Next Article in Journal
Correction: Olagunju et al. Pleurotus ostreatus Can Significantly Improve the Nutritive Value of Lignocellulosic Crop Residues. Agriculture 2023, 13, 1161
Previous Article in Journal
Social Capital, Trust, and Cultivation of Bioenergy Crops: Evidence from Germany and Greece
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Agriculture
Volume 14
Issue 3
10.3390/agriculture14030365
agriculture-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Geographical Types and Driving Mechanisms of Rural Hollowing-Out in the Yellow River Basin

by
Zhanhui Fu
1,2,
Yahan Yang
3,
Lijun Wang
1,
Xiaoyong Zhu
1,
Hui Lv
4 and
Jiajun Qiao
1,2,*
1
College of Geography and Environmental Science, Henan University, Kaifeng 475004, China
2
Key Laboratory of Geospatial Technology for Middle and Lower Yellow River Regions (Henan University), Ministry of Education, Kaifeng 475004, China
3
International Business School, Henan University, Zhengzhou 470000, China
4
College of Engineering, Zhengzhou Technology and Technology University, Zhengzhou 470000, China
*
Author to whom correspondence should be addressed.
Agriculture 2024, 14(3), 365; https://doi.org/10.3390/agriculture14030365
Submission received: 11 January 2024 / Revised: 19 February 2024 / Accepted: 22 February 2024 / Published: 24 February 2024
(This article belongs to the Section Agricultural Economics, Policies and Rural Management)
Browse Figures
Versions Notes

Abstract

:
Understanding the regional variations and mechanisms of rural hollowing-out in the Yellow River Basin (YRB) is crucial to guiding regional rural revitalization. However, further quantitative evaluation and analysis are essential to address the issue of rural hollowing-out caused by the decrease in rural population and expansion of residential land in the YRB at different spatio-temporal scales. Based on China’s census data and residential areas extracted from remote sensing images, the rural hollowing-out in the YRB is classified into five types: smart development type (SDT), human–-land recession type (HRT), population loss type (PLT), land expansion type (LET), and human–land–vacant waste type (HLW). Then, the influential features shaping the spatial diversity of rural hollowing-out types are examined, and the feature importance values at different spatio-temporal scales are assessed using the XGBoost model. The results of rural hollowing-out in the YBR indicate that (1) the geographical types of rural hollowing-out in the YRB are dominated by the HRT type and show significant heterogeneity and distribution at different spatio-temporal scales. At different time stages, the number of counties dominated by HRT in lower reaches accounts for 57% of the total counties, whereas the number of counties in the middle reaches is only 37%. Compared to the rural hollowing-out results from 2000 to 2010, the number of counties dominated by PLT and HLW from 2010 to 2020 in the middle reaches increased by 19% and 16%, respectively. (2) Precipitation had a positive effect on the variability of the rural hollowing-out distribution results based on the feature importance values, whereas agricultural productivity had a negative effect and exhibited a decreasing trend. In the entire study area, economic non-agriculturalization had a negative impact, but the topographic relief was positively correlated with the rural hollowing-out results of regional areas, and the intensity of its effect showed an increased trend from 2000 to 2020. (3) The evaluation results in the entire study area show that the main functional type and the distance from provincial capitals are the most important factors affecting rural hollowing out. Compared to other factors, the level of non-agricultural economic development has a greater impact on the spatio-temporal differentiation of rural hollowing types in different Yellow River basins. Given the significant effect of non-agricultural economic development in the evaluation of regional rural hollowing-out, it can provide methodological and indicator references for the evaluation of rural hollowing-out in different regions.

1. Introduction

Rural hollowing-out signifies an undesirable evolutionary phase amid the transformation and development of urban–rural areas. It manifests through declining rural populations, expanding village land usage, and an upsurge in abandoned rural settlements [1,2,3]. This phenomenon poses a significant hurdle to sustainable rural geographical systems and is prevalent in the developmental trajectories of numerous countries globally [4,5,6,7]. Presently, China is diligently promoting the rural revitalization strategy, igniting a focused exploration into the diverse forms of rural hollowing-out, its influencing factors, and potential remedial measures within rural geography research [8].
Research on rural hollowing-out in China has yielded fruitful results, providing scientific guidance for understanding the evolutionary process, pattern, and mechanism of rural hollowing-out in China. The existing research mainly focuses on the following aspects: (1) measurement of rural hollowing-out. There are analyses from the perspectives of changes in rural population size and demographic evolution [9,10], as well as assessments from the perspective of land expansion by focusing on the construction of new rural houses without demolishing old ones [11,12]. Other studies have constructed a comprehensive indicator system from the perspectives of population, land, and industry to conduct evaluations [13,14]. However, these studies have the shortcomings of having a single evaluation indicator or confusing the rural hollowing-out itself with its influencing factors. (2) Spatial characteristics of rural hollowing-out. In China, there is a significant correlation between the hollowing of rural populations and rural settlements. Specifically, areas in the east, plains, and urban peripheries have higher levels of hollowing in both rural settlements and populations. Meanwhile, areas in central and western China, hilly and mountainous regions, and villages distant from cities have lower hollowing levels in rural settlements and populations [15,16,17,18,19]. Yang et al. evaluated a comprehensive indicator system and concluded that the most severely affected areas are concentrated on the northern border of China, in traditional agricultural areas in central China, and on the eastern coast of China. Conversely, areas with a low degree of hollowing are concentrated in the backward mountainous regions, such as southern Xinjiang, the Qinghai–Tibetan Plateau, and the southwestern mountainous areas [14]. Existing studies have mainly focused on describing the spatial distribution of rural hollowing-out while lacking an in-depth analysis of the characteristics of the geographical types of rural hollowing-out. (3) Influencing factors and formation mechanism. The formation of hollow villages is a result of various factors, including the natural environment, and socioeconomic elements [20]. Rural depopulation, which involves the non-agricultural transfer of the rural population, is a direct cause of rural hollowing-out [14]. The absence of rural grassroots management and planning, traditional rural ideology, and residents’ behavior are the historical factors and driving forces of rural hollowing-out [21]. The urban-biased development strategy under the long-standing urban–rural dualistic system in China is the root cause of the rural hollowing-out [22,23,24]. For instance, Tan et al. explained how the formation of rural hollowing-out takes place in rapidly urbanizing areas in terms of the level of regional economic development, land use patterns, and changes in the employment structure [25]. Sun et al. pointed out that poor rural infrastructure and insufficient urbanization are the main causes of rural hollowing-out in China [26]. However, some studies suggest that the degree of population hollowing is higher in rural areas with convenient transportation and those adjacent to cities [27,28,29]. From the perspective of individual decision-making, Gao et al. found that rural housing abandonment is mainly influenced by the pull of the urban economy and the deterioration of rural housing conditions [6]. (4) Methodology for the research on factors influencing rural hollowing-out. The combination of quantitative and qualitative methods is typically used to analyze the factors contributing to rural hollowing-out. These include tools like linear regression models (built using SPSS software) [18], geographically weighted regression models (built using ArcGIS software) [30], spatial autocorrelation analysis [31], and Geodetector models [18,27]. In summary, the existing studies focus on the measurement of rural hollowing-out, spatial and temporal differences, and the analysis of the formation mechanism, but the research on the geographical types of rural hollowing-out is still relatively weak. Additionally, the analysis of influencing factors often fails to consider regional differences in rural hollowing-out and the non-linear characteristics of the factor action process [28].
The Yellow River Basin consists of three terraces that have varying natural environments, socioeconomic conditions, regional planning, and policy environments. As a result, the evolution of rural human–land relations and rural hollowing-out varies from place to place, and the pattern and mechanism also differ accordingly. Most current studies on rural hollowing-out in the YRB are based on small-scale empirical analyses, leaving insufficient research on geographical types, spatio-temporal differences, and influencing factors of rural hollowing-out throughout the basin. Given the background of ecological protection, high-quality development of the YRB, and the rural revitalization strategy, it is essential to explore the geographical types of rural hollowing-out and their influencing factors in the process of rural transformation and development of the YRB. Such an exploration will provide scientific decision support for the development of rural revitalization of the YRB in the new era.
This study aims to evaluate the geographical types of rural hollowing-out in the YRB and its spatio-temporal differences and influencing factors using machine learning methods. The main contents include (1) using quantitative methods to divide the geographical types of rural hollowing-out in the YRB, (2) revealing the spatio-temporal differences of the geographical types of rural hollowing-out, and (3) identifying the influencing factors based on the SHAP interpretation package of the XGBoost model and ranking the importance of the features. In short, the combination of qualitative and quantitative methods for dividing the geographical types of rural hollowing-out, and the identification of influencing factors based on the XGBoost model, demonstrate good performance in this study.
The structure of this paper is as follows: The second part introduces the overview of the research area and explains the data sources. The third section describes the research methodology. Part four describes the results of the study. The fifth part is a discussion. The sixth part is the conclusion.

2. Study Area

The Yellow River originates from the northern foothills of the Bayan Har Mountains on the Qinghai–Tibetan Plateau. The upper reaches are dominated by the Tibetan Plateau and the Inner Mongolian Plateau, with a small amount of arable land mainly distributed in the Ningxia Plain and the Hetao Plain. The middle reaches are dominated by the Loess Plateau, with arable land mainly distributed in the Guanzhong Plain and the Fenhe Plain. The lower reaches are characterized by low relief and extensive plains. There are 588 county-level administrative units in the YRB. As the main body of the “urban district” is the non-agricultural population and urban construction land, the research object of this paper is the 416 county administrative units excluding the urban districts (Figure 1).
After the reform and opening up, the rapid development of industrialization and urbanization in the areas along the Yellow River has accelerated the siphoning of production factors from the countryside by cities, especially the rapid agglomeration of the rural population to cities and non-agricultural economic sectors. Data show that the rural population of the YRB decreased by about 45.62 million people from 2000 to 2020 [32,33,34]. Rural revitalization is faced with dilemmas such as insufficient labor force and lack of talent, and the sustainable development of the rural territorial system has encountered serious challenges, which has become a major obstacle restricting the modernization of agriculture in the YRB and the high-quality development of the region.

3. Methodology

3.1. Data

The rural population data are from the fifth (2000), sixth (2010), and seventh (2020) China censuses [32,33,34]. The administrative division vector data were obtained from the Resource and Environmental Science Data Center of the Chinese Academy of Sciences. The topographic relief was derived from China’s 30 m resolution DEM data, the precipitation data were from the China Surface Climatic Data Daily Value Dataset (V3.0) of the National Meteorological Information Center of China (http://data.cma.cn/, accessed on 23 May 2023), and data on the area of rural settlements were derived from Chinese Academy of Sciences 30 m resolution land use remote sensing imagery. Socioeconomic factors include urbanization rate and value added of the three industries, which came from provincial statistical yearbooks. Location and planning factors include distance from provincial capitals, regional planning, and main functional area types. The distance from the provincial capital city was calculated using the Euclidean distance method. The data on regional planning and main functional area type mainly refer to the main functional area planning of each province and Miao’s research results on the YRB [35].

3.2. Methodology for Classifying Geographical Types of Rural Hollowing-Out

Population is the fundamental driving force for rural development, and land is the essential spatial medium for rural construction [36,37]. Therefore, population and residential land use are the core indicators for judging whether a village is hollowed or not. Based on the rate of change of the rural resident population per decade (2000–2010, 2010–2020), the rate of change of the rural population in the YRB was categorized into 4 levels, i.e., −15%~0, −30~−15%, <−30%, >0. Based on the rate of change in rural residential areas per decade (2000–2010, 2010–2020), the rate of change of the rural residential areas in the YRB was categorized into 4 levels, i.e., 0~50%, 50~100%, >100%, <0. Then, the spatial superposition of changes in the rural population and rural residential areas was conducted, and the geographical types of rural hollowing-out in the Yellow River Basin were divided accordingly (Table 1).

3.3. Indicator System of Influencing Factors

The geographical environment plays a fundamental role in the emergence and development of rural areas [38]. Under different conditions of topography, precipitation, and resource endowment, there are significant geographical differences in rural morphology, cultural concepts, and the level of rural economic development. Studies demonstrate that the degree of homogenization of rural land use decreases with increasing average elevation and ground slope [17]. The socioeconomic impact on rural population production and living behavior is the main driving force behind the evolution of rural hollowing-out [14,29,39,40]. With socioeconomic development, agricultural productivity has increased rapidly and there is a surplus of labor in the countryside. At the same time, the accelerated agglomeration of the rural population into urban areas and the non-farming sector of the economy, and the substantial increase in farmers’ wage income, has spawned a surge in the construction of new houses in the countryside. However, the construction of new houses without demolishing old ones has resulted in the emergence of rural land hollowing [2,3,6].
Location and planning are crucial elements that significantly influence regional socioeconomic development. They influence the mobility opportunities and costs for the rural population, land development models, and market participation by allocating production factors such as urban and rural capital, land, and technology in the region [29]. To comprehensively assess the influencing factors, this paper constructs an indicator system from three perspectives: geographical environment, socioeconomic factors, location, and regional planning (Table 2).

3.4. XGBoost Model and SHAP-Based Interpretability

XGBoost is an open-source machine learning project developed by Chen et al. [41] that effectively implements the gradient boosting decision tree algorithm. It makes many improvements and has been widely applied in the field of data mining. Although the XGBoost model can provide the importance of different factors based on existing data, it cannot explain the contribution of single factors, multiple factors, and factors.
SHAP (Shapley Additive explanation) is a “model explanation” package developed in Python, which is a game theory method to explain the output of black box models in machine learning. SHAP can complete single sample feature evaluation and evaluation of the degree of aggregation of different feature samples. To address the shortcomings of XGBoost models, this article introduces SHAP to evaluate the contribution of different factors to the model.
Assuming that the ith sample is xi, the jth feature of the ith sample is xij, the model’s predicted value for this sample is y 1 ^ , and the baseline for the entire model (usually the mean of the target variable across all samples) is ybase, then the SHAP value obeys Equation (1) [42,43]:
y 1 ^ = y base + f ( x i 1 ) + f ( x i 2 ) + ⋯ + f ( x i j )
where f(xij) is the SHAP value of xij. Intuitively, f(xi1) is the contribution value of the first feature in the ith sample to the final predicted value yi. When f(xi1) ≥ 0, it means that the feature enhances the predicted value and has a positive effect; conversely, it means that the feature makes the predicted value lower and has a negative effect.

4. Results

4.1. Characteristics of the Spatio-Temporal Evolution of the Rural Population and Residential Land Use

Compared with 2000–2010, the counties with rural population decline rates of more than 30% in the YRB in 2010–2020 rapidly increased and gradually clustered towards the Hetao Plain, Fenhe Plain, and Guanzhong Plain (Figure 2). The areas with a rural population change rate of −30% to −15% generally experienced a decline and gradually converged towards the upper reaches. The number of counties with a rural population decline rate of less than −15% gradually decreased, mainly in the upper Tibetan Plateau region.
In terms of changes in the area of rural settlements, the number of counties with a growth rate of more than 100% gradually increased and gradually clustered from the middle to the upper regions. The areas with a change rate of 50% to 100% expanded and gathered in the Hetao Plain in the upper reaches and in the Guanzhong Plain in the middle reaches. There were more counties with a rate of change of rural settlement area in the range of 0~50%, and the spatial distribution range was also larger; however, compared with 2000–2010, there was a decrease in 2010–2020, gradually clustering towards the lower and upper reaches.

4.2. Characteristics of Geographical Types of Rural Hollowing-Out

4.2.1. General Characteristics of the Geographical Types of Rural Hollowing-Out

In 2000–2010, compared to other basins, the middle reaches had the highest percentage of HRT and LET at 54% and 10%, respectively (Figure 3). However, the PLT and SDT accounted for the highest share of the upstream, with 17% and 17% of the total, respectively. The highest amount of HLW was found in the lower reaches, accounting for 18% of the total.
In 2010–2020, compared to other basins, the lower reaches had the highest percentage of HRT, accounting for 57% of the total, whereas the middle reaches had the lowest percentage of HRT, accounting for 37% of the total. The LET was low in all basins, accounting for no more than 2% of the total. Compared to the other basins, the middle reaches had the highest percentage of PLT and HLW, accounting for 28% and 31% of the total, respectively.

4.2.2. Spatio-Temporal Patterns of Types of Rural Hollowing-Out

The spatial distribution of the SDT and LET was decentralized and the spatial distribution gradually shrank (Figure 4). The HRT had a wider distribution range, and the spatial clustering pattern gradually manifested itself, i.e., clustered towards the upper reaches of the Yellow River in the provinces of Qinghai and Gansu, as well as the lower reaches of the border area between Henan and Shandong.
The spatial distribution range of the population loss type showed an expansion trend. From 2000–2010, the PLT was mainly distributed in the northern part of Shaanxi Province in the upper reaches; from 2010–2020, it rapidly expanded to the middle and upper reaches of the Yellow River. The spatial distribution characteristics of the human–land–empty waste type showed a decentralized layout from a small area, gradually spreading to the whole basin of the Yellow River and forming a catchment area in the middle reaches of the Yellow River in the Shanxi section and in the upper reaches of the Yellow River in the Inner Mongolia section.

4.3. Influencing Factors of Spatio-Temporal Differences in Geographical Types of Rural Hollowing-Out in the YRB

In terms of the influence of the eigenvalues on the model output (Figure 5), the model prediction results for 2000–2010 and 2010–2020 were both strongly influenced by factors such as the main functional types and the distance from the provincial capitals. Of these, the main functional types had the greatest impact. That is, when the feature value gradually increased, the corresponding Shapley value was greater than 0 and increased with the increase in the feature value; when the feature value decreased, the corresponding Shapley value was less than 0 and decreased with the decrease in the feature value, but the positive impact of this feature factor on the model prediction result was greater than the negative impact.
In terms of the sub-basin, characteristic factors such as the topographic relief, the distance from the provincial capitals, urbanization, and economic non-agriculturalization of the upper and lower reaches had a greater impact on the model prediction results in 2000–2010. Among them, the distance from the provincial capitals had the greatest influence on the model prediction results, i.e., when the distance from the provincial capitals decreased, the corresponding Shapley value was less than 0 and also decreased with the decrease in the eigenvalue; when the distance from the provincial capital increased, the corresponding Shapley value was greater than 0 and increased with the increase in the eigenvalue. However, the negative influence of the upper reaches was greater than the positive influence, and the negative influence on the lower reaches was greater than the positive influence. The reverse was true for the lower Yellow River.
Factors such as economic non-agriculturalization, topographic relief, and distance from provincial capitals had a greater impact on the prediction model results in the middle and lower reaches of the Yellow River in 2000–2010, as well as in 2010–2020. Among them, economic non-agriculturalization had the greatest influence. The Shapley value of this factor was less than 0 when economic non-agriculturalization increased, and it decreased with the increase in the eigenvalue. Conversely, when economic non-agriculturalization decreased, its corresponding Shapley value was greater than 0, and it increased with the decrease in the eigenvalue. However, the positive effect of this factor on the predicted results of the model for the middle reaches of the Yellow River was not as significant as the negative effect.

5. Discussion

5.1. Evolution of Rural Hollowing-Out Types and Transformation Characteristics

The YRB exhibited notable spatio-temporal differences in the geographical types of rural hollowing-out. There was an overarching trend indicating an increase in rural hollowing-out, particularly evidenced by a substantial surge in HLW instances, which escalated from 64 in 2000–2010 to 108 during 2010–2020, which mainly shifted from the SDT, HRT, and PLT (Figure 6).
In terms of the sub-basin, the upper reaches witnessed the most significant adverse shift, marked by a 13% decrease in SDT and primarily transmuting into HLW and HRT; conversely, the most prominent positive change unfolded in HLW, which increased by 10% and predominantly originated from HRT and PLT. In the middle reaches, the most considerable negative shift was registered in HRT, which plummeted by 16% and chiefly transformed into HLW and PLT. Counterbalancing this decline, the most notable positive change surfaced in PLT, which escalated by 19% and primarily evolved from HRT, HLW, and LET. In the lower reaches, the largest negative change materialized in SDT, which declined by 11% and mainly transitioned into HRT, PLT, and HLW. Conversely, the most substantial positive change was observed in HRT, which increased by 9% and primarily evolved from HLW and PLT.

5.2. Assessing the Feature Importance of the Model at Different Spatio-Temporal Scales

In terms of the evaluation of the feature importance of the model at different spatio-temporal scales (Figure 7), the features of distance from provincial capitals and main functional types always had a strong impact on the spatio-temporal differentiation of rural hollowing-out types throughout the basin, whereas the impact of precipitation, agricultural production efficiency, and regional planning was relatively weak. In addition, the influence of topographic relief and economic non-agriculturalization gradually increased, whereas precipitation, urbanization, and agricultural production efficiency had the opposite effect.
From the perspective of different basins, the influence of topographic relief and distance from provincial capitals on the spatio-temporal differentiation of rural hollowing-out types in the upper reaches of the Yellow River was relatively significant, and the degree of influence gradually increased. The impact of economic non-agriculturalization on the spatio-temporal differences in the geographical types of rural hollowing-out in the middle and lower reaches of the Yellow River was relatively large, and the degree of impact gradually increased. However, the influence of precipitation, agricultural production efficiency, and main functional types was relatively weak, and the degree of influence gradually decreased.
This study aligns with Liu’s research [17], which emphasizes topography’s influence on rural hollowing evolution. However, our findings unveil nuanced regional and developmental stage variations in topography’s impact. Notably, although topography positively affected the lower reaches in 2000–2010, this shifted to a negative influence in 2010–2020. Moreover, our research underscores the significant influence of distance from the provincial capital, notably impacting rural hollowing, particularly in the upper reaches. This corroborates findings by Wen, Luo, and Wang [27,28,29] and highlights the pivotal role of location factors in rural hollowing. Contrary to Gao and Tan’s conclusions [6,25], our study reveals spatial variability in the impact of economic non-agriculturalization. It notably affected the middle and lower reaches, whereas the upper reaches were more influenced by topography, geographical location, and other factors.
In summary, when evaluating the spatial heterogeneity factors affecting the rural hollowing-out types in the YRB, factors such as the distance from the provincial capitals, main functional types, topographic relief, and economic non-agriculturalization should be considered. For the upper and lower reaches of the Yellow River, the characteristics of the distance from the provincial capitals, economic non-agriculturalization, and topographic relief should be focused on. For the middle reaches, the characteristics of economic non-agriculturalization and topographic relief should be considered.

5.3. Shortcomings and Future Research

This paper discusses the different types of rural hollowing-out in the YRB, focusing on changes in the size of the rural resident population and the area of rural settlements at the county level. The SHAP method, based on the XGBoost model, identifies the main factors responsible for the spatio-temporal differentiation of rural hollowing-out geographical types in the YRB and its basin differences. This expands the application of machine learning and deepens our understanding of the mechanism behind the spatio-temporal differentiation of rural hollowing-out geographical types.
However, some limitations to this study require further exploration. For example, the assessment only considered the permanent rural resident population without taking into account demographic shifts such as the elderly, children, and the workforce. The classification criteria for rural populations and residential areas could also benefit from refinement, and there is a need for clarity in delineating geographical hollowing types. Additionally, it is essential to examine strategies for rural population development, land rehabilitation, and rural resource integration across different geographical types. Therefore, future research should focus on developing precise classification standards for rural population and residential area changes, articulating distinct geographical hollowing types, assessing challenges in regional agricultural and rural development, and charting future developmental trajectories.

6. Conclusions

Based on Chinese census data and land use remote sensing images with a 30 m resolution, the different types of rural hollowing-out in the counties of the YRB were assessed and analyzed in this study. In addition, the factors that influence the spatio-temporal differentiation of rural hollowing-out and the disparities in various basins were evaluated by using the XGBoost model. The main conclusions are as follows:
(1) The rural hollowing-out types in different basins of YBS were diverse from 2000 to 2020. The prevalence and spatial concentration of the PLT and HLW types consistently increased, tending to cluster in the upper reaches. Specifically, PLT was more pronounced in the upper reaches near the Shaanxi–Gansu border, whereas HLW predominantly concentrated in the Inner Mongolia section of the upper reaches. Conversely, the number of counties with SDT, HRT, and LET types showed a declining trend, with their distribution areas also shrinking.
(2) The upper, middle, and lower reaches of the Yellow River were mainly dominated by the HRT type, whereas the changes in the geographic type of rural hollowing-out varied greatly from basin to basin. From 2000 to 2020, HRT was the type with the highest proportion in each basin, but the middle reaches of the Yellow River showed a substantial decreasing trend, whereas the upper and lower reaches showed an increasing trend. In the middle reaches of YBS, the rural population declined at an accelerated rate, and the proportions of the PLT and HLW types grew rapidly; in the upper and lower reaches of YBS, the number of counties of the HRT type showed a more pronounced increase, whereas the number of counties with the SDT type rapidly decreased. The largest increases in the types of rural hollowing-out in the upper, middle, and lower reaches of the YRB were HRT (9%), PLT (19%), and HLW (9%), respectively, whereas the largest decreases were SDT (12%), HRT (17%), and SDT (11%), respectively.
(3) The importance of the characteristics varied significantly at different spatio-temporal scales. From 2000 to 2020, the factors affecting the results of the spatial and temporal differentiation of the geographic type of rural hollowing in the entire Yellow River Basin were dominated by the distance from the provincial capitals and the topographic relief. From 2000 to 2020, the main factors influencing the spatio-temporal differentiation of rural hollowing types in the middle reaches of the Yellow River Basin were economic non-agriculturalization and the degree of topographic relief. From 2000 to 2020, the main factors affecting the spatio-temporal differentiation of the rural hollowing-out types in the lower Yellow River Basin were the distance to provincial capitals and economic non-agriculturalization. The XGBoost-based feature importance evaluation method and results of this study can provide methodological and indicator references for regional rural hollowing evaluation.

Author Contributions

Z.F.: conceptualization, data curation, investigation, writing—original draft, writing—review and editing. Y.Y.: data curation, supervision, writing—review and editing. L.W.: conceptualization, data curation, investigation, methodology, validation, software. X.Z.: data curation, project administration, supervision. H.L.: conceptualization, project administration, software, supervision. J.Q.: project administration, funding acquisition, supervision, validation. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the Natural Science Foundation of Henan Province, China (grant No. 232300420430), the National Natural Science Foundation of China (grant No. 42071220), and the Natural Youth Science Foundation of Hebei Province, grant number D2022105002.

Institutional Review Board Statement

Not applicable.

Data Availability Statement

Data will be made available on request.

Conflicts of Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References

  1. Liu, Y.S.; Liu, Y.; Zhai, R.X. Geographical research and optimizing practice of rural hollowing-out in China. Acta Geogr. Sin. 2009, 64, 1193–1202. [Google Scholar]
  2. Chen, Y.F.; Sun, H.; Liu, Y.S. Reconstruction models of hollowed villages in key agricultural regions of China. Acta Geogr. Sin. 2010, 65, 727–735. [Google Scholar]
  3. Sun, J.W.; Liu, Y.S.; Ge, D.Z. Coordinated mechanism between comprehensive land consolidation and rural transformation development in plain agricultural areas of China. Acta Geogr. Sin. 2022, 77, 1971–1986. [Google Scholar]
  4. Liu, Y.; Zhou, Y.; Li, Y. Rural regional system and rural revitalization strategy in China. Acta Geogr. Sin. 2019, 74, 2511–2528. [Google Scholar]
  5. Li, Y.; Westlund, H.; Liu, Y. Why some rural areas decline while some others not: An overview of rural evolution in the world. J. Rural Stud. 2019, 68, 135–143. [Google Scholar] [CrossRef]
  6. Gao, X.; Xu, A.; Liu, L.; Deng, O.; Zeng, M.; Ling, J.; Wei, Y. Understanding rural housing abandonment in China’s rapid urbanization. Habitat Int. 2017, 67, 13–21. [Google Scholar] [CrossRef]
  7. Amartuvshin, A.; Chen, J.; John, R.; Zhang, Y.; Lkhagvaa, D. How does mining policy affect rural migration of Mongolia? Land Use Policy 2021, 107, 105474. [Google Scholar] [CrossRef]
  8. Li, Y.; Westlund, H.; Zheng, X.; Liu, Y. Bottom-up initiatives and revival in the face of rural decline: Case studies from China and Sweden. J. Rural Stud. 2016, 47, 506–513. [Google Scholar] [CrossRef]
  9. Zhou, Z. The hollowing out of China’s rural population and its challenges. Popul. Res. 2008, 32, 45–52. [Google Scholar]
  10. Wang, L.; Chen, K.; Li, N. The degree measure and spatial-temporal variation characteristics of rural population hollowing on county-level in China. Popul. J. 2017, 39, 14–24. [Google Scholar]
  11. Yu, L.J.; Sun, D.S.; Zhang, D.X.; Zheng, B.; Han, L.; Yu, Z. Rural household survey based analysis of hollowing of rural residential land in China. Sci. Geogr. Sin. 2016, 36, 1043–1049. [Google Scholar]
  12. Wang, L.; Wu, J. Study on the influencing factors of homestead hollowing in a household’s perspective. Geogr. Res. 2019, 38, 2201–2211. [Google Scholar]
  13. Qu, Y.B.; Zhao, L.J.; Chai, Y.F.; Li, Y.F.; Zhu, W.Y.; Ping, Z.L. Multidimensional form identification and targeted governance of hollow villages from the rural revitalization perspective: Taking Fangsi Town of Yucheng City in Shangdong Province as an example. Resour. Sci. 2021, 43, 776–789. [Google Scholar]
  14. Yang, R.; Liu, Y.S.; Chen, Y.F. Comprehensive measure and partition of rural hollowing-out in China. Geogr. Res. 2012, 31, 1697–1706. [Google Scholar]
  15. Song, W.; Chen, B.M.; Zhang, Y. Typical survey and analysis on influencing factors of village-hollowing of rural housing land in China. Geogr. Res. 2013, 32, 20–28. [Google Scholar]
  16. Li, T.; Long, H.; Wang, Y. Analysis of idleness of rural residential land and its causes in China. China Land Sci. 2019, 33, 64–71. [Google Scholar]
  17. Liu, S.; Xiao, W.; Ye, Y.; He, T.; Luo, H. Rural residential land expansion and its impacts on cultivated land in China between 1990 and 2020. Land Use Policy 2023, 132, 106816. [Google Scholar] [CrossRef]
  18. Zheng, D.; Wen, Q.; Wang, Y.; Mi, H. Differentiation mechanism and reconstruction strategy of rural population hollowing in China. Econ. Geogr. 2019, 39, 161–168. [Google Scholar]
  19. Li, Y.; Wang, H. Spatial distribution of rural population flow at the village level in China: Evidence from village samples in the third national agricultural census. Chin. Rural Econ. 2020, 4, 124–144. [Google Scholar]
  20. Qiao, J.; Liu, J.; Xie, M. Characteristics and Microcosmic Mechanism of Rural Hollow Villages in Less Developed Rural Area: A Case Study of Three Villages in Lankao County. Hum. Geogr. 2011, 26, 98–102. [Google Scholar]
  21. Wang, C.; Yao, S.; Chen, C. Empirical Study on Village-Hollowing in China. Sci. Geogr. Sin. 2005, 25, 3257–3262. [Google Scholar]
  22. Liu, Y. Geography of New Countryside Construction of China; Science Press: Beijing, China, 2011. [Google Scholar]
  23. Liu, Y.S.; Yan, B.; Wang, Y.F. Urban-rural development problems and transformation countermeasures in the new period in China. Econ. Geogr. 2016, 36, 1–8. [Google Scholar]
  24. Li, Y.R.; Liu, Y.S.; Long, H.L. Spatio-temporal analysis of population and residential land change in rural China. J. Nat. Resour. 2010, 25, 1629–1638. [Google Scholar]
  25. Tan, X.L.; Yu, S.Y.; Ouyang, Q.L.; Mao, K.B.; He, Y.H.; Zhou, G.H. Assessment and influencing factors of rural hollowing-out in the rapid urbanization region: A case study of Changsha-Zhuzhou-Xiangtan urban agglomeration. Geogr. Res. 2017, 36, 684–694. [Google Scholar]
  26. Sun, H.; Liu, Y.; Xu, K. Hollow villages and rural restructuring in major rural regions of China: A case study of Yucheng City, Shandong Province. Chin. Geogr. Sci. 2011, 21, 354–363. [Google Scholar] [CrossRef]
  27. Wen, Q.; Li, J.; Ding, J.; Wang, J. Evolutionary process and mechanism of population hollowing out in rural villages in the farming-pastoral ecotone of Northern China: A case study of Yanchi County, Ningxia. Land Use Policy 2023, 125, 106506. [Google Scholar] [CrossRef]
  28. Luo, X.; Tong, Z.; Liu, Y.; Xiong, Q.; Wu, Z.; Chen, H. Revealing the formation mechanism of rural housing hollowing via considering the nonlinear effects of spatial factors and planning variables. Appl. Geogr. 2023, 155, 102957. [Google Scholar] [CrossRef]
  29. Wang, F.; Ma, Y.L.; Qiao, J.J. Scale effect of center city impact on geographic distribution pattern of rural hollowing-out: A case study of Henan province. Areal Res. Dev. 2018, 37, 132–137. [Google Scholar]
  30. Yuan, Y.; Shi, X.; Niu, S.; Zhang, C.; Yan, S. A GWR-based study on Jincheng city hollow village driving force. Econ. Geogr. 2015, 35, 148–155. [Google Scholar]
  31. Chen, K.; Wang, L.; Li, N. Rural hollowing-out out in counties of China: Connotation, distribution and formation mechanism. Popul. Econ. 2018, 1, 28–37. [Google Scholar]
  32. Leadership Office of the Fifth National Population Census of the State Council. China’s Population Census Data by County-2000; Statistics Press: Beijing, China, 2002. [Google Scholar]
  33. Leadership Office of the Sixth National Population Census of the State Council. China’s Population Census Data by County-2010; Statistics Press: Beijing, China, 2012. [Google Scholar]
  34. Leadership Office of the Seventh National Population Census of the State Council. China’s Population Census Data by County-2020; Statistics Press: Beijing, China, 2022. [Google Scholar]
  35. Miao, C.; Zhang, B. Regulation strategy of Zoning-Gradation-Classification for high-quality development in the Yellow River Basin. Econ. Geogr. 2021, 41, 143–153. [Google Scholar]
  36. Long, H. Land use transition and land management. Geogr. Res. 2015, 34, 1607–1618. [Google Scholar]
  37. Long, H.; Chen, K. Urban-rural integrated development and land use transitions: A perspective of land system science. Acta Geogr. Sin. 2021, 76, 295. [Google Scholar]
  38. Zheng, D.Y.; Wen, Q.; Wang, X.J. Spatial differentiation and influencing factors of coupled development of rural poverty and hollowing. Hum. Geogr. 2020, 35, 74–80. [Google Scholar]
  39. Liu, Y. Research on the urban-rural integration and rural revitalization in the new era in China. Acta Geogr. Sin. 2018, 73, 637–650. [Google Scholar]
  40. Ge, D. The characteristics and multi-scale governance of rural space in the new era in China. Acta Geogr. Sin. 2023, 78, 1849–1868. [Google Scholar]
  41. Chen, T.; Guestrin, C. Xgboost: A scalable tree boosting system. In Proceedings of the 22nd ACM Sigkdd International Conference on Knowledge Discovery and Data Mining, San Francisco, CA, USA, 13–17 August 2016; pp. 785–794. [Google Scholar]
  42. Lundberg, S.M.; Lee, S.I. A unified approach to interpreting model predictions. Adv. Neural Inf. Process. Syst. 2017, 30. [Google Scholar] [CrossRef]
  43. Rui, C.; Bin, L.; Min, L.; Ruina, S. Predicting prices and analyzing features of online short-term rentals based on XGBoost. Data Anal. Knowl. Discov. 2021, 5, 51–65. [Google Scholar]
Agriculture 14 00365 g001
Figure 1. Location of the study area.
Figure 1. Location of the study area.
Agriculture 14 00365 g001
Agriculture 14 00365 g002
Figure 2. Spatio-temporal evolution of the rural population and residential land use.
Figure 2. Spatio-temporal evolution of the rural population and residential land use.
Agriculture 14 00365 g002
Agriculture 14 00365 g003
Figure 3. Percentage stacking of geographical types of rural hollowing-out in the YRB. (Up: upper reaches, Mi: middle reaches, Lo: lower reaches, Ye: Yellow River Basin; 1: 2000–2010 year, 2: 2010–2020 year. Same below.).
Figure 3. Percentage stacking of geographical types of rural hollowing-out in the YRB. (Up: upper reaches, Mi: middle reaches, Lo: lower reaches, Ye: Yellow River Basin; 1: 2000–2010 year, 2: 2010–2020 year. Same below.).
Agriculture 14 00365 g003
Agriculture 14 00365 g004
Figure 4. Spatial pattern of geographical types of rural hollowing-out in the YRB.
Figure 4. Spatial pattern of geographical types of rural hollowing-out in the YRB.
Agriculture 14 00365 g004
Agriculture 14 00365 g005aAgriculture 14 00365 g005b
Figure 5. The influence of feature values on model output. (From top to bottom, the Yellow River basin, the upper reaches, the middle reaches, and the lower reaches are indicated. From left to right are 2000–2010 and 2010–2020).
Figure 5. The influence of feature values on model output. (From top to bottom, the Yellow River basin, the upper reaches, the middle reaches, and the lower reaches are indicated. From left to right are 2000–2010 and 2010–2020).
Agriculture 14 00365 g005aAgriculture 14 00365 g005b
Agriculture 14 00365 g006
Figure 6. Evolution of rural hollowing-out types and transformation characteristics ((a–d) indicate the whole Yellow River basin, upper reaches, middle reaches, and lower reaches, respectively). The color corresponding to the letter on the left indicates the corresponding geographic type. The thickness of the line indicates the number of transformations corresponding to the geographic type.
Figure 6. Evolution of rural hollowing-out types and transformation characteristics ((a–d) indicate the whole Yellow River basin, upper reaches, middle reaches, and lower reaches, respectively). The color corresponding to the letter on the left indicates the corresponding geographic type. The thickness of the line indicates the number of transformations corresponding to the geographic type.
Agriculture 14 00365 g006
Agriculture 14 00365 g007
Figure 7. The evaluation of the feature importance of the model at different spatio-temporal scales. (From left to right indicates the whole Yellow River basin, the upper reaches, the middle reaches, and the lower reaches.)
Figure 7. The evaluation of the feature importance of the model at different spatio-temporal scales. (From left to right indicates the whole Yellow River basin, the upper reaches, the middle reaches, and the lower reaches.)
Agriculture 14 00365 g007
Table 1. Methodology for categorizing rural hollowing types.
Table 1. Methodology for categorizing rural hollowing types.
Rural Population Change RateChange Rate of Rural Residential AreasGeographical Types
>00~50%Smart development type (SDT)
<0<0
−30%~0<100%Human–land recession type (HRT)
−30%~050~100%
−15%~0 or <−30%>100%Land expansion type (LET)
<−30%0~50%Population loss type (PLT)
−30~−15%>100%Human–land–vacant waste type (HLW)
<−30%>50%
Table 2. Influence index system of spatio-temporal differentiation of rural hollowing-out geographical types in the Yellow River Basin.
Table 2. Influence index system of spatio-temporal differentiation of rural hollowing-out geographical types in the Yellow River Basin.
VariableIndicatorMeaning of IndicatorData Sources
Dependent variableRural hollow area type (Rht)SDT:1, MLR:2, LET:3, PLT:4, HLW:5—
Geographical environmentTopographic relief (Tor)Topographic relief30 m resolution DEM data
Precipitation variation (Prc)Change in average annual precipitationhttp://data.cma.cn/
SocioeconomicUrbanization (Urb)Urban population/total populationChinese census data
Changes in agricultural productivity (Acp)The difference in agricultural value added/rural population between different yearsProvincial statistical yearbooks
Economic non-agriculturalization (Nfe)The difference in non-farm GDP per capita between different yearsProvincial statistical yearbooks
Location and planningLocation (Trl)Distance from provincial capitalsEuclidean distance
Main function types (Mfo)Main agricultural product producing areas = 1, urbanized area = 2, key ecological function area = 3Provincial main functional area planning
Regional planning (Rep)Whether it belongs to urban agglomeration or not: yes: 1, no: 2Miao, C., Zhang, B. [35] (Miao and Zhang, 2021).
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Fu, Z.; Yang, Y.; Wang, L.; Zhu, X.; Lv, H.; Qiao, J. Geographical Types and Driving Mechanisms of Rural Hollowing-Out in the Yellow River Basin. Agriculture 2024, 14, 365. https://doi.org/10.3390/agriculture14030365

AMA Style

Fu Z, Yang Y, Wang L, Zhu X, Lv H, Qiao J. Geographical Types and Driving Mechanisms of Rural Hollowing-Out in the Yellow River Basin. Agriculture. 2024; 14(3):365. https://doi.org/10.3390/agriculture14030365

Chicago/Turabian Style

Fu, Zhanhui, Yahan Yang, Lijun Wang, Xiaoyong Zhu, Hui Lv, and Jiajun Qiao. 2024. "Geographical Types and Driving Mechanisms of Rural Hollowing-Out in the Yellow River Basin" Agriculture 14, no. 3: 365. https://doi.org/10.3390/agriculture14030365

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop