Assessing the Impact of Climate Change on the Landscape Stability in the Mediterranean World Heritage Site Based on Multi-Sourced Remote Sensing Data: A Case Study of the Causses and Cévennes, France

Abstract

Global climate fluctuations pose challenges not only to natural environments but also to the conservation and transmission of human cultural and historical heritage. World Heritage Sites are pivotal regions for studying climate change impacts and devising adaptation strategies, with remote sensing technology showcasing significant utility in monitoring these impacts, especially in the Mediterranean region’s diverse and sensitive climate context. Although existing work has begun to explore the role of remote sensing in monitoring the effects of climate change, detailed analysis of the spatial distribution and temporal trends of landscape stability remains limited. Leveraging remote sensing data and its derived products, this study assessed climate change impacts on the Causses and Cévennes Heritage Site, a typical Mediterranean heritage landscape. Specifically, this study utilized remote sensing data to analyze the trends in various climatic factors from 1985 to 2020. The landscape stability model was developed utilizing land cover information and landscape indicators to explore the landscape stability and its distribution features within the study area. Finally, we adopted the Geographical Detector to quantify the extent to which climatic factors influence the landscape stability’s spatial distribution across different periods. The results demonstrated that (1) the climate showed a warming and drying pattern during the study period, with distinct climate characteristics in different zones. (2) The dominance of woodland decreased (area proportion dropped from 76% to 66.5%); transitions primarily occurred among woodland, cropland, shrubland, and grasslands; landscape fragmentation intensified; and development towards diversification and uniformity was observed. (3) Significant spatiotemporal differences in landscape stability within the heritage site were noted, with an overall downward trend. (4) Precipitation had a high contribution rate in factor detection, with the interactive enhancement effects between temperature and precipitation being the most prominent. The present study delivers a thorough examination of how climate change affects the Causses and Cévennes Heritage Landscape, reveals its vulnerabilities, and offers crucial information for sustainable conservation efforts. Moreover, the results offer guidance for the preservation of similar Mediterranean heritage sites and contribute to the advancement and deepening of global heritage conservation initiatives.

1. Introduction

World heritage sites bear witness to the history of human development and are an essential component of human civilization. As products of ecological evolution under the influence of natural elements, particularly climate, they are closely linked to climate change. These sites not only encapsulate the localized knowledge accumulated by human societies in their adaptation to and utilization of the environment but also exhibit ecological fragility, making them susceptible to detrimental climate change impacts. Sudden meteorological disasters, including floods, hurricanes, and extreme temperature events, often pose direct threats to the safety of heritage sites [1,2,3], damaging physical structures and disrupting the balance of ecological environments. Long-term climate changes, especially alternating periods of drought and heavy rainfall [4], may accelerate the natural processes of erosion and weathering at heritage sites, thereby threatening their structural integrity and historical authenticity. Consequently, policy documents such as the “Policy Document on the Impacts of Climate Change on World Heritage” [5], the “World Heritage Convention” [6], and the “Global Research and Action Agenda on Culture, Heritage, and Climate Change” [7] have been introduced to heighten global recognition of climate change impacts and promote global conservation and adaptation efforts for heritage sites worldwide. Therefore, given the growing dangers of climate change to World Heritage Sites, there is an urgent need for in-depth exploration and research into its impacts. This topic has rapidly emerged as a focal point in the research field, with the aim of providing a robust scientific foundation and effective strategic guidance for the sustainable preservation of heritage sites worldwide. As the ways in which climate change affects heritage sites and the extent of these impacts are progressively revealed, these studies not only enhance our comprehension of this global issue but also supply crucial information for the formulation of adaptation and mitigation measures.

As a hotspot in climate change research [8,9], the Mediterranean region has demonstrated remarkable geographical characteristics in the past. Its intricate mountainous terrain and pronounced contrast between land and sea have collectively contributed to the diverse climatic features and substantial spatial variability [10]. A considerable amount of available information indicates that the Mediterranean region’s climate change has outpaced the global averages in the majority of climate variables. Statistics show that the Mediterranean basin’s annual average temperature has risen by 1.4 °C since the late 1800s, and for most areas in this region, a 1 °C rise in temperature corresponds to a 4% decrease in precipitation [11], indicating a shift towards a warmer and drier climate. Additionally, studies on drought changes have demonstrated a noticeable expansion of arid areas in Europe [12], particularly in southern Europe and the Mediterranean region, where the rates of warming and aridity are significantly higher than the global average [13,14,15,16]. Taking southern France as an example, Lespinas et al. observed through observational data that the warming rate in this region reached 0.49 °C per decade between 1979 and 2004 [17], a value markedly higher than the global average rate of approximately 0.27 °C per decade [18]. New environmental challenges threaten the maintenance of local natural ecological balance, and heritage conservation is facing impacts related to climate change. Understanding and effectively addressing the extensive impacts of climate change on World Heritage properties is crucial. Ensuring the long-term and systematic conservation of high-quality landscapes and promoting ecological sustainability in heritage areas are urgent issues that require immediate attention in today’s society.

Over thousands of years, landscapes have undergone fundamental changes under the influence of multiple factors [19,20]. These changes often vary due to the particularities of their regions, exhibiting different directions and rates of development. Understanding landscape change and the reasons behind it by tracing the response of landscapes to global change processes has been central to recent landscape research. Van Vliet et al. [21] analyzed multiple instances of land cover change across Europe, exploring the dynamics and determinants of agricultural land on a transnational level. Lasanta et al. [22] conducted a review of the widespread phenomenon of land abandonment in Europe over the past few decades and noted that analyzing the driving forces at a large spatial scale is particularly complex. Focusing on land dynamics and related driving factors in the Mediterranean basin, Debolini et al. [23] argued that the unique bioclimate of the Mediterranean region has created the conditions for landscape diversity, while the intensification of global climate change has further weakened the resilience of ecosystems in the Mediterranean countries and their basins, severely affecting both the stability and the benefits these ecosystems provide. Jiménez-Olivencia et al. [24] analyzed and summarized landscape changes in the Mediterranean mountains, emphasizing the prerequisite role of different bioclimatic regions in shaping landscape dynamics and highlighting the significance of regional specificity for comprehending the process of landscape change. Current research underscores that landscape change constitutes a complex, multifaceted process influenced by myriad factors, among which climate change stands out as a pivotal driver influencing the formation and evolution of landscape patterns. Despite this recognition, only a limited number of studies have taken an integrative approach to analyze landscapes as holistic systems. Moreover, our understanding of the mechanisms through which bioclimatic factors impact landscape dynamics in specific regions—particularly within World Heritage sites and other areas highly vulnerable to climate change impacts—remains relatively underexplored.

Landscape serves as a critical scale linking nature with human society in achieving sustainable development goals [25,26]. Its stability is directly related to the maintenance and enhancement of human well-being. However, due to differences in research scales, computational methods, and characterization approaches, current studies on landscape stability have not yet formed an independent research methodology or evaluation system. Land cover plays an important role in landscape change [27,28]. Previous studies have largely focused on diachronic analyses of land cover in the Mediterranean region [21,22,23,24,29], and some scholars have assessed landscape dynamics and stability by quantifying specific landscape indicators [30,31,32,33]. Nonetheless, the evaluation of the stability of World Heritage Landscapes in the Mediterranean region at the landscape scale remains relatively limited. For this purpose, the Causses and Cévennes World Heritage Site was selected in this study as a typical representative of the Mediterranean region. Given the area’s vulnerability to climate change and its role as a historical witness and valuable resource for the development of the region, a thorough investigation of this site is highly pertinent.

In the realm of driver analysis, traditional methods mostly rely on correlation analysis and other means, on the basis of which qualitative or semi-quantitative explanations are provided [34,35,36]. However, correlation does not equate to causation, and the credibility of purely quantitative studies remains under scrutiny [37]. Therefore, quantitative analysis should be complemented by qualitative reasoning [38] or introduced with econometric methods that test for causality [37]. The geodetector q-statistic [39] is a tool designed to assess spatial heterogeneity, identify explanatory factors, and analyze interactive effects between variables. Compared to general statistical measures, it demonstrates a stronger ability to reveal causal relationships and has been widely applied in several fields in natural and social sciences [40,41,42,43]. To more accurately unveil the driving mechanisms behind landscape stability, this study adopted the Geographical Detector as the primary tool for driver analysis, providing robust support for exploring how climatic factors and their interactions affect landscape stability in heritage sites.

In summary, climate change plays a significant role in shaping landscape stability. Understanding the specific impacts of various climatic factors on current landscape stability can not only enhance the practical value of ecological conservation and restoration measures in heritage sites but also provide necessary insights for optimizing the landscape patterns in similar heritage areas. Based on this, our study analyzed the trend of climate change and, from the perspective of landscape structure and pattern, employed the Geographical Detector to investigate how climatic factors influence landscape stability at the Causses and Cévennes Heritage Site. This study aims to inform ecological conservation, restoration, and landscape planning in heritage sites. The specific research objectives are: (1) to reveal the patterns of climate change at the heritage site between 1985 and 2020; (2) to assess the heritage site’s landscape pattern evolution during the past 36 years of longitudinal study; and (3) to explore and quantify the extent to which climate change impacts the heritage site’s landscape stability.

2. Materials

2.1. Study Area

The World Heritage Site of Causses and Cévennes, situated in south-central France (2°55′9″–4°8′58″E, 43°40′17″–44°34′52″N), spans the Southern Pyrenees and Languedoc-Roussillon regions and covers the departments of Gard, Hérault, Lozère, and Aveyron. The study area includes the Parc Naturel Régional des Causses and the Parc National des Cévennes, totaling 6324.37 km².

Due to the complex topography and altitudinal variations, the climate within the heritage site exhibits diversity and local differences. The Causses region consists of several plateaus at altitudes ranging from 750 to 1200 m, which are incised by gorges, creating a relatively uniform landscape. As a transitional zone between the Massif Central and the Languedoc Plain, the Causses region’s climate is influenced by a mix of Mediterranean, Atlantic, and continental climates. In contrast, the Cévennes region features diverse terrain, with altitudes rising from 300 to 1500 m, and is predominantly influenced by the Mediterranean climate. The significant altitudinal variations within the area further enhance the climate’s diversity. The spatial heterogeneity of climatic conditions renders the Causses and Cévennes an ideal location for studying the climate change impacts on landscape stability in Mediterranean World Heritage Sites. In this study, the heritage site is divided into sub-regions with reference to the Köppen–Geiger climate classification data. Figure 1 depicts the geographical location, landscape units, and climatic zones of the area.

2.2. Data and Pre-Processing

This study collected multi-source data spanning from 1985 to 2020, encompassing aspects of topography, land cover, and meteorological information. Boundary data for the Causses and Cévennes Heritage Site was sourced from UNESCO [44]. Topographic data were obtained from the NASADEM_HGT [45] to describe the terrain within the study area. The GLC_FCS30 land cover dataset [46] was selected as a key data source for analyzing landscape changes and assessing stability. Considering the actual condition and research need of the Causses and Cévennes Heritage Site, this study extracted land cover data for the years 1985, 2010, and 2020 and reclassified them into eight landscape types: cropland, woodland, shrubland, grassland, wetland, water body, impervious surfaces, and bare areas. Meteorological data were derived from the ERA5-Land reanalysis dataset of ECMWF [47], comprising monthly averaged data (~9 km resolution) for near-surface air temperature, dew point temperature, potential evaporation, and total precipitation. Using a raster calculator, we synthesized these into annual climatic indicators and calculated relative humidity from the saturation vapor pressure difference between air temperature and dew point temperature. The aim was to analyze the characteristics and trends of climatic factors over the study period. Additionally, the Köppen–Geiger climate classification data [48] was utilized to delineate climatic zones within the heritage site in this study.

3. Research Framework and Methods

3.1. Research Framework

The overall framework of this study is shown in Figure 2. Initially, ERA5-Land reanalysis data were utilized to systematically analyze the long-term trends and characteristics of temperature, precipitation, potential evaporation, and relative humidity in the study area from 1985 to 2020. Reanalysis data combine historical observations with numerical weather prediction models using data assimilation techniques, providing a long-term, continuous, and spatially complete dataset. This approach addresses issues such as unevenly distributed observation stations and missing data. Subsequently, using land cover data and transition matrix models, we analyzed the spatiotemporal changes in landscape composition and configuration at three specific time points—1985, 2010, and 2020. These analyses reveal the evolution of landscape patterns over the specified time span from 1985 to 2020. Key landscape indicators were selected based on the intrinsic link between landscape fragmentation and stability, leading to the construction of a landscape stability model. This model assessed the landscape stability levels of the Causses and Cévennes Heritage Site in different historical periods, providing a quantitative basis for the analysis of climatic factors’ impacts. Finally, the Geographical Detector model quantified climatic factors’ contributions to the spatial distribution of landscape stability. Our study provides analytical support for developing climate adaptation management strategies tailored to Mediterranean World Heritage Sites and offers valuable insights for conserving similar heritage sites globally.

3.2. Methods

3.2.1. Relative Humidity Calculation

ERA5-Land reanalysis data do not include relative humidity data directly. However, it can be calculated from the temperature $\left(T\right)$ and dew point temperature $\left(T_d\right)$ using the following formula:

$$RH=100\%\times {\displaystyle \frac{e_a\left(T_d\right)}{e_s\left(T\right)}}$$

(1)

where $e_a\left(T_d\right)$ is the actual vapor pressure at dew point temperature, and $e_s\left(T\right)$ is the saturation vapor pressure at the current temperature. These can be calculated using the following formula:

$$e_a\left(T_d\right)=0.6108\times exp\left({\displaystyle \frac{17.27\times T_d}{T_d+237.3}}\right)$$

(2)

$$e_s\left(T\right)=0.6108\times exp\left({\displaystyle \frac{17.27\times T}{T+237.3}}\right)$$

(3)

This approach allows for the estimation of relative humidity from ERA5-Land data, providing valuable insights into the moisture conditions of the study area.

3.2.2. Land Cover Change Analysis

The transition matrix, which originates from systems analysis, is employed to quantitatively characterize system states and transitions. The land cover transition matrix serves as an application of the Markov model to landscape-type change [49,50,51], which is crucial for studying landscape dynamics and environmental management. Utilizing the transition matrix to analyze the direction and scale of landscape type transfers between different periods, it reflects the compositional structure of landscape types within the region across various times, thereby revealing the specific processes of landscape pattern dynamics at the heritage site over the study period. Its expression is as follows:

$$P_{ij}=\left[\begin{array}{ccc}{P}_{11}& \cdots & P_{1n}\\ ⋮& \ddots & ⋮\\ P_{n1}& \cdots & P_{nn}\end{array}\right]$$

(4)

In the formula, $P$ signifies the landscape type’s area; $n$ signifies landscape types’ count; $P_{ij}$ (where $i$ and $j$ are indices) denotes the area that transitions from landscape type $i$ to type $j$ during a specific phase of the study. Specifically, when $i$ equals $j$, $P_{ij}$ refers to the area of landscape type $i$ that remains unchanged within that phase.

3.2.3. Landscape Stability Assessment

(1)

Landscape Index Selection and Calculation

In the ecological context, pattern primarily pertains to spatial patterning, which is constituted by the arrangement of landscape components that vary in size and shape across a given area [52]. Landscape pattern not only determines the distribution of resources and the formation of geographic environment but also have a profound impact on various ecological processes [53]. To conduct a quantitative assessment of the landscape’s structural composition and spatial configuration, this study selected a series of landscape indices (

Table 1. Landscape pattern index description.

Category	Landscape Metrics	Ecological Meaning
Area and Edge metrics	Largest Patch Index (LPI)	Determine the predominant landscape type within the landscape as its dominance intensifies with ascending values. Range: [0, 100].
Shape metrics	Perimeter-Area Fractal Dimension (PAFRAC)	Reflects landscape shape complexity, where higher values signify increased complexity. Range: [1, 2].
Contrast metrics	Total Edge Contrast Index (TECI)	Describes the landscape boundary contrast, with higher values indicating a more pronounced contrast. Generally, patches with different boundary types have higher contrast; additionally, the boundary contrast between landscape types with similar ecological functions is relatively low. Range: [0, 100].
Aggregation metrics	Contagion Index (CONTAG)	Reflects the extent of clustering or spreading tendencies among various patch types in the landscape. A higher value signifies stronger connectivity within the landscape. Range: (0, 100].
	Patch Density (PD)	Describes how fragmented the landscape is; a higher value signifies a greater level of fragmentation. Range: (0, +∞).
Diversity metrics	Shannon’s Diversity Index (SHDI)	Reflects the degree of landscape heterogeneity, with higher values indicating a more balanced distribution of landscape types. Range: [0, +∞).

), including the Largest Patch Index (LPI), Patch Density (PD), Perimeter-Area Fractal Dimension (PAFRAC), Total Edge Contrast Index (TECI), Shannon’s Diversity Index (SHDI), and Contagion Index (CONTAG). These indices collectively reflect the structural characteristics, shape complexity, heterogeneity, and aggregation of the landscape. By calculating and comparing landscape indices in different periods, the present study intends to deepen the understanding and evaluation of the current landscape state within the heritage site and reveal its trends and evolutionary patterns.

(2)

Construction of the Landscape Stability Assessment Model

Landscape is a dynamic natural space, and its changes are mainly reflected in the overall changes in the hierarchical system composed of landscape elements. Based on the theory of the hierarchical patch dynamics paradigm [54,55], the structural characteristics of the landscape are determined by the combined effects of patch dynamics at multiple scales, while landscape stability is directly influenced by these dynamic changes. Therefore, this study selected PD, TECI, and CONTAG as key indicators to assess landscape stability. Given that patches have different boundary types and varying degrees of contrast with adjacent patches, we quantified the contrast degree between boundaries of different landscape types based on principles proposed by Xu Q Y et al. [54]. Specifically, we defined the contrast degree for each type of boundary according to predefined category descriptions and edge contrast files.

Table 2. Edge contrast weights between landscape types.

Landscape Type	Cropland	Woodland	Shrubland	Grassland	Wetland	Water Body	Impervious Surfaces	Bare Areas
Cropland	0	0.3	0.3	0.3	0.3	0.3	0.3	0.3
Woodland	0.3	0	0.3	0.1	0.6	0.9	0.3	0.6
Shrubland	0.3	0.3	0	0.3	0.4	0.9	0.2	0.4
Grassland	0.3	0.1	0.3	0	0.6	0.9	0.3	0.6
Wetland	0.3	0.6	0.4	0.6	0	0.8	0.4	0.1
Water body	0.3	0.9	0.9	0.9	0.8	0	0.9	0.8
Impervious surfaces	0.3	0.3	0.2	0.3	0.4	0.9	0	0.4
Bare areas	0.3	0.6	0.4	0.6	0.1	0.8	0.4	0

lists the contrast values between categories, providing a basis for subsequent analyses.

Using ArcGIS 10.8 and Fragstats 4.2 tools, we employed the grid analysis method to calculate the stability index ($S$). The stability index ($S$) is shown below:

$$S={\displaystyle \frac{CONTAG}{PD\times TECI}}$$

(5)

In the formula, a higher $S$ value indicates that the landscape structure within the area is closer to a stable state, while a lower value suggests greater instability. The detailed meanings of the indices $CONTAG$, $PD$, and $TECI$ are shown in Table 1.

3.2.4. Geographical Detector for Climate Factor Analysis

Building on the insights of Wang et al. [39], the Geographical Detector can effectively reveal the spatial variability present in geographical events and identify key factors that significantly influence these events by comparing predictor and response variables’ spatial distribution patterns. To this end, this study quantified the effects of climatic factors on landscape stability’s spatial distribution within the heritage site, employing the methods of factor detection and interaction detection.

(1)

Factor detection. By testing the spatial variability of individual meteorological factors, the factors with significant influence on landscape stability were identified. The formula is shown below:

$$q=1-{\displaystyle \frac{{{\displaystyle \sum }}_{h=1}^L{N}_h{\sigma }_h^2}{N{\sigma }^2}},\left[0,1\right]$$

(6)

In the formula, the q-value signifies the extent to which climatic elements contribute to the spatial distribution of landscape stability within the heritage site, with higher values indicating a greater impact of the factor on landscape stability. $L$ denotes the count of driving factor categories. $N$ denotes the overall sample count, whereas $Nh$ specifies the count of samples belonging to the $h$ type. ${\sigma }^2$ and ${\sigma }_h^2$ are the variances of the landscape stability index for the whole region and the $h$ type, respectively.

(2)

Detection of Interactions. The relationships among various factors and their effects on the dependent variable were assessed and analyzed by computing and contrasting the q-values for individual and combined factor influences. The factors considered in this study include temperature, precipitation, potential evaporation, and relative humidity. As detailed in

Table 3. Interaction effects between each pair of factors (X1, X2).

Judgment Basis	Types of Interaction
q(X1∩X2) < Min (q(X1), q(X2))	Nonlinear reduction
Min (q(X1), q(X2)) < q(X1∩X2) < Max (q(X1), q(X2))	Univariate nonlinear reduction
q(X1∩X2) > Max (q(X1), q(X2))	Bivariate enhancement
q(X1∩X2) = q(X1) + q(X2)	Mutual independence
q(X1∩X2) > q(X1) + q(X2)	Nonlinear enhancement

, we evaluated their interactions by examining each factor in combination with every other factor, offering an analysis of all possible combinations.

4. Results and Analysis

4.1. Climate Change Characteristics

In this section, we first provide an overview of the climatic characteristics of the heritage site (Figure 3), which are derived from the monthly mean temperature and precipitation data from 1985 to 2020. Subsequently, we analyze climate change within the same period (Figure 4) and explore these changes in detail in the following subsections.

As illustrated in Figure 3, the Causses and Cévennes heritage site has exhibited distinct climatic features over the past 36 years (1985–2020), primarily marked by hot, dry summers and mild, wet winters. The summer temperatures, notably higher, average at 18.3 °C and reach a peak of 19.3 °C in July. In contrast, the winter temperatures are relatively mild, averaging 2.9 °C, with the coldest month being January at a minimum of 2.4 °C. The region experiences abundant precipitation, amounting to an annual total of 1120.1 mm. However, the distribution is uneven, exhibiting significant seasonal variations. Precipitation is predominantly concentrated from autumn to spring, with autumn being the season with the highest average precipitation at 134.1 mm. Conversely, summer rainfall is lower than in other seasons, averaging 59.4 mm of precipitation. After understanding the climatic characteristics of the heritage site, we further analyzed the specific changes from 1985 to 2020. Figure 4 illustrates the temporal dynamics of climatic factors during this period. Through an in-depth analysis of the long-term trends of selected climatic variables, we found that temperatures show a clear upward trend, while changes in precipitation are relatively moderate, with a slight increasing trend that is not significant. At the same time, potential evaporation also shows a significant increasing trend, while relative humidity exhibits a fluctuating downward characteristic.

4.1.1. Temperature

Climate data from the study area, as depicted in Figure 4a, reveal significant long-term temperature trends. Statistical analysis indicates that the multi-year average temperature of the region is 10.2 °C. Notably, the lowest recorded annual average temperature was observed in 2010, at 9.1 °C, which then climbed to the highest value of 11.4 °C by 2020. These data variations indicate a significant upward trend in the annual average temperature between 1985 and 2020, with an interannual rate of change of approximately +0.33 °C per decade, demonstrating significant signs of climate warming.

To more clearly delineate this long-term trend, the present study employed a moving average based on five-year intervals to smooth the annual mean temperature data. The results indicate that for the majority of the period from 1985 to 2010, the temperature at the heritage site maintained a relatively flat yet consistently increasing trend. Although there was a brief and minor dip around 2010, the temperature rapidly rebounded between 2010 and 2020. Moreover, the rate of warming during this latter period was significantly faster than in the preceding phase, suggesting that the impact of climate warming in the region may be intensifying in recent years.

4.1.2. Precipitation

Over the period from 1985 to 2020, the Causses and Cévennes Heritage Site received an average annual precipitation of 1121.8 mm. Within this timeframe, precipitation levels exhibited considerable interannual variability, with the minimum recorded value being as low as 738.1 mm and the maximum reaching up to 1749.3 mm. Figure 4b illustrates the trend in annual precipitation at the heritage site, which, although not markedly pronounced, is generally characterized by a slow increase, showing an average interannual rate of change of approximately +5.2 mm per decade.

Upon further analysis of the 5-year moving average values, it is evident that annual precipitation underwent a complex process of change from 1985 to 2010. Specifically, during the initial decade, there was a slight decrease in annual precipitation. Subsequently, precipitation levels began to steadily increase until around 2003, after which a fluctuating decline commenced. After 2010, precipitation entered a relatively stable phase, characterized by a significant reduction in variability and a tendency for a gradual increase.

4.1.3. Potential Evaporation

Figure 4c presents the long-term benchmark levels of potential evaporation at the Causses and Cévennes Heritage Site. The multi-year average potential evaporation for the region is 3012.9 mm, with values varying from a low of 2556.3 mm to a high of 3387.6 mm. Throughout the study period, the annual potential evaporation exhibited an overall increasing trend, as evidenced by an average interannual rate of change of approximately +100.61 mm per decade.

Upon in-depth analysis of the 5-year moving average curve, it was observed that during the first 15 years, the annual potential evaporation underwent a fluctuating process characterized by an initial decrease followed by an increase, demonstrating significant interannual variability. After the 21st century, especially from 2000 onwards, the annual potential evaporation began to steadily rise. Despite a short-term fluctuation around 2010, which led to a temporary decrease, this did not alter the overall long-term increasing trend.

4.1.4. Relative Humidity

The multi-year average relative humidity at the Causses and Cévennes Heritage Site is 72.5%, with the lowest recorded value at 68.7% and the highest at 75.8%. Figure 4d depicts the percentage variation in relative humidity at the heritage site from 1985 to 2020. Over this period, there is an observable downward trend in relative humidity. Although there is some interannual variability, the overall trend is consistent, with an average annual decrease in relative humidity of approximately 6.14%.

4.2. Landscape Pattern Evolution

4.2.1. General Landscape Changes

Figure 5 presents the geographic layout of the Causses and Cévennes Heritage Site’s landscape types for the years 1985, 2010, and 2020, highlighting the predominance of woodland as the key landscape type. The Cévennes region has long maintained its lush, wooded natural appearance, and extensive natural forests also cover the Causses region’s eastern and northern parts. The landscape structure within the heritage site exhibits a distinct gradient trend from southeast to northwest, with dense forests at the mountain bases giving way to open grassland landscapes at higher altitudes. This ecological gradient suggests a complex interplay of environmental factors that affect how different landscape types are distributed throughout the heritage site.

As indicated in

Table 4. Areas and proportions of the heritage site’s landscape types in different periods (1985, 2010, and 2020).

Landscape Type	1985		2010		2020		Rate of Change (%)
	Area (km2)	Percentage (%)	Area (km2)	Percentage (%)	Area (km2)	Percentage (%)	1985–2010	2010–2020
Cropland	597.512	9.448	926.802	14.654	913.5	14.444	55.11	−1.435
Woodland	4809.597	76.049	4264.94	67.437	4207.393	66.527	−11.324	−1.349
Shrubland	396.017	6.262	568.229	8.985	626.259	9.902	43.486	10.212
Grassland	503.63	7.963	533.823	8.441	542.196	8.573	5.995	1.568
Wetland	1.063	0.017	0.319	0.005	0.335	0.005	−70.038	5.268
Water body	1.785	0.028	2.023	0.032	1.984	0.031	13.314	−1.938
Impervious surfaces	14.749	0.233	28.119	0.445	32.579	0.515	90.643	15.864
Bare areas	0	0	0.099	0.002	0.107	0.002		8.447

, over the duration of the study, various landscape types’ areas underwent changes in different directions and magnitudes. Apart from the decline in the area of woodland and wetland, there was an increase in the area of cropland, grassland, shrubland, water bodies, impervious surfaces, and bare areas. Notably, the reduction in woodland area was particularly significant, exhibiting a continuous decreasing trend, with a total decrease of 602.2 km², and its proportion of the total area dropped from 76.0% to 66.5%. Concurrently, although cropland experienced a process of initial increase followed by a decrease, its net increase remained the largest among all types, with a net addition of 316.0 km². The increase in shrubland was also considerable, reaching 230.2 km², and maintained a high growth trend during both the 1985–2010 and 2010–2020 periods. Additionally, there was a certain increase in grassland and impervious surfaces, particularly the latter, which had the highest growth rate among the landscape types during the same period. In contrast, water bodies, wetlands, and bare areas occupied smaller areas, and their overall changes were not significant throughout the study. Comparing the change trends between the two periods revealed that all landscape types’ rate of change has slowed since 2010, suggesting that the natural landscapes and land cover patterns in the region may be approaching stability.

Based on the landscape classification map, this study calculated transition matrices for two distinct periods, 1985–2010 and 2010–2020, to analyze the changes in landscape types within the heritage conservation area. We utilized Sankey diagrams (Figure 6) to visually represent these transitions, a type of flow diagram where the width of the bands is proportional to the amount of change, illustrating the magnitude and direction of landscape-type changes. The majority of conversions occurred among cropland, woodland, shrubland, and grassland. Although transition activity significantly decreased from 2010 to 2020 compared to 1985–2010, there was continuity in the transfer characteristics. Woodland outflow dominated both periods, accounting for about half of the total outflow, leading to a decrease in woodland area due to a smaller inflow. Cropland saw the largest inflow, primarily from woodland, shrubland, and grassland conversions. Shrubland’s inflow was a close second, influenced by conversions from woodland, cropland, and grassland. Grassland and impervious surfaces also showed a slight increase, while other landscape-type conversions were minor.

From a spatial distribution perspective (Figure 7), landscape-type transition activities during the study period were predominantly observed in the heritage site’s western and northern areas. The southeastern region, influenced by the complex topography of the Cévennes, showed minimal changes and maintained relatively uniform and repetitive landscape patterns. Between 1985 and 2010, pronounced transitions between landscape types were observed, particularly on the Causses Plateau, where large-scale transformations occurred, primarily towards cropland, woodland, shrubland, and grassland. During this period, the migration of grassland was a predominant characteristic in the northern and southwestern parts of the site. In contrast, between 2010 and 2020, transition activities were more dispersed, lacking significant aggregation patterns. The migration of shrubland was notable only in the northwestern part of the site and the Monts du Lozère, where small-scale aggregation areas formed.

4.2.2. Landscape Metrics Analysis

In conjunction with the spatial distribution maps of landscape types across three phases of the heritage site, this study conducted a statistical analysis of landscape metrics for each period. The variation in these metrics indicates a growing landscape fragmentation and a trend towards diversification and uniformity within the Causses and Cévennes Heritage Site. As depicted in Figure 8 and detailed in

Table 5. Statistical characteristics of the landscape index in different periods.

Landscape Metrics	1985	2010	2020	Trend
Shannon’s Diversity Index (SHDI)	0.661	0.811	0.825	↑↑
Perimeter-Area Fractal Dimension (PAFRAC)	1.453	1.468	1.452	↑↓
Total Edge Contrast Index (TECI)	22.272	21.369	21.106	↓↓
Patch Density (PD)	32.206	38.513	34.499	↑↓
Contagion Index (CONTAG)	64.353	57.231	57.554	↓↑
Largest Patch Index (LPI)	72.742	63.475	62.886	↓↓

Note: “↑↑” signifies a sustained upward trend; “↑↓” signifies an initial rise followed by a decline; “↓↑” signifies an initial fall followed by a rise; “↓↓” signifies a sustained downward trend.

, the largest patch index showed a continuous decline from 72.7% in 1985 to 62.9% by 2020, suggesting a gradual decrease in the dominance of the prevalent landscape types and a weakening control of woodland over the overall landscape. Patch density peaked at 38.5 in 2010 and slightly decreased to 34.5 by 2020; however, the number of patches per unit area still increased, contributing to landscape fragmentation. The perimeter area fractal dimensions for 1985, 2010, and 2020 were 1.45, 1.47, and 1.45, respectively, indicating relatively simple patch shapes with no significant changes over time. A continuous decline in total edge contrast ratios across periods reflects a reduction in the number of artificial boundaries. The continuous rise in Shannon’s diversity index signifies an increase in landscape-type diversity, with landscape elements developing towards diversification and enhanced natural heterogeneity. Although the contagion index exhibited an initial decline followed by an increase, it overall decreased by 6.8%, indicating a reduction in landscape connectivity, more discrete and discontinuous patches, and further exacerbating landscape fragmentation.

4.2.3. Landscape Stability Analysis

As illustrated in Figure 9, significant geographical differences in landscape stability are observed within the Causses and Cévennes Heritage Site, with higher stability in the eastern part compared to the western part. Stable landscapes are predominantly concentrated in the core area of the Cévennes region and its northeastern extension, while unstable landscapes are extensively distributed across the Causses region.

Referencing

Table 6. Statistics of landscape stability area in the different levels from 1985 to 2020.

Level	1985		2010		2020		Change in Area (km2)
	Area (km2)	Percentage (%)	Area (km2)	Percentage (%)	Area (km2)	Percentage (%)	1985–2010	2010–2020	1985–2020
Instability	3409.480	54.069	3893.360	61.743	3735.480	59.239	483.880	−157.88	326.000
Less stability	126.590	2.008	13.780	0.219	22.730	0.360	−112.810	8.950	−103.860
Relatively stability	1439.150	22.823	1400.580	22.211	1463.330	23.206	−38.570	62.750	24.180
Stability	932.600	14.790	900.960	14.288	966.950	15.334	−31.640	65.990	34.350
Extremely stability	397.960	6.311	97.100	1.540	117.290	1.860	−300.860	20.190	−280.670

, the extent of stable landscapes within the heritage site underwent a notable reduction between 1985 and 2010, with a substantial decrease of 332.5 km² in the area of stable and extremely stable landscapes. Consequently, the proportion of unstable landscapes increased markedly, rising from 54.1% to 61.7%, leading to an overall decline in the heritage site’s landscape stability level. However, in the subsequent period from 2010 to 2020, a slight increase in the stable landscapes’ area indicated a partial recovery in landscape stability level, although unstable landscapes continued to predominate, with their proportion reaching 59.2%. This change is attributed to multiple factors, including cumulative effects, the lag in ecological processes, and the impact of climate change. Therefore, although the conversion between landscape types has become more stable between 2010 and 2020, the spatial structure and configuration of the landscape have already changed due to the previous period, affecting the function and stability of the landscape.

Further analysis of the transition characteristics of landscape stability levels across different periods (Figure 10) reveals that from 1985 to 2010 (Figure 10a), the heritage site was primarily characterized by a decline in stability levels. Notably, significant reductions in landscape stability levels occurred in the core area of the Cévennes and parts of the northwestern Causses, while the majority of other areas within the Causses remained relatively stable. Moving into the period from 2010 to 2020 (Figure 10b), subtle changes occurred in the transition characteristics of landscape stability within the heritage site. Although the stability levels of most areas remained relatively stable, a positive trend was observed towards the transformation of lower stability levels to higher stability levels at the junction of the Causses and the Cévennes. Compared with the long-term trend over the entire period from 1985 to 2020 (Figure 10c), it is evident that while there was a slight recovery in landscape stability from 2010 to 2020 (Figure 10b), the overall trend from 1985 to 2020 indicates a decline in landscape stability.

In summary, landscape stability within the Causses and Cévennes Heritage Site exhibited a declining trend, with a significant reduction in the area of strongly stable landscapes (approximately 280 km²) and the marked expansion of unstable landscapes, collectively reflecting an overall decline in the heritage site’s landscape stability level. Although the area of weakly stable landscapes partially recovered to the stability levels observed at the beginning of the study, the overall trend indicates a decrease in landscape stability.

4.3. Implications of Climate Change for Landscape Stability

Employing the Geographical Detector, this study integrated landscape stability data from the periods of 1985, 2010, and 2020 into the analytical model. Our aim was to assess the extent to which climatic factors contribute to variations in landscape stability across the Causses and Cévennes Heritage Site. Given the long-term cumulative response characteristics of landscape changes to climatic factors, this study employed the five-year cumulative averages of temperature, precipitation, potential evaporation, and relative humidity as key climatic factor sequences. This method enables a more thorough evaluation of how climate change affects the stability of landscapes.

Figure 11 presents the findings from the factor detection analysis. Although the intensity of the influence exerted by climatic factors on landscape stability varied across different periods, temperature, precipitation, potential evaporation, and relative humidity, all demonstrated strong explanatory power during the study period. Moreover, all driving factors were found to be statistically significant at the p < 0.01 threshold, which fully validates the model’s effectiveness and accuracy, thereby enhancing our understanding of the spatial determinants influencing landscape stability levels. Further analysis of the contribution of each factor revealed that precipitation is the most significant factor affecting landscape stability. Its contribution rate remains above 0.3 in all years, showing a significant dominant role. In addition, relative humidity and potential evaporation also exert a strong influence on landscape stability levels. In contrast, although temperature contributes to the distribution of landscape stability, its influence is relatively limited due to its lower contribution rate.

An interaction detection analysis was carried out in this study to elucidate the complex interrelationships between climatic factors, thereby enhancing our comprehension of their interactions. As detailed in

Table 7. The interactive effects of climatic factors on landscape stability across different periods.

Year	Factors	Temperature	Precipitation	Relative Humidity	Potential Evaporation
1985	Temperature	0.110
	Precipitation	0.547	0.321
	Relative humidity	0.398	0.494	0.248
	Potential evaporation	0.587	0.543	0.457	0.366
2010	Temperature	0.132
	Precipitation	0.617	0.439
	Relative humidity	0.506	0.541	0.353
	Potential evaporation	0.582	0.538	0.533	0.321
2020	Temperature	0.119
	Precipitation	0.565	0.397
	Relative humidity	0.488	0.491	0.300
	Potential evaporation	0.577	0.495	0.505	0.262

, the interactive effects of climatic factors exerted a more pronounced effect on landscape stability’s spatial distribution than did the factors acting in isolation. These effects showcased characteristics of either bivariate enhancement, where any pair of interactive effects surpassed the influence exerted by a single factor, or nonlinear enhancement, indicating a more complex, potentially exponential response. Notably, interactions involving temperature with other factors predominantly displayed nonlinear enhancement, while interactions among the other factors primarily exhibited bivariate enhancement. These findings underscore the intricate and significant dynamics at play among climatic factors, highlighting their collective influence on landscape stability.

5. Discussion

5.1. Climate Dynamics and Their Regional Manifestations

5.1.1. General Climate Change Trends in the Heritage Site

Over the last several decades, the Mediterranean region has undergone significant climate changes. Meteorological observations [17,56] and the results of model simulations [57,58] collectively indicate that the area has experienced a pronounced trend towards aridity and warming. The increasingly severe climate change dynamics pose unprecedented challenges to the conservation of the Mediterranean World Heritage Landscapes. Therefore, investigating the mechanisms of climate change impacts on heritage landscapes within the Mediterranean area is imperative and of great importance.

Our study shows that the Causses and Cévennes Heritage Site has exhibited an overall warming and drying climate pattern over the past 36 years. The temperature within the heritage site has increased significantly, with a warming trend reaching +0.33 °C per decade, which is higher than the average global terrestrial warming level. However, it is slightly lower compared to the range of temperature variations observed in southern France (+0.39 to 0.67 °C/10a) [56], possibly due to the complex topographical conditions of the area, making it more susceptible to local climatic processes. In stark contrast to the temperature changes, precipitation during the study period did not exhibit a clear long-term trend. This result aligns with previous research outcomes [14,46]. Despite the North Atlantic Oscillation (NAO) index’s effect on the interdecadal variability of Mediterranean precipitation throughout the 20th century having been acknowledged [59,60,61], variations in the NAO index alone do not adequately explain long-term changes in precipitation trends. Consequently, precipitation has not universally shown a decreasing trend; it has even slightly increased in some periods [17].

Since the midpoint of the twentieth century, global warming has intensified, and climate change has shown varying degrees of impact in different regions, with notable changes in wet and dry conditions in arid regions in particular. Over several decades, the long-term progression of drought in Europe has been especially pronounced. Potential evaporation, as the main driver affecting drought changes [62,63,64], has shown a clear increasing trend. Bešťáková et al. [12] estimated the long-term changes in the European aridity index using multiple gridded datasets, and their results confirmed the consistency between the rise in potential evaporation and the exacerbation of local drought conditions. In this study, potential evaporation within the heritage site also exhibited an increasing trend, particularly between 2000 and 2020. This finding supports previous research on the long-term development of drought in Europe. Additionally, with the significant rise in temperature, although the slight increasing trend in precipitation within the heritage site, this is not enough to fully offset the complex changes in atmospheric water vapor content caused by climate warming. As a result, the relative humidity in the Causses and Cévennes Heritage Site has shown an overall decreasing trend.

5.1.2. Sub-Regional Climate Variations Based on Köppen–Geiger Classification

The contrasts in climate and topography have created different ecological zones within the heritage site. To further analyze climate change at a regional level, this study divided the Causses and Cévennes Heritage Site into three sub-regions: temperate oceanic (Cf), Mediterranean (Cs), and temperate continental (Df) climatic zones, using the Köppen–Geiger climate data to explore trend differences among different climatic regions. The results (Figure 12) show that the trends of climate factors in the Cf, Cs, and Df sub-regions are consistent with those of the whole region, but there is a certain spatial heterogeneity between the different climatic regions due to the influence of multiple factors such as topography and atmospheric circulation. The Cf region, situated on the Causses Plateau, has a mild average temperature over the years, ranging from 9 to 10 °C. Its flat topography is not conducive to the formation of uplifting air currents, resulting in relatively scarce precipitation. In contrast, the Cs region benefits from the natural barrier of the Cévennes and its favorable location adjacent to the Mediterranean coast, leading to relatively higher temperatures. The area features complex terrain with widespread mountainous regions and abundant annual precipitation. The Df region, attributed to its higher altitude, experiences lower temperatures, typically below 9 °C, and its precipitation levels align with the overall regional average.

The Mediterranean region, acting as a transitional zone between North Africa’s arid climate and Central Europe’s temperate, rainy one, is deeply influenced by mid-latitude climate systems and dry, hot air currents from North Africa. The Causses and Cévennes World Heritage Site’s complex topography further intensifies the characteristics of the Mediterranean climate, with high temperature and intense solar radiation exacerbating evaporation within the region and making the arid conditions more severe. Our study found that the potential evaporation in the Cs region significantly exceeds the regional average and other climatic regions, with correspondingly the lowest relative humidity. On the one hand, the Cs region is strongly influenced by the Mediterranean climate, where high summer temperatures and intense solar radiation significantly increase the potential evaporation. On the other hand, landscape types also have an important influence on the potential evaporation. The landscape in the region is dominated by woodland and shrubland, which indirectly increase potential evaporation by releasing large amounts of water into the atmosphere through transpiration. However, it should also be noted that the data source (ERA5-Land) may overestimate potential evaporation [12], which could also contribute to the unusually high values observed in the Cs region. In contrast, the relative humidity in the Cf region is maintained at a high level due to the oceanic climate. The Df region, characterized by both continental and mountainous climate features, has higher altitudes and lower temperatures, resulting in weaker evaporation, with relative humidity levels between those of the Cf and Cs regions.

5.2. Impacts of Climatic Regional Differences and World Heritage Site Establishment on Landscape Stability

5.2.1. Impacts of Climatic Regional Differences on Landscape Stability

From the Mediterranean region to the mountainous areas and regions influenced by the sea, the Causses and Cévennes Heritage Site exhibits a diverse range of natural environments within a limited geographical scope. However, rising temperatures and drought have had numerous adverse effects on the sustainable management and conservation of landscapes, particularly in areas with high ecological quality and significant environmental value [52]. Although climatic variables in different regions generally exhibit trends that are in sync with the overall regional levels, each climate type possesses unique characteristics. The high sensitivity of plant growth to climate change results in landscape stability manifesting differently across various climatic regions. By comparing the climatic features of each region, we have found that certain types of climate change may exacerbate or mitigate the natural succession processes of landscapes.

The results of the driving force analysis suggest that precipitation significantly influences landscape stability. Compared to other climatic factors, precipitation exhibited notable fluctuations over the study period. Specifically, from 1985 to 2010, annual precipitation underwent a fluctuating pattern of initial increase followed by a decrease. After 2010, precipitation entered a more stable phase. Figure 12b indicates that the Cs region, characterized by abundant rainfall, has a multi-year average precipitation of 1219.8 mm. When analyzing the 5-year moving average curves of precipitation across various climatic zones, the Cs region demonstrates a higher degree of precipitation variability relative to other regions. Additionally, the Mediterranean climate characteristics are intensified by the region’s topography, with temperatures and potential evaporation higher than in other climatic regions and exceeding the overall regional average. The interplay of these climatic elements and the topographical influence renders the Cs region’s landscape stability particularly responsive to climate change, manifesting as significant variability throughout the study period.

5.2.2. Impact Assessment of the Establishment of the “Causses and Cévennes” World Heritage Site in 2011

Considering the complex interplay between the long-term impacts of climate change and the short-term effects of significant events on landscape stability, we analyzed the specific short-term effects brought about by the policy change when the Causses and Cévennes region was designated as a UNESCO World Heritage Site in 2011, alongside our assessment of the former.

In 2011, the ’Causses and Cévennes’ region was inscribed on the UNESCO World Heritage List, marking the entry of this area into a new phase of conservation and development. To assess the specific impact of this policy change on landscape dynamics, we incorporated a multi-temporal data analysis approach in our study design, selecting land cover data from the years 2000, 2010, and 2020 as the primary data sources. Through this method, we were able to effectively disentangle the relationship between the short-term effects caused by the establishment of the heritage site and long-term trends. The analysis (

Table 8. Areas and proportions of the heritage site’s landscape types in different periods (2000, 2010, and 2020).

Landscape Type	2000		2010		2020		Rate of Change (%)
	Area (km2)	Percentage (%)	Area (km2)	Percentage (%)	Area (km2)	Percentage (%)	2000–2010	2010–2020
Cropland	915.188	14.471	926.802	14.654	913.5	14.444	1.269	−1.435
Woodland	4314.303	68.217	4264.94	67.437	4207.393	66.527	−1.144	−1.349
Shrubland	546.128	8.635	568.229	8.985	626.259	9.902	4.047	10.212
Grassland	525.204	8.304	533.823	8.441	542.196	8.573	1.641	1.568
Wetland	0.297	0.005	0.319	0.005	0.335	0.005	7.161	5.268
Water body	2.048	0.032	2.023	0.032	1.984	0.031	−1.228	−1.938
Impervious surfaces	21.092	0.334	28.119	0.445	32.579	0.515	33.314	15.864
Bare areas	0.093	0.001	0.099	0.002	0.107	0.002	6.206	8.447

) revealed that the establishment of the heritage site in 2011 was not the primary cause of the abrupt changes observed in the landscape type conversion trajectory (Figure 6) and spatial distribution (Figure 7b) between 2010 and 2020. However, this policy fostered changes in land use practices within local areas, such as an increased trend towards sustainable agricultural practices. Therefore, we believe that while the establishment of the heritage site is not the sole decisive factor, it has indeed played a role in shaping the current landscape pattern to some extent.

5.3. Limitations and Future Work

Increasing weather fluctuations are continually threatening the stability of heritage landscapes. An effective monitoring program should be oriented towards a variety of objectives, ensuring that it supplies the necessary information for environmental policy decisions. This study provided an example of a potentially useful tool for integrating stability into the planning and management of heritage landscapes. By analyzing the spatial relationships between landscape stability levels and climatic variables, this study evaluated the specific contributions of climate change to the World Heritage Landscape and emphasized the importance of incorporating quantitative methods that test causality. This provides powerful support for developing effective heritage conservation strategies and advancing the preservation efforts of heritage landscapes.

Grounded in landscape ecological theory, this study integrated three key indices—CONTAG, TECI, and PD—to develop a model for evaluating landscape stability. This model is instrumental in revealing the complex relationship between landscape fragmentation and its changes in relation to landscape stability, yet it is not without limitations. Firstly, the universality and scientific connotation of the model need to be further verified in landscape stability assessments in different regions. Secondly, future research should explore the application of more landscape indices to reflect the characteristics of landscape patterns and ecological processes more comprehensively. Considering that landscapes are open systems susceptible to random events and various external and internal factors, the current study is limited by the finite time span of available data and the low resolution of climate data. Future work could attempt to employ a more diversified range of data sources and technical methods for verification. For instance, the development of multi-source data fusion strategies and deep learning frameworks has made target detection in complex backgrounds a reality [65,66]. These technologies can not only provide precise data on surface cover but also offer powerful tools for accurately distinguishing between natural processes and human interventions. Particularly, the Mediterranean region, with its climatic diversity and high spatial variability, holds significant importance in global change studies. For the climate factors, this work could be further improved by the consideration of soil moisture, particularly benefiting from some meaningful efforts on soil moisture mapping from satellite passive and active observations [67]. Finally, conducting cross-regional comparative studies will help deepen the understanding of landscape stability patterns under different climatic and ecological conditions, providing a scientific basis for targeted protection strategies.

6. Conclusions

Focusing on the Causses and Cévennes World Heritage Site, this study systematically analyzed the characteristics of climate change within the heritage site from 1985 to 2020 and provided a detailed description of the evolution of the landscape pattern during this period, utilizing reanalysis of meteorological data, land cover dataset, and additional auxiliary data products. Building on this foundation, we further assessed the levels of landscape stability in the heritage site for the periods of 1985, 2010, and 2020. Finally, the Geographical Detector model was used to investigate the extent to which climatic factors affect the landscape stability’s spatial distribution across different periods. The main findings are summarized thus:

(1)

From 1985 to 2020, the Causses and Cévennes World Heritage Site experienced noticeable climate changes characterized by rising temperatures, a slight but non-significant increase in precipitation, enhanced potential evaporation, and a decrease in relative humidity. These changes exacerbated drought conditions within the region and signaled an increased risk of future droughts. This trend has been corroborated in broader studies.

(2)

The landscape types in the Causses and Cévennes Heritage Site are predominantly woodland, cropland, shrubland, and grassland. Over the 36-year period, there was a significant reduction in woodland, while the areas of shrubland, grassland, and impervious surface and bare areas increased. Cropland, wetlands, and water bodies underwent initial increases followed by decreases. Landscape-type transitions mainly occurred between woodland, cropland, shrubland, and grassland, particularly between 1985 and 2010. After 2010, transition activities were more moderate. Statistical analysis revealed increased landscape fragmentation and enhanced heterogeneity within the heritage site across the study timeline.

(3)

The overall stability of the landscape at the Causses and Cévennes Heritage Site exhibited a downward trend characterized by pronounced spatiotemporal variations. The Causses region had lower stability but was relatively stable, while the Cévennes region, although more stable, experienced fluctuations across different periods. Specifically, stability in the Cévennes region decreased from 1985 to 2010 and then rebounded slightly between 2010 and 2020.

(4)

Climatic factors significantly influenced landscape stability, with precipitation being a key factor. Although the long-term trend in precipitation was not significant, its irregularity and high volumes contributed to the expansion of unstable landscape areas. While temperature had a smaller impact, the increase in potential evaporation and the decline in relative humidity due to rising temperatures also significantly affected landscape stability. Interaction detection analysis indicated that climate change affects landscape stability through combined effects, with the interaction between precipitation and temperature being the most pronounced.

This project seeks to deliver scientific insights for sustainably conserving the Causses and Cévennes World Heritage Site while also providing a reference for in-depth analysis of the stability dynamics and spatial relationships of analogous heritage landscapes affected by climate change. The anticipated outcomes of this study include offering scientific support for the formulation of climate adaptation management strategies tailored to World Heritage Sites in the Mediterranean region, as well as illuminating conservation efforts for analogous sites across other regions worldwide.