Relationships between Precipitation and Elevation in the Southeastern Tibetan Plateau during the Active Phase of the Indian Monsoon

Abstract

The precipitation gradient (PG) is a crucial parameter for watershed hydrological models. Analysis of daily precipitation and elevation data from 30 stations in the southeastern Tibetan Plateau (SETP) during the active phase of the Indian monsoon reveals distinct patterns. Below 3000 m, precipitation generally decreases with increasing altitude. Between 3000 and 4000 m, precipitation patterns are more complex; in western regions, precipitation increases with elevation, whereas in eastern regions, it decreases. Above 4000 m, up to the highest observation point of 4841 m, precipitation continues to decrease with elevation, with a more pronounced decline beyond a critical height. In the SETP, PGs for LYR and NYR are positive, at 11.3 ± 2.7 mm/100 m and 17.3 ± 3.8 mm/100 m, respectively. Conversely, PLZB exhibits a negative PG of −22.3 ± 4.2 mm/100 m. The Yarlung Zangbo River (YLZBR) water vapor channel plays a significant role in these PGs, with the direction and flux of water vapor potentially influencing both the direction and magnitude of the PG. Additional factors such as precipitation intensity, the number of precipitation days, precipitation frequency, and station selection also significantly impact the PG. Notable correlations between elevation and variables such as the number of precipitation days, non-precipitation days, and precipitation intensity. The precipitation intensity gradients (PIGs) are 0.06 ± 0.02 mm/d/100 m, 0.11 ± 0.04 mm/d/100 m, and −0.18 ± 0.04 mm/d/100 m for the three catchments, respectively. Future research should incorporate remote sensing data and expand site networks, particularly in regions above 5000 m, to enhance the accuracy of precipitation–elevation relationship assessments, providing more reliable data for water resource simulation and disaster warning.

1. Introduction

The precipitation gradient (PG), which describes the relationship between precipitation and elevation, is a crucial parameter for understanding and simulating spatial precipitation distribution in mountainous areas [1,2,3,4]. Accurate PGs and vertical temperature lapse rates significantly improve hydrological simulations, especially in cold regions [5,6,7]. Studying PGs provides an important scientific basis for understanding the water cycle, managing water resources, developing climate change adaptation strategies, tracking ecosystem succession, and assessing disaster risks [8,9,10,11,12,13,14].

However, the relationship between precipitation and elevation is far from uniform across the globe, and understanding its variability remains a key scientific challenge. For instance, in the Luquillo Mountains of northeastern Puerto Rico, studies have identified a strong positive correlation between precipitation and elevation, particularly during the rainy season [15,16,17]. Similar trends are evident in the arid regions of northwest China, Morocco, and Tunisia [3,18,19], where precipitation generally increases steadily with elevation. In contrast, the Andes Mountains present a different scenario, in which precipitation initially increases with elevation but decreases beyond a certain point, a phenomenon known as a reversed PG [20,21,22]. This reversal is influenced by terrain-induced physical barriers, variations in atmospheric circulation patterns, and differing moisture sources, resulting in significant regional differences in the elevation at which maximum precipitation occurs [8,20,21,22,23].

The Tibetan Plateau (TP), often referred to as the “Third Pole”, offers another unique setting for studying PGs [24,25,26]. Research on the TP and its surrounding areas has highlighted distinct regional variations in PGs [18,27]. Sun et al. (2020) suggest that atmospheric circulation patterns are the primary drivers of PGs on the TP, with positive PGs in regions influenced by westerlies and reversed PGs in areas affected by the Indian monsoon [26]. In the Himalayan region of the southern TP, during the active Indian monsoon phase, precipitation shows complex changes with elevation, featuring two distinct elevation bands and multiple shifts in the PG [28,29]. Outside the monsoon phase, an overall positive PG is observed, although this trend may diminish or even reverse at elevations above 4000 m [24]. Other studies have shown that PGs are mainly shaped by local atmospheric dynamics, precipitation rates, lifting condensation levels (LCLs), and convective available potential energy [24,25,29,30,31].

Although previous studies have established that the Indian monsoon significantly affects the PG [25,28,29], research on the relationship between monsoonal precipitation and elevation gradient in the Southeastern Tibetan Plateau (SETP), which is strongly influenced by the Indian monsoon, remains limited. Due to inadequate coverage of observation sites, obtaining reliable PG estimates in this region has been challenging, leading to divergent conclusions in earlier studies [25,32,33]. For instance, Guo et al. (2016) grouped stations in the SETP with those in the Hengduan Mountains, estimating a positive PG of 16.8 mm/100 m [32]. Conversely, using the same dataset, negative PGs were found when fitting data from SETP stations and those within the Yarlung Zangbo River (YLZBR) catchment [33].

Therefore, this study leverages precipitation observation data from 30 relatively dense stations in the SETP to accurately characterize the relationship between precipitation and elevation during the active phase of the Indian monsoon. It compares and analyzes the differences in PGs across three river catchments in the SETP and their influencing factors. These findings provide fundamental data for improving the accuracy of hydrological models and understanding the mechanisms behind regional geological hazards. Furthermore, this study aims to offer theoretical support for assessing the impacts of future climate change and enhance understanding of the water cycle under the complex terrain conditions of this region.

2. Materials and Methods

2.1. Study Area

The study area is located in the southeastern part of the TP, predominantly within Nyingchi City in the Tibet Autonomous Region. It lies at the junction of the Nyainqêntanglha Mountains and the Himalayas and is characterized by significant topographic variation. The highest peak, Namcha Barwa, reaches 7782 m, while the lowest points are just a few hundred meters above sea level. The SETP contains the largest distribution of maritime glaciers in China, but it is experiencing severe loss [34,35,36]. The Layue River (LYR), Niyang River (NYR), and Palong Zangbo (PLZB) all belong to the YLZBR system (Figure 1). This study area is directly exposed to the water vapor channel of the YLZBR, with a climate predominantly influenced by the Indian monsoon and the westerlies [37,38]. The region exhibits a significant vertical climate gradient across its mountains. From June to September, the Indian monsoon carries large amounts of moisture from the Bay of Bengal through the YLZBR valley into the central and eastern parts of the plateau, forming the well-known “warm tongue” and “wet tongue” on the plateau [39,40,41,42]. Mon-soon precipitation not only alters the distribution pattern of precipitation at different elevations within the region but also concentrates during the monsoon phase, with more than 70% of the annual precipitation occurring from June to September [37,43]. According to records from the Nyingchi Station of the National Meteorological Bureau over the past 60 years, the average annual temperature in Nyingchi is 8.9 °C, with an average annual precipitation of 680.8 mm.

2.2. Data

The data used in this study consist of quality-controlled daily precipitation records obtained from automatic weather stations (AWSs) set up at 30 different sites. Out of 30, 2 AWSs were provided by the Comprehensive Observation and Research Station of the Alpine Environment in Southeastern Tibet under the Chinese Academy of Sciences, and 28 AWSs were provided by the Meteorological Bureau of Nyingchi City, Tibet (Figure 1). These stations have all been continuously observed for more than five years during the active phase of the Indian monsoon. The altitudinal distribution of the stations in the LYR, NYR, and PLZB watersheds is shown in Figure 2. The LYR catchment has 7 stations with elevation ranging from 2438 m to 4553 m and data spanning from 2013 to 2018. The NYR catchment contains 10 stations with elevations ranging from 2980 m to 4841 m, with data also covering the 2013 to 2018 period. There is a shared station, Shanding, located on the boundary between NYR and LYR at an elevation of 4553 m. The PLZB catchment has 14 stations. All stations, except for Tongmai, Bomi, Midui, and Ranwu, were established in October 2013. To maintain data consistency across the catchment, data from all stations are used for the period from 2014 to 2018.

2.3. Methods

The PG, the number of precipitation days gradient (PDG), and the precipitation intensity gradient (PIG) in this study can be estimated according to the following formulas [32,43]:

$$Y_i=\beta X_i+C_0$$

(1)

where

Y represents precipitation P and X represents elevation. The equation thus represents the regression equation between the precipitation P at station i and the corresponding elevation X at station i.

β represents the slope, that is, the PG (mm/100 m).

$C_0$ represents the precipitation when the elevation is 0 m (sea level) (the intercept of the ordinate).

When Y represents the number of precipitation days and X represents the elevation, the β slope in Equation (1) is the precipitation day number gradient PDG (d/100 m).

When Y represents the precipitation intensity and X represents the elevation, the β slope in Equation (1) is the PIG (mm/d/100 m).

The PIG in this study differs from the daily PG defined by previous researchers, who calculated it by dividing the PG by the number of precipitation days at a single station (or by the average number of precipitation days at catchment stations) [7,30] (Zeng et al., 2020; Wang et al., 2018). In contrast, the PIG in this study is derived by linearly fitting the precipitation at each station divided by the number of precipitation days at that station against the corresponding elevation. In the regression analysis, this study uses the significance level P to determine correlation. A p-value of less than or equal to 0.05 indicates a significant correlation at the 95% confidence interval.

To enhance the comparability of PG values, this study also calculates the relative precipitation gradient (RPG) (%/100 m) using the following formula [26]:

$$RPG={\displaystyle \frac{PG}{\overline{P}}}$$

(2)

where $\overline{P}$ is the average precipitation across the stations with daily precipitation greater than 0.1 mm. Referring to the classification standard of the China Meteorological Administration, precipitation intensity is categorized as light rain (P ≤ 10 mm/d), moderate rain (10 mm/d < P ≤ 25 mm/d), and heavy rain (P > 25 mm/d).

Additionally, the empirical frequency of precipitation is determined by dividing the rank of precipitation (arranged from largest to smallest) by the total number of samples plus one.

3. Results

3.1. Precipitation in the SETP

The average precipitation during the active phase of the Indian monsoon from 2013 to 2018, observed at 30 stations in the SETP, ranged from 202.3 mm to 790.3 mm, with an overall average of 482.3 mm (Figure 3). The highest precipitation was recorded at Mila Station, which is located at the highest elevation, with a multi-year average of 790.3 mm. At this station, light rain accounted for a relatively small proportion, while moderate rain contributed to more than half of the total precipitation during the active monsoon period. In contrast, the lowest precipitation was observed at Ranwu Lake station, situated at an elevation of 3915 m, with a multi-year average of just 202.3 mm. At Ranwu Lake station, light rain constituted 57.7% of the total precipitation during the active monsoon period, despite its elevation ranking sixth among all the stations. Interestingly, the Tongmai station, which is at the lowest elevation, received a substantial amount of precipitation, totaling 731.7 mm, making it the third highest among the 30 stations (Figure 3).

From 2013 to 2018, the average monsoon precipitation at the seven stations in the LYR catchment ranged between 493.3 mm and 790 mm, with a mean of 621.7 ± 95.1 mm, the highest among the three catchments (Figure 4). In the NYR catchment, the average monsoon precipitation at the 11 stations for the same period was 571.3 ± 131.7 mm. The PLZB catchment, with data from 14 stations recorded from 2014 to 2018, had the lowest average precipitation of the three catchments, with a mean of 368.4 ± 129.7 mm. PLZB also exhibited the lowest average precipitation across all precipitation intensities. Among the three catchments, LYR had the highest average precipitation for light and heavy rain events, followed by NYR. However, for moderate rain intensity, NYR slightly exceeded LYR in average precipitation.

3.2. Precipitation Gradient (PG) and Relative Precipitation Gradient (RPG)

During the active phase of the Indian monsoon, precipitation across all stations in southeastern Tibet shows a weak positive correlation with elevation, with a PG of 9.2 ± 4.3 mm/100 m (R² = 0.14, p ≤ 0.05). Among different precipitation categories, only moderate rain shows a weak positive correlation with elevation (R² = 0.20, p ≤ 0.05). The relationships between monsoon precipitation and elevation during the active monsoon period for the LYR, NYR, and PLZB catchments are depicted in Figure 5 and

Table 1. The PG and RPG at different rain levels for the three study catchments during the active phase of the Indian monsoon.

Catchments	Grade (mm/d)	Rain Day (d)	Amount (mm)	PG (mm/100 m)	RPG (%/100 m)	R2	p
LYR	≤9.9	79.8	244.5	2.1 ± 0.6	0.9 ± 0.2	0.69	*
	10.0–24.9	15.9	255.1	5.1 ± 2.2	2.0 ± 0.9	0.52
	≥25	4	127.4	4.2 ± 1.7	3.3 ± 1.3	0.54
	Nongraded	99.7	627.1	11.3 ± 2.7	1.8 ± 0.4	0.78	**
NYR	≤9.9	73.3	210.6	2.9 ± 1.1	1.4 ± 0.5	0.46	*
	10.0–24.9	17.3	269.1	10.6 ± 2.5	3.9 ± 0.9	0.70	**
	≥25	2.9	91.6	3.8 ± 1.9	4.1 ± 2.1	0.32
	Nongraded	93.5	571.3	17.3 ± 3.8	3.0 ± 0.7	0.73	**
PLZB	≤9.9	73.7	182.3	−6.0 ± 1.5	−3.3 ± 0.8	0.56	**
	10.0–24.9	8.7	132.6	−10.3 ± 2.2	−7.7 ± 1.7	0.68	**
	≥25	1.6	53.4	−6.1 ± 1.6	−11.4 ± 3	0.55	**
	Nongraded	84	368.4	−22.3 ± 4.2	−6.1 ± 1.1	0.70	**

Note(s): * and ** indicate significance smaller than 0.05 and 0.01, respectively (significant coefficient).

.

When considering total precipitation, nongraded by intensity, precipitation in LYR and NYR showed a highly significant positive correlation with elevation, with PGs of 11.3 ± 2.7 mm/100 m (R² = 0.78, p ≤ 0.01) and 17.3 ± 3.8 mm/100 m (R² = 0.73, p ≤ 0.01), respectively. In contrast, the PLZB catchment exhibits a highly significant negative correlation, with a PG of −22.3 ± 4.2 mm/100 m (R² = 0.70, p ≤ 0.01) (Figure 5a).

For different precipitation intensities, the PLZB catchment consistently shows a highly significant negative correlation between precipitation and elevation. The PGs for light, moderate, and heavy rain are −6.0 ± 1.5 mm/100 m (R² = 0.56, p ≤ 0.01), −10.3 ± 2.2 mm/100 m (R² = 0.68, p ≤ 0.01), and −6.1 ± 1.6 mm/100 m (R² = 0.55, p ≤ 0.01), respectively (Figure 5b–d). Conversely, the LYR and NYR catchments show positive PGs. For light rain, the PGs in LYR and NYR are similar, at 2.1 ± 0.6 mm/100 m (R² = 0.69, p ≤ 0.05) and 2.9 ± 1.1 mm/100 m (R² = 0.46, p ≤ 0.05), respectively.

In the case of moderate rain, the LYR catchment shows a non-significant positive correlation between precipitation and elevation, while the NYR catchment demonstrates a highly significant positive PG of 10.6 ± 2.5 mm/100 m (R² = 0.7, p ≤ 0.01) (Figure 5c). For heavy rain, both the LYR and NYR catchments display a non-significant positive correlation with elevation (Figure 5d).

The relative RPG can reduce the heterogeneity of PGs caused by varying amounts of precipitation. The corresponding relative RPG values for LYR, NYR, and PLZB are 1.8%/100 m, 3.0%/100 m, and −6.1%/100 m, respectively (Table 1). For light rain, the relative RPG values for LYR, NYR, and PLZB are 0.9%/100 m, 1.4%/100 m, and −3.3%/100 m, respectively, which are approximately half of the values for the respective catchments when precipitation intensity is not considered. The RPG for different precipitation intensities in each catchment is highest for heavy rain, followed by moderate rain, and lowest for light rain.

3.3. Precipitation Days and Precipitation Day Number Gradient

The average number of precipitation days per station during the monsoon phase in LYR, NYR, and PLZB is 99.7 d, 93.5 d, and 84 d, respectively, accounting for 81.7%, 76.6%, and 68.9% of the total days in the entire monsoon phase (

Table 2. Statistical table of precipitation days, non-precipitation days, and percentage distribution of different rain levels for the three study catchments during the active phase of the Indian monsoon.

Catchments	Light Precipitation		Moderate Precipitation		Heavy Precipitation		Nongraded Precipitation		Non-Precipitation
	d	%	d	%	d	%	d	%	d	%
LYR	79.8	65.4	15.9	13.0	4	3.3	99.7	81.7	22.3	18.3
NYR	73.3	60.1	17.3	14.2	2.9	2.4	93.5	76.6	28.6	23.4
PLZB	73.7	60.4	8.7	7.1	1.6	1.3	84	68.9	38	31.1

). Across all catchments, the number of light rain days (<10 mm) exceeds 70 d, while the number of heavy rain days (>25 mm) is no more than 5 d. The number of light rain days in the LYR, NYR, and PLZB catchments constitutes 80.0%, 78.4%, and 87.7% of their respective precipitation days. In the entire study area, the station with the most precipitation days is at 4553 m, with an average of 108.2 precipitation days from 2013 to 2018. This station accounts for 88.7% of the entire monsoon phase and has the highest number of precipitation days in this study.

In the SETP catchments, the number of precipitation days and non-precipitation days is highly significantly correlated with elevation (Figure 6). The number of rainfall days in LYR and NYR shows a highly significant positive correlation with elevation, with gradients of 0.8 d/100 m (R² = 0.92, p ≤ 0.01) and 1.0 d/100 m (R² = 0.73, p ≤ 0.01), respectively. However, in PLZB, the number of rainfall days has a highly significant negative correlation with elevation, with a gradient of −1.8 d/100 m. The correlation between the number of non-precipitation days and elevation within the same catchment is opposite to that of precipitation days, with gradients similar to those of precipitation days (Figure 6b). The observations in LYR and NYR align with the findings of Garcia-Martino et al. (1996) in the Luquillo Mountains, where the number of non-rainy days decreases with increasing elevation [15].

For different precipitation grades, only the number of light rain days in LYR shows a significant positive correlation with elevation, with a PDG of 0.4 d/100 m (R² = 0.72, p ≤ 0.01). For moderate rain, the number of precipitation days in NYR and PLZB shows a highly significant correlation with elevation, with PDG values of 0.7 d/100 m (R² = 0.74, p ≤ 0.01) and −0.6 d/100 m (R² = 0.7, p ≤ 0.01), respectively. For heavy rain, only the number of precipitation days in PLZB shows a highly significant negative correlation with elevation, with a gradient of −0.2 d/100 m (R² = 0.56, p ≤ 0.01).

3.4. Precipitation Intensity Gradient (PIG)

There is a significant correlation between the ratio of precipitation amount to the number of precipitation days and elevation in the LYR, NYR, and PLZB during the active phase of the Indian monsoon (Figure 7a). The PIGs estimated using Equation (1) are 0.06 ± 0.02 mm/d/100 m, 0.11 ± 0.04 mm/d/100 m, and −0.18 ± 0.04 mm/d/100 m, respectively. Among the different rain levels in the three catchments, only the precipitation intensity of light and moderate rain in the PLZB catchment shows a statistically significant correlation with elevation, with PIGs of −0.06 ± 0.01 mm/d/100 m and −0.08 ± 0.03 mm/d/100 m, respectively (Figure 7b,c). For the light rain level, the precipitation intensities in LYR and NYR show an insignificant positive correlation with elevation (Figure 7b). In the moderate rain level, the correlation is weak (Figure 7c). For the heavy rain level in all three river catchments, the correlation between precipitation intensity and elevation is minimal (Figure 7d).

4. Discussion

4.1. Impact of Precipitation Level and Frequency on PGs

The SETP region, influenced by the Indian monsoon, experiences relatively humid conditions. Due to radiative cooling or orographic uplift, light rain constitutes the majority of precipitation in the valleys [44]. Precipitation intensity and frequency significantly impact the direction and magnitude of the PG [29,30]. Distribution maps of precipitation and cumulative frequency at each station across the LYR, NYR, and PLZB catchments show a consistent trend. At most stations, the cumulative probability of precipitation below 5 mm exceeds 60%, and the cumulative frequency below 10 mm exceeds 80% (Figure 8). In PLZB, precipitation is predominantly light rain. When precipitation reaches 10 mm, the cumulative probability at the 2088 mm station is about 80%, while at the other 13 stations, it reaches 90% (Figure 8c). The cumulative probability of precipitation at 10 mm in PLZB is significantly higher than in LYR and NYR, indicating that light rain is more concentrated and represents a larger proportion in PLZB.

When considering an 80% cumulative precipitation frequency as a threshold, precipitation at lower elevations in LYR and NYR is less than 10 mm, whereas at higher elevations, it exceeds 10 mm. This suggests that heavy rain is more prevalent at higher elevation stations, contributing to a positive elevation gradient in LYR and NYR. Conversely, in PLZB, where the cumulative frequency of precipitation at 2088 mm is about 10 mm, light rain is more prevalent at higher elevations, contributing to a negative PG [29].

Taking 10 mm of precipitation as an example, the difference in cumulative precipitation frequency between high and low elevations in NYR is greater than in LYR (Figure 8a,b). This implies a larger difference in precipitation between high and low elevations in NYR. Additionally, the elevation difference between high and low elevations in NYR (1861 m) is smaller than in LYR (2155 m), resulting in a steeper elevation gradient in NYR compared to LYR (Figure 5a).

In this study, the PGs at different precipitation levels in LYR, NYR, and PLZB exhibit consistent patterns. Among light rain, moderate rain, and heavy rain, light rain has the smallest PG, followed by heavy rain, with moderate rain having the largest. However, previous studies on the Zhuodi River on the northern slope of the Nyainqêntanglha Mountains and the Karuxung and Mabengnong Rivers on the northern slope of the Himalayas differ from these findings [30]. This discrepancy may be due to the proximity of the regions in this study, which are uniformly influenced by the Indian monsoon. In contrast, the locations of the catchments in the previous studies are more geographically diverse, with varying climate characteristics and precipitation mechanisms [30].

4.2. Comparison of the PG with PIG and RPG

The PG, RPG, PDG, and PIG for LYR, NYR, and PLZB exhibit common characteristics. Both LYR and NYR have positive gradients, with LYR’s gradient being smaller than NYR’s. In contrast, PLZB has a negative gradient, with the largest absolute value among the three. The gradient for PLZB, at −22 mm/100 m, is steeper than those observed in other regions: −8.0 mm/100 m in the Karakoram region in summer [45], −2.4 mm/100 m in the Northwest Himalayas from May to October [31], −19 mm/100 m during the rainy season (May to September) in the TP’s interior [25], and −3.1 mm/100 m before the Indian monsoon in the SETP (including some stations in Yunnan and Sichuan province) [32]. This indicates that precipitation in the PLZB catchment decreases more sharply with increasing elevation. The spatial heterogeneity of PGs is likely due to varying precipitation amounts across different areas [25,26].

The consistency between precipitation amount and the number of precipitation days with elevation observed in this study aligns with previous research [24,46]. Compared to the PG, RPG and PIG may better smooth out spatial heterogeneity, enhancing contrast between different areas. Jiang et al. (2022) estimated the relative PG of the entire TP in summer to be 4.2%/100 m. In this context, PLZB’s PG at 2.87%/100 m is similar to the results of this study for LYR’s 1.8%/100 m and NYR’s 3.0%/100 m [27]. However, these values are slightly lower than the 4.0%/100 m found in the Langtang Valley on the southern slope of the Himalayas, where monsoon precipitation is higher, and the station-specific precipitation rate is significantly greater than that in the LYR and NYR [5].

4.3. The Impact of the YLZBR Water Vapor Channel on the Variation of Precipitation with Elevation

The study of PGs reveals that the variation of precipitation with elevation during the active phase of the Indian monsoon in the SETP is complex. The three catchments in this study all belong to the YLZBR yet show contrasting trends in precipitation with elevation. The YLZBR Grand Canyon serves as a key channel for the Indian monsoon entering the TP, playing a decisive role in shaping precipitation patterns [39,42]. As the Indian monsoon passed through and enters the TP, the cumulative precipitation at Xirang Station, located at an elevation of 511 m at the channel’s entrance, recorded over 8000 mm of precipitation from November 2018 to October 2019, making it the station with the highest known annual precipitation in China [47]. The water vapor carried by the Indian monsoon rapidly decreases after entering the YLZBR Grand Canyon due to topographical barriers [48]. Figure 3 shows that from Tongmai Station to Yuxu Station (at elevations below 3000 m), the average monsoon season precipitation decreases gradually from 731.7 mm to 413.1 mm, confirming this observation. As the moisture bypasses the great bend of the YLZBR and is transported west or northwest, precipitation gradually increases again due to orographic lifting [39,47].

In the LYR catchment, the average monsoon precipitation increases from 493.3 mm at the lowest elevation station, Pailong, to 790.3 mm at Shanding. According to the correlation matrix of daily precipitation for various stations in the SETP (R²), the LYR stations, marked by the blue box in the figure, are arranged from low to high elevation, showing a strong correlation among all stations, indicating that precipitation is influenced by a uniform air mass (Figure 9). Additionally, there is a close relationship between the LYR and the NYH stations, from S113 to Linzhi, suggesting that they are likely influenced by the same moisture source. This also indicates that the moisture mass passes over Shanding Station, and the maximum precipitation altitude in LYR could exceed 4553 m, which is consistent with previous research findings [49].

In the NYR catchment, precipitation increases from 442.1 mm at the lowest elevation Mirui Station to 788.3 mm at the highest elevation Mila Station. However, as shown in Figure 9, the stations in the eastern part of the catchment, from S114 to Linzhi (marked by the purple box), may be affected by different air masses compared to the stations in the western part, from Gongbu to Mila. The western stations, Gongbu, Jinda, Songduo, and Mila, show a significant increasing trend in precipitation, with values of 514.2 mm, 540.2 mm, 676.1 mm, and 788.3 mm, respectively, suggesting that the maximum precipitation altitude exceeds 4841 m.

The low-altitude Tongmai station in the western part of PLZB is directly aligned with the YLZBR water vapor channel, resulting in high precipitation during the active monsoon period. While PLZB is also considered an important water vapor pathway in southeastern Tibet, its vapor transport flux is limited due to its direction being opposite to the main flow of water vapor [40,47,50]. As one moves eastward across PLZB, the influence of water vapor from the YLZBR decreases. Figure 9 illustrates a clear difference in water vapor sources between the high-altitude stations in the east, such as Songzong, Midui, and Ranwu, and the low-altitude stations in the west. Moreover, Gangrigabu Mountain, which exceeds 6000 m in height (Figure 2), acts as a barrier on the southern side of PLZB, resulting in a monsoon precipitation of only 202.3 mm at Ranwu station in the leeward valley [47,51,52]. Consequently, the precipitation gradient in PLZB is negative.

Thus, the pattern of precipitation variation with elevation in the SETP can be summarized as follows: at elevations below approximately 3000 m, precipitation generally decreases with increasing elevation. Between 3000 m and 4000 m, the relationship between precipitation and elevation becomes more complex and varies across different catchments [47,51,52]. In the western regions, such as LYQ and NYH, precipitation generally increases with elevation, while in the eastern region, such as PLZB, the opposite trend is observed (Figure 5). At elevations between 4000 m and the highest observed station at 4841 m, precipitation increases with elevation. The observations from LYR and NYR confirm that the maximum precipitation altitude in southeastern Tibet exceeds 4841 m. This finding differs from the study by Lan and Zhang (2017), which suggested that precipitation decreases with elevation above 4000 m due to the rise in the lifting condensation level (LCL) and a decline in total column water vapor (TCWV) [24]. This discrepancy may be due to the abundant moisture brought by the Indian monsoon in southeastern Tibet, providing a rich source of precipitation. At high elevations, strong vertical air movements and decreasing temperatures reduce the saturation level of water vapor, making it easier for condensation to form clouds and rain [25,26,51,52,53]. However, precipitation cannot increase indefinitely with elevation beyond a certain height; as the LCL rises and TCWV decreases, precipitation will inevitably decrease with further increases in elevation.

4.4. Uncertainty of the PG in the SETP

The PG is a crucial input parameter for catchment hydrological models [5,6]. The precipitation stations in this study are located in the middle and lower reaches of the YLZBR, a key area for hydrological and water resource modeling and prediction in the region [54,55,56]. Due to the limited availability of station data, most studies have likely relied on just two national meteorological stations in the SETP, Linzhi and Bomi, both situated at elevations below 3000 m [27,32,33,57]. The scarcity of data poses significant challenges to achieving accurate model simulations.

This study reveals that the pattern of precipitation with elevation in the SETP shows a high degree of regional heterogeneity in PGs. Consequently, using limited data to derive PGs or applying a uniform PG in models may lead to significant biases in the results. For example, precipitation between stations from Tongmai to Yuxu in the PLZB (at elevations below 3000 m) is strongly correlated (Figure 9, marked by the green box), indicating the influence of a common water vapor source. The PG in this region is −42.5 ± 9.2 mm/100 m. However, when more stations in the PLZB are included, the estimated PG becomes 22.3 ± 4.2 mm/100 m (Table 1), a nearly twofold difference. Similarly, Tang et al. (2023) found significant spatial heterogeneity in PGs within the Yamzho Yumco Catchment in southern Tibet, with positive PGs in the north and negative PGs in the south [43].

Moreover, using only the Linzhi and Bomi stations in the SETP to fit PGs for different nearby regions often yields varying results, with some PGs being positive and others negative [25,32,33]. Therefore, to accurately estimate PGs in a catchment, it is crucial to consider the influence of station selection and incorporate data from as many precipitation stations as possible to ensure reliable results.

5. Conclusions

Based on observations from 30 precipitation stations in the SETP during the active phase of the Indian monsoon, a weak positive correlation was identified between precipitation and elevation. However, the relationship between precipitation and elevation in this region is not straightforward. It does not consistently increase with altitude. The pattern of precipitation variation with elevation can be summarized as follows: at elevations below 3000 m, precipitation generally decreases with increasing altitude. Between 3000 m and 4000 m, the pattern becomes more complex. In the western region, influenced by the predominant direction of the YLZBR water vapor channel, precipitation tends to increase with elevation. In contrast, in the eastern region, precipitation decreases as elevation rises. Above 4000 m, up to the highest observation point at 4841 m, precipitation again decreases with increasing elevation. Beyond a certain critical elevation, the precipitation rate decreases further with any additional increase in altitude.

During the active monsoon period, the relationship between precipitation and elevation shows significant spatial heterogeneity across the three catchments, LYR, NYR, and PLZB. In the LYR and NYR catchments, precipitation generally increases with elevation, while in the PLZB catchment, it decreases. The PGs for LYR and NYR were 11.3 ± 2.7 mm/100 m and 17.3 ± 3.8 mm/100 m, respectively, while the PG for PLZB was negative, at −22.3 ± 4.2 mm/100 m. This study found that, in addition to the influence of the YLZBR water vapor channel, factors such as precipitation intensity, precipitation days, and precipitation frequency were identified as key factors influencing the PG in the SETP.

For light rain level, significant correlations with elevation were observed across all three catchments, with PGs of 2.1 ± 0.6 mm/100 m for LYR and 2.9 ± 1.1 mm/100 m for NYR, respectively, while the PG for PLZB was negative, at −6.0 ± 1.5 mm/100 m. The number of precipitation and non-precipitation days also correlated significantly with elevation across the three catchments. The PIG values for LYR, NYR, and PLZB were 0.06 ± 0.02 mm/d/100 m, 0.11 ± 0.04 mm/d/100 m, and −0.18 ± 0.04 mm/d/100 m, respectively. However, the accuracy of precipitation gradient estimates can be significantly influenced by the selection of observation stations.

Future efforts should focus on expanding the network of precipitation stations, particularly at elevations above 5000 m, to better assess the influence of various meteorological factors on the variation of precipitation with elevation in the SETP over different periods. Additionally, integrating data from remote sensing satellites, micro-rain radars, and other advanced technologies will allow for a more precise evaluation of the relationship between regional precipitation and elevation. This, in turn, will provide more reliable foundational data for studies on water resource modeling and disaster forecasting.