Policy evaluation in China’s national park system: integrating BERTopic modeling with a modular PMC framework

1 Introduction

Global biodiversity loss and climate change have evolved into interrelated, mutually reinforcing systemic ecological risks (IPBES, 2019; IPCC, 2022). National parks are a key regional conservation tool within the protection system, serving to safeguard large-scale natural ecosystems, biodiversity and critical ecological processes (Wu et al., 2023; Xu et al., 2022; Buenafe et al., 2025). Forest ecosystems provide a critical ecological foundation for national parks to conserve biodiversity, regulate climate, and safeguard water resources. National parks often encompass large areas and complex environmental gradients. Forest management within national parks should therefore extend beyond boundary-based protection and restrictive measures to include maintaining diversity in tree species, age structure, and spatial configuration; protecting natural and old-growth forests; restoring degraded habitats; retaining landscape heterogeneity generated by natural disturbance regimes; and sustaining functional connectivity among forest patches. Together, these measures can strengthen forest ecosystem resilience to drought, severe wildfire, pests and diseases, and extreme weather events (Mina et al., 2022; Kuuluvainen et al., 2021). Under accelerating climate change and increasing disturbance risks, forest management should no longer focus primarily on preserving fixed historical conditions. Instead, it should prioritize the maintenance of ecological processes, functional diversity, and adaptive capacity (Zhao et al., 2023; Li and Chen, 2024). This transition requires an iterative adaptive-management cycle that integrates long-term ecological monitoring, risk-based early-warning systems, scenario modeling, and the regular revision of management strategies.

In 2013, China incorporated the establishment of a national park system into its overall plan for comprehensively deepening reform; then, it has continued to advance pilot schemes, the formal establishment of national parks, and the development of the spatial layout for national parks (General Office of the CPC Central Committee, General Office of the State Council, 2017). According to the ‘Spatial Layout Plan for National Parks’ published in 2022, China has identified a total of 49 candidate areas for national parks, including the five national parks that had already been formally established. These areas cover a total area of approximately 1.1 million square kilometres, spanning the Qinghai-Tibet Plateau, the Yangtze River Basin and the Yellow River Basin, as well as major ecosystems such as forests, grasslands, wetlands, deserts and marine environments. This spatial layout has laid the foundation for the construction of a national park system (State Council of the People’s Republic of China, 2022).

Climate change challenges a central assumption of conventional spatial governance in national parks: that fixed boundaries will continue to encompass the habitats most important for conservation. Shifts in temperature and precipitation can alter species’ suitable ranges, vegetation zones, and natural disturbance regimes, meaning that critical habitats may move beyond existing park boundaries. Climate-adaptive conservation should therefore protect both current habitats and areas projected to remain or become suitable. It should also maintain or restore ecological corridors across elevational and latitudinal gradients and among different forest types, while coordinating conservation across park boundaries, adjacent buffer areas, collectively owned forestland, and other protected areas (Ranius et al., 2023; Pradhan et al., 2023; Gillingham et al., 2024). These spatial challenges are particularly relevant to four of China’s first five national parks: Giant Panda National Park, Northeast China Tiger and Leopard National Park, Hainan Tropical Rainforest National Park, and Wuyishan National Park. All four depend heavily on the integrity of forest ecosystems, and several extend across provincial boundaries. Species movements, wildfire spread, hydrological processes, and pest and disease outbreaks likewise transcend administrative jurisdictions. Effective conservation therefore requires governance arrangements that operate at the spatial scale of these ecological processes. Priorities include permanent cross-jurisdictional coordination mechanisms, harmonised monitoring standards, interoperable ecological information systems, joint law enforcement, coordinated emergency response, and ecological compensation arrangements across provinces.

Existing research has primarily focused on conservation objectives and governance performance. Relevant studies show that whether national parks can achieve their objectives of biodiversity conservation and ecosystem integrity depends not only on their protected area and legal status, but is also influenced by factors such as the quality of planning, financial and human resources, monitoring capacity, the rigor of law enforcement, and external anthropogenic pressures (Javeed et al., 2024; Nugraha et al., 2024; Lessmann et al., 2024). Research on China’s national parks has begun to develop an evaluation framework centred on ecological authenticity and integrity, monitoring of conservation outcomes, and adaptive management. These studies emphasise that the effectiveness of national park governance should be assessed across multiple dimensions, including ecosystem composition, structure, function, spatial patterns and social impacts (Jiao et al., 2022; Zhao and Yang, 2021). But, such research primarily focuses on the ecological or management outcomes resulting from the establishment of national parks, while paying relatively limited attention to the policy and institutional structures that shape these outcomes. Studies have examined the governance logic of China’s distinctive national park system from the perspectives of strengthened central authority, the restructuring of departmental responsibilities, the allocation of management and regulatory functions, community participation, and co-management (Zhao, 2022; Zhang, P., et al., 2023). Research generally suggests that China’s national park policies have promoted the integration and unified management of existing protected areas. Nevertheless, coordination challenges persist between central and local governments, between national park management agencies and territorial governments, and between ecological conservation and community development (Zhao et al., 2023; Ma et al., 2023). Community-based conservation, ecological compensation, and co-management can increase residents’ incomes and strengthen their recognition of and support for conservation. The effectiveness is influenced by rights to participation, benefit distribution, and local management capacity.

The ecological processes discussed above have direct implications for the design of national park policy. General conservation targets and restrictions on damaging activities are insufficient unless they are supported by operational rules for forest management. Such rules need to specify how ecological connectivity will be maintained and restored, how natural and old-growth forests will be protected, and how degraded forest areas will be rehabilitated. Additional provisions are required for wildfire and pest surveillance, climate-risk assessment, interregional coordination, stable long-term financing, and periodic policy review informed by monitoring evidence. Existing research has paid limited attention to whether and how forest-management and climate-adaptation mechanisms are reflected in national park policy design. Recent studies have used the PMC-Index model to evaluate national park policies issued at national and subnational levels (Li and Chen, 2024; Li et al., 2026). But the latent ecological-governance themes running through a broader corpus of policy documents remain insufficiently examined. The distinct roles played by different policy instruments in forest conservation and climate adaptation have also received little attention. High-level strategic documents, laws and regulations, and national park master plans serve distinct policy functions: setting the overall strategic direction, defining rights and obligations, and translating policy objectives into spatial plans and implementation measures, respectively. Applying a uniform set of indicators to all three document types risks treating differences in institutional function as differences in policy design quality. Conventional binary (0/1) coding methods have been unable to effectively distinguish between general statements and specific operational provisions. To account for this distinction, this study has designed a three-point rating scale (0, 0.5 and 1) for use within the modular PMC evaluation framework.

This study makes three four contributions. First, this study constructs a policy corpus covering national-level documents and policies related to China’s first five officially established national parks from 2013 to 2026. It conducts a full-corpus thematic analysis of 45 core policy documents and selects 13 representative policies for structural evaluation across three tiers—central top-level design documents, laws and regulations, and master plans. In doing so, the study broadens the temporal, hierarchical, and spatial coverage of national park policy research. Second, this study develops a modular PMC framework comprising a common core module and document type–specific modules. It further extends the conventional binary 0–1 scoring scheme to a three-level ordinal coding system of 0, 0.5, and 1. This approach helps reduce the structural bias that may arise when comparing policy documents with different functions. Third, this study links the relative prominence of policy themes derived from the BERTopic results with the strength of corresponding institutional provisions assessed using the modular PMC-Index framework. This comparison makes it possible to determine the extent to which attention to particular policy issues is matched by the completeness of their institutional design. Forth, The findings provide targeted evidence for optimizing the design, coordination, and implementation of national park policies.

2 Materials and methods

To systematically evaluate China’s national park policies, this study employs BERTopic to identify the core themes embedded in policy texts and utilises the PMC-Index to assess the structural completeness and internal consistency of representative policy documents (Figure 1).

This study has compiled policy documents from 2013 to 2026 at the national level, as well as those relating to China’s first five officially established national parks. The policy texts were primarily sourced from the Chinese Government Policy Document Repository, the Peking University Law Database, the National Forestry and Grassland Administration, and the official websites of the relevant provincial governments. ‘National Park’, ‘Giant Panda National Park’, ‘Northeast Tiger and Leopard National Park’, ‘Wuyishan National Park’, ‘Hainan Tropical Rainforest National Park’ and ‘Three-River Source National Park’ were used for full-text searches. By removing duplicate entries; following an initial screening of titles and abstracts to exclude documents such as notices, documents consisting solely of approvals, and those containing purely informational content; and then conducting a full-text assessment based on predefined inclusion and exclusion criteria, a policy corpus comprising 45 documents was ultimately constructed.

To ensure the relevance and comparability of the documents, we adopted the following selection criteria. First, only documents issued by the National People’s Congress and its Standing Committee, the State Council and its constituent departments, and the provincial governments of the provinces where the relevant national parks are located were retained. Secondly, priority was given to normative policy documents with strong legal or administrative force and clear implementation provisions, including laws, regulations, departmental rules and other normative documents; Documents primarily used for information dissemination or administrative communication, such as approvals, public notices and correspondence, were excluded. Third, the selected documents must directly concern the establishment, protection, governance or development of national parks. The dataset must cover the key policy documents relating to China’s first five officially established national parks. Finally, a total of 45 policy documents were ultimately selected for analysis.

2.1 BERTopic topic modeling

This study employs BERTopic for topic modeling and identification. Given that national park policy documents are characterized by dense policy-specific terminology, numerous compound concepts, and substantial semantic overlap, BERTopic is particularly well suited to uncovering latent themes and the semantic relationships among them (Grootendorst, 2022).

BERTopic was applied to the text units extracted from the 45 policy documents. Before semantic embedding, the policy text units were retained in their original sentence-level form and subjected only to basic text cleaning. Sentence embeddings were then generated using the paraphrase-multilingual-MiniLM-L12-v2 model. Model tuning began with repeated UMAP-HDBSCAN runs using alternative parameter combinations. The retained specification used n_neighbors = 25, n_components = 3, and min_dist = 0.0 for UMAP, together with min_cluster_size = 30 and min_samples = 3 for HDBSCAN. Selection did not rest on a single statistic. Topic coherence and diversity were considered alongside the outlier rate and the consistency of the clustering results across runs. Researchers familiar with management studies and national park conservation then examined the keywords and source passages to determine whether the topics supported defensible policy interpretations. Text processing differed between the embedding and topic-representation stages. Before embeddings were created, only light normalization was applied, preserving the semantic context of each text unit. Stopwords were removed later, when the c-TF-IDF representations were constructed. NLTK’s English list was supplemented with frequent corpus terms that contributed little to topic differentiation.

We reviewed the initial topics and consolidated them based on three criteria: semantic proximity, functional compatibility and consistency with the underlying policy text. To assess whether the expert-informed grouping was more semantically cohesive than expected by chance, we compared cosine similarities among the initial topic embeddings. Within-group similarity was calculated as the mean cosine similarity across all pairs of initial topics assigned to the same higher-order theme, whereas between-group similarity was calculated across pairs assigned to different themes. We then conducted a one-tailed Monte Carlo permutation test based on 100,000 random reallocations.

2.2 PMC index model

This study used PMC-Index to evaluate 13 policy documents (Table 1) (Ruiz Estrada, 2011). The 13 documents were not intended to constitute a statistically representative sample. Rather, they formed a purposively selected analytical subset covering three policy tiers within China’s national park policy system: national strategic documents, legal and regulatory instruments, and national park master plans. Taken together, the selected documents covered China’s first five national parks, established in 2021.

Table 1. Representative policies.

The first and corresponding authors independently screened and classified all 45 documents. For each document, they recorded the issuing authority, instrument type, normative authority, geographical scope, validity, and eligibility for one of the three policy tiers. Disagreements were resolved by consensus. To be included in the PMC analysis, a document had to: ① fall within one of the three policy tiers; ② have substantial normative authority or serve as a principal framework document within its tier; ③address national park protection and management as its primary regulatory or planning subject; ④ have been formally issued, adopted, or approved and remain in force as of the cut-off date; ⑤ be available in full text. Supporting documents relating solely to a single topic (such as financial support, ecological compensation, concession management or the ecological forest ranger programme) were excluded from the scope of the PMC analysis but were retained within the broader body of literature used for thematic modeling.

The availability of comprehensive park-specific regulations varied across the five national parks. Northeast China Tiger and Leopard National Park operates under a vertically integrated central administration, and no comprehensive park-specific provincial regulation was identified in either Jilin or Heilongjiang as of the cut-off date. By contrast, Wuyishan National Park extends across Fujian and Jiangxi, each of which has enacted a park-specific regulation. The regulatory subset therefore contains two documents for Wuyishan National Park but none for Northeast China Tiger and Leopard National Park. Consequently, the mean score for the regulatory tier represents the comprehensive regulations available for evaluation rather than all five national parks and gives greater weight to Wuyishan National Park.

At the national policy level, this study selected the ‘Overall Plan for Establishing a National Park System’ and the ‘Guiding Opinions on Establishing a System of Nature Reserves Centred on National Parks’; these two documents establish the fundamental institutional frameworks for the development of the national park system and the system of nature reserves, respectively. At the legislative level, the study selected the ‘Law of the People’s Republic of China on National Parks’ and five comprehensive local regulations that were in force as of the study’s cut-off date. At the master-planning level, this study selected the 2023–2030 master plans for the first five officially established national parks: Sanjiangyuan, Giant Panda, Northeast Tiger and Leopard, Hainan Tropical Rainforest and Wuyishan. These five master plans are similar in terms of their approval dates, planning periods, nature of the texts and basic structure, and thus possess strong horizontal comparability. They also cover a range of ecosystems—including alpine grassland wetlands, montane forests, temperate forests, tropical rainforests and central subtropical forests—as well as governance contexts such as cross-provincial management, species conservation, community development and the protection of natural and cultural heritage.

Drawing on relevant previous studies (Ruiz Estrada, 2011; Wang et al., 2026), this study develops a modular policy modeling consistency evaluation framework comprising a common core module and document type–specific modules. The framework is designed to reduce the structural bias that may arise when different types of policy documents are evaluated using the PMC approach. Conventional PMC studies generally employ binary coding. Although binary coding is transparent and readily replicable, it cannot adequately distinguish among principle-based statements, partial institutional arrangements, and relatively complete institutional chains. To improve the discriminatory capacity of the evaluation with respect to the institutional strength of policy provisions, this study assigns second-level indicators three ordinal scores: 0, 0.5, and 1 (Table 2).

Table 2. Scoring criteria for the PMC index.

In the PMC-Index model, second-level indicators are assigned scores of 0, 0.5, or 1 based on the content of each policy document. The common core module comprises six first-level dimensions, each consisting of four second-level indicators. The score of policy (p) for first-level dimension (i) in the common core module is calculated as follows:

X pi = 1 4 ∑ j = 1 4 x pij , i = 1 , … , 6

Therefore:

0 ≤ X pi ≤ 1

The common core PMC score is as follows:

PM C p core = ∑ i = 1 6 X pi

The theoretical range is defined as follows:

0 ≤ PM C p core ≤ 6

Each document type–specific module comprises three first-level dimensions, with each dimension consisting of two second-level indicators. The score of policy (p) for first-level dimension (k) in the document type–specific module is calculated as follows.

S pk = 1 2 ∑ l = 1 2 z pkl , k = 1 , … , 3

Where,

z pkl ∈ { 0 , 0.5 , 1 }

The specialized PMC score is as follows:

PM C p type = ∑ k = 1 3 S pk

The theoretical range is defined as follows:

0 ≤ PM C p type ≤ 3

The total modular PMC score for policy ppp is calculated as follows:

PM C p total = PM C p core + PM C p type

Therefore,

0 ≤ PM C p total ≤ 9

The values of the first-level dimensions are arranged in a PMC matrix, whose three-dimensional representation constitutes the PMC surface. Although scholars generally employ similar procedures to construct PMC surfaces, this study adopts the following approach. Nine points are positioned in three-dimensional space, with the (X)- and (Y)-coordinates determined by the row and column positions of the corresponding first-level dimensions in the PMC matrix and the (Z)-coordinate determined by their respective values. The surface generated by connecting these nine points constitutes the PMC surface.

PMC ( surface ) = [ X 1 X 2 X 3 X 4 X 5 X 6 X 7 X 8 X 9 ]

Because the theoretical maximum scores of the common core module, the document type–specific module, and the overall modular PMC-Index are 6, 3, and 9, respectively, their absolute scores are not directly comparable across modules. To facilitate cross-module comparisons, this study applies theoretical maximum-value normalization to rescale each score to the ([0, 1]) interval as follows:

N p core = PM C p core 6

N p type = PM C p type 3

N p total = PM C p total 9

The overall normalized score can also be expressed as follows:

N p total = 2 3 N p core + 1 3 N p type

Therefore, the overall normalized score is not a simple arithmetic mean of the normalized scores of the common core module and the document type–specific module. Instead, it is a weighted average, with the weights determined by the respective shares of the two modules in the nine first-level dimensions.

Drawing on previous research (Wang et al., 2026), the normalized scores are classified into five levels, as presented in Table 3.

Table 3. Policy evaluation levels based on the normalized PMC-index.

2.3 Integrated BERTopic-PMC analysis

This study then estimated the relative prominence of policy themes and assessed the strength of the corresponding institutional provisions.

• (1) The prominence of theme kk was calculated as follows:

S k = n k ∑ j = 1 n n j × 100 %

Where,

k denotes the higher-order theme being evaluated, k = 1,…, n_k = 1,…, n, and nn is the number of higher-order themes retained after consolidation;

j is the theme index used in the summation;

n k is the number of text units in the 45-document corpus assigned to the initial topics subsequently consolidated under higher-order theme kk;

∑ j = 1 n n j : n j is the total number of text units assigned to the nn retained higher-order themes;

S k represents the relative prominence of theme kk within the 45-document policy corpus.

• (2) Relative proportion of a theme within an individual core policy document

s pk = n pk ∑ j = 1 n n pj

Where,

p denotes the pth core policy document, p = 1, …, 13;

k denotes the kth theme;

n pk is the number of text units assigned to theme k in policy document p;

∑ j = 1 n n pj is the total number of text units in policy document p assigned to the n retained themes;

s pk represents the relative proportion of theme k within policy document p.

• (3) Standardised PMC score for an individual policy document

N p = PM C common , p + PM C specific , p 9

Where,

PMC common , p is the score for the common core module of policy document p, with a maximum score of 6;

PMC specific , pis the score for the document-type-specific module of policy document p, with a maximum score of 3;

N p: is the overall standardised PMC score for policy document p.

• (4) Theme-level institutional design strength

I k = ∑ p = 1 13 ( s pk × N p ) ∑ j = 1 13 s pk

Where,

s pk is the relative proportion of theme k within policy document p;

N p: is the standardised PMC score of policy document p;

s pk × N p is the product of the relative theme proportion and the policy design strength;

I k represents the institutional design strength associated with theme kk across the 13 core policy documents.

• (5) Four-quadrant classification

The median values across the n themes were used as relative cut-off points:

S ˜ = Median ( S 1 , … , S n )

I ˜ = Median ( I 1 , … , I n )

where S ˜ and I ˜ denote the median values of thematic prominence and institutional design strength, respectively (Table 4).

Table 4. Four-quadrant classification rules.

3 Results

3.1 Topic identification

BERTopic was applied to the segmented policy provisions, yielding 16 initial topics, as shown in Figure 2. Based on an examination of the keywords associated with each topic, semantically similar topics were consolidated, resulting in eight final topics, as presented in Table 5. The eight-theme grouping produced a mean within-group cosine similarity of 0.748, compared with a mean between-group similarity of 0.667, giving a difference of 0.081. In 100,000 random reallocations, the average similarity within groups was 0.674; while the 95 per cent confidence interval centred on the permutation distribution was 0.626–0.725. The observed value of 0.748 exceeds the upper limit of this interval. The results of the one-tailed Monte Carlo permutation test were statistically significant (p = 0.002), indicating that the observed grouping exhibits a higher level of intra-group semantic cohesion than would be expected from a random reallocation with the same distribution of group sizes. This result supports the semantic structure inherent in the ‘eight-theme framework’.

Table 5. Topic and keywords.

T1 can be interpreted as the regulation of the boundaries for the utilisation of natural resources within national parks and related development activities. The ‘Overall Plan for Establishing a National Park System’, established the principle of strict protection and proposed the demarcation of functional zones based on resource characteristics and management objectives, with the implementation of differentiated protection and management. Subsequent national park master plans and local regulations have further refined institutional arrangements concerning the approval and filing of construction projects, the construction and maintenance of infrastructure, and ecological conservation and restoration.

T2 highlights the central role of flagship species in national park conservation policy. The policy focus has gradually shifted from emergency conservation measures targeting individual rare and endangered species to a comprehensive conservation system encompassing flagship species, associated species, habitats and ecological corridors.

Key terms such as ‘management’, ‘institutions’, ‘ecosystems’, ‘concessions’, ‘rivers’, ‘forests’, ‘villages’ and ‘townships’ appear frequently in T3, primarily reflecting the integration of management functions and the coordination of spatial relationships. The ‘Overall Plan for Establishing a National Park System’ proposes integrating the management functions of nature reserves, with a unified body assuming responsibilities for ecological conservation, natural resource assets and concession management. The ‘Guiding Opinions on Establishing a System of Protected Areas Centred on National Parks’, issued in 2019, further promoted the integration and optimisation of protected areas, as well as their tiered management and zoned control. The ‘Law of the People’s Republic of China on National Parks’ clarifies the responsibilities and collaborative relationships between the competent national authorities, national park management bodies and local governments. T3 thus illustrates the process by which national park governance has shifted from decentralised management toward the integration of functions and cross-regional coordination.

T4 centers on tourism, ecological experiences, education, visitors, residents, industry, wellbeing, and wastewater, highlighting policy attempts to integrate ecological conservation with community development. Relevant policies permit nature education, ecological experiences and ecotourism on a moderate scale in areas consistent with conservation objectives. At the same time, measures such as visitor capacity limits, booking systems, designated recreational routes and improvements to environmental infrastructure are implemented to minimise the environmental impact of tourism activities. Local master plans further integrate gateway communities, ecological conservation posts, concession arrangements, ecological product certification and green industries into the development of national parks.

Ecological restoration defines T5, which brings together terms related to forests, rivers, wetlands, wildlife habitats, and human intervention. National park policies generally establish the principle of prioritizing natural regeneration while using necessary human-assisted restoration as a supplementary measure. Degraded forests, grasslands, wetlands, riverbanks, abandoned mining sites, and wildlife habitats are identified as key targets for ecological restoration.

Theme T6 outlines the monitoring, early warning and emergency response arrangements established by national parks to address various types of ecological risks. The relevant master plans, tailored to the ecological characteristics of each national park, cover forest and grassland fires, disease sources and epidemics, harmful organisms, invasive alien species, as well as meteorological and geological hazards. The ‘Master Plan for the Giant Panda National Park (2023–2030)’ also proposes the establishment of monitoring stations for disease sources and pests, as well as the construction of storage facilities for prevention, control and disaster relief supplies, a risk-level information database and emergency shelters. This reflects the evolution of ecological risk management from isolated prevention and control measures toward an integrated approach combining monitoring, early warning and emergency response.

The vocabulary associated with T7 centres on public finance, funding, investment, budgeting, performance assessment, compensation, expenditure, and insurance. The keywords for Theme T7 cover topics such as public finance, fund-raising, investment, budgeting, performance evaluation, compensation, expenditure and insurance. The 2017 ‘Overall Plan for Establishing a National Park System’ proposed the establishment of a diversified funding mechanism centred on fiscal investment, and set out guiding principles regarding ecological conservation compensation, fund management and supervision. The 2022 ‘Opinions on Several Fiscal Policies for Advancing the Development of National Parks’ further called for increased fiscal investment and better coordination, the refinement of ecological conservation compensation, the expansion of diversified funding channels, and the integration of budget performance management into the entire fund management process.

Theme T8 is characterized by terms such as ‘non-compliance’, ‘fines’, ‘illegal activities’, ‘remedial measures’ and ‘accountability’, indicating that the regulation of illegal activities, responsibility for ecological restoration and legal accountability have become key elements of national park policy.

3.2 PMC-based evaluation results

The general core module comprises six first-level dimensions and 24 s-level indicators, while the document-type-specific module comprises three first-level dimensions and six second-level indicators (Table 6).

Table 6. PMC Evaluation Indicator System.

PMC scores ranged from 5.875 to 7.625, with an average score of 6.558 and an average normalized score of 0.729. These results suggest that China’s national park policy is highly institutionalized.

Among the common core module, X1 received the highest score, reflecting the high level of comprehensiveness of the 13 policy documents regarding formal promulgation, scope of application, alignment with higher-level institutional frameworks, and the clarification of rights and obligations. X4 also scored highly, suggesting that China’s national park policy has established relatively systematic institutional arrangements in areas such as the demarcation of spatial boundaries, functional zoning, species conservation, ecological restoration and the management of resource utilisation. X3 and X6 also performed well, but there remain several issues concerning institutional coordination. Inter-departmental and inter-regional coordination often stays at principled cooperation or information sharing, while provisions for routine operational arrangements, dispute resolution mechanisms and joint decision-making procedures remain inadequate. And a complete and traceable closed-loop feedback mechanism linking policy objectives, specific tasks, responsible entities, financial resources, monitoring results and subsequent policy adjustments has also to be established. By contrast, X5 and X2 scored relatively low. The existing policy toolkit continues to place a marked emphasis on regulatory instruments such as prohibitive provisions, licensing management, land-use controls and legal liabilities, while the development of ecological compensation, benefit-sharing and community incentive mechanisms remains relatively underdeveloped. The policy objective framework is also predominantly qualitative in nature, with quantitative targets and outcome-oriented performance indicators still insufficient (Table 7).

Table 7. Results of PMC index.

As shown in Table 8, indicators assigned a value of 0 account for only 2.05 per cent of all observations, suggesting that there are few institutional elements that are entirely absent from the sample policy documents. Indicators assigned a value of 0.5 account for 0.497, which is slightly higher than the proportion of indicators assigned a value of 1. This implies that the primary challenges facing China’s current national park policy framework have shifted from whether a particular institutional element has been incorporated, to whether that element is supported by sufficiently well-developed quantitative standards, operational procedures, resource allocation and feedback mechanisms.

Table 8. Overall distribution of three-level institutional strength assignments.

As shown in Table 9, the normalized score for genre-specific functions in top-level design documents is 0.917, demonstrating relatively comprehensive provisions in areas such as strategic positioning, reform directions, institutional and governance mechanisms, and supporting policies. In contrast, the normalized score for the common core module was 0.583, reflecting that such documents remain largely principle-based as for quantitative targets, specific procedures, policy instruments and implementation feedback mechanisms. This characteristic primarily aligns with the functional positioning of this document type and should not be interpreted as indicating that its policy value is low. In the case of legal and regulatory documents, the normalized score for the common core module was 0.729, higher than the normalized score of 0.639 for document-specific functions. This shows that national park laws and regulations have established relatively comprehensive provisions regarding scope of application, governing bodies, spatial control, behavioural norms and ecological protection. But, there is still room for further improvement in document-specific legal functions such as law enforcement procedures, constraints on administrative discretion, legal remedies, supporting implementation measures and the adaptability of regulations. For master plan documents, the normalized score for the common core module was 0.800, the score for document-specific functions was 0.733, and the overall normalized score was 0.778. Master plans typically include functional positioning, spatial zoning, key projects and monitoring arrangements, and are therefore highly operational in nature. However, this does not mean that master plans take precedence over laws and regulations in terms of legal force or political significance.

Table 9. Mean PMC Scores by policy document type.

As depicted in Figure 3, the PMC surface for P1 exhibits relatively pronounced fluctuations, with notable localised depressions in the two dimensions of X2 ‘Policy Objectives and Public Value’ and X4 ‘Ecological Conservation and Resource Management’. P1 demonstrates relative strength in strategic positioning, unified management and overall institutional design; but, its provisions with regard to biodiversity targets, habitat conservation and quantitative performance evaluation remain relatively inadequate. The surface for P2 is generally higher and smoother, reflecting improvements in the alignment of regulatory and legal frameworks, policy objectives and public value, ecological conservation and resource management, as well as implementation safeguards and closed-loop policy feedback. But, both policies have only established partial institutional arrangements in areas such as collaborative governance, the alignment of tasks and resources, the operationalisation of ecological compensation mechanisms, and dynamic policy adjustments. Both policies scored 2.750 in the ‘Document Type-Specific Module’ of the PMC, indicating a high degree of continuity in their reform direction. But, both scored only 0.625 on the two dimensions X3 ‘Coordination of Governance Actors and Responsibilities’ and X5 ‘Coordination of Policy Instruments and Stakeholders’, suggesting that community participation, incentive and compensation mechanisms, and dynamic policy adjustment remain areas requiring joint improvement.

P3–P8 generally display moderate institutional maturity. P3 shows the most pronounced fluctuations in the PMC curve. Its relatively high scores for X1 and X3 reflect the overarching role of national legislation in terms of institutional authorisation, scope of application and governance arrangements. In contrast, X2 forms the deepest local trough, while the scores for X5, X6 and X9 are also relatively low, suggesting a relatively solid regulatory foundation but continuing shortcomings in stakeholder coordination, implementation feedback and supporting arrangements. The PMC surface for P4 is relatively smooth, with X1, X3, X4 and X6 forming a continuous high-value region, indicating good coordination between management responsibilities, ecological conservation and oversight mechanisms; however, the quantitative assessment indicators and implementation safeguards remain imperfect. The surface characteristics of P5 and P7 are similar; the high scores for X1, X3 and X4 indicate strong synergy between Fujian and Jiangxi Provinces in terms of the legal framework, collaborative governance and ecological conservation; whereas the relatively low scores for X2 and X6 reflect that performance feedback and dynamic adjustment mechanisms are not yet fully established. P6’s overall performance is relatively balanced, but its phased assessment indicators and supporting measures lack sufficient specificity. P8’s high scores are mainly concentrated in the X1 and X4 dimensions, while the scores for the X2, X5, X8 and X9 dimensions are relatively low (Figure 4).

The PMC scores for P9–P13 are generally high, with low values occurring only in isolated local areas; this indicates that the master plans for the five national parks have established a relatively consistent policy framework. High values are concentrated in regions X1, X4 and X7, reflecting the synergistic effects of support from three areas: legal regulation, ecological conservation and planning foundations; low values, on the other hand, are most commonly found in regions X5 and X9, indicating that there are still shortcomings in stakeholder coordination, implementation progress and adaptive management. P9 exhibits the greatest variation: while it has established robust institutional arrangements for ecological conservation and governance, it remains relatively weak in terms of quantified targets, compensation and incentive mechanisms, and dynamic adjustment mechanisms.

P10 has a high overall score, with good alignment in giant panda conservation, habitat connectivity and the sequencing of phased objectives; however, resource allocation and benefit-sharing require further improvement. P11 is similar to P10, but scores low on X6, indicating insufficient supervision, feedback and closed-loop implementation. P12 has the smoothest curve and the strongest coordination across dimensions, with only X9 being significantly low. P13 performs outstandingly in terms of rainforest conservation and planning objectives, but is weaker in inter-departmental coordination, community benefits and adaptive management (Figure 5).

3.3 Theme-level institutional design strength results

Table 10 and Figure 6 show that the relative prominence of themes across the 45-document policy corpus is not consistently matched by the institutional design strength of the corresponding themes in the 13 core policy documents. Instead, the relationship between policy attention and institutional design varies across the four quadrants. More specifically,

Table 10. Thematic prominence, institutional design strength, and four-quadrant classification of the eight themes.

High prominence–strong institutional design (T2 and T4). T2 is prominent and is addressed primarily in national park master plans and park-specific regulations that received comparatively high PMC scores. Considered alongside the relevant policy provisions, this pattern suggests that flagship species conservation, habitat protection, and ecological connectivity are supported by relatively well-developed institutional arrangements in the core policy documents. T4 also lies above the median for both thematic prominence and institutional design strength. Ecotourism, nature education, and community livelihoods receive comparatively detailed treatment in the relevant planning provisions. However, T4 lies only marginally above the median for thematic prominence and is therefore better treated as a borderline case within the high-prominence quadrant.

High prominence–weak institutional design (T1 and T3). T1 and T3 are addressed mainly in high-level strategic documents, general regulatory instruments, and administrative policy documents with moderate or comparatively low standardised PMC scores. A closer examination of the relevant provisions shows that the policies establish broad principles for natural resource conservation and use, governance responsibilities, and intergovernmental coordination. However, they provide less detail on cross-jurisdictional coordination procedures, the division of authority and responsibility between central and local governments, the allocation of resources in line with assigned responsibilities, dispute-resolution procedures, and mechanisms for monitoring, policy review, and accountability. T1’s institutional design strength lies only marginally below the median and should therefore be treated as a borderline case within this quadrant. T3 lies further below the median, indicating a clearer gap between its prominence and the level of institutional specification.

Low prominence–strong institutional design (T5 and T6). T5 and T6 account for a below-median share of the text units in the 45-document policy corpus. Nevertheless, a review of the core plans and park-specific regulations in which these themes appear shows comparatively detailed provisions on implementation tasks, restoration projects, responsible authorities, monitoring procedures, emergency arrangements, and resource commitments. Their position in this quadrant therefore appears to reflect the concentration of detailed provisions in a limited set of specialized documents, rather than an absence of institutional design. Low thematic prominence should not, however, be treated in itself as evidence of inadequate policy attention. Ecological restoration is often addressed through project-based planning, while forest-disaster preparedness and response may be concentrated in specialized or risk-specific instruments. The findings do not indicate an immediate need for wholesale redesign. Instead, future policy review could examine whether these provisions are consistently reflected across relevant policy tiers, regularly updated in response to monitoring evidence, and supported by effective implementation.

Low prominence–weak institutional design within the core-policy subset (T7 and T8). This result should not be interpreted as evidence that the relevant policy fields lack formal rules. In the case of T7, fiscal support and ecological compensation are addressed through dedicated national instruments that were retained in the broader 45-document corpus but excluded from the PMC assessment of the 13 core policies. Its position in this quadrant therefore reflects the dispersion of financial and compensation arrangements across specialized policy instruments, together with their limited integration into core strategic, regulatory, and planning documents. T8 also requires a qualified interpretation. The remaining concern is less the absence of formal provisions than their uneven specification across policy tiers and the limited operational links among monitoring results, policy revision, corrective action, enforcement, and accountability. T7 and T8 should therefore not be treated simply as fields requiring new institutional frameworks. Greater priority should be given to integrating existing specialized arrangements with core policies and improving coordination among funding, performance assessment, enforcement, remedial action, and policy feedback.

To assess the robustness of the four-quadrant classification, two one-at-a-time sensitivity checks were conducted. First, the corpus-wide prominence measure derived from all 45 policy documents was replaced with the equal-weighted mean of the document-level thematic proportions across the 13 core policies. Second, each document’s overall standardised PMC score was replaced with its standardised common-core-module score, thereby excluding the contribution of document-type-specific modules. In each check, only one component was changed, while the other was retained in its baseline form, and the median cut-off values were recalculated. All eight themes retained their above-or below-median status on both dimensions and remained in their original quadrants under both alternative specifications.

4 Discussion and implications

4.1 Discussion

4.1.1 Greater attention should be paid to the quality and effectiveness of institutional implementation in future research

The findings reveal that the basic framework for China’s national parks has largely taken shape. It is moving from a focus on the broad coverage of institutional elements to a greater emphasis on the quality and effectiveness of institutional implementation. This interpretation is broadly consistent with patterns reported in international research on protected-area management effectiveness. Leverington et al. found that, although many protected areas have attained statutory status and established planning frameworks, significant shortcomings remain regarding resource allocation, management processes, community relations and performance evaluation (Leverington et al., 2010). Watson et al. similarly pointed out that an increase in the number and spatial extent of protected areas does not necessarily lead to improved conservation outcomes. In contrast, management capacity, adequate funding, effective law enforcement and ongoing evaluation are the key factors determining whether protected areas can achieve their intended objectives (Watson et al., 2014). Moreover an analysis by Coad et al. of 2,167 protected areas showed that insufficient staffing and financial resources significantly constrain their ability to provide effective protection (Coad et al., 2019). The comparison shows that the relative prominence is not consistently matched by institutional design strength. The clearest discrepancies occur where highly prominent themes are accompanied by comparatively weak institutional provisions, or where detailed institutional arrangements are concentrated in themes with lower prominence. Future policy revisions should therefore respond to these different patterns rather than apply a uniform approach across all themes. Depending on the evidence for each theme, priorities may include consolidating established arrangements, adding operational detail, extending coverage across relevant policy instruments, or addressing clearly identified gaps in policy design.

4.1.2 Regulatory instruments are more developed than participatory and adaptive mechanisms

The provisions of these policies regarding regulatory instruments are significantly more detailed than those concerning participation, evaluation and adaptive mechanisms; this may be related to the phased objectives of China’s national park reform. In the early stages of the reform, the focus was on consolidating overlapping protected areas, harmonising standards, demarcating boundaries and restricting incompatible development; consequently, spatial zoning, land-use controls, administrative licensing and law enforcement measures were more readily prioritised for institutionalization. Participation mechanisms, on the other hand, rely on transparent selection of representatives, clearly defined resource rights, benefit-sharing, consultation and dispute resolution procedures; these entail higher coordination costs and must also accommodate differences among national parks in terms of socio-economic conditions, resource ownership and governance capacity.

Climate change has further exacerbated this imbalance in the system. As species distributions, vegetation dynamics and disturbance patterns continue to shift, fixed boundaries and static zoning are no longer sufficient to maintain the long-term integrity of ecosystems. National park policies should focus simultaneously on both current and future suitable habitats, maintain ecological connectivity across elevation gradients, forest landscapes and between adjacent protected areas, and coordinate the management of wildfires, pests and diseases, and cross-administrative ecological corridors at the landscape scale (Wiegant et al., 2022; Beazley et al., 2023; Hurtado et al., 2025). These tasks rely not only on spatial management but also require long-term monitoring, inter-governmental coordination, stakeholder participation and regular policy revisions. Existing policies have established a relatively comprehensive framework for spatial management and administrative oversight; however, specific provisions regarding participatory consultation, performance evaluation, feedback and adaptive adjustments remain lacking.

4.1.3 Community-related issues receive policy attention, but participation remains centred on compensation, employment, and project-based involvement

An equitable distribution of economic and non-economic benefits can improve local support for conservation and strengthen policy legitimacy (Poudyal et al., 2024; Zhang et al., 2025; Oldekop et al., 2016; Naughton-Treves et al., 2005). T4 lies above the median on both dimensions, although only narrowly so for thematic prominence. Its positioning is primarily influenced by livelihood projects, employment related to ecological conservation, compensation mechanisms and benefit-sharing provisions, rather than stemming from the community’s full right to participate in policy or institutional decision-making; this is consistent with the findings of existing research (He, 2024; Wang, P., et al., 2023; Zhang, Y., et al., 2023). The community co-management systems in Australia’s Kakadu, Uluru-Kata Tjuta and Budri National Parks offer some useful insights (Ross et al., 2009; Strickland-Munro and Moore, 2013). However, this does not necessarily imply that communities will have an equal say (Reardon-Smith, 2024). Consequently, future reforms should give greater consideration to the role of communities in resource monitoring, rule-making, impact assessment and standard-setting.

4.1.4 Institutional adaptation and the need for differentiated governance in different ecosystems

There are marked differences in the ecosystems, conservation targets and community resource utilisation patterns across the five national parks; consequently, similar policy shortcomings may have varying impacts in different national parks, necessitating the establishment of targeted institutional mechanisms. The Sanjiangyuan National Park faces a complex interplay of issues relating to climate change, permafrost degradation, hydrological processes, grassland conservation and herders’ livelihoods (Jin et al., 2022). The Sanjiangyuan National Park faces a complex interplay of issues relating to climate change, permafrost degradation, hydrological processes, grassland conservation and herders’ livelihoods (Jin et al., 2022). On pages 6 and 9, further steps should be taken to establish a comprehensive monitoring mechanism based on ecological thresholds, so that monitoring results can trigger adjustments to grazing practices, ecological restoration and revisions to compensation standards. The main issues facing the Giant Panda National Park are habitat fragmentation and insufficient population connectivity (Diao et al., 2025). P4 and P10 should further clarify the responsibilities, cost-sharing arrangements and restoration standards of the three provinces—Sichuan, Shaanxi and Gansu—in the construction of ecological corridors, and coordinate collective forest rights and community interests through ecological compensation or conservation easements. The Northeast Tiger and Leopard National Park needs to coordinate cross-border migration, prey restoration, the removal of hunting traps and the management of human–wildlife conflicts (Qi et al., 2021; Li et al., 2024). On page 11, population monitoring should be linked to management adjustments, and a governance mechanism should be established that includes risk early warning, compensation for losses, insurance and alternative livelihoods. Wuyishan National Park requires coordination of cross-provincial forest and watershed conservation with tea garden management, collective forest rights and tourism utilisation (Wang, B., et al., 2023). In P5, P7 and P12, ecological monitoring standards should be harmonised, and procedures for data sharing, joint law enforcement and rolling revisions should be improved; furthermore, ecological tea garden standards and benefit-sharing mechanisms should be utilised to reconcile conservation with community livelihoods. Hainan Tropical Rainforest National Park faces challenges relating to rainforest fragmentation and the recovery of critically small populations, such as the Hainan gibbon (Han et al., 2024). In P8 and P13, efforts should be stepped up to restore ecological corridors and the forest canopy, improve comprehensive monitoring and disaster early warning systems, and safeguard community rights through the clarification of land tenure, co-management and long-term compensation. Consequently, China’s national park policy should, on the basis of unified conservation standards, allocate differentiated institutional tools according to each park’s ecological processes, cross-regional relationships and community resource utilisation patterns.

4.1.5 Integrating centralised management with coordinated polycentric governance

This study has revealed three inconsistencies: regulatory instruments are given greater priority than participatory mechanisms; the broad scope of policy coverage is not matched by robust feedback mechanisms; and coordination mechanisms at central government level are not always closely aligned with local and community governance.

Socio-ecological systems often require multiple, interconnected decision-making centres (Ostrom, 2009). Effective governance of national parks necessitates close cooperation between the central government (ministries), cross-jurisdictional government bodies, park management authorities and local communities. The central government or ministries set standards for the conservation and development of national parks; cross-jurisdictional bodies coordinate responses to ecological processes; park management authorities provide specialist management services; local governments deliver public services; while local communities participate in relevant decision-making and monitoring activities. Mechanisms for information sharing, consultation, conflict resolution and accountability should be established between the various levels to ensure the effective functioning of the multi-centred governance system (Datta and Chaffin, 2022).

This study suggest that policy focus and institutional design do not necessarily evolve in tandem; the discrepancies between the two reflect gaps in the process of translating policy priorities into specific operational arrangements: while some high-profile issues remain poorly defined in terms of specific implementation rules, other issues—despite being supported by detailed provisions—have not yet been fully integrated into the broader policy framework. In the next phase, national park governance should place greater emphasis on optimizing vertical policy implementation mechanisms, strengthening coordination and collaboration across jurisdictions and between policy instruments, and refining the feedback mechanisms between the implementation process and policy revisions.

4.2 Implications

4.2.1 Institutionalise cross-regional coordination

For national parks extending across provincial boundaries, the priority is not simply to create additional coordinating bodies, but to establish stable working arrangements among central government agencies, park authorities, and relevant provincial and local governments. A permanent coordination mechanism or regularly convened intergovernmental meetings should clearly define the participating bodies, the rules for agenda-setting and decision-making, the division of responsibilities, arrangements for the sharing of funds and data, and procedures for dispute resolution and accountability. Transboundary ecological processes and ecological risks—including wildlife migration, river basin conservation, forest fire and pest control, and ecological restoration—likewise require uniform monitoring standards, joint work plans, shared information systems and coordinated emergency response procedures. These arrangements will closely link national-level conservation standards with cross-jurisdictional coordination and local implementation.

4.2.2 From project-based participation to institutionalized community co-governance

Effective community participation requires ensuring that communities have a proper say in the decision-making process for national parks—particularly where their access to land and resources, as well as their livelihoods, are directly affected. Community participation may begin at the stage of drafting or revising the master plan and continue throughout the entire process, encompassing decisions on land-use zoning, the management of concessions and ecotourism, compensation for wildlife damage, ecological restoration and conservation assessments. The institutional arrangements adopted by different parks need not be identical: some parks may establish representative committees, while others may adopt consultative co-management agreements or participatory monitoring mechanisms. Regardless of the arrangement adopted, the relevant rules must clearly specify the method of selecting representatives, the length of their terms of office, the scope of their powers, and how the park management authority will respond to recommendations put forward by the community. When decisions concern resource rights or livelihoods, the community must be provided with sufficient information before the final decision is taken, and viable channels for consultation, objection, review and oversight must be established.

4.2.3 Optimizing the policy instrument mix to close the adaptive management feedback loop

Policy refinement should now focus on how existing instruments work together within a practical adaptive-management cycle. Long-term monitoring needs to track the condition of natural forests, particularly remaining primary stands, as well as degraded habitats, variation in forest structure and function, ecological connectivity, and exposure to wildfire, pests, and disease. Disaster governance should follow the same logic, linking risk assessment and early warning with cross-jurisdictional preparedness, emergency response, post-event review, and ecological recovery. Implementation documents also need to show how funds correspond to defined tasks, how monitoring evidence enters policy and budget reviews, and which bodies are responsible for corrective action. Where fiscal rules allow, multi-year funding arrangements could reduce interruptions to restoration, species monitoring, and disaster prevention. Regular publication of performance results, followed by a formal requirement to respond, would connect monitoring and evaluation with resource reallocation, policy revision, and accountability.

5 Conclusion, limitations and directions for future research

5.1 Conclusion

Based on a corpus of 45 national park policy documents issued in China between 2013 and 2026, this study utilises BERTopic and the modular PMC-Index for topic identification and policy evaluation.

The results of the policy theme analysis indicate that China’s national park policy framework not only encompasses the integration of protected areas and ecological management, but also extends to areas such as ecological restoration, community livelihoods, ecotourism, financing, disaster management and collaborative governance. However, there are variations in the prominence of each theme: the conservation of flagship species, the regulation of resource utilisation and governance coordination received the most attention, while fiscal support, performance monitoring, legal accountability and forest disaster management were given relatively less coverage. These differences reflect the relative distribution of policy focus within the analysed literature.

The PMC evaluation results indicate that the core institutional framework of China’s national park system has largely taken shape. The 13 documents scored highly in areas such as the legal foundation, ecological conservation, spatial control, resource management and project planning. These results highlight the system’s strengths in terms of centralised coordination, uniform conservation standards, and spatial planning and control. However, scores were relatively low in terms of the alignment between policy objectives and public values, as well as the integration of stakeholder perspectives into policy instruments.

A combined analysis indicates that there is a certain degree of structural mismatch between thematic prominence and the robustness of institutional design. Themes T2 and T4 exhibit high thematic prominence and well-developed institutional frameworks, although T4 is close to the thematic prominence threshold; their positioning is primarily driven by provisions relating to compensation mechanisms, sheltered employment, ecotourism and community development, rather than by broad, formal rights for community participation in decision-making. T1 and T3 are prominent across the entire sample, but their specific provisions in core policies are relatively brief, particularly with regard to cross-jurisdictional coordination procedures; by contrast, T5 and T6 account for a smaller proportion of the sample, but their content is supported by detailed provisions in relevant plans and regulations. The findings regarding T7 and T8 should be treated with caution, as the issues they address—such as financing, ecological compensation, performance monitoring and legal accountability—are also covered in specialized legal documents outside the core sample; consequently, their position within the sample reflects the low degree of integration of these elements into the core policy framework, rather than an institutional gap.

The three types of document each fulfil distinct yet complementary functions; however, the vertical policy translation between them remains inadequate. Some strategic objectives have not yet been translated into clearly defined tasks, responsibilities, resources and timetables; furthermore, the level of detail in legal and planning documents regarding technical implementation, monitoring, revision and accountability mechanisms varies considerably. Consequently, the primary shortcoming lies not in a lack of policy objectives, but in the fact that these objectives have not yet been fully translated into specific arrangements applicable at all policy levels and conducive to implementation and review.

Overall, China’s national park policy framework places greater emphasis on regulatory instruments than on participatory ones. while the institutional coverage is extensive, the mechanisms for monitoring, evaluation, policy adjustment and accountability are not yet fully integrated, and the interplay between central coordination and local or community participation remains uneven. Consequently, further institutional development may depend less on the introduction of new policy elements and more on strengthening vertical coordination, cross-jurisdictional collaboration, implementing feedback mechanisms and promoting meaningful community participation.

5.2 Limitations and directions for future research

The sample used in this study does not yet provide comprehensive coverage of the technical standards issued by local government departments and the implementation documents at the grassroots level; consequently, the findings of this study are more suitable for elucidating the core institutional structure of China’s national park policy system, rather than being regarded as a comprehensive presentation of the overall situation regarding local implementation arrangements. The PMC-Index evaluates the structural integrity and internal consistency of policy texts, rather than the actual ecological and social outcomes generated following policy implementation. Consequently, future research should combine PMC evaluation results with empirical data—such as species population sizes, land cover changes, visitor pressure, community livelihoods and law enforcement performance—to further examine the relationship between policy design and implementation outcomes.

Generative AI statement

The author(s) declared that Generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/ffgc.2026.1950431/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: He Linbo, Hunan Agricultural University, China

Reviewed by: Xiangting He, Xihua University, China

Xiji Jiang,Xi’an University of Architecture and Technology, China