The Impact of Rural Population Decline on the Economic Efficiency of Agricultural Carbon Emissions: A Case Study of the Contiguous Karst Areas in Yunnan–Guizhou Provinces, China

Abstract

Amid global climate warming, agricultural low-carbon transition is critical for ecological governance. In China’s ecologically fragile contiguous karst areas of Yunnan–Guizhou, intensifying rural population decline poses unique challenges to emission reduction. This study analyzes population and agricultural production data from 25 cities (prefectures) (2013–2022) to quantify rural population decline rates and agricultural carbon emission efficiency. We map their spatiotemporal evolution patterns, apply spatial autocorrelation models to assess spatial dependencies, and investigate mechanisms through a mediation model integrated with agricultural modernization’s three core systems: industrial, production, and management. Key findings reveal (1) divergent trajectories of carbon emission efficiency across regions with varying population decline types; (2) a global Moran’s I of −0.3519, indicating significant negative spatial correlation between population decline intensity and emission efficiency; and (3) dual impact mechanisms where population decline directly alters emission efficiency and indirectly modulates it through interactions with agricultural systems, with mechanism heterogeneity across decline patterns. To reconcile carbon reduction and agricultural growth, region-specific strategies must align population decline gradients with dynamic adjustments to agricultural systems, ensuring synchronized demographic transition and modernization.

1. Introduction

In recent years, the combined impacts of human activities and natural processes have disrupted ecological balance through greenhouse gas and pollutant emissions. Agriculture, as a critical component in advancing global green development, exerts significant influence on climate change and ecological systems [1,2,3,4]. Current investigations reveal that agricultural carbon emissions account for 20–25% of global emissions. Notably, karst regions cover approximately 15% of the Earth’s land area and represent one of the planet’s most fragile ecosystems. Globally, agricultural development in karst areas faces dual challenges: ecological vulnerability and reliance on labor-intensive practices, compounded by the accelerating transformation of agricultural production systems due to the worldwide trend of rural population shrinkage. Compared to karst regions such as the Mediterranean and Southeast Asia, China’s contiguous Yunnan–Guizhou karst zone exhibits distinct characteristics. Driven by urbanization and ecological conservation policies, the region has experienced large-scale and sustained rural population loss. Traditional agricultural production, predominantly crop cultivation and livestock breeding, demands greater ecological resource consumption, resulting in disproportionately high agricultural carbon emissions. As rural depopulation intensifies—marked by outmigration, labor relocation, and declining birth rates—delays in agricultural modernization have led to extensive farming practices and inefficient resource utilization, profoundly exacerbating carbon emissions [5]. Under China’s dual-carbon goals, karst regions—where agriculture remains the primary livelihood—face heightened decarbonization responsibilities amid severe population contraction. The carbon reduction challenges in China’s karst agricultural systems now represent a global focus and bottleneck for rural emission mitigation, necessitating the urgent development of context-specific low-carbon pathways tailored to the interplay between population dynamics, ecological constraints, and agricultural transformation in fragile ecosystems [6,7].

Current research on agricultural carbon emissions primarily focuses on total emissions, intensity, efficiency, and performance [8,9,10]. Among these, agricultural carbon emission efficiency, which identifies factors influencing the development and changes in agricultural carbon emissions, is of practical significance [11,12]. Existing studies on agricultural carbon emission efficiency mainly concentrate on efficiency measurement, driving factor identification, and external environment exploration [13,14,15]. Research topics include the impact mechanisms of factors such as urbanization levels, natural disasters, and education levels on agricultural carbon emission efficiency, as well as the relationship between external environments like agricultural technological progress, food security, and industrial agglomeration with agricultural carbon emission efficiency [16,17,18]. Research scales range from provincial to regional and county levels, with existing achievements providing a solid reference for related carbon emission efficiency studies [19,20,21]. However, in terms of agricultural carbon emission efficiency measurement, most current studies construct models based on agricultural inputs and carbon emission outputs, with limited analysis of the linear relationship between agricultural carbon emission efficiency and agricultural economic efficiency. There is also a lack of attention to the synergy between regional agricultural carbon emissions and agricultural economic growth efficiency. For resource-dependent regions, improving the economic efficiency of agricultural carbon emissions plays a crucial role in achieving the dual goals of agricultural carbon reduction and economic growth.

Scholarly discourse on population shrinkage is mostly conducted from the perspectives of regional total population and urban population [22,23]. Existing studies have confirmed that population decline significantly impacts carbon emission issues, primarily in industrialized developed regions [24]. However, as the most fundamental and dynamic element in rural areas, the rural population deserves extensive attention. Furthermore, rural depopulation (a defining characteristic of karst mountain socio-ecosystems in the Yunnan–Guizhou Plateau) fundamentally restructures the trajectories of agrarian modernization [25,26,27]. It can directly or indirectly affect the economic efficiency of agricultural carbon emissions. However, there is limited academic research on the relationship between rural population decline and the economic efficiency of agricultural carbon emissions, and the role of rural population decline in this context remains unexplored [28,29,30]. Therefore, advancing agricultural modernization, it is essential to investigate the dynamic and differentiated impact mechanisms of rural population decline on the economic efficiency of agricultural carbon emissions in karst regions.

Based on this, this study focuses on karst regions to analyze the spatiotemporal evolution of agricultural carbon emission economic efficiency against the backdrop of rural population shrinkage. Integrating the “Three Major Systems” of agricultural modernization (industrial, production, and operational systems), we systematically examine how varying types of population contraction differentially impact carbon efficiency. Through empirical investigation of China’s contiguous Yunnan–Guizhou karst zone, we reveal the transmission mechanisms by which rural depopulation influences carbon efficiency through restructuring agricultural industrial frameworks, transforming production technologies, and reshaping operational models. This research contributes dual innovations: proposing targeted carbon reduction strategies tailored to distinct population shrinkage patterns in karst areas and establishing a theoretical framework elucidating how labor structure transitions drive low-carbon agricultural transformation in fragile ecosystems. Our findings bridge critical knowledge gaps in understanding the coupled dynamics between demographic shifts and emission efficiency optimization, offering both methodological and policy references for sustainable development in global karst regions.

The remaining sections of this paper are structured as follows: Section 2 elaborates the hypotheses and theoretical framework; Section 3 introduces the study area overview, data sources, and research methodology; Section 4 presents the empirical results and their detailed interpretation; Section 5 discusses the indirect effects of rural population decline on the economic efficiency of agricultural carbon emissions; and Section 6 concludes the study and proposes policy recommendations.

2. Hypotheses and Theoretical Framework

With the increasing loss of rural populations and the deepening phenomenon of rural hollowing, rural population decline has garnered widespread attention. While there is no unified academic definition of rural population decline, it is commonly operationalized as a reduction in population size, typically calculated using the registered population as a key variable to determine the population decline rate. However, this approach inadequately accounts for the temporal dynamics of population decline and overlooks the rural depopulation caused by labor force contraction—particularly in karst agricultural regions where farming remains the dominant mode of production [31]. As labor constitutes the most dynamic factor in agricultural production, the depopulation induced by workforce contraction holds significant implications [32]. Therefore, from the perspective of population reduction, rural population decline can be defined as a phenomenon within a specific region where both the registered population and the labor force population exhibit continuous decreases over a defined period.

Existing research primarily calculates agricultural carbon emission efficiency by constructing input–output functions, significantly enhancing the practical relevance of the conclusions. However, this approach disregards the fundamental orientation of agricultural activities toward economic growth, while failing to account for the direct linear correlation between agricultural economic growth and carbon emissions. By incorporating the economic efficiency framework, agricultural carbon emission efficiency emerges as a specialized application of economic efficiency theory within the domain of agricultural emissions. It measures the dependency of regional agricultural economic growth on agricultural carbon emissions by calculating the ratio of agricultural economic growth efficiency to agricultural carbon emission efficiency [33,34]. A higher ratio indicates a lower dependency of agricultural economic growth on carbon emissions and higher environmental benefits of agricultural production. For regions where agriculture is the economic pillar, compared to indicators such as total carbon emissions and carbon emission intensity, setting emission reduction targets based on carbon emission economic efficiency better aligns with the actual development needs of the region [35].

The agricultural industrial structure in karst regions is singular and the economic resilience of agriculture is low [36,37]. With the rapid advancement of agricultural and rural modernization, under the backdrop of rural population decline, the economic efficiency of agricultural carbon emissions in karst regions is influenced by numerous uncertain factors and complex processes. Based on related research and focusing on the “Three Major Systems” of agricultural modernization development, the impact of rural population decline on carbon emission economic efficiency can be divided into direct and indirect effects [38,39]. Population decline can directly affect carbon emission economic efficiency, while the changes in the agricultural industrial system, agricultural production system, and agricultural management system caused by rural population decline can also influence the evolution of carbon emission economic efficiency (Figure 1):

H1: 

Hypotheses on Rural Population Decline. With the continuous outflow of rural population, declining birth rates, and severe aging, the reduction in rural population leads to insufficient effective agricultural labor. Under the constraints of resource and environmental carrying capacity in karst regions, to ensure regional food supply and basic living needs, extensive agricultural production increases carbon emissions. In the context of rural depopulation, the marginal productivity of agricultural inputs diminishes, labor costs escalate, and agricultural economic growth encounters mounting constraints, collectively depressing carbon economic emission efficiency [40,41].

H2: 

Hypotheses on the Modern Agricultural Industrial System. Rural population decline leads to an aging rural population, inhibiting the optimization of the agricultural industrial structure and preventing the expansion of agriculture into leisure, cultural, and ecological fields. This dynamic accelerates the growth of non-agricultural sectors relative to agriculture, diminishing agriculture’s share of value-added in the economy while increasing its resource dependency, consequently reducing carbon emission economic efficiency [42,43]. On the other hand, it hinders the extension of the agricultural industrial chain, limiting traditional agriculture from extending to processing, cultural tourism, marketing, logistics, and other sectors, diminishing the value-added potential of agricultural outputs. As a result, agricultural economic growth becomes increasingly dependent on the environment, leading to lower carbon emission economic efficiency.

H3: 

Hypotheses on the Modern Agricultural Production System. Rural population decline, accompanied by rural hollowing, forces the transformation and upgrading of agricultural production technologies and accelerates the penetration of agricultural technology. Empirical studies substantiate that agricultural technologies exert dualistic impacts on carbon emission economic efficiency, manifesting as both synergistic enhancement and counterproductive effects. The positive impact is reflected in the optimization of resource utilization efficiency, which improves carbon emission economic efficiency. The negative impact is that improved resource utilization efficiency stimulates agricultural economic activities, increasing resource consumption and offsetting the resource-saving effects of efficiency improvements. Additionally, the penetration of agricultural technology has a time lag. On the other hand, rural hollowing reduces dependency on human labor in agricultural production, leading to changes in the labor market structure, reducing agricultural employment, and increasing non-agricultural employment [44,45,46]. However, the expansion of non-agricultural employment inevitably reshapes farmers’ carbon-intensive practices—including fertilizer and pesticide application patterns—ultimately degrading carbon emission economic efficiency through behavioral adaptation pathways.

H4: 

Hypotheses on the Modern Agricultural Management System. Rural population decline leads to labor loss, resulting in a lack of agricultural management entities, extensive land management, and increased difficulty in land consolidation, making it challenging to achieve large-scale land management. This phenomenon exacerbates arable land fragmentation, consequently elevating agricultural production complexity and resource input intensity. Conversely, substantial labor force contraction—accelerated by policies promoting non-grain and non-agricultural land use—drives land transfer to commercial farming entities, fostering scaled and intensive cultivation. These dual dynamics exhibit countervailing effects on carbon emission economic efficiency: the former degrades it through operational inefficiencies, while the latter enhances it via mechanization synergies. This can enhance land productivity and output efficiency to some extent, increasing agricultural carbon emission economic efficiency [47,48]. However, improved land use efficiency may lead to excessive resource consumption, causing a decline in agricultural carbon emission economic efficiency. The Environmental Kuznets Curve (EKC) postulates an inverted U-shaped relationship between environmental degradation and per capita income [49,50]. Within the agricultural sector, this implies that carbon emission economic efficiency exhibits a positive correlation with economic growth during early-stage development, consistent with the expansion phase of the EKC framework. When agriculture reaches a certain level of development, the relationship becomes negative.

3. Materials and Methods

3.1. Study Area

The Yunnan–Guizhou karst contiguous area, located between 21°8′ N–29°15′ N and 97°31′ E–109°35′ E, spans 570,267 km² across 25 prefectures in the Yunnan and Guizhou provinces. With an average elevation of approximately 1300 m, its subtropical climate—characterized by high temperatures and abundant rainfall—has shaped diverse landforms, including peak clusters, pinnacle karst, isolated hills, waterfalls, natural bridges, and canyons. Mountains and hills occupy 92.60% of the fragmented and ecologically fragile terrain. Agricultural production primarily focuses on drought-resistant crops like corn and potatoes, supplemented by specialty cash crops such as tea and flue-cured tobacco, while livestock farming revolves around mountain black goats, yellow cattle, and pigs, forming a composite “rock crevice cultivation + terrace farming” model. Agriculture remains pivotal to the regional economy, contributing 13.00% to the local GDP in 2022—5.7 percentage points above the national average. However, challenges including fragmented terrain, scarce and scattered arable land, thin soil layers, and alternating waterlogging-drought conditions have led to low agricultural efficiency, triggering population outflow and labor migration. From 2013 to 2022, rural population shrinkage reached 9.72%, forcing increased reliance on complementary resource inputs that escalate carbon emissions and hinder high-quality agricultural development. This dynamic has trapped the region in a vicious cycle: low agricultural carbon emission economic efficiency → rural population contraction → heightened agricultural inputs → further decline in carbon efficiency.

Simultaneously, the contiguous karst terrain spanning the Yunnan–Guizhou provinces, situated within China’s second topographic tier, functions as a critical ecological buffer zone for the Yangtze and Pearl Rivers’ mid-lower basins. Therefore, exploring the impact mechanism of rural population decline on the economic efficiency of agricultural carbon emissions in this region is of great strategic significance for promoting the organic integration of agricultural green transformation and high-quality economic development, narrowing regional disparities, and ensuring the sustained prosperity of the middle and lower reaches of the Yangtze and Pearl Rivers. Based on this, this paper selects the 25 cities (prefectures) in the contiguous karst area of Yunnan–Guizhou as the study area for agricultural carbon emission economic efficiency under the background of rural population decline, which is both representative and of practical significance.

3.2. Data Sources

The research data include geographic information data and socio-economic data. Socioeconomic data, including chemical fertilizers, pesticides, crop yields, livestock and poultry inventories, rural population, and agricultural economic output value, were obtained from publicly available sources: the Yunnan Statistical Yearbook https://stats.yn.gov.cn/List22.aspx, (accessed on 12 June 2024) Guizhou Statistical Yearbook https://hgk.guizhou.gov.cn/publish/tj/index.html (accessed on 20 June 2024), and the Statistical Communiqués on National Economic and Social Development of Guizhou and Yunnan provinces for the period 2013–2022. Data on cultivated land area and the number of cultivated land plots are obtained from the 30 m-resolution national land use database (CLCD) provided by Wuhan University. Missing data were supplemented using interpolation methods. The data preprocessing procedure was as follows: First, linear interpolation was employed to impute missing data. Second, to mitigate the impact of variable scale disparities, mediating variables including agricultural industrial structure, extension of agricultural industrial chains, and other relevant indicators were standardized through natural logarithm transformation.

3.3. Methods

3.3.1. Rural Population Decline Model

Traditional population change models only use permanent or registered population as variable indicators to calculate population change rates, neglecting the importance of changes in labor force size and deviating from the principle that “labor is the primary dynamic factor in agricultural production”. Therefore, based on relevant research, this paper adopts a dual-variable indicator to measure rural population decline [51]. The rural registered population and labor force size of 25 cities (prefectures) in the contiguous karst area of Yunnan–Guizhou are selected as variables to calculate the population decline rate of each city (prefecture) from 2013 to 2022. The calculation formula is as follows:

$$R_t={\mathrm{y}}_2-{\mathrm{y}}_1=\left(\sqrt[m]{{\displaystyle \frac{M_{p2}}{M_{p1}}}}-1\right)-\left(\sqrt[m]{{\displaystyle \frac{M_{l2}}{M_{l1}}}}-1\right)$$

(1)

In the formula, R represents the rural population decline rate during the study period; R ≥ 0 indicates no rural population decline, R < 0 indicates rural population decline, and the larger the absolute value of R, the higher the degree of rural population decline in the region. y₁ and y₂ represent the rural population change rates in the initial and final years, respectively; M_p1 and M_p2 represent the rural registered population in the initial and final years, respectively; M_l1 and M_l2 represent the labor force size in the initial and final years, respectively; and m represents the number of years between y₁ and y₂.

To systematically examine the impact mechanisms of varying rural population decline levels on agricultural carbon emission economic efficiency in the study area, we analyzed data from 2013 (baseline year) to 2022 (terminal year), calculating decade-long rural population decline rates across all prefecture-level divisions. Based on the Natural Breaks (Jenks) method in ArcGIS 10.2, we categorized these rates into three distinct types (Figure 2): severe decline (−6.37% ≤ R < −4.67%), moderate decline (−4.67% ≤ R < −2.89%), and mild decline (−2.89% ≤ R < −0.76%).

In terms of quantity, all 25 cities (prefectures) in the contiguous karst area of Yunnan–Guizhou exhibit rural population decline, including 6 cities (prefectures) with severe rural population decline, 11 cities (prefectures) with moderate rural population decline, and 8 cities (prefectures) with mild rural population decline. Spatially, severe-decline cities (prefectures) formed a V-shaped distribution pattern, radiating northwestward and northeastward from core areas including Kunming, Yuxi, and Chuxiong Yi Autonomous Prefecture. Moderate-decline cities (prefectures) clustered in two zones—western units adjacent to severe-decline areas and eastern concentrations centered on Zunyi, Guiyang, and Qiandongnan Miao-Dong Autonomous Prefecture. Mild-decline cities (prefectures) were primarily located in the northern (Diqing Tibetan Autonomous Prefecture) and southern (Xishuangbanna Dai Autonomous Prefecture) extremities of the western region, with additional distributions flanking moderate-decline zones in the east.

3.3.2. Agricultural Carbon Emission Economic Efficiency Model

Agricultural carbon emissions are diverse and complex, making it difficult to establish a unified measurement standard [52,53]. Based on the IPCC National Greenhouse Gas Inventory Guidelines and relevant research findings, and taking into account the current status of regional agricultural development, this study selects carbon emission sources from agricultural material inputs (fertilizers, pesticides, and agricultural plastic film), crop cultivation (rice, oil crops, vegetables, and edible fungi), and livestock breeding (pigs, cattle, sheep, and poultry). By referencing related research, an agricultural carbon emission accounting model is constructed for the Yunnan–Guizhou karst contiguous region [54,55,56]. The calculation formula is as follows:

$$C_t={\displaystyle \sum C_{\mathrm{kt}}}={\displaystyle \sum T_{k\mathrm{t}}}\times {\alpha }_k\times {\beta }_k$$

(2)

where C_t represents the total agricultural carbon emissions in year t; C_kt represents the carbon emissions from the carbon source k in year t; T_kt represents the quantity of the carbon source k in year t; α_k represents the carbon emission coefficient of the carbon source k; and β_k represents the conversion coefficient of the carbon source k.

To further measure the economic efficiency of agricultural carbon emissions in the contiguous karst area of Yunnan–Guizhou, the agricultural carbon emission economic efficiency model is constructed based on Equation (2), as follows:

$$F={\displaystyle \frac{G_{\mathrm{i}}}{G}}/{\displaystyle \frac{C_i}{C}}$$

(3)

where F represents the agricultural carbon emission economic efficiency index; G_i represents the agricultural economic added value of each city (prefecture); G represents the total agricultural economic added value of the contiguous karst area of Yunnan–Guizhou; C_i represents the agricultural carbon emissions of each city (prefecture); and C represents the total agricultural carbon emissions of the contiguous karst area of Yunnan–Guizhou. Referring to relevant studies, we selected 10 types of carbon sources and their emission and conversion coefficients, which are shown in

Table 1. Table of carbon emission factors for various carbon source factors.

Carbon Source Factor		Carbon Emission Type	Carbon Emission Coefficient (αk)	Conversion Coefficient (βk)
Agricultural material inputs	Fertilizers	CO2	0.8956 kg/kg	1
	Pesticides	CO2	4.9341 kg/kg	1
	Agricultural films	CO2	5.180 kg/kg	1
Crop cultivation	Rice	CH4	0.93 kg/kg	25
	Oil crops	NO2	0.93 kg/kg	25
	Vegetables and edible fungi	NO2	0.93 kg/kg	25
Livestock breeding	Pigs	CH4	5 kg·year−1·head−1	25
		NO2	0.53·year−1·head−1	298
	Cattle	CH4	56.39 kg·year−1·head−1	25
		NO2	1.28 kg·year−1·head−1	298
	Sheep	CH4	5.17 kg·year−1·head−1	25
		NO2	0.33 kg·year−1·head−1	298
	Poultry	CH4	0.02 kg·year−1·head−1	25
		NO2	0.02 kg·year−1·head−1	298

[47]:

3.3.3. Spatial Autocorrelation Model of Rural Population Decline and Agricultural Carbon Emission Economic Efficiency

The spatial correlation and significance of interaction between rural population decline and the economic efficiency of agricultural carbon emissions were analyzed using the Bivariate Global Spatial Autocorrelation (Bivariate Moran’s I) method in ArcGIS 10.2 [57,58]; A Moran’s I > 0 indicates positive spatial correlation and a Moran’s I < 0 indicates negative spatial correlation. The formula for Moran’s I is as follows:

$$I={\displaystyle \frac{n{\displaystyle \sum _{i=1}^n{\displaystyle \sum _{j=1}^n{{\displaystyle W}}_{ij}\left({{\displaystyle X}}_i-\stackrel{—}{{\displaystyle X}}\right)\left({{\displaystyle Y}}_j-\stackrel{—}{{\displaystyle Y}}\right)}}}{S^2{\displaystyle \sum _{i=1}^n{\displaystyle \sum _{j=1}^n{{\displaystyle W}}_{ij}}}}}$$

(4)

where I represent the overall correlation between the spatial distribution of rural population decline and agricultural carbon emission economic efficiency; X_i and Y_j represent the values of rural population decline rate and agricultural carbon emission economic efficiency in different cities (prefectures) (i and j); respectively; n is the number of cities (prefectures); and S² is the sample variance; W_ij is the spatial weight matrix.

The formula for bivariate local spatial autocorrelation (Bivariate Local Moran’s I) is as follows:

$$I_i=Z_i{\displaystyle \sum _{j=1}^n{w}_{ij}{z}_j}$$

(5)

where I_i represents the local correlation between the rural population decline rate and agricultural carbon emission economic efficiency in the city (prefecture) i; Z_i and Z_j are the standardized values based on the sample variance of different cities (prefectures) (i and j); W_ij is the spatial weight matrix; and n is the number of cities (prefectures).

3.3.4. Mediation Effect Model of Rural Population Decline and Agricultural Carbon Emission Economic Efficiency

Using Stata 18, we applied a mediation effect model to analyze the pathways and mechanisms through which rural population decline impacts the economic efficiency of agricultural carbon emissions. Based on theoretical foundations and aligned with the “Three Major Systems” of agricultural modernization—the agricultural industrial system, agricultural production system, and agricultural management system—we selected the following mediator variables by synthesizing insights from existing studies and prioritizing data availability: agricultural industrial structure, agricultural value chain extension, agricultural technology penetration, non-agricultural employment structure, cultivated land fragmentation, cultivated land intensification, and agricultural economic development [59,60] (see

Table 2. Index system of mediating effect variables.

Variable Type	Indicator Name	Indicator Variable	Variable Explanation	Unit
Dependent Variable	Agricultural carbon emission economic efficiency	EAC	Agricultural carbon emission economic efficiency index	%
Core Explanatory Variable	Rural population decline	RPS	Rural population decline rate during the study period	%
Mediating Variables	Agricultural industrial structure	ISL	Agricultural production value/total production value	%
	Agricultural industrial chain extension	EAL	Value of agricultural, forestry, animal husbandry, and fishery services/total value of agriculture, forestry, animal husbandry, and fishery	%
	Agricultural technology penetration	ATD	Total agricultural machinery power/cultivated land area	10,000 kWh/km2
	Non-agricultural employment structure	NFE	Non-agricultural employment/total employment	%
	Farmland fragmentation	DLF	Number of cultivated land plots/cultivated land area	10,000 plots/km2
	Farmland intensification	IIL	Agricultural production value/cultivated land area	10,000 yuan/km2
	Agricultural economic development	AEG	Agricultural production value/rural population	10,000 yuan/10,000 people

). The model is constructed based on the three-step mediation effect procedure proposed by Wen Zhonglin [61], with control variables excluded. The calculation formulas are as follows:

$$F=cR+{\epsilon }_1$$

(6)

$$M=aR+{\epsilon }_2$$

(7)

$$F=dR+bM+{\epsilon }_3$$

(8)

where F represents the agricultural carbon emission economic efficiency of each city (prefecture); R represents the rural population decline rate; M represents the mediating variable; c and d represent the total effect and direct effect, respectively; a × b represents the indirect effect; and ε₁, ε₂, and ε₃ represent the error terms.

4. Results and Interpretation

4.1. Temporal Evolution Characteristics of Agricultural Carbon Emission Economic Efficiency

The boxplot analysis of agricultural carbon emission economic efficiency across 25 cities (prefectures) in the Yunnan–Guizhou karst contiguous region (2013–2022, Figure 3) reveals three distinct patterns: (1) the regional efficiency demonstrates nonlinear fluctuations, initially increasing, then decreasing, before rebounding; (2) the gap between mean and median values follows similar cyclical variations, first widening, then narrowing, before expanding again; and (3) interregional disparities among cities (prefectures) exhibit comparable dynamics, initially intensifying, then subsequently reducing, before ultimately widening again during the study period.

The kernel density estimation (Figure 4, 2013–2022) reveals distinct spatial–temporal patterns of agricultural carbon emission efficiency in the Yunnan–Guizhou karst region: the density curve’s peak initially shifts leftward then rightward with decreasing height and widening breadth, indicating efficiency first declined then improved while spatial disparities dynamically expanded. In regions experiencing severe population decline, the density peak exhibits a characteristic left-then-right migration pattern accompanied by peak width contraction and amplitude reduction. Concurrently, multiple secondary peaks emerge to the right of the 2018 primary peak, indicating localized multi-tier demographic differentiation. Moderate population decline regions show rightward peak migration with an initial height increase followed by decrease and width expansion, accompanied by secondary peaks signaling shifting internal disparities that first contracted and then expanded. Severe population decline areas exhibit left-then-right peak movement with abrupt height reduction and significant width expansion, reflecting overall efficiency decline followed by marked recovery alongside intensifying interregional disparities and growing polarization trends.

4.2. Spatial Differentiation Characteristics of Agricultural Carbon Emission Efficiency

To clarify the spatial differentiation and evolution process of agricultural carbon emission economic efficiency in the contiguous karst area of Yunnan–Guizhou, three time periods (2013, 2018, and 2022) were selected. Using the natural breaks method, the agricultural carbon emission economic efficiency in the region was divided into four levels (Figure 5). The agricultural carbon emission economic efficiency in the contiguous karst area of Yunnan–Guizhou exhibits a spatial characteristic of “high in the northeast and low in the southwest”. Northeastern cities (prefectures) generally exhibit agricultural carbon emission economic efficiency values exceeding 1.765, reflecting higher agricultural economic output relative to carbon emissions. Conversely, southwestern cities (prefectures) demonstrate inverse efficiency patterns, with most falling below this threshold.

Specifically, the high-value and low-value areas are divided by the Wumeng Mountains. The high-value areas are mainly concentrated in Guizhou Province on the right side of the Wumeng Mountains, including Bijie, Zunyi, Tongren, Guiyang, Anshun, Liupanshui, Qiandongnan, Qianxinan, and Qiannan. The low-value areas are mainly concentrated in Yunnan Province on the left side of the Wumeng Mountains, including Zhaotong, Qujing, Wenshan, Pu’er, Diqing, Nujiang, Lijiang, and Dali. This indicates that, in the contiguous karst area of Yunnan–Guizhou, cities (prefectures) with high agricultural carbon emission economic efficiency are often those with economies highly dependent on agriculture and facing significant resource constraints on comprehensive agricultural production capacity.

Notably, the division between high-value and low-value areas along the Wumeng Mountains presents a stable and clear spatial pattern of “high in the northeast and low in the southwest”. This spatial pattern arises because Yunnan Province, situated southwest of the Wung Mountains, possesses an extensive land area where karst landscapes constitute 28.9% of its total territory, coupled with comparatively low population density. In contrast, Guizhou Province, situated northeast of the Wumeng Mountains, possesses merely half the land area of Yunnan, yet karst landscapes account for 61.9% of its territory, coupled with a higher population density. The greater population density leads to population agglomeration, which helps promote the optimization of agricultural production structures and improves agricultural production efficiency, thereby reducing the dependence of agricultural economic development on resources. As a result, the agricultural carbon emission economic efficiency is higher in the northeastern Guizhou Province.

By classifying different types of rural population decline and combining standard deviation ellipses, we further analyze the spatial differentiation characteristics of agricultural carbon emission economic efficiency in the contiguous karst area of Yunnan–Guizhou. From an overall perspective (Figure 6a), from 2013 to 2022, the standard deviation ellipses are oriented in a “northeast–southwest” direction, indicating that agricultural carbon emission economic efficiency grew faster in the northeast region from 2013 to 2018, while it grew faster in the southwest region from 2018 to 2022. In terms of the area of the standard deviation ellipses, the area in 2018 decreased by 9.32% compared to 2013, while the area in 2022 increased by 4.14% compared to 2018, suggesting a trend of spatial agglomeration followed by diffusion in agricultural carbon emission economic efficiency. Regarding the rotation angle of the standard deviation ellipses, the angles in 2013, 2018, and 2022 first increased and then decreased, expanding from 63.62° to 65.62° and then shrinking to 63.93°, indicating a “strengthening–weakening” change in the east–west spatial distribution. In terms of the lengths of the major and minor axes, both the major and minor axes in 2013, 2018, and 2022 were first shortened and then lengthened. Compared to 2013, the major and minor axes in 2018 decreased by 1.89% and 7.58%, respectively, while compared to 2018, the major and minor axes in 2022 increased by 0.28% and 3.85%, respectively. This indicates that the spatial distribution of agricultural carbon emission economic efficiency across prefecture-level cities in the Yunnan–Guizhou karst region exhibited a ‘contraction–expansion’ pattern, while the dispersion degree demonstrated a ‘decline–rise’ trend.

From a grouping perspective, the cities (prefectures) with mild rural population decline (Figure 6b) and those with moderate rural population decline (Figure 6c) exhibit changes in their standard deviation ellipses (SDEs) from 2013 to 2022 that are similar to the overall trend. The SDEs are oriented in a “northeast–southwest” direction, and the agricultural carbon emission economic efficiency shows a spatial pattern of “agglomeration–diffusion”. The east–west spatial distribution displays a “strengthening–weakening” variation, while the SDEs demonstrate a “shrinking–expanding” trend, with the degree of dispersion showing a “decreasing–increasing” pattern. In contrast, cities (prefectures) with severe rural population decline (Figure 6d) differ from those with mild and moderate decline. Their SDEs are skewed toward the northwest region and gradually shift southeastward, with smaller areas and lower flattening ratios. This indicates that the spatial distribution of agricultural carbon emission economic efficiency is more dispersed in these areas. This indicates that the evolution paths of economic efficiency in agricultural carbon emissions differ across various types of cities (prefectures) experiencing rural population decline.

4.3. Spatiotemporal Evolution Characteristics of Agricultural Carbon Emission Economic Efficiency in the Contiguous Karst Area of Yunnan–Guizhou

Using the spatial autocorrelation model of rural population decline and agricultural carbon emission economic efficiency, the global and local spatial autocorrelations were analyzed. Combined with GeoDa software, the spatial patterns were visualized. The results show that the global spatial autocorrelation Moran’s I is −0.3519, indicating a significant negative spatial correlation between the two. Specifically, a higher absolute value of the rural population decline rate corresponds to lower agricultural carbon emission economic efficiency.

The bivariate LISA cluster map (Figure 7) reveals distinct spatial patterns: High–High clusters are predominantly located in transitional zones between northeastern regions and Low–High clusters, where significant rural population decline coexists with relatively higher agricultural carbon emission efficiency due to productivity spillovers from adjacent advanced agricultural areas; High–Low clusters cover more extensive territories characterized by severe population decline and backward farming practices with resource-intensive and environmentally detrimental features, resulting in lower efficiency; discrete Low–High clusters, exemplified by Bijie and Liupanshui cities in the northeast, demonstrate improved efficiency through the “Three Reforms” initiative that optimized agricultural structures, retained rural labor, and developed integrated agro-industries to reduce emissions; while Low–Low clusters in central regions maintain stable populations but show lower efficiency due to their reliance on traditional agricultural practices. This indicates that rural population decline and the economic efficiency of agricultural carbon emissions exhibit a negative spatial correlation.

4.4. Direct Impact of Rural Population Decline on Agricultural Carbon Emission Economic Efficiency

To empirically examine the impact pathways of rural population decline on agricultural carbon emission economic efficiency in the Yunnan–Guizhou karst contiguous area, a mediation effect model was employed for regression analysis (

Table 3. Regression results of mediation effect.

Mediating Variable		Mild Rural Population Decline Group	Moderate Rural Population Decline Group	Severe Rural Population Decline Group
ISL	Model I	−2.8502 **	−1.1590 **	−1.0486 ***
	Model II	−0.1056 **	−0.2515 *	−0.1957 ***
	Model III	−9.6080 ***	0.2547	−2.4739 **
		−1.8360 ***	−1.2231 *	−0.1502
	Mediation effect:	Significant	-	-
EAL	Model I	−2.8502 **	−1.1590 **	−1.0486 ***
	Model II	−0.1388 **	−0.2428 **	−0.1942 **
	Model III	−1.1117	−2.1085 **	−1.9841 **
		−2.7684 **	− 0.9831 *	−1.7421 ***
	Mediation effect:	-	Significant	Significant
ATD	Model I	−2.8502 **	−1.1590 **	−1.0486 ***
	Model II	0.0019 *	0.0104 **	−0.0221
	Model III	−3.0584 **	−1.2496 ***	−2.4921 **
		−2.8561 **	−0.8711 *	−1.4571 ***
	Mediation effect:	Significant	Significant	-
NFE	Model I	−2.8502 **	−1.1590 **	−1.0486 ***
	Model II	0.21201	0.4976 ***	0.6976 **
	Model III	−0.9128	−1.2093 **	−1.5762 **
		−3.0437 **	−1.7608 **	−1.3773 **
	Mediation effect:	-	Significant	Significant
DLF	Model I	−2.8502 **	−1.1590 **	−1.0486 ***
	Model II	−43.7541 **	−53.7205 ***	−49.1957 ***
	Model III	−0.1444 ***	−0.0095 **	−0.1739 **
		−3.3255 **	−3.6480 ***	−3.1502 **
	Mediation effect:	Significant	Significant	Significant
IIL	Model I	−2.8502 **	−1.1590 **	−1.0486 ***
	Model II	−0.1438 **	0.0851 ***	0.0663 **
	Model III	−3.3371	−1.9925 ***	−0.5915 **
		−3.3296 **	−1.3286 **	−1.9031 **
	Mediation effect:	-	Significant	Significant
AEG	Model I	−2.8502 **	−1.1590 **	−1.0486 ***
	Model II	−3.2928	6.5864 ***	4.7961 **
	Model III	−0.0444 ***	− 0.1720 ***	−0.1732 **
		−3.3325 **	−2.2920 ***	−2.6263 **
	Mediation effect:	-	Significant	Significant

Note: *, **, and *** indicate significance at the 10%, 5%, and 1% levels, respectively.

, where Models I, II, and III correspond to Equations (6)–(8), respectively). The results demonstrate that the regression coefficient of rural population decline on agricultural carbon emission economic efficiency is positive and statistically significant at the 10% level, confirming a significant negative correlation—indicating that rural population decline exerts inhibitory effects on agricultural carbon emission economic efficiency. The direct impact manifests through two primary mechanisms: first, population decline undermines economies of scale and agglomeration effects, constraining the development of scaled agricultural operations and industrial structure optimization, thereby hindering green low-carbon transition; second, the loss of rural labor force coupled with aging population reduces agricultural economic vitality and technology adoption, leading to inefficient resource utilization. These factors collectively degrade agricultural development levels, escalate reliance on ecological resources, and ultimately diminish carbon emission economic efficiency in the agricultural sector. This suggests that rural population decline can directly affect the economic efficiency of agricultural carbon emissions and also indirectly influence it through its effects on the agricultural industrial system, agricultural production system, and agricultural management system.

This study referenced relevant research and adopted the bootstrap method to test the mediating effects [62]. In the analysis, no distinction was made between different types of rural population decline groups. As shown in

Table 4. Regression test.

Mediating Variable	Mediating Effect		95% Confidence Interval		Significance
	Coefficient	Std. Error	Lower Bound	Upper Bound
ISL	0.241	0.038	0.167	0.316	Significant
EAL	1.366	0.330	0.751	2.016	Significant
ATD	1.184	0.398	0.400	1.967	Significant
NFE	3.585	1.052	1.513	5.657	Significant
DLF	0.004	0.005	−0.03	−0.01	Significant
IIL	2.505	0.434	1.650	3.360	Significant
AEG	2.354	0.996	0.391	4.317	Significant

, the bootstrap test confidence intervals for the mediating variables did not include zero, indicating that the further validation criteria were met. This result demonstrates that rural population decline indeed indirectly affects the economic efficiency of agricultural carbon emissions by influencing the agricultural industrial system, agricultural production system, and agricultural management system.

5. Discussion

5.1. Indirect Impact of Pural Population Decline on Agricultural Carbon Emission Economic Efficiency

Using the mediation effect model, the indirect impact of rural population decline on agricultural carbon emission economic efficiency in the contiguous karst area of Yunnan–Guizhou was analyzed (Table 3). Rural population decline affects agricultural carbon emission economic efficiency by influencing the agricultural industrial system, agricultural production system, and agricultural management system. Unlike prior research that narrowly examines agricultural carbon emissions through specific lenses like crop production patterns in Latin America or agricultural management intensity in Chile [63,64], this study reveals that rural population decline impacts multiple agricultural dimensions: industrial restructuring, expansion of agricultural value chains, technology adoption rates, composition of non-farm employment, fragmentation patterns of arable land, land use intensity, and overall economic development. These multifaceted effects collectively drive shifts in the economic efficiency of agricultural carbon emissions.

Regarding the impact of rural population decline on the agricultural industrial system, rural population decline reduces agricultural carbon emission economic efficiency by inhibiting the optimization of agricultural industrial structure and hindering the extension of the agricultural industrial chain. However, agricultural industrial structure optimization shows limited progress in prefectures experiencing moderate-to-severe rural depopulation, while agricultural value chain extension remains insignificant in areas with severe population decline. This likely stems from substantial agricultural labor loss in moderate-to-severe depopulation zones, which leads to structural simplification and insufficient diversification, thereby obscuring observable impacts on industrial structure transformation. In contrast, in areas with mild rural population decline, the lower degree of population decline allows for more diversified agricultural industrial development and higher agricultural added value, resulting in a weaker negative impact on agricultural carbon emission economic efficiency through industrial chain extension.

Rural population decline influences agricultural production systems through two distinct pathways: enhanced technological adoption and expanded non-farm employment opportunities, which collectively contribute to reduced agricultural carbon emission economic efficiency. Nevertheless, this dynamic exhibits regional variations—agricultural technology penetration demonstrates limited progress in prefectures experiencing severe rural depopulation, while structural transformation of non-farm employment remains stagnant in areas with mild population decline. This paradoxical pattern emerges because, while labor shortages create demand for technological substitution of human capital, the concomitant loss of skilled workers simultaneously undermines local innovation capacity, ultimately impeding the modernization of agricultural production systems. Traditional agricultural production remains heavily resource-dependent and economically inefficient, leading to low agricultural carbon emission economic efficiency. Compounding this issue, the time-lag effects characteristic of cutting-edge agricultural technologies have hindered the immediate actualization of low-carbon farming advantages. In areas with severe rural population decline, the loss of high-quality labor suppresses agricultural technology penetration, making the regression insignificant. In areas with mild rural population decline, there is still a significant agricultural labor force, making the impact of non-agricultural employment structure on agricultural carbon emission economic efficiency insignificant.

Population decline influences agricultural management systems through three distinct pathways: reduced land fragmentation, heightened land use intensity, and stimulated agricultural economic development. While these factors collectively contribute to diminished agricultural carbon emission efficiency, the effects of land intensification and agricultural growth prove statistically insignificant in regions undergoing severe depopulation. Large-scale and intensive farmland use, as well as agricultural economic development, can improve agricultural production efficiency and promote efficient resource utilization, stimulating an increase in carbon emission economic efficiency. Nevertheless, the overexploitation of agricultural inputs can ultimately diminish agricultural carbon emission economic efficiency. In recent years, comprehensive land consolidation has significantly improved land fragmentation in karst regions, making the regression significant in areas with mild, moderate, and severe rural population decline. However, as human capital is a core element of agricultural production in impoverished areas, it directly affects farmland intensification and high-quality agricultural economic development in karst regions. Consequently, regions experiencing acute rural depopulation and substantial labor shortages demonstrate negligible progress in both farmland intensification and agricultural economic development.

5.2. Impact Mechanism of Rural Population Decline on Agricultural Carbon Emission Economic Efficiency

The magnitude of rural depopulation differentially influences both the baseline values and temporal trajectories of agricultural carbon emission economic efficiency across prefectures in the Yunnan–Guizhou karst region, amplifying regional heterogeneity. This dual-pathway mechanism aligns with theoretical expectations: directly through demographic effects on production systems, and indirectly through modernization pathways that vary with depopulation intensity (

Table 5. Key impact factors of rural population decline on the economic efficiency of agricultural carbon emissions.

Degree of Rural Population Decline	Agricultural Industrial System	Agricultural Production System	Agricultural Management System
Mild decline	Significant impact of agricultural industrial structure	Significant impact of agricultural technology penetration	Significant impact of cultivated land fragmentation
Moderate decline	Significant impact of agricultural value chain extension	Significant impacts of agricultural technology penetration; non-agricultural employment structure	Significant impacts of cultivated land fragmentation; cultivated land intensification
Severe decline	Significant impact of agricultural value chain extension	Significant impact of non-agricultural employment structure	Significant impacts of cultivated land fragmentation; cultivated land intensification; agricultural economic development

and Figure 8).

In areas with mild rural population decline, the phenomenon primarily influences agricultural carbon emission efficiency through agricultural industrial structure, technology penetration, and farmland fragmentation. Within the agricultural industrial system, rapid urbanization enables optimized industrial layouts that enhance value-added production through integrated “production–processing–marketing” chains and agro-cultural-tourism synergies, thereby increasing agricultural income, creating rural employment opportunities, mitigating population outflow, and improving carbon emission efficiency. Regarding the production system, even under severe population decline, agglomeration effects can accelerate agricultural technology diffusion, strengthen technological penetration capabilities, amplify positive technological impacts on emission efficiency, and facilitate green agricultural development. For the management system, comprehensive land consolidation promotes contiguous farmland parcels, enabling unified soil management and conservation that enhances organic carbon sequestration capacity while alleviating ecological fragility.

In prefectures with moderate rural depopulation, the principal mechanisms linking population decline to agricultural carbon emission economic efficiency encompass six dimensions: agricultural value chain expansion, technology adoption rates, non-farm employment restructuring, farmland consolidation, land use intensification, and agricultural economic growth. Within the agricultural industrial system, labor shortages constrain value chain extension into upstream and downstream sectors, depressing agricultural value-added potential, slowing economic returns, and impeding carbon efficiency improvements. The production system reveals a paradox: while agricultural technology diffusion mechanisms show progressive enhancement, concurrent growth in non-farm employment and selective outmigration of skilled laborers perpetuate extensive farming practices, obstructing the transition to sustainable agriculture. Regarding management systems, farmland intensification and scale expansion generate dual effects—stimulating economic output while escalating input dependency and emission intensity, ultimately exacerbating ecosystem vulnerability.

Severe rural population decline predominantly affects agricultural carbon emission efficiency through agricultural value chain extension, non-agricultural employment restructuring, farmland fragmentation, land use intensification, and agricultural economic development. Within the agricultural industrial system, intense population decline inevitably reduces production efficiency, compromising both output quantity and quality, which hinders value chain development and diminishes economic returns. In the production system, shrinking agricultural labor forces increase opportunity costs, prompting shifts toward less labor-intensive livestock farming—a sector with significantly higher carbon emissions than crop cultivation due to livestocks’ physiological characteristics, thereby impeding green agricultural development. The management system faces paradoxical effects: excessive land scale intensification and accelerated agricultural economic development correlate with intensified machinery usage and production inputs, exacerbating ecological pollution and consequently worsening agricultural carbon emission efficiency.

6. Conclusions and Suggestions

6.1. Conclusions

This paper focuses on the karst region as the research object, exploring the spatiotemporal evolution patterns and mechanisms of agricultural carbon emission economic efficiency under the influence of rural population decline. By using carbon source factors from agricultural material inputs, crop cultivation, and livestock breeding, as well as agricultural economic added value, the agricultural carbon emission economic efficiency is measured. From the perspective of rural population decline, the spatiotemporal evolution characteristics of carbon emission economic efficiency in the karst region are analyzed. Using the mediation effect model and based on the “Three Major Systems” of agricultural modernization development, the impact and mechanisms of rural population decline on carbon emission economic efficiency are explored.

Existing studies have predominantly focused on the impacts of regional development characteristics—such as demographic changes, industrial structure, energy mix, and economic growth—on agricultural carbon emissions. The economic efficiency of carbon emissions serves as a critical indicator for balancing agricultural development and emission reduction, making it a more pragmatic metric for quantifying decarbonization goals. This study innovatively integrates the realities of rural population decline with the “Three Major Systems” framework (agricultural industrial, production, and management systems) to expand the theoretical and empirical understanding of carbon emission efficiency in karst regions. Furthermore, by categorically analyzing mechanisms based on the severity of population decline (mild, moderate, and severe), this research deepens insights into the heterogeneous pathways through which rural depopulation shapes agricultural carbon efficiency, offering a nuanced typology absent in earlier literature.

This study comprehensively analyzes the evolution characteristics and impact mechanisms of rural population shrinkage on the economic efficiency of agricultural carbon emissions in karst regions, based on carbon emission data calculated from agricultural material inputs (fertilizers, pesticides, and plastic film), crop cultivation, and livestock breeding, as well as socioeconomic statistics. However, due to data availability constraints, the influence of factors such as energy structure is not addressed. While the research confirms the existence of mediating effects and verifies the impact of rural population shrinkage on agricultural carbon emission efficiency, future studies could incorporate more relevant indicators for robustness tests as data improves. More scientific spatial econometric methods [65] should also be employed to deepen theoretical and empirical research on carbon emission economic efficiency in karst regions. Additionally, although this study focuses on the 10-year impact of rural population shrinkage in karst areas, future research should extend to longer timeframes and broader geographical contexts, integrating regional characteristics to further explore underlying patterns and causal mechanisms.

6.2. Suggestions

For cities (prefectures) with mild rural population decline, fostering and expanding the agricultural carbon market can enhance the quality and efficiency of the agricultural industry. In terms of the agricultural industrial system, leveraging regional rural population advantages and agricultural development foundations, a government-led, socially participatory, and market-driven agricultural carbon trading system should be established. Regarding the agricultural production system, efforts should be accelerated to address deficiencies in key technological equipment for low-carbon agriculture, leverage the agglomeration effect of rural populations to enhance the penetration capacity of low-carbon agricultural technologies, and establish an intelligent fertilization system based on soil carbon sequestration monitoring. The revision of the Regulations on the Protection of Cultivated Land Quality should clearly set targets for reducing the application of nitrogen, phosphorus, and potassium fertilizers. In terms of the agricultural management system, it is essential to fully exploit the multi-functional expansion of agriculture—including its roles in product supply, ecological conservation, and cultural heritage preservation—while continuously optimizing the agricultural industrial structure, boosting agricultural economic performance, reducing carbon emissions, and enhancing the economic efficiency of agricultural carbon emissions.

Scientific agricultural development planning should be implemented in areas with moderate rural population decline to enhance carbon sequestration capacity. The agricultural industrial system demands customized development strategies that synergize regional resource endowments with demographic profiles to enhance structural transformation, optimize resource utilization, and expand value chain integration. Within the production system, strategic workforce planning should create agricultural employment opportunities while professional farmer training programs enhance technical capacities, building a skilled talent base for low-carbon development. In terms of agricultural management systems, we should strengthen the positive impact of large-scale and intensive cultivation of farmland on agricultural carbon emissions. This includes revising the Regulations on the Administration of Fertilizer Registration, implementing a value-added tax refund upon collection policy for enterprises engaged in the resource-based production of organic fertilizers from livestock and poultry manure, providing appropriate ecological subsidies based on organic fertilizer application areas, and thereby reducing carbon emissions to promote a steady improvement in the economic efficiency of agricultural carbon emissions.

For cities (prefectures) experiencing severe rural population decline, human capital enhancement and low-carbon agricultural intensification are imperative priorities. Regarding the agricultural industrial system, emphasis should be placed on cultivating new-type agricultural business entities and modernizing the entire supply chain—from agricultural input production to farming operations and product distribution—to boost production efficiency. Within the agricultural production system, the focus should center on supply-side structural reforms, optimizing non-farm employment configurations to reduce agricultural opportunity costs and foster green economic development. For the agricultural management system, leveraging advantages of scaled and intensive land use, innovative integrated crop–livestock production models should be established to promote green, low-carbon ecological cycles, thereby substantially improving agricultural carbon emission economic efficiency.