The Spatiotemporal Evolution of Wetlands Within the Yarlung Zangbo River Basin and Responses to Natural Conditions from 1990 to 2020

Abstract

The wetland in the Yarlung Zangbo River Basin is an important part of the ecological barrier zone of the Qinghai–Tibet Plateau and exerts a significant influence on the climate. To elucidate the evolutionary characteristics and potential causes of wetlands in the Yarlung Zangbo River Basin against the background of “warming-humidification” of the plateau, this study focused on the spatial–temporal changes of wetlands in the Yarlung Zangbo River from 1990 to 2020 and simultaneously discussed the contribution of natural factors to these wetland changes. The data used in this study encompassed meteorological observation, the Digital Elevation Model (DEM), land use remote sensing monitoring, the vegetation index and other relevant data, and the methods used were mainly hydrological analysis, landscape change dynamic analysis and GeoDetector. The research findings indicated the following: (1) The wetland area in the Yarlung Zangbo River Basin exhibits significant fluctuations. The wetland area increased steadily from 1990 to 2005, followed by a slight decline after 2005, reflecting the changing process of “humidification–drought–humidification–drought”. Nevertheless, the overall trend over the 30 years has been an increase in wetland area (a total increase of 14.92%), primarily driven by the conversion of forest and grassland. (2) The wetlands in the Yarlung Zangbo River Basin are mainly distributed in the lower river basin, especially in the Niyang River basin and the Yigong–Parlung Zangbo basin. The spatial distribution of these wetlands remained relatively stable over the 30 years studied. (3) The driving factor analysis results showed that the three main natural factors leading to the increase and reduction in wetland area include vegetation cover change, precipitation and evapotranspiration. Vegetation cover change contributed the most to the increase in wetlands in the Yarlung Zangbo River basin, and evapotranspiration played a decisive role in the reduction in wetland area. This study provided valuable perspectives for wetlands, water resources and ecosystem assessments in the Yarlung Zangbo River Basin and the broader Qinghai–Tibet Plateau region.

1. Introduction

The wetland system, often referred to as the “kidney of the Earth” due to its unique structure and diverse functions, is vital for human existence and development. Wetlands are sensitive areas that serve as critical indicators of global change. In the context of global change, the characteristics of wetland evolution can reveal whether the ecological environment of a region is developing healthily. Different types of wetlands, shaped by varying climatic conditions and developmental processes, respond to the impacts of climate change and human activities in distinct evolutionary ways [1]. This makes the study of the evolution characteristics of wetlands and their driving factors a core topic in the protection, restoration and management of wetland ecosystems.

A considerable amount of research has demonstrated that the evolution of alpine wetland groups, which possess unique climatic backgrounds [2], exhibits fluctuations in wetland areas across various spatial and temporal scales [3,4,5]. These fluctuations primarily reflect increases and reductions in wetland areas, rather than a simple decline. Changes in natural conditions and the spatial distribution of key meteorological elements within wetland areas may directly drive wetland change [6]. This suggests that the factors influencing wetland change result from the complex interplay of various natural elements against the backdrop of “warming and humidification” [7,8]. Rising temperatures have been shown to be significantly correlated with the overall degradation of alpine wetlands [9,10]. Additionally, wetland changes are notably responsive to variations in precipitation, evaporation, relative humidity, and the frozen soil environment [11]. In different regions, climatic differences can lead to distinct evolutionary characteristics of wetlands [12,13,14,15]. However, the existing studies still have certain limitations. Firstly, for specific river basins, such as the Yarlung Zangbo River Basin, there is a relative lack of systematic research on the relationship between wetland evolution and climate change over a long time span. Secondly, they mainly focus on a single or a few driving factors, lacking in-depth analysis of the comprehensive driving effect of natural factors. Thirdly, the research on the spatial evolution characteristics of the wetland landscape and the distribution features of sub-basins, as well as the trends of dry and wet environments, is not sufficiently detailed.

The Yarlung Zangbo River Basin, as an important drainage area in southwestern China, is located in the core area of the ecological barrier of the Qinghai–Tibet Plateau. Its high-altitude terrain, complex water system and unique climate conditions have created a rich and diverse wetland ecosystem. However, for the Yarlung Zangbo River Basin in the 1990–2020 period, for the Yarlung Zangbo River Basin, questions remain regarding whether the climate change conforms to the “warm and humid” trend of the Qinghai–Tibet Plateau, the temporal and spatial evolution laws of the wetlands, the distribution characteristics of the landscape changes in sub-basins, the trends of dry and wet environments and the differences in the contribution of natural factors to the evolution of the wetlands in different time periods.

Based on high-precision elevation data, long-term climate monitoring data and land use remote sensing data, this paper comprehensively uses methods such as hydrological analysis, landscape transformation amplitude calculation and geographic detectors to systematically explore the temporal and spatial evolution laws and dry–wet characteristics of the wetlands in the Yarlung Zangbo River Basin, as well as to deeply analyze the driving effects of natural factors on the changes in the wetlands. Compared with previous studies, this research not only extends the time series and expands the data sources, but also realizes the detailed analysis of the evolution process of the wetlands through the integration of multiple methods, aiming to provide a new perspective and solid theoretical basis for the scientific protection and precise restoration of the wetland ecosystem in this basin.

2. Study Area and Data

2.1. Research Area

The Yarlung Zangbo River Basin is situated in the Tibet Autonomous Region, encompassing a geographical range of 82°1′ to 97°6′ E and 27°49′ to 31°17′ N. It is the longest plateau river in China, with a length of 2057 km within the country and a total area of approximately 250,000 km². The average elevation in the basin exceeds 4600 m, with a maximum elevation variation of over 7000 m (Figure 1). The Yarlung Zangbo River originates from the Chemayungdung Glacier in the northern foothills of the Himalayas in southwestern Tibet. It flows from west to east across southern Tibet, veering south over the easternmost peak of the Himalayas towards Pasighat. The basin features numerous tributaries of the Brahmaputra River, five of which have drainage areas exceeding 10,000 square kilometers: the Dogxung Zangbo River, Nian-Chu River, Lhasa River, Niyang River and Parlung Zangbo River. The great Brahmaputra Valley, which is formed in the south, is a significant highland barley-producing region in China. The natural environment of the basin is complex, with diverse vegetation types. The upper reaches are primarily covered by alpine meadows and alpine steppes, while the middle reaches consist mainly of shrubby steppes, and the lower reaches are characterized by trees and secondary vegetation. Due to the influence of atmospheric circulation and topography, annual precipitation gradually reduces from the southeast to the northwest of Tibet, with the Heyuan area of Zhongba County receiving less than 300 mm of precipitation annually. The average annual temperature in the basin is approximately 6 °C, and evapotranspiration is around 700 mm. Water energy reserves are abundant. During the summer, the flood season of the Yarlung Zangbo River is driven by the melting of ice and snow on the Qinghai–Tibet Plateau, as well as the substantial precipitation brought by the summer monsoon from the southwest. The upstream river is primarily recharged by snowmelt, while the middle and lower reaches receive more rainwater recharge. The Yarlung Zangbo River Basin is a key development zone for the “One River and Two Rivers” agricultural comprehensive development project proposed by the Tibet Autonomous Region government.

2.2. Data

The primary spatial data utilized in this study included Digital Elevation Model (DEM), land use remote sensing data, meteorological data, vegetation index and other spatial data pertaining to the Yarlung Zangbo River Basin in China, along with relevant socio-economic statistics. The specific details of the data used are shown in

Table 1. The specific details of the data used.

Data Name	Source	Resolution	Purpose
Digital Elevation Model (DEM)	Geospatial Data Cloud Platform of the Chinese Academy of Sciences (http://www.gscloud.cn (accessed on 1 November 2021))	90 m	The topographic basic data are used for surface analysis and hydrological analysis based on ARCGIS to determine the study area and sub-basin differentiation.
Land Use Data	Research achievements of Professor Huang Xin’s team at Wuhan University [16]	30 m	Used for wetland extraction, land use analysis and landscape conversion calculation within the basin, and needs to be reclassified into specific land use types.
Meteorological Data (Temperature/Precipitation)	Resource and Environment Science Data Center of the Chinese Academy of Sciences	500 m	Generate raster distribution maps through spatial interpolation, calculate the changes every 5 years and analyze the explanatory power for wetland changes.
Meteorological Data (Evapotranspiration)	China 1 km Monthly Potential Evapotranspiration Data (1990–2020) of the National Tibetan Plateau Science Data Center [17]	1 km	Extract effective data and accumulate them into annual total evapotranspiration raster data to analyze the explanatory power for wetland changes.
Normalized Difference Vegetation Index (NDVI)	Global long-term data of MODIS and AVHRR products fused by Shen Huafeng’s team at Wuhan University [18]	1 km	Fuse growing season data for the analysis of influencing factors in wetland changes
Water System Data	Spatio-Temporal Three Poles Environmental Big Data Platform, Water System Dataset of Resource and Environment Science Data Center of the Chinese Academy of Sciences	-	Determine the study area and sub-basin differentiation through the ArcGIS hydrological analysis module based on DEM

.

3. Key Technologies and Methods

3.1. Hydrological Analysis

Using the ArcGIS hydrological analysis module and SRTM DEM terrain product data, this study referred to the Yarlung Zangbo River water resources spatio-temporal distribution dataset from the Spatio-Temporal Three Poles Environment Big Data Platform, as well as the China river system basin spatial distribution dataset from the Chinese Academy of Sciences Resource and Environment Science Data Center. The smallest watershed area was adjusted to 121,500 km² to delineate the Yarlung Zangbo River basin, resulting in a final study area of 251,854 km², as illustrated in Figure 1.

To better understand the characteristics of wetland changes in the study area over the past 30 years, the Yarlung Zangbo River Basin is divided into 11 relatively complete sub-basins. The spatial distribution of mean values and variability of vegetation indices and climate factors is analyzed as a unit to explore the spatial and temporal distribution characteristics of natural factors. To ensure the scientific validity and rationality of the spatial statistical analysis and research, as well as to increase the number of samples and reduce the threshold for valley formation, the Yarlung Zangbo River Basin is further subdivided into 329 complete sub-basins. This allows for a more detailed examination of the changes in natural driving factors and wetland area within each sub-basin. The driving forces of wetland evolution are then assessed using geographical detectors.

3.2. Dynamic Analysis of Landscape Transformation

Based on the research theme of regional wetland evolution and drawing on the work regarding landscape transformation and the dynamics of land cover types [1,19], this study defines the formula for calculating landscape transformation amplitude as follows:

$$C=\sum \left\{\left(j-i\right)·S_{ij}\right\}$$

(1)

where $C$ represents the landscape transformation amplitude (measured in km²), and $S_{ij}$ denotes the transformation area in the landscape transformation matrix across different years. The indices i and j refer to landscape types i and j, respectively. The term $(j-i)$ represents the conversion coefficient. The absolute value of the landscape conversion coefficient $(j-i)$ is smaller for similar land cover types and larger for different land cover types. A positive conversion range indicates that the landscape is developing toward increased wetness, while a negative range suggests a shift toward drought conditions. For further details, refer to

Table 2. Landscape conversion coefficient of Yarlung Zangbo River Basin.

Conversion Coefficient	Water	Swamp	Snow and Ice	Grassland	Cropland	Other Terrestrial Types
Water	0	1	2	3	4	5
Swamp	−1	0	1	2	3	4
Snow and ice	−2	−1	0	1	2	3
Grassland	−3	−2	−1	0	1	2
Cropland	−4	−3	−2	−1	0	1
Other terrestrial types	−5	−4	−3	−2	−1	0

. The $(j-i)$ conversion coefficient is analyzed over a 5-year cycle to assess the dynamic changes in wetlands within the Yarlung Zangbo River basin, utilizing the landscape transformation amplitude. With the aid of the ArcGIS spatial analysis tool, land use raster data from various years were reclassified according to the landscape classification outlined in the landscape conversion coefficient table (Table 2). The pairwise conversion results were generated using tabulate area tool, which was then matched and multiplied by the landscape conversion coefficient. The landscape conversion amplitude for all raster units was subsequently summed and analyzed statistically, revealing the characteristics of landscape dynamic change in the Yarlung Zangbo River basin.

3.3. Geographic Detector

GeoDetector [20,21] is a set of statistical methods used to identify spatial differentiation and uncover the underlying driving forces. The fundamental concept is as follows: if a study area is divided into several subregions, and the sum of the variances within these subregions is less than the total variance of the entire region, this indicates the presence of spatial differentiation. Furthermore, if the spatial distributions of two variables exhibit a consistent pattern, a statistical correlation between them can be inferred. The q statistic of GeoDetector (as shown in Equation (2)) serves to measure spatial differentiation, explore explanatory factors, and analyze the interactions between variables. GeoDetector has been applied across various fields in both the natural and social sciences.

$$q=1-{\displaystyle \frac{\sum _{h=1}^L{N}_h{\sigma }_h^2}{N{\sigma }^2}}$$

(2)

where h = 1…, L is the classification or partition of variable Y or factor X; N_h and N are the number of units in a certain category h and the whole region, respectively. ${\sigma }_h^2$ and ${\sigma }^2$ are the variances of the Y values of a certain category h and the whole region, respectively. ${\displaystyle \frac{\sum _{h=1}^L{N}_h{\sigma }_h^2}{N{\sigma }^2}}$ is, the ratio of the sum of variances of multiple classes to the total variances of the whole region. The range of q is [0, 1], and the larger the value, the stronger the driving effect of the factor, and the weaker the driving effect. A q value of 0 indicates that the factor has no relationship with the target factor, and the value of q indicates that the factor explains the 100 × q% target. Based on the raster data of each factor, this study uses the reclassification function of ArcGIS (Jenks Natural Breaks Method) to aggregate the raster data of the impact factors into 5 categories, and uses the zonal statistical (mean) function of 329 subbasins to process the attributes of each factor, and this is used in turn to match the increased area of wetlands (the area of Y1 converted to other land use types) and the reduced area of wetlands (Y2: area of other land use types converted to wetlands), wetland area change (Y3) and influence factors (X) of temperature change (X1), precipitation change (X2), vegetation index change (X3) and evapotranspiration change (X4) in each sub-basin during each period. We use Excel programming GeoDetector open-source geographic detector (http://www.geodetector.cn/ (accessed on 10 November 2021)), calculation of single factor and composite factor of wetland area, as well as the contribution rate of change.

4. Results and Analysis

4.1. Characteristics of Wetland Change in the Yarlung Zangbo River Basin

4.1.1. Spatial and Temporal Changes in Wetland Distribution in the Yarlung Zangbo River Basin

The primary wetlands in the Yarlung Zangbo River Basin consist of water bodies, including lakes and rivers, as well as swamps. Among these, the area of the water bodies is relatively extensive (Figure 2a), with an average size of 2727.58 km² from 1990 to 2020. The total area exhibits an overall increasing trend. It increased from 1990 to 2005, followed by a slight decline after 2005, returning to the total area recorded in 2000. In contrast, the swamp area is considerably smaller (Figure 2b), with an average size of 6.543 km². This area experienced a slight increase in 1995, but subsequently reduced, reaching a 30-year low in 2000. However, the total swamp area began to rise again from 2000 to 2020. The overall increase in wetland area within the 30a watershed accounted for 14.92% of the total wetland area in 1990.

Over the past 30 years, the total net area of five land use types, including cropland, grassland, snow and ice, impervious surfaces and swamp, has been positive, indicating that these land use type conversions have led to an increase in water area. The primary sources of this increase are attributed to the conversion of snow and ice (328.284 km²) and grassland (112.656 km²), which account for 97.95% and 33.61% of the final net increase in water area, respectively.

Conversely, the cumulative net areas for forest land, shrubs and barren land have shown negative values, suggesting that transitions involving these three land use types have resulted in a reduction in water area. Notably, forest land has consistently shown negative values across all six analyzed conversion periods; cumulatively, 89.217 km² of water has been converted into forested areas—with significant conversions occurring between 1990 and 2000, before gradually declining from 2000 to 2020. Barren lands were characterized by positive values before 2005, indicating a continuous conversion from barren lands to water bodies during the period from 1990 to 2005. However, post-2005 data indicate a reversal, in which water is being converted back into barren lands. Other land use types also exhibit dynamic interactions with water bodies at various intervals: croplands experienced conversions from water bodies between 1990 and 1995 but transitioned predominantly from croplands back into water bodies thereafter, starting in 1995. The exchanges involving construction land, swamps and shrubs relative to water bodies remained relatively minimal throughout all six periods examined. According to the data analysis from over three decades in

Table 3. Area conversion between wetlands (water and swamp) and other land use types in the Yarlung Zangbo River (unit: km²).

Land Use Type	Period	Cropland	Forest	Shrub	Grassland	Water	Snow and Ice	Barren Land	Impervious Surface	Swamp	Summary
water	1990–1995	−0.109	−26.635	−0.015	−10.511	0.000	46.202	27.627	0.047	0.016	36.622
	1995–2000	1.646	−34.861	0.000	178.655	0.000	70.380	47.465	0.986	0.525	264.796
	2000–2005	0.635	−4.980	0.005	284.385	0.000	78.041	65.864	−0.503	0.117	423.563
	2005–2010	0.292	−9.514	0.000	−45.945	0.000	76.652	−62.800	−0.392	0.078	−41.630
	2010–2015	0.194	−11.585	−0.004	−240.575	0.000	35.256	−25.339	0.212	0.043	−241.797
	2015–2020	0.168	−1.643	−0.021	−53.353	0.000	21.754	−73.004	−0.302	0.009	−106.390
	Summary	2.827	−89.217	−0.035	112.656	0.000	328.284	−20.187	0.047	0.788	335.163
swamp	1990–1995	−0.084	0.000	0.000	3.452	−0.016	0.000	0.000	0.000	0.000	3.353
	1995–2000	−0.094	−0.002	0.000	−2.786	−0.525	0.000	−0.002	0.000	0.000	−3.407
	2000–2005	−0.307	−0.021	0.000	1.547	−0.117	0.000	0.000	0.000	0.000	1.103
	2005–2010	0.030	−0.001	0.000	1.152	−0.078	0.000	0.000	0.000	0.000	1.103
	2010–2015	−0.101	0.000	0.000	3.195	−0.043	0.000	0.000	0.000	0.000	3.051
	2015–2020	0.032	0.000	0.000	6.111	−0.009	0.000	0.000	0.000	0.000	6.134
	Summary	−0.523	−0.023	0.000	12.672	−0.788	0.000	−0.002	0.000	0.000	11.336

, swamp land increased by a total area of approximately (11.3355 km²), representing 3.5 times its total extent in 1990. This was most notably driven by conversions from grasslands into swamp lands, while other transformations reflect shifts from swamp lands into alternative uses. The overall findings suggest that there is an ongoing trend, wherein swamp lands are increasingly converting into water bodies.

4.1.2. Spatio-Temporal Analysis of Landscape Dry–Wet Conversion Amplitude in the Yarlung Zangbo River Basin

The primary key indicators of the dry–wet conversion of wetland landscapes are the net changes from dry to wet (or vice versa) [3]. To analyze the changes in landscape types within the basin, a five-year time period was selected. Based on Equation (1) and Table 1, the statistical map illustrating the amplitude of landscape transformation in the Yarlung Zangbo River Basin from 1990 to 2020 was calculated and analyzed (Figure 3), which showed that the cumulative conversion range within 30 years is greater than 0, indicating that the Yarlung Zangbo River Basin has been developing toward wet habitats since 1990, and the landscape transformation can be roughly divided into four periods. The first period was from 1990 to 2000, during which, the range of landscape transformation in the basin was large. In the two five-year intervals, the areas were 2299.02 km² and 15,112.04 km², respectively, and the cumulative value was 17,411.06 km², all of which were greater than 0, indicating that more patches of land use type changed to wet habitats in this period. During the second period, from 2000 to 2005, the conversion amplitude of the whole basin was less than 0 (−1169.17 km²), indicating that the conversion amplitude of the whole basin was developing in the direction of dry habitat. However, the conversion amplitude was smaller than that of the previous period, so the cumulative landscape conversion amplitude was still greater than 0 (16,241.89 km²). During the third period, from 2005 to 2015, the conversion amplitude of the whole basin began to be greater than 0 again, indicating that the transformation direction of the whole basin began to develop in the direction of wet and biological habitat, and the conversion amplitude was about 2000 km². The accumulation of the two five-year intervals further increased the overall conversion amplitude. In the fourth period, from 2015 to 2020, the conversion amplitude of the whole basin began to be less than 0 (−2891.10 km²), which was a small transition to dry habitat compared with the third period.

Based on the changes in spatial distribution, ArcGIS overlay analysis and regional statistics were employed to visualize the landscape transformation and further analyze the spatial distribution changes in the Yarlung Zangbo River wetlands. In comparison to the total basin area, the changes in the spatial distribution of the Yarlung Zangbo River wetlands were not significant; therefore, only the spatial distribution of the wetlands (water bodies and swamps) in 2020 is presented (Figure 4a). The figure illustrates that the sub-basins with higher concentrations of wetland distribution are the lower river basins of the Yarlung Zangbo River, the Niyang River basin and the Yigong–Parlung Zangbo Basin. All the areas with wetland distribution in the past 30 years were merged and reclassified for visualization (Figure 4b). It was found that the areas with relatively stable wetland spatial distribution accounted for a large proportion, mostly distributed along the Brahmaputra River valley and mountain gullies, accounting for 48% of the statistical area and for 43% of the areas with constant wetland distribution in the past 30 years. The areas with the least conversion frequency (wetlands with 1–2 changes) accounted for 29%, and their spatial distribution was relatively uniform in the basin. The areas with medium conversion frequency (wetlands with 3–5 changes) accounted for 23%, and most of them were distributed in the lower reaches of the Yarlung Zangbo River basin.

4.2. Research on the Influencing Factors in Wetland Change in the Yarlung Zangbo River Basin

4.2.1. Influencing Analysis of Single Factor

The Yarlung Zangbo River Basin was divided into 329 sub-basins, and the influence of temperature changes, precipitation changes, evapotranspiration changes and vegetation cover changes (four independent variables, and the subsequent ones were abbreviated as follows: TC; PC; EC; and VCC) on the change in wetland (water body, swamp and the total amount of both) area (three dependent variables) in the Yarlung Zangbo River Basin were analyzed. Among the four natural factors, differences between the two periods, reclassification and zonal statistics were carried out, and 329 sample types were obtained. Wetland area change refers to the area of wetland (water body and swamp) increase, reduction and total change every 5 years within the sub-basins. The results of the GeoDetector (

Table 4. Results of geographic detector driven by a single factor.

Periods	1990–1995						1995–2000						2000–2005
Dependent Variables	Increased Area of Wetlands		Reduced Area of Wetlands		Total Area of Wetland Change		Increased Area of Wetlands		Reduced Area of Wetlands		Total Area of Wetland Change		Increased Area of Wetlands		Reduced Area of Wetlands		Total Area of Wetland Change
Independent Variables	q	p	q	p	q	p	q	p	q	p	q	p	q	p	q	p	q	p
TC	0.07	0.00	0.01	0.76	0.03	0.03	0.15	0.00	0.00	0.86	0.10	0.00	0.13	0.00	0.03	0.09	0.12	0.00
PC	0.05	0.01	0.01	0.35	0.02	0.13	0.03	0.05	0.03	0.03	0.03	0.04	0.05	0.00	0.01	0.37	0.04	0.01
VCC	0.10	0.00	0.01	0.54	0.07	0.00	0.05	0.01	0.01	0.54	0.04	0.01	0.09	0.00	0.01	0.45	0.07	0.00
EC	0.02	0.16	0.04	0.02	0.02	0.28	0.06	0.00	0.01	0.58	0.05	0.01	0.02	0.18	0.01	0.62	0.02	0.12
Periods	2005–2010						2010–2015						2015–2020
Dependent Variables	Increased Area of Wetlands		Reduced Area of Wetlands		Total Area of Wetland Change		Increased Area of Wetlands		Reduced Area of Wetlands		Total Area of Wetland Change		Increased Area of Wetlands		Reduced Area of Wetlands		Total Area of Wetland Change
Independent Variables	q	p	q	p	q	p	q	p	q	p	q	p	q	p	q	p	q	p
TC	0.08	0.00	0.01	0.42	0.08	0.00	0.05	0.00	0.05	0.00	0.03	0.04	0.05	0.01	0.03	0.04	0.03	0.10
PC	0.06	0.00	0.04	0.01	0.07	0.00	0.05	0.00	0.01	0.68	0.03	0.03	0.04	0.03	0.02	0.13	0.03	0.10
VCC	0.06	0.00	0.00	0.86	0.05	0.00	0.10	0.00	0.02	0.16	0.10	0.00	0.06	0.00	0.02	0.22	0.06	0.00
EC	0.01	0.67	0.02	0.27	0.01	0.52	0.05	0.00	0.02	0.11	0.04	0.02	0.04	0.02	0.01	0.61	0.04	0.01

) were as follows: (1) According to the p value, significant differences were only found in the increase in wetlands and the total change of wetlands in the detection of the four factors (p < 0.05), and the frequency was in the order of VCC >TC > PC > EC. (2) The contributions of the four factors to the increase and reduction in and total change of wetlands in the six periods in the study area were different. However, the factor intensity ranking and the factor with the strongest explanatory power were consistent with the total change of the wetlands at each period. Table 4 shows that from 1990 to 1995, with an interval of 5 years, the explanatory power of the factors influencing wetland increase was ranked as VCC > TC > PC > EC, which is consistent with the explanatory power of the factors influencing wetland change, but inconsistent with the following explanatory power ranking of the factors influencing wetland reduction: EC > PC > VCC > TC. The following five periods are roughly similar. Meanwhile, according to the p and q statistics, the strongest factors influencing wetland increase in the six five-year periods were VCC, TC, TC, TC, VCC and VCC. The strongest factors influencing wetland reduction were EC, PC, TC, PC, TC and TC. The strongest factors influencing the overall change of wetlands were VCC, TC, TC, TC, T and VCC.

4.2.2. Influence Analysis of Pairwise Factor Interactions

In the interaction detection results, the top two groups of interaction factors ranked by q-values were selected to generate

Table 5. Statistical table of pairwise factor interaction effect.

Periods	1990–1995			1995–2000
	Increased Area of Wetlands	Reduced Area of Wetlands	Total Area of Wetland Change	Increased Area of Wetlands	Reduced Area of Wetlands	Total Area of Wetland Change
Strongest interaction	VCC ∩ EC	VCC ∩ EC	VCC ∩ EC	TC∩ VCC	PC ∩ EC	TC∩ VCC
q	0.187	0.102	0.148	0.299	0.087	0.219
Second interaction	PC ∩ EC	PC ∩ EC	PC ∩VCC	TC∩ EC	TC∩ EC	TC∩ EC
q	0.173	0.086	0.121	0.272	0.075	0.190
Periods	2000–2005			2005–2010
	Increased Area of Wetlands	Reduced Area of Wetlands	Total Area of Wetland Change	Increased Area of Wetlands	Reduced Area of Wetlands	Total Area of Wetland Change
Strongest interaction	TC ∩ VCC	VCC ∩ EC	TC ∩ VCC	TC ∩ VCC	PC ∩ EC	PC ∩ VCC
q	0.262	0.060	0.228	0.186	0.139	0.206
Second interaction	TC ∩ EC	PC ∩ VCC	TC ∩ EC	PC ∩ VCC	TC ∩ PC	TC ∩ VCC
q	0.212	0.059	0.188	0.178	0.124	0.181
Periods	2010−2015			2015−2020
	Increased Area of Wetlands	Reduced Area of Wetlands	Total Area of Wetland Change	Increased Area of Wetlands	Reduced Area of Wetlands	Total Area of Wetland Change
Strongest interaction	TC ∩ VCC	TC ∩ VCC	TC ∩ VCC	TC ∩ EC	TC ∩ VCC	TC ∩ EC
q	0.170	0.466	0.213	0.247	0.094	0.203
Second interaction	PC ∩ VCC	TC ∩ EC	PC ∩ VCC	PC ∩ EC	TC ∩ PC	PC ∩ EC
q	0.167	0.166	0.163	0.220	0.091	0.190

. By integrating Table 4 and Table 5, it is evident that most pairwise interactions among the six types of factors enhance the explanatory power for wetland expansion, wetland reduction and total wetland dynamics. Over the six five-year periods within the 30-year timeframe, the strongest interaction factor combinations explaining wetland expansion and total wetland change showed relatively consistent patterns, whereas the factor combinations influencing wetland reduction diverged from those of wetland expansion and total change.

The statistical analysis reveals that among the interaction factor combinations with the highest explanatory power for wetland expansion, the “TC ∩ VCC” combination appeared consecutively in four five-year periods, from 1995 to 2015. In contrast, the “VCC ∩ Evapotranspiration changes” and “TC ∩ EC” combinations each occurred once, in the first and last five-year intervals of the study period. For total wetland change, the “TC ∩ VCC” combination dominated in three five-year periods (1995–2005 and 2010–2015), while “VCC ∩ EC”, “TC ∩ EC” and “PC ∩ VCC” each emerged as the strongest combination in the one remaining period. In the context of wetland reduction, the strongest explanatory combinations—” VCC ∩ EC”, “PC ∩ EC”, and “TC ∩ VCC”—each appeared twice, without discernible temporal regularity.

In terms of secondary explanatory power, the interaction factor combinations for wetland expansion were ordered as “PC ∩ VCC” (the occurrence frequency is 3.), “PC ∩ EC” (the occurrence frequency is 2.) and “TC ∩ EC” (the occurrence frequency is 1.). For total wetland change, the secondary rankings were “PC ∩ VCC” (the occurrence frequency is 2.), “TC ∩ EC” (the occurrence frequency is 2.), “TC ∩ VCC” (the occurrence frequency is 2.) and “PC ∩ EC” (the occurrence frequency is 1.). For wetland reduction, the secondary hierarchy comprised “TC ∩ EC” (the occurrence frequency is 2.), “TC ∩ PC” (the occurrence frequency is 2.), “PC ∩ EC” (the occurrence frequency is 1.) and “VCC ∩ PC” (the occurrence frequency is 1.).

Synthesizing the magnitude of q-values and occurrence frequency, the “TC ∩ VCC” combination demonstrated the strongest explanatory capacity for wetland expansion, reduction and total dynamics in the Yarlung Zangbo River wetland across all the detected factor interactions.

5. Discussion

5.1. The “Warming–Humidification” Trend in the Yarlung Zangbo River Basin

Based on various land cover types, this study developed a dry–wet conversion coefficient to assess the degree of landscape conversion and to explore the characteristics of dry–wet changes in the Yarlung Zangbo River basin. The findings indicate that the basin generally transitioned into a wet habitat between 1990 and 2020, with a notable dry–wet conversion point occurring around the year 2000. This observation is closely aligned with the results obtained by Liu et al. [22,23] and Li et al. [24], using standardized precipitation evapotranspiration index and vegetation index methods. Further analysis of the trend of annual mean temperature change in the Yarlung Zangbo River Basin during 1990–2020 (Figure 5) showed that the annual mean temperature in the Yarlung Zangbo River Basin during the past 30 years has ranged from 5.29 °C to 7.02 °C, with an average of 6.04 °C and an extremely significant upward trend, with a change tendency rate of 0.254 °C/(5a). This is consistent with the results of other relevant studies on the Tibetan Plateau [25]. The spatial differences in temperature gradually increased from the upper reaches to the middle and lower reaches of the basin (Figure 6a), but the regions with higher mean temperatures were concentrated in the lower reaches of the mainstream of the Yarlung Zangbo River basin, and the average temperature in these areas was higher than 7.36 °C. In the past 30 years, the regions with the fastest spatial temperature growth in the basin were the Niyang River basin, the middle reaches of the Yarlung Zangbo River and the Lhasa River basin (Figure 6b), and the five-year average increase rate exceeded 0.02 ° C. The trend of temperature rise and the change of humid habitat were consistent with the warming and humidification trend of the Qinghai–Tibet Plateau.

Furthermore, by comprehensively analyzing Figure 4a and Figure 6a and combining with statistical data, it was found that the distribution of wetlands in the lower reaches of the Yarlung Zangbo River, where the average temperature is higher, is also more extensive; however, Figure 4b and Figure 6b show that the distribution areas where the temperature rose rapidly do not match the areas with more frequent changes in wetlands.

5.2. Temporal and Spatial Changes of Wetlands in the Yarlung Zangbo River Basin

The total area of water and swamps in the Yarlung Zangbo River Basin showed an increasing trend from 1990 to 2020. The main contributors to this change are two land use types: ice and snow, and grassland. In contrast, the spatial distribution of wetlands has remained relatively stable, which aligns with the findings of Li Peijun’s analysis of land use and landscape patterns in the Yarlung Zangbo River Basin based on topographic gradients [24]. To further explore the spatio-temporal change characteristics of the wetlands, the spatial distribution of the four land use types of converted ice, snow, water, swamp and grassland was visualized, as shown in Figure 7. It was found that the snow and ice, grassland and swamp converted into wetlands were mostly distributed around the water body, and most of the swaps converted into water bodies were also located around the water body. Considering the overall warming and wetting background of the Qinghai–Tibet Plateau, melting water from ice and snow has increased and merged into streams, submerged the grassland and swamps around the original water body and, finally, presented the transformation of ice and snow and grassland into water body, which can also explain the continuous transformation of swamps into water bodies in the past 30 years, to a certain extent. Many studies have also confirmed the relationship between glacial meltwater and alpine wetlands [26,27,28,29]. The drastic area conversion between wetland and grassland is consistent with the relevant research results for the Tarim River basin [30], other areas of the Qinghai–Tibet Plateau [31,32] and the Sichuan Nanmoqie Wetland National Nature Reserve [33].

5.3. Factors Influencing the Increase and Reduction in Wetland Areas

In this study, GeoDetector was used to evaluate the degree to which natural factors influence wetland changes in the Yarlung Zangbo River Basin. The findings revealed that vegetation coverage is the most significant factor contributing to the increase in wetland area, while evapotranspiration is the primary factor leading to wetland reduction. Research indicated that a warm and wet climate promotes an increase in the total volume of water resources in the basin and enhances the coverage of surface vegetation [34,35], thereby impacting wetland area changes. Consequently, although the contribution of vegetation cover is relatively substantial among the factors influencing the increase and reduction in wetland areas, temperature and humidity changes ultimately drive changes in vegetation cover [36], making it the fundamental factor influencing wetland area expansion [4]. Additionally, increased evaporation has been shown to be a crucial driving force behind wetland reduction [37], with a more pronounced effect on alpine wetlands [38].

5.4. Limitation

A complete water collection network forms a natural basin. The natural geographical features within the basin are similar and consistent [39], so that the natural factors influencing wetland changes can be examined through the lens of watershed differentiation. By calculating the mean value of each index within the basin, the driving factors can be summarized while minimizing the impact of outliers. Utilizing data from the DEM, this study was divided into 326 sub-basins by adjusting the area threshold during the hydrological analysis. The average values of the index in each sub-basin were calculated and extracted to determine the contribution rate of the impact factors. Compared with previous studies on pixel-by-pixel raster operation [40] and the method of extracting index values from fishing net mesh points [41], this study was unable to further measure the differences caused by the research results, which is the direction to be explored in the future. In addition, the scale of this study was relatively large, and the evolution of wetlands within each sub-basin was not analyzed in depth. The next step is to consider refining the wetland within each sub-basin with the support of high-resolution data and incorporating distinct natural factors. For example, the evolution of wetlands at different elevation levels, the interaction between vegetation species evolution and wetland evolution, and the regions and driving forces of frequent changes in wetland spatial distribution need to be further investigated and discussed.

6. Conclusions

In this paper, we employed hydrological analysis, land use transfer matrices, landscape transformation amplitude, and other methods to examine the spatial and temporal changes of wetlands in the Yarlung Zangbo River Basin from 1990 to 2020. Furthermore, we investigated the natural factors influencing wetland changes using GeoDetector. The main findings of this study can be summarized as follows:

(1)

Since 1990, the Yarlung Zangbo River Basin has undergone a changing process of “humidification–drought–humidification–drought”. Over 30 years, this has led to the development of a wet habitat. The critical period for the transformation of this habitat occurred between 1995 and 2000.

(2)

The changes in the wetland water bodies and swamp areas of the Yarlung Zangbo River Basin over the past 30 years have been significant. Initially, the water area increased, followed by a subsequent reduction, with most of the wetland areas converted from snow and ice and grassland. From 1990 to 2000, the swamp area experienced only slight changes; however, it began to increase after 2000, primarily due to the conversion of grassland.

(3)

The spatial distribution of the wetlands in the Yarlung Zangbo River Basin has remained relatively stable over the past 30 years. Most of the wetlands are concentrated in the lower river basins of the Yarlung Zangbo River, the Niyang River, and the Yigong–Parlung Zangbo basin. Notably, the location of half of the total wetland area has remained unchanged.

(4)

The single-factor driving analysis indicated that vegetation cover played a significant role in the increase in wetlands and the overall change in wetland areas. In contrast, precipitation and evapotranspiration, which contributed to the reduction in wetlands, were ranked higher than vegetation cover. Furthermore, the interaction analysis of pairwise driving factors revealed that vegetation cover is the most influential factor.