Evaluating the Ecological Restoration Effectiveness of Poverty Alleviation Relocation through Carbon Storage Analysis: Insights from Karst Regions

Abstract

The Poverty Alleviation Relocation (PAR) policy is widely regarded as an effective approach for breaking the cycle of ecological vulnerability and poverty. However, quantitative research on the ecological restoration effectiveness of PAR lacks sufficient experimental data support. This study focuses on the karst region and employs analysis methods such as volume-derived biomass and correlation analysis to evaluate the impact of PAR on carbon storage in forest ecosystems using on-site experimental data. The objective is to enhance and broaden the research framework for assessing PAR’s ecological restoration effectiveness. The findings reveal that, compared to the pre-PAR implementation period in 2015, the study area experienced an 8.16% increase in forest land area and a 6.57% increase in carbon storage after six years of PAR implementation in 2021. Following PAR implementation, carbon storage in the stone desertification area surged by 14.31%, indicating a significant correlation between PAR households and carbon storage variables. In the karst area, carbon storage increased by 4.34%, exhibiting a significant correlation between the two variables. Conversely, in the non-karst area, carbon storage rose by 5.01%, but no significant correlation was observed between the variables. Furthermore, post-PAR implementation, there is a discernible trend of stronger carbon storage enhancement with increasing distance from the relocated PAR households.

1. Introduction

The southern karst region of China stands as one of the world’s 30 major biodiversity hotspots and is a global priority for ecological conservation [1,2]. This region serves as a crucial ecological barrier for the upstream areas of China’s major rivers, the Yangtze and Pearl Rivers [3]. Over the past two decades, the region has achieved international recognition as a hotspot for ecological restoration research [4,5]. In addition, the southern karst mountainous region of China is distinguished by its high mountains, steep slopes, severe rocky desertification, and fragile ecological environment. Usable land resources are scarce in this area. Unlike other karst regions around the world, the population density here is notably higher, exacerbating the challenges associated with land use and environmental sustainability [6,7]. The region is characterized by scarce land resources, a dense population, and pronounced conflicts between human activities and the land. This stark contradiction between the limited natural resources and the burgeoning population has spawned an insurmountable dilemma, perpetuating a vicious cycle of poverty and environmental fragility. The intricate relationship between poverty and ecological vulnerability has culminated in this region harboring the highest concentration of impoverished populations in China [8,9]. Investigating the issues of poverty and ecological fragility in the karst region of southern China carries both theoretical and practical significance in fostering a harmonious coexistence between humanity and nature [10,11]. In response to the fragile ecological environment and the imperative to alleviate local poverty, the Chinese government has devised numerous policies tailored to address poverty and ecological fragility in the karst regions [12,13]. Among these policies, the relocation of farmers from ecologically vulnerable areas to regions with better conditions is considered the most effective strategy. This approach aims to simultaneously alleviate poverty and enhance the ecological environment [14,15]. The PAR initiative involves relocating farmers from severely impoverished or ecologically vulnerable areas to urban or town centers. This relocation provides them with free housing, convenient healthcare, high-quality education, abundant employment opportunities, and other amenities. Implemented during China’s “13th Five-Year Plan” period (2015–2020), PAR represents a continuation of ecological migration projects. The overarching goal of this approach is to eradicate poverty while concurrently enhancing the ecological environment of the areas from which people are relocated, achieved through spatial relocation [16,17]. During the “13th Five-Year Plan” period, China allocated a total investment of USD 10 billion for the PAR initiative. This funding facilitated the relocation of approximately 10 million of the most impoverished households. In the southern karst region, over 60% of the population has been relocated as part of this effort. PAR is considered the most effective method for addressing poverty in ecologically fragile areas and conforming to relevant theories such as the Environmental Kuznets Curve [18,19]. PAR serves as a potent tool for advancing the United Nations’ Sustainable Development Goals. Currently, research on PAR primarily emphasizes aspects such as voluntary relocation, the sustainability of farmers’ livelihoods, and the preservation of cultural heritage [20,21,22]. Our team has also engaged in studies on these topics [23]. However, there is a dearth of research specifically addressing the ecological restoration outcomes of PAR.

In order to comprehensively evaluate the ecological restoration effects of PAR, it is imperative to quantitatively analyze its contribution to ecological restoration and its regional suitability. Existing research on PAR’s ecological restoration predominantly relies on remote sensing data across various scales for research methods and data analysis. Some studies analyze the enhancement of Ecological Service Value facilitated by PAR at the county level [24]. Others evaluate the reduction in indoor PM2.5 concentrations in urban and rural areas due to PAR [25]. Additional research calculates PAR’s enhancement effect on carbon sequestration [26] or quantifies its promotion effect on the remote sensing ecological index [27]. Existing studies on PAR’s ecological restoration effectiveness heavily rely on observational data, particularly remote sensing data. These studies often utilize models to infer the extent of PAR’s contribution. However, most research is conducted at the provincial level or above, with some studies focusing on typical areas but still utilizing similar data sources. Notably, there is a lack of validation and support from field survey data in typical regions [26,27]. The current research framework on PAR’s ecological restoration effectiveness is characterized by a relatively limited range of data sources, a lack of experimental data support, and a deficiency in diverse research methods. These limitations potentially impact the credibility of research conclusions. To further enhance the research framework for PAR ecological restoration, this paper will build upon existing studies [26,28] by incorporating field survey data as a primary source of information to quantify the effectiveness of PAR ecological restoration. Considering that the carbon storage service function of terrestrial ecosystems is fundamental to human survival and the development of modern civilization [29,30], with forests playing a crucial role in fostering harmony between humans and nature [31,32], this study will focus on the carbon storage service function of forest ecosystems as an analytical indicator. Through field surveys in typical regions to collect data on forest systems, we aim to quantify their carbon storage service function. This approach will enable us to experimentally quantify the impact of PAR on the carbon storage service function of forest ecosystems. This paper can effectively enhance the methodology of PAR ecological restoration studies, providing validation and support for existing PAR ecological restoration research systems. It aims to offer scientific support for more effective and systematic research on the ecological restoration effectiveness of population migration. The findings have valuable insights for promoting harmonious coexistence between humans and nature in ecologically vulnerable areas.

2. Materials and Methods

2.1. Study Area

Guizhou Province, located in the core area of the southern karst region in China, has been designated by the Chinese government as a demonstration area for poverty alleviation and ecological civilization construction. This designation highlights its significant role in poverty eradication and ecological restoration [33]. PAR serves as a primary means for poverty alleviation and ecological restoration in Guizhou. This initiative involved the relocation of 1.92 million people, which represents approximately one-fifth of the total number relocated under PAR nationwide. The program spans 96.59% of the province’s counties (Figure 1a). The majority of the relocated population (97.25%) has been concentrated in urban areas, with 125 large resettlement points accommodating over 5000 people each and resettling 1.12 million individuals (Figure 1b). From 2000 to 2022, Guizhou Province has witnessed a significant increase in forest coverage, rising from 23.8% to 62.8% (Figure 1c). As the province with the most severe rocky desertification in the country (Figure 1d), Guizhou has now become one of the provinces with the highest forest coverage nationwide. Notably, Guizhou is the only province in China without plains, with a population density reaching 219 people/km² (Figure 1e). The majority of the population in mountainous areas relies on forests for their livelihoods, giving rise to the saying “living off the mountains”. The forest ecosystem in the mountainous regions of Guizhou is intricately linked to the lives of nearly 40 million residents in the province.

This article focuses on specific research areas in Guizhou, including severely rocky desertification areas, typical karst regions, and non-karst areas. The three main types of karst mountainous regions chosen for study are located in the western, northern, and eastern parts of Guizhou Province. The specific townships selected for study in these regions where PAR has been implemented are Angu Township in Qinglong County, Fuyan Town in Zheng’an County, and Wangfeng Township in Leishan County, respectively (Figure 2). The township of Angu is characterized by 70.59% of its area falling within the karst region, with 18.15% designated as moderately to severely rocky desertified terrain, rendering it a quintessential township representative of such ecological fragility. Consequently, it harbors the highest relocation population among the three townships. Fuyan Township boasts 96.81% of its territory situated within the karst region, with only 9.82% classified as mildly rocky desertified, thus representing a predominantly mild desertification landscape. Consequently, it accommodates the fewest relocation populations among the three townships. Conversely, Wangfeng Township resides outside the karst region, with a notable forest coverage rate of 67%, indicative of its superior ecological foundation. It accommodates a relocation population of 2698 individuals (

Table 1. Basic overview statistics.

Study Area	Total Area (km2)	Registered Population (Individuals)	Cultivated Land (Hectares)	Ecological Conditions		PAR
				Karst	Rocky Desertification	Household	Individuals
Angu	101.99	21.55	1000	70.59%	36.92%	933	4317
Fuyan	164.52	20.53	1530	96.81%	9.82%	577	2065
Wangfeng	98.69	15.64	750	0	0	671	2698

).

2.2. Experimental Methods

The experiment was conducted in the karst mountains of Guizhou. The forest survey data enriched the study’s data sources. Although the workload is substantial and the work period is lengthy, the method is straightforward, and the operation is simple and standardized. Currently, research in this field predominantly focuses on small to medium scales [34]. Considering the complexity and accessibility of experimental data and recognizing the shortcomings in previous research in typical areas, this study will focus on several townships in Guizhou Province, the core zone of the karst region in southern China, as the units for analysis. Given the diverse ecological vulnerabilities present in the karst region, the study will select one township with severe rocky desertification in the western part of Guizhou Province, one township with developed agriculture in the northern part, and another non-karst township in the southeastern part as representative areas. These selections aim to investigate the evolution of the carbon storage service function of forest ecosystems induced by PAR in different ecological settings. Through on-site investigation data analysis, the evolution of carbon storage in forest ecosystems will be analyzed before (2015) and after (2021) the implementation of PAR. This study will assess the correlation between the number of people relocated through PAR and carbon storage variables, determine the driving force of PAR implementation on regional carbon storage capacity, analyze the correlation between the distance change of migration relocation points and regional carbon storage evolution, and examine the differences in carbon storage capacity evolution caused by PAR under different levels of ecological fragility, aiming to quantify the ecological restoration effectiveness of PAR through systematic experimental studies (Figure 3).

2.2.1. Forest Biomass Survey

In accordance with the requirements of technical standards such as the “National Comprehensive Monitoring Technical Regulations for Forest and Grassland Ecology”, and taking into consideration data from forest resource surveys, the growth status of forest stands, forest stand maps, vegetation distribution maps, and soil type distribution maps were found. Field surveys and sample collection were conducted in these plots. The sampling was carried out in early July 2015 and mid-August 2021. Within the established standard sample plots, trees were measured with a starting diameter at breast height (DBH) of 5 cm (trees below 5 cm DBH were treated as shrubs for measurement). Survey factors included DBH, tree height, tree species, canopy closure, stand age, crown width, and growth status of the forest stand.

2.2.2. Experimental Methods

This study specifically focuses on the biomass of standing trees in the forest, including mature trees, sparse forests, Shrubland, and nurseries. It excludes the biomass of the herbaceous layer and deadwood. The volume-derived biomass method employed in this research has proven effective in estimating forest biomass and is considered one of the most crucial methods currently available [35,36]. The estimation of biomass for mature trees and sparse forest in this study follows the methodology provided by the Intergovernmental Panel on Climate Change (IPCC).

$$C_i=B_i\times F_i$$

(1)

$$B_i={\displaystyle \sum V_i}\times E_i\times F_{BE}\times (1+R_i)\times A_i$$

(2)

$$V_i=0.0000657\times D_i^{1.94106}\times H_i^{0.84929}$$

(3)

In the equation, C_i represents the carbon storage of tree species i (Mg); B_i represents the total biomass of tree species i (Mg); F_i represents the carbon content coefficient of tree species i; V_i represents the standing volume of tree species i per unit area (m³·hm⁻²); E_i represents the wood density of tree species i (Mg·m⁻³); F_BE represents the biomass expansion factor of tree species i; R_i represents the root-to-shoot ratio of tree species i; A_i represents the area covered by tree species i (hm²); D represents the average diameter at breast height (DHB) of tree species i (cm); H represents the average height of tree species i (m). During forestry surveys, the accumulation-related data of Shrubland, bamboo forests, nurseries, and other undeveloped forest lands were not statistically recorded. According to the research findings of Fang **gyun, the biomass values for Shrubland and nurseries are set to 19.76 Mg·hm⁻², and for Sparse Woodland, it is 9.88 Mg·hm⁻². The carbon content ratios are 0.465 for Shrubland and 0.4705 for Sparse Woodland. This paper involves specific values for biomass expansion factors, basic wood density, root-to-shoot ratio, and carbon content ratio of Dominant Species, as detailed in

Table 2. Carbon storage quantification model for Dominant Species (groups).

Dominant Tree Species	Biomass Expansion Factor	Wood Basic Density	Root-to-Shoot-Ratio	Carbon Content
Masson Pine	1.8763	0.4482	0.1886	0.5271
Cunninghamia lanceolata	1.8611	0.3071	0.2338	0.5127
Broadleaf Deciduous Forest	1.8451	0.4222	0.2175	0.4536
Broadleaf Evergreen Forest	1.7127	0.6062	0.215	0.4822
Oak–Hickory Forest	1.7609	0.6119	0.2424	0.4798
Cupressus funebris Endl	1.8064	0.2893	0.2277	0.5331
Liquidambar formosana Hance	2.6871	0.644	0.2148	0.4502
Betula spp.	1.8082	0.527	0.27	0.4914
Paulownia	1.8855	0.4754	0.2218	0.6811

[37].

2.2.3. Correlation Analysis

The Pearson correlation coefficient was utilized to validate the degree of correlation between PAR and carbon storage variables. This coefficient, also known as the Pearson product moment correlation coefficient, assesses the linear correlation between two variables by analyzing their deviations from their respective mean values [38].

$$P=\frac{{\displaystyle \sum _{i=1}^n(x_i-\overline{X})(y_i-\overline{Y})}}{\sqrt{{\displaystyle \sum _{i=1}^n{(x_i-\overline{X})}^2{\sqrt{{\displaystyle \sum _{i=1}^n(y_i-\overline{Y})}}}^2}}}$$

(4)

In the formula, $\overline{X}$ represents the mean of sequence X, $\overline{Y}$ represents the mean of sequence Y, and P represents the Pearson correlation coefficient. The Pearson correlation coefficient, denoted as P, ranges between −1 and 1. The larger the absolute value of p, the higher the degree of correlation between sequences X and Y. When p > 0, its value is divided into the following intervals: 0–0.2 (very weak correlation), 0.2–0.4 (weak correlation), 0.4–0.6 (moderate correlation), 0.6–0.8 (strong correlation), 0.8–1 (very strong correlation). This categorization results in five levels.

3. Results

3.1. Overall Change Trends

The results showed conditions before PAR implementation in 2015 and after PAR implementation in 2021. The analysis examines the evolution of basic forest conditions. According to the results from three townships’ stand compartments, in 2015, the total forest area of the three townships was 20,852.99 hectares, involving 13 types of Dominant Species, including Masson Pine, Cunninghamia lanceolata, Cupressus funebris Endl, Betula spp., shrub forest, and others. The combined area of Masson Pine and Cunninghamia lanceolata was 8694.5 hectares, accounting for 41.69% of the total forest area; shrub forest accounted for 19.09%, and other types of forest accounted for 39.22%. In 2021, the forest area increased to 22,553.69 hectares, a growth of 1700.70 hectares or 8.16% compared to the pre-PAR implementation period. The Dominant Species involved 13 types, with Masson Pine and Cunninghamia lanceolata covering 22,553.69 hectares, representing 43.21% of the total forest area; shrub forest accounted for 19.14%, and other types of forest accounted for 37.65%. Detailed data on Dominant Species, age groups, canopy closure, average tree height, and other specific aspects of the forest ecosystem for two years are provided in

Table 3. On-site survey statistics.

Area	Dominant Tree Species	2015					2021
		Area (hm2)	DHB (cm)	Age of Class	Canopy Closure	Height of Tree	Area (hm2)	DHB (cm)	Age of Class	Canopy Closure	Height of Tree
Wangfeng	Masson Pine	2550.89	15.56	2	0.51	1.26	2778.33	16.15	2	0.51	1.31
	Cunninghamia lanceolata	3201.02	10.82	3	0.43	0.99	3381.09	10.29	3	0.43	0.99
	Broadleaf Deciduous Forest	183.6	16.2	2	0.58	0.81	271.98	12.36	2	0.49	0.72
	Broadleaf Evergreen Forest	21.34	10.4	2	0.45	1.31	43.1	8.6	2	0.45	1.31
	Camellia sinensis	92.12					158.29
	Sparse Woodland	48.34
	Shrubland	389.25					501.41
	Bamboo Thicket	61					75.06
	Total	6547.56					7209.26
Fuyan	Masson Pine	573.16	7.56	1	0.45	0.83	600.84	9.2	1	0.45	0.83
	Cunninghamia lanceolata	1211.55	9.78	2	0.42	0.81	1891.13	7.8	2	0.41	0.68
	Broadleaf Deciduous Forest	839.54	7.85	1	0.47	0.78	733.13	8.24	2	0.49	0.72
	Broadleaf Evergreen Forest	3913.74	6.36	1	0.51	0.51	4073.78	5.32	2	0.47	0.53
	Camellia sinensis	158.93					118.58
	Oak–Hickory Forest	479.3	9.7	1	0.56	0.63	440.8	10.1	1	0.58	0.63
	Cupressus funebris Endl	93.51	7.7	1	0.43	0.65	75.94	7.5	1	0.41	0.65
	Sparse Woodland	174	4.36	1	0.42	0.48	175.89	4.36	2	0.42	0.48
	Shrubland	1177.94					1217.39
	Bamboo Thicket	1001.63					725.11
	Total	9623.3					10,052.59
Angu	Masson Pine	257.44	15.36	2	0.53	1.18	273.03	16.21	2	0.53	1.23
	Cunninghamia lanceolata	900.44	12.5	2	0.58	1.11	820.97	12.6	2	0.58	1.11
	Broadleaf Deciduous Forest	524.08	11.7	2	0.5	0.95	926.77	7.5	2	0.5	0.95
	Oak–Hickory Forest	351.11	14.6	3	0.66	1.23	337.87	15.1	3	0.66	1.23
	Liquidambar formosana Hance	88.84	17.2	2	0.47	1.23	96.49	17	2	0.5	1.21
	Betula spp.	107.83	15.3	2	0.49	1.21	97.31	17	2	0.49	1.21
	Paulownia	42.36	15.9	3	0.51	1.25	58.55	15.5	3	0.51	1.25
	Sparse Woodland						82.87	6	1	0.5	0.5
	Shrubland	2410.03					2597.98
	Total	4682.13					5291.84

Note: In the table, the age classes represented by 1, 2, 3, respectively, correspond to young forests, middle-aged forests, near-mature forests.

.

3.2. Temporal and Spatial Evolution of Carbon Storage

The results of this study reveal a significant improvement trend in the forest ecosystem of the study area from 2015 to 2021. During this period, there was a notable increase in forest land area by 7.89% and in carbon storage by 6.57%. The biomass model was utilized to calculate the biomass of the study area, and carbon storage for the region was then computed based on the carbon content rates of various forest stands. In 2015, before the implementation of PAR, the total wood biomass in the study area was 969,200 tons, with a carbon storage value of 481,900 tons. Among these, the carbon storage contributed by Masson Pine and Cunninghamia lanceolata was 255,700 tons, accounting for 53.07% of the entire forest ecosystem’s carbon storage. In 2021, after the implementation of PAR, the total biomass of the forest ecosystem in the study area was 1,032,000 tons, with a total carbon storage of 513,600 tons, an increase of 31,700 tons or 6.57% compared to 2015. Among these, the carbon storage contributed by Masson Pine and Cunninghamia lanceolata was 279,200 tons, representing 54.36% of the total carbon storage of the entire forest ecosystem. After the implementation of PAR, the forest area increased by 8.16% and carbon storage increased by 6.57%. Specifically, the carbon storage of Masson Pine and Cunninghamia lanceolata increased by 23,400 tons, accounting for 73.96% of the total increase in carbon storage of the entire forest ecosystem. The spatial evolution trends of carbon storage in three typical townships are as follows (

Table 4. Forest ecosystem carbon storage evolution statistics.

Area	Dominant Tree Species	2015		2021		Variables
		Area (Hectare)	Carbon Storage (t)	Area (Hectare)	Carbon Storage (t)	Area (Hectare)	Rate (%)	Carbon Storage (t)	Rate (%)
Wangfeng	Masson Pine	2550.89	92,303	2778.33	99,717	227.44	8.92%	7415	8.03%
	Cunninghamia lanceolata	3201.02	80,142	3381.09	80,487	180.07	5.63%	345	0.43%
	Broadleaf Deciduous Forest	183.6	5863	271.98	5627	88.38	48.14%	−235	−4.02%
	Broadleaf Evergreen Forest	21.34	559	43.1	953	21.76	101.97%	393	70.34%
	Camellia sinensis	92.12	846	158.29	1454	66.17	71.83%	608	71.83%
	Sparse Woodland	48.34	444			−48.34		−444
	Shrubland	389.25	3577	501.41	4607	112.16	28.81%	1031	28.81%
	Bamboo Thicket	61	567	75.06	698	14.06	23.05%	131	23.05%
	Total	6547.56	184,301	7209.26	193,544	661.7	10.11%	9242	5.01%
Fuyan	Masson Pine	573.16	16,676	600.84	17,475	27.68	4.83%	799	4.79%
	Cunninghamia lanceolata	1211.55	31,023	1891.13	44,838	679.58	56.09%	13,814	44.53%
	Broadleaf Deciduous Forest	839.54	19,668	733.13	14,446	−106.41	−12.67%	−5221	−26.55%
	Broadleaf Evergreen Forest	3913.74	92,997	4073.78	96,002	160.04	4.09%	3005	3.23%
	Camellia sinensis	158.93	1460	118.58	1090	−40.35	−25.39%	−371	−25.39%
	Oak–Hickory Forest	479.3	16,300	440.8	15,934	−38.5	−8.03%	−366	−2.25%
	Cupressus funebris Endl	93.51	2522	75.94	1840	−17.57	−18.79%	−682	−27.06%
	Sparse Woodland	174	1599	175.89	1616	1.89	1.09%	17	1.09%
	Shrubland	1177.94	10,823	1217.39	11,186	39.45	3.35%	362	3.35%
	Bamboo Thicket	1001.63	9312	725.11	6741	−276.52	−27.61%	−2571	−27.61%
	Total	9623.3	202,381	10,052.59	211,168	429.29	4.46%	8787	4.34%
Angu	Masson Pine	257.44	7670	273.03	8206	15.59	6.06%	537	7.00%
	Cunninghamia lanceolata	900.44	27,926	820.97	28,428	−79.47	−8.83%	502	1.80%
	Broadleaf Deciduous Forest	524.08	13,780	926.77	23,100	402.69	76.84%	9319	67.63%
	Oak–Hickory Forest	351.11	14,155	337.87	13,919	−13.24	−3.77%	−237	−1.67%
	Liquidambar formosana Hance	88.84	0	96.49	0	7.65	8.61%	0
	Betula spp.	107.83	0	97.31	0	−10.52	−9.76%	0
	Paulownia	42.36	0	58.55	0	16.19	38.22%	0
	Sparse Woodland			82.87	761	82.87		761
	Shrubland	2410.03	22,144	2597.98	23,871	187.95	7.80%	1727	7.80%
	Total	4682.13	85,675	5291.84	98,285	609.71	13.02%	12,610	14.72%
Total		20,852.99	472,358	22,553.69	502,997	1700.7	8.16%	30,639	6.49%

and Figure 4).

In Wanfeng Township, a non-karst township, the carbon storage of the forest ecosystem was 184,300 tons in 2015, with the highest carbon storage per unit area, approximately 1862 tons/km². Masson Pine and Cunninghamia lanceolata contributed 172,400 tons, accounting for 93.57% of the total carbon storage in the area. Carbon storage per unit area in all 17 villages was at a relatively high level, showing an evenly distributed pattern, with slightly lower values in the area around the central township government, Wanfeng Village. In 2021, carbon storage increased to 193,500 tons, a rise of 9242 tons or 5.01% compared to 2015. Masson Pine and Cunninghamia lanceolata contributed 180,200 tons, accounting for 93.57% of the total carbon storage in the area. The spatial distribution remained evenly distributed, with a slightly higher increase in carbon storage in the northwest and southwest regions. After the implementation of PAR, Wanfeng Township, located in a non-karst area, showed a certain increase in carbon storage. The carbon storage of the entire forest ecosystem was evenly distributed, with a high proportion of dominant tree species carbon storage and strong per-unit area carbon storage.

In Angu Township, where desertification is relatively severe, the carbon storage of the forest ecosystem was 95,200 tons in 2015, which is approximately half of the area of Wanfeng Township with a similar land area. The highest carbon storage tree species were Cunninghamia lanceolata and shrub forest, contributing 50,100 tons, accounting for 52.59% of the total carbon storage in the township. Spatial distribution showed a significantly lower carbon storage in the northern part, the highest carbon storage in the central part, and a relatively balanced state in the southern part. In 2021, carbon storage increased to 108,800 tons, with a growth of 13,600 tons and an increase of 14.31%, approximately three times the growth rate of Wanfeng Township. Among these, the carbon storage of Cunninghamia lanceolata and shrub forest was 52,300 tons, accounting for 48.05% of the total carbon storage in the township. The spatial distribution showed a more significant increase in carbon storage in the northern region, particularly in Baiyan Village, and a relatively obvious increase in carbon storage in the southwest region, with the central region remaining relatively stable. Angu Township had a relatively weak foundation for carbon storage in the forest ecosystem, and after the implementation of PAR, there was a significant increase in carbon storage. The increase was most noticeable in the northern and southwest regions.

In Fuyan Township, a karst region, the carbon storage of the forest ecosystem was 202,400 tons in 2015, the highest among the three townships, with a per-unit land area carbon storage of 1226 tons/km², accounting for 66% of that of non-karst Wanfeng Township. The highest carbon storage tree species were other hardwoods and Cunninghamia lanceolata, with a total of 124,000 tons, accounting for 61.28% of the entire system’s carbon storage. Spatial distribution showed a significantly higher carbon storage value in the central and southern parts and relatively lower values in other regions. In 2021, carbon storage increased to 211,200 tons, with a growth of 8787 tons and an increase of 4.34%. Among these, the carbon storage of other hardwoods and Cunninghamia lanceolata was 140,800 tons, accounting for 66.70% of the entire system’s carbon storage. The spatial distribution remained relatively high in the central and southern regions, with a noticeable increase in carbon storage in the northwest region. After the implementation of PAR, Fuyan Township’s forest system showed a certain increase in carbon storage.

After the implementation of PAR, among the three typical townships in Guizhou Province, Angu Township, where desertification is most severe and the forest ecosystem’s carbon storage is the weakest, showed the most significant increase in carbon storage. At the same time, its dominant tree species contributed the least to carbon storage, and shrub forest carbon storage ranked second, indicating a significant increase in carbon storage after the implementation of PAR. However, the stability of carbon storage in the forest ecosystem is still low, and the foundation is still weak. In Fuyan Township, a karst region with the largest land area, the highest total carbon storage in the forest ecosystem, and the highest per-unit area carbon storage, the increase in carbon storage after PAR implementation was relatively low, but dominant tree species contributed strongly compared to Angu Township. In non-karst Wanfeng Township, carbon storage increased to a certain extent after PAR implementation, with dominant tree species contributing significantly, and the forest ecosystem’s carbon storage remained relatively stable.

3.3. Relationship between PAR and Carbon Storage

In the three typical towns in the karst region, the implementation of PAR has led to a certain degree of increase in carbon storage. However, the change in carbon storage is influenced by various factors, and we need to further analyze the correlation between carbon storage variables and PAR (Figure 5).

In the non-karst town of Wangfeng, among the 17 village areas, the forest ecosystem carbon storage increased in 7 villages, while it decreased in 10 villages. Through bivariate analysis of the village relocation population and carbon storage variables for each village area, the calculated Pearson correlation coefficient is 0.284, with a significance level (p value) greater than 0.05. This indicates that there is no significant correlation between the increase in carbon storage in the local forest ecosystem and PAR implementation.

In the rocky desertification town of Angu, carbon storage increased in all five village areas. Through bivariate analysis of the population relocated within the village area and the carbon storage within the village area, the calculated Pearson correlation is 0.948, and the significance level (p value) is 0.014. In this typical rocky desertification town, there is a significant positive correlation between carbon storage variables in the forest ecosystem and the population relocated due to PAR, with a correlation coefficient of 0.948 at a significance level of 0.05.

In the karst town of Fuyan, carbon storage increased in all six village areas. Through bivariate analysis of the number of local relocated farmers and the increment in carbon storage, the calculated Pearson correlation coefficient for Fuyan is 0.838, and the p value is at the 0.05 level, indicating a significant correlation. As a karst town, the implementation of PAR in Fuyan has led to a decrease in the local population, reducing disturbances to forest resources and, to some extent, promoting an increase in carbon storage in the forest ecosystem. There is a significant correlation between PAR implementation and an increase in carbon storage.

Overall, in the typical karst province of Guizhou, Angu, which has the most severe rocky desertification, shows the most significant increase in carbon storage, with a significant correlation with the number of relocated farmers due to PAR. In Fuyan, a karst town, although the increase in carbon storage is relatively low, there is still a significant correlation with the number of relocated farmers due to PAR. On the contrary, in the non-karst area of Wangfeng, there is no significant correlation between PAR implementation and changes in carbon storage, indicating a weak overall correlation between local carbon storage increase and PAR implementation.

3.4. Impact of PAR on Carbon Storage at Different Distances from Affected Forest Areas

The impact of human activities on ecology varies with distance. To analyze the ecological impact of PAR on the region, the relationship between PAR and regional carbon storage changes was analyzed based on the distance of various stand compartments from the PAR initiative’s relocation points. In the towns where PAR was implemented, the study area’s stand compartments were categorized into four levels based on their distance from the relocation point: 200 m, 500 m, 1000 m, and >1000 m. The analysis aimed to understand the trends in stand compartment carbon storage capacity changes at different distances from PAR relocation points. After the implementation of PAR, the area within 200 m of the relocation points covered 8285.87 hectares, increasing by 87.96 hectares compared to before PAR implementation. Carbon storage decreased by 1065 t, representing a reduction of 0.59%. This distance zone is mainly around farmers’ households and villages, with minimal changes in forest ecosystem carbon storage and virtually no change. In the 200–500 m range, the forest area increased by 648.7 hectares after PAR implementation, with a growth rate of 10.94%. Carbon storage increased by 10,301 t, showing a rise of 7.64%. This area is relatively close to villagers’ homes, and after PAR implementation, most farmers still live in the villages. Some cultivated land has been converted to forest land, promoting an increase in regional forest ecosystem carbon storage. In the 500–1000 m range, the forest area increased by 648.7 hectares after PAR implementation, with a growth rate of 13.27%. Carbon storage capacity increased by 13,108 t, representing an increase of 11.51%. This area is relatively farther from villages and is a concentrated region of farmers’ cultivated land and forest land. More conversion of cultivated land to fallow and afforestation occurred in this area, contributing to an increase in regional carbon storage. In the >1000 m range, the forest area increased by 343.1 hectares, with a growth rate of 16.84%. Carbon storage capacity reached 61,993 t, increasing by 9561 t, with a growth rate of 18.24%. This area has the largest increase in carbon storage capacity, which is the least manageable by farmers. The implementation of PAR had a significant impact on the forest ecosystem carbon storage in this area. Overall, after PAR implementation, different distances from PAR relocation points showed varying changes in forest ecosystem carbon storage, generally indicating that the farther the distance from the PAR relocation points, the more significant the increase in carbon storage (Figure 6).

4. Discussion

The volume-derived biomass method employed in this study is widely recognized as a highly accurate approach for analyzing forest ecosystems. The calculated carbon storage per unit for each tree species is consistent with findings from existing research [39,40]. The results of this study indicate a significant improvement in the forest ecosystem of Guizhou from 2015 to 2021, corroborating closely with previous research conclusions [41]. The analysis reveals an overall increase of 6.57% in carbon storage in the regional forest ecosystem following the implementation of PAR. In comparison with existing research, PAR facilitates a 5.28% increase in the Remote Sensing Ecological Index or a 2.98% increase in carbon storage. The impact of PAR on regional ecological restoration is essentially similar to that observed in related studies [26,27]. Unlike methods using large-scale remote sensing data to calculate carbon storage [42], this study validates the effectiveness of PAR in regional ecological restoration using experimental data. The findings align with earlier research supported by remote sensing data [26], and the inclusion of experimental data from typical areas has significantly bolstered the study of PAR ecological restoration effectiveness, thus enhancing the research framework for PAR ecological restoration effectiveness to a considerable extent.

The driving factors behind the evolution of carbon storage in the forest ecosystems of ecologically fragile areas are complex. In addition to the influence of human activities, natural factors such as temperature and precipitation also play a role [43,44]. Guizhou Province, characterized by a subtropical monsoon climate with abundant water and heat conditions, provides favorable natural conditions for vegetation growth [45,46]. Our research results indicate that, in three typical towns involved in PAR, the area of Cunninghamia lanceolata and Cunninghamia lanceolata increased by 1050.89 hectares, accounting for 61.79% of the total forest area increment. The carbon storage increased by 23,412 tons, representing 73.96% of the total carbon storage increment in the entire forest ecosystem. This suggests that after the implementation of PAR, not only did the forest area increase, but the forest species also gradually transitioned to types more conducive to enhanced carbon storage. We observed that the impact of PAR implementation on the carbon storage of forest ecosystems is greatest in typical townships with rocky desertification, with the highest correlation coefficient. The increase in tree species is predominantly Shrubland, resulting in enhanced carbon storage in forest ecosystems, although the contribution per unit area is not substantially higher. Similarly, in karst typical townships, a significant correlation exists, albeit with a smaller correlation coefficient compared to rocky desertification areas, and the p value is also smaller. The predominant type of forest increase is dominated by Dominant Species such as Chinese fir. In non-karst regions, the enhancement of forest carbon storage is not as significant as in other areas. The increased forest is primarily dominated by Chinese fir and other Dominant Species, but the results show no correlation between forest carbon storage enhancement and PAR. This suggests that the implementation of PAR has a more pronounced effect on promoting the improvement of carbon storage in forest ecosystems in ecologically vulnerable areas, with further room for enhancement. In regions with relatively good ecological conditions, there is no significant promotion effect. This conclusion is generally consistent with previous research findings [27]. Some current research results suggest that relocating rural populations to centralized areas may cause fragile ecosystems to be unable to withstand large-scale settlement, leading to further ecological degradation. However, it is worth noting that such studies often focus on arid and semi-arid regions, and the relocation methods are not primarily centered around urban consolidation [47,48]. The conclusion of this study, combined with existing research, demonstrates that the improvement of ecological restoration through PAR is geographically specific. It is not universally applicable that ecological conditions will improve in all regions after the implementation of PAR. More research suggests that the impact of migration and relocation on ecology is complex, exhibiting varying effects in different geographic locations and regions [49]. Our research findings also demonstrate the diverse and complex effects of PAR on the evolution of carbon storage in forest ecosystems. In ecologically fragile areas such as rocky desertification regions, the ecological restoration effectiveness of PAR is more pronounced. In non-karst areas, the ecological restoration effectiveness of PAR is weaker than in rocky desertification regions, aligning with our earlier research conclusions [27]. We attribute it to regions with poorer ecological foundations, where the magnitude of ecological improvement may be greater. We also found that the impact of PAR implementation on carbon storage in forest ecosystems is somewhat correlated with distance. The farther from the PAR relocation point, the more pronounced the enhancement in forest ecosystem carbon storage, a conclusion that is generally consistent with previous research [50]. In summary, the effects of PAR on regional forest ecosystem carbon storage are relatively complex and exhibit certain regional characteristics. PAR and afforestation complement each other, with PAR indirectly promoting an increase in forested areas in relocation regions. Overall, implementing PAR in mountainous areas is beneficial for ecological restoration [51].

This study has successfully achieved the goal of quantifying the contribution of PAR to ecological restoration through experimental data, providing an effective supplement to systematically analyze the effectiveness of PAR ecological restoration. We selected three typical towns as our study areas, each located in different ecologically vulnerable regions with distinct PAR households. This approach brings a certain level of representativeness to the study of the ecological restoration effectiveness of PAR implementation in karst mountain areas. However, the number of sample plots is relatively limited, and in subsequent phases, we aim to use field survey data on a larger scale for validation. This will enable us to enhance the experimental research on the impact of PAR on ecological restoration more effectively. Moreover, our study focused specifically on Guizhou Province, and natural environments vary significantly across different regions in China or globally. Particularly in the arid and semi-arid areas of northern China, which are also key regions for PAR implementation, the conditions for natural vegetation restoration may not be as favorable. After reducing human disturbances or management in these areas, will it lead to the promotion of local ecological degradation? This will be a key issue for our future attention. Expanding the scope of our research to investigate the ecological restoration impact of PAR on a larger scale, both nationally and globally, will also be a focal point in our future studies. This paper only utilizes data from two periods, namely, before and after the implementation of PAR, to infer the ecological restoration contributions of PAR based on the variable changes observed in these two periods. This approach indeed exhibits a certain degree of bias. Numerous studies have demonstrated that carbon storage is influenced by various factors. However, we attempted to quantify the ecological restoration contributions of PAR using remote sensing data [28], benefiting from the ease of access to remote sensing data and the ability to perform multivariate regression model calculations based on long time series remote sensing data. Drawing on the research findings of Fan [48] and Zhang [49], among others, when analyzing the ecological impact of migration relocation using experimental data, we utilized social survey data to analyze the impact of PAR on regional ecology, dissecting the reasons for the ecological impact of population migration. The results of the social questionnaire survey indicate that a significant amount of land abandoned by relocated farmers is returned to forestation or left fallow, which is a crucial factor driving the substantial increase in forest land in agricultural areas (Supplementary Table S1).

5. Conclusions

(1)

The results of this study reveal a significant improvement trend in the forest ecosystem of the study area from 2015 to 2021. During this period, there was a notable increase in forest land area by 7.89% and carbon storage by 6.57%. The total carbon storage in the forest ecosystem of the study area was 48.19 million tons before the implementation of the Poverty Alleviation and Relocation (PAR) policy in 2015. After PAR implementation in 2021, the total carbon storage increased to 51.35 million tons, marking a gain of 3.17 million tons and a growth rate of 6.57% across the three typical townships.

(2)

After the implementation of PAR, the increase in carbon storage was most significant in rocky desertification areas, followed by karst areas, with no noticeable impact in non-karst areas. Following PAR implementation in the three typical townships, in the desertification-prone township of Angu, carbon storage increased by 13,626 tons, representing a growth of 14.31%. The Pearson correlation coefficient, calculated at a significance level where p < 0.05, indicates a significant correlation (coefficient: 0.948) between PAR households and carbon storage variables. In the karst township of Fuyan, carbon storage increased by 8787 tons, showing a growth rate of 4.34%. The Pearson correlation coefficient, also significant (coefficient: 0.838), suggests a noteworthy correlation between PAR households and carbon storage variables. In the non-karst township of Wangfeng, carbon storage increased by 9242 tons, with a growth rate of 5.01%. However, the Pearson correlation coefficient (coefficient: 0.284) between PAR households and carbon storage variables did not exhibit a significant correlation.

(3)

Post-PAR implementation, the forest ecosystem’s carbon storage demonstrated different trends based on proximity to the relocation point, and the farther the area is from the relocation point, the more pronounced the increase in forest carbon storage. Within 200 m of the PAR relocation point, carbon storage decreased by 1065 tons, representing a reduction of 0.59%. In the 200–500 m range, carbon storage increased by 10,301 tons, showing a growth rate of 7.64%. Within the 500–1000 m range, carbon storage increased by 13,108 tons, reflecting a growth rate of 11.51%. Beyond 1000 m from the relocation point, carbon storage increased by 9561 tons, indicating a growth rate of 18.24%.