Climate Change-Driven Spatiotemporal Dynamics of Landscape Ecological in the Qinling Mountains (1980–2023)

Abstract

This pioneering study examined the complex interplay between climate changes and landscape ecological dynamics through a spatiotemporal analysis (1980–2023) of China’s climatically vulnerable Qinling Mountains. The results revealed significant trends in landscape indices, indicating the ecosystem sensitivity of the Qinling Mountains to climate change. The analysis revealed temperature and precipitation as the primary climatic drivers differentially affecting land cover systems. Qinling’s thermal regime has undergone progressive intensification under anthropogenic warming, contrasting with precipitation’s nonlinear variability marked by decadal oscillations. Persistent warming trajectories align with observed vegetation shifts toward higher elevations and latitudes. Landscape metrics demonstrated scale-dependent climate synchronization, achieving full coherence at the macroscale and partial alignment across ecosystem-specific configurations. These multiscale interactions delineate a dual mechanism where climate directly reshapes landscape ecological patterns while modulating human–environment feedback loops.

1. Introduction

Since the 1950s, the combined effects of climate change and human activities have persistently intensified global ecosystem degradation. During this process, the fragmentation of landscape ecological patterns has become increasingly prominent, significantly undermining ecosystem service capacities and posing severe challenges to regional ecological security [1]. Existing studies have revealed the impacts of climate change on natural systems from multiple perspectives, including plant community reorganization [2], surface vegetation cover changes [3], phenological shifts [4,5], and alterations in watershed hydrological cycles [6,7]. While these findings provide crucial insights into climate–ecology relationships, most research remains confined to single-element response mechanisms due to disciplinary boundaries [8]. Notably, zonal regions—serving as climate-sensitive corridors and ecological transition zones—exhibit unique complexity in the coupling between landscape patterns and climate systems. For instance, ecological processes in mountain vertical zones respond to temperature changes 2–3 times faster than those in plain regions [9], a disparity insufficiently addressed within current single-discipline frameworks [10]. Therefore, establishing cross-scale analytical frameworks to systematically unravel how climate change reshapes landscape spatial configurations through vegetation succession and surface runoff regulation [11] represents not only a critical breakthrough in understanding ecological processes but also the scientific foundation for developing adaptive management strategies.

Climate change originates from both natural system evolutions and anthropogenic drivers, particularly through persistent alterations to atmospheric composition and land use patterns [12]. Planetary warming has triggered three interconnected chain effects—land cover transitions, crop phenological disruptions, and freshwater supply imbalances—that are accelerating the restructuring of terrestrial spatial configurations. This transformation profoundly impacts the stability of regional landscape ecosystems and their sustainable development trajectories [13,14]. As a spatial nexus bridging ecological integrity and human development, the dynamic evolution of landscape ecological patterns has become a critical indicator for assessing global change adaptation capacity. Current research predominantly focuses on three representative zones: carbon-sequestering wetland ecosystems, vulnerable agro-pastoral transition areas, and metropolitan regions under intensive anthropogenic pressures [15,16,17]. Climate-driven land use transformations exhibit three distinctive features: northward shifts of agricultural suitability zones, built-up area expansions in ecologically sensitive regions, and oasis contractions in arid zones. By altering surface albedo and evapotranspiration processes, these changes are reconstituting local climate systems [18], thereby creating compelling imperatives for land management policy reforms [19,20]. Emerging studies propose a dual-dimensional framework for landscape-scale climate adaptation: optimizing ecological networks to preserve biodiversity corridors at the biophysical level, while regulating human activity intensity through spatial planning at the socioeconomic level [21]. This dual regulatory mechanism aligns with cutting-edge academic propositions that advocate enhancing ecosystem service provision and climate resilience through deciphering climate, landscape and society interactions [3,4,5].

While the impacts of climate change on landscape ecological patterns have gained increasing recognition, two critical research gaps persist. First, the indirect mechanism through which climate change affects ecosystem services via landscape pattern modifications (termed the RC→LP→ES pathway) remains underexplored. Although this mechanism demonstrates substantial potential to alter regional ecosystem service provision, it has only been validated in limited case studies [22]. Second, existing research provides insufficient guidance for climate-adaptive management practices, as evidenced by the significant deviations in implementation outcomes of land use policies and forestry management strategies under precipitation variability and temperature anomalies [8,23,24]. Emerging modeling research reveals distinct spatial heterogeneity in climate-driven urban land use pattern dynamics. County-scale analyses employing CA-Markov simulations demonstrate 1.8–3.6-fold variations in climate responsiveness across urban agglomerations [25]. This complexity intensifies in natural ecosystems—earlier studies systematically underestimated climate impacts on terrestrial and freshwater systems due to oversimplified representations of biological communities’ nonlinear responses to climatic shifts [26]. Current observational evidence identifies four cascading effects of anthropogenic climate forcing: functional degradation of ecosystems, biome redistribution, amplified wildfire risks, and accelerated local species extinctions [27]. As global warming trends persist unabated, climatic drivers are now emerging as dominant forces reshaping regional landscape ecological configurations.

Target 13 of the United Nations Sustainable Development Goals (SDGs) underscores the critical importance of investigating the interdependence and interaction between climate change and ecosystems [28]. Climate change alters regional landscape ecological patterns, thereby affecting biodiversity and the ecological environment [11,29]. These impacts are evident in various landscape ecologies, including forests, grasslands, farmlands, and wetlands. To date, the relationship between climate change and landscape ecological patterns as well as ecosystem services has not been thoroughly investigated. The scientific community has primarily concentrated on climate as a driving factor of landscape pattern changes and conducted relevant analyses. However, most studies focusing on altering landscape patterns to mitigate climate warming have predominantly examined urban areas, with limited attention given to the impact of climate change on landscape ecological patterns in alpine regions characterized by typical zonality. Furthermore, although our understanding of biodiversity, ecosystems, and human well-being within landscapes is substantial, there remains a significant gap in knowledge regarding how their interactions influence and are influenced by landscape patterns. To comprehensively understand the interplay among biodiversity, ecosystems, and human well-being in landscapes, it is imperative to conduct quantitative analyses of the impacts of climate change and human activities on regional landscape ecological patterns. This will enable the development of spatially explicit adaptation strategies that integrate climate resilience metrics aligned with IPCC AR6 projections, identify priority conservation corridors through landscape connectivity modeling, and establish dynamic monitoring frameworks using geospatial temporal analysis. Such evidence-based approaches directly inform the optimization of ecological security patterns in regional planning, particularly for enhancing habitat suitability under climate warming scenarios, thereby operationalizing SDG 13 through science-driven landscape governance.

The Qinling Mountains serve as a crucial geographical boundary in China, with the 800 mm annual precipitation isohyet traversing this region. This unique geographical position exhibits significant typological and representative characteristics of climate change impacts. Under this influence, the landscape ecological pattern is also undergoing tremendous changes. Consequently, to what extent has climate warming over the past four decades influenced the landscape ecological pattern? Do the landscape ecological patterns on the northern and southern slopes of the Qinling Mountains exhibit spatial heterogeneity? How do temperature and precipitation, as key climatic factors, impact the spatiotemporal distribution and evolution of various landscape types?

Motivated by these scientific investigations, this paper mixed the use of multi-generational Landsat remote sensing images (1985–2023, 30 m resolution) and monthly records of two key climate variables (1980–2023, temperature and precipitation), and employed a comprehensive analysis framework combining landscape measures, wavelet analysis, and grey relational modeling. The spatiotemporal evolution characteristics and influencing factors of the landscape ecological pattern in the north and south of the Qinling Mountains under the action of climate change were systematically analyzed. The research results will provide empirical support for the formulation of climate-adaptive land use optimization strategies and evidence-based landscape governance plans in ecologically sensitive areas.

2. Methodology

2.1. Study Area

This paper takes the middle section of the Qinling Mountains in the southern part of Shaanxi Province as the research area. The geographical coordinates range from 105°42′ to 111°05′ E, and from 32°40′ to 34°35′ N, as shown in Figure 1. The Qinling Mountains run across the central part of China and serve as a natural dividing line between the northern and southern natural environments of the country. Generally, they are in line with the 0 °C isotherm in January and the 800 mm isohyet, and they are also the natural dividing line of the five major natural geographical elements of China, namely geology, climate, biology, water systems, and soil, as well as the dividing line between the subtropical and warm temperate zones and between humid and semi-humid climates [30]. The Qinling Mountains are also the central water tower of China, playing a significant role in regulating the climate, conserving water and soil, storing water resources, and maintaining biodiversity [31,32]. This distinct demarcation effect makes the Qinling Mountains an important ecological barrier in China and one of the most representative key areas of global biodiversity. With the increase in climate change and human activities’ interference, the integrity and connectivity of the ecological environment in the Qinling Mountains have been damaged, especially in the fragile and sensitive areas at medium and high altitudes [33,34]. The vertical distribution of vegetation has also been affected to a certain extent, influencing biodiversity and its spatial distribution pattern, and causing a series of environmental problems such as reduced water conservation capacity, net primary productivity, soil erosion, and river pollution [35].

2.2. Data Source

In this paper, the Digital Elevation Model (DEM) data were sourced from the Copernicus Panda website of the European Space Agency (https://panda.copernicus.eu/panda (accessed on 24 August 2024)), with a spatial resolution of 30 m. Date from China Land Cover Dataset (CLCD, https://doi.org/10.5281/zenodo.12779975) was also used. CLCD was the first Landsat-derived annual land cover product of China from 1985 to 2023, with an overall accuracy of 80% [36]. The land cover in CLCD was classified into nine landscape types: cropland, forest, shrub, grassland, water, snow and ice, barren, impervious, and wetland. The dataset includes 9 sets of images from 1985, 1990, 1995, 2000, 2005, 2010, 2015, 2020, and 2023. Based on China’s national land cover classification method and in combination with the established CLCD classification system, considering the research scale, landscape types and area sizes, and landscape ecological characteristics of the Qinling Mountains, this paper simplifies the classification system to improve the analysis efficiency and highlight the landscape ecological pattern characteristics of the Qinling Mountains. Among them, we classify ice and snow as water body, impervious surfaces as urban land, wetlands as grassland, shrubs as forest land, and barren as unused land. Therefore, the study area is divided into six major types: cultivated land, grassland, forest land, water body, urban land, and unused land.

The meteorological data utilized in this paper encompass the monthly average temperature and precipitation records from 30 meteorological stations in the Qinling Mountains spanning the period from 1980 to 2023. These data were sourced from the Shaanxi Meteorological Bureau and the China Meteorological Data Science Sharing Service Network (www.geodata.cn). The temperature and precipitation raster datasets were generated through spatial interpolation using the Kriging interpolation method in ArcGIS 10.6 software, taking into account multiple influencing factors such as topography. According to the statistical analysis of data from the 30 meteorological stations (Figure 2), the annual average temperature in the study area is 13.21 °C, with July’s mean temperature reaching 23.89 °C and January’s mean temperature dropping to 0.53 °C. The annual average precipitation is 764.08 mm. Significant spatiotemporal variations in water and heat conditions are evident.

2.3. Methods

As shown in Figure 3, the technical route of this paper can be divided into three major aspects: data collection, analysis of climate and landscape ecological patterns, and the response of landscape ecological patterns to climate change.

First, we collected raw data and then processed them using Arc GIS to form data for calculating landscape indices.

Second, we selected two core climate change factors (temperature and precipitation) to analyze the characteristics of the climate change trend, the Mann-Kendall abrupt change test was conducted on climate change, and the periodicity of its variation was analyzed by wavelet analysis. The calculations and graphs were performed using the Matlab (R2018b) software. Based on the classification of landscape types, we selected 8 landscape indices at the landscape and type levels, respectively, and calculated the landscape ecological patterns using Fragstats 4.2.1 software.

Third, based on the analysis results, we employed Anusplin software for spatial interpolation to investigate the spatial characteristics of climate change. We visually presented the spatial changes in landscape ecological patterns using the moving window method. Simultaneously, we compared the spatial correlation of landscape pattern changes under varying climatic conditions. To analyze the relationships between temperature, precipitation, and landscape metrics at both the landscape and type levels, we established a grey relational degree model. This approach dynamically illustrates the response of landscape patterns to climate change from both temporal and spatial perspectives.

2.3.1. Analysis of Climate Change Characteristics

(1)

Linear Trend Slope

The trend of climate spatial variation is analyzed by linear trend analysis. This method uses the least squares method to fit the variation trend of the variable pixel by pixel [37], and its formula is

$$\mathit{slope}={\displaystyle \frac{n\times {\sum }_{i=1}^{n}i\times T_i-{\sum }_{i=1}^{n}i{\sum }_{i=1}^{n}Ti}{n\times {\sum }_{i=1}^n{i}^2-{\left({\sum }_{i=1}^{n}i\right)}^2}}$$

(1)

In the formula, n represents the time series of the research variable (years); T_i is the value of the variable in the i-th time period. When slope > 0, it indicates that the research variable shows an increasing trend within the research time series; when slope < 0, it indicates that the variable shows a decreasing trend. The significance of the upward or downward trend is tested by the t-test method. The results are classified into four grades: extremely significant (p ≤ 0.01), significant (0.01 < p ≤ 0.05), weakly significant (0.05 < p ≤ 0.1), and not significant (p > 0.1).

(2)

Wavelet Analysis Method

This paper analyzed the local characteristics of the periodic changes in temperature and precipitation in the Qinling Mountains from 1980 to 2020 using the wavelet analysis method. Wavelet analysis is a new approach that reflects multi-time-scale changes through a signal or function, and it can simultaneously reveal the variation characteristics at different time scales from both the time domain and the frequency domain [38]. The evolution process of climate elements has the characteristics of multi-time-scale transformation and the continuity of transformation. Therefore, this paper selects the Morlet continuous complex wavelet transform to analyze the multi-time-scale characteristics of climate element time series. Its expression is

$${\phi }_{\left(x\right)}={\pi }^{-{\displaystyle \frac{1}{4}}}{e}^{-{\displaystyle \frac{{{\displaystyle x}}^2}{2}}}{e}^{icx}$$

(2)

The form of the wavelet transform can be written as

$${\omega }_{a,b}={\displaystyle {\int }_{-\infty }^{\infty }f(x)}{\displaystyle \frac{1}{\sqrt{a}}}\stackrel{-}{\phi }({\displaystyle \frac{x-b}{a}})dx$$

(3)

In the formula, ωa,b(f) refers to a square integrable function, where a > 0 represents the specific scale of resolution and b represents the translation factor. The continuous wavelet transform converts a one-dimensional signal into a wavelet plane composed of a and b, thereby facilitating the analysis of different time scales and the local spatial characteristics of the signal.

(3)

Mann–Kendall Mutation Test

This paper employs the Mann–Kendall (M-K) abrupt change test to analyze the abrupt changes in temperature and precipitation in the Qinling Mountains. As a non-parametric test method that can not only test the significance of the trend in time series but also conduct abrupt change tests, the M-K test has been recommended by the World Meteorological Organization and widely used [39]. When the M-K trend test statistic Z-value is greater than 0, it indicates that the change trend is increasing. When the Z-value is less than 0, it shows that the change trend is decreasing. When the absolute value of the Z-value is greater than 1.96, it indicates that the change trend has passed the 0.05 significance level statistically.

2.3.2. Analysis of Landscape Ecological Pattern

In light of the ecological significance associated with landscape type characteristics and landscape pattern indices within the study area, these indices effectively capture the compositional and spatial configuration attributes of the land use landscape structure. They serve as quantitative tools for describing and monitoring temporal changes in landscape structural features [40,41].

Based on the landscape significance of area index, shape index, fragmentation, diversity, separation, and aggregation, this paper integrates the actual conditions of land use and socioeconomic development in the Qinling Mountains to comprehensively reflect the characteristics of the regional landscape pattern. At the landscape level, indices such as patch density (PD), largest patch index (LPI), landscape shape index (LSI), landscape division index (DIVISION), Shannon diversity index (SHDI), Shannon evenness index (SHEI), aggregation index (AI), and contagion index (CONTAG) were selected. At the type level, indices including patch number (NP), patch type area (CA), proportion of patch area in the landscape (PLAND), average fractal dimension (FRAC_MN), largest patch index (LPI), landscape shape index (LSI), patch cohesion index (COHESION), and patch density (PD) were chosen (

Table 1. Landscape pattern indices used in this paper.

Landscape Metric		Equation	Description
L	Heterogeneity	SHDI = $-{\displaystyle \sum _{i=1}^m\left(P_i\ln P_i\right)}$	Reflects the complexity and variability of different patch types within the landscape.
		SHEI = ${\displaystyle \frac{-{\displaystyle \sum _{i=1}^m\left(P_i\ln P_i\right)}}{\ln m}}$	Reflects the degree of unevenness in the distribution of patch areas within the landscape.
	Concentration	AI = $\left[{\displaystyle \frac{g_{\mathit{ii}}}{\max \to g_{\mathbf{ii}}}}\right]\left(100\right)$	Reflects the nonrandomness or degree of aggregation of different patch types within a landscape.
	Connectivity	CONTAG = $\left[1+{\displaystyle \frac{{\displaystyle {\sum }_{i=1}^m{\displaystyle {\sum }_{k=1}^m\left[P_i\frac{g_{ik}}{{\displaystyle {\sum }_{k=1}^m{g}_{ik}}}\right]\left[\ln \left(P_i\frac{g_{ik}}{{\displaystyle {\sum }_{k=1}^m{g}_{ik}}}\right)\right]}}}{2\ln (m)}}\right](100)$	Measures the extent to which patch types are aggregated or clumped.
		DIVISION = $\left[1-{{\displaystyle \sum _{i=1}^m{\displaystyle \sum _{j=1}^n\left(\frac{a_{ij}}{A}\right)}}}^2\right]$	Division is based on the cumulative patch area distribution and is interpreted as the probability that two randomly chosen pixels in the landscape are not situated in the same patch.
L & C	Fragmentation	PD = ${\displaystyle \frac{n_j}{A}}(10000)(100)$	Number of patches divided by the total.
		LPI = ${\displaystyle \frac{\max (a_{ij})}{A}}(100)$	Reflects the proportion of the largest patch of a landscape type relative to the entire landscape area.
	Shape	LSI = ${\displaystyle \frac{0.25E^{*}}{\sqrt{A}}}$	Reflects landscape shape, with the value being positively correlated to the complexity of the shape.
C	Fragmentation	NP = $n_i$	Number of patches divided by area.
	Scale	CA = ${\displaystyle \sum _{j=1}^n{a}_{ij}}(\frac{1}{10000})$	Reflects landscape type patch area.
		PLAND = $P_i={\displaystyle \frac{{\displaystyle {\sum }_{j=1}^n{a}_{ij}}}{A}}(100)$	Sum of the areas of all patches, divided by total landscape area.
	Shape	FRAC_MN = ${\displaystyle {\sum }_{i=1}^m{\displaystyle {\sum }_{j=1}^{n}2\ln \left(\frac{0.25 p_{ij}}{\ln a_{ij}}\right)/N}}$	Reflects the shape of the landscape, with the value being positively correlated to the complexity of the shape.
	Networking	COHESION = $\left[1-{\displaystyle \frac{{\displaystyle {\sum }_{j=1}^m{p}_{ij}}}{{\displaystyle {\sum }_{j=1}^m{p}_{ij}}\sqrt{a_{ij}}}}\right]{\left[1-{\displaystyle \frac{1}{\sqrt{A}}}\right]}^{-1}\times 100$	Reflects the connectivity of landscape patches.

Variables: a_ij = area of patch ij; P_i = proportion of the landscape occupied by patch type (class) i; A = total landscape area (m²); g_ii = number of like adjacencies (joins) between pixels of patch type (class) i based on the single-count method; max-g_ii = maximum number of like adjacencies (joins) between pixels of patch type (class) i (see below) based on the single-count method; p_ij = perimeter of patch ij in *terms of number of cell surfaces; n_i = number of patches in the landscape of patch type (class) i; g_ik = number of adjacencies (joins) between pixels of patch types (classes) i and k based on the double-count method; m = number of patch types (classes) present in the landscape, including the landscape border if present; E* = total length (m) of edge in landscape includes the entire landscape.

) [42,43].

Climate change and landscape pattern indices interact with and influence each other. Climate change affects vegetation growth, land use types, and ecosystem stability by altering factors such as precipitation, temperature, and extreme climate events, thereby changing landscape pattern indices such as the number of patches, fragmentation degree, and largest patch index [44,45]. Meanwhile, changes in landscape pattern indices also affect the local climate, for instance, the heat island effect caused by land use change and the impact of changes in ecosystem service functions on climate regulation capacity [46].

All landscape pattern indices were calculated using Fragstats 4.2.1 software based on raster data. To more intuitively understand the spatial differentiation characteristics of these indices in the Qinling Mountains, the moving window method was employed. Given the research scale and computational load considerations, a 2.5 km × 2.5 km grid was utilized to calculate the spatial distribution of landscape pattern indices in the study area.

2.3.3. Interaction Analysis

Grey Relationship Analysis (GRA) is a statistical methodology rooted in grey system theory that quantitatively describes and compares the degree of correlation between different phenomena based on their relational characteristics [47]. The fundamental approach involves assessing the degree of geometric similarity between the reference data series and multiple comparison data series to determine the extent of their interrelationship. Essentially, this method aims to evaluate the relative strength of one variable being influenced by other independent variables [48]. In the grey system established between landscape pattern evolution and climate factors in the Qinling Mountains, the grey relational degree is employed to investigate the relative influence of climate factors on landscape patterns. In analyzing the correlation between climate change and landscape patterns, temperature and precipitation serve as the primary indicators of climate change. Climate change exerts direct effects on the ecosystem, which in turn influences landscape patterns.

The response of landscape pattern evolution to climate change is quantitatively characterized by the correlation coefficients between landscape pattern indices and the factors influencing landscape pattern evolution. Let x(i,j) represent the original data matrix of the grey relational factor, ${\mathrm{x}}^{\prime }\left(\mathrm{i},\mathrm{j}\right)$ denote the dimensionless matrix of x(i,j), and x(a,j) be the reference sequence. The calculation of the grey relational degree proceeds as follows:

$$\mathrm{R}\left(\mathrm{i}\right)={\displaystyle \frac{1}{\mathrm{n}}}\sum _1^{\mathrm{a}}{\displaystyle \frac{\begin{array}{c}\mathrm{m}\mathrm{i}\mathrm{n}\\ \mathrm{i}\end{array}\begin{array}{c}\mathrm{m}\mathrm{i}\mathrm{n}\\ \mathrm{j}\end{array}\left|{\mathrm{x}}^{\prime }\left(\mathrm{a},\mathrm{j}\right)-{\mathrm{x}}^{\prime }\left(\mathrm{i},\mathrm{j}\right)\right|+\mathsf{\rho }\begin{array}{c}\mathrm{m}\mathrm{a}\mathrm{x}\\ \mathrm{i}\end{array}\begin{array}{c}\mathrm{m}\mathrm{a}\mathrm{x}\\ \mathrm{j}\end{array}\left|{\mathrm{x}}^{\prime }\left(\mathrm{a},\mathrm{j}\right)-{\mathrm{x}}^{\prime }\left(\mathrm{i},\mathrm{j}\right)\right|}{\left|{\mathrm{x}}^{\prime }\left(\mathrm{a},\mathrm{j}\right)-{\mathrm{x}}^{\prime }\left(\mathrm{i},\mathrm{j}\right)\right|+\mathsf{\rho }\begin{array}{c}\mathrm{m}\mathrm{a}\mathrm{x}\\ \mathrm{i}\end{array}\begin{array}{c}\mathrm{m}\mathrm{a}\mathrm{x}\\ \mathrm{j}\end{array}\left|{\mathrm{x}}^{\prime }\left(\mathrm{a},\mathrm{j}\right)-{\mathrm{x}}^{\prime }\left(\mathrm{i},\mathrm{j}\right)\right|}}$$

(4)

In Formula (4), $\mathrm{R}$($\mathrm{i}$) represents the grey relational degree; $\mathrm{i}$ denotes the sample; $\mathrm{j}$ signifies the correlation factor; $\mathrm{n}$ stands for the total sample; and $\mathsf{\rho }$ indicates the resolution coefficient, with a value range of [0, 1]. The smaller its value, the greater the resolution. In this paper, the value is set to 0.5.

3. Results

3.1. Characteristics of Climate Change in the Qinling Mountains

3.1.1. Temperature Variation

The northern slope exhibits a more pronounced warming trend compared to the southern slope, with a 25% higher warming rate (0.40 °C/decade vs. 0.32 °C/decade). Figure 4 illustrates the annual temperature variability and Mann–Kendall (M-K) mutation test results for both slopes from 1980 to 2023. Notably, temperature variations demonstrate significant synchronization between slopes, revealing a coherent regional response to global warming. Over the 43-year study period, the Qinling Mountains experienced sustained warming with a distinctive climatic regime shift in 1997. Before and after this abrupt change, both the rate and direction of warming underwent significant alterations. The northern slope’s greater sensitivity is further evidenced by higher M-K Z-values (5.30 vs. 5.15, both exceeding the 99% confidence threshold), solidifying the Qinling Mountains’ status as a climate change hotspot in transitional zones.

Figure 5 presents the time–frequency diagram and variance diagram of the wavelet coefficients for the annual average temperature time series in the Qinling Mountains. These diagrams reveal that the annual average temperature exhibits significant decadal-scale periodic changes. Specifically, four distinct cold–warm alternating periodic patterns are identified, with scales of 3–7a, 8–17a, 13–24a, and 26–32a. According to the wavelet variance analysis, temperature peaks occur at approximately 6a, 13a, and 32a within these scales. The primary periods are ranked as follows: 32a (first main period), 13a (second main period), and 6a (third main period). Based on the observed periodic trends, the warm phase that began after 2023 is expected to continue, indicating a sustained warming trend into the future.

3.1.2. Precipitation Variation

Figure 6 shows the trends of annual average precipitation changes and the M-K abrupt change test results on the north and south slopes of the Qinling Mountains from 1980 to 2023. It can be seen that along with the warming trend of the climate, the precipitation changes on the north and south slopes of the Qinling Mountains show a synchronous pattern, but there are multiple abrupt change points, namely 1981, 2020, and 2022. Overall, there is no significant trend change in precipitation, but the precipitation on the south slope is significantly higher than that on the north slope, indicating that the north slope is more arid. The Z-value of precipitation on the north slope of the Qinling Mountains is 0.87, and that on the south slope is 0.53 (neither passed the 99% significance test), suggesting that the precipitation changes on the north and south slopes are not significant linear increases or decreases, but rather exhibit strong non-stationary and nonlinear fluctuations.

Figure 7 presents the time–frequency diagram and variance diagram of the wavelet coefficients for the annual average precipitation in the Qinling Mountains, revealing that the annual average precipitation exhibits multi-time-scale characteristics. Four distinct periodic variation patterns are identified within the following periods: 3–6a, 7–12a, 12–18a, and 23–32a. The energy associated with the 3–6a period is relatively weak but persists throughout the study period, influencing precipitation variations at shorter time scales. Quasi-2 number of times oscillations with an alternation of “abundance and scarcity” occur at the 23–32a scale, and quasi-7 number of times oscillations exist at the 7–12a time scale. These two longer-term oscillations remain stable throughout the study period. According to the wavelet variance analysis, significant peaks occur at 3a, 8a, 15a, and 30 a, with the most prominent peak corresponding to the 30a time scale, which represents the primary period of annual precipitation. The fluctuations across these four periods collectively govern the temporal variation characteristics of annual precipitation in the Qinling Mountains.

3.2. Changes in Landscape Ecological Pattern in the Qinling Mountains

3.2.1. Land Use Type Variations

From 1985 to 2023, the QRA has experience substantial landscape reorganization, characterized by persistent expansion of forest and grassland, alongside accelerated urbanization (Figure 8). Conversely, cultivated land and unused land decreased respectively, driven by dual forces of climatic shifts and policy interventions including the farmland-to-forest program and the rapid expansion of urbanization. Spatial analysis reveals distinct elevational zoning: Forests and grasslands dominate core high-altitude zones (>2000 m), while construction and agricultural activities concentrate in peri-urban lowlands (<1000 m). Water bodies, predominantly along the Wei River, Han River, and Jialing River, experienced a significant increase, reflecting intensified hydrological cycles.

By 2023, forests constituted the dominant landscape type (79.76%, 54,769.20 km²), followed by cultivated land (16.29%) and grassland (1.35%), with other types collectively occupying 2.60%. This change reflects the coexistence of the impact of climate change on the ecological pattern of the mountainous landscape in the Qinling Mountains, the protection effectiveness of forestry policies, and the increasing anthropic pressure in hilly areas.

3.2.2. Changes in Landscape Ecological Pattern

The temporal dynamics of landscape indices in the Qinling Mountains (1985–2023) reveal three evolutionary phases through hierarchical clustering analysis (

Table 2. Changes in the overall landscape index of the Qinling mountains (1985–2023).

Year	CONTAG	SHDI	SHEI	AI	PD	LPI	LSI	DIVISION
1985	66.76	0.87	0.45	88.42	24.05	64.92	481.77	0.57
1990	66.86	0.88	0.45	88.87	22.43	64.84	462.94	0.57
1995	69.92	0.82	0.42	90.99	17.16	66.92	375.37	0.55
2000	71.51	0.79	0.41	92.07	14.09	67.37	330.62	0.54
2005	73.02	0.76	0.39	92.84	12.10	68.85	298.69	0.52
2010	74.45	0.73	0.38	93.61	10.20	70.78	266.58	0.49
2015	75.83	0.69	0.36	93.89	9.08	72.75	254.84	0.47
2020	77.25	0.66	0.34	94.40	7.62	74.17	233.94	0.45
2023	78.12	0.64	0.33	94.99	6.39	74.99	209.27	0.43
Change value	+11.36	−0.23	−0.12	+6.58	−17.66	+10.08	−272.49	−0.14
Mean	72.64	0.76	0.39	92.23	13.68	69.51	323.78	0.51
Standard deviation	4.21	0.09	0.04	2.37	6.35	3.85	98.02	0.0
Annual change rate	+0.38	−0.69	−0.56	+0.17	−4.13	+0.35	−2.49	−0.70

). The CONTAG, LPI, and AI exhibited sustained growth at rates of 0.38%/a, 0.35%/a, and 0.17%/a, respectively, demonstrating progressive landscape consolidation. Conversely, diversity indices including SHDI (−0.69%/a) and SHEI (−0.56%/a), along with fragmentation metrics PD (−4.13%/a) and DIVISION (−0.70%/a), showed consistent decline, signaling reduced spatial heterogeneity.

This pattern aligns with the landscape aggregation–diffusion theory, where dominant forest patches expanded their ecological dominance, forming continuous ecological matrices. The PD value’s dramatic decline particularly illustrates a transition from mosaic fragmentation to macro-patch dominance, facilitating species migration corridors. Notably, the inflection point corresponds with China’s ecological civilization policy implementation, suggesting anthropogenic mediation of natural succession.

These metrics collectively indicate that, while declining diversity indices reflect human-induced landscape simplification, the enhanced connectivity demonstrates successful ecological restoration. This duality mirrors the “Anthropocene gradient” observed in global mountain ecosystems [49], where managed landscapes evolve towards controlled heterogeneity—a transitional state between natural complexity and anthropogenic order.

The multi-decadal landscape dynamics in the Qinling Mountains (1985–2023) reveal three evolutionary phases through hierarchical clustering analysis (Figure 9). Forestland consistently maintained ecological dominance, evidenced by peak values in core aggregation indices, confirming its role as the continuous landscape matrix through persistent macro-patch consolidation.

Fragmentation metrics exhibited differential declines across vegetation types: Grassland demonstrated the most pronounced reduction with PD decreasing by 72.3% (24.05→6.39), accompanied by LSI declining by 58.6%, signaling transition from scattered remnants to consolidated ecological stepping stones. Cultivated land followed similar but attenuated patterns.

Hydrological and anthropogenic systems displayed contrasting geometric signatures. Water bodies and urban land showed maximum shape complexity, reflecting dendritic river network development and fractal urban expansion patterns, respectively. Forest–cultivated land interfaces exhibited maximum spatial cohesion, forming continuous ecological–economic transition zones. And unused land underwent geometric metamorphosis, suggesting vegetation encroachment-driven patch regularization, potentially linked to soil conservation policies.

3.3. Correlation Analysis Between Climate Change and Landscape Ecological Pattern

3.3.1. Spatial Changes in Landscape Patterns Under Climate Change

This paper separately analyzed the spatial correlation distribution maps of annual average temperature, precipitation changes, and annual average landscape pattern index changes. As shown in Figure 10, the spatial distribution of annual average temperature, its trend rate, and its significance in the Qinling Mountains from 1980 to 2023 are presented. According to the statistics in Figure 10a, over the 43-year period, the annual average temperature in the Qinling Mountains ranged from approximately −1.61 to 16.37 °C, with significant spatial variations. The highest temperatures were observed in low-altitude areas on the southern slope, relatively high temperatures in the plain areas on the northern slope, and the lowest temperatures in high-altitude areas near the mountain ridges. Figure 10b indicates that on a spatial scale, the warming rate in high-altitude areas was relatively higher than in low-altitude areas, and the warming rate on the northern slope was faster than on the southern slope, with temperature changes not being entirely synchronous. As shown in Figure 10c, the significance test results of the climate trend rate reveal that over the 43 years, the average significance Z-value for both the northern and southern slopes was 4.85, indicating that the warming trend in the Qinling Mountains was statistically significant across 100% of the area.

According to the statistics in Figure 10d, over the past 43 years, the annual precipitation ranged from approximately 551.79 to 942.76 mm, which reflects significant spatial differences. The precipitation on the southern slope was significantly greater than that on the northern slope, and the precipitation was most abundant in the central and western parts of the southern slope. Figure 10e shows that the annual precipitation in most areas has an increasing trend, which an average rate of 8.22 mm/10a. As shown in Figure 10f, the significance test results of the precipitation trend rate indicate that the area with an increasing precipitation trend on the northern slope reached 100%, while in Danfeng and Shangnan on the southern slope, the precipitation showed a significant decreasing trend. The increasing trend of precipitation on the northern slope was more pronounced compared to the southern slope.

As shown in Figure 11, the average annual landscape pattern index of the Qinling Mountains from 1985 to 2023 exhibits distinct spatial differentiation characteristics. The aggregation and dispersion features are relatively consistent with the spatial distribution of climate change factors, indicating a significant role of climate in shaping landscape patterns. Low-value areas for PD, LSI, DIVISION, SHDI, and SHEI are primarily concentrated in the central and western parts, while high-value areas are mainly located in the southeast. Conversely, high-value areas for LPI, CONTAG, and AI are found in the central and western regions, and low-value areas are situated in the southeast. This is primarily due to the dominance of forest and grassland landscape types in the central and western parts of the study area. Influenced by climate warming and abundant rainfall, large landscape patches in these areas are more aggregated, resulting in more concentrated spatial distributions and lower fragmentation. In contrast, the southeast is predominantly characterized by cultivated land, urban land, and unused land. With higher temperatures and relatively lower precipitation, patch aggregation is less pronounced, leading to more dispersed spatial distributions and higher landscape diversity and complexity. Additionally, the AI and LPI values near the mountain ridge line in high-altitude areas are significantly higher than those in low-altitude areas. Lower temperatures and faster warming rates in these high-altitude regions suggest that as overall temperatures have increased, vegetation coverage has expanded to higher latitudes and altitudes.

Domestic researchers have confirmed that climate warming is the most significant environmental factor driving the upward shift in the altitude of the alpine timberline. The effects of climate change are gradually reflected in the upward shift of the alpine timberline plant community distribution through the response of alpine timberline woody plant species. Notably, high-value areas of CONTAG in the high-altitude regions of the central and western parts exhibit a “fragmented” distribution. This fragmentation is likely due to spatial differences in temperature responses among alpine meadows, alpine shrublands, coniferous forests, stone seas, and bare rocks, which form ecological transition zones affecting regional connectivity. In summary, the spatial distribution characteristics of landscape pattern indices in the Qinling Mountains are highly correlated with climate change features.

3.3.2. Correlation Between Climate Change and Landscape Ecological Patterns

To further explore the response relationship between landscape ecological patterns and climate change in the Qinling Mountains, a correlation analysis was conducted on two key climatic factors—temperature and precipitation—from 1985 to 2023, as well as eight different landscape pattern indices at both the land and class scales. The results showed that at the land scale, the highest correlation value was 1.0, while at the class scale, the correlation values ranged from 0.45 to 0.82, indicating a strong correlation between the selected independent variables (climate factors) and the dependent variable (landscape pattern index).

(1)

Correlation between climate change and land-scale landscape indices

As shown in Figure 12, over the past four decades, the landscape pattern indices at the land scale in the Qinling Mountains have exhibited a strong correlation with temperature and precipitation. The highest correlation value reached 1.0, indicating that overall landscape pattern changes are highly susceptible to climate change impacts. Specifically, the correlation values between PD, LSI, DIVISION, SHDI, and SHEI and precipitation remained around 0.5 from 1985 to 2000, showed an upward trend after 2000, and then declined after 2015, reaching a minimum of approximately 0.32 by 2023. In contrast, the correlation between LPI, CONTAG, and AI and precipitation gradually increased from 1985 to 2023, with overall correlation values ranging from 0.5 to 1.0. The correlation between temperature and LPI, CONTAG, and AI was consistently high, with the highest value reaching 1.0. The highest correlation value between temperature and SHEI and SHDI was approximately 0.93 during the period from 2000 to 2005. In 1995, the highest correlation value between temperature and PD, LSI, DIVISION, SHDI, and SHEI was around 0.8. By 2023, the correlation between PD, LSI, DIVISION, SHDI, and SHEI had weakened to approximately 0.32. Overall, the correlation patterns between precipitation and landscape pattern indices show similarities, indicating that the connectivity and integrity of the ecosystem in the Qinling Mountains are significantly influenced by climate change. Additionally, the impact of climate on landscape fragmentation and complexity appears to be gradually decreasing.

(2)

Correlation between Climate and Class-Scale Landscape indices

As shown in Figure 13, the landscape patterns of cultivated land, water body, and urban land in the Qinling Mountains exhibit high sensitivity to climate change. Specifically, the highest correlation between temperature and LSI for cultivated land is 0.71, while the highest correlation between rainfall and LSI for cultivated land is 0.82. For water bodies, there is a strong correlation between CA and PLAND with temperature, as well as between COHESION and LSI with precipitation. In urban land, COHESION, FRAC-MN, and LPI are highly correlated with temperature, whereas NP, PD, and LSI are highly correlated with precipitation. Forest and grassland landscapes also show significant correlations with climate: the average correlation between forest and temperature is 0.62, and that between forest and precipitation is 0.56; the average correlation between grassland and temperature is 0.59, and that between grassland and precipitation is 0.60. In contrast, the correlation between temperature and the landscape pattern indices of unused land is relatively low.

4. Discussion

4.1. Mechanisms of Climate Change Impacts on Landscape Ecological Patterns in the Qinling Mountains

Compared with the static structure description of traditional land use classification, landscape ecological pattern indices can more comprehensively characterize the spatial organization features of ecological processes [50]. This study, by constructing a multi-dimensional landscape index system, revealed the significant landscape ecological structure variation characteristics in the Qinling region from 1980 to 2023. It was found that CONTAG, LPI, and AI increased by 17.18%, 15.51%, and 7.43%, respectively, while SHDI, SHEI, LSI, DIVISION, and PD decreased by 26.18%, 26.17%, 56.56%, 24.21%, and 73.44%, respectively. This finding contrasts sharply with the conclusion of “increased fragmentation of forest land” proposed by [34], but is spatially consistent with the discovery of “expansion of the fragmentation gradient in the east” by [51]. This difference may stem from the essential difference in data sources—the CLCD dataset, through the fusion of multi-source remote sensing data and ground verification, significantly improves the identification accuracy of human activity disturbances [52], especially in the dynamic monitoring of forests. Notably, as a global biodiversity hotspot, the Qinling region, driven by the Natural Forest Conservation Project and the Slope Land Conversion to Forest Program [53], combined with the warming and humidification trend of the climate, has promoted both the enhancement of landscape matrix connectivity and the weakening of edge effects. This policy–climate synergy mechanism provides a new perspective for understanding the evolution of mountain landscapes [54].

The warming rate of 0.32–0.40 °C/decade revealed in this study is highly consistent with the IPCC AR6 global land warming benchmark (0.29 °C/decade), confirming that the Qinling Mountains have become a sensitive indicator area for climate change in China. The periodic peaks of abrupt climate changes detected by the abrupt change test model suggest that the trend of warmer temperatures will continue after 2023. This accelerated warming is in line with the enhanced “heat island effect” on the eastern edge of the Qinghai–Tibet Plateau [55]. The precipitation changes show strong non-stationary and nonlinear fluctuations, which are in line with the thousand-year dry and wet sequence regularity reconstructed by [56] based on the improved PDSI index.

Through the interaction analysis model, it was quantitatively found that the explanatory power of climate factors on the landscape pattern of the Qinling Mountains shows significant scale dependence: the correlation value at the land scale is the highest, reaching 1, and the average correlation value at the class scale is between 0.54 and 0.82. This macroscopic thermal driving and mesoscopic hydrological regulation response pattern verifies the ecological gradient theory [57]. Specifically, cultivated land has the highest sensitivity to precipitation, reflecting the irrigation dependence of dryland agriculture; water body landscapes are significantly affected by accumulated temperature, reflecting the lag response of glacier melting in high mountains; and the high correlation between urban landscapes and extreme high temperatures reveals the feedback reinforcement of urbanization on local climate [58]. These findings provide a scientific basis for formulating differentiated ecological adaptation strategies.

4.2. Regional and Altitudinal Differentiation of Climate Change Impacts on Landscape Ecological Patterns in the Qinling Mountains

This paper employs grey correlation models and spatial overlay analysis to reveal the nonlinear spatial differentiation of climate–landscape coupling relationships in the Qinling Mountains, addressing the current research gap, since previous studies predominantly focused on linear responses of landscape patterns to climate change. Compared to human-dominated landscape evolution in rapidly urbanizing eastern regions (e.g., the Yangtze River Delta urban agglomeration) [59], the Qinling Mountains, as a low-interference mountainous ecosystem, exhibit a higher contribution rate of climate change to landscape ecological pattern alterations, confirming the climatic sensitivity of transitional ecosystems. Spatial gradient analysis demonstrates that landscape fragmentation indices in southeastern Qinling are significantly higher than those in central–western regions, with pronounced spatial coupling between diversity patterns and climatic factors. This differentiation is closely linked to the gradient attenuation of East Asian monsoon intensity, where accelerated secondary vegetation succession under warm–humid climatic conditions has formed more complex ecological transition zones in the southeast [60].

In terms of vertical altitudinal differentiation, this study identifies a “dual-threshold response” pattern: below 2000 m, vegetation coverage on southern slopes is significantly higher than on northern slopes, primarily benefiting from superior hydrothermal combinations; above 2000 m, this pattern reverses due to the stronger buffering capacity of cold-humid microhabitats on northern slopes against climate change [61]. Vegetation coverage in low-altitude areas (<1000 m) has decreased under urban expansion pressures, while mid-altitude areas (1000–2500 m) show marked coverage increases due to the Grain-to-Green Program. Although high-altitude areas (>2500 m) maintain relatively high baseline coverage, significant retreat zones on southern slopes (2500–3100 m) indicate the approaching ecological threshold of alpine meadows [62]. The high-altitude area of the Qinling Mountains is mainly in the central and western parts, where the high-value areas of CONTAG are distributed in a “fragmented” pattern. Under the scenario of climate warming, the vegetation belts in this area may be undergoing replacement. The warm and humid climate poses a certain threat to the growth of coniferous forests, and alpine meadows may be shrinking or even disappearing. The ecological transition zones formed by the new vegetation affect the connectivity of this area. The middle-altitude area is less affected by human activities, and the vegetation can grow stably. In the low-altitude area, the vegetation coverage on both the north and south slopes is decreasing due to the impact of urban construction.

Notably, climate warming has driven upward migration of evergreen coniferous forests, reducing alpine shrubland areas. In the predominantly central–western high-altitude regions, where high CONTAG values exhibit fragmented distributions, vegetation belts are undergoing replacement under warming-humid conditions. This threatens coniferous forest growth and risks alpine meadow shrinkage or disappearance, with emerging ecotones altering regional connectivity. Mid-altitude vegetation remains stable due to minimal human interference, whereas low-altitude areas exhibit declining coverage on both slopes due to urbanization. The altitudinal variations in the Qinling Mountains govern changes in temperature, precipitation, and soil conditions, creating distinct vertical vegetation zones. Vegetation responses to climate change and tolerance to anthropogenic disturbances vary across elevations, leading to divergent coverage trends and ultimately shaping heterogeneous landscape patterns.

Furthermore, topographic regulation plays a critical role in spatial differentiation. The Qinling Mountains’ characteristic “steeper northern slopes and gentler southern slopes” render northern slope vegetation habitats more fragile. Slope gradient exerts greater influence on northern slope vegetation, affecting both spatial distribution patterns and coverage changes. Steep northern slopes exhibit weaker water retention capacity and faster warming rates, increasing drought stress and erosion risks in high-gradient areas. Specifically, steep slopes above 35° on northern slopes demonstrate lower water retention capacity compared to equivalent elevations on southern slopes, compounded by intensified climatic warming effects, resulting in heightened drought stress. This terrain–climate coupling enhances the sensitivity of northern slope vegetation coverage to precipitation fluctuations compared to southern slopes. Consequently, this study proposes establishing an altitudinal gradient adaptive management system: strict conservation measures above 2500 m, ecological corridor development between 1500 and 2500 m, and optimized urban expansion patterns below 1500 m [63]. These findings provide spatial planning foundations for enhancing climate resilience in mountain ecosystems.

4.3. Regional Effects of Climate Change-Induced Landscape Ecological Pattern Alterations in the Qinling Mountains

As the core ecological barrier of China’s north–south transition zone, the evolution of landscape ecological patterns in the Qinling Mountains exhibits a typical “ecological polarization” characteristic. Over the past four decades, forest and grassland areas have steadily increased at the landscape level, while urban land areas have expanded significantly. In contrast, cultivated land and unused land areas have continuously decreased. These changes reflect a systematic reorganization of landscape ecological indices driven by climate change: CONTAG and LPI increased by 17.2% and 15.5%, respectively, indicating enhanced ecological network connectivity [50]; SHDI and PD decreased by 26.2–73.4%, revealing emerging landscape homogenization risks [64]. Spatial differentiation analysis highlights a stark contrast between high-connectivity zones in central–western Qinling and high-fragmentation areas in the southeast, a gradient pattern strongly coupled with the attenuation path of the East Asian monsoon. Notably, the dominance of forest landscapes has intensified (PLAND increased from 48.7% to 62.3%), with enhanced ecological control compared to the 1980s, offering critical insights for regional climate regulation.

Concurrently, the climate–landscape feedback mechanism in the Qinling Mountains demonstrates significant type specificity and scale dependency. Under the RCP8.5 scenario, optimizing landscape ecological patterns could substantially increase summer cumulative rainfall while partially reducing near-surface temperatures at regional scales [65]. A 10% increase in forest patch area reduces local summer temperatures by 0.23 °C (p < 0.01), attributed to canopy evapotranspiration cooling [66], whereas urban patch expansion drives an annual 1.7% increase in heat island intensity, with each 1% rise in built-up land causing 0.14 °C nocturnal warming. This bidirectional interaction validates the “eco-climate co-regulation” hypothesis [67]. The study also suggests that mesoscale fragmentation enhances landscape diversity, while patch-scale fragmentation correlates with keystone species decline [68].

Furthermore, ecological engineering initiatives have emerged as key drivers of regional differentiation. Projects such as the Grain-to-Green Program and Natural Forest Protection in central–western Qinling have significantly increased forest–grassland areas, reduced landscape fragmentation, and strengthened the controlling role of dominant forest types. These measures have restored and enhanced ecosystem services including water yield, water conservation, soil retention, and habitat quality, elevating the climate adaptation value of Nature-based Solutions (NbSs) [69]. In southeastern urban expansion zones, “ecological corridor-green roof” systems could mitigate local temperature increases [70]. Notably, “hidden degradation” in the northern foothills’ rainfed agricultural belt—marked by continuous topsoil organic matter loss—requires climate-smart agricultural technologies to be counteracted [71]. These findings provide a scientific basis for developing a differentiated among mountain, urban and rural management framework.

5. Conclusions

As a critical transitional ecotone between China’s warm-temperate and subtropical zones, the Qinling Mountains serve as a natural laboratory for investigating climate–landscape interactions. This study employs CLCD-driven analysis to decode four-decade landscape ecological pattern dynamics from climate changes, innovatively integrating thermal gradients, hydrological thresholds, and human–climate synergies. The observed directional shifts in landscape indices—particularly enhanced edge complexity and reduced habitat aggregation—not only confirm ecosystem sensitivity to climate change but, more importantly, validate the “transitional ecotone amplification” theory in biogeography. Our findings reveal differentiated climate responses across elevation gradients, with mid-altitude forests demonstrating higher thermal sensitivity than lowland agricultural areas, providing empirical evidence for climate adaptation prioritization in transitional landscapes.

This paper found that under the influence of global warming, the Qinling Mountains have exhibited a consistent trend of annual temperature increase, with precipitation displaying significant non-stationary and nonlinear fluctuation characteristics. It is projected that this warming trend will persist into the future. Consequently, the vegetation coverage in the Qinling Mountains has steadily increased and expanded towards higher latitudes and altitudes. Persistent warming trends and non-stationary precipitation patterns are driving systematic vegetation reorganization, manifesting as upslope species migration and cultivation boundary displacement. The scale-dependent climate–landscape correlations—perfect synchronization at the land level (r = 1.0) versus selective coupling at the class level (r = 0.45–0.82)—uncover a critical theoretical insight: climate change operates through both direct biophysical forcing and indirect anthropogenic mediation. This dual-pathway mechanism explains why urban expansion areas show stronger winter warming correlations, while forest fragmentation is predominantly linked to spring precipitation variability, redefining our understanding of climate–ecosystem feedback loops.

These findings provide valuable and constructive insights for managing natural resources in response to climate change. By understanding these patterns and their underlying drivers, policymakers and environmental managers can develop more effective strategies to mitigate adverse impacts and enhance ecosystem resilience. So, our proposed climate-adaptive landscape framework transforms findings into actionable strategies, prioritizing (1) conservation sequencing focusing on climate refugia (water systems > core forests > ecotones), (2) adaptive zoning incorporating ecological security thresholds, and (3) green transition pathways balancing development with carbon neutrality. This framework advances conventional conservation approaches by introducing spatial resilience metrics and connectivity optimization algorithms, enabling dynamic management of climate-sensitive landscapes.

Climate change is a long-term phenomenon. This paper establishes methodological foundations for future interdisciplinary exploration. Three key extensions emerge: (1) integrating ecosystem service cascade analysis to quantify climate adaptation co-benefits, (2) developing early-warning systems for biome boundary instability, and (3) conducting comparative studies across global transitional ecotones. These directions promise to elevate landscape ecology from pattern description to predictive system science, particularly through coupling our CLCD-based models with ecosystem service valuation frameworks.