The Influence Mechanism of Agricultural Heritage Systems Conservation on Farmers’ Sustainable Livelihoods: Evidence from Tea Globally Important Agricultural Heritage Systems in China

Abstract

Agricultural Heritage Systems (AHS), which embody economic, ecological, and social sustainability, are critically important for leveraging these dimensions to advance the sustainable development of farmers’ livelihoods. Utilizing a newly developed evaluation index system for sustainable livelihoods, this study applies mediating effect models to cross-sectional survey data collected from farmers to investigate the mechanisms through which AHS conservation influences their sustainable livelihoods. The analysis focuses on two Tea Globally Important Agricultural Heritage Systems (Tea-GIAHS) sites in Fujian Province, China: the Anxi Tieguanyin Tea Culture System (ATTCS) and the Fuding White Tea Culture System (FWTCS). The findings indicate that Tea-GIAHS conservation significantly enhances farmers’ sustainable livelihoods, primarily by fostering the scaling and industrialization of traditional agricultural operations. Furthermore, the positive effects are more pronounced among households with higher initial livelihood levels, higher incomes, and those primarily engaged in agriculture. Consequently, this study recommends that AHS sites enhance support for large-scale traditional farming practices among farmers and continuously improve and extend the industrial and value chains of traditional agriculture to further promote livelihood sustainability.

1. Introduction

Agricultural Heritage Systems (AHS) represent quintessential “socio-ecological” production landscapes that have remained functional for centuries or even millennia. This enduring resilience stems from their inherent stable and sustainable mechanisms, operating across multiple dimensions—primarily economic, ecological, and social [1]. These multifaceted pillars enable AHS to perform not only sustainable production functions but also roles beyond production, such as ecological conservation, cultural inheritance, and social cohesion. This multifunctionality manifests externally as multiple values spanning economic, ecological, socio-cultural, and livelihood domains [2,3,4]. AHS are thus of paramount importance in addressing major global challenges, including ensuring food security, protecting biodiversity, combating climate change, and safeguarding cultural diversity [5,6], thereby contributing to the implementation of the United Nations 2030 Agenda for Sustainable Development [7].

However, the ongoing advancement of urbanization and modernization has led to the gradual abandonment of traditional agriculture by farmers, driven by factors such as its relatively low comparative efficiency [8]. This trend has resulted in labor resource depletion, changes in traditional farming practices, and the erosion of agricultural culture at numerous AHS sites [9]. Consequently, reconciling the conservation of traditional agricultural production with the imperative of ensuring sustainable livelihoods for farmers has emerged as a core challenge in the sustainable development of AHS [10,11]. In response, dynamic conservation approaches, conceptualized based on the multifunctional value and living nature of traditional agriculture, are considered effective strategies for balancing these dual objectives. These approaches primarily encompass the development of eco-agricultural products, the promotion of sustainable tourism, and the fostering of industrial integration [12]. Nevertheless, the certification of eco-agricultural products in AHS sites often encounters low farmer participation, attributable to the transition period from conventional to ecological practices during which farmer incomes cannot be guaranteed [13]. The development of AHS tourism frequently compromises farmer interests due to inadequate participation mechanisms [14]. Similarly, while the impact of integrated development in traditional agriculture on farmer participation and benefits has been extensively theorized, it remains under-explored empirically [15]. Some studies and practical experiences even indicate that AHS conservation can negatively impact farmers’ livelihoods due to factors such as unreasonable development and utilization [16]. Therefore, determining how to fully leverage the multifunctional value of AHS to enhance farmers’ sustainable livelihoods has become a critical and urgent issue.

Tea-AHS refers to agricultural heritage systems centered on tea-planting, integrating tea culture throughout the entire process of production, processing, and consumption [17]. These systems are directly linked to tea cultivation and production, with tea leaves being their primary output. Tea-AHS has become a significant category within AHS, recognized for its outstanding economic and livelihood value. First, many tea culture systems embody the core values of AHS, leading to their designation as either China’s Nationally Important Agricultural Heritage Systems (NIAHS) or FAO’s Globally Important Agricultural Heritage Systems (GIAHS). Specifically, China hosts 23 Tea-NIAHSs and 4 Tea-GIAHSs, while there are 6 Tea-GIAHSs worldwide [18,19]. These systems share similar and well-defined key elements, most notably a strong reliance on farmer participation [20]. Second, as a cash crop, tea offers relatively higher returns compared to grain crops. In Tea-GIAHS sites, farmers typically manage tea plantations on a relatively large scale, which constitutes their primary livelihood source. Finally, the tea industry generally encompasses multiple segments—including cultivation, processing, sales, and consumption—forming a relatively long value chain. This allows farmers in Tea-GIAHS sites to engage simultaneously in multiple segments [21].

Farmers’ sustainable livelihoods serve as a primary criterion for evaluating AHS conservation, with extensive research examining the impact of such conservation efforts on livelihoods. In existing literature, AHS conservation primarily refers to the nomination of sites, their official designation, and the implementation of dynamic conservation measures [22]. The identification of AHS sites can promote county-level economic development by advancing grain processing industries, improving infrastructure, building AHS brand recognition, and fostering agricultural economic entities [23,24]. Furthermore, it enhances rural residents’ income levels through increased non-agricultural employment, industrial integration, agricultural technological advancement, the development of agricultural entities, and AHS brand building [25,26]. Additionally, AHS conservation strengthens agricultural resilience by leveraging rural financial support and attracting industrial and commercial capital to rural areas [27]. However, most previous studies have focused on single dimensions or specific indicators—such as livelihood capitals, income, or well-being—rather than providing a holistic assessment of farmers’ sustainable livelihoods levels within AHS sites. Moreover, existing research has primarily investigated the influence mechanisms of AHS conservation on rural residents’ income from a macro perspective, with limited evidence from micro-level analyses. Therefore, this study focuses on farmers from two Tea-GIAHS sites—the Anxi Tieguanyin Tea Culture System (ATTCS) and the Fuding White Tea Culture System (FWTCS) in Fujian Province, China. Employing mediating effect models, it examines the mechanism through which AHS conservation influences farmers’ sustainable livelihoods, aiming to provide theoretical and practical support for formulating effective AHS conservation and development policies. Compared to existing research, the marginal contributions of this paper are twofold. First, it offers a multidimensional and systematic assessment of farmers’ sustainable livelihoods in AHS sites. Second, it investigates the specific pathway through which AHS conservation enhances farmers’ sustainable livelihoods.

2. Theoretical Analysis and Research Hypotheses

Following the designation of AHS sites, governments typically enact protective and developmental measures and conduct continuous monitoring and evaluation. These actions aim to strengthen the conservation of key components and encourage participation from farmers and broader agricultural households [28,29]. This initiative leverages the multifunctional value and resource advantages of traditional agriculture to advance large-scale and industrialized farming practices, thereby enhancing the sustainability of farmers’ livelihoods (Figure 1).

2.1. The Impact of AHS Conservation on Farmers’ Sustainable Livelihoods

In China, the conservation criteria for AHS show a high degree of alignment with the core requirements of the Rural Revitalization Strategy. AHS conservation can act as a catalyst for comprehensive rural revitalization, thereby promoting socio-economic development in rural areas and increasing employment and income opportunities for farmers [30,31]. Research indicates that AHS conservation significantly impacts farmers’ livelihood development. Specifically, it positively influences the sustainable livelihoods of participating households by diversifying their livelihood pathways and activities, enriching their income sources, and safeguarding their rights related to conservation and management [32,33]. Empirical studies from various AHS contexts substantiate this positive direction. For instance, research on the Hani Rice Terraces demonstrated that conservation efforts led to measurable improvements in household income through agri-tourism and premium product markets [34]. Similarly, conservation initiatives in the Fenghuangdancong Tea Cultural System were found to enhance livelihood diversification and resilience among participating farmers [35].

Synthesizing previous research and practical experiences, AHS conservation primarily improves farmers’ sustainable livelihoods through three key approaches. First, it effectively enhances farmers’ capacity to cope with risk shocks. It facilitates the effective utilization of diverse traditional agricultural resources, techniques, and knowledge held by farmers. This directly promotes the accumulation and optimization of their livelihood capitals, thereby strengthening household resilience to risks and providing foundational conditions for implementing effective livelihood strategies [36]. Second, AHS conservation can provide farmers with more economic opportunities. It generates additional employment opportunities and income-generating avenues, promotes the effective deployment of household labor, diversifies farmers’ economic activities, mitigates agricultural operational risks, enhances livelihood stability, and thereby improves sustainable livelihoods [37,38,39]. Third, AHS conservation helps leverage the multifunctional value and resource advantages of traditional agriculture, thereby improving farmers’ living standards and quality of life. It enables farmers to capitalize on these aspects, facilitating the development of agri-cultural tourism, the creation of distinctive agricultural product brands, and the enhancement of agricultural management efficiency. Consequently, it diversifies farmers’ income sources, improves their quality of life, and directly promotes the enhancement of sustainable livelihoods [40,41].

Hence, Research Hypothesis 1 is proposed as follows:

Hypothesis 1.

The conservation of AHS has a significant positive impact on improving the sustainable livelihoods of farmers.

2.2. Analysis of the Mechanism by Which AHS Conservation Affects Farmers’ Sustainable Livelihoods

The productive function is fundamental to AHS and represents their key distinction from other heritage types. Accordingly, a primary objective of AHS conservation is to achieve the effective enhancement of this productive function. As a significant category of ecological agriculture, the scaling and industrialization of traditional agricultural operations are viewed as key aspects and developmental trends in realizing the productive functions of AHS [42].

2.2.1. The Scaling of Agricultural Operations

Farmers at many AHS sites typically operate on a comparatively small scale, characterized by high labor intensity and low mechanization levels. This results in relatively low comparative efficiency for traditional agriculture, making it difficult for farmers to meet their needs for sustainable livelihoods improvement solely through traditional production. However, traditional agricultural production and management models can yield a wider variety of products and higher-priced ecological goods. Moreover, scaled operation of traditional agriculture not only enhances agricultural economic efficiency but also boosts labor productivity, thereby reducing the labor input required [43]. Empirical evidence supports the positive livelihood outcomes of scaling. For example, a study on grassland ecological reward mechanisms found that scaling through cooperative management significantly increased herders’ income and improved grassland conditions, demonstrating a clear positive pathway [44]. Similarly, research on specialty crop systems indicates that scaled operations facilitate better access to technology, credit, and markets, leading to enhanced capital accumulation and livelihood security for smallholders [45].

On the one hand, the scaling of traditional agricultural operations directly promotes the accumulation of natural and cultural capitals. Economies of scale drive farmers to invest in production and living facilities, enhancing physical capital, while also strengthening production ties among them, boosting social capital, and facilitating access to government subsidies, thereby increasing financial capital. On the other hand, it releases surplus agricultural labor, facilitating a shift toward specialized agricultural livelihood strategies. This increases opportunities for farmers to engage in other income-generating activities, raising income levels. Simultaneously, it encourages farmers to optimize resource allocation, ultimately improving livelihood sustainability [46].

Based on this, Research Hypothesis 2 is proposed as follows:

Hypothesis 2.

The conservation of AHS can improve the sustainable livelihoods of farmers by promoting the scaling of agricultural operations.

2.2.2. The Industrialization of Agricultural Operations

In most AHS sites, farmers typically participate in relatively short traditional agricultural supply chains, with most primarily engaged in crop cultivation. This results in characteristics such as unstable yields and low product value-added, leading to significant natural and market risks in traditional agricultural production. Consequently, farmers face challenges in enhancing the sustainability of their livelihoods [47]. However, with the multifunctional development of traditional agriculture and the extension of industrial chains, an increasing number of farmers are engaging not only in cultivation but also in processing, sales, tourism, and other related sectors. Traditional agricultural production is gradually transitioning toward industrialized operations to meet farmers’ growing material and spiritual needs, achieve sustainable development in traditional agriculture, create more employment opportunities, diversify livelihood sources, and promote the improvement of sustainable livelihoods [48]. The positive effect is corroborated by empirical studies. Research on the industrialization of traditional tea systems showed that deeper integration into processing and branding activities directly and positively correlated with higher farm-gate prices and greater income stability for households [49]. Another study on agro-food value chains demonstrated that farmer participation in extended industrial activities (e.g., eco-labeling and direct sales) led to a significant increase in value capture and livelihood diversification [50].

On the one hand, increased participation in traditional agricultural supply chains directly promotes the accumulation of farmers’ human and cultural capitals. This enables more efficient utilization of physical capital, enhances the added value of agricultural products (improving financial capital), reduces transaction costs with enterprises and cooperatives (boosting social capital), and incentivizes farmers to continue agricultural cultivation, thereby enhancing natural capital [51]. On the other hand, increased participation absorbs more household labor into traditional agricultural operations, diversifying farmers’ livelihood strategies. This enhances opportunities for engaging in agricultural extension industries and independent entrepreneurship, enriches agricultural activities, improves operational profitability, and helps diversify operational risks, thereby elevating livelihood standards [52].

Therefore, Research Hypothesis 3 is proposed as follows:

Hypothesis 3.

The conservation of AHS can enhance the sustainable livelihoods of farmers by promoting the industrialization of agricultural operations.

3. Materials and Methods

3.1. Research Areas

The ATTCS and FWTCS are comprehensive agricultural production systems. Each system is built around traditional tea varieties—Tieguanyin and white tea, respectively—along with their associated cultivation practices, plant protection management, harvesting and processing techniques, and tea culture. These systems encompass the biodiversity fostered through production, the ecological functions they support, and the cultural and natural landscape features of their respective regions. The ATTCS and FWTCS were designated as China’s NIAHS in 2014 and 2017, respectively, and recognized as FAO’s GIAHS in 2022 and 2025, respectively. Among the existing six Tea-GIAHSs, they are the only oolong and white Tea-GIAHS, respectively. The ATTCS and FWTCS sites are located in Anxi County and Fuding City, Fujian Province, Southeast China, respectively (Figure 2). They share remarkably similar climatic, topographical, and soil conditions, both being quintessential tea-producing regions in Fujian Province. Specifically, Anxi County is the birthplace of Tieguanyin tea and China’s largest tea-producing county, while Fuding City is the world-famous origin of white tea and China’s top producer and exporter of white tea. The core zone of the ATTCS encompasses Lutian, Xiping, and Huqiu Towns, while the core zone of the FWTCS includes Guanyang, Diantou, Bailin, Panxi, and Taimushan Towns. Within the entire heritage region, these core zones are where traditional tea cultivation practices are most preserved and tea garden areas are relatively large. For instance, in 2019, the tea garden areas in Lutian, Xiping, and Huqiu Towns of Anxi County were 1903.93 ha, 4081.80 ha, and 4041.27 ha, respectively. Meanwhile, the figures for Guanyang, Diantou, Bailin, Panxi, and Taimushan Towns of Fuding City reached 2165.80 ha, 3029.53 ha, 2815.73 ha, 2412.73 ha, and 1779.20 ha in 2022, respectively. Furthermore, the tea industry functions as a pillar of the local economy in these regions, with tea-related income accounting for a substantial share of farmers’ earnings.

3.2. Data Sources

Data on farmers from both the ATTCS and FWTCS were collected in the fifth year following their designation as China’s NIAHS. Specifically, surveys for the ATTCS were conducted in November 2019 and July 2020, while those for the FWTCS were conducted in March and July 2023.

A multi-stage sampling method was employed across both studies. Given that AHS are inherently endangered—with their core elements concentrated in limited, often shrinking areas—a purposive sampling strategy targeting these elements is methodologically justified. To ensure the sample administrative villages were typical and relevant for studying AHS-livelihood linkages, the first stage involved purposive selection within the heritage core zones of each study area based on two key criteria: (1) the size of tea gardens area, and (2) the richness and distinctiveness of their tea cultural resources. These criteria directly align with the defining, endangered core elements of the Tea-GIAHS. Consequently, our sample captures the focal communities of Tea-GIAHS policy—where tea production and culture are predominant and where conservation interventions are most salient. We note that this targeted approach may limit the statistical representativeness of our sample to all villages but is appropriate for achieving the study’s objective of in-depth analysis within characteristic heritage contexts. Subsequently, from the official list of farming households provided by the village committee in each selected village, 10 to 20 households were randomly chosen to ensure unbiased selection at the household level.

The questionnaire survey was administered through face-to-face interviews. A team of trained enumerators, familiar with the local dialect and context, conducted the interviews. Each interview lasted approximately 20–40 min. To ensure data quality, the enumerators explained each question clearly, avoided leading responses, and reviewed completed questionnaires on-site for consistency and completeness before leaving the household.

In total, 402 valid household questionnaires were collected: 214 from households within the ATTCS (spanning 14 villages across 3 towns) and 188 from the FWTCS (covering 16 villages across 5 towns) (

Table 1. The sample distribution at two Tea-GIAHS sites.

Tea-GIAHS	Core Zone	Sample Villages	Sample Households
ATTCS	Lutian Town	3	42
	Xiping Town	8	131
	Huqiu Town	3	41
FWTCS	Guanyang Town	3	35
	Diantou Town	3	38
	Bailin Town	4	43
	Panxi Town	4	49
	Taimushan Town	2	23
Total		30	402

). The questionnaires covered three main areas: basic household information, sustainable livelihoods status, and recognition of and behaviors related to participation in AHS conservation.

Given that men are typically the primary decision-makers in agricultural production and generally have greater familiarity with input-output dynamics, the majority of respondents were male (89.55%). The largest age group comprised respondents between 46 and 65 years old (60.95%). In terms of educational background, those with a junior high school education formed the largest cohort (44.53%). Household sizes were predominantly in the range of 4 to 6 members (59.20%). Regarding labor and income dependence on the tea industry, 83.58% of respondents reported that tea-related activities accounted for 50% or more of their household labor input, while 71.14% indicated that tea income constituted over half of their total household earnings. Furthermore, a significant majority (76.62%) perceived the village committee as having a relatively high or very high level of enthusiasm toward heritage conservation efforts (

Table 2. The basic characteristics of sample households in Tea-GIAHS sites.

Items	Grouping	Proportion (%)
Gender	Female	10.45
	Male	89.55
Age (years)	≤35	7.46
	36~45	18.66
	46~55	32.34
	56~65	28.61
	>65	12.94
Level of education	Below elementary school	8.71
	Elementary school	21.14
	Junior high school	44.53
	High school	18.41
	College and above	7.21
Number of household members (persons)	1~3	13.93
	4~6	59.20
	7~9	17.66
	≥10	9.20
Tea industry labor force accounts for household labor force (%)	<10	0.25
	10~50	16.17
	50~90	39.80
	>90	43.78
Tea industry income accounts for household income (%)	<10	1.99
	10~50	26.87
	50~90	45.02
	>90	26.12
Village committee’s initiatives for conservation efforts *	Very low	1.49
	Relatively Low	1.74
	Average	20.15
	Relatively high	47.76
	Very high	28.86

* In rural China, the village committee is a mass grassroots organization elected by villagers. It serves as a platform for their self-management, self-education, and self-service.

).

3.3. Variable Specifications

3.3.1. Dependent Variables

Farmers’ sustainable livelihoods were evaluated based on a three-dimensional framework encompassing livelihood capitals, livelihood strategies, and livelihood outcomes [53]. This framework aligns with the widely recognized Sustainable Livelihoods Approach (SLA) and its adaptations in AHS contexts [54,55]. The selection of specific measurement indicators under each dimension (

Table 3. The indicator system and weighting for evaluating sustainable livelihoods of farmers.

First-Level Dimension Indicators	Second-Level Dimension Indicators	Third-Level Dimension Indicators	Weights
Livelihood capitals	Natural capital	Area of cultivated tea gardens	0.078
		Quality of cultivated tea gardens	0.006
		Plot conditions of cultivated tea gardens	0.019
		Distance between residence and cultivated tea gardens	0.019
	Physical capital	Housing structure	0.006
		Quantity of durable consumer goods	0.003
		Conditions of agricultural production tools	0.051
	Financial capital	Conditions of household savings	0.067
		Conditions of household debts	0.101
		Conditions of livestock and poultry	0.190
	Human capital	Quantity of family laborers	0.016
		Head of the household’s education	0.013
		Head of the household’s health condition	0.004
		Times of agricultural skills training	0.060
	Social capital	Interacted with neighborhood	0.003
		Joined a cooperative or association	0.089
		Served as a village official 1	0.099
	Cultural capital	Understanding of traditional agricultural knowledge	0.005
		Understanding of indigenous folk customs	0.006
		Tea planting years	0.011
Livelihood strategies	Livelihood diversity	The proportion of livelihood activity types among all livelihood activity types	0.014
	Livelihood dependency	The proportion of agricultural income in household income	0.008
Livelihood outcomes	Farmers’ income	Household income	0.040
		Per capita household income	0.038
	Farmers’ well-being	Income level of planters in their social circle	0.012
		Satisfaction with the scale of the tea gardens operations	0.033
		Satisfaction with tea planting benefits	0.009

¹ In rural China, village officials are local governance personnel tasked with organizing villagers and carrying out administrative and public service duties.

) was informed by both the SLA theoretical constructs and prior empirical studies on rural livelihoods and AHS evaluations [56].

To determine the relative importance of these indicators objectively and mitigate subjective bias in weighting, the entropy method was employed [57]. This method was preferred over alternative composite index construction techniques, such as Principal Component Analysis (PCA), or equal weighting, for several reasons relevant to our study context. First, unlike PCA, which transforms original variables and may complicate the interpretation of individual indicators, the entropy method retains the original indicator structure and is particularly suitable when indicators are not highly correlated. Second, compared to simple equal weighting, the entropy method accounts for the actual variability in indicator values across samples—assigning higher weights to indicators with greater variation, which often reflect more discriminative power among farmers’ livelihood conditions. Thus, it enhances the sensitivity and accuracy of the composite index in capturing disparities within the sample. This method calculates weights based on the information utility or variation degree of each indicator across all samples: a smaller entropy value indicates greater variation in the data for that indicator and thus assigns it a higher weight in the composite index. The resulting weights for our sample are presented in Table 3.

3.3.2. Core Independent Variable

The measurement of Tea-GIAHS conservation follows established research practices focused on farmer participation [58,59]. Consistent with these studies, we conceptualize conservation as a multi-faceted engagement process. It is measured through farmers’ self-reported participation levels in four key conservation activities: (1) conserving tea tree germplasm resources, (2) inheriting tea-making techniques, (3) participating in tea cultural activities, and (4) constructing ecological tea gardens. These activities collectively represent the core behavioral dimensions of on-farm heritage conservation identified in the literature [60].

A five-point Likert scale was employed [61]. Each activity is assessed on a 5-point scale (1 = None to 5 = A lot). To form a composite index of overall participation, the scores of these four indicators are averaged, assigning equal weight (0.250 each) to each dimension (

Table 4. The indicator system and weighting for evaluating farmers’ participation in Tea-GIAHS conservation.

Core Independent Variable	Indicators	Assignment	Weights
Farmers’ participation in Tea-GIAHS conservation	Conserving tea tree germplasm resources	1 = None, 2 = relatively little, 3 = average, 4 = more, 5 = a lot	0.250
	Inheriting tea-making techniques		0.250
	Participating in tea cultural activities		0.250
	Constructing ecological tea gardens		0.250

). This equal weighting scheme is adopted based on the rationale presented in the source studies [18], which argue that these four aspects are theoretically distinct yet equally vital components of holistic heritage conservation participation by farmers.

3.3.3. Control Variables

Drawing on existing research [62,63], control variables were selected based on three key dimensions: individual, household, and village characteristics. Individual characteristics include age, gender, and whether the respondent serves as a village official. Household characteristics comprise the dependency ratio and household size. Village-level controls include the village’s economic development level and its distance to the nearest town (Table 5).

3.3.4. Mechanism Variables

This study examines the scaling and industrialization of agricultural operations as mechanism variables [64]. Given that tea production is the dominant agricultural activity in Tea-GIAHS sites, and local governments often classify tea farmers into ordinary households and large-scale growers when promoting the industry and implementing policies (e.g., subsidies for expanded cultivation), whether a farmer operates at a large scale is used as a proxy for the scaling of agricultural operations (Table 5). Concurrently, to strengthen the tea industry’s role in employment and income generation through integrated development, local initiatives aim to extend the industrial chain. Farmers are encouraged through policies—such as subsidies for tea processing and e-commerce support—to engage in multiple segments of the tea value chain. Therefore, the proportion of segments in which a farmer operates serves as an indicator of the level of industrialization of agricultural operations (Table 5).

3.3.5. Instrumental Variable

To address potential endogeneity, this study employs an instrumental variable (IV) based on farmers’ self-assessed mastery of conservation-related knowledge and skills, which is used to predict their actual participation in Tea-GIAHS conservation [65]. Further, in our analysis, we assume that after controlling for observable household and village characteristics, any residual correlation between the instrument and the error term is minimal.

The description of each variable and the corresponding descriptive statistics are provided in Table 5. To assess potential multicollinearity, we calculated the variance inflation factor (VIF) for all selected variables. The results show that all VIF values are below 3, with a mean value less than 2, which is well below the common threshold of 10. This indicates the absence of severe multicollinearity, supporting the suitability of the variables for inclusion in the regression model analysis.

Table 5. Description of variables and descriptive statistics.

Variables	Name	Description	Mean	S.D.	VIF
Dependent variables	Farmers’ sustainable livelihoods	Calculated using the indicator system of farmers’ sustainable livelihoods	0.225	0.106	2.370
	Farmers’ livelihood capitals		0.208	0.113	2.230
	Farmers’ livelihood strategies		0.504	0.220	2.260
	Farmers’ livelihood outcomes		0.309	0.162	1.330
Core independent variable	Tea-GIAHS conservation	Calculated using the indicator system of farmers’ participation in Tea-GIAHS conservation	3.374	0.987	1.560
Control variables	Age	Respondents’ age (years)	52.766	11.223	1.290
	Gender	1 = Male, 0 = Female	0.896	0.306	1.070
	Served as a village official	1 = Yes, 0 = No	0.279	0.449	1.710
	Family dependency ratio	The proportion of household members under 18 and over 65 years of age	0.230	0.233	1.180
	Household size	Number of household members (persons)	5.771	2.556	1.250
	Level of village economic development	1 = Very low, 2 = Low, 3 = Moderate, 4 = High, 5 = Very high	3.229	0.775	1.120
	Distance between village and town	1 = Very far, 2 = Fairly far, 3 = Average, 4 = Fairly near, 5 = Very near	3.030	1.171	1.070
	Village dummy variable	1 = Yes, 0 = No
Mechanism variables	The scaling of agricultural operations	Is the respondent a large-scale tea grower 1: 1 = Yes, No = 0	0.246	0.431	1.270
	The industrialization of agricultural operations	Proportion of business operations across tea industry chain segments	0.285	0.188	1.460
Instrumental variable	Farmers’ assessment of their mastery of basic knowledge and professional skills for conservation participation	1 = Very unskilled, 2 = Fairly unskilled, 3 = Average, 4 = Fairly skilled, 5 = Very skilled	3.383	0.919	1.360

¹ In the ATTCS site, farmers with tea gardens exceeding 10 mu are classified as large-scale growers, whereas the threshold at the FWTCS site is 15 mu. (1 ha = 15 mu).

Table 5. Description of variables and descriptive statistics.

Variables	Name	Description	Mean	S.D.	VIF
Dependent variables	Farmers’ sustainable livelihoods	Calculated using the indicator system of farmers’ sustainable livelihoods	0.225	0.106	2.370
	Farmers’ livelihood capitals		0.208	0.113	2.230
	Farmers’ livelihood strategies		0.504	0.220	2.260
	Farmers’ livelihood outcomes		0.309	0.162	1.330
Core independent variable	Tea-GIAHS conservation	Calculated using the indicator system of farmers’ participation in Tea-GIAHS conservation	3.374	0.987	1.560
Control variables	Age	Respondents’ age (years)	52.766	11.223	1.290
	Gender	1 = Male, 0 = Female	0.896	0.306	1.070
	Served as a village official	1 = Yes, 0 = No	0.279	0.449	1.710
	Family dependency ratio	The proportion of household members under 18 and over 65 years of age	0.230	0.233	1.180
	Household size	Number of household members (persons)	5.771	2.556	1.250
	Level of village economic development	1 = Very low, 2 = Low, 3 = Moderate, 4 = High, 5 = Very high	3.229	0.775	1.120
	Distance between village and town	1 = Very far, 2 = Fairly far, 3 = Average, 4 = Fairly near, 5 = Very near	3.030	1.171	1.070
	Village dummy variable	1 = Yes, 0 = No
Mechanism variables	The scaling of agricultural operations	Is the respondent a large-scale tea grower 1: 1 = Yes, No = 0	0.246	0.431	1.270
	The industrialization of agricultural operations	Proportion of business operations across tea industry chain segments	0.285	0.188	1.460
Instrumental variable	Farmers’ assessment of their mastery of basic knowledge and professional skills for conservation participation	1 = Very unskilled, 2 = Fairly unskilled, 3 = Average, 4 = Fairly skilled, 5 = Very skilled	3.383	0.919	1.360

¹ In the ATTCS site, farmers with tea gardens exceeding 10 mu are classified as large-scale growers, whereas the threshold at the FWTCS site is 15 mu. (1 ha = 15 mu).

3.4. Model Design

3.4.1. Baseline Regression Model

In this study, farmers’ sustainable livelihoods are treated as a continuous variable. A multiple linear regression model is employed to estimate the impact of participation in Tea-GIAHS conservation on farmers’ sustainable livelihoods. The specific configuration of Model 1 is as follows:

$$S{L}_i={\alpha }_0+{\alpha }_{1}F{P}_i+{\alpha }_{2}contro{l}_i+{\epsilon }_i$$

(1)

In Equation (1), SL_i denotes farmers’ sustainable livelihoods. FP_i represents farmers’ participation in Tea-GIAHS conservation. The item control_i refers to a set of control variables, while i indexes individual farming households. α₀ is the constant, α₁ captures the net effect of participation on sustainable livelihoods, α₂ represents the coefficients for the control variables, and ε_i is the error term.

3.4.2. Mechanism Test Model

To investigate the mechanism, this paper incorporates mediating variables—specifically, the scaling and industrialization of agricultural operations—into the analysis. Models 2 and 3 are constructed as follows:

$$A{O}_i={\beta }_0+{\beta }_{1}F{P}_i+{\beta }_{2}contro{l}_i+{\mu }_i$$

(2)

$$S{L}_i={\lambda }_0+{\lambda }_{1}F{P}_i+{\lambda }_{2}A{O}_i+{\lambda }_{3}contro{l}_i+{\nu }_i$$

(3)

In Equations (2) and (3), AO_i denotes the mediating variable (scaling or industrialization of agricultural operations), while FP_i, SL_i, and control_i retain the same meanings as in Equation (1). β₀ and λ₀ represent are constants, and β₁, β₂, λ₁, λ₂, and λ₃ are coefficients to be estimated. μ_i and ν_i are error terms.

If the estimated coefficients α₁, β₁, and λ₂ are all statistically significant, then the scaling or industrialization of agricultural operations serves as a mediator. If mediation exists, the significance of coefficient λ₁ is examined: if λ₁ is not significant, it suggests full mediation; if significant, it indicates partial mediation [66].

4. Results

4.1. Baseline Regression

Table 6. Effects of Tea-GIAHS conservation on farmers’ sustainable livelihoods.

	(1)	(2)	(3)	(4)	(5)
Variable	SL	SL	Livelihood Capitals	Livelihood Strategies	Livelihood Outcomes
FP	0.040 ***	0.031 ***	0.030 ***	0.107 ***	0.043 ***
	(0.005)	(0.004)	(0.004)	(0.009)	(0.008)
Age		−0.001 ***	−0.002 ***	−0.003 ***	0.001
		(0.000)	(0.000)	(0.001)	(0.001)
Gender		0.018	0.021 *	−0.000	0.003
		(0.011)	(0.012)	(0.030)	(0.025)
Served as a village official		0.126 ***	0.141 ***	−0.029	0.034 **
		(0.009)	(0.010)	(0.021)	(0.017)
Family dependency ratio		0.002	0.005	−0.089 *	−0.029
		(0.018)	(0.020)	(0.046)	(0.036)
Household size		0.004 **	0.004 **	0.009 **	0.002
		(0.002)	(0.002)	(0.004)	(0.003)
Level of village economic development		0.019 ***	0.014 **	−0.015	0.048 ***
		(0.006)	(0.006)	(0.015)	(0.011)
Distance between village and town		−0.008 *	−0.008 *	0.028 **	−0.011
		(0.005)	(0.005)	(0.011)	(0.010)
Village dummy variable	Yes	Yes	Yes	Yes	Yes
Constant	0.086 ***	0.077 **	0.093 **	0.217 **	−0.066
	(0.025)	(0.036)	(0.039)	(0.085)	(0.066)
Observations	402	402	402	402	402
R2	0.198	0.506	0.507	0.343	0.230

Note: The values in parentheses represent robust standard errors; *** 1% significance, ** 5% significance, and * 10% significance.

reports the impacts of participating in Tea-GIAHS conservation on farmers’ sustainable livelihoods and its three dimensions. The coefficient of determination (R²) for each model ranges from 0.2 to 0.6, indicating that the selected independent variables adequately explain a substantial portion of the variation in the dependent variables.

Column (1) considers only the core variable of participation in Tea-GIAHS conservation. The coefficient of 0.040 (significant at the 1% level) provides an initial, uncontrolled estimate of its positive impact. Column (2) incorporates control variables for individual, household, and village characteristics. The coefficient for participation decreases to 0.031 but remains significant at the 1% level. This refinement indicates that approximately 22.5% of the initial estimate was attributable to correlated observable factors, and failing to account for these characteristics would overestimate the effect of Tea-GIAHS participation. The results confirm that Tea-GIAHS conservation effectively promotes improvements in farmers’ sustainable livelihoods, thereby validating Hypothesis 1.

To interpret the practical significance, we consider the estimated coefficient of 0.031 for participation (FP) in the full model of Column (2). Given that the dependent variable (SL) is a standardized index with a mean of 0 and a standard deviation of 1, a coefficient of 0.031 implies that participation in Tea-GIAHS conservation increases a household’s sustainable livelihoods index by 0.031 standard deviations, ceteris paribus.

Decomposing the overall sustainable livelihoods index into its three dimensions—livelihood capitals, strategies, and outcomes—columns (3)–(5) reveal that Tea-GIAHS conservation exerts a significant positive impact on all three aspects. The strongest association is with livelihood strategies, followed by livelihood outcomes and livelihood capitals, with coefficients of 0.107, 0.043, and 0.030, respectively. The impact on the sub-dimensions is particularly noteworthy for farmers’ livelihood strategies (Column 4), suggests that Tea-GIAHS participation is a major driver of shifts towards more sustainable and diversified livelihood strategies.

Regarding the control variables, households with older heads show a small but statistically significant reduction in sustainable livelihoods levels. The gender of the household head (where male = 1) shows a positive association with sustainable livelihoods. Households with members serving as village officials exhibit significantly higher sustainable livelihoods levels. The effects of the child-rearing and elderly support ratio (family dependency ratio) are generally insignificant, as many elderly individuals in the Tea-GIAHS sites remain economically active in tea production and childcare. Furthermore, larger household sizes contribute positively to sustainable livelihoods, likely due to enhanced labor availability and resilience to risks. In terms of village characteristics, both higher village economic development and closer proximity to towns are correlated with higher levels of household sustainable livelihoods, emphasizing the importance of external economic and market access.

4.2. Mechanism Tests

Building upon the baseline regression results, we employed Models 2 and 3 to examine the respective mechanisms through which the scaling and industrialization of agricultural operations influence the relationship between participation in Tea-GIAHS conservation and farmers’ sustainable livelihoods. The regression results are presented in

Table 7. Mechanism tests: The scaling and industrialization of agricultural operations.

	(1)	(2)	(3)	(4)
Variable	The Scaling of Agricultural Operations	SL	The Industrialization of Agricultural Operations	SL
FP	0.084 ***	0.026 ***	0.063 ***	0.021 ***
	(0.020)	(0.004)	(0.008)	(0.004)
The scaling of agricultural operations		0.060 ***
		(0.009)
The industrialization of agricultural operations				0.165 ***
				(0.022)
Control variables	Yes	Yes	Yes	Yes
Constant	−0.169	0.087 **	0.089	0.063 *
	(0.191)	(0.035)	(0.082)	(0.033)
Observations	402	402	402	402
R2	0.145	0.558	0.234	0.572

Note: The significance levels of 1%, 5%, and 10% are denoted by ***, **, and *, respectively. Robust standard errors are reported in parentheses.

, where columns (1) and (3) display the results from Model 2, and columns (2) and (4) show the results from Model 3.

As shown in Column (1) of Table 7, participation in Tea-GIAHS conservation exerts a significant positive effect on the scaling of agricultural operations, with a coefficient of 0.084 significant at the 1% level. This suggests that Tea-GIAHS conservation effectively facilitates the expansion of farmers’ agricultural activities. Column (2) reveals that participation in Tea-GIAHS conservation also has a statistically significant effect on sustainable livelihoods at the 1% level. Moreover, the scaling of agricultural operations is positively associated with farmers’ sustainable livelihoods, showing a coefficient of 0.060, which is also significant at the 1% level. These results indicate that the scaling of agricultural operations serves as a partial mediator in the relationship between participation in Tea-GIAHS conservation and farmers’ sustainable livelihoods. Specifically, Tea-GIAHS conservation can enhance farmers’ sustainable livelihoods by promoting agricultural expansion, thereby providing empirical support for Hypothesis 2.

As presented in Column (3) of Table 7, participation in Tea-GIAHS conservation has a significant positive impact on the industrialization of agricultural operations, with a coefficient of 0.063 significant at the 1% level. This implies that Tea-GIAHS conservation effectively facilitates the industrialization process of farmers’ agricultural activities. Furthermore, Column (4) indicates that participation in Tea-GIAHS conservation remains significantly associated with farmers’ sustainable livelihoods at the 1% level. Additionally, the industrialization of agricultural operations exhibits a positive effect on farmers’ sustainable livelihoods, with a coefficient of 0.165, also significant at the 1% level. These results suggest that the industrialization of agricultural operations serves as a partial mediator in the relationship between participation in Tea-GIAHS conservation and farmers’ sustainable livelihoods. In other words, Tea-GIAHS conservation can enhance farmers’ sustainable livelihoods by promoting agricultural industrialization, thereby providing support for Hypothesis 3.

4.3. Heterogeneity Analysis

4.3.1. Heritage Sites

To further examine whether the impact of AHS conservation on farmers’ sustainable livelihoods differs across AHS sites, this study conducted regression analyses separately for the ATTCS and the FWTCS sites. The results presented in

Table 8. Heterogeneity analysis: Heritage sites.

	ATTCS				FWTCS
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Variable	SL	Livelihood Capitals	Livelihood Strategies	Livelihood Outcomes	SL	Livelihood Capitals	Livelihood Strategies	Livelihood Outcomes
FP	0.034 ***	0.033 ***	0.060 ***	0.046 ***	0.034 ***	0.034 ***	0.061 ***	0.028 **
	(0.007)	(0.008)	(0.020)	(0.015)	(0.006)	(0.006)	(0.012)	(0.014)
Control variables	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Constant	0.099	0.106	0.365 **	−0.002	0.114 **	0.145 ***	0.237 **	−0.102
	(0.068)	(0.074)	(0.163)	(0.126)	(0.048)	(0.053)	(0.112)	(0.098)
Observations	214	214	214	214	188	188	188	188
R2	0.604	0.584	0.326	0.342	0.517	0.524	0.226	0.246

Note: The significance levels of 1% and 5% are denoted by *** and **, respectively. Robust standard errors are reported in parentheses.

.

As shown in Table 8, the estimated coefficient of the impact of participation in Tea-GIAHS conservation on farmers’ sustainable livelihoods is 0.034 in both heritage sites, significant at the 1% level. This indicates that the overall effect of Tea-GIAHS conservation on farmers’ sustainable livelihoods does not differ significantly between the two sites, supporting the robustness of treating them as comparable tea-related GIAHS in the main analysis. The ATTCS and the FWTCS in Fujian Province are typical representatives of oolong and white Tea-GIAHS, respectively. While there are similarities and differences in conservation models and policies between the two systems, the effects of heritage policies are generally comparable.

When examined by dimension, the influence of participation in Tea-GIAHS conservation on farmers’ livelihood capitals and livelihood strategies shows little variation between the two sites. However, its impact on livelihood outcomes differs more noticeably, with coefficients of 0.046 for the ATTCS and 0.028 for the FWTCS. This disparity is primarily attributed to the fact that the Anxi Tieguanyin tea industry has a longer development history, larger scale and production volume, and a more complete industrial chain encompassing cultivation, processing, and sales. Consequently, farmers in the ATTCS are more likely to benefit from participating in heritage conservation, thereby achieving higher income and well-being.

4.3.2. Livelihood Levels

To investigate the impact of participation in Tea-GIAHS conservation on farmers with varying levels of sustainable livelihoods, this study draws on the methodology employed in relevant research [62], while accounting for sample size limitations. Households were classified into high-level and low-level sustainable livelihoods groups based on the median value of sustainable livelihoods, and separate regressions were performed for each group. The results are reported in

Table 9. Heterogeneity analysis: Livelihood levels.

	(1)	(2)
	Low-Level	High-Level
Variable	SL	SL
FP	0.010 ***	0.024 ***
	(0.003)	(0.006)
Control variables	Yes	Yes
Constant	0.084 ***	0.194 ***
	(0.026)	(0.046)
Observations	201	201
R2	0.318	0.340

Note: The significance levels of 1% is denoted by ***. Robust standard errors are reported in parentheses.

.

As farmers’ sustainable livelihoods level increases, the estimated coefficient for participation in Tea-GIAHS conservation also becomes larger. This suggests that farmers with higher sustainable livelihoods levels are more likely to benefit from engaging in Tea-GIAHS conservation, whereas those with lower levels find it difficult to gain similar advantages. Tea-GIAHS conservation exhibits a “Matthew effect” in enhancing sustainable livelihoods for farmers. The disparity arises because farmers with greater livelihood assets are better positioned to participate in and benefit from conservation activities, owing to more developed capabilities and favorable conditions. In contrast, farmers with lower sustainable livelihoods often face constraints such as traditional farming techniques and limited experience, which hinder their ability to achieve meaningful benefits from participation in Tea-GIAHS conservation.

4.3.3. Income Levels

The capacity of households at different income levels to access and utilize AHS resources may vary. Accordingly, we further examine how participation in Tea-GIAHS conservation affects the sustainable livelihoods of households across income brackets. Based on median annual household income, sample households are divided into high-income and low-income groups. The variable “income level” is defined as 1 for the high-income group and 0 for the low-income group. Model 1 incorporates an interaction term between participation in Tea-GIAHS conservation and income level, and the regression results are presented in

Table 10. Heterogeneity analysis: Income levels.

	(1)	(2)	(3)	(4)
Variable	SL	Livelihood Capitals	Livelihood Strategies	Livelihood Outcomes
FP	0.021 ***	0.019 ***	0.105 ***	0.029 ***
	(0.004)	(0.004)	(0.010)	(0.008)
FP × Income level	0.017 ***	0.016 ***	0.003	0.021 ***
	(0.002)	(0.003)	(0.006)	(0.004)
Control variables	Yes	Yes	Yes	Yes
Constant	0.112 ***	0.126 ***	0.223 **	−0.022
	(0.035)	(0.038)	(0.087)	(0.066)
Observations	402	402	402	402
R2	0.569	0.557	0.344	0.272

Note: The significance levels of 1% and 5% are denoted by *** and **, respectively. Robust standard errors are reported in parentheses.

.

As shown in Column (1), the interaction coefficient between farmers’ participation in Tea-GIAHS conservation and income level is positive and statistically significant at the 1% level, indicating that Tea-GIAHS conservation has a stronger positive effect on the sustainable livelihoods of high-income farmers compared to their low-income counterparts. Similarly, the interaction term coefficients in columns (2) and (4) are also significantly positive at the 1% level, suggesting that Tea-GIAHS conservation enhances both livelihood capitals and livelihood outcomes more markedly for high-income farmers. This disparity may be attributed to the fact that higher-income households are more likely to participate in Tea-GIAHS conservation, possess higher educational attainment, engage more actively in conservation activities, and have access to more abundant heritage resources. As a result, Tea-GIAHS conservation exerts a more substantial impact on these households, leading to greater improvements in sustainable livelihoods.

In contrast, although the interaction term in Column (3) is positive, it is not statistically significant. This implies that the effect of Tea-GIAHS conservation on improving livelihood strategies does not differ significantly between high-income and low-income households. A possible explanation is that households within Tea-GIAHS sites tend to choose livelihood strategies based on a broader assessment of their individual conditions, policy support, and other contextual factors, where income level may not play a decisive role.

4.3.4. Livelihood Types

Households with different types of livelihoods are likely to engage in AHS conservation through distinct approaches. Accordingly, this paper assesses the impact of participation in Tea-GIAHS conservation on the sustainable livelihoods of households classified by livelihood type. Building upon the methodology employed in prior research practices [67], households are categorized as either agricultural or non-agricultural based on whether agricultural income accounts for at least 50% of total household income. Regression analyses were performed separately for each group, and the results are presented in

Table 11. Heterogeneity analysis: Livelihood types.

	(1)	(2)
	Agricultural Households	Non-Agricultural Households
Variable	SL	SL
FP	0.031 ***	0.022 ***
	(0.005)	(0.007)
Control variables	Yes	Yes
Constant	0.069	0.144 **
	(0.049)	(0.060)
Observations	301	101
R2	0.513	0.621

Note: The significance levels of 1% and 5% are denoted by *** and **, respectively. Robust standard errors are reported in parentheses.

.

As shown in Table 11, participation in Tea-GIAHS conservation exerts a significant positive impact on the sustainable livelihoods of both livelihood types. Notably, the estimated coefficient is larger for agricultural households and is statistically significant at the 1% level, suggesting that Tea-GIAHS conservation plays a more pronounced role in enhancing their sustainable livelihoods compared to non-agricultural households. This difference can be attributed to the fact that agricultural households typically possess richer traditional agricultural resources and knowledge, allowing them to derive greater benefits from participating in Tea-GIAHS conservation and thereby achieve more substantial improvements in sustainable livelihood outcomes.

4.4. Endogeneity Test

Participation in Tea-GIAHS conservation may enhance the sustainable livelihoods of farmers, while improved sustainable livelihoods could further encourage participation, creating a potential bidirectional relationship indicative of endogeneity. To address this, we employ an IV approach.

The instrument is “farmers’ self-assessed level of mastery regarding the professional knowledge and basic skills required for participating in heritage conservation.” Theoretically, this variable is closely related to actual participation, as involvement in conservation activities depends on relevant competency. Regarding exogeneity, we argue that as a perceptual measure specific to conservation capability, it aligns with the Theory of Planned Behavior (TPB) constructs of perceived behavioral control and attitude. Empirical studies grounded in TPB consistently show that such cognitive factors directly shape behavioral intention and subsequent action (e.g., participation in conservation practices), but their effect on final socio-economic outcomes (e.g., livelihood welfare) is channeled primarily through the enacted behavior itself, rather than through other direct pathways [68]. For instance, knowledge influences technology adoption, which in turn affects productivity and income, with no significant direct effect of knowledge on income once adoption behavior is accounted for [69]. Therefore, after controlling for observable covariates (e.g., basic household characteristics, village features), we contend that any potential correlation between our instrument and the error term—stemming from unobserved factors like innate ability—is minimal. This is because such latent traits are theorized to express themselves through, and thus be captured by, the participation behavior variable. Consequently, the instrument can be reasonably assumed to satisfy the exogeneity condition for identifying the causal effect of conservation participation on sustainable livelihoods.

As shown in columns (2) and (3) of

Table 12. The results of endogeneity test.

	(1)	(2)	(3)
Variable	SL(OLS)	FP(First Stage)	SL(IV)
FP	0.031 ***		0.049 ***
	(0.004)		(0.008)
Farmers’ assessment of their mastery of basic knowledge and professional skills for conservation participation		0.522 ***
		(0.048)
Control variables	Yes	Yes	Yes
F-statistic value of first stage		120.63 ***	120.63 ***
Cragg-Donald Wald F Test			119.511
Kleibergen-Paap rk LM			10.910 ***
Observations	402	402	402

Note: The significance levels of 1% is denoted by ***. Robust standard errors are reported in parentheses.

, the first-stage regression results using the instrumental variable indicate a statistically significant positive effect of the instrument on farmers’ participation in Tea-GIAHS conservation (p < 0.01), confirming that the instrumental variable satisfies the relevance condition. Moreover, the first-stage F-statistic is 120.63, rejecting the null hypothesis of a weak instrument. The Cragg-Donald Wald F-statistic of 119.511 exceeds the 10% critical value of 16.38, and the Kleibergen-Paap rk LM statistic of 10.910 is significant at the 1% level, further supporting the absence of a weak instrument problem. With only one instrumental variable, there is no need to examine the issue of over-identification. Compared with the baseline regression results in column (1) without instrumental variables, the IV estimates show that the coefficient of participation in Tea-GIAHS conservation on sustainable livelihoods increases from 0.031 to 0.049. After accounting for endogeneity, participation in Tea-GIAHS conservation continues to exhibit a statistically significant positive impact on sustainable livelihoods, reinforcing the conclusion that such involvement enhances farmers’ sustainable livelihood outcomes.

4.5. Robustness Tests

To ensure the reliability of the findings, we conducted robustness checks by modifying the measurement approaches of both the dependent and core independent variables, as well as by replacing the core independent variable. The results, presented in

Table 13. The results of robustness tests.

	(1)	(2)	(3)
Variable	SL (PCA Method)	SL (Entropy Method)	SL (Entropy Method)
FP (Averaging method)	0.163 ***
	(0.020)
FP (Entropy method)		0.119 ***
		(0.015)
Village committees’ participation in Tea-GIAHS conservation			0.020 ***
			(0.005)
Control variables	Yes	Yes	Yes
Constant	−0.844 ***	0.112 ***	0.112 ***
	(0.177)	(0.035)	(0.040)
Observations	402	402	402
R2	0.424	0.502	0.455

Note: The significance levels of 1% is denoted by ***. Robust standard errors are reported in parentheses.

, consistently support the main conclusions.

First, the measurement method of the dependent variable was altered. While the entropy method was initially used to quantify the sustainable livelihoods indicators, PCA was additionally employed as an alternative approach to re-quantify these indicators. As shown in column (1), participation in Tea-GIAHS conservation remains significantly and positively associated with farmers’ sustainable livelihoods, corroborating the initial results.

Second, the quantification method for the core independent variable was modified. Originally calculated using the arithmetic mean method, the variable “participation in Tea-GIAHS conservation” was re-quantified using the entropy method to maintain comparability and enhance validity. The results in column (2) confirm that the positive effect persists, further validating the robustness of the estimated relationship.

Finally, the core independent variable was replaced altogether. Instead of using individual household participation, the enthusiasm of the village committee toward Tea-GIAHS conservation was adopted as a proxy, considering the critical role of grassroots entities in conservation efforts and their broader impact on household livelihoods. As illustrated in column (3), this variable also exhibits a significant positive effect on sustainable livelihoods, confirming that the main findings are not sensitive to variable selection.

In summary, all three robustness tests reinforce the empirical conclusions of this study, affirming that participation in Tea-GIAHS conservation has a robust and positive influence on farmers’ sustainable livelihoods.

5. Discussion

Based on a comparative analysis of how AHS conservation affects farmers’ sustainable livelihoods in two distinct tea-related AHS sites—the ATTCS and the FWTCS [53]—this study validates and elaborates the mechanisms through which Tea-GIAHS conservation influences livelihood sustainability. The findings not only confirm but also extend previous understandings in several key aspects, offering insights that invite comparison with other AHS types.

First, consistent with the broader literature on AHS livelihoods [32], Tea-GIAHS conservation demonstrates a significant positive impact on farmers’ sustainable livelihoods. This study advances the discussion by operationalizing sustainable livelihoods through a comprehensive, three-dimensional framework—encompassing livelihood capitals, strategies, and outcomes—thereby enhancing the indicator’s scope and general applicability for future research.

Second, the analysis identifies the scaling-up and industrialization of traditional tea production as a core mechanism through which Tea-GIAHS conservation enhances livelihoods. This supports existing research highlighting scaling and industrialization as vital pathways for AHS sustainability [15,43]. A fruitful avenue for further discussion lies in comparing this mechanism with those observed in other AHS types. For instance, while tea culture systems may leverage product branding and vertical integration, paddy-field systems (e.g., rice-fish co-culture) might emphasize ecological synergies and landscape-based tourism [70], and folk-custom-based systems could rely more on cultural performance and artisan craftsmanship. Such a cross-type comparison would help clarify whether the drivers of livelihood improvement are context-specific or share commonalities across different AHS categories. These comparative insights carry direct implications for international GIAHS policy, suggesting that policy support could be categorized and tailored—for instance, facilitating market access and Geographical Indication (GI) branding for product-based systems like tea, promoting agro-ecotourism partnerships for landscape-based systems, and safeguarding intellectual property for craft and performance-based systems [71,72]. This advocates for a more typology-sensitive approach within the global GIAHS support framework.

Finally, this study effectively addresses a common methodological limitation noted in prior AHS case studies—often constrained by small sample sizes that hinder robust heterogeneity analysis [11]. By integrating survey data from two major Tea-GIAHS sites, we not only reveal significant differential impacts of conservation on distinct farmer subgroups but also provide a more reliable empirical basis for exploring the factors shaping livelihood sustainability. This approach offers a replicable model for future multi-site comparative research within and across different AHS typologies.

However, this study also has some limitations. First, research on other influence mechanisms and the heterogeneity of AHS conservation’s effects on farmers’ sustainable livelihoods remains insufficient. Beyond the scaling and industrialization of agricultural operations, conservation efforts may influence farmers’ sustainable livelihoods through other channels, such as agricultural production factor allocation [73] and agricultural management models [74]. Similarly, beyond livelihood status, income levels, and livelihood types, heterogeneity may also exist in other dimensions, such as labor force quality [75].

Second, the cross-sectional design adopted in this study imposes inherent limitations on causal inference. While we employ an IV approach to mitigate potential endogeneity concerns—such as reverse causality or omitted variables—the findings should be interpreted as robust associations rather than conclusive causal evidence. In addition, our chosen IV—farmers’ assessment of their mastery of basic knowledge and professional skills for conservation participation—may itself correlate with other unobserved livelihood attributes (e.g., intrinsic ability, social capital, or entrepreneurial motivation) that could also influence livelihood outcomes, potentially introducing residual confounding [76]. Therefore, future research using longitudinal data, panel surveys, or natural experimental designs would help to further validate the mechanisms and strengthen causal interpretation.

Third, research on the long-term influence mechanisms of AHS conservation on farmers’ sustainable livelihoods remains limited. As in previous studies, this research uses cross-sectional data to examine short-term influence mechanisms. Since short-term and long-term mechanisms may differ, further exploration of the long-term impacts is warranted.

6. Conclusions and Policy Recommendations

This study utilizes survey data collected from households in two Tea-GIAHS sites during the fifth year following their designation. By quantifying farmers’ sustainable livelihoods, it empirically assesses the impact of Tea-GIAHS conservation on their sustainable livelihoods, while further examining the roles of agricultural scale and industrialization in this process. The research also investigates the heterogeneous effects of Tea-GIAHS conservation on farmers’ sustainable livelihoods across four dimensions: heritage sites, livelihood status, income levels, and livelihood types. The main findings are as follows. First, Tea-GIAHS conservation has a significant positive impact on farmers’ sustainable livelihoods and its constituent dimensions, including livelihood capitals, strategies, and outcomes. Second, mechanism analysis reveals that Tea-GIAHS conservation enhances farmers’ sustainable livelihoods by facilitating the scaling and industrialization of traditional agricultural operations. Third, heterogeneity analysis shows that the positive effects of Tea-GIAHS conservation are more pronounced among households with higher livelihood levels, higher incomes, and those engaged in agricultural-based livelihoods, reflecting a “Matthew effect” in its influence on farmers’ sustainable livelihoods.

Based on the above quantitative findings, the following targeted policy recommendations are proposed. Given the currently insufficient subsidies for traditional tea cultivation at the farmer level and inadequate financial support for industrialized operations within Tea-GIAHS sites, it is essential to enhance support for large-scale traditional tea planting and continuously improve and extend the related industrial and value chains. At the same time, both the short-term effectiveness and long-term sustainable impact of policies must be taken into consideration.

First, increase investment to facilitate the scaling of traditional tea operations. The mechanism test identified a significant positive mediating effect of agricultural scaling. To activate this pathway, policy emphasis should be placed on scaled cultivation subsidies, processing support, and marketing assistance. These measures will directly encourage farmers to expand tea garden areas, thereby leveraging the verified mechanism to enhance sustainable livelihoods. Furthermore, to ensure the long-term sustainability of scaled operations, it is necessary both to develop rural infrastructure (such as irrigation and roads) to support larger-scale production and operation, and to foster tea farmer cooperatives or associations to reduce costs and enhance bargaining power.

Second, promote value chain development to advance the industrialization of traditional tea operations. The analysis identified an even stronger mediating effect through industrialization. In the short term, efforts should focus on advancing local eco-friendly tea brands and GIs, launching tea culture tourism projects through partnerships with local tourism operators, and facilitating basic vertical integration, such as connecting tea farms with local processing facilities. In the long term, the focus should shift to fostering deep-processing enterprises for high-value products (e.g., tea extracts and functional foods) and promoting cross-sectoral convergence with the tourism, cultural, and health industries. These strategies are directly supported by the quantitative evidence showing that industrialization is a powerful channel for improving livelihood outcomes.

Third, implement differentiated conservation and management policies tailored to various farmer types. The heterogeneity analyses provide clear guidance for targeting. For farmers with different livelihood levels, the significant “Matthew effect” indicates that farmers with higher initial livelihood levels benefit more. Therefore, complementary capacity-building and resource-access programs are crucial for low-level households to ensure equitable benefits and prevent widening disparities. For agricultural and non-agricultural households, given the larger coefficient for agricultural households compared to non-agricultural ones, support for the former should focus on improving quality and efficiency within existing tea production, while the latter may benefit from support for diversified operations within the tea sector. To build systematic resilience and equity safeguards, it is essential to strengthen community governance and benefit-sharing mechanisms, thereby ensuring that the long-term benefits of AHS conservation are distributed equitably among all farmer groups.