Introduction

Humankind adapted to different and difficult environmental conditions during the centuries to obtain food, materials, and services for its livelihood, applying agricultural, silvicultural, and pastoral practices that modified the natural environment contributing to the creation of different traditional rural landscapes (TRLs)1. TRLs belong to the group of cultural landscapes, a term that was introduced in scientific literature about 100 years ago2. TRLs received increasing attention in the last decades, both at the scientific and at the institutional level, due to their multifunctionality and to the capacity of providing various ecosystem services to the rural communities. These landscapes are capable of providing high-quality products3, as well as services related to (agro)biodiversity conservation4,5, hydrogeological risks mitigation6,7 and improved control and mitigation of plant pests and diseases thanks to the application of traditional practices5. In addition, TRLs contribute to strengthening the local identity8,9, beside representing areas with a high touristic potential due to their aesthetic and cultural values10,11. TRLs also represent examples of sustainable adaptation to various environmental conditions, assuming a new role in the context of climate change and of the related challenges that humanity will have to face in the near future12,13.

At the institutional level, the first initiative for the protection and valorization of cultural landscapes was established in 1992 by UNESCO. During the 16 th UNESCO Committee, it was decided to include the cultural landscape category in the World Heritage List (WHL), defining these landscapes as the result of the “combined works of nature and of man”, “illustrative of the evolution of human society [.] over time, under the influence of the physical constraints and/or opportunities presented by their natural environment and of successive social, economic and cultural forces”14. While some authors highlighted that the category “cultural landscape” is underrepresented within the UNESCO WHL, especially in some parts of the world15, as for October 2024, 121 properties with 6 transboundary properties are included in the WHL under the category “cultural landscapes”, of which 5 are located in China. UNESCO divides cultural landscapes in different categories16:

  1. 1.

    Landscape designed and created intentionally by man: including garden and parks constructed for aesthetic reasons, often associated with religious or monumental buildings and ensembles.

  2. 2.

    Organically evolved landscapes: resulting from an initial social, economic, and/or religious need, and developed as a form of adaptation to the surrounding natural environment. These landscapes are divided into two sub-categories:

  • 2.a Relict (or fossil) landscapes: landscapes where the evolutionary process came to an end, but distinguishing features are still visible.

  • 2.b Continuing landscapes: landscapes that still retain an active social role in contemporary society, being associated with the traditional way of life and in which the evolutionary process is still in progress.

  1. 3.

    Associative cultural landscapes: landscapes whose inclusion in the WHL is justified by the powerful religious, artistic or cultural associations of the natural element rather than material cultural evidence, which may be insignificant or absent.

According to these UNESCO definitions, TRLs are included in the 2.b subcategory, as they preserve an active role in the rural society, providing benefits and services17. It is also necessary to highlight that UNESCO is only interested in cultural landscapes of Outstanding Universal Value (OUV), but also everyday rural landscapes, as defined by the European Landscape Convention18, can provide benefits and services to rural communities. In this paper we use the term TRLs referring to cultural rural landscapes not only of Outstanding Universal Value, but including all the rural landscapes created through the centuries (traditional landscapes) by the traditional activity of farmers, foresters, and shepherds.

Another crucial international initiative related to TRLs, even if not specifically including this term in its name or aims, is the GIAHS (Globally Important Agricultural Heritage Systems) Programme established in 2002 by the Food and Agriculture Organization (FAO) of the United Nations. The aims of the GIAHS Programme are to identify and safeguard agro-silvo-pastoral systems resulting from the adaptation of rural communities to the surrounding environments, and shaped and managed through the application of traditional and sustainable techniques. These systems, therefore, can be considered TRLs. GIAHS sites do not only contribute to the preservation of cultural rural landscapes, but are also characterized by high levels of agrobiodiversity, by a rooted presence of traditional knowledge, and by cultural and social values. GIAHS sites are not relics of the past but are still fundamental for providing multiple products and services to local rural communities19. As per October 2024, FAO has designated 89 systems in 28 countries as GIAHS, of which 22 are located in China.

The high number of GIAHS sites located in China compared to other countries testifies, on one side, the great and increasing interest at the national and ministerial level for this initiative and, on the other side, the huge potential (in terms of number, variety and variability) of TRLs that are still present. Despite the intensity, rapidity, and the surface affected by socio-economic changes and consequent land use transformations, China is still particularly rich in terms of TRLs. The growing importance of TRLs at the institutional level is demonstrated by the NIAHS (Nationally Important Agricultural Heritage Systems) initiative established in 2012 by the Ministry of Agriculture, with the aims of identifying, protecting, and valorizing agricultural heritage systems. NIAHS are defined as “ingenious agricultural production systems inherited through generations and formed by long-term co-adaptation of local people with their environment” and are considered particularly important for the “sustainable development and multi-functionality of agriculture in China”. NIAHS management approach is not based on their “musealization” or strict conservation, but, as for the GIAHS sites, on dynamic conservation principles. NIAHS list also represents the “basin” from which selected agricultural heritage sites are proposed to the FAO for the inclusion in the GIAHS Programme; to date 188 sites are recognized as NIAHS.

The NIAHS initiative is not the only national programme established following the GIAHS approach and principles, as other countries have developed similar programmes at the national level, including Italy20, Korea21, and Japan22, testifying a growing attention to the topic due to the connections and opportunities related to sustainable rural development, rural economy diversification, landscape and agrobiodiversity conservation. In addition, TRLs are also important attractors for implementing the rural tourism, that in many rural areas can represent a crucial driver to counteract depopulation, offering farmers diversified sources of incomes as they can be involved in direct product selling, accommodation, or catering activities23,24. Another important issue is related to the fact that traditional food products produced in NIAHS sites are commonly sold at a higher price on the market as they are perceived to be of higher quality by consumers (also thanks to a label specifically developed in China for NIAHS and GIAHS agricultural products), compared to similar products produced outside recognized traditional landscapes25, even if this trend has significant inequalities depending on the counties26.

However, there are subjective approaches regarding the selection of these systems, while a data-based methodology for identifying potential distribution of TRLs across larger regions based on high-resolution spatial analyses has not yet been developed. Taking China’s NIAHS as an example, the identification is mostly based on administrative boundaries and data, while each province is not allowed to propose more than three sites per year27. Therefore, it is necessary to develop a framework for TRLs identification integrating environmental, socio-economic and cultural dimensions. In fact, it is important to emphasize that the characterization of these landscapes cannot be based solely on environmental and biophysical factors, as cultural factors are the main drivers of these systems; in this regard, Guo, et al.28state that in the long term, cultural and social factors play an important role in determining the character of NIAHS and should play a key role in the development of management strategies. Indeed, according to Maurer, et al.29, changes in management and agricultural practices due to cultural and socio-economic transformations are responsible for altering the local environmental features, i.e. the plant diversity of traditional landscapes in the Alpine steppe. In southern Europe, chestnut trees have been widely planted by farmers in hilly and mountainous areas over the past centuries, even far beyond their ecological optimum, being multipurpose species capable of providing a wide range of essential products and materials, creating unique cultural forests and cultural landscapes30; this has also affected the species composition of current forests, even if traditional management of chestnuts ceased many decades ago. Despite the growing interest towards TRLs, these systems are affected by different threats, including the abandonment of rural areas and of traditional agricultural practices, especially in marginal areas31, and agricultural intensification in more mechanizable and suitable areas32,33. Another major threat is represented by urban sprawl and construction of new infrastructures that do not only determine land use changes and shrinkage of agro-silvo-pastoral surfaces, but also negatively impact the aesthetic and perceptive values of the sites34.

This study intends to contribute to fill the knowledge gap related to the identification of potential TRLs of different types across large territories by applying a standard methodology. The aims of this research are the following: (i) proposing a conceptual framework based on high-resolution spatial analyses of environmental, socio-economic, and cultural indicators, to provide a data-based methodology to support the identification of TRLs at the national scale; (ii) testing the proposed framework through its application to the entire territory of China as a study case, validating the results; (iii) discussing the effectiveness and the limitations of the proposed methodological framework.

Materials and methods

Development of a conceptual framework for identifying TRLs

The identification of TRLs requires the integration of environmental, socio-economic and cultural factors. As highlighted by different national and international programmes dealing with TRLs (FAO GIAHS, UNESCO cultural landscapes, China NIAHS, Italian Register of Historical Rural Landscapes, Japanese satoyama) as well as by scientific research19,27, the environmental factors represent the basis on which TRLs develop as a result of human activities through the centuries (cultural factor) and which evolve according to socio-economic conditions. Therefore, it is necessary to integrate environmental, cultural, and socio-economic dimensions to properly develop a methodological framework for the identification of potential TRLs across large regions. The socio-economic and cultural dimensions are deemed more significant than the environmental dimension in identifying historical rural landscapes28. This can be attributed to the human capacity to utilize a wide range of resources and adapt to diverse environments. Consequently, while the environmental dimension provides useful context or specific environmental limitations to agro-silvo-pastoral practices, the socio-economic and cultural indicators offer deeper insights into the presence and preservation of TRLs:

Environmental Dimension: The environmental dimension is useful for understanding the physical attributes and limitations, which influence its overall character and usability35,36. Indicators such as morphology, including slope, exposure, and elevation differences, are crucial for assessing the terrain’s structure and its suitability for traditional agro-silvo-pastoral activities37,38. Other additional indicators include soil composition, climate, and vegetation cover, which offer insights into the natural resources available and their impact on agro-silvo-pastoral practices and land uses.

Socio-Economic Dimension: The socio-economic dimension examines the human and economic aspects that impact the landscape, by analyzing indicators such as GDP, population density, and land ownership patterns. Other indicators can also be related to demographic changes, migration patterns, and economic developments, that may help in identifying the socio-economic forces shaping the landscape over time and space, that led to the current socio-economic situation of a specific rural area.

Cultural Dimension: The cultural dimension encompasses the heritage, traditions, and practices defining the landscape’s historical identity. Indicators such as the distribution of specific ethnic groups within the territory, as well as local traditional agricultural products and the presence of Geographical Indications39, offers critical insights into the cultural richness and diversity40,41. Other indicators such as locations of historical land use practices, cultural landmarks, historical settlements, locations of cultural events, and areas with significant folklore or community practices further enrich the analysis42,43.

The methodology applied in the present study was based on the following consecutive steps, summarized in Fig. 1:

  1. 1.

    Selection of the relevant spatial data, preparation of the spatial indicators.

  2. 2.

    Assigning a specific score to the values of the different indicators.

  3. 3.

    Assigning a specific weight to the different indicators.

  4. 4.

    Multicriteria analysis for the identification of potential traditional landscapes.

  5. 5.

    Validation of the results.

Due to differences in the specific indicators and sources used for analysis at the national level in different countries, the specific data handling processes and methods are described in detail in the following sections referring to the specific case study.

Fig. 1
Fig. 1
Full size image

Methodological framework for the identification of potential TRLs of agricultural, forest, and pastoral type.

Case study

The study area used for testing the proposed framework corresponds to the entire terrestrial surface of China. Being one of the largest country in the world, China encompassess diverse environments and climates, including Tundra (ET), Subarctic (Dfc), Cold semi-arid (BSk), Cold desert (BWk), Monsoon-influenced hot-summer humid continental (Dwa), Monsoon-influenced humid subtropical (Cwa), Subtropical highland (Cwb), Humid subtropical (Cfa), Temperate oceanic (Cfb), Tropical monsoon (Am), and Tropical wet (Aw)44. Based on climatic, morphological, and socio-economic features, China’s vast territory is commonly divided into seven major geographic regions: Northeast, North, Northwest, Southwest, Central, East, and South China (Fig. 2).

Fig. 2
Fig. 2
Full size image

Geographic divisions of China. Chinese national boundaries have been downloaded from the official governmental database https://www.webmap.cn.

The methodology

Data sources

The choice of the layers and databases used in this study was made after having explored different possibilities based on various and recent layers and databases available at a national level and according to the results of previous studies28,36. Overall, 8 different databases and layers have been used, of which 2 related to environmental characteristics (Digital Elevation Model, Land-Use/Cover), 3 to cultural characteristics (Traditional villages, Typical Agricultural Productions, China’s planted forest map), and 3 to socio-economic characteristics (Census Data, Gross Domestic Production, Population) (Table 1). Starting from these layers, different indicators have been extracted as explained in the following section.

Table 1 List of original databases and layers used to perform the study.
Full size table

Step 1: selection of the relevant Spatial data and Preparation of the Spatial indicators

The first step focused on the preparation of the different layers through different processing performed with QGIS 3.28 and GRASS 7.8.7 software. These elaborations, summarized in Table 2, were necessary to homogenize the spatial data in terms of resolution and coordinate reference system and to obtain the different indicators.

Table 2 List of elaborations applied to the original databases and layers and resulting indicators.
Full size table

The first consideration concerned the fact that TRLs are characterized by very different features depending on their type. Traditional forest landscapes, for example, present morphological, socio-economic, and cultural aspects that differ from those mainly devoted to pasture or agriculture. Consequently, the scores and weight of each indicator would be not pertinent if applied similarly to all the three types. For this reason, the workflow was divided into three distinct, yet very similar, procedures. The methodology takes into account agricultural, forestry (including shrublands), and grassland land uses, respectively for traditional agricultural landscapes, forestry landscapes and pastoral landscapes; the weights of the indicators are calibrated based on these types of landscapes.

The methodology applied to the three different landscape types was very similar, and the differences for each phase of the elaborations are highlighted in this section. Each layer was then reprocessed to obtain a raster with a 1 km resolution for each indicator, so that they could be overlaid to obtain a single value, derived by the sum of the different indicators.

The first phase of the elaborations regarded the preparation of land use layers. For each category, the types of land use that characterize the specific traditional landscapes were identified. Remote inspection revealed that tea gardens were initially misclassified as ‘shrubland’. Since this category contains a wide range of vegetation types, tea plantations were included in agricultural land only in the traditional tea-producing townships, while bush forests in the other townships were categorised as forest land. The layers were eventually processed into a 1 km pixel grid to record the percentage of agricultural, forest and grassland land use.

The second phase focused on the preparation of the morphology layer. For this indicator, it was decided to consider only the slope, classified it according to the following classes: 0–5%, 5–10%, 10–20%, 20–30%, > 30%. The slope layer was then clipped using the land use layer (specific to each landscape type) to obtain the relative slope class for each land use pixel. Finally, the 30 m resolution slope raster map was resampled in a 1 km raster, calculating for each pixel the average slope of the specific land use.

The spatial information about the census data was transferred to the vector layer of the county administrative boundaries. Each county was assigned the relative values about (i) Percentage of ethnic minority on the total county population, (ii) Percentage of people engaged in agriculture, forestry, pastoral and fisheries activities on the total county population, (iii) Percentage of population living in rural and urban areas on the total county population. The vector layer was rasterized into three different raster layers according to the same 1 km grid to obtain a layer for the three indicators.

The distance from traditional villages was processed using a 1 km resolution raster, where each pixel reported the minimum distance from the nearby traditional village.

Regarding the information on GDP, it was considered that the simple use of the total GDP data per km² could be not very significant, as it does not indicate the real wealth condition of the population. It was also observed that the cartography was very similar to that of population density. It was therefore decided to use the GDP per capita, collecting the values of the 1 km raster by county, and then dividing these values for the total county population to obtain the average GDP per capita. The vector layer was then rasterized with the same 1 km grid of the previous layers.

The layer related to the population density was not modified as the original database already consisted of a raster with 1 km pixel grid.

Regarding the presence of local traditional products, the original data corresponding to a table was reported and merged with the township shapefile. The products were first categorized based on their type: Tea, Rice, Tree products (products harvested from permanent tree cultivations excluding tea), Field products (products cultivated in open fields, i.e. annual species including non-tree fruits, such as melons), Animal origin products (any product derived from livestock, excluding fish), Fish products (products derived from aquaculture or fishing; marine fish products were not considered), Forest-related products (products harvested in forest environments, including wood and medicinal herbs), Mushrooms products. After this categorization, two indicators were obtained:

  1. 1.

    The number of products related to each specific landscape category for each Township.

  2. 2.

    The number of types (agricultural-forestry-pastoral) of products present in each Township

For the first indicator, data was reclassified as follow: for the agricultural landscape type the first four categories have been considered; for the forest landscape type only the forest-related products and the mushroom products have been considered; for the pastoral landscape type only the products of animal origin have been considered; fish products have not been considered in this study. For the second indicator, data was reorganized to highlight how many types of products (agricultural-forestry-pastoral) are present in each township, obtaining values from 0 (no products) to 3 (all three kinds of products are present). This information has been used to underline the presence of rich and diverse landscapes, where different kinds of traditional activities are carried out.

Finally, China’s planted forest map layer was used exclusively for the characterization of forest landscapes, without further elaborations as the original layer was already provided in a 1 km resolution raster.

At the end of the first step, 12 different raster files were produced according to a 1 km resolution grid. Each raster represents a specific indicator.

Step 2: assigning a specific score to the values of the different indicators

The second step was to establish a scoring scale for the different values of each indicator. This score, ranging from 0 to 1, represents how much a given value contributes to the potential presence of a traditional landscape, and its evaluation sometimes differs based on the type of landscape considered (agricultural, forest, pastoral). A score of 0 indicates that the specific value for a particular indicator does not contribute to the likelihood of identifying a traditional landscape, whereas a score of 1 signifies that the specific value is optimal for the presence of a traditional landscape. Values between 0 and 1 means that there is a certain probability of finding a traditional landscape. Some indicators have continuous values, while others have discrete values. In the first case, values were assigned linearly by specifying thresholds and using a graph through the graph function of the mapcalc tool in GRASS GIS 7.0. In the second case, specific scores were assigned to each value of the different indicators. Following, the rules for choosing these scores are described in detail and their values on a minimum-maximum (0–1) scale are presented in Fig. 3.

Percentage of land use

A minimum threshold of 5% for each specific land use (agricultural, pastoral, or forest) was set, below which the pixel is considered equivalent to a pixel without land use. Pixels without the relevant land use are excluded from all processing and considered null in every field. Despite being a continuous value, this indicator has been treated just as a boolean value.

Dominant slope class

This indicator has been elaborated from a continuous value (the slope) in a discrete one (slope classes).

  • For agricultural landscapes, the likelihood that a given traditional agricultural area has undergone intensification and modernization was considered. Therefore, lower slope classes have lower scores as it is more likely that these areas have undergone agricultural intensification due to their suitability for mechanization, while higher slope classes have higher scores as these areas are more difficult to mechanize and can be effectively cultivated only applying traditional techniques45.

  • For forest landscapes, the likelihood that a forest primarily serves for hydrogeological protection rather than being traditionally managed for obtaining specific products was considered, meaning that on steep slopes it is less likely to have traditional management46. Therefore, higher slope classes result in lower scores.

  • For pastoral landscapes, slope can represent a limiting factor as livestock usually prefer gentle slopes47, while slopes above 60% are considered unsuitable48; scores are, therefore, similar to those applied to traditional forest landscapes.

Traditional village proximity

For this indicator, common to all three landscape types, the highest score was given to areas within 5 km of a traditional village, while the lowest score was given to areas beyond 15 km. Values in between were distributed with the graph function of mapcalc. This choice is supported by research findings showing that the closer to traditional villages, the more the landscape is shaped by traditional agro-silvo-pastoral practices, while beyond a certain distance (10–15 km) the direct influence of traditional human activities on the landscape is limited49. Although the proximity to traditional villages was a continuous value, it was elaborated in classes, thus it has a discrete score distribution.

Population density

For this indicator, the highest score was considered where human presence is lower, an average value was considered where the density reached 100 inhabitants/km² and the lowest value was set for densities of 1000 inhabitants/km² or above. This choice aims to exclude more urbanized areas, and to underline the areas where the presence of sparse population is prevalent. No different scores were given to the different landscape types. This is a continuous indicator.

GDP per capita

As reported in the literature, there is an inverse correlation between increasing GDP and the presence of traditional rural landscapes and villages28,50, as higher GDP is commonly related to more urbanized areas. The GDP per capita was valued based on the quartiles, giving the maximum score for the areas that were included in the first quartile and the lowest value for the areas over the third quartile. This is a continuous indicator.

Amount of traditional products

For this indicator, the presence of a single traditional product (related to the specific type of landscape) was considered a positive score, and the presence of more than one as the highest possible score. This is a discrete indicator.

Variety of traditional products

For this indicator, a growing score was chosen based on the number of types of products (agricultural - forestry - animal) present in that specific township. This is a discrete indicator.

Percentage of population belonging to an ethnic minority on the total county population

To determine a score for this indicator, various approaches were attempted. The initial attempt involved using the average value of all counties as the maximum score (1). However, the distribution of this value across the different counties, generally very low for most of them but with very high peaks in few counties, resulted in very low and meaningless scores. Given that the presence of ethnic minorities, even if not particularly abundant in percentage terms, is an important indicator for the potential presence of TRLs51,52, an approach using the median value as the maximum score (1) was then attempted. In this case as well, the result was meaningless, as it assigned a maximum value to half of the China counties. It was therefore decided to use a hybrid method, assigning the intermediate score (0.5) to the median value, and then reaching the maximum score (1) at the average value. This is a continuous indicator.

Percentage of people living in rural areas on the total county population

For this indicator, the mean and median are relatively similar. Therefore, the highest score was assigned to all values above the mean. This is a continuous indicator.

Percentage of population involved in agricultural activities on the total population

As with the previous indicator, the mean and median values are very similar. The highest score was assigned to all values above the mean. This is a continuous indicator.

Planted forest map

This indicator was considered only for the forest landscapes. The presence of natural forest has been considered as a positive indicator, while no value was assigned to planted forest. China, in fact, has promoted large afforestation programmes during the 1980 s and 1990 s, such as the Natural Forest Protection and the Sloping Land Conversion Programs, to reduce runoff, control erosion, and stabilize local livelihoods, and more recently to increase carbon sequestration53. These programmes only partly met the original expectations54,55, but caused significant land use and landscape changes by introducing large planted forests with no traditional and heritage values56. This is a boolean indicator, the presence of natural forest has a value of 1, the absence has a value of 0.

Fig. 3
Fig. 3
Full size image

Scores assigned to the different indicators considered in the study from 0 to 1.

Step 3: assigning a specific weight to the different indicators

The third step focused on assigning specific weights to all the different indicators, by applying the Analytical Hierarchy Process (AHP), a methodology developed by Saaty57to quantify relative priorities for a given set of alternatives on a ratio scale. AHP is a method for determining weights by constructing judgment matrices of multiple types of data with pairwise comparisons, and has the characteristics of ease of use, systematic construction of the problem, and calculation of standard weights and alternative priorities58,59. It consists of three main steps: (1) construct a hierarchical structure model to determine the target, criterion and factor levels; (2) construct a judgment matrix, make pairwise comparisons at each node and quantify the relative importance of the indicators; (3) heterogeneity test and cross-check of the judgment matrix using Yaahp software and the consistency index, that should be < 0.1 to satisfy consistency of the assigned weights60. The study invited 30 experts (from relevant research fields in China and European countries) to score the different indicators according to the judgment matrix that compared the different indicators in pairs; a total of 25 forms were collected of which 23 were valid and passed the consistency test to obtain the final weight for each indicator (in the supplementary materials).

Step 4: multicriteria analysis for the identification of potential traditional landscapes

The fourth step focused on combining the weights and the scores of each indicator to perform a multicriteria analysis in order to obtain a value for each pixel of the 1 km raster grid linked to the potential presence of traditional landscapes of different types, according to the following formula:

$$M_n =W_1i_1 + w_2i_2 \:+\cdots \:+ w_pi_p$$

Where: Mn is the final value of potential presence of traditional landscape, considering that the higher the value the higher is the potential presence; wp is the weight of the p indicator; ip is the score of the p indicator.

Step 5: validation of the results

The fifth step focuses on the comparison of our results with the localisation of the 188 China-NIAHS, whose focus is, however, mainly on production systems or typical agricultural products rather than on TRLs. In order to better compare and validate the results of this study, the 188 NIAHS have been classified in three different classes according to the main type of production system: agriculture related NIAHS, forest related NIAHS, and pasture related NIAHS (Fig. 4). For each landscape type, the Zonal Statistics tool in QGIS was used to calculate the average pixel value of the raster within the boundaries of the corresponding NIAHS sites. For example, for the agricultural NIAHS, the average pixel value of the agricultural raster map within the boundaries of these sites was calculated. This allowed us to obtain an average landscape evaluation value for each type of landscape within the NIAHS sites. Subsequently, this average value was compared with the average landscape evaluation value for the entire national surface. This comparison allowed for the assessment of the validity of the proposed methodology by comparing the results obtained in the NIAHS with those of the rest of the country. The same process was then repeated for forest and pastoral landscapes, using the respective shapefiles of the corresponding category of NIAHS. In this way, the likelihood of each type of landscape within the NIAHS was evaluated and compared with the national average.

Fig. 4
Fig. 4
Full size image

Map of the NIAHS sites according to the main type of traditional landscape. Chinese national boundaries have been downloaded from the official governmental database https://www.webmap.cn.

Results

Traditional agricultural landscapes

Traditional agricultural landscapes are more likely to be found in the eastern part of the country (Fig. 5), especially in Northeast, Central, and East China, that, together with Guizhou province, can be considered as potential hotspots for traditional agricultural landscapes. The central eastern part of China is characterized by a higher population density and urbanization, but despite this there still are some parts of it that might preserve important traditional agricultural landscapes, even if scattered among more intensive agricultural systems. Northeast China and part of North China could also be interesting for the presence of traditional agricultural landscapes on quite large surfaces, while in Northwest China the potential presence of traditional agricultural landscapes results to be rarer and scattered due to serious environmental limitations.

Fig. 5
Fig. 5
Full size image

Map of potential presence of traditional agricultural landscapes. Chinese national boundaries have been downloaded from the official governmental database https://www.webmap.cn.

Traditional forestry landscapes

Traditional forest landscapes are less scattered than agricultural ones, and are mainly concentrated on specific parts of the country (Fig. 6). Significant hotspots can be found in Northeastern China, along the borders with Russia and North Korea. The southern part of China also has large portions of forests that might have significant cultural values. In Northwest China and in most of North China traditional forest landscapes have a reduced potential, corresponding to mainly small and scattered forest surfaces due to environmental limitations, while in Central and East China traditional forest landscapes seem less common due to urbanization and agricultural intensification, except for the southern provinces that might preserve valuable traditional forest landscapes.

Fig. 6
Fig. 6
Full size image

Map of potential presence of traditional forestry landscapes. Chinese national boundaries have been downloaded from the official governmental database https://www.webmap.cn.

Traditional pastoral landscapes

Traditional pastoral landscapes are complementary to the previous types. Significant hotspots can be found along the North China-Mongolian border or in Qinghai, Gansu province and Ningxia Hui Autonomous Region in Northwestern China (Fig. 7), being more related to large and sparsely populated areas. Other potentially interesting areas include large surfaces in the Qinghai-Tibet plateau, and north-western and western China, along the border with Kazakhstan, Kirghizstan, and Tajikistan. Smaller and scattered surfaces can instead be found in completely different environments, as in the Yunnan province, but in general this type of traditional landscape is more linked to large and contiguous patches rather than to scattered and small ones, as it is related to extensive grazing.

Fig. 7
Fig. 7
Full size image

Map of potential presence of traditional pastoral landscapes. Chinese national boundaries have been downloaded from the official governmental database https://www.webmap.cn.

Traditional landscapes potential presence according to the province boundaries

Comparing the area share of the potential distribution of traditional landscape types within each province in combination with the province boundaries (Fig. 8), it is found that, except for the four municipalities directly under the central government, namely Beijing, Tianjin, Shanghai and Chongqing, all other provinces possess high potential with two or more traditional landscape types, with the difference being that there are large variations in the area share of the various types of traditional landscapes. Among them, the high potential areas of the three types of traditional landscapes in Liaoning province are relatively close to each other, each accounting for about 10% of the province’s area. The seven provinces of Shanxi, Inner Mongolia Autonomous Region, Heilongjiang, Shandong, Sichuan, Guizhou, and Yunnan all have potential of preserving significative portions of different types of traditional landscapes, but with differences in terms of percentage of each type, with Qinghai Province standing out regarding the potential presence of traditional pastoral landscape.

Fig. 8
Fig. 8
Full size image

Graph of the potentiality of each province regarding the presence of traditional landscapes of agricultural, forest, and pastoral type.

Validation of TRLs potential presence

As shown in Table 3, the average potential score of agriculture related landscapes in China-NIAHS is 51.0, forest related landscape is 48.4, and pasture related landscape is 69.3, while the average potential scores of these three types of TRLs on the entire national surface are equal to 42.4, 41.1, and 49.8, respectively. The average potential scores of the agriculture and pastoral related landscapes within the boundary of China-NIAHS are much higher than the overall average potential scores outside, while the average score of the forestry related landscape inside and outside the boundary of China-NIAHS is similar, but with the first one anyhow higher than the national average.

Table 3 Average scores of potential NIAHS and at the national level for traditional agricultural, forest, and pastoral landscapes.
Full size table

The analysis of the results of this study according to the Heihe-Tengchong line also validates the applied methodology (Fig. 9). The Heihe-Tengchong line (also known as Hu line) runs from the city of Heihe in Heilongjiang province to the city of Tengchong in Yunnan province, and was conceived by the Chinese geographer Hu Huanyong in 193561. This line divides the country in two very different parts, according to the population density and also to the economic development, as the southeastern part accounts only for 36% of China’s total area but hosts 96% of the total population62. These two parts of the country have also been characterized by different historical events, with the southeastern part that underwent a major agricultural development, due to the suitable environmental conditions, to the water network and to the needs of obtaining different agricultural products for the growing population through the centuries. The northwestern part of the line, which has a very low population density, is mainly characterized by a harsh climate, less suitable for cultivating a wide range of crops and, in large part, less suitable also for forests, and has been the home of nomadic herders since ancient times. The Heihe-Tengchong line thus divides the country into distinct environmental and cultural regions, with very different TRLs. The results of this study directly and quantitatively support the theoretical basis of the Heihe-Tengchong line, which is manifested in the fact that the high-potential hotspot areas of traditional agricultural and forestry landscapes are mostly concentrated on the east side of the line, and the high-potential hotspot areas of traditional pastoral landscapes are mostly distributed on the west side.

Fig. 9
Fig. 9
Full size image

Maps of potential traditional landscapes with the Heihe-Tengchong Line in red (from left to right: traditional agricultural landscapes, traditional forestry landscapes, traditional pastoral landscapes). Chinese national boundaries have been downloaded from the official governmental database https://www.webmap.cn.

Discussion

This study is the first to map the spatial potential distribution of traditional rural landscapes throughout the land territory of China. The developed framework turned out to be innovative as it is based on the integration of various environmental, socio-economic, and cultural indicators that are applied over a large territory with high spatial resolution. The methodology is, in fact, based on a complex approach that allowed to identify with a high resolution (1 km grid) the potential presence of traditional landscapes of different types (agricultural, forest, pastoral) across all the country and for each province. The spatial distribution of 188 China NIAHS, especially with the agricultural and pastoral NIAHS, confirms the validity and reliability of the proposed methodology. Despite the fact that the methodology has been tested across a single, but highly diversified in terms of environment and society, country, the framework has been developed taking into account its replicability in different countries. While the exact list of indicators could differ from one country to another due to the specificity of the environmental, cultural, and socio-economic context, the use of indicators of different types (environmental, socio-economic, and cultural indicators) represents the basis of this methodology and can be replicated in other contexts.

In terms of spatial distribution, the eastern coastal region is a. hotspot for the potential presence of traditional agricultural and forestry landscapes, but both show fragmentation and insularity. However, in China as well as in most of flat areas around the world, the rapid urbanization process has fragmented the traditional landscape, representing a major threat for different traditional agricultural landscapes of great cultural and historical values. This is the case of the Mulberry-Dykes and Fishponds System of Huzhou City, whose importance was also recognized at the international level as it was included in the GIAHS Programme of the FAO34,63, but also of other traditional landscapes in flat areas of the European continent64.

The spatial distribution characteristics of 7586 national forest villages in two batches published by the State Forestry and Grassland Administration of China also validate the methodology adopted in this study65. Traditional forest landscapes often have a high cultural, spiritual, and ecological value, as especially among some ethnic groups, traditional practices contributed to select specific tree species that were not used to provide firewood or timber, but were preserved for religious/spiritual reasons (sacred tree, fengshui tree) and nowadays represent a cultural heritage but also a refuge or source of food for different fauna species52. According to Sinthumule66, sacred forests have a great potential for in situ conservation, but their management should integrate formal protection with traditional practices in order to be effective for biodiversity preservation. Similarly, in Europe regularly managed chestnut groves were particularly rich in biodiversity, while with the abandonment of the traditional management and their successions towards mixed broadleaved forests the level of biodiversity decreased67. In China, different national forest restoration programmes have been implemented in the last decades, including the Natural Forest Conservation Program (NFCP), the Grain for Green Project (GGP), and the Beijing-Tianjin Sandstorm Source Treatment Project (BTSST). Unfortunately, these initiatives emphasize the natural values and the plantation of large forests, without considering the cultural values and the role of Traditional Ecological Knowledge for sustainable rural development. These large scale programmes led to significant land use changes, raising concerns on their effective sustainability from the environmental and cultural point of view68. It is necessary to consider that, on a global scale, cultural forests are poorly considered by public awareness and by institutions, that often neglect the cultural values and the related Traditional Ecological Knowledge, in favor of values connected to nature and to the ecological values69,70. Cultural forests managed through traditional practices should receive more attention by local institutions as they combine cultural and ecological values.

Regarding traditional pastoral landscapes potential distribution, several studies validate the results of this research, highlighting a significative higher concentration in western and northern China71,72. Pastoral landscapes and grasslands are particularly important for preserving open spaces and habitats for birds and other small species, especially in Europe where their reduction in the last decades led to a decrease in their ecological role, in particular in relation to biodiversity and to different ecosystem services73.

In many initiatives related to traditional rural landscapes protection and valorization, the choice of the boundaries represents a key issue. Usually, two different approaches can be used: the one based on administrative boundaries or the one based on physical boundaries. NIAHS sites and some of the GIAHS, as well as Geographical Indications, are based on administrative boundaries, that makes them more easy to manage; other initiatives, as the Italian National Register of Historical Rural Landscapes, instead, decided to use physical boundaries to be more realistic, but with increasing difficulties regarding their management27. Geographical Indications and TRLs share common goals in protecting regional cultures and resources, but they differ significantly in their protection targets, methods, and scopes. Geographical Indications focus on product protection and market management, using legal measures (e.g., trademark registration) to prevent counterfeiting. In contrast, TRLs protect traditional agricultural systems and their associated cultural landscapes, biodiversity, and traditional knowledge, promoting community participation and sustainable development through integrated management. The two approaches complement each other and can contribute to the protection of cultural heritage.

Some limitations of the methodological framework and of its application need to be reported. The primary limitation of the proposed framework lies in the selection of spatial layers, encompassing environmental, social, and cultural data. Specifically, it includes: (1) The collection of social and cultural digital spatial data is challenging; on the one hand, quantifiable indicator data for social and cultural dimensions are very limited and mostly based on administrative areas, and there is a lack of high-resolution data at the full scale of China. On the other hand, due to changes in administrative boundaries or names of administrative units, it takes a long time to reorganize the data, which leads to some errors. In the future, the spatialization study of the indicators of social and cultural dimensions of TRLs can be strengthened to provide richer and more accurate indicators for the identification; (2) In the case study, the classification of the existing land use/cover data did not consider the characteristics of the cultivation, so the tea cultivations were included in “shrubland” when they are actually cultivated areas. In the future, land use/cover types can be decoded using methods such as random forests in combination with high-resolution remote sensing imagery to obtain a more accurate classification; (3) By calculating and comparing the potential average of TRLs across China with the potential average of three types of TRLs in the NIAHS, this paper finds that the scores for TFLs are only slightly higher than that in the NIAHS. This is because there are fewer forest-related spatial data indicators available. However, despite the closeness of these two scores, it can be argued that identifying the spatial distribution of TRLs using the current methodology is effective and can be used as a census of TRLs; (4) Some of the selected indicators can present a meaningful collinearity, as in the case of population density and GDP; to reduce this problem, we decided to use the GDP per capita instead of the total GDP, that is more correlated to the average level of economic development rather that to the total population. (5) The proposed methodological framework is applicable only to potential areas of traditional agricultural, forestry, and pastoral landscapes, excluding traditional fishery landscapes. Traditional fisheries and coastal landscapes (seascapes) are receiving increasing attention and some of them are already included among the China’s NIAHS, GIAHS, and UNESCO programmes. Their identification, however, should be based on different indicators compared to the ones used in this study, and require future research to further expand the applicability of the TRLs identification framework.

Conclusions

The proposed methodology proved to be effective in identifying the potential presence of TRLs over a large territory with high spatial resolution. The inclusion of environmental, socio-economic, and cultural indicators is crucial to allow a correct identification of TRLs that are the results of cultural activities performed by local rural communities through the centuries. This methodology can represent a guideline for the identification of TRLs in other countries, although its replicability must be carefully evaluated, especially by choosing the proper indicators and assigning them different scores depending on the country and on the environmental, socio-economic, and cultural specific context.

The present methodology could also provide reliable high resolution spatial data that can be used in different countries at the national, provincial, or local level for identifying TRLs or to develop specific initiatives for their conservation and valorization. At the local level, in fact, the correct identification of areas rich in traditional rural landscapes can support the valorization of an important heritage and of the related traditional agricultural products, with positive impact for the local rural economy, especially considering the potential role of rural tourism. At the national level, instead, the results that this methodology can offer can be used to identify possible traditional rural landscapes, where to implement bottom-up approaches for their effective valorization and conservation, also thanks to the opportunities related to the NIAHS initiative or to international programmes, such as the FAO GIAHS Programme or the UNESCO World Heritage List. The effectiveness of this important initiatives, in fact, is strictly connected to the active involvement of the local communities, in particular of local farmers and shepherds; in many cases, in fact, top-down approaches proved to be unsuccessful in the long term as they are not capable of meeting the real needs and expectations of the local communities.