The Disparity in Normalized Difference Vegetarian Index Response to Climate Warming and Humidification in the Tibetan Plateau before and after 1998

Abstract

The Tibetan Plateau (TP) serves as a crucial ecological barrier in Asia, with vegetation playing a pivotal role in the terrestrial ecosystem by facilitating energy exchange between the land and atmosphere, regulating climate, and participating in the carbon cycle. In this study, we analyze the characteristics of surface vegetation on the TP in the growing season during 1982–2018 using satellite remote sensing data obtained from the National Oceanic and Atmospheric Administration (NOAA) and China Meteorological Forcing Dataset (CMFD). We investigate how these characteristics respond to climate change under different warming and humidification conditions across the TP. The main conclusions are as follows. (1) The normalized difference vegetation index (NDVI) values on the TP exhibit a gradual decrease from southeast to northwest during the growing season. There is a significant overall increasing trend at a climate tendency rate of 0.01·decade⁻¹ (p < 0.01) from 1982 to 2018, characterized by a notable mutation of around 1998. Over the past 37 years, a polarized trend of vegetation was observed on the TP, with notable improvement in its central and eastern regions. However, there has been noticeable degradation in northwestern TP, specifically within the Kunlun Mountains and Qaidam Basin. (2) The climate of the TP demonstrates distinct regional disparities in terms of warming and humidification characteristics before and after 1998. During the period of 1982–1998 (1998–2018), the temperature increase is primarily concentrated in the northern (southern) TP, while precipitation increase is mainly observed in the southern and northwestern (northeastern and western) regions of the TP. (3) The responses of surface vegetation to climate factors exhibit significant variations across diverse climatic backgrounds. It is noteworthy that moisture conditions have a substantial impact on the response of vegetation to air temperature on the TP. During the period of 1982–1998, under relatively insufficient moisture conditions, a positive correlation was observed between air temperature and surface vegetation in the humid and semi-humid regions of the southeastern TP, while a negative correlation was found in the semi-arid regions of northeastern TP. During 1998–2018, as moisture conditions became relatively sufficient, surface vegetation in the semi-arid regions showed positive correlations with both temperature and precipitation. However, surface vegetation in the humid and semi-humid regions exhibited a significant negative correlation with precipitation. During this period, the synergistic effects between warm and humid climates in the semi-arid regions of northeastern TP and warm and dry climates in humid and semi-humid regions of southeastern TP substantially enhanced surface vegetation on the TP. Furthermore, our results indicate that thermal factors (air temperature) primarily influence variations in surface vegetation within the high-altitude arid region of the TP. During 1998–2018, a significant cooling trend was observed in the northwestern TP, which could potentially account for the degradation of surface vegetation in the Kunlun Mountains. The findings of this study establish a scientific basis for the sustainable development of grassland ecosystems on the TP.

1. Introduction

The Tibetan Plateau (TP) is the world’s largest and highest plateau [1,2], with an average elevation exceeding 4000 m and a total area surpassing 2.5 million km². It serves as a crucial ecological barrier for China [3], encompassing expansive alpine grasslands, meadows, deserts, and forests. However, due to its elevated topography and intricate underlying conditions, it exhibits high sensitivity to climate variations and is an ecologically fragile region [1,2,4]. Vegetation plays a vital role in the terrestrial ecosystem by facilitating the terrestrial carbon cycle, energy exchange, and climate regulation [5,6,7,8]. With global climate warming, the TP’s climate exhibits evident signs of warming and humidification [9,10,11], manifested as temperature increase, precipitation rise, accelerated permafrost degradation, and the expansion of lake areas, among others [12,13,14,15,16,17]. In recent years, intensified climate change has added more uncertainty to the fragile ecosystem of the TP, and the response of vegetation to climate change has become increasingly complex [4,18]. Therefore, investigating the response of vegetation growth on the TP to climate change holds immense significance in protecting the ecological environment of the TP and maintaining global ecosystem stability.

The normalized difference vegetation index (NDVI) is an important indicator of vegetation growth, providing an effective technical means for studying large-scale and long-term characteristics of vegetation change and its response to climate change [19,20]. Research indicates that temperature and precipitation, as the main manifestations of hydrothermal conditions, are important environmental factors affecting dynamic changes in vegetation [21], and the warming and humidification of climate leads to the intensification of vegetation activities, which is the main reason for the increase in NDVI [22,23]. In the context of global warming, vegetation activity in the northern hemisphere has exhibited an increase since the 1980s, with a corresponding rise in NDVI observed across most regions of China [24,25,26]. However, due to variations in climate across latitudes, there are significant differences in how NDVI responds to climate change. At higher latitudes, low temperatures and insufficient water availability limit vegetation growth, while warming conditions facilitate it [27,28,29]. Conversely, in middle and low latitudes, the relationship between mountain vegetation and temperature gradually weakens or exhibits a negative correlation [21,30]. Notably, on the TP, researchers have demonstrated an overall improvement trend with localized degradation characteristics in vegetation [31,32], where interannual variations in alpine grassland coverage are significantly influenced by climate factors [17,18,33]. However, the impact of different climatic factors may vary across different spatiotemporal scales and types of vegetation [26]. Generally speaking, plant growth in alpine meadows exhibits higher sensitivity to temperature variations, whereas those in drier steppes are predominantly determined by moisture conditions [34].

Research indicates that following the deceleration of global warming in 1998, the TP experienced a more pronounced warming trend [35]. Concurrently, there has been an increasing trend in precipitation and potential evaporation [10,36,37]. Therefore, this study aims to investigate regional disparities and interdecadal variations in surface vegetation on the TP response to climate factors during different stages of global warming. The findings of this research provide a scientific foundation for sustaining the long-term development of grassland ecosystems on the TP.

2. Data and Methods

2.1. Study Area

The TP is located in the southwestern region of China, spanning from 73°E to 104°E and from 26°N to 39°N (Figure 1a). It not only holds the distinction of being the highest plateau globally but also enjoys a reputation as Asia’s Water Tower [38] due to its vast frozen water storage, second only to polar regions. The TP exhibits a distinctive climate characterized by consistently low temperatures, with an annual average air temperature ranging between −2.2 and 0 °C [39]. Precipitation over the TP is significantly influenced by the westerlies and Asian summer monsoon, with approximately 73% occurring during the summer monsoon period [40]. Total precipitation during the vegetation growing season exceeds 600 mm in southeastern TP and is below 50 mm in the northwest (Figure 1b), resulting in a corresponding humidity gradient from humidity in the southeast to arid conditions in the northwest [41]. According to the classifications of vegetation type and the eco-geographical region on the TP [2,42], this study defines an arid climatic zone as having a total precipitation below 200 mm during the vegetation growth season. A semi-arid climatic zone is defined as having a total precipitation ranging from 200 to 500 mm, while a humid and sub-humid climatic zone is characterized by a total precipitation greater than 500 mm (Figure 1c). This classification facilitates the analysis of surface vegetation responses to climate change across diverse climatic zones on the TP.

2.2. Remote Data

NDVI, defined as the ratio of the difference between near-infrared reflectance and red visible reflectance to their sum, is an indicator of vegetation greenness [43]. NDVI data used in the present study have a horizontal resolution of 0.05° × 0.05° and are derived from the National Oceanic and Atmospheric Administration (NOAA) based on the surface reflectivity climate data record (CDR) of the grid dataset (https://www.ncei.noaa.gov/products/climate-data-records/normalized-difference-Vegetarian-index, accessed on 5 October 2023). The dataset is daily observed by NOAA’s polar-orbiting satellites, which utilize either the Advanced Very-High-Resolution Radiometer (AVHRR) or the Visible Infrared Imaging Radiometer (VIIRS) [44]. As a preliminary step, the maximum value compositing method was used to process daily data into monthly data in order to further reduce cloud contamination and enhance the vegetation signal [45]. According to Du et al. [46] and Zhang et al. [47], NDVI data obtained from the AVHRR exhibit a robust correlation alongside that derived from the Moderate Resolution Imaging Spectrometer (MODIS). Both data sources consistently respond to various climatic variables, indicating their reliability for analyzing vegetation changes and their association with climate [23].

2.3. Climatic Data

The climatic data used in the present study include the following: (1) 3-hourly near-surface air temperature and precipitation rate data at a spatial resolution of 0.1° from China Meteorological Forcing Dataset (CMFD), which is provided by the National TP Scientific Data Center (https://data.tpdc.ac.cn/zh-hans/data/8028b944-daaa-4511-8769-965612652c49/, accessed on 5 October 2023). CMFD is the first high spatial-temporal resolution-gridded near-surface meteorological dataset developed specifically for studies of land surface processes in China. This dataset is a fusion product based on remote sensing products, reanalysis datasets, and in-situ data from 740 China Meteorological Administration (CMA) stations [48]. It was found that the CMFD exhibits superior quality compared to that of GLDAS (Global Land Data Assimilation System) when validated against independent station observations, primarily because it utilizes a larger number of stations for data generation [48]. (2) Monthly mean temperature and total precipitation observations were obtained from 105 CMA stations located in central–eastern TP (Figure 1b). These data are collected, quality-controlled, and provided by the CMA [49]. All datasets include records from January 1982 to December 2018; furthermore, May through September is considered the vegetation growing season on TP.

2.4. Methods

The linear relationships between surface vegetation and climate factors over the TP are conducted through regression, and Pearson’s correlation, and partial correlation analyses. Statistical significance is assessed using a two-tailed Student’s t-test [50,51]. In the analysis of unary linear regression, the climatic tendency rate of a certain variable is defined as b × 10, where b represents its linear trend changing over time [52]. Additionally, the Mann–Kendall (M-K) mutation test [53] was employed to examine the mutation characteristics of the climate variables discussed in this study.

Analysis of variance is used to measure the contribution rates of climate factors to the change in surface vegetation [54]. In multiple linear regression analysis, the regression variance equals the sum of the variance contributions from each factor, which can be expressed as follows:

$${s_{\widehat{y}}}^2={\displaystyle \sum _{k=1}^P{b}_k{s}_{ky}},$$

(1)

where y is the predicted variable; $\widehat{y}$ is the regression value of the predicted variable; k is the forecast factor; P is the number of forecast factors; b_k is the regression coefficient of the factor; and s_ky is the covariance of the forecast factor and the predicted variable. Therefore, the contribution rate of a certain forecast factor k compared to the change in the predicted variable y is expressed as

$$R_k=\frac{b_k{s}_{ky}}{s_y^2}\times 100\%,$$

(2)

where $s_y^2$ is the variance of the predicted variable. This method can accurately determine the key meteorological factors that affect the interannual variation in NDVI [52]. Additionally, the linear trends of the variables were eliminated when employing this method.

3. Results

3.1. Spatiotemporal Variability of NDVI on the TP in the Growing Season

The spatial variation in NDVI in the growing season on the TP and its distribution of the climate tendency rate from 1982 to 2018 is depicted in Figure 2. As can be observed, the NDVI of the TP during the growing season ranges from 0.1 to 0.7, gradually decreasing from southeast to northwest, with an average value of 0.28. (Figure 2a). The regions with higher NDVI values (>0.4) are predominantly located in the humid and semi-humid zone in the eastern part of the TP, at the intersection of Sichuan, Qinghai, and Gansu provinces, encompassing approximately 26.4% of the total area of the plateau (Figure 1c and Figure 2a). The regions characterized by lower NDVI values (<0.2) are primarily situated in the arid zone in the northern part of TP, including the Kunlun Mountains and Qaidam basin, covering approximately 47.2% of its total area (Figure 1c and Figure 2a). The central part of TP features a transitional zone extending from grassland to the desert in a northeast-to-southwest direction, exhibiting NDVI values ranging from 0.2 to 0.4. This transitional region encompasses about 26.4% of the TP’s total area and falls within a semi-arid climatic zone (Figure 1c and Figure 2a).

The spatial distribution of the NDVI climate tendency rate (Figure 2b) reveals a predominant increase in vegetation across the TP. Notably, significant increases are observed near Qinghai Lake in the northeast TP, the Sanjiangyuan region, the western Sichuan Plateau, and southeast Tibet. These regions collectively account for 41.1% of TP’s total area. Among them, the most prominent increase is observed in the Sanjiangyuan region with a climate tendency rate of 0.079·decade⁻¹, passing the t-test at a 99% confidence level (i.e., p < 0.01). Conversely, there is a significant decreasing trend in vegetation within the northern part of TP, specifically in the Kunlun Mountains and Qaidam Basin (−0.040·decade⁻¹; p < 0.01), accounting for 13.6% of TP’s total area. This indicates the polarization of surface vegetation on the TP over the past 37 years (1982–2018). While notable improvement has been observed in central and eastern parts of TP, substantial degradation is evident in northwestern TP, particularly within the Kunlun Mountains and Qaidam Basin.

The interannual variation in regional mean NDVI on the TP during 1982–2018 and its M-K mutation test results are presented in Figure 3. As shown in Figure 3a, the average NDVI of the TP exhibits a linear increasing trend with a climate tendency rate of 0.01·decade⁻¹ (p < 0.01), varying from 0.24 to 0.30. Prior to the mid-1990s, the NDVI values were small, followed by a rapid increase, peaking in 2002. Subsequently, there was a downward fluctuation until 2014, with a slight recovery thereafter. The M-K mutation test reveals that the growing season’s NDVI on the TP exhibited a continuously increasing trend from 1982 to 2018. A significant mutation occurred in 1994 and reached a significant level of 0.05 in 1998 (Figure 3b). Therefore, it is crucial to investigate both the extent and regional variations in this abrupt change in vegetation across the TP. To comprehensively address this objective, we divide the period from 1982 to 2018 into the following two distinct periods, P1 (1982–1998) and P2 (1998–2018), enabling us to analyze specific characteristics of NDVI variation in the TP during the growing season.

The distributions of mean NDVI in the growing season on the TP during P1 and P2, as well as their difference, are depicted in Figure 4. A comparison between Figure 4a,b reveals a significant westward expansion of the contours with an NDVI value equal to 0.2 during P2 compared to that in P1. Furthermore, there is a notable enhancement in surface vegetation in the southeastern and central regions of TP, accompanied by an increase in the area exhibiting an NDVI value greater than 0.4 from 23.8% to 28.4%. However, it is noteworthy that within the northern area of TP, particularly encompassing the Kunlun Mountains and Qaidam Basin, substantial degradation in vegetation can be observed (Figure 4b), resulting in a rise from 4.8% to 10.8% in the area displaying an NDVI value below 0.1.

The positive difference in NDVI between P1 and P2 (P2 compared to that in P1) accounted for 77.0% of the total TP area, whereas the negative difference accounted for 23.0% (Figure 4c). The magnitude of vegetation increase was 3.3 times greater than that of vegetation decrease, which is consistent with the overall enhancement of the plateau vegetation shown in Figure 2b. In conclusion, significant regional disparities and interdecadal variations have been evident in surface vegetation changes during the growing season on the TP over the past 37 years.

3.2. Applicability Evaluation of CMFD to TP

Firstly, we assess the suitability of CMFD for the TP. Due to limited observation stations in the western TP, we primarily compared the temperature and precipitation distributions derived from interpolation data from stations in the central and eastern TP with those obtained from CMFD (Figure 5). The spatial distribution of near-surface air temperature over the TP depicted in Figure 5a,b closely corresponds to observed patterns based on meteorological station data. This correspondence effectively captures spatial variations in higher air temperatures across the southeast region of the TP and Qaidam Basin in northern TP, as well as lower air temperatures within the central region of the TP. Over the past 37 years, changes in regional average CMFD temperature over the TP were highly consistent with those of meteorological stations (Figure 5c), exhibiting a correlation coefficient of 0.98 (p < 0.01), and their warming rates were 0.46 °C·decade⁻¹ and 0.40 °C·decade⁻¹, respectively, showing a significant warming trend (p < 0.01). However, due to the sparse meteorological stations on the TP, particularly in the higher-altitude regions, we found that average temperatures reported by meteorological stations were approximately 5.23 °C higher than those reported by CMFD.

Comparing Figure 5d,e, it is evident that the CMFD surface precipitation rate over the TP aligns remarkably well with the interpolated precipitation center derived from meteorological station observations. Moreover, their spatial distributions exhibit striking similarities, effectively capturing the gradual decrease in precipitation from the southeast to northwest during the growing season over the TP. The correlation coefficient between the average precipitation of meteorological stations on the TP and the average precipitation rate of CMFD is 0.75 (p < 0.01). Overall, both datasets demonstrate a significant increasing trend in precipitation during the growing season over the TP. The normalized climate tendency rates for these datasets are 0.59·decade⁻¹ and 0.30·decade⁻¹, respectively (Figure 5f).

In conclusion, the temperature and precipitation distributions over the central and eastern TP obtained from CMFD data and meteorological station observations exhibited remarkable similarity, with significantly correlated inter-annual variations (p < 0.01). Therefore, high-resolution CMFD data demonstrate excellent applicability for studying the climatic characteristics of the TP. Both datasets effectively capture the observed warming and humidification trends over the TP during 1982–2018.

3.3. Characteristics of Climate Warming and Humidification over the TP

The above analysis reveals significant regional disparities and interdecadal abrupt shifts in surface vegetation during the growing season on the TP from 1982 to 2018. Therefore, it is crucial to investigate whether there are also regional variations and inter-decadal abrupt changes in TP warming and humidification. Consequently, we use CMFD data to clarify this matter. The M-K mutation test results for near-surface air temperatures and precipitation rates over the TP are shown in Figure 6. It is evident that both air temperature and precipitation experienced abrupt changes in 1990 and 1994, respectively. Subsequently, significant warming and humidification trends were observed in 1995 and 1999, reaching significant levels of 0.01 and 0.05, respectively (Figure 6a,b). These findings are consistent with the significant increase in NDVI observed in 1998 (Figure 3b), indicating a warmer and wetter climate on the TP since the mid-to-late 1990s. Furthermore, this substantial vegetation growth can be attributed to the combined influence of temperature and precipitation. However, it remains unclear whether there are regional variations in the warming and wetting patterns across different parts of the TP. To address this question, we examined the disparities in warming and humidification characteristics between P1 and P2.

Figure 7 illustrates the distributions of climate tendency rates of near-surface air temperature and precipitation over the TP in different periods, as well as the difference between P1 and P2 (P2 compared to that in P1). The analysis of Figure 7a,b reveals a consistent warming and humidification trend across the majority of the TP from 1982 to 2018. Importantly, approximately 66.1% of the total area exhibited a statistically significant increase (p < 0.05) in temperature, with particularly pronounced changes observed in the southwestern TP and Qaidam Basin in the northern TP, where the maximum climate tendency rate reached 2.54 °C·decade⁻¹. Conversely, approximately 13.2% of the total area, including the Kunlun Mountains in northwestern TP, experienced a significant decrease in temperature (Figure 7a). Moreover, significant increases in precipitation were mainly observed in the western, northern, and southern regions of TP, covering approximately 58.5% of its total area with a maximum climate tendency rate of 0.11 mm·h⁻¹·decade⁻¹ (Figure 7b).

During P1, the temperature of almost the entire TP exhibits a consistent upward trend, particularly in the northern TP, where the climate tendency rate can reach 2.7 °C·decade⁻¹ (Figure 7c). The precipitation pattern shows significant deviations from that observed during 1982–2018, with approximately 68.2% of the region showing an increasing trend and around 12.7% experiencing significantly higher rainfall, mainly in the southern and northwestern parts of the TP. However, both northeastern and southwestern regions of the TP witness a declining trend in precipitation, with approximately 2.1% of these areas becoming considerably drier (Figure 7d).

During P2, the temperature trend over the TP exhibits a distinct deviation from that observed during P1. The regions experiencing a significant increase in temperature are primarily situated in the eastern and southwestern parts of the TP, accounting for 38.5% of its total area. In contrast, significant decreases in temperature are observed in the northwestern and southern portions of the TP, which represent 10.9% of its total area (Figure 7e). The precipitation trend exhibits an almost opposite pattern compared to that during P1. There is an increasing trend in the northeast and western part of the TP, encompassing 51.3% of its total area; meanwhile, a significant decreasing trend is observed in the southern and north–central part of the TP, representing 9.9% of its total area (Figure 7f). The spatial distributions of the temperature and precipitation rate differences between P1 and P2 (P2 compared to that in P1) are depicted in Figure 7g,h. It is evident that when compared to P1, more than 60% of the TP experiences significant warming and humidification during P2. Notably, approximately 14.8% of these areas exhibit a temperature difference exceeding 1.5 °C, primarily concentrated in the southwestern region. The areas with significant humidification are mainly located in the western, northern, and southern parts of the TP, with the precipitation rate differences reaching up to 0.20 mm·h⁻¹. These findings are consistent with the overall warming and humidification trends observed on the TP from 1982 to 2018 (Figure 7a,b).

To summarize, from 1982 to 2018, the TP experienced a significant warming and humidification trend, accompanied by abrupt changes in temperature and precipitation during the mid and late 1990s. There are distinct regional variations in patterns of warming and humidification before and after these abrupt shifts. Regarding warming, the temperature increase primarily occurred in the northern (southern) TP during P1 (P2). Concerning humidification, increased precipitation mainly concentrated in the southern and northwestern (northeastern and western) TP during P1 (P2).

3.4. Responses of NDVI to Climatic Factors over the TP

The consistency analysis of the NDVI trend alongside that of the near-surface air temperature and precipitation shows that (Figure 8) the variation trend of NDVI was highly consistent with that of near-surface air temperature and precipitation rate during the vegetation growing season from 1982 to 2018, except for the Kunlun Mountains and Qaidam Basin in the northwest part of the TP. The proportion of areas where the NDVI trend aligns with near-surface air temperature was found to be 72.6%, while areas exhibiting significant increases in both parameters accounted for 51.1%. These regions are primarily concentrated in the Sanjiangyuan region located in the central–eastern part of the TP (Figure 8a). Similarly, there is a high degree of consistency between the NDVI trend and near-surface precipitation rate, with approximately 73.5% of areas displaying alignment. Among these regions, about 37.8% exhibited significant increases in both variables, mainly situated in the central and northeastern parts of the TP (Figure 8b).

In general, the response of surface vegetation to the near-surface air temperature trend is more pronounced compared to that of the precipitation rate trend. The significant enhancement of vegetation during the growing season from 1982 to 2018 can be attributed to long-term climate warming and humidification on the TP. However, during P1 and P2, there is a lack of consistency between NDVI and both near-surface air temperature and precipitation trends, resulting in a notable reduction in areas exhibiting consistent increases for both factors. This suggests that short-term fluctuations in temperature and precipitation do not elicit a significant response from the vegetation. Therefore, it is crucial to comprehensively investigate the interannual response of surface vegetation to climate change while also examining their differences during different stages of climate warming and humidification over the TP. To address this issue comprehensively, we first eliminated the linear trend of all utilized data and subsequently performed partial correlation analysis along with variance analysis.

The results of partial correlation analysis between NDVI on the TP and near-surface air temperature (with the influence of precipitation removed) and precipitation rate (with the influence of temperature removed) in different periods are presented in Figure 9. During P1, a positive–negative–positive pattern was observed between NDVI and near-surface air temperature, showing a gradient from northeast to southwest (Figure 9a). Conversely, an insignificant negative partial correlation was found between NDVI and precipitation rate (Figure 9b). This indicates that temperature plays a pivotal role in driving changes in surface vegetation across the TP during this period. In the humid and semi-humid regions located in the southeast, as well as the high-altitude arid areas situated in the northwest of the plateau, there existed a positive correlation between air temperature and surface vegetation. Consequently, warming conditions are conducive to vegetation growth within these areas. On the other hand, within the semi-arid regions positioned centrally on the TP, a negative relationship was observed between air temperature and surface vegetation. Therefore, warming conditions do not favorably affect vegetation growth within this region.

During P2 (Figure 9c,d), compared to P1, the partial correlations between NDVI and near-surface air temperature (with the influence of precipitation removed), as well as precipitation rate (with the influence of temperature removed), exhibited significant enhancements during the growing season. Approximately 52.8% of the area demonstrated a positive correlation between vegetation and near-surface air temperature, particularly in the Sanjiangyuan region located in the central–eastern part of TP (Figure 9c). Moreover, there was a positive correlation between surface vegetation and precipitation rate in approximately 54.6% of the area, primarily concentrated in the northeast, central, and western parts of the TP (Figure 9d). This indicates that both temperature and precipitation influence variations in surface vegetation across TP during this period. It is worth noting that a significant negative correlation exists between NDVI and the precipitation rate within the southeastern TP. During this period, as depicted in Figure 7e,f, the temperature in this region experienced a substantial increase while precipitation decreased. The reduced precipitation resulted in an extended duration of available light for vegetation, leading to a pronounced enhancement in thermal conditions for surface vegetation growth and consequently causing a notable increase in vegetation within the southeastern TP.

The contribution rates of near-surface air temperature and precipitation to the NDVI on the TP gained from the multiple regression analysis between P2 and P1 (P2 compared to that in P1) are illustrated in Figure 10. During P2, there was a significant increase in the contribution rate of temperature to NDVI, particularly in southern Qinghai and central Tibet, with the Sanjiangyuan region showing an increase surpassing 20% (Figure 10a). Additionally, there was an increased contribution rate of precipitation to NDVI observed in 74.4% of the regions, especially in the western Sichuan Plateau and central-eastern Qinghai, where it rose by more than 20% (Figure 10b).

By comparing Figure 9 and Figure 10, it was evident that the increasing contribution rates of temperature and precipitation to NDVI on the TP (Figure 10) aligned with the significant correlation areas depicted in Figure 9. The outcomes of comprehensive partial correlation analysis and variance analysis revealed notable differences in surface vegetation response to climate change across different climatic zones on the TP. During P1, under conditions of relatively low precipitation, air temperature was the primary factor that influenced vegetation growth. In the southeastern humid and semi-humid regions of the TP, as well as in the high-altitude arid areas in the northwest, surface vegetation predominantly exhibited a positive correlation response to changes in air temperature. Conversely, in the northeastern TP and the semi-arid regions within TP’s central area, surface vegetation primarily demonstrated a negative correlation response to variations in air temperature. However, during P2, when water conditions became relatively sufficient, the growth of surface vegetation was influenced by both heat (air temperature) and water (precipitation). In the humid and semi-humid areas of southeastern TP, surface vegetation was negatively correlated with precipitation. There was a significant increase in temperature and a decrease in precipitation in this region (Figure 7e,f), which greatly enhanced light and heat conditions conducive to surface vegetation growth. In the semi-arid area of central TP, surface vegetation showed a positive correlation with both air temperature and precipitation factors, leading to substantial improvement in surface vegetation over a large area of this region. In the arid region of northwestern TP, while surface vegetation still demonstrates a positive correlation with air temperature factors, there was a significant cooling trend during this period that hindered its growth. As a result, notable degradation trends were observed for surface vegetation within this area.

4. Discussion

4.1. Trends of Vegetation on the TP

In this study, based on the NOAA NDVI dataset, we observed a significant regional variation in the overall increasing trend of growing season NDVI on the TP. These findings further support the conclusion that there has been an overall increase in vegetation on the TP, except for certain localized areas that have experienced degradation [55,56,57]. However, variations exist in both the growth rate and spatial extent of vegetation in the TP. Our results indicate that the growing season for NDVI on the TP exhibits a climate tendency rate of 0.01·decade⁻¹, which is consistent with previous studies by Li et al. [42] and Chen et al. [51], albeit slightly lower than Liu’s [23] findings. We found that approximately 77% of the analyzed areas exhibited improved vegetation during 1982–2018, primarily concentrated in central and southern TP. Conversely, around 23% of areas showed a declining trend in vegetation, predominantly observed in northern TP, including the Kunlun Mountains and Qaidam Basin. This closely corresponds to the results of Chen et al.’s study for the period between 2000 and 2019 [51] but exceeds those reported by Liu et al. [23] and Zhang et al. [1] for the period between 1981 and 2000. These discrepancies can mainly be attributed to differences in study periods; however, variations in NDVI data sources may also contribute to some extent.

4.2. Abrupt Change in Climate and Warming and Humidification in the TP

We observed abrupt changes in vegetation and climatic factors on the TP around 1998, which are consistent with Easterling and Wehner’s [58] proposed slowdown of global warming. Additionally, numerous studies have also confirmed that all temperature-related variables over the TP, such as wind speed, sunshine duration, vapor pressure, and precipitation, reversed their trends in 1998 [35,59]. A debate exists regarding whether there has been a deceleration or acceleration of temperature increase on the TP since 1998 [35,59,60,61,62]. Our findings indicate that post-1998 warming slowed down in northern TP while experiencing an acceleration in southern TP (south of 35°N). These differing conclusions about temperature trends on the TP can be attributed to regional variations.

4.3. Response of Surface Vegetation to Climate Factors across Different Climatic Backgrounds

We observed significant variations in how surface vegetation responds to climate factors across different climatic backgrounds. Before 1998, precipitation on the TP was relatively lower. During this period, changes in surface vegetation within the humid and semi-humid regions in the southeast of the plateau exhibited a positive correlation with air temperature; however, changes in surface vegetation within the semi-arid areas located in northeastern TP showed a negative correlation with air temperature. These findings align with the previous studies conducted by Zhou et al. [63] and Chen et al. [51]. Therefore, it is clear that water limits how vegetation responds to changes in air temperature on the TP. Since 1998, there has been a significant increase (as reported by reference [36]) in precipitation, which has resulted in improved moisture conditions. During this period, surface vegetation in the semi-arid regions of the northeast and southwest of the plateau showed positive correlations with temperature and precipitation, while humid and semi-humid surface vegetation in the southeast exhibited a significant negative correlation with precipitation. These findings are consistent with those reported by Liu et al. [23] and Zhang et al. [17]. Precipitation plays a crucial role in vegetation growth within the northeastern and southwestern parts of TP [42,51]. The increased precipitation after 1998 led to a shift in the response of surface vegetation to air temperature from a negative correlation before 1998 to a positive correlation thereafter within this region. In contrast, summer rainfall is abundant in southeastern TP; however, an increase in precipitation often coincides with cloud cover that reduces incident solar radiation [64], thereby impeding vegetation growth. Nevertheless, there has been a decrease in precipitation since 1998, and the warming and drying of the climate significantly stimulated surface vegetation activities within this area.

The results also indicate that the variation in surface vegetation within the high-altitude arid region of the TP is primarily influenced by thermal factors (air temperature). After 1998, a significant cooling trend was observed in the northwest area of the TP, which could potentially account for the degradation of surface vegetation in the Kunlun Mountains. This finding is inconsistent with the conclusions drawn by Chen et al. [51], who employed Pearson correlation analysis and identified a negative relationship between vegetation and air temperature in the northern part of TP. However, our study employed partial correlation analysis to exclude precipitation effects when examining changes in vegetation response to temperature. It is plausible that these divergent analytical approaches contribute to discrepancies between our findings and those reported by Chen et al. [51].

Our research demonstrates that the warming and humidification of the TP’s climate play a crucial role in significantly enhancing surface vegetation. Previous studies have indicated that climate warming has resulted in permafrost degradation [13] and increased soil moisture content [65], thus influencing surface energy distribution and strengthening land–atmosphere interactions [66]. Regarding increased soil moisture, there is greater evapotranspiration and reduced sensible heat flux, leading to cloud formation and precipitation [67], ultimately impacting vegetation growth [68]. In addition, Sun et al. suggested that the amplification of climate warming and humidification on the TP is linked to weakened westerly winds over the plateau and intensified southwest monsoon activity [36]. Further investigation into the synergistic effects of atmospheric circulation patterns, such as westerly winds and monsoons, on vegetation dynamics in this region is necessary.

4.4. Possible Effects of Unnatural Factors on Vegetation of the TP

The growth of vegetation is influenced by numerous factors, encompassing both natural and anthropogenic influences. Within the realm of natural factors, climate plays an indispensable role [33], as it determines the composition and distribution patterns of surface vegetation. While precipitation and temperature are paramount hydrothermal conditions affecting plant growth [21], other climatic factors such as solar radiation, relative humidity, and surface wind speed should not be disregarded [69,70,71]. Therefore, further investigation is needed to explore the synergistic effects of various climatic factors on the surface vegetation of the TP. Additionally, human activities significantly influence vegetation growth. For example, deforestation, overgrazing, and urban expansion can lead to a decline in vegetation cover, while rational land use practices can enhance vegetation coverage [71,72]. Towards the end of the 20th century, the Chinese government implemented policies, such as converting farmland to forest and grassland and the establishment of ecological and environmental protection projects to safeguard the ecological environment. These initiatives have played a constructive role in restoring and enhancing vegetation on the TP. Furthermore, regional variations in vegetation response to climate change are influenced by unique geographical conditions and soil characteristics. Notably, this research reveals evident degradation trends in surface vegetation within arid regions such as the Kunlun Mountains and Qaidam Basin; however, their underlying causes differ significantly. The higher altitude of the Kunlun Mountains coupled with lower temperatures primarily influences surface vegetation growth through changes in thermal factors (temperature). The decrease in near-surface air temperature since 1998 has been unfavorable for vegetative development in this area, resulting in degradation trends. Conversely, the lower altitude of the Qaidam Basin experiences limited precipitation alongside extensive bare surfaces with high evaporation rates. Under water-stressed conditions exacerbated by warming temperatures, surface vegetation growth is impeded within this region. Additionally, abundant salt lakes dominate this area along with saline–alkali soils that differ from other regions, hence leading to distinct responses of surface vegetation towards climate change.

5. Conclusions

The NDVI values on the TP exhibit a gradual decrease from southeast to northwest during the growing season. There is a significant overall increasing trend at a climate tendency rate of 0.01·decade (p < 0.01) from 1982 to 2018, characterized by a notable mutation of around 1998. Over the past 37 years (1982–2018), a polarized trend of vegetation has been observed on the TP, with notable improvement in central and eastern parts of the TP but with substantial degradation also evident in northwestern TP, particularly within the Kunlun Mountains and Qaidam Basin.

The climate of the TP exhibits significant regional disparities regarding warming and humidification characteristics before and after 1998. In period P1 (1982–1998), there is a predominant temperature increase in concentration in the northern region, while in period P2 (1998–2018), it shifts to become concentrated in the southern region. Additionally, an increase in precipitation is mainly observed in the southern and northwestern regions during period P1 (1982–1998); however, during period P2 (1998–2017), it is primarily seen in the northeastern and western regions.

The response of surface vegetation to climate factors varied significantly across diverse climatic backgrounds. It is noteworthy that moisture conditions have a substantial influence on the response of vegetation to air temperature. During P1, when moisture conditions were relatively insufficient, surface vegetation in the humid and semi-humid regions of southeastern TP showed a positive correlation with air temperature, while surface vegetation in the semi-arid regions of northeastern TP displayed a negative correlation with air temperature. During P2, as water conditions became relatively sufficient, surface vegetation in the semi-arid regions showed positive correlations with both temperature and precipitation. However, surface vegetation in the humid and semi-humid regions exhibited a significant negative correlation with precipitation. During this period, synergistic effects between warm and humid climates in the semi-arid regions of northeastern TP and warm and dry climates in humid and semi-humid regions of southeastern TP substantially enhanced surface vegetation on the TP. Furthermore, our results indicate that thermal factors (air temperature) primarily influence variations in surface vegetation within the high-altitude arid region of the TP. After 1998, a significant cooling trend was observed in the northwestern TP, which could potentially account for the degradation of surface vegetation in the Kunlun Mountains.