ZHOU Lian, HUANG Xueyuan, ZHAO Chunmei, PU Tiancun, ZHANG Lei

College of Landscape Architecture, Sichuan Agricultural University, Chengdu, 611130, China

Keywords：

Sustainability

Rural landscape

Homegardens

Agricultural heritage

Social ecosystem

Chengdu Plain

A B S T R A C T

Traditional rural homegarden agroforestry system in the Chengdu Plain of China, called “Linpan” in Chinese, integrates the ecological functions of the landscape with human production activities. Studying the driving mechanisms of rural landscape changes in the Chengdu Plain is of great significance from stakeholders’ perspective. Taking the Pidu Linpan Farming System (PLFS) in the suburban area of Chengdu (designated as one of China-Nationally Important Agricultural Heritage Systems in 2020) as a case, this study combined remote sensing image analysis, oral history interviews, and focus group interviews to elucidate the driving forces of landscape changes in the PLFS. The results show that tourism development, traffic accessibility,economic demand, and agricultural heritage protection measure are the main driving forces promoting the stability and maintenance of the traditional homegardens. In contrast, population ageing, land circulation, centralized residence, climate change, and living and recreational need are the forces leading to adverse changes in the traditional homegardens. In addition, these driving forces have led to the gradual abandonment of traditional agricultural activities and critical issues related to rural landscape planning and management. The current research indicates that infrastructure construction and increased traditional agricultural income are considered as the best practices of local stakeholders, promoting the development of the protected homegardens of heritage sites and tourist destinations. Finally, we put forward some suggestions to improve and maintain the traditional rural landscape: (1)establishing a benefit-sharing mechanism; (2) establishing a training system with traditional technology and culture; (3) strengthening infrastructure construction; (4) promoting the development of the agricultural industry; (5)improving the cultural quality of farmers; and (6) establishing a management system with legal effects. This research can provide a basis for the formulation of rural landscape planning and the orderly and healthy development of agricultural heritage in Chengdu Plain.

1. Introduction

Declining of rural areas has become a global problem (Tenza-Peral et al., 2022). In this context, the transformation of rural land use is concerned by researchers worldwide (Yang et al., 2021; Yirgu et al., 2022). In the early stage of the 21stcentury, the research of rural land use transformation was mainly covered by economics and geography, but it was then gradually integrated into sociology, landscape ecology, biology, and other disciplines (Flores-Díaz et al.,2014). With the continuous advancement of economic globalization and regional integration, the internal and external environments of rural regions have changed significantly, and the countryside has entered a new stage of transformation and development, showing a trend of diversified development (Han et al., 2021). Rural areas are gradually being understood as multifunctional spaces to protect traditional cultural values and natural resources. In addition to providing space for production activities, such as agriculture and forestry, rural areas also have new functions of leisure tourism, environmental protection, and preservation, reminder, and heritage of rural traditions(Zhang et al., 2018).

In the early stage of rural land use transformation research, qualitative methods such as field investigation and case analysis were employed to understand the views of farmers and determine the main factors influencing rural landscape development (Okello et al., 2009). The main characteristics of rural land use transformation research are deepening from qualitative analysis to quantitative analysis and from single social survey to multiple methods (Martínez-Sastre et al., 2017). The United States, European Union (EU), Australia, and other countries and regions have used various policy tools and mechanisms to encourage positive environmental actions to create a sustainable rural landscape environment. The Australian Land Conservation campaign focuses on rural environmental degradation and weighs the pros and cons of different approaches of environmental protection (Royal, 2021). The new rural policy of the EU harmonizes environmental standards and subsidies through the limits of minimum environmental standards, essential environmental resource payments, and higher-level environmental compensation regulations to minimize trade disputes among World Trade Organization (WTO) members and protect environmental resources (Kiryluk-Dryjska and Baer-Nawrocka, 2021).

In the case of limited resources, comprehensive land renovation is an essential method of agricultural transformation and development to achieve large-scale agricultural management (Pašakarnis et al., 2021), but it also dramatically impacts the traditional rural landscape. The loss of agricultural land caused by urbanization has been emphasized by many researchers (Handayani, 2013; Appiah et al., 2014; Debolini et al., 2015). In the Chengdu Plain of China, this general change endangers the versatility of rural landscape and erodes the contribution of nature to human well-being, i.e., ecosystem services, especially those related to ecosystem regulation and cultural services (Liu et al., 2019a; Wu et al., 2020). Therefore, it is necessary to understand the causes of rural landscape changes in the Chengdu Plain and provide information on hindrance and stability factors.

The rural landscape of the Chengdu Plain is generally called “Linpan” in Chinese. It is a type of rural homegarden agroforestry system, which is regarded as a multifunctional landscape system (Wu et al., 2020). It maintains ecosystem services and biodiversity while integrating the ecological functions of the landscape with human production activities (Liu et al., 2019b). The landscape of the homegardens is composed of five elements: dwelling,woodland, farmland, road, and river and/or irrigation canal (Fig. 1). Some scholars have summarized the three main characteristics of the homegardens. Specifically, the homegardens are usually small and flexible in scale, reasonable distribution, and adaptable to the carrying capacity of the natural environment; there is an apparent transition zone between the homegardens and the surrounding farmlands; and the residents living in the homegardens are harmonious coexistence with trees (Min et al., 2020). At present, two opposing trends influence the authenticity and versatility of the rural landscape in the Chengdu Plain. First, the local government implements protection measures of agricultural heritage and sets the development tone of “protection priority”, but the effect is not obvious. Moreover, most studies have focused on controlling the form and texture of this kind of settlement macroscopically. Second, urban expansion and its impact on the livelihood of rural residents have led to the degradation and loss of productive agricultural land and agro-ecosystem functions, as well as the changes in the intensification of land use and land cover (LULC).

Fig. 1. Schematic diagram of rural landscape unit (a) and rural landscape settlement (b).

Therefore, some scholars have analyzed the climate regulation function, ecosystem cultural services, spatial patterns, and distribution characteristics of the homegardens from a macroscopic point of view (Li et al., 2019; Liu et al., 2019b; Wu et al., 2020), and revealed the driving mechanisms of the landscape pattern change and the characteristics of the typical cultural landscape. However, most previous studies are based on the quantitative analysis(Wan et al., 2021a, b) and rarely combine qualitative and quantitative methods. Besides, previous researches also lack views on rural planning and construction from the perspective of stakeholders.

The harmony between human and nature in the rural landscape comes from the social-ecological interaction generated by the knowledge of local people. Failure to recognize these interrelationships may lead to conflicts among stakeholders, as this interaction shapes cultural identity and drives land management choices to maximize human well-being (Plieninger and Bieling, 2013). The agricultural landscapes of the Chengdu Plain are undergoing substantial changes, reflecting the physical and cultural changes they have experienced. However, until now, we know very little about why local stakeholders choose to change the rural landscape and how to deal with the changes in the rural landscape. Therefore, agricultural landscape conservation policies should ensure the rights and interests of local landowners and encourage the participation of more stakeholders. This can not only improve the ecosystem service efficiency of rural landscapes but also help to establish more specific policy management guidelines in this region.

The present study used qualitative and quantitative methods, focusing on the regional landscape transformation and sustainability of the homegardens of the Pidu Linpan Farming System (PLFS) in the Chengdu Plain, which is one of the China-Nationally Important Agricultural Heritage Systems. Specifically, we investigated the stakeholders’landscape cognition of the homegardens in the core area of the Chengdu Plain. We further combined the quantitative LULC information from modern maps, government mapping information, and aerial images with the qualitative information from oral history interviews and focus group interviews, to diachronically reveal the rural landscape values of local stakeholders (Atwell et al., 2011; Zurita-Benavides et al., 2016; Bürgi et al., 2017; Gullino et al., 2018).The aims of the present study are to (1) sort out the types of heterogeneous changes in the homegardens, (2) determine the main driving forces of rural landscape changes and their effects, and (3) analyze the possible future scenes of the homegardens and propose a development strategy for the sustainability of the homegardens in the study area.

2. Materials and methods

2.1. Study area

This study was conducted in the northwest of Chengdu, Sichuan Province, China. The core research area is the Pidu District, where the PLFS is located. The surrounding areas include the Dujiangyan City, Wenjiang District,Shuangliu District, and Chongzhou City. Our research area belongs to the core irrigation area of the Dujiangyan irrigation system, a World Cultural Heritage Site. Investigating the surrounding areas of the PLFS, not just the heritage site, can provide a more comprehensive understanding of the homegardens. With suitable climate, soil, and water resource conditions, long-standing farming culture, and unique artesian irrigation system, the agricultural landscapes of the homegardens in the Chengdu Plain are typical and representative.

2.2. Methodological framework

Traditionally, landscape change research mainly focuses on the spatial pattern changes using multi-temporal satellite images, and rarely considers the perception of local communities (Cushman and Wallin, 2000; Moseley,2006; Sohl et al., 2019). Therefore, the combination of cartographic and social researches can better elucidate landscape changes (Krajewski, 2019). We divided the research framework into three parts: (1) qualitative analysis of focus group, (2) quantitative analysis of geographic information, and (3) landscape scenario assessment.

2.2.1. Qualitative analysis of focus group

We conducted field research in new rural communities in the study area for a long time since 2017. In 2019, we began a new stage of the study using the “snowball” method (Luyet et al., 2012; Storie and Bell, 2017) to conduct oral history interviews with 59 social identities (41 ordinary residents and 18 members of rural cultural groups; i.e.,village cadres, peasant entrepreneurs, clan leaders, and teachers) in the study area (Table 1). The interview time was approximately 1 h per person. In-depth interview is considered as the most suitable method for monitoring the cultural heritage (Eiter and Vik, 2015). Though in-depth interviews, we gained a better understanding of the local traditional customs, history, culture, and regional development patterns, how local people perceive landscape and environmental changes, and the social and political impact of their life experiences (Persson et al., 2018).

Table 1Basic statistics information of respondents.

From October 21 to 23 in 2019, we participated in the 6thChina Agricultural Cultural Heritage Conference and the 1stSymposium on the Protection and Development of the PLFS. Under the leadership of local government, we gained a more systematic understanding of the research and utilization of agricultural heritage, agricultural culture, and landscape characteristics in the PLFS, and identified the objects for follow-up focus group interviews. Gullino et al.(2018) explained that identifying focus group is a qualitative research technique that involves stakeholders, including the representatives of key stakeholders; it should cover a wide range of interests, aspirations, expectations, and perspectives. Conrad et al. (2011) suggested that the existing landscape research is severely limited in the extent to which it involves stakeholders or develops innovative methods. The participation and understanding of stakeholders are core tasks in landscape management (Prager, 2015).

In this exploratory study, participants with different opinions and roles participated as stakeholders. The participants in the focus group of this study included (1) people with economic interest (resident, returning entrepreneur, returning worker, forestry producer, and tourism developer); (2) people interested in governance and planning (government staff and rural planner); (3) people with an interest in nature conservation and agricultural heritage (member of non-governmental organization, compiler of local chronicle, and member of rural cultural group);and (4) people interested in research (academic researcher) (Table 2). All stakeholders involved in the focus group interviews are related to the traditional rural landscape and can directly or indirectly influence landscape changes.During field visits and interviews, participants may voluntarily provide more information and are less likely to conceal knowledge of land use and landscape characteristics (Wartmann and Purves, 2017).

Table 2Number of stakeholders involved in the focus group interviews.

We first introduced the purpose of the study to participants of the focus group interviews, after collecting their personal background information. Subsequently, they recalled the driving forces that led to the landscape changes in the homegardens over the past 70 years. We asked participants of each focus group to describe the significant possible impacts of driving forces, and each participant explained his or her opinions and described the positive and negative aspects of these driving forces. This section focused on the importance of the driving forces perceived by each participant. Finally, we asked the participants to discuss their scores related to the importance of the driving force from 1 (low importance) to 5 (high importance). The distribution of importance values was the result of open discussion among the participants.

2.2.2. Quantitative analysis of geographic information

Satellite photography has long been used to analyze landscapes and has recently been used to monitor landscape changes (Dramstad et al., 2002; Fensham and Fairfax, 2002; Bewley, 2003; Ståhl et al., 2011). Cloud-free satellite images captured in the dry seasons of 1973, 2002, 2010, and 2020 were acquired from the U.S. Geological Survey Earth Resources Observation and Science website (https://www.usgs.gov/). We selected 1997, 2002, 2010, and 2020 as study years comprehensively according to the situation of data collection. First, we visually interpreted the satellite images of the Pidu District from 1973, 2002, 2010, and 2020 and compared them with the land use map of the Pidu District in 2018 provided by the Agricultural, Rural, and Forestry Bureau of Pidu District. Second, referring to the urban-rural gradient analysis method of Wadduwage et al. (2017) and Liu et al. (2019b), we continued to interpret the remote sensing images of the near suburbs, middle suburbs, and outer suburbs (i.e., linear distances from the central urban area of Chengdu was 17.4, 24.9, and 32.4 km, respectively). In process of interpreting the satellite images, we identified the area and location of “non-agricultural” lands to better understand the direction and rate of the current rural landscape transformation. Our criteria for “non-agricultural” lands are those that occupy permanent basic farmland to plant seedlings, turfs and ornamental plants used for greening decoration and other plants that destroy the tillage layer.

2.2.3. Landscape scenario assessment

In the last part of the focus group interviews, we asked the participants to imagine the homegardens in the future.In accordance with the methods of Ramos (2010) and Ramirez et al. (2015), we showd several aerial photos to the focus group participants to determine the future scene of the landscape. These aerial photos included satellite images of 1997, 2002, 2010, and 2020. Then, we asked the participants to think creatively about what strategies could be used to better protect the appearance and connotation of the PLFS over the next 20 years. To analyze the views of stakeholders, we constructed the transformation and persistence scenarios related to the homegardens (Fig. 2).

Fig. 2. Representation of transformation (a) and permanence (b) landscape scenarios related to the homegardens.

3. Results and discussion

3.1. Rural landscapes of different types

In the process of development and transformation, the agricultural landscape in the near and middle suburbs hasbeen seriously damaged, and the agricultural landscape in the outer suburbs has also been threatened to a certain extent. There are apparent differences in the direction and degree of landscape changes in these three areas, which can be classified into four landscape development models: the forestry development model, industrial development model, facility forestry development model, and traditional agricultural development model.

There are three reasons for the different development models. First, the near suburbs are greatly affected by urbanization (Appiah et al., 2014), and the expansion of construction land and road has blocked part of the canal system (Wu et al., 2020), resulting in insufficient water supply to farmland and reduced efficiency of traditional crop farming. Second, China’s large-scale urban construction and green space transformation in recent years have caused a sharp increase in the demand for trees, so the relative economic return of planting trees is greater, and the traffic accessibility between the near suburbs and the middle suburbs is more robust. Third, because of the limited sources of livelihood in rural areas, a large number of young residents leave rural areas to work (Li et al., 2019), and the lack of rural labor force leads to changes in land use. Facing to the lower input and higher output, an increasing number of residents choose to plant fields full of trees or rent land to forestry companies.

3.1.1. Near suburbs: forestry development model and industrial development model

There are two main directions of LULC change in the near suburbs that significantly impact the homegardens. The first is the forestry development model (Fig. 3). The main feature of this model is that the production objects of the residents have changed from crops to ornamental shrubs, trees and fruit trees. The second is the industrial development model (Fig. 4). The main feature of this model is the change in land use types caused by the changes of government land use planning. The government gradually converts farmland into construction land by land expropriation. There is also a common reason for the loss of rural human resources. Residents adjacent to urban built-up areas are more likely to be attracted to the city, resulting in the outward migration of residents in the near suburbs. Many residents plant trees with high economic value on abandoned land, generally showing an irregular state, but for the land leased to forestry companies, trees are arranged in an orderly manner. Moreover, owing to the land merger brought about by the development of township enterprises, some residents need to move to other places to live, and residents with better economic conditions will choose to move to the city.

Fig. 3. Schematic diagram of the forestry development model of land use and land cover (LULC) change in the near suburbs in 1973 (a), 2002 (b), 2010 (c), and 2020 (d).

Fig. 4. Schematic diagram of the industrial development model of LULC change in the near suburbs in 1973 (a), 2002(b), 2010 (c), and 2020 (d).

3.1.2. Middle suburbs: facility forestry development model

Unlike the near suburbs, there are many greenhouses in the middle suburbs, in addition to an increase in woodland.This is the facility forestry development model, and this kind of landscape development model has a significant impact on the homegardens (Fig. 5). The main feature of this model is that the production objects of the residents have changed from crops to ornamental flowers, shrubs, trees, and fruit trees, and also the production mode of the residents has changed from the traditional agricultural mode to the modern industrialized agricultural mode. Land rent in the middle suburbs is relatively low. With consideration for a certain degree of traffic accessibility, more and more companies choose to conduct large-scale facility forestry in the middle suburbs. The plant varieties are mainly ornamental flowers, shrubs, trees, and a small number of vegetables and fruit trees. Compared to those in the near suburbs, most of the new woodlands in the middle suburbs are regular, and the new irregular woodlands are shelterbelts on both sides of the river. This also reflects that most of the land in the middle suburbs has centralized management after renting.

Fig. 5. Schematic diagram of the facility forestry development model of LULC change in the middle suburbs in 1973(a), 2002 (b), 2010 (c), and 2020 (d).

3.1.3. Outer suburbs: traditional agricultural development model

At present, most of the land in the outer suburbs retains the traditional agricultural development model (Fig. 6),which generally maintains the characteristics of the traditional homegarden pattern. In government planning, the outer suburbs are the key protected areas of the PLFS, and the agricultural landscape and irrigation canal system have been well protected. Because it is far from the built-up area of the city and is less affected by urbanization, the residents in the outer suburban areas often continue the traditional agricultural way of life, the division of land in here is more precise, and there is still a close relationship between the residents and the PLFS. However, some of the land still has large areas of regular woodland, indicating that some forestry companies carry out intensive land use in the outer suburbs.

Fig. 6. Schematic diagram of the traditional agricultural development model of LULC change in the outer suburbs in 1973 (a), 2002 (b), 2010 (c), and 2020 (d).

3.2. Driving forces and landscape effects

3.2.1. Main driving forces of landscape changes

The driving force of the homegardens can be determined by analyzing the results of the focus group interviews. In the focus group interviews, participants identified many driving forces that may have a negative, positive, or variable impact on the traditional homegardens. Participants of the focus group identified the driving forces based on the landscape effect and importance score (see Table 3). Population ageing, land circulation, tourism development, and centralized residence are the most important driving forces of the homegardens (importance score of 5). Traffic accessibility (importance score of 4), economic demands (importance score of 4), climate change (importance score of 3), living and recreational needs (importance score of 3), and agricultural heritage protection measures (importance score of 2) were considered secondary driving forces of the homegardens.

Table 3Importance scores of the driving forces and their impacts on the traditional rural landscape changes.

The above driving forces show that the widespread problem is the abandonment of traditional agricultural activities and the change of residents’ livelihoods. We confirmed to the focus group that the biggest problem of the rural landscape changes was the abandonment of traditional crops by residents. People had a negative view of the abandonment of traditional agricultural activities because traditional farming methods were considered as an important part of the local landscape and knowledge. The impacts of population ageing, land circulation, traffic accessibility, economic demand, and the migration of young people to urban areas in search of better employment opportunities have led to the farmland rental and the loss of local farming culture. Population ageing, centralized residence, climate change (extreme weather such as drought and flood), economic demand, and living and recreational needs have led to the spontaneous expansion of woodland by residents, usually at the expense of their farmland. In recent years, recurrent drought and flood disasters, as well as cold and other natural forces, have become important driving factors of landscape changes. Residents in the focus group believe that precipitation has declined significantly since the 1990s. Agricultural heritage protection measures scored the lowest (importance score of 2) because participants all believe that there is a lack of clear and practical action on the protection measures.

Abandonment of traditional crop cultivation seems to be the most crucial process in landscape changes. Based on the results of remote sensing research and comparing these perceived changes with those obtained from satellite image interpretation (Sohl et al., 2019), it is confirmed that the effects of forest land and greenhouse increased significantly(Fig. 7). Dujiangyan City, Wenjiang District, Shuangliu District, and Chongzhou City show almost the same situation in the field investigation process. Statistical analysis indicates that 28.12% and 4.70% of the 225.02 km2of farmland in Pidu District have been transformed into woodland and greenhouse, respectively. A greenhouse of 100 plots was randomly selected for verification, with 87 planted ornamental flowers and shrubs, 8 planted vegetables, and 5 planted fruit trees. We determine that greenhouse exists as a derivative of woodland and occupy farmland space with overwhelming superiority.

Fig. 7. Distribution of land use in Pidu District from 1973 to 2020. This map is based on the standard map (川审(2016) 027) of the Map Service System (https://scsm.mnr.gov.cn/) marked by the Sichuan Bureau of Surveying,Mapping and Geoinformation, and the base map has not been modified.

3.2.2. Analysis of main driving forces in different stages

According to the literature review and oral history interview materials, we divided the modern evolutionary process of the homegardens into three periods (Fig. 8). There are two important time points for changes in the homegardens.First, Sichuan Province began to implement the new policy of reform and opening-up in 1981, which realized a complete change in the mode of agricultural land use. Second, the rural economy and culture of Sichuan Province entered a period of rapid development after 1994. Before and after these two time points, villages in the Chengdu Plain have different development patterns.

Fig. 8. Evolution of LULC change from farmland to woodland in the Pidu District from 1949 to 2020.

From the perspective of residents, the development of the rural landscape is guided by productivity to address the problem of food shortage during 1949-1980. In this period, the main driving forces of rural landscape development were population pressures and government policies. The rural landscape presented a single pastoral landscape development model. From 1981 to 1993, consumerism guided the development of rural landscapes. During this period,the main driving forces of rural landscape development were economic demands, traffic accessibility, and living and entertainment needs. As the contradiction between the growing leisure and entertainment needs of the public and the protection of the rural landscape with the plural economy and the construction and development of public resources,an increasing number of people left the countryside. From 1994 to 2020, the development of rural landscapes was guided by humanism. During this period, the main driving forces of rural landscape development were tourism development, population ageing, land circulation, and centralized residence. Since 2003, local governments have actively promoted the overall planning of urban and rural areas to improve the rural environment. In 2007, through the comprehensive consolidation and circulation of land, rural settlements gradually “demolished small buildings and merged large settlements”, which gradually pulled up the prelude to constructing a new countryside.

3.2.3. Impact of driving forces on landscape changes

Concerning the driving forces of landscape changes, the participants described the interactions between the driving forces (Prager, 2015). Participants of the focus group outline that the abandonment of traditional crop cultivation had caused the PLFS to change, resulting in severe environmental problems and threatening the local traditional culture.In addition, centralized residences have led to the abandonment of large-scale homegardens, which are no longer carefully managed. The problems of management and land planning, the loss of traditional skills and local knowledge,and the lack of traditional culture in new rural settlements are the main negative consequences of centralized residence.

The land circulation policies aimed at increasing residents’ income have also had three adverse effects: (1) reducing the planting area of traditional crops; (2) destroying traditional planting methods and living habits; and (3) blurring the boundaries of the homegardens. Land circulation policies have helped to increase farmers’ incomes, but they are not conducive to protecting the traditional planting system. Climate change has increased the workload of residents in maintaining crop yields, and climate disasters have led to the destruction of infrastructure, crops, and vegetation.In addition, the relatively high in non-traditional agricultural income of resident is one of the reasons. In this case,traditional products have not received sufficient attention, so the traditional production model has been abandoned.

Centralized residence is a planning behavior that is led by the government, which has dramatically changed the spatial pattern of the PLFS. Residents of concentrated residential areas believe that the distance between residence and farmland increases, while the living space decreases. There is no place for drying grain or storing farm tools and captive livestock, which is not conducive to agricultural production. In addition, some scattered residents want the centralized houses supported by the government to away from muddy and incomplete roads. They believe that concentrated residence will help to create a large-scale landscape, attract more tourists, and increase their employment opportunities and incomes.

With the increasing living needs of residents, an increasing number of families abandon traditional agriculture, and they cannot continue their social and cultural norms and practices. Farmland management, irrigation canal management, and mechanisms for the organization and utilization of woodland resources, such as the Dujiangyan irrigation canals, have disappeared or become less important. Some practices that used to encourage protection of the environment and agricultural heritage have not been effectively implemented.

In contrast, traffic accessibility and economic demands have two positive effects on the homegardens: (1) the development of the homegardens’ tourist destinations; and (2) the promotion of the homegardens’ cultural value.Participants agree that tourism development could bring about many positive effects, such as increasing residents’income, maintaining the traditional landscape, and promoting rural planning development. Moreover, they believe that agricultural heritage protection measures should be combined with the development of tourism. Some respondents and forestry producers who are not involved in tourism management said that there is no apparent connection between tourism and the sale of trees, and that tourism and agricultural heritage protection were not good for them. Alternatively, different planning and protection methods bring uneven benefits, resulting in the dissatisfaction among low-income stakeholders.

In summary, farmland in the traditional PLFS is being replaced by woodland, resulting in the loss of agricultural heritage. The emergence of new elements has changed the landscape of the Chengdu Plain, i.e., regular woodland and greenhouse is new elements recently introduced in the PLFS of the Chengdu Plain. Therefore, this type of LULC change has a hostile landscape and environmental impact on the PLFS, which is characterized by traditional farming life.

3.3. Future landscape scenarios and development strategies

Participants of the focus group discussed the future landscape of the PLFS. The transformation and persistence scenarios were built and evaluated. Most participants (60.9%) pointed out that the abandonment of traditional agricultural cultivation is the most likely perception scenario. Currently, rural residents in the area do not have enough time and labor to maintain the excellent state of traditional agriculture to meet the requirements of agricultural heritage protection. Abandoned and unmanaged irregular woodland and regular woodland and greenhouse are considered to be the main landscape scenarios for transforming the traditional homegardens. The participants assumed that mixed forests would become a feature of the study area after the abandonment of traditional agriculture. Of the participants,30.4% think that the land area transferred to other types or rented by the companies would increase. Only 8.7% of the participants believe that the traditional homegardens may be restored.

Participants of the focus group were asked to list critical strategies for improving and maintaining the PLFS over the next 20 years (Fig. 9). Infrastructure construction (78.3%) and increasing traditional agricultural income (69.6%)are considered to be the significant landscape schemes related to the persistence of the traditional homegardens.Nearly one-third (34.8%) of the participants believe that strengthening the identifiability of rural landscape is the primary strategy to protect the traditional homegardens, as the traditional homegardens have a clear boundary; that is,flat farmland, which is now occupied by woodland, resulting in blurred boundaries. Preserving the original architectural style (26.1%), resuming the production of bamboo handicrafts (21.7%), and brand management of agricultural products (21.7%) are considered appropriate strategies to protect the traditional homegardens. All-forone tourism development, inheritance of intangible cultural heritage, and agricultural life experience are other strategies identified by the participants (21.7%, 17.4%, and 17.4%, respectively).

Fig. 9. Strategies identified by the participants of the focus group to maintain the Pidu Linpan Farming System (PLFS).

In particular, residents agree that the hardening of roads in infrastructure construction must be combined with a strategy to enhance the identifiability of the homegardens. In the focus group interviews, residents and researchers debated the road system. Residents believe that the road is not as beautiful and poetic as described in literary works.The most significant perception of the road system among residents is that it is muddy and difficult to use during the rainy season. Moreover, residents believe that the first task of infrastructure construction in a heritage site is that the roads need to be hardened, which is very important for foreign tourists to experience the entire agricultural heritage environment. There are different views on which hardening paths should be given priority, but it is agreed that trees in the fields around the protected homegardens of heritage sites need to be cleared first to improve the physical and visual accessibility.

Farmland structure before industrialization is also regarded as an essential aspect of the protected landscape. The participants propose that the farmland structure around the protected homegardens and tourism destinations of the heritage site can be restored so that it is not the same as the farmland landscape in other parts of the country, whilst becomes a characteristic agricultural landscape similar to the Hani terraces (Zhang et al., 2017), one of the globally important agricultural heritage systems of China. Participants have different acceptance levels for using modern tools and methods to manage specific parts and buildings.

In addition to their views on the physical elements of the agricultural heritage environment, all participants expressed their strong attachment to the rural landscape and said that the homegardens system is of great significance to them. The homegardens have inseparable cultural values such as homesickness, family consciousness, and childhood memories. With the migration of young people to cities and towns, the middle-aged and older people in the homegardens have a different view from young people. In particular, they know what traditional production and life are like, including not only physical practices, but also potential customs and unwritten “rules”. It may include local knowledge and skills transferred through labor rather than formal training. Examples include historical farming practices, building construction and maintenance, and manufacturing and maintenance of tools. Good task coordination and neighborhood cooperation are essential when residents live in the homegardens and manage their private buildings, farmlands, and public land.Participants warry that these cultural values may no longer exist in the future.

4. Conclusions and recommendations

4.1. Conclusions

Discussions among stakeholders involved in the process help to assess the relative importance of changes in agroforestry systems. However, participants may have been more inclined to pay attention to the relatively small areas of their settlements and surrounding sites. Simultaneously, mapping from remote sensing images can identify changes in the landscape of the entire region and highlight the cumulative effects of many small changes over time. The results show that tourism development, traffic accessibility, economic demands, and agricultural heritage protection measures are the main driving forces that promote the stability and maintenance of the traditional homegardens. In contrast,population ageing, land circulation, centralized residence, climate change, and living and recreational needs are the driving forces leading to the adverse changes in the traditional homegardens. In addition, these driving forces have led to the gradual abandonment of traditional agricultural activities and critical issues related to rural landscape planning and management.

The current research shows that infrastructure construction and increasing traditional agricultural income are considered the best practices of local stakeholders, promoting the development of the protected homegardens of heritage sites and tourist destinations. When discussing and evaluating possible landscape scenarios in the future,participants recognized that the restoration of some standard features is a development strategy to maintain the homegardens. Social and economic needs can strengthen and encourage traditional agricultural activity. Two possible scenarios are identified by the participants of the focus group. The income level of some residents has increased due to tourism development. However, some observed phenomena have shown that tourism development, which has improved farmers’ livelihoods, may also threaten agricultural heritage protection (Zhang et al., 2019). In this context, the participation of local populations is necessary to determine common and appropriate development strategies.

4.2. Recommendations

The core of agricultural heritage is the function of the traditional agricultural system and the internal mechanism of sustainable development. The rural landscape of the Chengdu Plain seems to have recreated the situation of many traditional agricultural areas in Europe in previous years (Eiter and Vik, 2015; Prager, 2015; Bürgi et al., 2017; García de Jalón et al., 2018). The modernization and specialization of agriculture are regarded as reasonable models, but this model cannot preserve many cultural heritage sites that do not bring economic benefits. In a similar context, we need to learn from existing achievements and past lessons. To maintain the traditional homegardens, we can translate the research results into specific rural planning and the development strategy into local actions. In order to meet the needs of agricultural heritage protection and promote sustainable development in rural areas, we put forward the following suggestions.

(1) Establishing a benefit-sharing mechanism. Farmers are the most vulnerable group among the stakeholders.However, they are also the most critical part of agricultural heritage protection; therefore, it is necessary to establish a benefit-sharing mechanism to ensure the interests of farmers in heritage sites. Farmers should be encouraged to use traditional farming methods through incentive mechanisms, such as ecological compensation.

(2) Establishing a training system that inherits traditional technology and culture. It is necessary to develop relevant technical manuals or protection guides for young people and adults to raise awareness of agricultural heritage through television, radio, the internet, and other media.

(3) Strengthening infrastructure construction. Without destroying the authenticity of traditional villages,infrastructures should be built to strengthen education, culture, medicine, sports, and other living and leisure facilities,and the livability of traditional villages should be enhanced.

(4) Promoting the development of the agricultural industry. To avoid excessive commercialization and capitalization, the integration of agriculture, culture, and tourism is commonly used to industrialize part of the ecological space and commercialize the living space to form a compound production space. It is very important to enrich rural economic forms and promote the healthy development of regional economy.

(5) Improving the cultural quality of farmers. Farmers are direct stakeholders in improving rural environments. By cultivating capable people in rural areas to participate in village governance, we can stimulate the enthusiasm and initiative of farmers and village committees to improve grassroots government.

(6) Establishing a management system with legal effects. At present, most local agricultural cultural heritage management methods lack pertinence and maneuverability, which is not conducive to the systematic protection of heritage.

Declaration of competing interest

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

Acknowledgements

This research was funded by the National Natural Science Foundation of China (52008278). We gratefully acknowledge the interviewees for their support and the government staff of the Wenjiang District, Pidu District,Shuangliu District, Dujiangyan City, and Chongzhou City for their kindly help.