Urease activity and sucrase activity were expressed by the mass of specific substrate (or production of specific substrate) consumed by soil enzymes in 1 g of dried soil per unit time [30]. Catalase activity was expressed by the volume of 0.02 mol/L KMNO4 consumed by soil enzymes in 1 g of dried soil per unit time [31].

We randomly selected ten *P. yunnanensis* pines from five 5 m × 5 m sample plots in each forest type and measured their ground diameter and plant height.

### *2.4. Statistical Analysis*

At the beginning of afforestation (2016) and after two years of afforestation (2018), we selected 10 soil fertility indexes of six forest types for principal component analysis. The PCs receiving high eigenvalues and comprising variables with high factor loading were assumed to be the variables that best represent the system attributes. Therefore, we selected only PCs with eigenvalues *n* > 1.0. The score of each principal component is calculated according to the factor load of each principal component and the standardized mass fraction data. The calculation formula of principal component score (F) is:

$$\mathbf{F} = \mathbf{F} \mathbf{A} \mathbf{C} \times \lambda \tag{1}$$

where FAC is the normalized data and λ is the arithmetic square root of the eigenvalue.

The composite principal component score (F overall score) is the sum of the product of each principal component score and its corresponding contribution rate:

$$\mathbf{F}\_{\text{overall score}} = \left(\mathbf{F1} \times \mathbf{R1} + \mathbf{F1} \times \mathbf{R1} + \dots + \mathbf{Fn} \times \mathbf{Rn}\right) / \left(\mathbf{R1} + \mathbf{R2} + \mathbf{Rn}\right) \tag{2}$$

where R is the variance contribution rate.

The comprehensive score of each forest type was taken as a new index to evaluate its comprehensive quality, the difference of soil quality between different forest types was measured by Euclidean distance, and the classification average method and the shortest distance method were used to systematically cluster each treatment.

All statistical data analyses were performed using Microsoft Excel 2010 (Microsoft Redmond, WA, USA), Origin 8.0 software (Origin Lab Corp., Northampton, MA, USA), and SPSS software (Ver. 22.0; SPSS Inc., Chicago, IL, USA) for Windows. One-way analyses of variance (ANOVA) and Duncan's multiple comparison test were used to assess statistically significant differences (*p* < 0.05) among different soil depths under each forest stand. Pearson correlation analysis was conducted to identify relationships among measured the soil's properties.
