Increasing Spatial Mismatch of Cropland-Grain Production-Population in China over the Past Two Decades

Abstract

Identifying the spatiotemporal coupling characteristics of cropland-grain production-population is essential for the rational utilization of cropland and the evaluation of national and regional food security. Based on the grain production statistical data, GlobeLand30, and WorldPop data in the years 2000, 2010, and 2020, the spatiotemporal changes in China’s cropland area, grain production, and population and their coupling characteristics over the past two decades were detected at the grid level using the models of barycenter fitting and coupled dynamic analysis. The results showed that spatial change of cropland area in China was roughly characterized by the increase in the northwest and the decrease in the southeast; while grain production was characterized by an increase in the north and a decrease in the south, and population was roughly characterized by an increase in urban areas of the southeast coastal regions and a decrease in traditional agricultural areas. The barycenter of cropland area and that of grain production moved toward the northwest and the northeast, respectively, which mismatch the spatial pattern of hydro-thermal conditions of cropland resources in China and thus result in the increased risk of the national grain production system. Meanwhile, the barycenter of grain production and that of population continued to move in opposite directions overall, and the distances between their barycenters increased from 119.65 km in 2000 to 455.16 km in 2020, indicating that the phenomenon of ‘north-to-south grain diversion’ is intensifying. Our results highlight that the spatial mismatch of cropland-grain production-population in China has increased over the past two decades.

1. Introduction

The cropland–grain–population system is based on the utilization of cropland resources and aims to meet human grain demand through grain production, consumption, and circulation [1]. Cropland resources affect grain production, and grain production and population together determine the per capita grain possession and consumption. With the continuous growth of China’s population and rapid urbanization, the reverse change between the decreasing high-quality cropland resources and the increasing demand for grain consumption has a negative impact on per capita grain possession, thus threatening China’s grain security [2,3,4]. The change in cropland resources, grain production, and population in China are spatially heterogeneous [5,6]. Given that the spatial coordination of the change in the cropland–grain–population system is an important guarantee for regional and national food security [7,8], it is important to comprehensively explore the spatiotemporal coupling characteristics of the cropland–grain–population system.

‘Storing grain in the cropland’ is one of China’s cropland and grain security strategies [9]. Under the background of rapid urbanization, the total cropland areas in China showed a remarkable decrease trend, especially in the southeast coastal areas [10,11,12], which has shifted the pattern of national grain production [13,14]. Previous studies showed that the barycenters of both cropland area and grain production moved toward the northwest in the period 1990–2005 [15], while they moved toward the southwest and the northeast during 2000–2010, respectively [16]. These results suggest that their moving directions roughly shifted from the same direction to the opposite direction. Although China’s grain production increased, the barycenter of the new cropland gradually shifted from northeast to northwest [17], which resulted in a spatial mismatch in cropland resources and grain production [18]. With the spatial changes in China’s grain production and population, the spatial patterns of grain circulation and consumption have also changed [6]. Previous studies pointed out that China’s grain circulation pattern has evolved from ‘south-to-north grain diversion’ to ‘north-to-south grain diversion’ since the mid-1990s [19,20,21]. Meanwhile, with the decrease in grain production in Southern China and China’s population gradually moving toward the south [22], the provinces in Southeast China have evolved from grain export areas to grain import areas [23]. Most of the previous studies were based on the datasets before 2010, whereas the spatial coupling characteristics in the cropland-grain-population system after 2010 are less well known.

Furthermore, most of the above studies focused on the change in cropland area were based on the grid scale, while that in grain production or population were generally based on the provincial and county scales. To keep the consistency with grain production, some studies counted the cropland area into administrative units and afterward conducted the spatial analysis [15,16]. These results may vary among scales due to the scale effect caused by the modifiable areal unit problem (MAUP) [24]. Meanwhile, China’s administrative units vary greatly among regions, and there are significant regional differences in natural and social conditions across China. Thus, using a superior research scale can reflect more detailed information on the spatial patterns of the change in cropland area, grain production, and population [4,13,25]. Recently, the increase in open access to gridded datasets and the development of GIS technology provides an opportunity for the analysis of the coupling relationship of the cropland-grain-population system at the grid scale.

The aim of this study is to explore the spatial and temporal pattern of the interactive coupling characteristics among China’s cropland area, grain production, and population at the grid scale over the past two decades. In detail, there are three major tasks: (1) to detect the spatial pattern of the change in cropland area, grain production, and population; (2) to identify the change trajectories in barycenters of cropland area, grain production, and population, respectively; (3) to further analyze spatial coupling relationship of these barycenters. The results will enhance our understanding of the trend of ‘north-to-south grain diversion’ and provide references for ensuring grain security.

2. Data Source and Methods

2.1. Data Source and Preprocessing

2.1.1. Cropland Datasets

Cropland datasets for 2000, 2010, and 2020 were extracted from GlobeLand30 products (http://www.globeland30.org (accessed on 18 September 2022)), which are referred to as a snapshot of the Earth’s land cover with a 30-m spatial resolution [26]. This data is composed of ten types of land covers such as cropland, forest, and waterbody. The accuracy evaluation showed that the overall classification accuracy of the cropland data layer in GlobeLand30 is more than 85%. Spatially, the classification accuracy of the cropland data layer in northern China is higher than that in southern China, and the classification accuracy in Fujian and Guizhou Province and Chongqing Municipality is relatively low [16].

2.1.2. Grain Production Data

The statistical data of grain production at the county level in the years 2000, 2010, and 2020 were obtained from the statistical yearbooks and statistical bulletins of provinces (autonomous regions or municipalities) in mainland China. A few missing grain production statistical data were replaced by the data for adjacent years. To match with spatial consistency of cropland area and population data, it is necessary to convert the statistical data of grain production into gridded data. Given that the grain production per unit area of cropland is roughly the same in one certain county from the perspective of a national scale, this study thus disaggregated grain production evenly to the cropland areas. Specifically, we initially linked the statistical data of grain production at the county level to the corresponding administrative units. The grain production per unit area of cropland was then calculated based on the grain production and the total areas of cropland at the county level and subsequently redistributed to the cropland data with a 30-m spatial resolution.

2.1.3. Gridded Population Datasets

Gridded population datasets for the years 2000, 2010, and 2020 were downloaded from the WorldPop population datasets (https://www.worldpop.org/doi/10.5258/SOTON/WP00645 (accessed on 18 September 2022)). The datasets contain global and national population density products with a spatial resolution of 3 arcsec (~100 m) in the WGS84 geographic projection [27,28]. Compared with sets of gridded population datasets in China, WorldPop gridded population datasets displayed higher accuracy [29,30]. However, the datasets still showed an overestimation in some areas with low population density. Therefore, the gridded datasets were adjusted proportionally to match census data at the county/city level [31] (http://www.citypopulation.de/ (accessed on 18 September 2022)).

Based on the above datasets, the gridded data of cropland area, grain production, and population density with 10 km × 10 km spatial resolution were generated by the summation method via the spatial aggregation tool in ArcGIS. Then, per capita cropland area and per capita grain possession with 10 km × 10 km spatial resolution were further calculated. The per capita cropland area will help in revealing the changes both in cropland area and population [25], and the per capita grain possession could reflect the evolution of grain production and consumption [32].

Additionally, agricultural meteorological datasets such as averaged annual precipitation and accumulative temperature above 10 °C across China, agro-ecological zones, and vector data of provincial and county administrative units were obtained from the Resource and environment science and data center of Chinese Academy of Sciences (https://www.resdc.cn/ (accessed on 18 September 2022)). Agro-ecological zones roughly reflect the spatial differences in the natural conditions of agricultural production such as averaged annual precipitation and accumulative temperature above 10 °C (Figure 1) and thus reveal the basic pattern of the cropping system across China. Therefore, agricultural regionalization is an important guide for agricultural production. The information on the agro-ecological zones is shown in

Table 1. Summary of 11 agro-ecological zones in China.

Code	Agro-Ecological Zone	Code	Agro-Ecological Zone
NED	Northeast district	SWD	Southwest district
HHH	Huang-Huai-Hai plain	LPD	Loess Plateau district
YRD	Yangtze River Delta	IMD	Inner Mongolia district
MYR	Middle-lower reaches of Yangtze River	GXD	Gansu-Xinjiang district
PRD	Pearl River Delta	TPD	Tibetan Plateau district
SCD	South China district

. In addition, the Hu Huanyong line [33], named Hu-line, is considered to be the dividing line between agriculture and pastoral areas in China.

2.2. Methods

2.2.1. Spatial Barycenter Model

The spatial barycenter model is referred to as a useful analysis tool for understanding the evolution characteristics of geographic elements in two-dimensional space [16,34,35]. That is, the analysis of the barycenter movement helps in visually and accurately tracing the process of spatial change in geographic elements and revealing the spatial mismatch between them from a macro perspective [35,36]. This study employed this model to explore the spatial mismatch of China’s cropland area, grain production, and population by analyzing the moving distances and directions of their barycenters.

(1)

Barycenter calculation

Barycenter calculation aims to identify the barycenter coordinates of China’s cropland area, grain production, and population. Taking the values of certain elements (cropland area, grain production, population) $z_i$ as for the weight, the calculation formula of barycenter coordinates is as follows:

$$X={\sum }_{i=1}^n{z}_i{x}_i/{\sum }_{i=1}^n{z}_i$$

(1)

$$Y={\sum }_{i=1}^n{z}_i{y}_i/{\sum }_{i=1}^n{z}_i$$

(2)

where X and Y are the coordinates of the barycenter, n is the number of pixels, and $x_i$ and $y_i$ are the center coordinates of the ith pixel.

(2)

Displacement distance of barycenters

The displacement distance of barycenters could be calculated by the following formula:

$$D_{m-n}=\sqrt{{\left(X_m-X_n\right)}^2+{\left(Y_m-Y_n\right)}^2}$$

(3)

where $D_{m-n}$ denotes the displacement distance of the barycenter of a certain element (cropland area, grain production, population) between two different years (m and n). ($X_m$, Y_m) and (X_n, Y_n) represent the corresponding barycenter coordinates in the mth and nth year, respectively.

(3)

Deviation angle of the barycenter movement

The deviation angle of the barycenter movement is then calculated by the following formula:

$${\theta }_{m-n}=\left[\frac{k\times \pi }{2}+arctan\left(\frac{Y_m-Y_n}{X_m-X_n}\right)\right]\times \frac{180°}{\pi } k=0,1,2$$

(4)

where ${\theta }_{m-n}$ is the deviation angle of the barycenter movement in the mth and nth year, and the east is 0°.

2.2.2. Spatial Coupling Analysis of the Barycenters

(1)

Spatial overlap of barycenters

The spatial overlap of the barycenter was measured by the distance between the barycenters of two elements (cropland area, grain production, population) in the same year, which is used to reveal the coordination of two elements from a static perspective [37]. The closer distance denotes the higher spatial overlap and vice versa. The calculation formula can be expressed as:

$$S_{i-j}=\sqrt{{\left(X_i-X_j\right)}^2+{\left(Y_i-Y_j\right)}^2}$$

(5)

where $S_{i-j}$ is the spatial overlap of two elements, and ($X_i$, $Y_i$) and ($X_j$, $Y_j$) are the barycenter coordinates of two elements in the same year.

(2)

Change consistency of barycenters

Change consistency refers to the consistency of the deviation angle of the barycenter movement of two elements as compared with that in the previous year, which is used to measure the consistency of the moving trajectory of the barycenter of two elements from a dynamic perspective [37]. Change consistency can be expressed by vector angle, that is, the cosine value ($C$) of the angle ($\alpha$) is used as the consistency index. The value of $C$ is between −1 and 1. The larger the value, the more consistent the change. The two directions are the same when $C$ is equal to 1, while the two directions are opposite when $C$ is equal to −1. The cosine value ($C$) of angle ($\alpha$) was produced using Equation (6), as follows:

$$C=cos\alpha =\frac{\left(\Delta X_i^2+\Delta Y_i^2\right)+\left(\Delta X_j^2+\Delta Y_j^2\right)-\left[\left(\Delta X_i^2-\Delta X_j^2\right)+\left(\Delta Y_i^2-\Delta Y_j^2\right)\right]}{2\sqrt{\left(\Delta X_i^2+\Delta Y_i^2\right)\left(\Delta X_j^2+\Delta Y_j^2\right)}}=\frac{\Delta X_i\Delta X_j+\Delta Y_i\Delta Y_j}{\sqrt{\left(\Delta X_i^2+\Delta Y_i^2\right)\left(\Delta X_j^2+\Delta Y_j^2\right)}}$$

(6)

where $\Delta X$ and $\Delta Y$ denote the variation in longitude and latitude of the barycenters as compared with that of the previous time, respectively.

3. Results

3.1. Spatial Pattern of the Changes in Cropland Area, Grain Production, Population

China’s cropland area decreased by 484 × 10⁴ hm² in 2020 as compared with that in 2000. The proportion of pixels (10 km × 10 km) with the decrease of 100 hm² in cropland area is 22% of the total number of pixels, which is approximately 1.5 times the number of pixels with the cropland area increase of 100 hm² (Figure 2a). Spatially, massive cropland in the southeastern half of the Hu-line showed a decreasing trend, especially in the regions of HHH, YRD, MYR, LPD, and the southeast coastal areas of China. On the contrary, the cropland area in the farming-pastoral ecotones along the Hu-line (e.g., the west part of NED and the east part of IMD) and northwestern China (e.g., the GXD) showed a remarkable increasing trend. These results suggest the spatial pattern of change in China’s cropland area was characterized by the increase in the northwest and the decrease in the southeast.

Compared with the year 2000, China’s grain production increased by more than 200 × 10⁶ tons in 2020. The number of pixels (10 km × 10 km) with 100 tons of grain production increase accounted for 34% of the total number of pixels, which is approximately twice the number of pixels with the grain production decrease of 100 tons (Figure 2b). Spatially, the increased trend of grain production was detected in the regions of NED, HHH, LPD, and the eastern part of IMD. Contrarily, the grain production in the regions of YRD, SWD, and the southeast coastal areas (e.g., YRD and PRD) showed an obvious decrease trend. That is, the spatial pattern of change in China’s grain production showed an increasing trend in the north and a decreasing trend in the south.

China’s population increased by more than 110 × 10⁶ persons in 2020 as compared with that in 2000. The number of pixels (10 km × 10 km) with population decrease of 100 people accounts for 35% of the total number of pixels, which is about 1.5 times the number of pixels with population increase of 100 people. The urban agglomerations in the regions of YRD, PRD, and the northern HHH showed a remarkable increasing trend, followed by the provincial capital cities and their surrounding areas (Figure 2c). The areas with population displaying an obvious decrease trend are mainly distributed in the traditional agricultural areas of the regions of SWD, MYR, the eastern NED, and the central and southern parts of HHH. The above analysis showed that the patterns of change in China’s population were regional differences, which were roughly characterized by the increase in urban areas of the southeast coastal regions and the decrease in traditional agricultural areas.

Additionally, the spatial pattern of changes in per capita cropland area and per capita grain possession showed similar characteristics with cropland area and grain production overall, respectively (Figure A1 and Figure 2). The rapidly increasing trend of per capita cropland area is detected in the region around the Hu-line (e.g., NED, IMD, and LPD). However, in the southeast part of Hu-line, except for the region of NED and SWD, the areas where the per capita cropland showed a decreasing trend were larger than those that showed an increasing trend. Besides, the areas where the per capita grain possession showed an increasing trend were larger than those that showed a decreasing trend, especially in the northern part of China.

3.2. Change Trajectories of Barycenters

The barycenters of China’s cropland area, grain production, and population were all distributed in the southeast half of the Hu-line in the years 2000, 2010, and 2020, while their moving trajectory displayed obvious spatial differences (Figure 3). The barycenters of China’s cropland area during 2000–2020 were distributed in Jincheng City of Shanxi Province, the east part of LPD (Figure 3c). This barycenter moved slowly toward the west in the first decade, and then accelerated toward the northwest in the next decade, and the distances of the latter (35.81 km) was about 2.5 times of the former (

Table 2. Moving distances and directions of barycenters.

Study Period	Barycenter of Cropland Area		Barycenter of Grain Production		Barycenter of Population
	Distance (km)	Direction (°)	Distance (km)	Direction (°)	Distance (km)	Direction (°)
2000–2010	13.35	−172.25	197.49	75.02	11.69	−39.01
2010–2020	35.81	144.09	91.53	70.09	42.84	−110.32
2000–2020	46.45	155.42	288.61	73.46	47.71	−96.82

). In summary, the barycenter of the cropland area moved 46.45 km toward the northwest from 2000 to 2020.

The barycenters of China’s grain production were distributed in the region of HHH and continually moved toward the northeast over the past two decades (Figure 3b). The barycenter of grain production in 2000 was located in the middle of Henan Province and then moved to the northeast of Henan Province in the year 2010 and 2020. The moving distance in the first decade (197.49 km) is about twice that in the next decade (Table 2), and the total moving distance was 288.61 km.

The barycenter of China’s population in 2000 was located in the south of the HHH, and then moved to the north part of the MYR in 2010 and 2020 (Figure 3d). That is, the barycenters in 2000 and 2010 were located in the south of Henan Province, and that was distributed in the north of Hubei Province in 2020. The barycenters of China’s population in the first decade moved slightly toward the southeast, while it accelerated moved toward the southwest in the next decade, and the moving distance of the latter is about four times as much as the former. The total moving distance of the barycenters is 47.71 km with the moving direction of southwest overall during the period 2000–2020 (Table 2). The increase in population reflects the growth of demand for grain consumption to a certain extent, therefore, the barycenter of China’s grain consumption has moved southward.

3.3. Spatial Coupling Relationship of Barycenters

Apart from the slight increase in the spatial overlap between the barycenters of China’s cropland area and that of grain production, spatial overlap among the barycenters of population, cropland area and grain production displayed continued decrease trends in the period 2000–2020 (

Table 3. Spatial overlap of barycenters.

Study Period	Cropland Area and Grain Production	Grain Production and Population	Cropland Area and Population
2000	232.31	119.65	337.58
2010	158.90	321.99	348.27
2020	222.25	455.16	411.01

). Specifically, the spatial overlap of the barycenter of cropland area and that of grain production increased first and then decreased, resulting in the spatial overlap of their barycenters being similar in the years 2000 and 2020. The distances between the barycenter of grain production and that of population increased from 119.65 km in 2000 to 455.16 km in 2020, and the distance between the barycenter of cropland area and that of population increased from 337.58 km in 2000 to 411.01 km in 2020. As a result, both the barycenter of China’s per capita cropland and per capita grain possession moved toward the northeast over the period 2000–2020 (Figure A2). It is worth mentioning that the barycenters of per capita grain possession moved 382.65 km to the northeast from 2000 to 2020, which is longer than the moving distance of the barycenters of China’s grain production.

The change consistency of the barycenter of grain production and that of population is close to −1, indicating that the moving directions of the two are roughly opposite over the past two decades (

Table 4. Consistency of spatial changes in the barycenters.

Study Period	Cropland Area and Grain Production	Grain Production and Population	Cropland Area and Population
2000–2010	−0.38	−0.39	−0.70
2010–2020	0.28	−0.99	−0.27
2000–2020	0.14	−0.98	−0.30

). The change consistency between the barycenter of cropland area and that of grain production is 0.14, and the change consistency between the barycenter of cropland area and that of population is −0.30. These results indicate that the angle between the movement direction of the barycenter of cropland area and that of grain production is close to a right angle, and a larger angle was displayed between that of cropland area and population. Particularly, the change consistency indexes of the barycenter of grain production and that of population were negative and decreased, that is from −0.39 during 2000–2010 to −0.99 during 2010–2020, indicating that they gradually tended to move in the opposite direction. Additionally, the change consistency index of the barycenter of cropland area and that of grain production shifted from negative to positive, which suggests that their moving directions were gradually closing.

In summary, the movement distances and directions of the barycenter of China’s cropland area, grain production, and population displayed remarkable differences over the past two decades. Firstly, the movement distance of the barycenters of grain production is about six times that of both the cropland area and population. Secondly, the barycenters of grain production and that of population continuously moved in opposite directions overall, and the movement direction of the barycenter of the cropland area is roughly perpendicular to the above two. Furthermore, the spatial overlap between the barycenter of grain production and that of population showed a decreasing trend, and a similar trend occurs between barycenters of cropland area and population. These results indicate that there is an increasing mismatch in the changes in the barycenters of China’s cropland area, grain production, and population, and led to an intensifying trend of ‘north-to-south grain diversion’ across China. Additionally, although the spatial overlap and change consistency between the barycenters of cropland area and that of grain production increased slightly, they moved toward the areas with less annual precipitation and lower accumulative temperature above 10 °C (Figure 1), where the cropland productivity is low and the risk natural risk is high, thus increasing the uncertainty of China’s grain security.

4. Discussion

This study initially explored the spatial pattern of changes in China’s cropland area and grain production since the new millennium from the perspective of grain production. The barycenter of cropland area and that of grain production moved toward the southwest and the northeast, respectively, which is consistent with the results of Wang et al. [16] and Wang et al. [34]. We further found that the barycenter of cropland area moved toward the northwest overall in the period 2000–2020, while the barycenter of grain production continually moved toward the northeast (Figure 3 and Table 2). The main reasons for this movement of the barycenter of China’s cropland may be the following two aspects. On the one hand, China has implemented national strategies such as ‘the Dynamic Balance of Total Farmland Area policy’, and ‘Northeast Revitalization’ [4], which contributed to the massive reclaiming of farmland surrounding the Hu-line (Figure 2a). On the other hand, the construction occupation and ecological occupation in eastern China led to the continuous decline of cropland area [4,17,38], especially the high-quality cropland in the region of HHH and the central and north part of MYR (Figure 2a). Meanwhile, the barycenter of grain production continually moving toward the northeast might be related to the improvement of agricultural production conditions, farmland reclamation, and the adjustment of planting structure [13,16,39], especially for the region of NED and the east of IMD (Figure 2b). It is also related to the decline in rice planting area and non-grain production in the South part of China [40,41,42,43].

Although the spatial overlap and change consistency between the barycenter of cropland area and that of grain production increased slightly, their moving direction mismatched the spatial pattern of hydro-thermal conditions of cropland resources across China (Figure 1 and Figure 2a). Even though the rapid expansion of cropland in the northwest and northeast China partially offset the loss of cropland in other regions of China such as southern China [4,44], the barycenter of the cropland area and that of the grain production moved toward the northwest and the northeast, respectively, which will likely increase the risk for China’s grain production system. Firstly, the productivity of dryland crops in northwest and northeast China is unable to reach the paddy field loss in southern and eastern China [39]. Secondly, the negative impact of climate conditions on grain production should not be ignored. For example, the severe water shortage in northern China, especially in the regions of GXD and IMD, will aggravate the conflict between grain production and environmental sustainability [4], which in turn leads to inefficient use of cropland [18]. Besides, natural disasters seem to be more frequent in Northeast and Northwest China than in the South, and climate disasters such as extreme drought may be more frequent in the future [45]. Therefore, instead of reclaiming new cropland with low quality to maintain the total cropland area, it is better to make full use of high-quality cropland in the eastern and southern parts of China. For example, the governments implement effective policies to increase the multiple cropping index in high-quality cropland [40], accelerate the construction of high-standard cropland, and improve the degree of agricultural mechanization [46] to make up for the lack of weakening agricultural labor force (Figure 2c).

This study also analyzed the spatial coupling characteristics between China’s grain production and the demand for consumption from the perspective of production and consumption. The barycenter of grain production and that of population continually moved in an opposite direction, and their spatial overlap gradually decreased (Figure 3 and Table 3), resulting in China’s per capita grain possession moving toward the northeast (Figure A2). These results indicate that the phenomenon of ‘north-to-south grain diversion’ that occurred since the mid-1990s has intensified over the past decades [19,20,21]. These also suggest that the spatial dislocation of China’s grain production and consumption (that is population aggregation) has intensified, and the pressure on cross-regional circulation and storage has also increased, which might result in a decrease in macro-grain supply capacity [13,47]. Thus, it is suggested for governments upgrade the storage system of important agricultural products, enhance the ability of grain offsite storage, and improve the grain circulation system.

The spatial-temporal coupling characteristics of China’s cropland area, grain production, and population were explored at the pixel scale, however, some limitations exist in our study. For example, the accuracy of cropland data in southern China is relatedly low, which may cause uncertainties in the results of cropland area change [16]. Besides, the data of grain production at the county level was disaggregated evenly to the cropland areas at the grid level, which might overlook the phenomenon of non-grain production and multiple cropping index. Additionally, the barycenter analysis can reveal the spatial imbalance of China’s cropland area, grain production, and population from the macro perspective, but it is unable to explore the local distribution pattern and the evolution of these spatial imbalances [48]. Thus, more attention in future works should be paid to exploring the local distribution pattern of spatial mismatch in China’s cropland-grain production-population system with high precision datasets, and then deepen the study of theoretical and policy to achieve the spatial coordination of this system.

5. Conclusions

This study explored spatial-temporal coupling characteristics of China’s cropland area, grain production, and population over the past two decades based on the 10 km spatial resolution gridded dataset. The results showed that the spatial mismatch of cropland-grain production-population in China has increased. The barycenter of cropland area and that of grain production moved toward the northwest and the northeast, respectively, which mismatch the spatial pattern of hydro-thermal optimization conditions of cropland resources across China. Furthermore, the barycenter of grain production and that of population continuously moved in opposite directions overall, and the barycenter of China’s per capita grain possession moved toward the northeast, suggesting that the phenomenon of ‘north-to-south grain diversion’ is intensifying. Therefore, it is suggested to make full use of high-quality cropland in eastern and southern China with adequate hydro-thermal conditions, strengthen the ability of grain offsite storage, and further improve the grain circulation system.