**1. Introduction**

Soybean is an important crop in regard to global food security and sustainable development due to its dual properties as a protein food ingredient and oilseed [1]. Soybean, a native plant of China and one of its most important crops, has been known to man for over 5000 years [2]. China's meat consumption and demand for soybean are rapidly increasing with a growing population, rising per capita income, and changing dietary preferences [3,4]. As the main soybean producer worldwide, China has transitioned from a net exporter of soybeans to a net importer since 1996, with soybean imports increasing from 1.11 Mt in 1996 to 100.33 Mt in 2020 [5]. China's soybean imports account for 60.57% of the global soybean trade volume, making the country the world's largest soybean importer and highly

**Citation:** Chen, W.; Zhang, B.; Kong, X.; Wen, L.; Liao, Y.; Kong, L. Soybean Production and Spatial Agglomeration in China from 1949 to 2019. *Land* **2022**, *11*, 734. https:// doi.org/10.3390/land11050734

Academic Editors: Le Yu and Richard Cruse

Received: 30 March 2022 Accepted: 12 May 2022 Published: 13 May 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

dependent on imports from countries, such as Brazil, the United States, and Argentina [6,7]. At present, China's soybean consumption heavily depends on international imports; however, the total population in China will peak by approximately 2030, and if the current soybean production and consumption trends persist, the soybean production and demand gap in China will continue to expand in the future [8]. In addition, soybean yields have been projected to decline by 7–19% in 2100 against the backdrop of global warming [9]. Therefore, China must urgently optimize its soybean planting area and increase soybean production to ensure its national food security.

Mastering the spatial-temporal change in soybean planting advantages and their geographical agglomeration patterns is fundamental to optimizing the spatial layout of soybean production and ensuring national food security [1,10]. In terms of soybean planting spatial changes [5,10], Sun et al. [11] studied the spatial-temporal patterns of the soybean sown area in China in response to soybean imports from 1980 to 2012, and the results demonstrate that the soybean sown area decreased in southeastern China while it increased in northwestern China. Regarding the soybean cultivation's influencing factors, Liu et al. [12] analyzed the factors causing farmers to increase soybean production, and the study found that the age of farmers, farm income, land topography, and ease of sale positively influence the behavior of farmers. In addition, soybean imports were identified as another important factor influencing soybean cultivation [11,13–15]. In terms of timing the changes in soybean planting advantages, for political and economic reasons, soybean production in China has lost its competitiveness and has been declining since the early 2000s [15]. In terms of space optimization in soybean planting, land suitability [16], climate suitability [17], and climate production potential [18] have mainly been considered. Most areas of the Sanjiang Plain are suitable for soybean cultivation, except for areas with slopes of ≥30% [16]. Zhao et al. [17] determined that the areas of high climatic suitability for soybean planting are mainly located in the northeastern and northern-central regions and that the total area of high suitability covers 1.2988 × <sup>10</sup><sup>8</sup> ha. In addition, the effects of conservation tillage [19], wheat straw mulching [20], temperature [21], CO2 [22], and drought [23,24] on soybean yields have been studied. Existing research plays a key role in optimizing the soybean production space and increasing soybean production. However, little research has studied the spatial-temporal changes in patterns of soybean production on a national scale and over a long time series; the spatial difference between the comparative advantages of soybean planting efficiency and soybean planting scale and their spatial agglomeration characteristics remains unclear.

Soybean trade exerts a negative impact on the resources and environments of both importing and exporting countries [25]. Land expansion for soybean production has increased since 2000 by 160% in Brazil and by 57% in Argentina [4], resulting in deforestation [26], greenhouse gas emissions [27], and ecological damage [28]. Across South America, 9% of the forestland lost was converted into soybean planting areas from 2000 to 2016 [4]. Simultaneously, the soybean cultivation space in China is constantly being replaced by land for the cultivation of crops, such as rice, corn, vegetables, and fruit, resulting in irrigation water usage increasing by 96.42% (3.05 km3), and the application of N fertilizer has increased by 256.65 thousand tons (almost 5 times) [15,25,29]. The optimization of the soybean planting space and enhancement of domestic soybean production to relieve pressure on resources and the environment in China and other soybean-exporting countries require immediate solutions.

With the frequent occurrence of global extreme weather hazards, the trade war between China and the United States, and outstanding structural contradictions in domestic food security, as a country with a large population, China's food security must be firmly controlled at all times. The research of this paper consists of three parts: first, this paper analyzes the spatial-temporal evolution of patterns of soybean production from 1949 to 2019; second, this paper analyzes the spatial-temporal evolution of comparative advantages in soybean production and its spatial agglomeration characteristics; third, this paper provides relevant policy implications based on the research results. The objective of this paper is to

provide a means to optimize the layout of soybean production and alleviate the structural contradictions of food security in China.

The remainder of the paper is organized as follows. Section 2 introduces the data sources and methods used. Section 3 describes the spatial-temporal evolution of patterns of soybean production and the spatial-temporal evolution of the comparative advantages of soybean production and its spatial agglomeration characteristics. Section 4 presents a discussion of the results and limitations of this study. Finally, Section 5 provides the research conclusions and policy implications.

#### **2. Data Sources and Methods**

#### *2.1. Data Sources*

A total of 31 provinces of China were selected as the study area (excluding Hong Kong, Macao, and Taiwan). Statistical and raster data were used. Statistical data used include panel data on soybean and grain crop yields, sown area, and production in 31 provinces in China from 1949 to 2019. Data were drawn from the official website of the National Bureau of Statistics (https://data.stats.gov.cn/index.htm, accessed on 20 December 2020). As raster data, we used data on China's cropland potential productivity (CPP) in 2010 from the Resource and Environment Science and Data Center, Chinese Academy of Sciences (https://www.resdc.cn/data.aspx?DATAID=261, accessed on 22 April 2022). The CPP data are based on China's cultivated land distribution, soil, and DEM data from the Global Agro-Ecological Zones model, comprehensively considering light, temperature, water, CO2 concentration, pests and diseases, agroclimatic restrictions, soil, terrain, etc. Using 1949 as the starting point, and 10-year intervals, this paper analyzed the characteristics of the spatial-temporal patterns of soybean production and sown areas and the comparative advantages of the planting efficiency and planting scale in China over eight periods.

To measure the differences in patterns of soybean production on a regional scale, China was divided into nine agricultural zones (Figure 1): the Northeast China Plain (NECP, including Heilongjiang, Jilin, and Liaoning), the Northern arid and semiarid region (NASR, including Inner Mongolia, Ningxia, Gansu, and Xinjiang), the Huang-Huai-Hai Plain (HHHP, including Beijing, Tianjin, Hebei, Shandong, and Henan), the Loess Plateau (LP, including Shanxi and Shannxi), the Middle-Lower Yangtze Plain (MLYP, including Shanghai, Jiangsu, Zhejiang, Anhui, Jiangxi, Hubei, and Hunan), the Sichuan Basin and surrounding regions (SBSR, including Chongqing and Sichuan), the Yunnan-Guizhou Plateau (YGP, including Yunnan, Guizhou, and Guangxi), Southern China (SC, including Fujian, Guangdong, and Hainan), and the Qinghai Tibet Plateau (QTP, including Qinghai and Tibet).

**Figure 1.** Spatial distribution map of CPP in China.

#### *2.2. Methods*

#### 2.2.1. Coefficient of Variation Method

The coefficient of variation method can eliminate the influence of different units and average values on results and is widely used in the analysis of spatial differences within a geographical community [30]. The spatial variations in the soybean yield, production, and sown areas in China were analyzed by calculating the coefficient of variation over different periods. The established equations are given as follows:

$$\text{CV} = \frac{\sigma}{\mu} \tag{1}$$

where CV is the coefficient of variation, *σ* is the standard deviation, and *μ* is the mean.

#### 2.2.2. Comparative Advantage Model

The soybean yield level and planting area are the results of the interactions between the regional agricultural natural resource endowment, socioeconomic and local conditions, planting system, and market demand. The soybean yield and sown area were chosen as factors of the comparative advantages of soybean cultivation efficiency and scale, respectively, in each province. The established equations are given as follows:

$$\text{SAI}\_{ij} = \frac{s\_{ij}}{s\_i} / \frac{s\_j}{s} \tag{2}$$

$$\text{EAI}\_{ij} = \frac{t\_{ij}}{t\_i} / \frac{t\_j}{t} \tag{3}$$

where *i* and *j* denote province *i* and crop *j*, respectively; *Sij* and *Sj* denote the planting area of crop *j* in province *i* and China, respectively; *Si* and *S* denote the planting area of all grain crops in province *i* and China, respectively; *tij* and *tj* denote the yields of crop *j* in province *i* and China, respectively; *ti* and *t* denote the yields of all grain crops in province *i* and China, respectively; and SAI*ij* and EAI*ij* denote the comparative advantages of the planting scale and efficiency, respectively, of crop *j* in province *i*.

#### 2.2.3. Spatial Autocorrelation Model

The spatial autocorrelation model usually includes global and local spatial autocorrelation aspects. Global spatial autocorrelation determines whether aggregation exists in the spatial distribution of the comparative advantages of soybean planting scales and the efficiency of various provinces. Local spatial autocorrelation determines the state of the spatial agglomeration or dispersion based on the similarities in values across provinces. The established equations are given as follows:

$$\text{GlobalNorm's I} = \frac{\sum\_{i=1}^{n} \sum\_{j=1}^{n} W\_{ij} (\mathbf{x}\_i - \overline{\mathbf{x}}) (\mathbf{x}\_j - \overline{\mathbf{x}})}{S^2 \sum\_{i=1}^{n} \sum\_{j=1}^{n} W\_{ij}} \tag{4}$$

$$\text{LocalNorm's I} = \frac{n(\mathbf{x}\_i - \overline{\mathbf{x}}) \sum\_{j=1}^{n} \mathcal{W}\_{ij} (\mathbf{x}\_j - \overline{\mathbf{x}})}{\sum\_{i=1}^{n} \left(\mathbf{x}\_i - \overline{\mathbf{x}}\right)^2} \tag{5}$$

where Global Moran's I is the global spatial autocorrelation index; Local Moran's I is the local spatial autocorrelation index; *n* is the number of provinces; *xi* and *xj* denote the attribute values of a certain element in provinces *i* and *j* (*i* = *j),* respectively; and *Wij* is the spatial weight matrix. The value range of the Global Moran's I index is [−1, 1]. When the significance level is provided, if the Global Moran's I index value is significantly

positive, this indicates a spatially significant clustering of regions with large (small) values of the comparative advantages of soybean planting efficiency or comparative advantages of the soybean planting scale. Conversely, if Global Moran's I is significantly negative, this indicates significant spatial differences in the comparative advantages of soybean planting efficiency or soybean planting scale between a specific region and its neighbors. If Global Moran's I = 0, no spatial correlation occurs.

#### 2.2.4. Contribution Model

The interannual variation in soybean production is the result of the combined effect of the interannual variation in the soybean sown area and soybean yield. Therefore, the contribution model is used to determine the contribution of the soybean sown area and yield to production. The established equations are given as follows:

$$A\_c = \frac{\left(A\_{\dot{j}} - A\_i\right) \cdot Y\_i}{P\_{\dot{j}} - P\_i} \tag{6}$$

$$Y\_{\mathbb{C}} = \frac{\left(Y\_{\bar{j}} - Y\_{\bar{i}}\right) \cdot A\_{\bar{i}}}{P\_{\bar{j}} - P\_{\bar{i}}} \tag{7}$$

where *Ac* is the area contribution (%); *Yc* is the yield contribution (%); *Ai* and *Aj* represent the soybean sown area in year *i* and *j* (*j* > *i*), respectively; *Yi* and *Yj* represent the soybean yield in year *i* and *j*, respectively; and *Pi* and *Pj* represent the soybean production in year *i* and *j*, respectively.

### **3. Results**
