Article

Spatiotemporal Dynamics of Landscape Pattern and Vegetation

Ecological Quality in Sanjiangyuan National Park
Xiangbin Peng 1,2,3 , Ruomei Tang 1,2,3, *, Junjie Li 4 , Huanchen Tang 4 and Zixi Guo 1,2,3

1 College of Art & Design, Nanjing Forestry University, Nanjing 210037, China; [email protected] (X.P.);
[email protected] (Z.G.)
2 Jinpu Research Institute, Nanjing Forestry University, Nanjing 210037, China
3 Digital Innovation Design Center, Nanjing Forestry University, Nanjing 210037, China
4 College of Fashion and Design, Donghua University, Shanghai 201620, China; [email protected] (J.L.);
[email protected] (H.T.)
* Correspondence: [email protected]

Abstract: As one of China’s largest national parks, Sanjiangyuan National Park (SNP) plays
a crucial role in preserving ecological security and biodiversity. Conducting a scientific
evaluation of dynamic changes in vegetation ecological quality and landscape patterns
within the park is essential for ensuring its sustainable development and conservation
as a national ecological security barrier. This study analyzed the spatial and temporal
dynamics of vegetation ecological quality index (VEQI) and Landscape Pattern Metrics
(LPM) in SNP using the VEQI model and Fragstats 4.2.1, along with spatial correlation
analyses spanning from 2007 to 2022. The findings indicated an overall upward trend in
VEQI, with a notable increase of approximately 38.88% over the 15-year period. Particularly
in the Yangtze River Source Park, VEQI exhibited the most significant increase, reaching
48.99%. Furthermore, forest and shrub cover types displayed higher VEQI values and
demonstrated an increasing trend, signifying significant ecological improvement in these
ecosystems. Regarding landscape patterns, patch density (PD) and landscape shape index
(LSI) demonstrated an increasing trend, while average patch area and edge density (ED)
gradually decreased, indicating a rising level of landscape fragmentation. High values of
the largest patch index (LPI) were primarily concentrated in the Lancangjiang source park,
Academic Editor: Richard Ross the Yellow River source park, and the southern part of the Yangtze River source, suggesting
Shaker greater ecological connectivity in these regions. Spatial autocorrelation analysis between
Received: 29 November 2024 VEQI and LPM revealed significant spatial heterogeneity. Specifically, VEQI exhibited
Revised: 30 December 2024 positive correlations with LPI and mean patch area, while showing negative correlations
Accepted: 31 December 2024 with PD, ED, PR, TE, NP, and mean shape index. This indicates that areas with lower
Published: 6 January 2025
vegetation ecological quality tend to exhibit higher landscape fragmentation and complexity.
Citation: Peng, X.; Tang, R.; Li, J.; The study’s findings highlight the increasing trend in VEQI and changing landscape
Tang, H.; Guo, Z. Spatiotemporal
fragmentation within SNP, offering a scientific foundation for ecological protection policy
Dynamics of Landscape Pattern and
formulation and sustainable park development.
Vegetation Ecological Quality in
Sanjiangyuan National Park.
Sustainability 2025, 17, 373. https:// Keywords: landscape pattern; national park; land use changes; spatiotemporal change;
doi.org/10.3390/su17010373 vegetation quality
Copyright: © 2025 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and 1. Introduction
conditions of the Creative Commons National parks have been established worldwide to protect significant natural re-
Attribution (CC BY) license
sources and ecological environments [1–4]. National parks, recognized as a harmonized
(https://creativecommons.org/
management model for ecological conservation and resource utilization, have gained
licenses/by/4.0/).

Sustainability 2025, 17, 373 https://doi.org/10.3390/su17010373

Sustainability 2025, 17, 373 2 of 17

widespread international acceptance. As a critical area for biodiversity conservation, the

Vegetation Ecological Quality Index (VEQI) of national parks is directly related to ecological
balance and the sustainable development of these protected areas [5,6]. The Sanjiangyuan
National Park (SNP), one of China’s first national parks, is not only the largest in the country
but also serves as a significant repository of highland biological germplasm resources and
a crucial ecological security barrier with national and global conservation importance [5,6].
The VEQI of a national park reflects a comprehensive evaluation of the vegetation status
within the park. Concurrently, Landscape Pattern Metrics (LPM) aid in the assessment and
monitoring of the integrity and connectivity of ecosystems within national parks by quan-
tifying the spatial structure of landscapes. Understanding the spatiotemporal dynamics
of LPM and VEQI within national parks, as well as their spatial correlations, is essential
for elucidating the complex relationships between VEQI changes and the landscape spa-
tial structure of national parks. This not only facilitates the identification of areas within
the SNP that exhibit vegetation degradation or health but also enables the formulation
of scientific ecological restoration and protection measures tailored to different regions.
Furthermore, it fosters a deeper exploration of the relationship between vegetation quality
and changes in landscape spatial patterns within national parks.
LPM describe the spatial arrangement and distribution characteristics of surface
features, such as forests, wetlands and grasslands resulting from the diverse sizes, shapes,
and functions of landscape patches [7–9]. Advances in geospatial information technology
have led to the sophisticated integration of tools such as ArcGIS 10.8 and Fragstats 4.2.1
for comprehensive landscape analysis. LPM have emerged as a primary quantitative
analysis tool for assessing landscape patterns, effectively summarizing and quantifying
landscape structure, spatial configuration, and other pattern characteristics. They serve as
concentrated reflections of landscape ecological status [7–9]. Presently, pertinent studies
predominantly employ patch density (PD), average shape index, aggregation index, spread
index, and Shannon diversity index within LPM [10–12]. For instance, Xiao et al. discovered
a significant correlation between landscape pattern changes and surface temperature
rise in built-up areas across 26 Chinese cities, offering valuable scientific insights for
urban sustainability and environmental planning [13]. Researchers utilized principal
component analysis and inflection point analysis to determine the most suitable LPM and
scale for Guiyang, a city in the karst mountainous region. Additionally, they examined
the spatial and temporal characteristics of the area’s landscape patterns [14]. Building on
previous research, to ensure minimal similarity among the selected Landscape Pattern
Metrics (LPMs), this study comprehensively selects eight indicators from three perspectives:
landscape fragmentation, landscape aggregation, and landscape patch richness. These
indicators include the Number of Patches, Patch Density, Largest Patch Index, Total Edge,
Edge Density, Landscape Shape Index, Mean Patch Area, and Mean Shape Index. Through
these metrics, the study aims to thoroughly explore the evolution of landscape patterns
within the SNP.
Vegetation provides essential ecological services for both humans and wildlife, such
as carbon sequestration, oxygen production, soil stabilization, and water source regulation.
The quality and growth condition of vegetation in a region are directly related to the health
of the ecosystem and biodiversity in that area. By studying vegetation, we can understand
the health status of the ecosystem and its capacity to support complex biological networks.
Utilizing the Net Primary Productivity (NPP) and Fractional Vegetation Cover (FVC)
metrics to assess the quality and growth condition of vegetation is a currently effective
method of evaluation. NPP and FVC are essential indicators of the functional services
provided by terrestrial ecosystems. These parameters are critical for assessing the ecological
quality of vegetation and its growth status. Vegetation NPP, an integral component of the
Sustainability 2025, 17, 373 3 of 17

terrestrial carbon cycle, plays a pivotal role in regulating ecological processes, determining
ecosystem carbon sink capacity, and directly mirroring the growth status and productivity
potential of vegetation communities [15]. Notable advancements have been achieved in
the study of vegetation NPP. For instance, Yang et al. utilized the FVC index to assess
vegetation growth status and extent of vegetation cover [16]. Additionally, Yang et al.
utilized MODIS NDVI data to unveil an uptrend in NPP within the Yangtze River Basin
driven by changes such as forest expansion and grassland reduction [17]. Similarly, Zarei
employed NPP analysis to examine the impacts of climate change on grassland ecosystems
in Tanzania and projected trends in grassland NPP for 2050 [18]. However, relying solely on
FVC or NPP to evaluate vegetation ecological quality in a region often resulted in subjective
outcomes. To address this issue, Qian Shuan et al. employed an inversion method using
MODIS monthly NDVI synthetic data and daily meteorological observation station data
in Chinese regions to derive vegetation NPP and FVC. Subsequently, they developed a
VEQI model that comprehensively integrates vegetation NPP and FVC, thus offering a
more objective assessment of vegetation ecological quality. Currently, this model has been
applied in relevant studies assessing regional vegetation ecosystems [19,20].
Previous studies have not addressed the VEQI and LPM in the Sanjiangyuan region.
In this study, vegetation is defined as all plant communities, including both natural veg-
etation (such as grasslands, shrubs, wetlands, and forests) and artificial vegetation (such
as plantations and cultivated lands). This comprehensive inclusion ensures that both
human-managed and natural ecosystems are analyzed to provide a holistic assessment
of vegetation ecological quality. This study focuses on SNP, analyzing the spatial and
temporal evolution of VEQI and LPM from 2007 to 2022. Utilizing the VEQI model and
Fragstats 4.2.1, along with bivariate spatial autocorrelation analysis, it examines the spatial
correlation characteristics of dynamic changes in landscape patterns and vegetation ecolog-
ical quality, as well as the changing aggregation characteristics. The primary contributions
of this research are threefold: (1) By leveraging remote sensing data, this study analyzes
the spatiotemporal dynamics of the VEQI in the SNP from 2007 to 2022. This analysis aims
to provide scientific evidence for the ecological protection of vegetation in the SNP region.
(2) Utilizing land use data from four periods (2007, 2012, 2017, and 2022) at a 1 km × 1 km
grid scale, this research comprehensively examines the spatiotemporal characteristics of
Landscape Pattern Metrics (LPM) within the SNP. The metrics include Number of Patches
(NP), Patch Density (PD), Largest Patch Index (LPI), Total Edge (TE), Edge Density (ED),
Landscape Shape Index (LSI), Mean Patch Area (MPA), and Mean Shape Index (MSI).
The findings are intended to offer significant support for landscape ecological planning
and management decision-making in the SNP region. (3) Based on the spatial correlation
characteristics between Vegetation Ecological Quality Index (VEQI) and the eight landscape
pattern indices (NP, PD, LPI, TE, ED, LSI, MPA, and MSI), this study provides policy
recommendations for biodiversity conservation and sustainable development in the SNP.

2. Materials and Methods

2.1. Research Area
SNP, located in the southern part of Qinghai Province in western China, is the source
of the Yangtze, Yellow, and Lancang rivers. The park’s geographical coordinates are
between 89◦ 50′ 57′′ to 99◦ 14′ 57′′ east longitude and 32◦ 22′ 36′′ to 36◦ 47′ 53′′ north latitude
(Figure 1). The pilot area of the Sanjiangyuan National Park (SNP) spans a total area of
123,100 square kilometers, crossing two Tibetan Autonomous Prefectures, namely Yushu
and Guoluo, encompassing Zhiduo County, Qumali County, and Maduo County, as well
as 12 townships and 53 administrative villages. The Sanjiangyuan National Park consists
of three geographically discontinuous areas: the Yangtze River Source Park, Yellow River
 Guoluo, encompassing Zhiduo County, Qumali County, and Maduo County, as well as
1217,
Sustainability 2025, townships
373 and 53 administrative villages. The Sanjiangyuan National Park consists of 4 of 17
three geographically discontinuous areas: the Yangtze River Source Park, Yellow River
Source Park, and Lancang River Source Park. This discontinuity arises because the three
parks are locatedSource Park, regions,
in distinct and Lancang
each River Sourcethe
protecting Park. This discontinuity
headwaters of a majorarises
riverbecause
and the three
parks are located in distinct regions, each protecting the headwaters
preserving unique ecological systems, such as glaciers, wetlands, and canyon landscapes. of a major river and
The scope of thepreserving
SNP, basedunique
on theecological systems, suchareas
typical representative as glaciers,
of the wetlands,
sources ofandthe canyon
Yang- landscapes.
The scope of the SNP, based on the typical representative areas
tze River, Yellow River, and Lancang River, has been optimized and integrated with the of the sources of the Yangtze
Hoh Xil National River, Yellow
Nature River,and
Reserve and five
Lancang River,areas
protected has been optimized
within and integrated
the Sanjiangyuan Na-with the Hoh
Xil National Nature Reserve and five protected areas within the Sanjiangyuan National
tional Nature Reserve. These include Xingxinghai, Zhaling-Eling Lake, Angsai, Suojia-
Nature Reserve. These include Xingxinghai, Zhaling-Eling Lake, Angsai, Suojia-Qumaha
Qumaha River, and Guozongmu Cha. Ultimately, this integration has resulted in the for-
River, and Guozongmu Cha. Ultimately, this integration has resulted in the formation
mation of the Yangtze River Source Park, Yellow River Source Park, and Lancang River
of the Yangtze River Source Park, Yellow River Source Park, and Lancang River Source
Source Park. The SNP boasts a diverse array of natural landscapes, including majestic
Park. The SNP boasts a diverse array of natural landscapes, including majestic glaciers
glaciers and snow-capped mountains, rivers and lakes, wetlands, as well as grasslands
and snow-capped mountains, rivers and lakes, wetlands, as well as grasslands and forests.
and forests. These ecosystems
These ecosystems nurture
nurturea arich
richbiodiversity, comprising760
biodiversity, comprising 760species
species ofof vas- plants and
vascular
cular plants and125125 species
species of wild
of wild animals.
animals. ManyMany of species
of these these species are endemic
are endemic to the
to the Qinghai-Tibet Plateau,
Qinghai-Tibet Plateau, such as the Tibetan antelope, wild yak, Tibetan wild
such as the Tibetan antelope, wild yak, Tibetan wild ass, blue sheep, Tibetan ass, bluegazelle, white-
sheep, Tibetan gazelle, white-lipped
lipped deer, deer, snow
snow leopard, leopard,
and brown bear.and brownas
It serves bear. It serves as
an ecological an
safeguard for the
ecological safeguard for the downstream
downstream areas of China’s areas of China’s
rivers and the rivers and the environmental
environmental security of Southeast Asian
security of Southeast Asian
countries countries [21,22].
[21,22].

Figure 1. LocationFigure Location

of the1.study areaof the study
(labels “Thearea (labels
source “The
of the source River”,
Yangtze of the Yangtze River”,of“The
“The source the source of the
Lancang River”, and “The source of the Yellow River” refer to the entire sections of the park associ-park associated
Lancang River”, and “The source of the Yellow River” refer to the entire sections of the
with these rivers).
ated with these rivers).
2.2. Methods
2.2. Methods 2.2.1. Measuring Spatial Associations
2.2.1. Measuring Spatial
The Associations
global bivariate Moran’s I statistic was used to assess the linear association
The global between
bivariateVEQI and ILPM
Moran’s across
statistic theused
was studytoarea [23].the
assess This statistic
linear be- −1 to 1: −1
ranges from
association
tween VEQI andindicates strongthe
LPM across negative
studyspatial autocorrelation,
area [Error! Reference where similar
source notvalues
found.].are This
widely dispersed;
0 signifies no spatial autocorrelation, indicating a random distribution;
statistic ranges from −1 to 1: −1 indicates strong negative spatial autocorrelation, where and 1 denotes strong
similar values are widely dispersed; 0 signifies no spatial autocorrelation, indicating a To examine
positive spatial autocorrelation, where similar values are closely clustered.
spatialand
random distribution; correlations
1 denoteswithin
strongspecific
positivelocalities, the local bivariate
spatial autocorrelation, whereMoran’s I was utilized.
similar
values are closely clustered. To examine spatial correlations within specific localities, the and high
Clustering diagrams classified spatial units into four types: “H-H” (high VEQI
LPM), “L-L” (low VEQI and low LPM), “H-L” (high VEQI and low LPM), and “L-H” (low
local bivariate Moran’s I was utilized. Clustering diagrams classified spatial units into
VEQI and high LPM). The “H-H” and “L-H” categories were identified as potential sites
four types: “H-H” (high VEQI and high LPM), “L-L” (low VEQI and low LPM), “H-L”
of high ecological integrity, based on the significant local correlation between LPM and
(high VEQI and low LPM), and “L-H” (low VEQI and high LPM). The “H-H” and “L-H”
VEQI. Bivariate Moran’s I was calculated, and local spatial correlations were visualized
using GeoDa 1.14, with significance set at 0.01.
Sustainability 2025, 17, 373 5 of 17

2.2.2. Fractional Vegetation Cover

FVC indicates the proportion of an area covered by vegetation relative to its total extent.
Through the utilization of MODIS satellite data and the hybrid image decomposition
method, FVC estimation becomes accurate by analyzing NDVI values to differentiate
between vegetated and non-vegetated areas [24]. The calculation formula for FVC is
as follows:
FVC = (NDVI − NDVIsoil )/(NDVIveg − NDVIsoil ) (1)

where: NDVI stands for Normalized Difference Vegetation Index, which assesses vegetation
vigor by measuring the difference in reflectance between the infrared and visible bands.
Its values range from −1 to 1, with higher values indicating denser vegetation; NDVIsoil
represents the average NDVI value over bare soil surfaces, reflecting areas devoid of
vegetation; NDVIveg denotes the average NDVI value over surfaces entirely covered by
vegetation, representing the most thriving vegetation state [25].

2.2.3. Vegetation Ecological Quality Index

The VEQI is a comprehensive metric utilized to assess the ecological condition of
vegetation within a defined area. It is formulated through the integration of two pivotal
factors: vegetation cover and net primary productivity of vegetation, employing a weighted
summation approach [24,25]. The specific calculation formula is as follows:

VEQI = [ f 1 × FVC × f 2 × ( NPP/NPPmax )] × 100 (2)

where: f 1 and f 2 represent weighting coefficients that are determined based on the study’s
objectives and specific environmental conditions. They are employed to adjust the relative
contribution of FVC and NPP in the VEQI. For this study, f 1 and f 2 are both set to 0.5;
NPPmax denotes the maximum NPP achievable within an ecosystem. Typically, this value
is utilized to normalize the actual NPP into a percentage format, facilitating comparisons;
The VEQI value ranges from 0 to 100, with higher values indicating superior ecological
quality of vegetation within the area.

2.2.4. Landscape Pattern Metrics

This study utilizes the land use dataset published by Yang et al. in 2023 for its
research. The dataset encompasses nine distinct land use types, including Serac, grassland,
forest, bare ground, cropland, impervious ground, water, wetland, and shrub. Building
on previous research [7–9], to ensure minimal overlap among the selected LPM, this study
comprehensively selects eight indicators from the perspectives of landscape fragmentation,
landscape aggregation, and landscape patch richness. These indicators include the NP,
PD, LPI, TE, ED, LSI, MPA and MSI to thoroughly investigate the evolution of landscape
patterns within the SNP. The calculation formulas for these metrics are presented in Table 1.
These indices were computed solely at the Landscape level using the Landscape metrics
module in Fragstats 4.2.1, providing a holistic perspective of the entire landscape structure.
Class-level metrics were not utilized in this study, as the focus was on overall landscape
patterns rather than specific land use classes. This study employed a grid sampling method
to calculate LPM. Utilizing ArcGIS 10.8, the study area was divided into 1428 one-kilometer
by one-kilometer grids. The land use data were then clipped according to these 1428 grids,
and the LPM within each grid were calculated accordingly.
Sustainability 2025, 17, 373 6 of 17

Table 1. The calculation steps and explanations for Landscape Pattern Metrics used in research.

Index Formula Descriptions

Reflects the proportion of the largest patch
Largest Patch Index (LPI) LPI = Amax/Atotal relative to the entire landscape, with higher
values indicating dominance of large patches
Reflects the density of edges in the landscape,
Edge Density (ED) ED = T/Atotal with higher values indicating a higher degree
of fragmentation
Reflects the average size of patches per hectare
Mean Patch Area (MPA) MPA = Atotal/NP in the landscape, which helps understand the
uniformity of patch distribution
Directly reflects the total number of patches in
Number of patch (NP) NP = total number of patches the landscape, an indicator of
landscape diversity
Reflects the total length of edges of all patches
Total Edge (TE) TE = total length of all patch edges in the landscape, related to
landscape complexity
Reflects the complexity of patch shapes, with
Mean Shape Index (MSI) MSI = 4A/P
higher values indicating irregular shapes
Reflects the density of patches in the landscape,
Patch Density (PD) PD = NP/Atotal with higher values indicating a greater number
of patches and a more dense landscape
Reflects the diversity of patch types in the
Patch Richness (PR) PR = S/NP landscape, with higher values indicating a
richer variety of patch types
Note: Amax: Represents the area of the largest patch in the landscape; Atotal: Represents the total area of
the landscape; T: Represents the total length of all patch edges in the landscape; A: Represents the area of an
individual patch; P: Represents the perimeter of an individual patch; S: Represents the total number of patch
types in the landscape.

2.3. Data Sources

Land-use data for the SNP for the years 2007, 2012, 2017, and 2022, with a spatial
resolution of 30 m per pixel (each pixel representing a 30 m × 30 m area on the ground),
were sourced from the ZENODO database [26]. This resolution provides sufficient spatial
detail to capture the diverse landscape features of the SNP while maintaining computa-
tional efficiency for large-scale analyses. The vector boundary data were derived from the
Resource and Environment Science Data Center [27]. Environmental datasets encompassed
the MOD13A2 Normalized Difference Vegetation Index (NDVI) [28], the MOD17A3 Net
Primary Productivity (NPP) of Vegetation provided by NASA [29], and associated climatic
data [30]. To ensure consistency in data processing and to improve analytical precision, all
datasets were reprojected to the Krasovsky 1940 Albers coordinate system. Furthermore,
they were resampled to a uniform spatial resolution of 1000 m. This approach ensured
consistent row and column alignment across all datasets.

3. Results
3.1. Analysis of Spatiotemporal Variations in VEQI
As illustrated in Figure 2, the overall VEQI of the SNP exhibited an upward trend,
increasing by approximately 38.88% between 2007 and 2022. Specifically, the VEQI of the
Yangtze River Source Park experienced a notable increase of about 48.99% during this
15-year period, indicating a significant enhancement in environmental quality within the
Yangtze River source area. Notably, from 2017 to 2022, the growth rate accelerated to
31.99%. Similarly, the VEQI of the Yellow River Source Park increased by 35.38% from 2007
to 2022, albeit with a slightly lower growth rate compared to the Yangtze River Source
Sustainability 2025, 17, x FOR PEER REVIEW 7 of 18

Sustainability 2025, 17, 373 7 of 17

2007 to 2022, albeit with a slightly lower growth rate compared to the Yangtze River
Source Park. Nevertheless, it still demonstrates a substantial improvement in environ-
Park.
mentalNevertheless,
quality. Over itthestill
lastdemonstrates a substantial
five years, the VEQI of theimprovement
Lancang River in Source
environmental
Park in-
quality. Over the last five years, the VEQI of the Lancang River Source
creased by approximately 22.02% between 2007 and 2022, showing a growth rate slightly Park increased
by approximately
lower than that of 22.02% between
the Yangtze and2007 andRiver
Yellow 2022,Source
showing a growth
Park, yet stillrate slightly an
indicating lower
en-
than that of the Yangtze and Yellow River Source Park, yet still indicating
hancement in environmental quality. A comprehensive comparison reveals that the an enhancement
in environmental
Lancang quality.
River Source ParkAhas comprehensive comparison
consistently maintained reveals high
relatively that the Lancangecolog-
vegetation River
Source Park has consistently maintained relatively high vegetation ecological
ical quality over the years, whereas the VEQI of the Yellow River Source Park and Yangtze quality over
the
Riveryears, whereas
Source Park, the VEQI relatively
although of the Yellow
low,River Source
exhibit Park and
an overall Yangtze
upward River
trend. Source
Moreover,
Park,
all three parks collectively demonstrate a higher growth trend during the periodparks
although relatively low, exhibit an overall upward trend. Moreover, all three from
collectively
2017 to 2022.demonstrate a higher growth trend during the period from 2017 to 2022.

Figure 2. Temporal
Figure 2. Temporaltrends
trendsofofVegetation
VegetationEcological Quality
Ecological Index
Quality (VEQI)
Index variation
(VEQI) between
variation 20072007
between and
2022 in Sanjiangyuan
and 2022 National
in Sanjiangyuan ParkPark
National based on remote
based sensing
on remote data.data.
sensing

As depicted in Figure 3, the VEQI in the Lancang River Source Park and Yellow River
As depicted in Figure 3, the VEQI in the Lancang River Source Park and Yellow River
Source Park is higher than in the Yangtze River Source Park, with the southern regions of
Source Park is higher than in the Yangtze River Source Park, with the southern regions of
the Lancang River Source Park and Yellow River Source Park showing the highest VEQI
the Lancang River Source Park and Yellow River Source Park showing the highest VEQI
levels within the park. The southeastern section of the SNP consistently demonstrates
levels within the park. The southeastern section of the SNP consistently demonstrates
higher VEQI values compared to other areas. Regarding land use types within the SNP, the
higher VEQI values compared to other areas. Regarding land use types within the SNP,
VEQI is highest in grassland, forest, and shrub cover types. Across the years 2007, 2012,
the VEQI is highest in grassland, forest, and shrub cover types. Across the years 2007,
2017, and 2022, the VEQI for forest cover is 70.85, 68.38, 71.64, and 72.39, respectively, while
2012, 2017, and 2022, the VEQI for forest cover is 70.85, 68.38, 71.64, and 72.39, respec-
for shrub cover, it is 78.81, 73.68, 73.79, and 80.81. Grassland cover exhibits a consistent
tively, while for shrub cover, it is 78.81, 73.68, 73.79, and 80.81. Grassland cover exhibits a
upward trend, with VEQI values of 21.24, 22.65, 22.39, and 25.79. The VEQI of forest and
consistent upward trend, with VEQI values of 21.24, 22.65, 22.39, and 25.79. The VEQI of
shrub cover types remains consistently high throughout the study period, reflecting their
forest and shrub cover types remains consistently high throughout the study period, re-
ecological stability. Overall, the VEQI in the SNP has shown a continuous increasing trend
flecting their ecological stability. Overall, the VEQI in the SNP has shown a continuous
from 2007 to 2022. While the ecological quality of the Yangtze River Source Park, Yellow
increasing trend from 2007 to 2022. While the ecological quality of the Yangtze River
River Source Park, and Lancang River Source Park has improved significantly, particularly
Source Park, Yellow River Source Park, and Lancang River Source Park has improved
in the last five years, regional disparities in VEQI levels persist across the parks.
significantly, particularly in the last five years, regional disparities in VEQI levels persist
across
3.2. the parks.
Analysis of Landscape Pattern
3.2.1. Overall Changes in Landscape Pattern Metrics
Different Landscape Pattern Metrics reflect various landscape pattern characteristics.
As depicted in Table 2, from 2007 to 2022, both the patch density and the landscape shape
index within the SNP exhibited a gradual increase. Specifically, the patch density increased
from 2.9173 to 3.4808 (19.38%), indicating a rise in the number of patches per unit area and
a corresponding increase in landscape fragmentation. Conversely, the mean patch area
and edge density displayed a decreasing trend. The mean patch area values declined from
34.2785 km2 to 28.7291 km2 during the study period, representing a decrease of 5.5494 km2 ,
consistent with the landscape fragmentation observed through patch density. Similarly, the
edge density and the landscape shape index exhibited a similar pattern, initially increasing
before gradually decreasing. The edge density rose from 24.0093 to 24.8283 and then
gradually decreased to 21.4671, while the mean shape index increased from 214.1305
to 221.3198 before gradually decreasing to 191.8188. These trends suggest fluctuating
Sustainability 2025, 17, 373 8 of 17

changes in both the edge complexity of the landscape and the shape complexity of patches
throughout the study period. Additionally, the largest patch index decreased gradually
Sustainability 2025, 17, x FOR PEER REVIEW 8 of 18
from 57.9642 to 53.1715, indicating a reduction in the index of the largest connectivity patch
over time.

Figure Spatialdistribution
3. Spatial
Figure 3. distributionofof Vegetation
Vegetation Ecological
Ecological Quality
Quality Index
Index (VEQI)
(VEQI) andand Sanjiangyuan
Sanjiangyuan Na-
National Park
tional Park from
from 2007
2007 to 2022.
to 2022.

Table 2. Landscape Pattern Metric of Sanjiangyuan National Park in 2007–2022.

3.2. Analysis of Landscape Pattern
Number 3.2.1. Overall Changes
of Patches in Landscape
Patch Density PatternLargest
MetricsPatch Index Total Edge
(NP) Different Landscape(PD) Pattern Metrics reflect various
(LPI) landscape pattern characteristics.
(TE)
2022 As depicted in Table 2, from
359,626 2.91732007 to 2022, both the53.1715
patch density and the landscape shape
264,633,560.9
2017 index
411,998 within the SNP exhibited
3.3421 a gradual increase. Specifically,
56.9563 the patch density in-
291,449,679.2
2012 445,796 3.6163(19.38%), indicating 57.0581
creased from 2.9173 to 3.4808 a rise in the number of306,067,646.5
patches per unit
2007 429,089
area and a corresponding 3.4808
increase in landscape57.9642 295,970,655.5
fragmentation. Conversely, the mean
patch area andLandscape
Edge Density edge density displayed
Shape Index a decreasing
Mean Patchtrend. The mean
Area Meanpatch areaIndex
Shape values
declined
(ED) from 34.2785 km 2 to 28.7291 km2 during the study period, representing a decrease
(LSI) (MPA) (MSI)
of 5.5494 km2, consistent with the landscape fragmentation observed through patch den-
2022 21.4671 191.8188 34.2785 1.2457
2017 sity. Similarly, the edge210.9114
23.6425 density and the landscape29.9209
shape index exhibited a 1.2776
similar pattern,
2012 initially increasing before
24.8283 gradually decreasing. 27.6524
221.3198 The edge density rose from1.280624.0093 to
2007 24.0093
24.8283 and then gradually214.1305decreased to 21.4671,28.7291 1.2794increased
while the mean shape index
from 214.1305 to 221.3198 before gradually decreasing to 191.8188. These trends suggest
fluctuating
3.2.2. changes
Temporal in both Dynamics
and Spatial the edge complexity of the
of Landscape landscape
Pattern and the shape complex-
Metrics
ity of patches throughout the study period. Additionally, the largest patch index de-
At the spatial scale, areas with high Largest Patch Index (LPI) values are predominantly
creased gradually from 57.9642 to 53.1715, indicating a reduction in the index of the largest
concentrated in the Lancang River Source Park, the Yellow River Source Park, and the
connectivity patch over time.
Sustainability 2025, 17, 373 9 of 17

southern section of the Yangtze River Source Park (Yangtze River Source Park) (Figure 4).
These high LPI values reflect robust ecosystem connectivity, which supports biodiversity
preservation and enhances the stability of ecological functions in these regions. Conversely,
regions with elevated Edge Density (ED) values are primarily located in the northern sector
of the Yangtze River Source Park, suggesting a higher degree of boundary interactions
between landscape elements. This area also exhibits high values of Number of Patches (NP)
and Total Edge (TE), indicating a proliferation of boundary lines and increased landscape
complexity and diversity. The distribution of high Patch Density (PD) values further
highlights areas with a denser arrangement of patches, particularly in the northern Yangtze
River Source Park. This pattern points to significant landscape fragmentation, potentially
driven by human activities such as grazing or infrastructure development, or natural
influences like climate variability. Similarly, high Patch Richness (PR) values are observed
in the southern region of the Lancang River Source Park and the northern part of the Yangtze
River Source Park, signaling a rich diversity of ecological landscape types. These areas
demonstrate a strong potential for biodiversity conservation and ecosystem stability due to
the variety of habitats available. The southern segment of the Lancang River Source Park
emerges as a critical zone for ecological conservation within the SNP. With its high PR values
and ongoing ecological restoration efforts, this region plays a pivotal role in maintaining
ecological integrity. Restoration initiatives in this area have significantly contributed
to enhancing habitat connectivity, mitigating landscape fragmentation, and preserving
Sustainability 2025, 17, x FOR PEER REVIEW 10 of 18
ecological functions, underscoring its importance as a focal point for conservation activities
in the SNP.

Figure 4. Spatial distribution of Landscape Pattern Metrics of Sanjiangyuan National Park from 2007
Figure 4. Spatial distribution of Landscape Pattern Metrics of Sanjiangyuan National Park from 2007
to 2022, categorized using the Natural Breaks (Jenks) method.
to 2022, categorized using the Natural Breaks (Jenks) method.
We investigated the evolutionary trends of the Landscape Pattern Metrics (LPM)
within theinvestigated
We three parks the evolutionary
comprising trends of the
the Sanjiangyuan Landscape
National Pattern
Park (SNP) Metrics
from (LPM)
2007 to 2022
within the three parks comprising the Sanjiangyuan National Park (SNP)
(Figure 5) and analyzed the relationship between LPM changes and Vegetation Ecological from 2007 to
2022 (Figure
Quality Index5) (VEQI).
and analyzed the relationship
The results between
indicate spatial LPM changes
variability and Vegetation
in different Eco-
metrics across
logical Quality Index (VEQI). The results indicate spatial variability in different
the parks. The Largest Patch Index (LPI) exhibited a slight decline across all three parks. metrics
across
In the the parks.
Yellow TheSource
River LargestPark
Patch(YRSP),
Index (LPI) exhibited afrom
LPI decreased slight decline
93.14 across
in 2007 to all three
91.94 in
parks. In the Yellow River Source Park (YRSP), LPI decreased from
2022 (−1.29%). In the Yangtze River Source Park (YZRSP) and Lancang River Source93.14 in 2007 to 91.94
in 2022
Park (−1.29%).
(LRSP), In the Yangtze
LPI dropped RivertoSource
from 78.45 77.28 (−Park (YZRSP)
1.49%) and95.06
and from Lancang River
to 93.99 (−Source
1.13%),
Park (LRSP), LPI dropped from 78.45 to 77.28 (−1.49%) and from 95.06 to
respectively. However, specific regions within the parks, such as the southern YRSP and 93.99 (−1.13%),
respectively. However, specific regions within the parks, such as the southern YRSP and
eastern LRSP, displayed an increasing trend in LPI, corresponding to areas with improv-
ing VEQI. The Edge Density (ED) showed an increase in the YRSP from 6.895 in 2007 to
7.289 in 2022 (+5.71%), while it decreased in the YZRSP and LRSP, dropping from 29.594
to 26.650 (−9.96%) and from 13.121 to 9.812 (−25.21%), respectively. Areas such as the east-
Sustainability 2025, 17, 373 10 of 17

eastern LRSP, displayed an increasing trend in LPI, corresponding to areas with improving
VEQI. The Edge Density (ED) showed an increase in the YRSP from 6.895 in 2007 to 7.289
in 2022 (+5.71%), while it decreased in the YZRSP and LRSP, dropping from 29.594 to
26.650 (−9.96%) and from 13.121 to 9.812 (−25.21%), respectively. Areas such as the eastern
LRSP and southern YRSP exhibited the most pronounced decreases in ED, coinciding
with increasing VEQI values. The Patch Density (PD) exhibited different trends across the
parks. The YRSP and LRSP showed increases in PD, rising from 1.227 to 1.311 (+6.84%) and
from 2.898 to 3.942 (+35.99%), respectively. In contrast, the YZRSP saw a decrease in PD
from 4.592 to 4.048 (−11.84%). The northern YZRSP and eastern LRSP displayed the most
significant changes in PD, with decreases in these regions corresponding to higher VEQI.
The Number of Patches (NP) increased in the YRSP from 79.049 to 89.367 (+13.07%) but
decreased in the LRSP and YZRSP, from 190.067 to 121.605 (−35.98%) and from 412.117
to 357.715 (−13.20%), respectively. The southern YRSP and eastern LRSP had a marked
reduction in NP, while the northern YZRSP exhibited the most stable patterns. The Total
Edge (TE) increased slightly in the YRSP, from 48,590.068 to 53,395.952 (+9.89%), while
decreasing in the YZRSP, from 279,813.785 to 252,970.927 (−9.59%), and in the LRSP, from
92,145.980 to 70,184.437 (−23.82%). The southeastern YRSP and eastern LRSP experienced
the most significant decreases in TE. The Mean Shape Index (MSI) decreased slightly in
all three parks. In the YRSP, MSI declined from 1.23 to 1.22 (−0.82%); in the YZRSP, it
dropped from 1.29 to 1.26 (−2.33%); and in the LRSP, it fell from 1.25 to 1.24 (−0.80%). The
southeastern LRSP and southern YRSP displayed the most stable MSI values. The Patch
Richness (PR) increased slightly in the YRSP, rising from 2.78 to 2.84 (+2.16%). In contrast,
PR decreased in the YZRSP and LRSP, dropping from 3.51 to 3.17 (−10.25%) and from 3.91
to 3.52 (−9.97%), respectively. The southeastern YRSP and eastern LRSP had consistently
higher PR values compared to other regions. The Mean Patch Area (MPA) declined across
all parks. In the YRSP, MPA decreased from 1291.83 to 1021.49 (−20.97%), while in the
YZRSP and LRSP, it dropped from 401.85 to 353.22 (−12.10%) and from 786.51 to 687.19
(−12.60%), respectively. The southern YRSP and eastern LRSP exhibited slightly increasing
MPA values compared to other areas.

3.3. Analysis of the Correlation Between Vegetation Ecological Quality Index and Landscape
Pattern Metric
3.3.1. Temporal Correlation Between Vegetation Ecological Quality Index and Landscape
Pattern Metric
As depicted in Table 3, there exists a strong spatial coupling between VEQI and four
LPM: LPI, ED, TE, and NP within the study area. Conversely, the spatial coupling with
PD, PR, MPA, and MSI is relatively weak. The relationship between VEQI and LPI, as well
as MPA, exhibits a significant positive spatial correlation, which was further accentuated
over time. Throughout the study period, the bivariate global Moran’s I absolute value
for VEQI and LPI was the highest, indicating a correlation between VEQI and landscape
connectivity and cohesion, influencing the spatial distribution of environmental quality
based on landscape characteristics. VEQI displayed a negative correlation with PD, ED, PR,
TE, NP, and MSI. Areas with higher vegetation ecological quality (high VEQI) generally
exhibited fewer patches (low NP), lower patch density (low PD), fewer edges (low ED),
lower total edge length (low TE), and more regular patch shapes (low MSI). This suggests
that regions with better environmental quality typically feature more continuous and
regular landscapes with larger, fewer patches, and fewer edges.
 higher PR values compared to other regions. The Mean Patch Area (MPA) declined across
all parks. In the YRSP, MPA decreased from 1291.83 to 1021.49 (−20.97%), while in the
YZRSP and LRSP, it dropped from 401.85 to 353.22 (−12.10%) and from 786.51 to 687.19
Sustainability 2025, 17, 373 (−12.60%), respectively. The southern YRSP and eastern LRSP exhibited slightly increas-
11 of 17
ing MPA values compared to other areas.

Figure
Figure 5.5. Correlation
CorrelationBetween
BetweenChanges
Changesinin
Vegetation Ecological
Vegetation Quality
Ecological Index
Quality (VEQI)
Index and Landscape
(VEQI) and Land-
Pattern Metrics in Sanjiangyuan National Park (2007–2022). Note: The y-axis scale does
scape Pattern Metrics in Sanjiangyuan National Park (2007–2022). Note: The y-axis scale doesnot start at
not0
to better highlight trends and variations.
start at 0 to better highlight trends and variations.

Table 3. Global Moran’s I values of Vegetation Ecological Quality Index (VEQI) degree and Landscape
3.3. Analysis of the Correlation Between Vegetation Ecological Quality Index and Landscape
Pattern Metric in 2007–2022.
Pattern Metric
Years LPI PD TemporalED
3.3.1. PR
Correlation Between TE
Vegetation NP Quality
Ecological Shape_MN
Index andArea_MN
2007 0.490 Landscape
− 0.093 Pattern
−0.512Metric −0.069 −0.516 −0.410 −0.232 0.297
2012 0.493 −0.155 − − − − −
As depicted in Table 3, there exists a strong spatial coupling between VEQI0.268
0.475 0.141 0.523 0.448 0.236 and four
2017 0.467 − 0.157 − 0.465 − 0.151 − 0.511 − 0.474 − 0.254 0.321
LPM: LPI, ED, TE, and NP within the study area. Conversely, the spatial coupling with
2022 0.480 −0.083 −0.523 −0.069 −0.513 −0.514 −0.235 0.302
PD, PR, MPA, and MSI is relatively weak. The relationship between VEQI and LPI, as well

3.3.2. The Spatial Correlation Between Vegetation Ecological Quality Index and Landscape
Pattern Metric
From a spatial perspective (Figure 6), the clustering pattern between VEQI and LPM in
SNP from 2007 to 2022 remains relatively stable, showing significant spatial heterogeneity.
The clustering types of VEQI and six LPM—ED, PD, PR, TE, NP, and MSI—are mainly
evident in spatial aggregation types. Specifically, the aggregations of VEQI with ED, PR,
TE, NP, and MSI predominantly exhibit “H-L” and “L-H” types. In the northwestern region
 From a spatial perspective (Figure 6), the clustering pattern between VEQI and LPM
in SNP from 2007 to 2022 remains relatively stable, showing significant spatial heteroge-
Sustainability 2025, 17, 373
neity. The clustering types of VEQI and six LPM—ED, PD, PR, TE, NP, and MSI—are
12 of 17
mainly evident in spatial aggregation types. Specifically, the aggregations of VEQI with
ED, PR, TE, NP, and MSI predominantly exhibit “H-L” and “L-H” types. In the north-
western
of region
Yangtze RiverofSource
Yangtze River
Park, Source
“L-H” Park, “L-H”
aggregation aggregation
prevails, prevails,
indicating indicating
relatively poor VEQIrel-
atively
but poor VEQI
complex butlandscape
and rich complex and rich landscape
patterns. patterns.
Conversely, Conversely,
the spatial the spatial
aggregation ag-
of VEQI
gregation of VEQI with MPA and LPI mainly consists of “H-H” and “L-L”
with MPA and LPI mainly consists of “H-H” and “L-L” types. The “H-H” aggregation is types. The “H-
H” aggregation
primarily is primarily
concentrated in theconcentrated
Lancang River in the Lancang
Source River
Park and SourceRiver
Yellow ParkSource
and Yellow
Park,
River Source Park, signifying areas with good environmental quality and
signifying areas with good environmental quality and high landscape uniformity and high landscape
uniformity and
connectivity. connectivity.
Meanwhile, theMeanwhile, the “L-L”
“L-L” aggregation aggregation
is mainly is mainly
observed in theobserved in the
northwestern
northwestern part of Yangtze River Source Park, characterized by poor environmental
part of Yangtze River Source Park, characterized by poor environmental quality and low
quality anduniformity
landscape low landscape uniformity and connectivity.
and connectivity.

Figure 6. Local Bivariate LISA agglomeration of Vegetation Ecological Quality Index (VEQI) and
Figure 6. Local Bivariate LISA agglomeration of Vegetation Ecological Quality Index (VEQI) and
Landscape Pattern Metric in 2007–2022.
Landscape Pattern Metric in 2007–2022.
4. Discussion and Limitation
4. Discussion and Limitation
4.1. Discussion
In this study, we analyzed the spatial evolution of VEQI and LPM in SNP from
2007 to 2022 using multi-source remote sensing data and LPM. Overall, the VEQI of SNP
exhibited an upward trend, increasing by approximately 38.88%. This finding aligns
with previous research in other ecological conservation regions, such as the Loess Plateau
and the Three-River Headwaters, which reported significant improvements in vegetation
quality following large-scale ecological restoration projects [31,32]. However, our study
uniquely highlights the spatial disparities in VEQI growth among different sub-regions
of the SNP and their correlations with landscape fragmentation metrics. Notably, the
VEQI of the Yangtze River Source Park demonstrated the most significant growth, rising
by about 48.99%, particularly between 2017 and 2022, with a remarkable growth rate of
31.99%. Similarly, the VEQI of the Yellow River Source Park experienced a growth of
35.38%, albeit slightly lower than that of the Yangtze River Source Park, it still manifested a
significant trend of environmental quality improvement, particularly in the last five years.
Conversely, the VEQI of the Lancang River Source Park increased by 22.02%, representing
a relatively lower growth rate; however, it still indicated a positive trend in environmental
quality enhancement. Through a comprehensive comparison, it was observed that the
Lancang River Source Park maintained relatively high vegetation ecological quality over
the study period, while the Yellow River Source Park and Yangtze River Source Park
exhibited relatively low VEQI. Nonetheless, both parks displayed an overall growth trend,
particularly notable between 2017 and 2022. Since the initiation of the SNP construction
project in 2016, China has directed substantial ecological resources towards the region,
leading to significant advancements in local ecological conservation [33,34]. This result
aligns with findings in other restoration-focused regions, such as the Loess Plateau and the
Three-River Headwaters, where large-scale ecological engineering projects contributed to
vegetation quality improvements over similar timescales [35,36]. However, our study adds
unique insights into the spatial variability of VEQI within a single integrated protected area,
Sustainability 2025, 17, 373 13 of 17

which remains underexplored in previous research. The park encompasses five protected
sub-areas, namely Zaling-Erling Lake, Xingxinghai, Soga-Qumahe, Gezongmucha, and
Angsai, alongside the Coco-Cecily National Nature Reserve. Through collaborative efforts
between national parks and nature reserves, the ecological quality of vegetation in the
Sanjiangyuan region has notably improved in recent years [37]. Furthermore, there has been
a discernible upward trajectory in the region’s improvement, attributable to the combined
efforts of national park and nature reserve construction [38–40]. Our findings indicate that
metrics such as Largest Patch Index (LPI) and Mean Patch Area (MPA) showed positive
correlations with VEQI, emphasizing the importance of large, connected landscape patches
in enhancing ecosystem stability. Conversely, negative correlations were observed between
VEQI and metrics such as Patch Density (PD) and Edge Density (ED), which are consistent
with prior studies indicating that fragmentation adversely affects ecological complexity
and biodiversity [41,42]. These correlations highlight the dual importance of reducing
fragmentation and enhancing connectivity to improve vegetation quality.
From the perspective of spatial distribution characteristics, the Vegetation Ecological
Quality Index (VEQI) exhibits distinct spatial patterns, with lower values predominantly
concentrated in the northwest direction of the study area, while higher values are pri-
marily found in the southeast direction. Particularly, the southern regions of the Lancang
River Source Park and the Yellow River Source Park stand out as areas with exceptionally
high vegetation ecological quality within the entire park. This can be attributed to robust
ecological protection measures, such as the establishment of the Dzongmucha Protection
Sub-area and the Angsay Protection Sub-area, along with the implementation of the eco-
logical red line policy [43,44]. These protection zones have effectively safeguarded local
ecosystems and enhanced vegetation ecological quality, consequently elevating the VEQI
in these areas. Similar practices in other biodiversity hotspots, such as the Amazon Basin,
have also demonstrated the effectiveness of strengthening protected area management
to maintain ecological quality [45]. In the future construction and management of the
SNP, it is imperative to adopt customized ecological protection measures tailored to the
vegetation ecological quality and landscape fragmentation characteristics of different park
areas. For the northwest region with lower VEQI values, specific vegetation restoration
policies need to be designated. This involves implementing ecological engineering and
adaptive management to enhance the resilience and stability of the ecosystem. The es-
tablishment of ecological corridors in this region, a strategy that has been successfully
implemented in areas such as Yellowstone National Park, could significantly improve
connectivity among fragmented patches, thereby supporting biodiversity and ecosystem
functionality [46]. Concurrently, although the southeast region has higher VEQI values, it
also requires the formulation of corresponding protection policies to preserve its ecological
advantages. This includes strengthening protected area management, optimizing land use
planning, and restricting human activities that could potentially degrade ecological quality.
By integrating these strategies with insights from this study, future conservation efforts
can more effectively balance ecological restoration and protection across diverse regions of
the SNP.
Between 2007 and 2022, there were significant changes in the Landscape Pattern
Metric and vegetation ecological quality indicators within SNP. The multi-year mean
values of Moran’s I for LPI, Mean Patch Area, and VEQI were 0.48 and 0.29, respectively,
indicating a positive correlation. Conversely, PD, ED, PR, TE, NP, and MPI showed negative
correlations with VEQI, with Moran’s I multi-year means ranging from −0.13 to −0.49.
This suggests that as the number of connected patches increases, so does the MPA, likely
due to advancements in ecological restoration projects promoting ecosystem connectivity
and stability [47,48]. Conversely, the negative correlations indicate an increase in landscape
Sustainability 2025, 17, 373 14 of 17

fragmentation, alongside a decrease in ecological complexity and diversity [49–51]. The

findings of this study reveal that the southeast region of the study area has lower VEQI
and a lower degree of landscape fragmentation. The primary land use type in these areas is
grassland, which exhibits good landscape connectivity. Therefore, future development in
this area should be minimized to ensure the maintenance of landscape connectivity and
the ecological quality of the vegetation. In addition, the northwest region of the study
area exhibits a higher level of landscape fragmentation and lower VEQI. Previous research
indicates that the construction of ecological corridors can effectively mitigate the degree
of landscape fragmentation in a region. Consequently, it is recommended that ecological
corridors be established in this area in the future to connect the more fragmented landscape
patches, thereby ensuring the region’s landscape connectivity and level of biodiversity.

4.2. Limitation
Based on land use data, this study analyzes the spatial and temporal characteristics of
LPM and VEQI from 2007 to 2022 at three scales: the national park as a whole, subdivided
park, and grid. This analysis enriches theoretical insights into the relationship between
landscape pattern change and vegetation ecological quality, providing a theoretical founda-
tion for optimizing landscape patterns and protecting the ecological environment in SNP.
However, there are notable limitations. Firstly, the spatial correlation analysis focused solely
on Moran’s I, characterizing the spatial correlation between vegetation ecology and LPM.
This approach didn’t capture the direct influence of spatial-temporal changes in landscape
patterns on vegetation ecological quality. Future studies should integrate other research
methods to visually illustrate the direct impact and its magnitude. Secondly, although
this study selected landscape pattern metrics based on their theoretical relevance and
applicability, no correlation analysis was conducted to eliminate redundant metrics. While
the current approach did not involve weighted overlay analysis, where such correlations
might directly influence results, performing a correlation analysis could help refine the
selection of representative metrics in future research. This would enhance the objectivity
and robustness of the metric selection process. By addressing these limitations, future
studies can further improve the comprehensiveness and depth of the research.

5. Conclusions
As China’s largest national park, the Sanjiangyuan National Park (SNP) serves a
crucial role in ecological conservation and biodiversity protection within its regional scope.
This study employs the Vegetation Ecological Quality Index (VEQI) model and Landscape
Pattern Metrics (LPM) to comprehensively analyze the temporal and spatial dynamics of
vegetation quality and landscape fragmentation in the SNP from 2007 to 2022. Furthermore,
this research elucidates the spatial correlation between VEQI and landscape fragmentation,
offering a robust scientific basis for guiding ecological conservation strategies in the SNP.
Our findings reveal a significant overall improvement in vegetation ecological quality
within the SNP, with the VEQI increasing by approximately 38.88% over the study period.
Among the three major sub-parks, the Yangtze River Source Park exhibited the most pro-
nounced growth in VEQI at 48.99%, particularly between 2017 and 2022, reflecting the
substantial impact of national ecological restoration initiatives since the park’s establish-
ment in 2016. The Yellow River Source Park and Lancang River Source Park also displayed
considerable VEQI improvements, growing by 35.38% and 22.02%, respectively. These
findings underscore the success of ecological restoration measures but also highlight the
spatial disparities in VEQI growth, driven by differences in landscape fragmentation and
ecological restoration efforts. The spatial distribution of VEQI reveals that higher values are
predominantly concentrated in the southeast regions of the park, where robust ecological
Sustainability 2025, 17, 373 15 of 17

protection measures, such as the Dzongmucha and Angsay Protection Sub-areas, have been
implemented. Conversely, the northwest regions exhibited lower VEQI values and a higher
degree of landscape fragmentation. This emphasizes the need for targeted conservation
strategies that address the unique ecological challenges of different regions within the SNP.
Landscape fragmentation analysis showed a gradual increase in metrics such as Patch
Density (PD) and Landscape Shape Index (LSI), alongside decreases in Mean Patch Area
(MPA) and Edge Density (ED), indicating rising fragmentation across the SNP. Regions with
high Largest Patch Index (LPI), particularly in the southern parts of the Yangtze, Yellow,
and Lancang River Source Parks, demonstrated better ecosystem connectivity, whereas the
northern regions, with higher Edge Density (ED), exhibited greater landscape complexity
and fragmentation. These trends reinforce the importance of maintaining connectivity in
high-quality vegetation areas while addressing fragmentation in more vulnerable regions.
Based on these findings, we propose targeted conservation strategies for different
areas of the SNP: Northwest Regions: With lower VEQI values and higher landscape
fragmentation, these areas require active vegetation restoration policies, including the
establishment of ecological corridors. Drawing from successful implementations in regions
like Yellowstone National Park, these corridors can enhance connectivity and ecosystem
stability. Southeast Regions: Despite their higher VEQI values and lower fragmentation,
these areas require continued protection through measures such as land use planning, the
restriction of human activities, and the strengthening of protected area management to
preserve their ecological advantages. This study uniquely contributes to understanding
the spatial variability of VEQI and its correlations with landscape metrics within a single
integrated protected area. It highlights the need for tailored, region-specific ecological
conservation strategies that balance restoration and protection efforts across diverse sub-
regions of the SNP. Future conservation policies should leverage these insights to enhance
the resilience and sustainability of the park’s ecosystems, ensuring long-term ecological
integrity and biodiversity conservation.

Author Contributions: Conceptualization, X.P., R.T. and H.T.; Methodology, X.P., R.T., Z.G. and J.L.;
Software, X.P., R.T., Z.G. and H.T.; Validation, X.P., R.T., J.L. and H.T.; Formal analysis, X.P., R.T., Z.G.
and H.T.; Investigation, X.P., R.T., J.L. and H.T.; Resources, X.P., R.T. and J.L.; Data curation, X.P. and
R.T.; Writing—original draft, X.P. and R.T.; Writing—review & editing, X.P. and Z.G.; Visualization,
X.P.; Supervision, X.P. and R.T.; Project administration, X.P.; Funding acquisition, X.P. All authors
have read and agreed to the published version of the manuscript.

Funding: This project was funded by the Jiangsu Provincial Postgraduate Research and Practice
Project: SJCX24_0388.

Institutional Review Board Statement: Not applicable.

Informed Consent Statement: Not applicable.

Data Availability Statement: The original contributions presented in this study are included in the
article. Further inquiries can be directed to the corresponding author.

Conflicts of Interest: The authors declare no conflict of interest.

References
1. Everhart, W. The National Park Service; Routledge: London, UK, 2019.
2. Runte, A. National Parks: The American Experience; U of Nebraska Press: Lincoln, NE, USA, 1997.
3. Wang, G.; Innes, J.L.; Wu, S.W.; Krzyzanowski, J.; Yin, Y.; Dai, S.; Zhang, X.; Liu, S. National park development in China:
Conservation or commercialization? Ambio 2012, 41, 247–261. [CrossRef] [PubMed]
4. D’Arco, M.; Presti, L.L.; Marino, V.; Maggiore, G. Is sustainable tourism a goal that came true? The Italian experience of the
Cilento and Vallo di Diano National Park. Land Use Policy 2021, 101, 105198. [CrossRef]
Sustainability 2025, 17, 373 16 of 17

5. Ma, T.; Swallow, B.; Zhong, L.; Xu, K.; Sang, W.; Jia, L. Local perspectives on social-ecological transformation: China’s San-
jiangyuan National Park. Environ. Dev. Sustain. 2024, 26, 1809–1829. [CrossRef]
6. Ma, T.; Swallow, B.; Foggin, J.M.; Sang, W.; Zhong, L. Developing co-management for conservation and local development in
China’s national parks: Findings from focus group discussions in the Sanjiangyuan Region. Front. Conserv. Sci. 2023, 4, 903788.
[CrossRef]
7. O’Neill, R.V.; Krummel, J.R.; Gardner, R.E.A.; Sugihara, G.; Jackson, B.; DeAngelis, D.L.; Milne, B.T.; Turner, M.G.; Zygmunt, B.;
Christensen, S.W.; et al. Indices of landscape pattern. Landsc. Ecol. 1988, 1, 153–162. [CrossRef]
8. Wu, J. Effects of changing scale on landscape pattern analysis: Scaling relations. Landsc. Ecol. 2004, 19, 125–138. [CrossRef]
9. McGarigal, K. Landscape pattern metrics. In Encyclopedia of Environmetrics; John Wiley & Sons, Ltd.: Hoboken, NJ, USA, 2006.
10. Zhu, C.; Zhang, X.; Zhou, M.; He, S.; Gan, M.; Yang, L.; Wang, K. Impacts of urbanization and landscape pattern on habitat
quality using OLS and GWR models in Hangzhou, China. Ecol. Indic. 2020, 117, 106654. [CrossRef]
11. Arroyo-Rodríguez, V.; Martínez-Ruiz, M.; Bezerra, J.S.; Galán-Acedo, C.; San-José, M.; Fahrig, L. Does a Species’ Mobility
Determine the Scale at Which It Is Influenced by the Surrounding Landscape Pattern? Curr. Landsc. Ecol. Rep. 2023, 8, 23–33.
[CrossRef]
12. Hou, L.; Wu, F.; Xie, X. The spatial characteristics and relationships between landscape pattern and ecosystem service value along
an urban-rural gradient in Xi’an city, China. Ecol. Indic. 2020, 108, 105720. [CrossRef]
13. Xiao, R.; Cao, W.; Liu, Y.; Lu, B. The impacts of landscape patterns spatio-temporal changes on land surface temperature from a
multi-scale perspective: A case study of the Yangtze River Delta. Sci. Total Environ. 2022, 821, 153381. [CrossRef] [PubMed]
14. Hu, C.; Wu, W.; Zhou, X.; Wang, Z. Spatiotemporal changes in landscape patterns in karst mountainous regions based on the
optimal landscape scale: A case study of Guiyang City in Guizhou Province, China. Ecol. Indic. 2023, 150, 110211. [CrossRef]
15. Cramer, W.; Kicklighter, D.W.; Bondeau, A.; Iii, B.M.; Churkina, G.; Nemry, B. Comparing global models of terrestrial net primary
productivity (NPP): Overview and key results. Glob. Change Biol. 1999, 5 (Suppl. S1), 1–15. [CrossRef]
16. Yang, L.; Jia, K.; Liang, S.; Liu, M.; Wei, X.; Yao, Y. Spatio-temporal analysis and uncertainty of fractional vegetation cover change
over northern China during 2001–2012 based on multiple vegetation data sets. Remote Sens. 2018, 10, 549. [CrossRef]
17. Yang, H.; Zhong, X.; Deng, S.; Xu, H. Assessment of the impact of LUCC on NPP and its influencing factors in the Yangtze River
basin, China. Catena 2021, 206, 105542. [CrossRef]
18. Zarei, A.; Chemura, A.; Gleixner, S.; Hoff, H. Evaluating the grassland NPP dynamics in response to climate change in Tanzania.
Ecol. Indic. 2021, 125, 107600. [CrossRef]
19. Cui, L.; Chen, Y.; Yuan, Y.; Luo, Y.; Huang, S.; Li, G. Comprehensive evaluation system for vegetation ecological quality: A case
study of Sichuan ecological protection redline areas. Front. Plant Sci. 2023, 14, 1178485. [CrossRef]
20. Wang, Z.; Bai, T.; Xu, D.; Kang, J.; Shi, J.; Fang, H.; Nie, C.; Zhang, Z.; Yan, P.; Wang, D. Temporal and Spatial Changes in
Vegetation Ecological Quality and Driving Mechanism in Kökyar Project Area from 2000 to 2021. Sustainability 2022, 14, 7668.
[CrossRef]
21. Zhao, X. The five integrative management strategies of Sanjiangyuan National Park. Biodivers. Sci. 2021, 29, 301. [CrossRef]
22. Cao, W.; Wu, D.; Huang, L.; Liu, L. Spatial and temporal variations and significance identification of ecosystem services in the
Sanjiangyuan National Park, China. Sci. Rep. 2020, 10, 6151. [CrossRef] [PubMed]
23. Han, M.; Sun, R.; Feng, P.; Hua, E. Unveiling characteristics and determinants of China’s wind power geographies towards
low-carbon transition. J. Environ. Manag. 2023, 331, 117215. [CrossRef] [PubMed]
24. Xu, D.; Yang, F.; Yu, L.; Zhou, Y.; Li, H.; Ma, J.; Huang, J.; Wei, J.; Xu, Y.; Zhang, C.; et al. Quantization of the coupling mechanism
between eco-environmental quality and urbanization from multisource remote sensing data. J. Clean. Prod. 2021, 321, 128948.
[CrossRef]
25. Bai, T.; Cheng, J.; Zheng, Z.; Zhang, Q.; Li, Z.; Xu, D. Drivers of eco-environmental quality in China from 2000 to 2017. J. Clean.
Prod. 2023, 396, 136408. [CrossRef]
26. Yang, J.; Huang, X. The 30 m annual land cover datasets and its dynamics in China from 1985 to 2022 [Data set]. Earth Syst. Sci.
Data 2023, 13, 3907–3925. [CrossRef]
27. Resource and Environment Science Data Center. Available online: https://www.resdc.cn/ (accessed on 15 January 2024).
28. Didan, K. MOD13A2 MODIS/terra Vegetation Indices 16-Day L3 Global 1 Km SIN Grid V006 [Data Set]. NASA EOSDIS Land
Processes DAAC; 2015. Available online: https://lpdaac.usgs.gov/products/mod13a2v006/ (accessed on 15 January 2024).
29. Running, S.; Mu, Q.; Zhao, M. MOD17A3H MODIS/Terra Net Primary Production Yearly L4 Global 500 m SIN Grid V006 [Data
Set]. NASA EOSDIS Land Processes DAAC, 2022-09-12; 2015. Available online: https://lpdaac.usgs.gov/products/mod17a3hv006/
(accessed on 15 January 2024).
30. Abatzoglou, J.T.; Dobrowski, S.Z.; Parks, S.A.; Hegewisch, K.C. TerraClimate, a highresolution global dataset of monthly climate
and climatic water balance from 1958–2015. Sci. Data 2018, 5, 1–12. [CrossRef]
31. Mou, X.; Chai, H.; Duan, C.; Feng, Y.; Wang, X. An Assessment of Vegetation Changes in the Three-River Headwaters Region,
China: Integrating NDVI and Its Spatial Heterogeneity. Plants 2024, 13, 2814. [CrossRef]
Sustainability 2025, 17, 373 17 of 17

32. Zhang, X.; Ning, J. Patterns, Trends, and Causes of Vegetation Change in the Three Rivers Headwaters Region. Land 2023, 12,
1127. [CrossRef]
33. Xia, C.Z.; Zhou, L.H.; Pei, X.D.; Wang, Y.; Li, J.H. Performance research of Pastureland Rehabilitation Project in Sanjiangyuan
National Park from the perspectives of local governments and herdsmen. J. Nat. Resour. 2023, 38, 1570–1587. [CrossRef]
34. Ma, T.; Swallow, B.; Foggin, J.M.; Zhong, L.; Sang, W. Developing Co-management for Conservation and Local Development in
China’s National Parks: Findings from Focus Groups in the Sanjiangyuan Region. preprint, 2021. [CrossRef]
35. Li, J.; Peng, X.; Tang, R.; Geng, J.; Zhang, Z.; Xu, D.; Bai, T. Spatial and temporal variation characteristics of ecological environment
quality in China from 2002 to 2019 and influencing factors. Land 2024, 13, 110. [CrossRef]
36. Liu, T.; Peng, X.; Li, J. Evaluation of Ecological Sensitivity and Spatial Correlation Analysis of Landscape Patterns in Sanjiangyuan
National Park. Sustainability 2024, 16, 5294. [CrossRef]
37. Yin, B.F.; Huai, H.Y.; Zhang, Y.L.; Le, Z.; Wei, W.H. Trophic niches of Pantholops hodgsoni, Procapra picticaudata and Equus
kiang in Kekexili region. Ying Yong Sheng Tai Xue Bao = J. Appl. Ecol. 2007, 18, 766–770.
38. Zhang, Y.; Hacker, C.; Zhang, Y.; Xue, Y.; Wu, L.; Dai, Y.; Luo, P. An analysis of genetic structure of snow leopard populations in
Sanjiangyuan and Qilianshan National Parks. Acta Theriol. Sin. 2019, 39, 442.
39. Wang, X.; Zhang, X. Study on the Coordinated Development of Economy, Tourism, and Eco-Environment in Sanjiangyuan. Math.
Probl. Eng. 2022, 2022, 9463166. [CrossRef]
40. Huangqingdongzhi; Chen, L.; Li, S.; Sun, Z.; Yang, M.; Zhang, D.; Li, C.; Mao, R.; Li, Z.; Zhang, L. Effects of roads on the
abundance and habitat of Tibetan wild ass and Tibetan gazelle in the Yellow River Source Zone of Sanjiangyuan National Park.
Acta Theriol. Sin. 2022, 42, 34.
41. Liu, Y.; Xu, W.; Hong, Z.; Wang, L.; Ou, G.; Lu, N.; Dai, Q. Integrating three-dimensional greenness into RSEI improved the
scientificity of ecological environment quality assessment for forest. Ecol. Indic. 2023, 156, 111092. [CrossRef]
42. Shaker, R.R. The well-being of nations: An empirical assessment of sustainable urbanization for Europe. Int. J. Sustain. Dev. World
Ecol. 2015, 22, 375–387. [CrossRef]
43. Liu, G.; Shao, Q.; Fan, J.; Huang, H.; Liu, J.; He, J. Assessment of restoration degree and restoration potential of key ecosystem-
regulating services in the three-river headwaters region based on vegetation coverage. Remote Sens. 2023, 15, 523. [CrossRef]
44. Yu, Z.; Chen, X.; Wu, J. Calibrating a Hydrological Model in an Ungauged Mountain Basin with the Budyko Framework. Water
2022, 14, 3112. [CrossRef]
45. Shaker, R.R. Examining sustainable landscape function across the Republic of Moldova. Habitat Int. 2018, 72, 77–91. [CrossRef]
46. Mengqiao, W.; Zhongjun, W. Ecological integrity conservation and management of group of national parks of Yellowstone to
Yukon in North America. Landsc. Archit. 2021, 28, 113–118.
47. Jiang, F.; Zhang, J.; Song, P.; Qin, W.; Wang, H.; Cai, Z.; Gao, H.; Liu, D.; Li, B.; Zhang, T. Identifying priority reserves favors
the sustainable development of wild ungulates and the construction of Sanjiangyuan National Park. Ecol. Evol. 2022, 12, e9464.
[CrossRef] [PubMed]
48. Zhang, Y.; Yao, X.; Zhou, S.; Zhang, D. Glacier changes in the Sanjiangyuan Nature Reserve of China during 2000–2018. J. Geogr.
Sci. 2022, 32, 259–279. [CrossRef]
49. Peng, J.; Lyu, D.N.; Dong, J.Q.; Liu, Y.X.; Liu, Q.Y.; Li, B. Processes coupling and spatial integration: Characterizing ecological
restoration of territorial space in view of landscape ecology. J. Nat. Resour. 2020, 35, 3–13.
50. Qu, Y.; Zong, H.; Su, D.; Ping, Z.; Guan, M. Land use change and its impact on landscape ecological risk in typical areas of the
Yellow River Basin in China. Int. J. Environ. Res. Public Health 2021, 18, 11301. [CrossRef] [PubMed]
51. Wang, C.; Yu, C.; Chen, T.; Feng, Z.; Hu, Y.; Wu, K. Can the establishment of ecological security patterns improve ecological
protection? An example of Nanchang, China. Sci. Total Environ. 2020, 740, 140051. [CrossRef] [PubMed]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.