The Dynamics of Vegetation Evapotranspiration and Its Response to Surface Meteorological Factors in the Altay Mountains, Northwest China

Abstract

The Altay Mountains’ forests are vital to Xinjiang’s terrestrial ecosystem, especially water regulation and conservation. This study evaluates vegetation evapotranspiration (ET) from 2000 to 2017 using temperature, precipitation, and ET data from the China Meteorological Data Sharing Service. The dataset underwent quality control and was interpolated using the inverse distance weighted (IDW) method. Correlation analysis and climate trend methodologies were applied to assess the impacts of temperature, precipitation, drought, and extreme weather events on ET. The results indicate that air temperature had a minimal effect on ET, with 68.34% of the region showing weak correlations (coefficients between −0.2 and 0.2). Conversely, precipitation exhibited a strong positive correlation with ET across 98.91% of the area. Drought analysis, using the standardized precipitation evapotranspiration index (SPEI) and the Temperature Vegetation Dryness Index (TVDI), showed that ET was significantly correlated with the SPEI in 96.47% of the region, while the TVDI displayed both positive and negative correlations. Extreme weather events also significantly influenced ET, with reductions in the Simple Daily Intensity Index (SDII), heavy precipitation days (R95p, R10), and increases in indicators like growing season length (GSL) and warm spell duration index (WSDI) leading to variations in ET. Based on the correlation coefficients and their significance, it was confirmed that the SII (precipitation intensity) and R95p (heavy precipitation) are the main factors causing vegetation ET increases. These findings offer crucial insights into the interactions between meteorological variables and ET, essential information for sustainable forest management, by highlighting the importance of optimizing water regulation strategies, such as adjusting species composition and forest density to enhance resilience against drought and extreme weather, thereby ensuring long-term forest health and productivity in response to climate change.

1. Introduction

Evapotranspiration (ET) is the process by which water is transferred from soil and open water surfaces to the atmosphere. This transfer is influenced by interactions between land surfaces and atmospheric conditions [1,2]. This process significantly contributes to monitoring water usage within agricultural ecosystems and plays a vital role in ecological conservation efforts [3,4]. The accurate estimation of ET and its driving factors enhances our understanding of ecosystems and water resources.

Climate change is a predominant factor influencing forest growth, necessitating comprehensive examinations of its impacts on forest ecosystems. Such an understanding is crucial for rehabilitating degraded lands and fostering sustainable forestry development [5,6]. In addition to meteorological factors including temperature and precipitation, other factors also affect the distribution of forests and other species. [7,8,9]. The Altay Mountains’ forest resources are pivotal for water conservation and regulation within Xinjiang’s terrestrial ecosystem [10,11,12]. Understanding the interplay between climate change and ET in the Altay Mountains is vital for enhancing forest land quality and productivity and promoting sustainable forest development [13,14]. Given the intimate link between forests and climate, climate variations directly impact forest ecosystems [15,16,17]. The accurate prediction of future climate conditions hinges on precise estimations of vegetation ET, which integrates processes from the planetary water, carbon, and energy cycles [18,19]. Vegetation, by altering how energy is distributed between the canopy and soil, determines evapotranspiration (ET) and its components. As global warming accelerates and human activities intensify, vegetation greens at both the regional and global scales [20].

Research on forest ecosystems in China since the 1990s has largely focused on assessing climate change impacts on these systems. However, much of this research has primarily examined annual average changes in climate indicators rather than delving into the intricate relationships between long-term variations in meteorological factors and vegetation evapotranspiration (ET) [21,22,23].

Using satellite data from 1981 to 1999, Zhou et al. (2001) observed significant increases in vegetation vitality and growth periods [24]. Zhao et al. (2009) [21] conducted a study on temporal and spatial variations in vegetation phenology and their impact on ET in the temperate regions of China using MODIS data. Furthermore, Yu Xiaozhou developed an equation for estimating daytime dark respiration in various tree species by taking into account the characteristics of dark respiration and the intensity of light inhibition in individual leaves [25]. It has been noted that ET and temperature are significantly correlated in Chinese terrestrial ecosystems [26,27] due to climate change.

Additionally, vegetation ET serves as a crucial indicator of ecosystem productivity, vegetation health, and moisture availability, making it invaluable for drought-related studies [28,29,30]. Accurate ET estimation at high spatiotemporal resolutions is essential for effective water resource management, ecosystem conservation, and agricultural planning [31,32,33].

Unlike direct measurements of precipitation and runoff, estimating ET over large areas, such as regions or ecosystems spanning multiple climatic zones, remains challenging. This is due to variability in vegetation types, topography, and meteorological conditions. These complexities contribute to uncertainties in understanding water cycles at the regional and global scales, as models must account for diverse environmental factors that directly influence ET, often with limited or inconsistent data across expansive areas [34].

By leveraging the MODIS vegetation ET model, this study aims to uncover detailed relationships between vegetation ET and meteorological factors. It combines meteorological data with on-ground observations. This research evaluates variations in vegetation ET to assess ecosystem degradation, identify critical restoration areas, and mitigate meteorological disasters in mountain forest ecosystems. Despite the challenges posed by limited systematic data on natural disasters and human disturbances, integrating remote sensing data with field investigations and plant physiology experiments can provide invaluable insights into forest ecosystem dynamics in the Altay Mountains. We hypothesize that variations in vegetation ET in the Altay Mountains are significantly influenced by long-term changes in meteorological factors, including temperature, precipitation, and drought indices. The Altay Mountains’ forest ecosystems are particularly sensitive to climate change and human disturbances due to their alpine climate, semi-arid environment, and distinct vertical zonality.

This study represents one of the first comprehensive efforts to explore the relationship between vegetation ET and meteorological factors in the Altay Mountains. It employs various methodologies such as MODIS vegetation ET models, trend analysis, linear regression, and climate trend rate methods.

This research integrates field monitoring, computer simulation, and remote sensing to comprehensively analyze these interactions.

2. Data and Methods

2.1. Description of Study Area

The Altay Mountains, spanning approximately 2000 km across China, Kazakhstan, Russia, and Mongolia, were selected for this study, focusing on the Chinese section. This region, situated on the southern slope of the Altay Mountains, extends roughly 450 km east–west and 80–150 km north–south, encompassing a total area of 2.6 × 10⁴ km² [35]. The topography is characterized by a gradual narrowing from northwest to southeast, with a pronounced elevation profile of high and wide in the northwest and low and narrow in the southeast (Figure 1).

There are extensive forests, grasslands, and wetlands in the area, with forests covering 9802 km², which is 37.7% of the total. Wetlands cover about 800 km², or 3% of the total area, and feature Siberian larch (Larix sibirica) and Siberian spruce (Picea obovata), forming various forest types and peat swamps in abundance. As a consequence of these swamps serving as “carbon pools,” northern Xinjiang has an important natural forest area for ecological protection. The Altay Mountains exhibit a continental climate, characterized by rapid temperature increases in spring, short summers, swift temperature drops in autumn, and prolonged, cold winters. The annual average temperature is approximately 2 °C, with a maximum recorded temperature of 33.3 °C. Areas at altitudes between 3100 and 3300 m are snow-covered year-round. Mid- and high-mountain areas, at altitudes of 1400–2600 m, experience average annual temperatures below 9 °C, with July being the warmest month at 15 °C. Temperature variations are significant, with annual differences around 30 °C and daily differences around 12 °C. Precipitation varies with altitude, increasing by 30–80 mm per 100 m rise in elevation. Low-mountain areas receive 200–300 mm annually, middle-mountain areas 300–500 mm, and high-mountain areas 600–800 mm. Precipitation distribution decreases from north to south and east to west, underlining the Altay Mountains’ role as a crucial ecological barrier and natural forest area in northern Xinjiang [36,37].

2.2. Data Sources

Daily meteorological data were collected from seven stations surrounding the Altay Mountains (Altay, Burqin, Habahe, Jimunai, Fuhai, Fuyun, and Qinghe) from 2000 to 2017, sourced from the China Meteorological Data Sharing Service Network (https://data.cma.cn/ (accessed on 15 August 2023)). To maintain data integrity, we applied the inverse distance weighted (IDW) method for spatial interpolation, but we acknowledge that this method can introduce errors, especially in areas with sparse data points or varying topography. To address these potential issues, we conducted cross-validation to ensure the robustness of the IDW results by comparing them with alternative interpolation methods, such as kriging. As a result, spatial interpolation is more reliable and reduces errors. This comprehensive dataset allowed for the exploration of relationships between vegetation ET and meteorological factors through the daily averages of ET, temperature, and precipitation.

Further analysis was conducted using MODIS MOD13Q1 data with a 250 m spatial resolution spanning from January 2000 to December 2017. In order to ensure consistency and accuracy across datasets, the grid size for IDW interpolation was specifically selected to align with the 250 m resolution. Considering the complex terrain of the Altay Mountains, this spatial resolution was considered sufficient for capturing the relevant spatial variability within the study area. In the pre-processing steps, data formats were converted, mosaics were created, projections were converted, and the study area was extracted. Data clarity was enhanced through Savitzky–Golay filtering [38] and MVC synthesis [39]. Additionally, TVDI (Temperature Vegetation Dryness Index) data were derived, offering insights into surface soil moisture conditions.

2.3. Statistical Methods

2.3.1. Definition of Extreme Weather Events

Climate is a statistical distribution of weather events [15]. Statistically, less likely events are considered extreme events when the weather deviates significantly from its average state.

There are two main categories of extreme weather events: The first is the common weather extremes that occur every year. Secondly, severe events are determined directly by the occurrence of an event that does not occur every year at a particular location. First, the frequency of the event is relatively low; second, the event has a relatively large or small intensity value; and third, the event caused significant social and economic damage. Extreme weather climate events are those with a very low probability of occurrence (usually 10% or less of such weather phenomena) based on probability distribution.

2.3.2. Definition of Extreme Temperature Indicators

Based on the criteria developed by the Climate Commission of the World Meteorological Organization (WMO), the extreme climate index is defined and calculated using the “Climate Change Detection and Indicators [15]”. Scholars from both domestic and foreign countries have used this method to study extreme climate events for a long time. This classification system classified extreme temperature indices into 16 types, as shown in

Table 1. Definitions of extreme temperature indicators.

Types	Abbreviation	Extreme Climate Index	Unit	Definition
Relative index	TX10	Number of cold days	d	The number of days on which the daily maximum temperature in a year was less than the 10th percentile value for 1961–2017
	TN10	Number of cold nights	d	The number of days on which the daily minimum temperature in a year was less than the 10th percentile value for 1961–2017
	TX90	Number of warm days	d	The number of days on which the daily maximum temperature in a year was less than the 90th percentile value for 1961–2017
	TN90	Number of warm nights	d	The number of days on which the daily minimum temperature in a year was less than the 90th percentile value for 1961–2017
Absolute index	ID	Freezing days	d	The number of days during the year when the maximum daily temperature is below 0 °C
	FD	Number of frost days	d	The number of days during the year when the daily minimum temperature is below 0 °C
	SU	Summer days	d	The number of days during the year when the maximum daily temperature is greater than 25 °C
	TR	Number of hot nights	d	The number of days during the year when the daily minimum temperature is greater than 20 °C
External index	TXn	Monthly maximum temperature	°C	The minimum value of the daily maximum temperature of each month
	TNn	Monthly minimum temperature	°C	The minimum value of the daily minimum temperature of each month
	TXx	Daily maximum temperature	°C	The maximum value of the daily maximum temperature of each month
	TNx	Daily minimum temperature	°C	The maximum value of the daily minimum temperature of each month
Other index	WSDI	Consecutive warm days	d	The number of days with a daily maximum temperature greater than the 90th percentile value for 1961–2017 and for more than 6 consecutive days
	CSDI	Consecutive cold days	d	The number of days with a daily minimum temperature was less than the 10th percentile value for 1961–2017 and for more than 6 consecutive days
	GSL	Biological growing season	d	The number of first-appearing days in which the average daily temperature was above 5 °C for at least 6 consecutive days and the number of days in which the average temperature was below 5 °C for at least 6 consecutive days after 1 July
	DTR	Daily temperature ranges	°C	The difference between the maximum daily temperature and the minimum temperature during the year

. A weak extreme, low noise, and strong significance are characteristics of these indices, which reflect changes in different aspects of extreme climate.

2.3.3. Definition and Calculation of Extreme Precipitation Indicators

Since Bonsal’s non-parametric scheme is robust in handling non-normally distributed data, this study uses it to calculate the extreme precipitation threshold due to its ability to detect extreme values under complex weather conditions. The following are the steps for calculating extreme precipitation thresholds: Let a meteorological element have N values, and arrange these N values in ascending order; then, the probability that a value is less than or equal to another value is as follows:

$$x_m: P=\left(m-0.31\right)/\left(n+0.38\right)$$

where m is the serial number of x.

For this percentile threshold method, R95 and R99 are also used for the extreme precipitation index. According to the method used, the annual daily precipitation of a station from 1960 to 2013 is arranged in ascending order, and the mean 95th percentile over the past 50 years is defined as the threshold for extreme precipitation events [30]. The precipitation on a given day must exceed this threshold to be considered an extreme precipitation event. RClimDex (1.0) software was used to calculate 10 extreme precipitation indices (

Table 2. Definitions of extreme precipitation indicators.

Code	Name	Unit	Definition
RX1day	Maximum 1d precipitation	mm	The monthly maximum precipitation of 1 d
RX5day	Maximum 5 d precipitation	mm	The maximum precipitation for 5 consecutive days per month
R95p	Heavy precipitation	mm	The daily precipitation is greater than the total precipitation of the 95th percentile in the base period
R99p	Extreme precipitation	mm	The daily precipitation is greater than the total precipitation of the 99th percentile in the base period
PRCPTOT	Total annual precipitation	mm	Total annual daily precipitation
SDII	Precipitation intensity	mm/d	The ratio of total precipitation with daily precipitation ≥1.0 mm to the number of precipitation days
R10 mm	Number of moderate rain days	d	The number of days with daily precipitation ≥ 10 mm
R20 mm	Number of heavy rain days	d	The number of days with daily precipitation ≥ 20 mm
R25 mm	Number of rainstorm days	d	The number of days with daily precipitation ≥ 25 mm
CDD	Continuous dry days	d	The maximum number of consecutive days with daily precipitation < 1.0 mm
CWD	Continuous wet day	d	The maximum number of consecutive days with daily precipitation ≥1.0 mm

) based on the research needs.

2.3.4. Definition and Calculation of Standardized Precipitation Evapotranspiration Index (SPEI)

In the SPEI, total precipitation (P) is summed with potential evapotranspiration (PET) to give a monthly time series of total precipitation. To calculate the monthly mean day length, the Thornthwaite equation relies only on the monthly mean temperature (T) and latitude (L). The SPEI is calculated as follows:

(1) The monthly potential evapotranspiration was calculated by using the Thorn Thwaite method

$${PET=16K\left({\displaystyle \frac{10T}{I}}\right)}^m$$

where K is the correction coefficient, which is calculated according to the dimension; T is the average monthly temperature; I is the annual total heating index; m is the coefficient determined by I.

(2) The difference between monthly precipitation and potential evapotranspiration was calculated:

$$D_i=P_i-{PET}_i$$

where P_i is monthly precipitation. and PET_i is monthly potential evapotranspiration.

(3) The Log-logistic distribution of three parameters was used to fit D_i, and the cumulative function was obtained:

$$f\left(x\right)={\displaystyle \frac{\beta }{\alpha }}{\left({\displaystyle \frac{x-\gamma }{\alpha }}\right)}^{\beta -1}{\left[1+{\left({\displaystyle \frac{x-\gamma }{\alpha }}\right)}^{\beta }\right]}^{-2}$$

$$F\left(x\right)={\int }_0^{x}f\left(t\right)dt={[1+{\left({\displaystyle \frac{\alpha }{x-\gamma }}\right)}^{\beta }]}^{-1}$$

where α is the scale parameter, β is the shape parameter, γ is the origin parameter, f(x) is the probability density function, and F(x) is the probability distribution function.

(4) Standardized normal processing was performed on the sequence to obtain the corresponding SPEI:

$$SPEI=W-{\displaystyle \frac{C_0+C_1+C_2{W}^2}{1+d_{1}W+d_2{W}^2+d_3{W}^3}}$$

$$W=\sqrt{-2\mathit{ln}P}$$

where when P ≤ 0.5, P = F(x). When P > 0.5, P = 1 − F(x); the other parameters are C₀ = 2.515 517, C₁ = 0.802 853, C₂ = 0.010 328, d₁ = 1.432 788, and d₂ = 0.001 308.

The SPEI has the characteristics of multiple time scales (1, 3, 6, 12, 24, 48 months, etc.). The 3-month SPEI can reflect seasonal drought, while the 12-month SPEI can reflect interannual drought more clearly.

We selected 6-month and 12-month SPEI scales for the analysis and comparison of the temporal and spatial evolution characteristics of the growing season and interannual drought in the study area and categorized them according to the criteria in

Table 3. Classification of drought grade of standardized precipitation evapotranspiration index.

Grade	Types	SPEI Value
1	Non-drought	(−0.5, +∞)
2	Minor drought	(−1.0, 0.5]
3	Moderate drought	(−1.5, 1.0]
4	Severe drought	(−2.0, −1.5]
5	Extreme drought	(−∞, −2.0]

.

2.3.5. Climate Tendency Method

When using the time series of meteorological factors, Y as a meteorological variable and t as the time variable, and setting t as the time variable, a linear regression equation between Y and t was established. To determine the linear relationship between vegetation ET and meteorological factors, Pearson correlation coefficients were calculated. Meteorological factors contribute to climate tendency through the following formula:

$$Y_{\mathrm{i}}=a_0+a_1{t}_i$$

where Y_i is the meteorological factor, t_i is time, α₁ is the linear trend term, and α₁ × 10 is the climate tendency rate of meteorological elements every 10 years, with the unit of 10 years. When it is less than 0, this means that the meteorological factor sequence decreases with time; otherwise, it increases. The larger the absolute value of α₁, the more significant the trend [40,41,42].

2.3.6. Correlation ANALYSIS

In this study, the dimensionless climate trend coefficient r_xt is obtained by using the following equation:

$$r_{xt}={\displaystyle \frac{{\displaystyle {\sum }_{t=1}^n(x_t-x)(t-\frac{n+1}{2})}}{\sqrt{{\displaystyle {\sum }_{t=1}^n{(x_t-x)}^2{(t-\frac{n+1}{2})}^2}}}}$$

In the equation, r_xt is the trend coefficient, and its significance can be tested by t-distribution statistics. The relationship between tendency rate b and trend coefficient r_xt is as follows:

$$\mathrm{b}=r_{xt}({\sigma }_x/{\sigma }_t)$$

where σ_x and σ_t are the mean square deviations of a factor sequence and natural sequence, respectively.

To visualize the relationship between vegetation productivity and precipitation and temperature, the Pearson correlation coefficient (R) is employed. In general, a high R-value indicates a strong correlation, while a low R-value indicates a weaker correlation. With MATLAB 24.1.0.2508561 (R2024a) software, custom code was written to perform a grid-by-grid analysis, evaluating the correlation between each grid and providing a comprehensive spatial examination of the relationships between variables. The technology roadmap of this study is shown in the Figure 2.

3. Results

3.1. Effect of Air Temperature on Evapotranspiration of Vegetation

The correlation coefficient between annual average temperature and ET in different parts of the study area was obtained through linear regression analysis, and the distribution of these coefficients is illustrated in Figure 3.

Figure 3 illustrates the correlation between ET and the annual mean temperature across the Altay Mountains. The majority of the region demonstrates a weak positive correlation, with correlation coefficients ranging between −0.2 and 0.2. Specifically, 37.92% of the area exhibits a weak positive correlation (0 to 0.2), while 42.63% of the region shows a negative correlation, with 30.42% displaying a moderate negative correlation (−0.2 to 0). Figure 4 further details the spatial distribution of these correlation coefficients. This provides a clearer understanding of the influence of temperature on ET across different parts of the Altay Mountains.

Additionally, Figure 3 offers a spatial analysis of vegetation ET in relation to the annual mean temperature. This highlights the variability in ET responses to temperature gradients across the region. This analysis provides valuable insights into temperature–ET interactions’ spatial patterns and their implications for regional ecosystem dynamics.

In Figure 4 (with the left panel showing the spatial distribution and the right panel showing the percentage and ratio), vegetation ET responded differently in different parts of the Altay Mountains to temperature changes. With decreasing temperature, vegetation ET decreased in the majority of mountainous regions (50.02% of the total area). Conversely, in 28.38% of the area, vegetation ET decreased as temperature increased. A decrease in temperature led to an increase in vegetation ET in 14.75% of the area, while an increase in temperature resulted in increased vegetation ET in 6.85% of the area. These findings highlight the complex and diverse responses of vegetation ET to temperature variations across the Altay Mountain region.

3.2. Effect of Precipitation on Evapotranspiration of Vegetation

The correlation between annual cumulative precipitation and ET was assessed using the same analytical method, with the results depicted in Figure 5.

Based on the map, 98.91% of the region shows a significant positive correlation between annual precipitation and ET. There are areas of weak to strong positive correlations, with some regions showing stronger relationships than others. The area with negative correlations is minimal, with only 1.09% exhibiting negative relationships. Based on Figure 5, vegetation ET in the Altay Mountains generally exhibits a positive correlation with annual cumulative precipitation, with 98.91% of the area showing this trend. Only 1.09% of the area exhibited a negative correlation. Specifically, there is a notable distribution of correlation coefficients: 44.42% of areas exhibit coefficients between 0.4 and 0.6, 37.06% show coefficients greater than 0.6, and 15.18% show coefficients between 0.2 and 0.4. These proportions detailing the different correlation coefficients are depicted in Figure 5, while the specific response of vegetation ET to precipitation is detailed in Figure 6.

In 77.06% of the Altay Mountain region, vegetation ET decreased with decreasing precipitation, as illustrated in Figure 6. Conversely, in 21.19% of the area, vegetation ET increased with increasing precipitation. These findings underscore the sensitivity of vegetation ET to changes in precipitation levels across different parts of the Altay Mountains.

3.3. Effect of Drought on Vegetation ET

Under drought conditions, the impact of the standardized precipitation evapotranspiration index (SPEI) and Temperature Vegetation Dryness Index (TVDI) on vegetation ET was assessed using climate trend analysis and correlation coefficients across the study area, as shown in Figure 7. The effects of drought on vegetation ET were analyzed using data with drought SPEI values (except non-drought).

From Figure 7, it is evident that vegetation ET in the Altay Mountains shows a strong positive correlation with the SPEI across 96.47% of the area. Within this, approximately 39.96% and 47.14% of the areas exhibit correlation coefficients between 0.2–0.4 and 0.4–0.6, respectively, indicating a moderate to strong influence of the SPEI on vegetation ET in these regions.

In contrast, vegetation ET does not exhibit a clear trend in response to the Temperature Vegetation Dryness Index (TVDI). Positive and negative correlations with the TVDI were observed in 47.55% and 52.44% of the area, respectively. Furthermore, increased and decreased vegetation ET were identified using climate trend analysis, detailed in Figure 8.

Based on Figure 8, it is evident that in the majority of the Altay Mountains, vegetation ET decreases with greater climate wetting. This covers 76.11% of the region. Conversely, in 20.35% of mountainous areas, vegetation ET increases with increasing wetness. Vegetation ET increases during drought conditions in 1.25% of the areas. This indicates the need for further investigation into these anomalies.

Regarding the relationship between vegetation ET and soil moisture conditions, our analysis reveals that in 41.88% of the area, vegetation ET decreases with reduced soil moisture. In contrast, in 36.51% of the area, vegetation ET increases with higher soil moisture. Additionally, there is an increase in vegetation ET despite decreased moisture in 10.91% of the area, while 10.69% of the area shows increased vegetation ET during drought conditions. These findings highlight the complex interactions between vegetation ET and soil moisture dynamics in the Altay Mountains.

3.4. Effect of Extreme Weather Conditions on Vegetation ET

We examined various extreme weather indices to uncover vegetation ET variability trends in the Altay Mountains under extreme weather conditions. These indices include growing season length (GSL), warm spell duration index (WSDI), daily maximum temperature (TX90), occurrence of daily minimum temperature (TNx), summer days (SU25), Simple Daily Intensity Index (SDII), very wet days (R95p), and heavy precipitation days (R10). Figure 8 displays the calculated correlation coefficients between these indices and vegetation ET.

As depicted in Figure 9, vegetation ET exhibits a negative correlation with extreme temperature indices such as TX90, TNx, SU25, and WSDI. Specifically, negative correlation coefficients are observed in 68.79%, 71.87%, 66.09%, and 69.43% of the area, respectively. Conversely, vegetation ET shows a positive correlation with extreme precipitation indices including SDII, R95p, and R10. The positive correlation coefficients are 86.14%, 92.26%, and 96.19%, respectively.

Additionally, vegetation ET demonstrates both positive and negative correlations with the growing season length (GSL), with 50.82% and 49.18% of the area exhibiting positive and negative correlation coefficients, respectively. For a detailed breakdown of correlation coefficients between vegetation ET and extreme climate indices, refer to

Table 4. The ratio of the area with different correlation coefficients between vegetation ET and extreme climate indices in the Altay Mountains (%).

Index	Range of Correlation Coefficients
	<−0.6	−0.6~−0.4	−0.4~−0.2	−0.2~0	0~0.2	0.2~0.4	0.4~0.6	>0.6
GSL	0.09 **	2.59 **	18.38 *	28.12 *	26.43 *	21.95 *	2.33 *	0.11 *
WSDI	0.00	1.26 *	11.37 *	27.54 *	36.72 **	21.36	1.72	0.04
TX90	0.00	3.77 **	20.34 *	44.68 **	24.62 *	5.72	0.80	0.07
TNx	0.38 **	13.25 **	28.95 **	29.29 *	19.48	7.78	0.83	0.03
SU25	0.00	7.40 **	28.89 **	29.80 *	26.97 *	6.22	0.68	0.04
SDII	0.00	0.03	1.14	12.69	24.58 *	39.51	19.21 **	2.84 **
R95p	0.00	0.04	0.55	7.15	29.73 *	48.20 **	13.97 **	0.36 *
R10	0.01	0.07	0.80	2.93	35.50 **	54.92 **	5.71 *	0.06

Note: ‘**’ indicates p < 0.001, ‘*’ indicates p < 0.01, and other values indicate p < 0.05. The climate tendency model was employed to examine vegetation ET variation in response to various extreme weather indices. As illustrated in Figure 9, the ratio between the areas where vegetation ET changes and extreme climate indices was calculated and mapped.

.

Figure 10 shows a significant relationship between vegetation ET and various extreme climate indices. The areas where vegetation ET declined showed higher percentages associated with decreases in the warm spell duration index (WSDI), Simple Daily Intensity Index (SDII), very wet days (R95p), and heavy precipitation days (R10), with values of 52.64%, 64.86%, 71.19%, and 75.68%, respectively. Conversely, vegetation ET increased in areas where the SDII, R95p, and R10 rose, with corresponding percentages of 21.27%, 21.08%, and 21.52%.

Regarding the influence of growing season length (GSL), 47.26% of the area experienced a decline in vegetation ET with an increase in GSL, while 31.13% of the area showed a decline in vegetation ET with a decrease in GSL. A detailed analysis of vegetation ET changes due to extreme weather indices is provided in

Table 5. Vegetation ET changed area under the influence of extreme weather indices%.

Extreme Weather Indices	ET Rises as the Index Rises	ET Declines as the Index Declines	ET Rises as the Index Declines	ET Declined as the Index Rises
GSL	19.69	31.13	1.92	47.26
WSDI	7.19	52.64	14.41	25.75
TX90	8.21	23.00	13.39	55.39
TNx	1.54	26.59	20.06	51.81
SU25	2.15	31.75	19.45	46.65
SDII	21.27	64.86	0.33	13.54
R95p	21.08	71.19	0.52	7.21
R10	20.52	75.68	1.09	2.72

.

4. Discussions

To obtain spatially distributed data on regional ET from land surfaces, remote sensing is widely recognized as an effective method. As a result of remote sensing data from polar-orbiting satellites, researchers are able to monitor surface biophysical variables such as albedo, biome type, and leaf area index (LAI), which all influence ET [43]. We used MODIS MOD13Q1 data, which provides the NDVI (Normalized Difference Vegetation Index) and EVI (Enhanced Vegetation Index) at a high spatial resolution, covering the period from 2000 to 2017. Previously, these datasets have been used to monitor vegetation health and productivity over large regions [44,45]. Through the use of MODIS data, we were able to assess ET dynamics and the relationship between vegetation and key meteorological factors, contributing to the growing understanding of vegetation–water interactions.

4.1. Precipitation as the Primary Driver of ET

Based on our findings, precipitation drives vegetation ET in 98.91% of the study area, with a strong positive correlation. In other semi-arid regions, water availability determines vegetation productivity and water use more than temperature [46,47]. A study by Yang et al. (2014) highlighted the dominant role precipitation plays in controlling ET in Xinjiang grasslands where water stress affects plant transpiration [48]. In addition, Cao et al. (2018) found that precipitation plays a key role in vegetation ET in arid and semi-arid ecosystems, especially during growing seasons [49]. These findings reinforce the idea that, in regions like the Altay Mountains, water availability through precipitation governs ET variability more than any other meteorological factor [50,51].

4.2. Complex Role of Temperature

ET was influenced less strongly by temperature, with weaker correlations observed in 68.34% of the area. The results of this study are in agreement with findings from previous studies showing that temperature influences energy fluxes, but its influence on ET is secondary to water availability in semi-arid ecosystems [52]. The effect of temperature on ET in China’s arid regions is largely modulated by the availability of water, with the strongest ET responses occurring in areas with more precipitation [53]. Although temperature plays a role in ET in the Altay Mountains, its influence is constrained by the region’s semi-arid conditions, where precipitation is important.

4.3. Impact of Drought and Soil Moisture

To analyze soil moisture and drought conditions, we also used the Temperature Vegetation Dryness Index (TVDI). According to TVDI soil moisture data, vegetation ET is sensitive to moisture availability, supporting previous studies indicating that soil moisture regulates ET [54,55]. It has been shown by Tijdeman et al. (2021) and Liu et al. (2021) that moisture stress in soil can reduce ET, particularly during drought periods [56,57]. In accordance with these studies, we found that vegetation ET declines significantly in regions with low soil moisture or drought conditions.

4.4. Effect of Extreme Weather Events

Precipitation and temperature are not the only factors that influence ET. Extreme weather events also play a significant role. Extreme weather indices, such as the Simple Daily Intensity Index (SDII), the number of heavy precipitation days (R95p), and the frequency of days with rainfall exceeding 10 mm (R10), showed strong correlations with ET. There was an inverse correlation between declines in these indices and declines in ET. This is consistent with the results of Wang et al. (2021), who observed a significant reduction in ET in European forests during extreme droughts and heat events [58]. In northwestern China, Ying et al. (2020) reported similar trends regarding the impact of extreme precipitation events on ET [59]. Therefore, extreme weather events play a significant role in influencing vegetation water use and ecosystem health.

4.5. Comparative Analysis with Previous Studies

Several previous studies have investigated the interaction between temperature, precipitation, and ET in semi-arid environments. Yao et al. (2022) found that precipitation was responsible for up to 30% of ET variability in arid mountainous ecosystems in northwestern China, similar to our findings in the Altay Mountains [60]. Additionally, Abbas et al. (2021) observed that precipitation had a stronger effect on ET than temperature in the Yellow River Delta, which further supports our conclusion that precipitation is the dominant factor controlling ET in the Altay Mountains [61]. Moreover, Zhang et al. (2021) explored the sensitivity of forest productivity to air temperature and precipitation in Inner Asian forests, finding that the interaction between precipitation and temperature can lead to soil moisture stress, influencing ET and carbon uptake [62]. This complex interaction is also evident in our study, where the weaker impact of temperature on ET can be attributed to the overriding influence of water availability.

4.6. Ecological and Management Implications

This study has significant implications for forest resource management in the Altay Mountains and other semi-arid ecosystems. The strong dependence of ET on precipitation underscores the need for water resource management strategies that account for future changes in precipitation patterns and the increased frequency of extreme weather events due to climate change [63]. Reduced water availability could exacerbate forest ecosystems’ vulnerability, threatening their capacity to regulate water and support biodiversity. Furthermore, our results indicate that forest management practices must adapt to the increasing variability in extreme weather events. This can drastically affect vegetation water use and ecosystem stability.

5. Conclusions

The Altay Mountains in Xinjiang are critical for regulating and safeguarding regional water resources, playing a central role in the ecosystem. In the context of climate change, our study reveals significant shifts in ET dynamics, closely linked to key meteorological factors. Through correlation analysis and climate trend methodologies, our results emphasize that precipitation is the primary driver of vegetation ET, with 98.91% of the study area showing a strong positive correlation. Temperature also influences ET, but its impact is weaker than that of precipitation.

The standardized precipitation evapotranspiration index (SPEI) was found to be a key drought indicator, correlating strongly with vegetation ET, underscoring the importance of moisture availability in determining ET dynamics in this semi-arid region. Rising atmospheric and evaporative demands, influenced by factors such as vapor pressure deficit and wind turbulence, further modulate ET patterns. These demands interact synergistically with precipitation and humidity to drive ET fluctuations.

Although our study focused on precipitation as the dominant meteorological factor, the role of temperature variations in regulating ET dynamics requires further exploration. This is particularly true as climate change progresses. These findings highlight the importance of high-resolution, accurate ET measurements to guide effective water resource management and ecosystem conservation efforts in the Altay Mountains and similar temperate forest regions globally.