Landscape Pattern Changes and Climate Response in Nagqu Hangcuo National Wetland Park in the Tibetan Plateau

Abstract

Wetlands are vital ecosystems in the Tibetan Plateau that play a crucial role in water conservation, flood storage, and biodiversity maintenance. They are sensitive to climate change and typically have high ecological and environmental quality levels due to minimal human disturbance. This study aimed to quantify landscape pattern changes within the Nagqu Hangcuo National Wetland Park (NNWP) and identify the impact of climate on wetland change. Using remote sensing data from six periods (1990, 1995, 2000, 2005, 2010, and 2015), dynamic change analysis, landscape pattern analysis, and correlation analysis were employed to determine the evolutionary features of the wetland landscape and explore their relationship with climatic factors. The results showed that the total wetland area increased from 15.11 km² in 1990 to 15.23 km² in 2015. The meadow area increased the most among landscape types, primarily converted from swamps. Over 25 years, the fragmentation of the NNWP’s landscape increased while diversity decreased and its shape became more complex. Meadows were more sensitive to climatic factors than other landscape types, with correlation coefficients between wetland separation and sunshine duration being more significant than other climatic factors. Therefore, it is imperative to monitor landscape pattern changes and the effects of climate change to better protect wetland parks through long-term planning, suitable mechanisms, and advanced technology.

1. Introduction

Wetlands serve as vital transitional ecosystems between terrestrial and aquatic environments, playing a crucial role in drought mitigation, climate regulation, and ecological conservation [1,2,3,4,5]. In recent years, the study of wetland landscape ecology has gained significant attention in both the academic and public spheres. Central to this field of research is the wetland landscape pattern, which refers to the spatial arrangement of wetland landscape patches of varying sizes, shapes, and types resulting from diverse ecological processes at multiple scales. These landscape patterns have considerable influence on the ecological processes and functions of wetlands [6,7,8,9]. Consequently, the conservation and sustainable utilization of wetland ecosystems have become pressing concerns [10,11,12,13].

In arid alpine regions, climate plays a crucial role in sustaining the ecological processes and functions of wetlands. These regions are significantly influenced by factors such as hydrology and temperature [14,15,16]. The Tibetan Plateau, for instance, is experiencing a more rapid warming process compared to other areas worldwide, leading to swift changes in the ecological structure and function of its wetland ecosystems [17]. As a result, the protection of the ecological environment in arid alpine wetland areas faces unprecedented challenges.

In the 21st century, as satellite remote sensing (RS) and Geographic Information System (GIS) technologies continue to evolve, they have proven to be invaluable tools in wetland distribution research. These technologies provide cost-effective solutions for monitoring shifts in wetland landscapes and contribute significantly to their effective management [18,19,20,21,22,23,24]. Numerous researchers are currently employing technology and wetland landscape ecology methods to investigate the evolutionary features of the wetland landscape pattern and the climate response to wetland changes. However, much of this research has focused on plains wetlands or coastal wetlands [25,26,27,28,29,30,31,32,33,34], with relatively limited studies on the ecologically fragile and climate-sensitive lake wetlands of the Tibetan Plateau.

In December 2016, the Chinese government established the Nagqu Hangcuo National Wetland Park (NNWP) within the Tibetan Plateau. This park, characterized by minimal human disturbance and a high-quality ecosystem, is highly susceptible to climate change [35]. Understanding the wetland landscape pattern changes in the NNWP and identifying their causes are essential for the conservation of plateau wetlands and biodiversity, the enhancement of wetland protection systems, and the development of a national ecological security barrier. This study, therefore, examines the NNWP as a case study to analyze its landscape changes over a 25 year period, highlighting evolutionary features and elucidating the relationship between wetland landscape pattern evolution and climate change. The ultimate aim is to provide a theoretical foundation for safeguarding wetland landscape ecology and optimizing wetland landscape patterns.

The main contributions of this study are as follows:

• Quantifying the landscape pattern changes within the NNWP and identifying the influence of climate on wetland changes;
• Discovering that meadows are more sensitive to climatic factors than other landscape types and proposing targeted conservation measures to avoid meadow degradation;
• Concluding that increased attention must be given to landscape pattern changes and the effects of climate change in order to better protect wetland parks through long-term planning, appropriate mechanisms, and advanced technology.

The remainder of this paper is organized as follows. Section 2 introduces the general characteristics of the study region. Section 3 details the data and analysis methods, including data obtained from Landsat-5TM, Landsat ETM+, and Landsat OLI satellite images from 1990, 1995, 2000, 2005, 2010, and 2015. Section 4 discusses the results of the data analysis and interpretation. Finally, Section 5 presents several key conclusions drawn from the research.

2. Generality of the Study Region

NNWP (31°16′~31°19′ N, 91°45′~91°49′ E) is located in the northwestern part of Luoma Town in Nagqu County, between the Tanggula and Nianqingtanggula mountain ranges (Figure 1). The park covers an area of 15.66 km² and is home to a variety of wetland types, including permanent river wetlands, permanent freshwater wetlands, swamp meadow wetlands, and non-wetland prairies. The park’s water sources include mountain snow and rainfall.

The park is situated at an altitude of 4516–4556 m and features subrigid vegetation. The park’s wetlands are natural and rich in biodiversity. The average annual temperature is −1.06 °C, with a maximum temperature of 20.7 °C and a minimum of −7.9 °C. The annual precipitation is 514.2 mm, the average relative humidity is 56.72%, and the average sunshine duration is 7.57 h. It is located in an arid alpine region and contains Nagqu Lake, which has a crescent shape. The southern part of the park lies beneath the southern Qinghai–Tibet railway, while the eastern part is at the foot of Dakar. The park’s belt is 6 km long in an east-west direction and 4 km wide in a north-south direction.

3. Data and Analysis Methods

3.1. Data

3.1.1. RS Data

To ensure accurate extraction of remote sensing information, we utilized Landsat-5TM satellite imagery from the years 1990, 1995, 2005, and 2010, in addition to ETM+ images from 2000 and OLI images from 2015. The images were captured on 16 October 1990, 8 October 1995, 6 October 2000, 9 October 2005, 20 October 2010, and 18 October 2015, respectively. The data for our study was procured from the USGS website. These datasets, processed to Level 1T, have undergone systematic radiometric and geometric corrections and boast a spatial resolution of 30 m. We utilized ENVI 5.1 software for the initial preprocessing of these remote sensing images. To enhance the accuracy of interpretation, we undertook an individual band fusion operation for the Landsat images corresponding to each time point. This process culminated in the creation of a singular, multi-band image amalgamating information across multiple spectral bands, spanning the mid-infrared, near-infrared, and visible ranges. Subsequently, we applied the FLAASH atmospheric correction method to precisely offset the atmospheric effects on radiation. Concluding our preprocessing, these images were cropped in alignment with the boundaries of the NNWP. As there is no unified wetland classification standard, wetlands were classified based on the landscape types and characteristics of the study area. NNWP is small and its landscape is not complex, so classification was based on land use and cover status and correlative references [36,37,38] (

Table 1. NNWP landscape type and interpretation mark.

Types	Landscape Types	Landscape Features	Video Display
Wetland	Waterbody	Covering the surface without emergent plants; clear boundary, regular shape, image color band combination black or blue, dark blue, and uniform texture.
	Swamp	The land surface is covered with emergent vegetation (mainly Hippuris alginate, three-lobed base buttercup, and Kobresia) and perennial water accumulation, which displays patchy distribution and an irregular shape. The combination of image bands was green, light green, and rough texture.
	Meadow	The surface is covered with snow in winter and generally vegetation (mainly Carex and Kobresia) in other seasons. The soil is perennially moist and has a bulk distribution of strip-shaped, irregularly shaped, and uniform texture.
Non-wetland	Prairie	Surface area covered with snow in winter, and there are advantages to plant species, including the Artemisia community with dwarf plants and other communities. In the alpine soil all year round, in a dry state, clear border strip chunks are distributed, irregularly shaped, pale pink, and have an uneven texture.

).

The remote sensing interpretation of the NNWP was based on factors such as image color, shape, texture, surrounding relationships, and field observations. The study area was divided into non-wetland and wetland areas. Non-wetland areas included prairies, while wetland areas included waterbodies, swamps, and meadows. Combined with Google Earth remote sensing image data, a total of 225 training points were selected for interpretation, including 54 meadows, 84 waterbodies, 51 swamps, and 36 prairies. A support vector machine (SVM) was used for supervised classification, and misclassified features were reclassified through visual interpretation to obtain a wetland information distribution map. The results were validated using 55 field survey sample points with an overall classification accuracy of 87.24%, 85.45%, 92.72%, 89.09%, 89.09%, and 87.27%, respectively, and were found to be reliable. Fragstats 4.2 was used to analyze the landscape pattern index; ArcGIS 10.2 was used to obtain the NNWP landscape pattern classification; and SPSS 22.0 was used for statistical correlation analysis of meteorological data, wetland area, and landscape index.

3.1.2. Meteorological Data

Meteorological data include the daily average air temperature, precipitation, relative humidity, and sunshine hours from 1990 to 2015. It was obtained from the NAGQU meteorological station, which is also located in the NAGQU region (91°12′–93°02′ E, 30°31′–31°55′ N), with a similar altitude and 29 km from the study area. The National Meteorological Data Sharing Network maintains them all. Based on NAGQU meteorological weather data, the average temperature (October–November), the average annual rainfall (November of the previous year–November of the base year), the average sunshine duration (October–November), and the average relative humidity (October–November) were statistics.

3.2. Analysis Methods

3.2.1. Comprehensive Land-Use Dynamic Degree

The comprehensive land-use dynamic degree can reflect the degree of change of all land types in the NNWP over a certain period. The formula is as follows [27]:

$$LC=\frac{{\displaystyle {\sum }_{i=1}^n\Delta L{U}_{i-j}}}{{\displaystyle {\sum }_{i=1}^{n}L{U}_i}}\times \frac{1}{T}\times 100\%$$

(1)

where $L{U}_i$ is the area of land use type i at the starting time of monitoring; $\Delta L{U}_{i-j}$ is the absolute value of the area of land use type i converted to non-i land use type during the monitoring period; and T is the monitoring period.

3.2.2. Landscape Pattern Analysis Method

Landscape metrics were used to quantitatively study the evolutionary feature of the spatial pattern in the NNWP to analyze the landscape structure, the feature, and the rule of landscape pattern development. Since there was minimal variation in the features of the landscape in the NNWP, landscape boundaries tended to be simplified. To accurately obtain the landscape’s structural information, we combined the landscape analysis method to select different indices based on type and landscape level. For the type level, we considered Patch Density (PD), Percent of Landscape (PLAND), Landscape Shape Index (LSI), and Splitting Index (SPLIT). For the landscape class, we also considered Number of Patches (NP), Shannon’s Diversity Index (SHDI), Aggregation Index (AI), and Landscape Shape Index (LSI) [39,40]. The indicator values were calculated by the Fragstats 4.2 software.

3.2.3. Correlation Analysis Method

Changes in alpine wetland landscapes are primarily driven by both anthropogenic activities and climatic factors. However, in the case of the NNWP, the influence of human behavior is minimal due to its sparse population. Therefore, climatic factors are regarded as the prevailing contributors to shifts in its landscape patterns. Therefore, four climatic factors were selected to represent the four dimensions of temperature, moisture, light, and humidity: average temperature (October–November), average annual precipitation (November of the previous year–November of the base year), average sunshine hours (October–November), and average relative humidity (October–November). With SPSS 22.0 statistical software, Pearson correlation analysis was used to explore the relationship between changes in the landscape pattern and climatic factors in the NNWP. The formula is as follows:

$$\frac{1}{n-1}{\displaystyle \sum _{i=1}^n\left(\frac{X_i-\overline{X}}{{\sigma }_X}\right)}\left(\frac{Y_i-\overline{Y}}{{\sigma }_Y}\right)$$

(2)

where $\overline{X}$ is the sample mean of the sample $X_i$; ${\sigma }_X$ is the sample standard deviation; $\frac{X_i-\overline{X}}{{\sigma }_X}$ is the standard deviation fraction of the sample; and $Y$ is similarly.

4. Results, Analysis, and Discussion

4.1. Land Use and Land Cover in the NNWP

The change in the NNWP landscape reflects the conversion of the different types of wetlands. By analyzing the changes in various landscape areas, we obtained the evolutionary trend of the structure of wetlands. According to Figure 2 and

Table 2. Area of different types of alpine wetlands in the NNWP during the period 1990–2015.

Wetland Type		Meadow	Swamp	Waterbody	Wetland Area
1990	Total area (km2)	7.33	3.67	4.11	15.11
	Proportion	46.81%	23.44%	26.25%	96.49%
1995	Total area (km2)	7.88	2.93	4.19	15
	Proportion	50.32%	18.71%	26.76%	95.79%
2000	Total area (km2)	7.76	2.83	4.55	15.14
	Proportion	49.55%	18.07%	29.05%	96.68%
2005	Total area (km2)	7.2	3.74	4.06	15
	Proportion	45.98%	23.88%	25.93%	95.79%
2010	Total area (km2)	7.92	2.9	4.32	15.14
	Proportion	50.57%	18.52%	27.59%	96.68%
2015	Total area (km2)	8.52	2.29	4.42	15.23
	Proportion	54.41%	14.62%	28.22%	97.25%
1990–2015	Rate of change (km2/a)	0.0476	−0.0552	0.0124	-

, the meadow area—which accounts for the largest proportion of the NNWP’s overall area—showed an overall increasing trend over the course of 25 years. It had an increase of 1.19 km² and a growth rate of 16.23%. The swamp area fluctuated the most. In 2005, it reached a maximum of 3.74 km², accounting for 23.44% of the overall area of the park. However, by 2015, it had been reduced to a minimum of 2.29 km², accounting for only 14.62%. Overall, the swamp area decreased by 37.60% with a deceleration of 0.0552 km²/a in the past 25 years. Similarly, the overall fluctuation in the area of waterbodies between 1990 and 2015 was large. In 2000, it reached a maximum of 4.55 km², which accounted for 29.05% of the park’s overall area. However, it suddenly dropped to a minimum value of 4.06 km² in 2005. Over the next ten years, it slowly increased again. Generally, it indicated no wetland shrinkage in the NNWP in the past 25 years, and it seemed that the landscape of the NNWP has been transformed from terrestrial grassland landscape to original wetland landscape, based on the evolution of various wetland types.

According to the statistics in

Table 3. The NNWP landscape area transition matrix (km²) for 1990–2015.

Landscape Types in 1990	Landscape Types in 2015				Comprehensive Dynamic Degree (%)
	Meadow	Prairie	Waterbody	Swamp
Meadow	6.96	0.00	0.20	0.14	0.48
Prairie	0.13	0.38	0.00	0.03
Waterbody	0.01	0.00	4.10	0.01
Swamp	1.39	0.04	0.13	2.11

, the large dynamic degree of wetlands in the NNWP was consistently less than 1% from 1990 to 2015. During this period, the meadow and water areas increased while the prairie and swamp areas decreased. The most substantial change observed was in the meadow area, which increased by 1.19 km². This enlargement of the meadow area primarily resulted from the conversion of 1.39 km² of marshland. Meanwhile, the area of waterbodies increased by 0.31 km², largely stemming from the conversion of 0.2 km² of meadow and 0.12 km² of marsh, accounting for 61.29% and 38.71% of the total converted area, respectively. Conversely, the marshland area diminished, with a large discrepancy between the areas gained and lost. Specifically, 1.39 km² was converted into meadows, 0.12 km² into waterbodies, and 0.04 km² into grassland. Finally, the grassland area slightly decreased, with 0.13 km² being converted into meadow and 0.03 km² into marsh.

4.2. The Wetland Landscape Pattern Change

4.2.1. Landscape Pattern Changes in the Type Level

Landscape indicators such as Patch Density (PD), Percent of Landscape (PLAND), Landscape Shape Index (LSI), and Splitting Index (SPLIT) were used to analyze the patch area and structure of the NNWP at the class level (Figure 3). Figure 3a shows that the meadow has the highest PLAND value and is the dominant land cover in the NNWP. From 1990 to 2015, its PLAND value initially increased, then decreased, before increasing again. The lowest value was 46.01% in 2005, and the highest was 54.37% in 2015. A waterbody had a similar PLAND evolution to a meadow, while a swamp had the opposite trend, with a decrease of 37.06% over the period. The prairie had the lowest PLAND value but remained stable during this time. Figure 3b shows that PD increased for all types, indicating increasing fragmentation in the NNWP’s landscape. The PD change fluctuated significantly between types, with meadow and prairie steadily increasing while swamp and water rapidly increased and decreased mid-period. Figure 3c shows that LSI for meadow and swamp had similar trends, with an overall increase reaching their highest values of 5.68 and 6.89, respectively, in 2005 before slowly declining and slightly recovering at the end of the period. Waterbody’s LSI remained relatively stable except for a sudden high value in 2005. Prairie’s LSI gradually increased at a slightly accelerated rate in the later period. Overall, except for the waterbody, all landscape types became more complex in shape. Figure 3d shows that the waterbody and meadow had low SPLIT values, indicating small distances between their patches. Swamp’s SPLIT value fluctuated slightly over the period, with an initial increase followed by a decrease before increasing again. Prairie’s SPLIT value increased with strong fluctuations and remained higher than other landscape types, indicating intensified dispersion of its spatial distribution.

4.2.2. Landscape Pattern Changes in the Entire Area

At the landscape level, four indices—Number of Patches (NP), Shannon’s Diversity Index (SHDI), Aggregation Index (AI), and Landscape Shape Index (LSI)—were used to analyze the NNWP’s landscape pattern (Figure 4). According to Figure 4a, NP experienced an increase over time, with a peak value of 67 between 1990 and 2005. It then decreased before increasing again from 2010 to 2015. This overall increase indicates increasing fragmentation in the NNWP’s landscape. Figure 4b shows that SHDI decreased over the period except for a slight increase from 2000 to 2005, indicating reduced landscape heterogeneity in the NNWP. Figure 4c shows that LSI first increased and then decreased, with a maximum value of 5.63 in 2005 and an overall increase of 6.01%, indicating more complex patch shapes. Figure 4d shows that AI decreased then increased from 2005 to 2015, with a minimum value of 96.50% in 2005 and an overall decrease of 0.23%, indicating increased fragmentation. Over time, the landscape of the NNWP experienced increased fragmentation and complexity in patch shapes, while its diversity decreased.

4.3. Correlation Analysis of Wetland Evolution and Meteorological Factors

4.3.1. Climatic Factors and Correlation Analysis of the Wetland Landscape Area

The correlation coefficient between wetland landscape area and climatic factors such as annual average precipitation, average temperature, average sunshine duration, and average relative humidity was analyzed using SPSS 22.0 statistical software.

Table 4. Correlation between wetland type area and climatic factors in the NNWP from 1990 to 2015.

Landscape Types	The Average Annual Precipitation		Average Temperature		Sunshine Duration		Average Relative Humidity
	Correlation	Significance	Correlation	Significance	Correlation	Significance	Correlation	Significance
Meadow	0.695	0.125	0.567	0.241	0.962	0.002 2	−0.917	0.01 1
Swamp	0.584	0.535	0.380	0.458	−0.531	0.279	0.660	0.563
Waterbody	0.180	0.773	−0.263	0.615	0.427	0.398	−0.182	0.731

¹ indicates a significant correlation at 0.05. ² indicates a significant correlation at 0.01.

shows a significant positive correlation between the meadow area and sunshine duration, with a correlation coefficient of r = 0.962 (p < 0.01). According to Table 3′s land transformation direction, increased solar radiation may cause swamps to dry up and evolve into meadows. It suggests that sunshine may be a key driving factor for changes in the meadow area. The meadow area also had a significant negative correlation with average relative humidity, which may explain why some meadows changed into swamps due to increased humidity. No significant correlation existed between swamp or water area and any of the four climatic factors. A series of models were fitted to analyze the relationships between wetland area, precipitation, temperature, sunshine duration, and relative humidity in the NNWP. (Figure 5).

Upon analyzing Figure 5, we identified a sequential correlation in the swamp area with annual average precipitation, first exhibiting a positive correlation, followed by a negative one. The swamp area experienced an initial increase and then decreased with the annual average precipitation increase. The reason may have been that the meadow was converted to swamps as the annual average precipitation increased. When the average precipitation continued to increase, the swamps were flooded by the lake and converted into waterbodies. Along with the increment in annual average precipitation, water area increased, meadow area decreased, and some meadows were transformed into swamps and waterbodies. There was a negative correlation between the waterbody and the average temperature and a positive correlation between the swamp and meadow and the average temperature. The increased evaporation due to the temperature increment caused the transformation from a waterbody to a swamp or meadow. There was a negative correlation between the swamp area and sunshine duration and a positive correlation between the meadow area and sunshine duration. The reason may have been that the increase in sunshine duration caused the intensification of water evaporation, which indirectly promoted the reduction of the moisture content of the swamps and the transformation of swamps into meadows, as observed in the northern part of the NNWP. The area of the waterbody had a quadratic correlation with sunshine duration. A plausible justification for the expansion of waterbody areas might be attributed to the distinctive geographic conditions of the Tibetan Plateau. Situated at a higher altitude with thinner atmospheric conditions, the plateau experiences less attenuation of solar radiation. This intensified solar exposure prompts the accelerated melting of snow and ice in the surrounding mountainous regions, which subsequently contributes to increased runoff into the lake. However, with the further enhancement of solar radiation, the water loss caused by increased evaporation is greater than the increase in water from snow and ice melt [41]. There was a quadratic correlation between swamp area and average relative humidity, a positive correlation between meadow area and average relative humidity, and a negative and weak correlation between waterbody area and average relative humidity. Among the three types of wetland landscapes, the absolute value of the correlation coefficients between meadow area and climatic factors was the largest. It indicates that meadows were more sensitive to climatic factors than other landscape types. The absolute values of the correlation coefficients between waterbody area and climatic factors were the most minor, indicating that the waterbody area was affected less by changes in climatic factors. Alterations in wetland ecosystems are typically the outcome of an intricate interplay of various factors. Climatic determinants significantly impinge on these environments. For instance, prior research has demonstrated a notable augmentation of evaporation rates and extension of plant growth cycles within the wetlands of the Qinghai-Tibet Plateau [42], consequently influencing the extent of the wetland area. Over recent years, anthropogenic factors have increasingly come to the fore. Rapid advancements in tourism, population growth, and escalating livestock numbers have all catalyzed the degradation of these fragile ecosystems. It is also crucial to consider the variability in wetland areas stemming from seasonal fluctuations.

4.3.2. Analysis of Climatic Factors and Wetland Landscape Pattern Index

According to the correlation coefficients of wetland separation and climatic factors (annual average precipitation, average temperature, sunshine duration, and average relative humidity) in

Table 5. Correlation analysis of wetland landscape index separation and climatic factors in the NNWP from 1990 to 2015.

Landscape Index Separation	Average Annual Precipitation		Average Temperature		Average Relative Humidity		Average Sunshine Duration
	Correlation	Significance	Correlation	Significance	Correlation	Significance	Correlation	Significance
Meadow	0.446	0.375	−0.0114	0.83	0.652	0.161	−0.865 1	0.026
Swamp	−0.472	0.344	0.563	0.245	−0.773	0.071	0.856 1	0.03
Waterbody	0.083	0.876	−0.374	0.465	0.563	0.245	−0.774	0.071

¹ indicates a significant correlation at 0.05.

, the correlation coefficients between wetland separation (meadow separation, swamp separation, water separation, and average) and sunshine duration were −0.865 (p < 0.05), 0.856 (p < 0.05), and −0.774, respectively, and more significant than the correlation coefficients between wetland separation and other climatic factors (the annual average precipitation, average temperature, and average relative humidity). It indicated that the feedback of the wetland-type separation to average sunshine duration was more sensitive than the feedback of the wetland-type separation to annual average precipitation, average temperature, and average relative humidity. A model was established to show the relationship between wetland-type areas and climatic factors (average precipitation, temperature, sunshine duration, and relative humidity) in the NNWP (Figure 6). The results showed a positive correlation between meadow separation and annual average precipitation and a negative correlation between separation degree and annual average precipitation. The reason may have been that when the rainfall increased, the number of meadow patches increased because the swamp divided some of the meadows into more patches. It means the degree of meadow separation has changed to be higher. As the swamp was relatively concentrated and connected, swamp patches gathered up easily, and the degree of swamp separation reduced when the rainfall increased. Swamp separation had a positive correlation with the average temperature. The temperature increase led to reduced swamp water content, and some swamps turned into meadows. Water separation negatively correlates with the average temperature. It may be due to the supplementary water from the melting of snow and ice caused by the temperature increase. The waterbody was less fragmented, and the landscape separation was reduced. The correlation between meadow separation and the average temperature was not obvious because meadows have the largest proportion in 25 years, and a relatively small number of transformations from swamps to meadows only have a little effect on the landscape separation of the meadow. Swamp landscape separation had a negative correlation with average relative humidity, while meadow separation and water separation had a positive correlation with average relative humidity. The main reason was that the increment in average relative humidity caused the meadows to be transformed into swamps, and the swamp divided the meadow and waterbodies due to the increasing separation between meadows and waterbodies. Based on the analysis, it can be inferred that as the global climate warms, wetland landscape separation and fragmentation in the NNWP will increase. The NNWP’s geographical position within an ecologically vulnerable zone, coupled with its significant exposure to natural elements, results in the complexity of its landscape fragmentation process and influencing factors. This complexity is further compounded by periodic disturbances, variations in wetland types, and the impact of human activities. Each of these elements not only affects the NNWP but also contributes to the degree of landscape fragmentation.

5. Conclusions

Since the 21st century, the impact of climate change and human factors on the evolution of wetlands has become a hot issue. Climate change has a massive impact on the natural water supply in alpine regions, resulting in a change in the wetland structure. Based on RS images and climate data from 1990 to 2015, this study analyzed the evolutionary features of the wetland landscape patterns and their relationship with the climate in the NNWP in the past 25 years with the RS technology and related analysis methods. From the above study, the main conclusions are as follows:

(1)

The total wetland area increased from 15.11 km² in 1990 to 15.23 km² in 2015, with meadow area increasing the most among landscape types primarily converted from swamps;

(2)

Fragmentation of the NNWP’s landscape increased while diversity decreased and shape became more complex over 25 years;

(3)

Meadows were more sensitive to climatic factors than other landscape types, with correlation coefficients between wetland separation and sunshine duration being more significant than other climatic factors;

(4)

It is imperative to monitor landscape pattern changes and the effects of climate change to better protect wetland parks through long-term planning, suitable mechanisms, and advanced technology.

Based on the research conducted in this article, it is crucial to pay attention to the change in landscape patterns and the impact of climate change. A long-term plan, suitable mechanisms, and advanced technology are necessary to protect wetland parks better. Future research could focus on identifying the specific mechanisms that contribute to the observed changes in landscape patterns and how these changes affect the ecosystem services provided by wetlands. Additionally, it would be interesting to investigate how different management strategies could be used to mitigate the impacts of climate change on wetland ecosystems. Overall, this study provides valuable insights into the dynamics of wetland ecosystems in the Tibetan Plateau and highlights the importance of continued research in this area.