Next Article in Journal
Estimation of Solöz River Water Balance Components and Rainfall Runoff Pattern with WEAP Model 1
Previous Article in Journal
Laboratory Experiments of Surge Pressure Loads Acting on Downstream Dams and Reservoir Banks Caused by Landslides in Narrow, Elongated Reservoirs
Previous Article in Special Issue
Distribution Characteristics of Spring Maize Meteorological Drought in Different Climatic Regions of Inner Mongolia Based on Standardized Precipitation Evapotranspiration Index
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Water
Volume 17
Issue 8
10.3390/w17081134
water-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Spatio-Temporal Variation Characteristics of Grassland Water Use Efficiency and Its Response to Drought in China

by
Mengxiang Xing
1,
Liang Liu
1,
Jianghua Zheng
1,2,*,
Xinwei Wang
1 and
Wei Li
1
1
College of Geography and Remote Sensing Sciences, Xinjiang University, Urumqi 830046, China
2
Xinjiang Key Laboratory of Oasis Ecology, Xinjiang University, Urumqi 830046, China
*
Author to whom correspondence should be addressed.
Water 2025, 17(8), 1134; https://doi.org/10.3390/w17081134
Submission received: 15 February 2025 / Revised: 1 April 2025 / Accepted: 8 April 2025 / Published: 10 April 2025
(This article belongs to the Special Issue Effect of Climate Change on Water Availability in Arid and Semi-arid Regions)
Browse Figures
Versions Notes

Abstract

:
Understanding the impact of drought on the water use efficiency (WUE) of grasslands is essential for comprehending the mechanisms of the carbon–water cycle in the context of global warming. Nevertheless, the cumulative and lagged effects of drought on WUE across different grassland types in China remain unclear. This study investigates the cumulative and lagged effects of drought on WUE across different grassland types in China from 1982 to 2018. We employed the Sen-MK trend test and correlation analysis to identify the primary factors influencing the temporal effects of drought on WUE. The results indicated that WUE in Chinese grasslands, across various grassland types, exhibited an upward trend over time, with the most rapid increase observed in meadow. Drought had both cumulative and lagged effects on WUE, with cumulative effects lasting an average of 5.2 months and lagged effects lasting 6.1 months. Specifically, the cumulative effects of drought on WUE lasted for 5.6 months for alpine and subalpine meadow, slope, and desert grassland, whereas the lagged effects lasted 9 months for alpine and subalpine plain grassland. Furthermore, the influence of drought on WUE in grasslands varied across different grassland types and intensified with increasing altitude. The trends observed in the cumulative and lagged impacts of drought on WUE across various aridity index (AI) zones were consistent with those for grasslands as a whole. Our findings underscore that the response of WUE to drought in grasslands and their distinct types is primarily characterized by lagged effects. This research provides an important reference value for enhancing the stability of grassland ecosystems.

1. Introduction

Drought is characterized by a reduction in soil and atmospheric moisture resulting from insufficient effective precipitation over a specific period. This phenomenon significantly impacts soil moisture and vegetation transpiration, thereby indirectly influencing the photosynthetic activity of plants [1,2]. By the year 2100, it is anticipated that global drought events will become more intense and frequent [3]. Water use efficiency (WUE) is commonly defined as the ratio of total primary productivity (GPP) of vegetation to evapotranspiration (ET) [1,4,5]. This metric is essential for understanding the relationship between carbon and water cycles, as it highlights the interactions between water availability and plant growth [6]. Additionally, WUE is utilized to assess ecosystem responses to drought conditions, prolonged high temperatures, and water deficits [7]. The mechanisms by which meteorological drought affects vegetation WUE are complex and can vary in duration [2]. With global warming and the increasing prevalence of drought, understanding the link between drought and vegetation WUE plays a vital role in predicting future carbon–water cycles and their associated feedback mechanisms.
Grasslands are extensively spread across the alpine climate zone of the Tibetan Plateau, northern China, and the arid and semi-arid regions of western China. As an essential part of ecosystems, grasslands act as an environmental buffer against global environmental changes, significantly contributing to the maintenance of ecological balance and the management of climate and water resources [2,8]. Drought significantly affects the grassland ecosystem’s carbon and water cycle, and the degree of this impact varies considerably across different types of grasslands [6]. The differences in the response of different grassland types to drought mainly manifest in their physiological, morphological, and ecological characteristics [9]. Drought-tolerant grasslands have a greater ability to retain and utilize water, while wet grasslands may face more severe ecological stress in times of drought [10]. However, the relationship between WUE and drought in different grassland types is not clear and needs to be further explored. Therefore, a deeper understanding of the relationship between WUE and drought in different grassland types will help provide an important theoretical basis for the impact of climate change on the ecological water–carbon cycle in arid regions.
Grassland typically close their stomata during short-term droughts to minimize evaporation and improve WUE, but prolonged drought conditions can result in sustained stomatal closure, which subsequently decreases photosynthesis and undermines carbon sequestration [11]. Research has demonstrated that historical droughts exert a more significant influence on vegetation growth than contemporary conditions, with early studies showing that including temporal effects enhances the explanatory power of vegetation responses to climate dynamics [12,13,14]. These temporal effects of drought on vegetation growth encompass both lagged and cumulative impacts [15]. The lagged effects refer to the impacts of historical droughts on vegetation, while cumulative effects denote the gradual accumulation of impacts on vegetation resulting from consecutive droughts over time [16]. Lagged effects can be utilized to characterize the sensitivity of vegetation growth to drought conditions [17]. Earlier droughts diminish the carbohydrate reserves necessary for vegetation growth, thereby inhibiting growth and leading to negative consequences such as an increased risk of pests and diseases [18]. Unlike lagged effects, cumulative effects reflect the tolerance of vegetation during prolonged droughts, highlighting its adaptability and critical role in ecosystem stability [17]. The cumulative effects of drought restrict grassland’s water absorption and photosynthesis through sustained water stress, leading to a decrease in WUE [19]. Under sustained drought conditions, grassland’s water use efficiency decreases due to the limited availability of water resources [20]. In addition, prolonged drought can alter the physiological adaptation mechanisms of grassland, leading to a decrease in its adaptability to water scarcity. However, in the Chinese context, the lag and cumulative effects of drought on WUE across different grassland types, as well as the dominant role of drought on WUE in various grasslands, remain unclear. Therefore, further investigation is needed, which will help to reveal the long-term and delayed effects of drought on grassland WUE.
External environmental changes have a more pronounced impact on the sensitivity and vulnerability of grassland ecosystems [2,21,22]. The lagged and cumulative effects are often asymmetric, influenced by factors such as the severity and duration of dry conditions, climatic zones, plant types, and elevation [23,24]. The altitude affects temperature and precipitation, while the wetness–dryness degree directly determines water supply [25,26]. The two factors jointly influence grassland growth and species diversity. Currently, the cumulative and lagged effects of drought on grassland WUE across different altitudes and wet–dry regions still require further investigation. Further study of the cumulative and lagged effects of grassland water use efficiency on drought from the perspective of different climatic zones and altitudinal gradients is needed. This will help to elucidate the complexity of WUE responses to drought from a multidimensional perspective.
In this study, we utilized the 1982–2018 grassland WUE estimated from NIRv-GPP and PEW-ET data and standardized precipitation evapotranspiration index (SPEI) data on 1–12 month time scales to investigate the cumulative and lagged effects of drought on the WUE of grasslands and different grassland types in China under different elevations and aridity index (AI) zones and also to identify the main roles of drought on the WUE of grasslands. The objectives of this study are to answer the following questions: (1) How does the spatio-temporal distribution characteristic of WUE in different grassland types in China from 1982 to 2018 vary across different elevation and AI zones? (2) From multiple perspectives, how does the response of drought to the WUE of different grassland types in China change at different altitudes and AI zones? (3) Is the main effect of drought on the WUE of different grassland types in China a cumulative or lagged effect?

2. Materials and Methods

2.1. Study Area

China is geographically located in the eastern part of Eurasia, characterized by a three-step gradient in terrain that gradually decreases from west to east. This significant altitudinal variation contributes to the diverse types of grasslands found in the region (Figure 1a). Grasslands in China include alpine and subalpine meadow (ASM), slope grassland (SG), plain grassland (PG), desert grassland (DG), meadow (MD), and alpine and subalpine plain grassland (ASPG) (Figure 1b) [27]. ASM (26.5%), ASPG (18.2%), and PG (13.2%) are mainly distributed in high and sunlit areas such as the Tibetan Plateau, Tianshan, and Altai Mountains [12]. DG (16.6%) is located in the desert fringes of Xinjiang, Tibet, and central Inner Mongolia. SG (5.9%) is primarily in southwestern China, characterized by low-altitude hilly areas, lowlands, valleys, and mountain plains. MD, covering 19.5%, is mainly in northeastern Qinghai Province and northeast China (Figure 1b).

2.2. Data

The data we used included NIRv-GPP, PEW-ET, SPEI, AI, and GLC2000 land use data and DEM data (Table 1). To calculate WUE, PEW-ET and NIRv-GPP data were used in this study. The NIRv-GPP dataset is generated based on approximately 40 years of remote sensing AVHRR data and observations from hundreds of global flux stations. This dataset has broad global coverage and a long GPP time series (1982–2018). Compared to other GPP products, it demonstrates superior performance in capturing interannual variations [28]. The PEW-ET model is a water–energy balance evapotranspiration model based on the proportionality hypothesis. This model simultaneously considers the effects of water balance constraints and energy budget processes, improving simulation accuracy compared to previous models. It provides a foundation for studying long-term water cycles and climate change [29]. To investigate the impact of drought on grassland WUE, this study used SPEI data for analysis. The SPEI data are based on water balance calculations, considering the effects of climatic factors and effectively monitoring the occurrence of drought, making it an excellent index for quantifying drought severity [30,31]. In addition, SPEI data can also describe the intensity and duration of droughts, and they are widely used in drought monitoring and vegetation studies [10]. Consequently, this study selected SPEI data from 1 to 12 months to compute cumulative and lagged effects. The AI was an indicator that characterizes the degree of wetness and dryness of a region [32]. This study classified the mean AI data from 1982 to 2018 into the categories listed below: humid (AI < 1), semi-humid (1 ≤ AI < 1.5), semi-arid (1.5 ≤ AI < 4), and arid (AI ≥ 4), in order to analyze the WUE values and cumulative and lagged effects under different degrees of the AI [33]. Grassland type data were obtained from GLC2000, a global coverage data product of the Institute for Space Applications of the European Union Joint Research Center, and grassland types in the Chinese region mainly include ASM, SG, PG, DG, MD, and ASPG (Figure 1b) [27]. The grassland type boundary data are converted based on the raster data of GLC2000. The specific operation is to use the raster-to-polygon tool of the ArcGIS software (version 10.8) to convert different grassland type raster data into polygon data and add them to the map. The digital elevation model (DEM) is derived from the Advanced Spaceborne Thermal Emission and Reflection Radiometer Global Digital Elevation Model. The dataset is a digital elevation model of the globe obtained using satellite-based thermal emission and reflectance radiometers with high vertical and horizontal accuracy. Eventually, for subsequent analysis and discussion, the data were resampled to a uniform 0.05° × 0.05° resolution using a bilinear resampling method. The bilinear interpolation method achieves high interpolation accuracy by determining the new values of the image elements based on the weighted average distance between the centers of gravity of the four nearest input image elements. This method is better suited to be applied to continuous data and is capable of smoothing the data during resampling [34,35].

2.3. Data Validation

The ET and GPP validation data used in this study are long-term observation data from the ChinaFLUX network of land ecosystem flux observations and field measurement data from other observation stations in China (Table 1) [36]. Grassland ecosystem flux data in China were extracted by screening data. The results show a significant correlation between the PEW-ET model and the observed ET values (R2 = 0.73, p < 0.01) (Figure 2a). Due to the lack of measured GPP data at some flux tower sites, the sample size for the correlation study between NIRv-GPP and the observed GPP values was reduced. The results showed that the correlation coefficient was R2 = 0.63 (p < 0.01) (Figure 2b). The validation of both data shows that PEW-ET and NIRv-GPP accurately reflect the actual situation.

2.4. Method

2.4.1. Estimation of WUE

In the context of the ecosystem, W U E was typically identified as the rate of carbon acquisition to water loss [37]. In this paper, G P P was chosen to quantify carbon acquisition, and E T represented water loss. W U E was calculated as follows [38]:
W U E = G P P / E T
where the unit of G P P data is gC·m−2 month−1, E T is in mm month−1, and W U E is in gC·kg−1H2O.

2.4.2. Calculating the Cumulative Effect of Drought on Water Use Efficiency

By evaluating the cumulative effect of drought on grassland W U E , we determined which time scale of S P E I has the greatest correlation with grassland W U E . For example, if the correlation between S P E I (1 ≤ n ≤ 12) and grassland W U E is maximized at an n -month time scale, the cumulative effect of drought on grassland W U E for that pixel is determined to be n months. This suggests that grass W U E during the first n months is a critical time point for drought accumulation to affect grasses. The calculation steps are as follows:
We used Pearson correlation analysis to evaluate the cumulative effects and the cumulative time of drought on grassland W U E in China. Firstly, we selected a total of 259 images of month-by-month W U E data of grassland in China and 259 corresponding images of S P E I data on an m -month time scale (1 ≤ m ≤ 12) for the growing period (from April to October) from 1982 to 2018. Then, according to Equation (1), we conducted correlation analyses between the W U E of grassland and a S P E I time scale of 1 month ( m = 1), resulting in the correlation coefficient r 1 ; the W U E of grassland was correlated with the S P E I time scale of 2 months ( m = 2), resulting in the correlation coefficient r 2 . And so on until the calculation with the S P E I time scale of 12 months is completed, resulting in a correlation coefficient r 12 and finally a total of 12 correlation coefficients ( r 1 , r 2 … r 12 ). According to Equation (2), we used the maximum value as the cumulative time effects value of drought on the W U E of grassland for this pixel [24].
r m = c o r r W U E , m S P E I   1 ≤ m ≤ 12
R m a x _ c u m = m a x r m   1 ≤ m ≤ 12
where m denotes the cumulative time scale of S P E I , m S P E I is the S P E I time series with m cumulative months, and r m is the correlation coefficient between W U E and m S P E I .
Then, we used the grassland type data to mask the spatial raster data of the cumulative effect, obtaining the cumulative effect values of drought on different grassland types. Finally, we used the grassland type data to mask the cumulative months spatial raster map. By calculating the number of pixels for 1–12 cumulative months in different grassland types, we derived the area distribution map for different cumulative months.

2.4.3. Calculating the Lagged Effect of Drought on Water Use Efficiency

By evaluating the lagged effect of drought on grass W U E , the lagged effect of drought on grassland W U E for the pixel was determined to be n months for the first n months (1 ≤ n ≤ 12) when the correlation with the S P E I on a 1-month time scale was greatest. This indicates that drought in the first n months affects grassland W U E in the current month. The calculation steps are as follows:
We selected a total of 259 monthly images of W U E data of grassland in China during the growing period from 1982 to 2018 and the S P E I data for the time scale of 1 month for the previous i months (0 ≤ i ≤ 12) to compute the lagged effects of drought on grassland W U E . Firstly, according to Equation (4), we correlated the W U E of grassland with the SPEI time scale of 1 month match the current month to obtain the correlation coefficient r 0 . After that, we calculated the correlation coefficients of months 1–12 sequentially and obtained 13 correlation coefficients ( r 0 , r 1 … r 12 ) in total. Finally, the maximum value was taken as the value of the lagged effect of drought on the W U E of grassland for that pixel [24].
r i = c o r r W U E , S P E I i   0 ≤ i ≤ 12
R m a x _ l a g = m a x r i   1 ≤ i ≤ 12
where i denotes the time series of W U E at the i month-lagged W U E from 1982 to 2018 during the growing season, S P E I i represents the S P E I time series of 1 month at the i month lag, and r i is the correlation coefficient at the i month.
Then, we used the grassland type data to mask the spatial raster data of the lagged effects, obtaining the lagged effect values of drought on different grassland types. We then applied pixel statistics tools to calculate the average values of the raster data, yielding the average lagged effect values for different grassland types. Finally, we used the grassland type data to mask the lagged months spatial raster map. By calculating the number of pixels for 0–12 lagged months in each grassland type, we derived the area distribution map for different lagged months.

2.4.4. Trend Analysis

The Theil–Sen median is a non-parametric statistical method that effectively mitigates the impact of outliers in time series, making it ideal for trend analysis of long-term data. The formula is as follows [39,40]:
S e n = m e a n X j − X i j − i , ∀ j > i
where Sen represents the slope, and the Mann–Kendall significance test was further used to analyze the significance of the trend changes [41,42]. Given a significance level α, a trend in the time series data is considered significant when | Z | > Z (1 − α)/2. Specifically, if | Z | exceeds 1.65, 1.96, or 2.58, it indicates that the test passes the significance threshold with confidence levels of 90%, 95%, and 99%, respectively.

3. Results

3.1. Spatial and Temporal Variation in Grassland WUE

We analyzed the annual WUE trends for grasslands and various grassland types in China from 1982 to 2018. The mean multi-year grassland WUE was 1.32 gC·kg−1H2O, peaking at 1.44 gC·kg−1H2O in 2017 and reaching a low of 1.17 gC·kg−1H2O in 1989. The overall WUE of the grassland exhibited an upward trajectory in interannual variation, fluctuating in the range of 1.17 to 1.44 gC·kg−1H2O. When examining annual WUE variations across different grassland types, all six types displayed similar patterns and showed a decline from 1988 to 1989. The interannual variations of ASM and ASPG are relatively stable, but the trend of change is not significant (p > 0.05). For the remaining four grassland types, the interannual variations of WUE show a fluctuating upward trend, and the change is significant (p < 0.01). Notably, MD WUE increased the fastest interannually, increasing at a rate of 0.0085a−1, followed by SG, PG, and DG (Figure 3).
The spatial distribution of WUE in Chinese grasslands showed a west-to-east pattern of increase, decrease, and increase again, demonstrating clear spatial differentiation (Figure 4a). The low-WUE area (0 gC·kg−1H2O–2 gC·kg−1H2O) was primarily located in the Ningxia, central Inner Mongolia, and western Tibetan Plateau, comprising 64.65% of Chinese total grassland area. The high-WUE area (WUE > 4 gC·kg−1H2O) was concentrated in the Ili Valley of Xinjiang and the eastern Tibetan Plateau, accounting for only 4% of the total. The median-WUE area (2 gC·kg−1H2O–4 gC·kg−1H2O) was concentrated in northern Xinjiang, the eastern Tibetan Plateau, northwestern Sichuan Province, and the northeastern region. The multi-year average WUE across various grassland types in China was as follows: MD (1.97 gC·kg−1H2O) > ASM (1.72 gC·kg−1H2O) > SG (1.44 gC·kg−1H2O) > PG (1.26 gC·kg−1H2O) > DG (0.65 gC·kg−1H2O) > ASPG (0.39 gC·kg−1H2O) (Figure 4b).
Combined with different altitudes and AI zones, the WUE of grassland showed a trend of decreasing (at 1000 m) then increasing (at 1500 m) and then decreasing sharply (above 4000 m) with increasing altitude, and the minimum value appeared at 6000 m above sea level (Figure 5a). SG WUE showed an increasing trend with elevation, reaching a minimum at 2000 m and increasing thereafter (Figure 5c). The WUE of DG showed a continuous decreasing trend with increasing elevation (Figure 5e). Meanwhile, the rest of the grassland WUE with increasing elevation gradient basically showed a similar trend to the overall grassland WUE (Figure 5b,d,f,g). The grassland WUE under different AI zones from high to low: semi-humid (1.96 gC·kg−1H2O) > humid (1.78 gC·kg−1H2O) > semi-arid (1.19 gC·kg−1H2O) > arid (0.75 gC·kg−1H2O) (Figure 5h). The different grassland types’ WUE showed similarity to grassland as a whole under different AI zones, but MD WUE was higher than other grassland types’ WUE, and the DG WUE was basically the same under different AI zones (Figure 5i).

3.2. Spatial Distribution of Grassland WUE Trends

The findings indicated considerable spatial variability in the WUE trends of grasslands across China (Figure 6a). The WUE of grassland in China as a whole was mainly increased (57.90% of the area), of which the proportion of areas with dramatic increases (p < 0.01) and significant increases (p < 0.05) were 17.40% and 8.50%, respectively, and were largely concentrated in the Ili Valley, the Yunnan, and Inner Mongolia. Grassland WUE showed a dramatic decrease (p < 0.01) in only 2.40% of the area, mainly in the Tacheng region of Xinjiang and the southwestern part of the Tibetan Plateau. Specifically for the trend changes in WUE in different grasslands, we found that the trends of WUE changes in ASM were mainly unchanged (35%) (Figure 6b). WUE trends had a significant increasing trend over 51% of the SG area. WUE trends in PG were similar to those of MD, with increasing trends dominating WUE for both grassland types. The trend of WUE change in DG was mainly dominated by a decline, with a notable decrease representing 22% of the area. The trends of increasing (42%) and decreasing (41%) trends in WUE for ASPG are generally consistent as a percentage.

3.3. Spatial Distribution of Cumulative Effects of Drought on Grassland WUE

Grassland WUE was mostly positively correlated with the strength of the cumulative effects of SPEI at different time scales (72.15% of the area, of which 13.37% passed the test of significance at the 0.05 level) (Figure 7a). The mean value of cumulative positive correlation R m a x _ c u m was 0.06, and in northwestern Sichuan cumulative positive correlation R m a x _ c u m was mostly greater than 0.2. The mean value of the cumulative positive correlation R m a x _ c u m for ASM is the highest in the different grassland types (0.077), while MD is the lowest (0.042) (Figure 7c). The cumulative negative correlation R m a x _ c u m was mainly distributed in the northeastern part of Inner Mongolia and the western part of the Tibetan Plateau (Figure 7a). The mean value of the cumulative negative correlation ∆ R m a x _ c u m for SG is the highest (0.038), while MD is the lowest (0.029) (Figure 7c). Combining the different grassland types, the cumulative negative correlation R m a x _ c u m area proportion was greater than the positive correlation for PG, while the opposite was true for other grassland types. ASM, DG, and ASPG accounted for more than 70% of the area of cumulative positive correlation R m a x _ c u m , followed by MD and SG with 65.17% and 57.09%, respectively (Figure 7c).
Grassland WUE exhibited notable spatial variability in response to cumulative drought, displaying an average of 5.2 months. This response showed a spatial trend of increasing to decreasing from west to east. About 44.6% of the region’s cumulative months were concentrated in 1–3 months, mainly in the Ili Valley of Xinjiang, the Tibetan Plateau, the northwestern part of Sichuan Province, and the northeast region (Figure 7b). Looking at the different grassland types, there were differences in the percentage of the cumulative month area. The WUE of ASM and ASPG accumulates over 5.6 months and 5.4 months of drought, respectively, with similar area proportions. The cumulative drought effect on DG and SG averaged 5.6 months, 0.6 months longer than that on PG. The average cumulative months of MD WUE was 4.5 months and was mainly concentrated in 2–3 months (Figure 7d).
The cumulative months of drought on grassland WUE varied with altitude, showing a trend of increase followed by decrease, especially at high altitudes (above 4000 m), with the fluctuation range concentrated in 4.2–6.4 months (Figure 8a). In SG, the months of drought response showed a gradual increase followed by a rapid decline with rising elevation. In contrast, ASPG showed the opposite trend to SG, decreasing initially and then increasing gradually (Figure 8c,g). The remaining four grassland types displayed a trend similar to that of grasslands overall (Figure 8b,e,f). Under different AI zones, the variation in the average cumulative months of grassland WUE in response to drought was minimal (Figure 8h). The cumulative number of months of WUE response to drought for different grassland types showed a similar trend to the overall grassland in different AI zones. However, in arid regions, SG showed higher cumulative months compared to other types, while MD had lower values in semi-humid and semi-arid regions (Figure 8i).

3.4. Spatial Distribution of Lagged Effects of Drought on Grassland WUE

The area proportion of the intensity of the lagged effects of grassland WUE with SPEI at different time scales was 66.99% (p < 0.05), of which the area proportion of R m a x _ l a g > 0 was 66.59%, and the mean value of the lagged positive correlation R m a x _ l a g was 0.13, whereas the mean value of the lagged negative correlation R m a x _ l a g was −0.02. The mean value of the lagged positive correlation R m a x _ l a g for ASPG is the highest (0.195), while for SG it is the lowest (0.106) (Figure 9c). The mean value of the lagged negative correlation ∆ R m a x _ l a g for MD is the highest (0.018), while for SG it is the lowest (0.01). The strength of the lagged effects of grassland WUE on SPEI was concentrated in the range 0–0.2 (88.22% of the area). The western part of the Tibetan Plateau was the area where R m a x _ l a g > 0.2 image elements were concentrated (Figure 9a). Combining the intensity of the lagged effects of WUE and SPEI for different grassland types, we found that R m a x _ l a g > 0 accounted for more than 99% of the area (Figure 9c).
The lagged effects of drought on grassland WUE averaged 6.1 months, with a distribution of increasing and then decreasing lagged months from west to east (Figure 9b). The lag of 9 months, accounting for 14.83% of the area, was primarily distributed in the western part of the Tibetan Plateau. Areas with a lag of 7 months and 8 months had similar proportions, while lags of 10–12 months constituted only 5% of the area, predominantly located in the northwestern part of Sichuan Province (Figure 9b). Combined with the WUE response lag of different grassland types to drought, the results showed that the WUE lag time of ASM was 6.4 months on average, and the area with a 9-month lag accounted for the largest proportion (13%). SG and DG had similar lag time area proportions, with an average lag of 5.9 and 5.7 months, respectively. The average months of lag for ASPG was the longest (7.3 months) among all the grassland types (Figure 9d).
Combined with different elevations and AI zones, the months of lag for grassland WUE in response to drought tended to fluctuate and increase with increasing elevation, with the fluctuation range concentrated in the 4.8–7.3 month range. The lagged months value showed a slowly decreasing trend at altitudes of 500–1000 m and 1500–3000 m (Figure 10a). Although there was difference in the lagged months for different grassland types, they all showed a fluctuating trend of increasing with elevation (Figure 10b–g). The difference in the mean lagged months for grassland WUE in response to drought was less under different AI zones, and the mean lagged months were ranked from high to low: arid (6.3 months) = semi-humid (6.3 months) > Humid (6.2 months) > semi-arid (5.8 months). The lagged months for WUE in response to drought for different grassland types showed similar trends to grassland as a whole under different AI zones, but the lag for SG was much higher in arid regions, and for ASPG in semi-arid and arid regions than for other grassland types (Figure 10i).

3.5. Compared with the Cumulative Effect, the Lagged Effects of Drought Dominate Grassland WUE in China

Using the difference between R m a x _ l a g and R m a x _ c u m [16], we explored the dominant factor of drought on grassland WUE at the image element scale. The results revealed that the lagged effect of drought on grassland WUE predominated, affecting 91.3% of the area, with an average Δ R m a x of 0.104. Among them, the Δ R m a x values exceeding 0.2 were concentrated in the southwestern region of the Tibetan Plateau. The average Δ R m a x , reflecting the cumulative impact of drought on grassland WUE in China, ranged primarily from −0.05 to 0 (Figure 11a).
For different grassland types, the WUE response to drought was primarily characterized by a lagged effect. The average Δ R m a x for the lagged effects of WUE in PG were the highest at 0.957, followed by ASM at 0.162. The average | Δ R m a x | for lagged and cumulative effects were consistent across ASM, SG, DG, and MD (Figure 11b). The lagged effects of WUE on drought were larger than the cumulative effects across different grassland types. In ASM, the lagged effect accounted for 82.46% of the area, while in other grassland types, it was around 95% (Figure 11c).

4. Discussion

4.1. Cumulative and Lagged Effects of Drought on Grassland WUE

Our results revealed spatial heterogeneity in both the lagged and cumulative effects of drought on grassland WUE. Grassland WUE exhibited a greater dependence on cumulative and lagged droughts; however, the lagged and cumulative effects were asymmetric (Figure 7 and Figure 9). This asymmetry is primarily attributable to reduced grassland growth and photosynthesis, which result from insufficient nutrient availability under varying intensities of drought. When the physiological status of grasslands returns to normal, GPP typically recovers more slowly than ET, resulting in a significant increase in soil evaporation [43]. Adaptive measures, such as adjusting stomatal closure, reducing water evaporation, and lowering respiration rates, are generally implemented to cope with drought stress, often requiring a period of adaptation [24].
This study found an average lag of 6.1 months for drought effects on grassland WUE, with concentrations occurring between 3 months and 7–9 months. The results indicated that the lagged effects of drought on grassland WUE tended to be prolonged, suggesting that grassland growth requires more time to adjust and adapt in response to drought conditions. Globally, grasslands encompass 25% of the Earth’s land area, with dry and semi-arid ecosystems predominantly found in Asia and Europe. In contrast, other regions display a coexistence of dry, semi-arid, and humid grassland ecosystems [44]. Ji et al. [18] found that the average lag for the effect of soil moisture drought on global grassland WUE is 0–4 months, with central North American grasslands (1–4 months), Asian grasslands (5–12 months), and Australian grasslands (0 months) demonstrating different averages. Thi s finding is inconsistent with our observation of an average lag; however, the results align with those observed in the Chinese region. The discrepancies may stem from various factors influencing global grassland distribution, including climate zones, latitude and longitude, topographical features, and ecological conditions [45]. The Central Asian region, which borders Xinjiang, China, is predominantly composed of grassland yet faces significant water scarcity due to extreme temperatures, low precipitation, and high evaporation rates. These conditions result in frequent drought stress, which further exacerbates the lagged effects of drought on the grasslands. In contrast, the Central North American grasslands, also situated at high latitudes, exhibit a shorter average lag period of WUE in response to drought compared to their Chinese counterparts. This North American region is characterized by increasing water availability from west to east, a progressively wetter climate, and grassland vegetation that includes a mixture of shortgrass and tallgrass prairie. This diversity in vegetation is crucial for enhancing resilience to drought stresses [46]. The variability observed in Chinese grasslands may be attributed to their topographic diversity, whereas central North American grasslands are typically characterized by flat terrain [47]. Regarding lagged effects, the study revealed that the majority of grassland WUE responses to drought on the Tibetan Plateau were primarily concentrated around longer durations, specifically 9 months. This finding contrasts with the short-term lagged responses of grassland GPP to drought documented by Wei et al. [16]. The results indicate that grassland WUE necessitates a longer duration to adapt to drought conditions, involving physiological processes such as leaf stomatal closure, root modifications, and changes in soil moisture levels. In contrast, grassland GPP reacts more swiftly to drought, as photosynthesis is impeded under such conditions, resulting in a rapid decline in GPP. The Tibetan Plateau, situated in an alpine climate zone defined by high altitude, low temperatures, and intense solar radiation, is home to fragile grassland ecosystems. In this environment, soil moisture evaporation is heightened under drought conditions, leading to a relatively prolonged adaptive capacity and response time to drought [48,49]. The lagged response time of grassland WUE to drought in the northwestern part of Southwest China was mainly 12 months, likely resulting from the exceptional drought conditions that affected Southwest China in 2009 and 2010, which resulted in frequent short-term, weakly severe, and high-frequency droughts in northwestern Sichuan [50], and consequently a long lagged response time in this region. Additionally, months with shorter lagged responses were predominantly observed in northeastern Inner Mongolia, as well as in the Ili River Valley and Tacheng region of Xinjiang, where elevations primarily ranged from 500 to 2000 m. This phenomenon is likely attributable to the abundant water resources, favorable temperatures, and significant precipitation characteristic of these low-altitude areas, which alleviated growth constraints and enabled grasslands to swiftly adapt to and alleviate drought stress [51]. Our findings indicate that the cumulative effects of drought on grassland WUE in China are primarily concentrated within a period of 1 to 3 months, with an average cumulative duration of 5.2 months. These results align closely with those of Zhang et al. [52], who reported an average cumulative duration of 5.28 months for the impact of drought on global grassland GPP. The cumulative monthly distribution observed in China aligns with the findings of this study. Although different vegetation growth indicators were employed (this study analyzed both GPP and ET), the consistency underscores the cumulative impact of drought on grassland growth. Notably, the cumulative effects of drought on grassland WUE in China tended to manifest over a brief period, suggesting a rapid response to drought during grassland growth, alongside an increased adaptability and sensitivity to drought [53]. As grasslands become increasingly sensitive to environmental changes, drought significantly diminishes their carbon sequestration capacity and shortens the early warning window due to the rapid accumulation of cumulative responses [54]. Bai and Li. [55] found that the cumulative months of drought’s impact on GPP in the Mongolian Plateau closely resemble the cumulative months of WUE observed in the Inner Mongolian grasslands in this study. In the grasslands of central Inner Mongolia, the positive cumulative effects of drought identified in this study differ from the findings of Bai and Li. [55], implying that the grasslands in this region possess a reduced capacity to withstand water stress compared to those with higher WUE. Different vegetation types exhibit significant variability in their drought tolerance. Grasslands generally respond to drought stress more rapidly than forests; however, forests are more vulnerable to recurrent droughts due to their capacity to access water through deep root systems and their greater moisture storage capacity, which enhances their adaptability in arid environments [56]. Moreover, this short-term accumulation of drought may prompt grasslands to adjust their strategies quickly, enabling them to respond effectively to drought impacts and maintain the ecosystem’s functionality and stability.

4.2. Response of Grassland WUE to Drought in China Under Different Influencing Factors

Our findings indicate that the response in months of grassland WUE to drought conditions varied across different grassland types and AI zones (Figure 8h,i and Figure 10h,i). This variability can be attributed to the diversity of grassland types, which results in differences in vegetation root structures, photosynthesis rates, and transpiration rates. These factors collectively influence the patterns of WUE response to drought [57]. Different vegetation types exhibit unique adaptations and resilience in response to drought conditions. For instance, forests utilize deep root systems to access water, grasslands regulate their growth to conserve resources, desert plants develop drought-resistant traits, and wetland plants rely on saturated water. These adaptations significantly influence their growth patterns and ecological functions [58,59,60]. Research has shown that in the West Asia region, the bare soil and shrubs’ WUE exhibits a rapid response to drought, and in areas characterized by high water demand and rapid growth, vegetation generally possesses small, shallow root systems, which constrain their ability to resist drought [61]. In the Chinese region, the cumulative effects of drought on WUE across various grassland types were predominantly shorter in duration, indicating a more rapid response in grass growth to persistent drought. The duration of lagged effects of drought on WUE for ASM and ASPG was concentrated over longer time spans. Furthermore, these grassland types, which offer significant ecosystem services, are highly susceptible to climate change and are experiencing irreversible degradation [60]. ASPG primarily existed in arid and semi-arid climates, where it was particularly responsive to water availability, and this sensitivity led to extended response times during periods of drought stress. In ASM regions, climate warming led to increased vegetation transpiration and soil evaporation, thereby worsening drought conditions in the northwest [62]. Different AI zones influence factors such as precipitation distribution, temperature, and humidity, which subsequently affect the adaptation and recovery of grassland WUE to drought. The response times of WUE to drought across various grassland types were largely consistent within AI zones, likely due to minimal variation in temperature and precipitation within the same zone, resulting in similar response times.
The cumulative effects of drought on grassland WUE increased with elevation, consistent with the overall trend of lagged responses across elevation gradients (Figure 8a–g and Figure 10a–g). The number of cumulative months of effect were less at lower elevations and stable at middle and high elevations, with notable differences observed between grassland types across the elevation gradient. Intensive human activities, particularly overgrazing, have significantly impacted grassland growth and development. These activities exacerbated degradation, reduced productivity, exposed the soil surface, and heightened soil moisture loss through evaporation [38]. The presence of bare ground influenced climate change, further intensifying climatic droughts, necessitating that grass growth adapt its strategies to cope with drought stress. The lagged effect was stronger at higher elevations, indicating that the lagged effects of drought on grassland WUE were stronger at higher elevations with lower temperatures, a result that aligns with the findings of Du et al. [51]. This may be related to the fact that grasslands at high altitudes are affected by snow and ice for a lengthy duration; precipitation is mainly in the form of snowfall, which exhibits a long lag period, and snow and ice melting requires a continuous accumulation of heat in the surrounding area, thus leading to lower air temperatures in the region [58].

4.3. Lagged Effects of Drought Dominate Grassland WUE in China

This study found that approximately 91.3% of grassland’s WUE response to drought was characterized by significantly larger lagged effects compared to cumulative effects, based on the analysis of drought’s impact on cumulative and lagged effects, as well as a comparison of the Δ R m a x coefficients. Furthermore, this study concluded that lagged effects predominantly influenced the response of WUE to drought across various grassland types. However, Wei et al. [16] found a predominance of cumulative effects of drought on global grassland GPP, which contrasts with our findings. This discrepancy may stem from our study’s greater focus on the impact of soil moisture evaporation on grassland growth. Drought stress not only inhibits grassland growth but also increases soil moisture evaporation. Research has demonstrated that soil moisture significantly influences 70% of vegetation growth, with the most pronounced effects observed in semi-arid regions [63]. Wang et al. [28] highlighted the importance of fully understanding ecosystem responses to drought for the efficient management and conservation of grassland water resources. This understanding also is essential for addressing future extreme alterations in the climate. The generation of drought lagged effects involves both biotic and abiotic factors [64], which may include plant physiological processes, soil water redistribution, changes in ecosystem structure, and influences such as climate and topography. As the climate continued to warm, resulting in drier conditions, soil moisture migrated to deeper soil layers or root zones, and as the drought ended, it took time for soil moisture to be redistributed to the surface layer of rooted soils where vegetation is located, which subsequently delayed the recovery of vegetation WUE [65]. Moreover, drought has intensified soil evaporation, causing a notable decline in moisture levels and essential nutrients such as N, P, and K, which adversely affect plant growth [66]. Therefore, accounting for the lagged effects in past studies of drought on grassland WUE is crucial. This approach deepens insights into the linkage between climate change and grassland WUE, offering valuable scientific understanding for drought risk mitigation in grassland ecosystems.

4.4. Uncertainty Analysis

Firstly, this study utilized GLC2000 land cover data due to the absence of recent and detailed grassland classification data in China. Grasslands have undergone changes over time, which may introduce errors into the results. Future research should prioritize high-precision land cover classification. In this study, only the effect of drought on grass WUE was considered, and no other factors were taken into account. In future research, the impact of multiple factors (such as soil properties, land use, human activities, etc.) on grassland WUE will be considered, and the pathways of these influencing factors will be explored to better balance the growth demands of grasslands with the sustainable use of water resources. Finally, the latest WUE data for grasslands cannot be studied due to the data source covering the period from 1982 to 2018, which presents challenges for current grassland resource efficiency analysis. Therefore, future research should use field sampling data to model GPP and ET to study the latest grassland WUE data, providing the scientific foundation for optimizing grassland resource use.

5. Conclusions

This study quantified the cumulative and lagged effects of drought on grassland WUE across different grassland types, assessing the spatial-temporal variations, altitudinal gradients, and AI zones. It also identified the dominant means by which drought affected grassland WUE. The key findings of this study are as follows:
(1)
Temporal analysis indicated that the annual WUE of grassland showed an increasing trend (0.0033 gC·kg−1H2O·a−1), with all grassland types showing upward trends, particularly MD, which increased the fastest. Spatially, WUE values gradually increased from west to east, decreasing with altitude;
(2)
The average cumulative and lagged effects of drought on grassland WUE lasted 5.2 and 6.1 months, respectively. For different grassland types, cumulative effects lasted 5.6 months for ASM, SG, and DG, while ASPG had a lagged time of 9 months. Both cumulative and lagged effects’ duration increased with altitude. The cumulative and lagged effects of drought on WUE across different grassland types under different AI zones showed a similar trend to that of grasslands as a whole;
(3)
We found that the response of WUE to drought was dominated by lagged effects (more than 95%) in Chinese grasslands and different grassland types except for ASM.
This study emphasizes the importance of future assessments on the significant impact of lagged drought effects on grassland WUE, providing a vital theoretical foundation for understanding climate change’s effects on ecosystem carbon and water cycles.

Author Contributions

Conceptualization, Methodology, Software, Writing—review and editing, M.X.; Methodology, Resources, and Supervision, L.L.; Writing—review and editing, Visualization, Supervision, Funding acquisition, J.Z.; Conceptualization and Software, X.W.; Conceptualization and Investigation, W.L. All authors have read and agreed to the published version of the manuscript.

Funding

This study was supported by the Xinjiang “Tianshan Yingcai” cultivation plan (2023SNGGGGCC004) and the Key Laboratory of Xinjiang Science and Technology Department (No. 2022D04009).

Data Availability Statement

NIRv-GPP is sourced from https://doi.org/10.6084/m9.figshare.12981977.v2 (accessed on 15 July 2024). PEW-ET is sourced from https://doi.org/10.11888/Terre.tpdc.272874 (accessed on 13 July 2024). SPEIbase v2.8 is sourced from http://hdl.handle.net/10261/288226 (accessed on 20 June 2024). AI data are sourced from https://doi.org/10.11888/Atmos.tpdc.300560 (accessed on 14 March 2024). GLC2000 is sourced from https://forobs.jrc.ec.europa.eu (accessed on 3 January 2024). ASTER GDEM V3 is sourced from https://www.gscloud.cn/sources/accessdata/310?pid=302 (accessed on 13 July 2024). The observed dataset is sourced from https://www.nesdc.org.cn/sdo/detail?id=5fa53684042ebb70d0c833ff&subjectCode=5fa53cea042ebb70d0c8340f (accessed on 13 July 2024).

Acknowledgments

Thanks for the support and help from the Xiniang Uyqur Autonomous Region Grassland General Station and the industry-University-Research Joint Training Base.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Song, X.; Chen, H.; Chen, T.; Huang, Q.; Deng, S.; Yang, N. Spatial and temporal variations of spring drought in Southwest China and its possible teleconnection with the global climate events. J. Hydrol. Reg. Stud. 2024, 51, 101655. [Google Scholar] [CrossRef]
  2. Zhang, T.; Quan, W.; Tian, J.; Li, J.; Feng, P. Spatial and temporal variations of ecosystem water use efficiency and its response to soil moisture drought in a water-limited watershed of northern China. J. Environ. Manag. 2024, 355, 120251. [Google Scholar] [CrossRef] [PubMed]
  3. Liu, W.; Sun, F.; Lim, W.H.; Zhang, J.; Wang, H.; Shiogama, H.; Zhang, Y. Global drought and severe drought-affected populations in 1.5 and 2 °C warmer worlds. Earth Syst. Dyn. 2018, 9, 267–283. [Google Scholar] [CrossRef]
  4. Naeem, S.; Zhang, Y.; Zhang, X.; Rehman, A.U.; Tang, Z.; Xu, Z.; Li, C.; Azeem, T. Recent change in ecosystem water use efficiency in China mainly dominated by vegetation greening and increased CO2. Remote Sens. Environ. 2023, 298, 113811. [Google Scholar] [CrossRef]
  5. Yang, S.; Zhang, J.; Han, J.; Wang, J.; Zhang, S.; Bai, Y.; Cao, D.; Xun, L.; Zheng, M.; Chen, H.; et al. Evaluating global ecosystem water use efficiency response to drought based on multi-model analysis. Sci. Total Environ. 2021, 778, 146356. [Google Scholar] [CrossRef]
  6. Hou, Q.; Pei, T.; Yu, X.; Chen, Y.; Ji, Z.; Xie, B. The seasonal response of vegetation water use efficiency to temperature and precipitation in the Loess Plateau, China. Glob. Ecol. Conserv. 2022, 33, e01984. [Google Scholar] [CrossRef]
  7. Chen, T.; Wang, Y.; Peng, L. Advanced time-lagged effects of drought on global vegetation growth and its social risk in the 21st century. J. Environ. Manag. 2023, 347, 119253. [Google Scholar] [CrossRef]
  8. Yao, Y.; Ren, H. Estimation of grassland aboveground biomass in northern China based on topography-climate-remote sensing data. Ecol. Indic. 2024, 165, 112230. [Google Scholar] [CrossRef]
  9. Zhang, W.; Luo, G.; Hamdi, R.; Ma, X.; Termonia, P.; De Maeyer, P. Drought changes the dominant water stress on the grassland and forest production in the northern hemisphere. Agric. For. Meteorol. 2024, 345, 109831. [Google Scholar] [CrossRef]
  10. Li, M.; Wu, P.; Ma, Z.; Liu, J. Climate change reshapes bioclimatic environments in China’s dry–wet transition zones. J. Hydrol. 2024, 634, 131122. [Google Scholar] [CrossRef]
  11. Huang, M.; Zhai, P.; Piao, S. Divergent responses of ecosystem water use efficiency to drought timing over Northern Eurasia. Environ. Res. Lett. 2021, 16, 045016. [Google Scholar] [CrossRef]
  12. Liu, L.; Zheng, J.; Guan, J.; Han, W.; Liu, Y. Grassland cover dynamics and their relationship with climatic factors in China from 1982 to 2021. Sci. Total Environ. 2023, 905, 167067. [Google Scholar] [CrossRef] [PubMed]
  13. Xu, S.; Wang, Y.; Liu, Y.; Li, J.; Qian, K.; Yang, X.; Ma, X. Evaluating the cumulative and time-lag effects of vegetation response to drought in Central Asia under changing environments. J. Hydrol. 2023, 627, 130455. [Google Scholar] [CrossRef]
  14. Wang, S.; Zhou, Q.; Wu, Y.; Wang, F.; Li, M.; Zhang, Y. Drought lag and its cumulative effects on vegetation dynamics and response to atmospheric circulation factors in Yinshanbeilu, Inner Mongolia. Glob. Ecol. Conserv. 2024, 54, e03087. [Google Scholar] [CrossRef]
  15. Li, P.; Wang, J.; Liu, M.; Xue, Z.; Bagherzadeh, A.; Liu, M. Spatio-temporal variation characteristics of NDVI and its response to climate on the Loess Plateau from 1985 to 2015. Catena 2021, 203, 105331. [Google Scholar] [CrossRef]
  16. Wei, X.; He, W.; Zhou, Y.; Ju, W.; Xiao, J.; Li, X.; Liu, Y.; Xu, S.; Bi, W.; Zhang, X.; et al. Global assessment of lagged and cumulative effects of drought on grassland gross primary production. Ecol. Indic. 2022, 136, 108646. [Google Scholar] [CrossRef]
  17. Zhan, C.; Liang, C.; Zhao, L.; Jiang, S.; Niu, K.; Zhang, Y. Drought-related cumulative and time-lag effects on vegetation dynamics across the Yellow River Basin, China. Ecol. Indic. 2022, 143, 109409. [Google Scholar] [CrossRef]
  18. Ji, Y.; Li, Y.; Yao, N.; Biswas, A.; Zou, Y.; Meng, Q.; Liu, F. The lagged effect and impact of soil moisture drought on terrestrial ecosystem water use efficiency. Ecol. Indic. 2021, 133, 108349. [Google Scholar] [CrossRef]
  19. Su, J.; Fan, L.; Yuan, Z.; Wang, Z.; Wang, Z. Quantifying the drought sensitivity of grassland under different climate zones in Northwest China. Sci. Total Environ. 2024, 910, 168688. [Google Scholar] [CrossRef]
  20. Zhao, Y.; Xiong, L.; Yin, J.; Zha, X.; Li, W.; Han, Y. Understanding the effects of flash drought on vegetation photosynthesis and potential drivers over China. Sci. Total Environ. 2024, 931, 172926. [Google Scholar] [CrossRef]
  21. Huang, F.; Liu, L.; Gao, J.; Yin, Z.; Zhang, Y.; Jiang, Y.; Zuo, L.; Fang, W. Effects of extreme drought events on vegetation activity from the perspectives of meteorological and soil droughts in southwestern China. Sci. Total Environ. 2023, 903, 166562. [Google Scholar] [CrossRef] [PubMed]
  22. Kwon, H.; Law, B.E.; Thomas, C.K.; Johnson, B.G. The influence of hydrological variability on inherent water use efficiency in forests of contrasting composition, age, and precipitation regimes in the Pacific Northwest. Agric. For. Meteorol. 2018, 249, 488–500. [Google Scholar] [CrossRef]
  23. Hoover, D.L.; Pfennigwerth, A.A.; Duniway, M.C. Drought resistance and resilience: The role of soil moisture–plant interactions and legacies in a dryland ecosystem. J. Ecol. 2021, 109, 3280–3294. [Google Scholar] [CrossRef]
  24. Liu, L.; Guan, J.; Zheng, J.; Wang, Y.; Han, W.; Liu, Y. Cumulative effects of drought have an impact on net primary productivity stability in Central Asian grasslands. J. Environ. Manag. 2023, 344, 118734. [Google Scholar] [CrossRef]
  25. Wang, Z.; Cao, S.; Cao, G. The Effect of Vegetative Coverage and Altitude on the Vegetation Water Consumption in the Alpine Inland River Basin of the Northeastern Qinghai–Tibet Plateau. Water 2022, 14, 1113. [Google Scholar] [CrossRef]
  26. Wang, S.; Zhang, Q.; Yue, P.; Wang, J. Effects of evapotranspiration and precipitation on dryness/wetness changes in China. Theor. Appl. Climatol. 2020, 142, 1027–1038. [Google Scholar] [CrossRef]
  27. Bartholomé, E.; Belward, A.S. GLC2000: A new approach to global land cover mapping from Earth observation data. Int. J. Remote Sens. 2005, 26, 1959–1977. [Google Scholar] [CrossRef]
  28. Wang, M.; Ding, Z.; Wu, C.; Song, L.; Ma, M.; Yu, P.; Lu, B.; Tang, X. Divergent responses of ecosystem water-use efficiency to extreme seasonal droughts in Southwest China. Sci. Total Environ. 2021, 760, 143427. [Google Scholar] [CrossRef]
  29. Fu, J.; Wang, W.; Shao, Q.; Xing, W.; Cao, M.; Wei, J.; Chen, Z.; Nie, W. Improved global evapotranspiration estimates using proportionality hyp othesis-based water balance constraints. Remote Sens. Environ. 2022, 279, 113140. [Google Scholar] [CrossRef]
  30. Vicente-Serrano, S.M.; Beguería, S.; López-Moreno, J.I. A Multiscalar Drought Index Sensitive to Global Warming: The Standardized Precipitation Evapotranspiration Index. J. Clim. 2010, 23, 1696–1718. [Google Scholar] [CrossRef]
  31. Beguería, S.; Serrano, S.M.V.; Reig-Gracia, F.; Garcés, B.L. SPEIbase, v.2.8 [Dataset]. 2023. Available online: http://hdl.handle.net/10261/288226 (accessed on 20 June 2024).
  32. Kimura, R.; Moriyama, M. Recent Trends of Annual Aridity Indices and Classification of Arid Regions with Satellite-Based Aridity Indices. Remote Sens. Earth Syst. Sci. 2019, 2, 88–95. [Google Scholar] [CrossRef]
  33. Peng, S. 1 km Annual Arid Index Dataset in China (1901–2022). 2023. Available online: https://data.tpdc.ac.cn/en/data/b7a5c363-3f63-4921-b657-8a4229476ec1 (accessed on 14 March 2024).
  34. Xue, Y.; Bai, X.; Zhao, C.; Tan, Q.; Li, Y.; Luo, G.; Wu, L.; Chen, F.; Li, C.; Ran, C.; et al. Spring photosynthetic phenology of Chinese vegetation in response to climate change and its impact on net primary productivity. Agric. For. Meteorol. 2023, 342, 109734. [Google Scholar] [CrossRef]
  35. Zarei, A.; Chemura, A.; Gleixner, S.; Hoff, H. Evaluating the grassland NPP dynamics in response to climate change in Tanzania. Ecol. Indic. 2021, 125, 107600. [Google Scholar] [CrossRef]
  36. Zheng, H.; Yu, G.R.; Zhu, X.J.; Wang, X.F.; Zhang, L.M.; Chen, Z.; Sun, X.M.; He, H.L.; Su, W.; Wang, Y.F.; et al. Datasets on actual evapotranspiration and water use efficiency of typical terrestrial ecosystems in China, 2000–2010. Chin. Sci. Data 2019, 4, 55–68. [Google Scholar] [CrossRef]
  37. Liu, Y.; Xiao, J.; Ju, W.; Zhou, Y.; Wang, S.; Wu, X. Water use efficiency of China’s terrestrial ecosystems and responses to drought. Sci. Rep. 2015, 5, 13799. [Google Scholar] [CrossRef] [PubMed]
  38. Guo, X.; Arshad, M.U.; Zhao, Y.; Gong, Y.; Li, H. Effects of climate change and grazing intensity on grassland productivity-A case study of Inner Mongolia, China. Heliyon 2023, 9, e17814. [Google Scholar] [CrossRef]
  39. Sen, P.K. Estimates of the Regression Coefficient Based on Kendall’s Tau. J. Am. Stat. Assoc. 1968, 63, 1379–1389. [Google Scholar] [CrossRef]
  40. Wang, M.; Zhao, J.; Zhang, H.; Zhang, Z.; Guo, X.; Zhang, T.; Wu, R. Detecting the response characteristics and thresholds of grassland spring phenology to climatic factors in the Mongolian Plateau. Ecol. Indic. 2023, 153, 110440. [Google Scholar] [CrossRef]
  41. Kendall, M.G. Rank Correlation Methods; APA: Washington, DA, USA, 1949. [Google Scholar]
  42. Mann, H.B. Nonparametric Tests Against Trend. Econom. J. Econom. Soc. 1945, 13, 245–259. [Google Scholar] [CrossRef]
  43. Gao, X.; Sun, S.; Meng, P.; Cai, J.; Pei, S.; Huang, H.; Zhang, J. Carbon fluxes and water-use efficiency in a Pinus tabuliformis plantation in Northeast China and their relationship to drought. Sci. Total Environ. 2024, 946, 174258. [Google Scholar] [CrossRef]
  44. Strmberg, C.; Staver, A.J.S. The history and challenge of grassy biomes. Science 2022, 377, 592–593. [Google Scholar] [CrossRef] [PubMed]
  45. Zhang, Z.; Zhang, F.; Wang, L.; Lin, A.; Zhao, L. Biophysical climate impact of forests with different age classes in mid- and high-latitude North America. For. Ecol. Manag. 2021, 494, 119327. [Google Scholar] [CrossRef]
  46. Harrington, J.A., Jr.; Harman, J.R. Climate and Vegetation in Central North America: Natural Patterns and Human Alterations. Great Plains Q. 1991, 11, 103–112. [Google Scholar]
  47. Ippolito, J.A.; Blecker, S.W.; Freeman, C.L.; McCulley, R.L.; Blair, J.M.; Kelly, E.F. Phosphorus biogeochemistry across a precipitation gradient in grasslands of central North America. J. Arid Environ. 2010, 74, 954–961. [Google Scholar] [CrossRef]
  48. Lin, S.; Wang, G.; Hu, Z.; Sun, X.; Song, C.; Huang, K.; Sun, J.; Yang, Y. Contrasting response of growing season water use efficiency to precipitation changes between alpine meadows and alpine steppes over the Tibetan Plateau. Agric. Water Manag. 2023, 289, 108571. [Google Scholar] [CrossRef]
  49. Wu, W.; Sun, R.; Liu, X.; Li, L.; Qi, M.; Zhang, F.; Zhou, G.; Zhu, J.; Xu, Z. Driving mechanisms of community biomass allocation along environmental gradients in different grasslands in China. Ecol. Indic. 2024, 160, 111886. [Google Scholar] [CrossRef]
  50. Liu, Y.; Ding, Z.; Chen, Y.; Yan, F.; Yu, P.; Man, W.; Liu, M.; Li, H.; Tang, X. Restored vegetation is more resistant to extreme drought events than natural vegetation in Southwest China. Sci. Total Environ. 2023, 866, 161250. [Google Scholar] [CrossRef]
  51. Du, Q.; Guan, Q.; Sun, Y.; Wang, Q.; Zhang, J.; Xiao, X.; Liu, H.; Yang, X.; Zhang, E. Legacy effects of extreme drought and wetness events on mountain grassland ecosystems and their elevation dependence. J. Hydrol. 2024, 630, 130757. [Google Scholar] [CrossRef]
  52. Zhang, Z.; Ju, W.; Zhou, Y.; Li, X. Revisiting the cumulative effects of drought on global gross primary productivity based on new long-term series data (1982–2018). Glob. Change Biol. 2022, 28, 3620–3635. [Google Scholar] [CrossRef]
  53. Barnes, M.L.; Moran, M.S.; Scott, R.L.; Kolb, T.E.; Ponce-Campos, G.E.; Moore, D.J.P.; Ross, M.A.; Mitra, B.; Dore, S. Vegetation productivity responds to sub-annual climate conditions across semiarid biomes. Ecosphere 2016, 7, e01339. [Google Scholar] [CrossRef]
  54. Tang, J.; Niu, B.; Hu, Z.; Zhang, X. Increasing susceptibility and shortening response time of vegetation productivity to drought from 2001 to 2021. Agric. For. Meteorol. 2024, 352, 110025. [Google Scholar] [CrossRef]
  55. Bai, Y.; Li, S. Growth peak of vegetation and its response to drought on the Mongolian Plateau. Ecol. Indic. 2022, 141, 109150. [Google Scholar] [CrossRef]
  56. Xu, H.-J.; Wang, X.-P.; Zhao, C.-Y. Drought sensitivity of vegetation photosynthesis along the aridity gradient in northern China. Int. J. Appl. Earth Obs. Geoinf. 2021, 102, 102418. [Google Scholar] [CrossRef]
  57. Tang, X.; Li, H.; Desai, A.R.; Nagy, Z.; Luo, J.; Kolb, T.E.; Olioso, A.; Xu, X.; Yao, L.; Kutsch, W.; et al. How is water-use efficiency of terrestrial ecosystems distributed and changing on Earth? Sci. Rep. 2014, 4, 7483. [Google Scholar] [CrossRef] [PubMed]
  58. Li, Q.; Gao, X.; Li, J.; Yan, A.; Chang, S.; Song, X.; Lo, K. Nonlinear time effects of vegetation response to climate change: Evidence from Qilian Mountain National Park in China. Sci. Total Environ. 2024, 933, 173149. [Google Scholar] [CrossRef] [PubMed]
  59. Li, W.; Shen, Y.; Wang, G.; Ma, H.; Yang, Y.; Li, G.; Huo, X.; Liu, Z. Plant species diversity and functional diversity relations in the degradation process of desert steppe in an arid area of northwest China. J. Environ. Manag. 2024, 365, 121534. [Google Scholar] [CrossRef] [PubMed]
  60. Zhang, Q.; Wang, Z.; Xia, S.; Zhang, G.; Li, S.; Yu, D.; Yu, X. Hydrologic-induced concentrated soil nutrients and improved plant growth increased carbon storage in a floodplain wetland over wet-dry alternating zones. Sci. Total Environ. 2022, 822, 153512. [Google Scholar] [CrossRef]
  61. Alsafadi, K.; Bashir, B.; Mohammed, S.; Abdo, H.G.; Mokhtar, A.; Alsalman, A.; Cao, W. Response of Ecosystem Carbon–Water Fluxes to Extreme Drought in West Asia. Remote Sens. 2024, 16, 1179. [Google Scholar] [CrossRef]
  62. Zhang, J.; Fang, S.; Liu, H. Estimation of alpine grassland above-ground biomass and its response to climate on the Qinghai-Tibet Plateau during 2001 to 2019. Glob. Ecol. Conserv. 2022, 35, e02065. [Google Scholar] [CrossRef]
  63. Liu, L.; Gudmundsson, L.; Hauser, M.; Qin, D.; Li, S.; Seneviratne, S.I. Soil moisture dominates dryness stress on ecosystem production globally. Nat. Commun. 2020, 11, 4892. [Google Scholar] [CrossRef]
  64. Zhao, J.; Feng, H.; Xu, T.; Xiao, J.; Guerrieri, R.; Liu, S.; Wu, X.; He, X.; He, X. Physiological and environmental control on ecosystem water use efficiency in response to drought across the northern hemisphere. Sci. Total Environ. 2021, 758, 143599. [Google Scholar] [CrossRef]
  65. Domec, J.C.; King, J.S.; Noormets, A.; Treasure, E.; Gavazzi, M.J.; Sun, G.; McNulty, S.G. Hydraulic redistribution of soil water by roots affects whole-stand evapotranspiration and net ecosystem carbon exchange. New Phytol. 2010, 187, 171–183. [Google Scholar] [CrossRef]
  66. Szejgis, J.; Nielsen, U.N.; Dijkstra, F.A.; Carrillo, Y. Prolonged drought moderates flood effects on soil nutrient pools across a rainfall gradient. Soil Biol. Biochem. 2024, 193, 109404. [Google Scholar] [CrossRef]
Water 17 01134 g001
Figure 1. Elevation spatial distribution (a) and grassland type (b).
Figure 1. Elevation spatial distribution (a) and grassland type (b).
Water 17 01134 g001
Water 17 01134 g002
Figure 2. Evaluation of observed ET and PEW-ET accuracy (a); observed GPP data and NIRv-GPP accuracy evaluation (b).
Figure 2. Evaluation of observed ET and PEW-ET accuracy (a); observed GPP data and NIRv-GPP accuracy evaluation (b).
Water 17 01134 g002
Water 17 01134 g003
Figure 3. Interannual changes in WUE of grassland and different grassland types in China from 1982 to 2018 (total is grassland overall; ASM is alpine and subalpine meadow; SG is slope grassland; PG is plain grassland; DG is desert grassland; MD is meadow; ASPG is alpine and subalpine plain grassland).
Figure 3. Interannual changes in WUE of grassland and different grassland types in China from 1982 to 2018 (total is grassland overall; ASM is alpine and subalpine meadow; SG is slope grassland; PG is plain grassland; DG is desert grassland; MD is meadow; ASPG is alpine and subalpine plain grassland).
Water 17 01134 g003
Water 17 01134 g004
Figure 4. Spatial distribution of average WUE of grassland (a) and average WUE of different grassland type (b).
Figure 4. Spatial distribution of average WUE of grassland (a) and average WUE of different grassland type (b).
Water 17 01134 g004
Water 17 01134 g005
Figure 5. The mean annual WUE of grassland and different grassland types under different elevation gradients. (a–g) represent the total, ASM, SG, PG, DG, MD, and ASPG, respectively, and the horizontal and vertical coordinates of the graphs denote the mean annual WUE values under the 500 m elevation intervals. The mean annual WUE values of grassland (h) and WUE values of different grassland types (i) under different AI zones.
Figure 5. The mean annual WUE of grassland and different grassland types under different elevation gradients. (a–g) represent the total, ASM, SG, PG, DG, MD, and ASPG, respectively, and the horizontal and vertical coordinates of the graphs denote the mean annual WUE values under the 500 m elevation intervals. The mean annual WUE values of grassland (h) and WUE values of different grassland types (i) under different AI zones.
Water 17 01134 g005
Water 17 01134 g006
Figure 6. Spatial distribution of grassland WUE trends (a) and area proportion of trend in different grassland types (b).
Figure 6. Spatial distribution of grassland WUE trends (a) and area proportion of trend in different grassland types (b).
Water 17 01134 g006
Water 17 01134 g007
Figure 7. Spatial distribution of cumulative effect (a) and cumulative months (b). Mean cumulative effect values (c) and cumulative month area proportion (d) for different grass types. The blue × indicates a positive correlation with significance ( R m a x _ c u m > 0), while the red + indicates a negative correlation with significance ( R m a x _ c u m < 0).
Figure 7. Spatial distribution of cumulative effect (a) and cumulative months (b). Mean cumulative effect values (c) and cumulative month area proportion (d) for different grass types. The blue × indicates a positive correlation with significance ( R m a x _ c u m > 0), while the red + indicates a negative correlation with significance ( R m a x _ c u m < 0).
Water 17 01134 g007
Water 17 01134 g008
Figure 8. Mean cumulative months for grassland and different grassland types under different elevation gradients. (a–g) represent the total, ASM, SG, PG, DG, MD, and ASPG, respectively, and the horizontal and vertical coordinates of the graphs denote the mean cumulative months under the 500 m elevation intervals. The mean cumulative months of grassland (h) and cumulative months of different grassland types (i) under different AI zones.
Figure 8. Mean cumulative months for grassland and different grassland types under different elevation gradients. (a–g) represent the total, ASM, SG, PG, DG, MD, and ASPG, respectively, and the horizontal and vertical coordinates of the graphs denote the mean cumulative months under the 500 m elevation intervals. The mean cumulative months of grassland (h) and cumulative months of different grassland types (i) under different AI zones.
Water 17 01134 g008
Water 17 01134 g009
Figure 9. Spatial distribution of lagged effect (a) and lagged months (b). Mean lagged effect values (c) and lagged month area proportion (d) for different grass types. The blue × indicates a positive correlation with significance ( R m a x _ l a g > 0), while the red + indicates a negative correlation with significance ( R m a x _ l a g < 0).
Figure 9. Spatial distribution of lagged effect (a) and lagged months (b). Mean lagged effect values (c) and lagged month area proportion (d) for different grass types. The blue × indicates a positive correlation with significance ( R m a x _ l a g > 0), while the red + indicates a negative correlation with significance ( R m a x _ l a g < 0).
Water 17 01134 g009
Water 17 01134 g010
Figure 10. Mean lagged months for grassland and different grassland types under different elevation gradients. (a–g) represent the total, ASM, SG, PG, DG, MD, and ASPG, respectively, and the horizontal and vertical coordinates of the graphs denote the lagged months under the 500 m elevation intervals. The mean lagged months of grassland (h) and lagged months of different grassland types (i) under different AI zones.
Figure 10. Mean lagged months for grassland and different grassland types under different elevation gradients. (a–g) represent the total, ASM, SG, PG, DG, MD, and ASPG, respectively, and the horizontal and vertical coordinates of the graphs denote the lagged months under the 500 m elevation intervals. The mean lagged months of grassland (h) and lagged months of different grassland types (i) under different AI zones.
Water 17 01134 g010
Water 17 01134 g011
Figure 11. Dominant factors for cumulative and lagged effects of drought (a), mean absolute values of R m a x _ l a g and R m a x _ c u m (b), and area proportion of dominant factors (c) for different grassland types.
Figure 11. Dominant factors for cumulative and lagged effects of drought (a), mean absolute values of R m a x _ l a g and R m a x _ c u m (b), and area proportion of dominant factors (c) for different grassland types.
Water 17 01134 g011
Table 1. The introduction of GPP, ET, SPEI, AI, grassland type data, and observational data used in this study.
Table 1. The introduction of GPP, ET, SPEI, AI, grassland type data, and observational data used in this study.
Data TypeNameTime ScaleSpatial-Temporal ResolutionData Website
NIRv-GPP datasetGross Primary Production (GPP)1982 − 20180.05° × 0.05°, Monthlyhttps://doi.org/10.6084/m9.figshare.12981977.v2 (accessed on 15 July 2024)
PEW-ET datasetEvapotranspiration (ET)1982 − 20180.1° × 0.1°, Monthlyhttps://doi.org/10.11888/Terre.tpdc.272874 (accessed on 13 July 2024)
Drought datasetStandardized Precipitation Evapotranspiration Index (SPEI)1981 − 20180.5° × 0.5°, Monthlyhttp://hdl.handle.net/10261/288226 (accessed on 20 June 2024)
Wet and dry datasetAridity index (AI)1982 − 20180.00833° × 0.00833°, Yearlyhttps://doi.org/10.11888/Atmos.tpdc.300560 (accessed on 14 March 2024)
Grassland datasetGlobal Land Cover 200020001 km × 1 km, Yearlyhttps://forobs.jrc.ec.europa.eu (accessed on 3 January 2024)
Elevation datasetASTER GDEM V3 30 m × 30 mhttps://www.gscloud.cn/sources/accessdata/310?pid=302 (accessed on 13 July 2024)
China grassland carbon fluxes datasetObserved ET and GPP2000–2010Yearlyhttps://www.nesdc.org.cn/sdo/detail?id=5fa53684042ebb70d0c833ff&subjectCode=5fa53cea042ebb70d0c8340f (accessed on 13 July 2024)
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Xing, M.; Liu, L.; Zheng, J.; Wang, X.; Li, W. Spatio-Temporal Variation Characteristics of Grassland Water Use Efficiency and Its Response to Drought in China. Water 2025, 17, 1134. https://doi.org/10.3390/w17081134

AMA Style

Xing M, Liu L, Zheng J, Wang X, Li W. Spatio-Temporal Variation Characteristics of Grassland Water Use Efficiency and Its Response to Drought in China. Water. 2025; 17(8):1134. https://doi.org/10.3390/w17081134

Chicago/Turabian Style

Xing, Mengxiang, Liang Liu, Jianghua Zheng, Xinwei Wang, and Wei Li. 2025. "Spatio-Temporal Variation Characteristics of Grassland Water Use Efficiency and Its Response to Drought in China" Water 17, no. 8: 1134. https://doi.org/10.3390/w17081134

APA Style

Xing, M., Liu, L., Zheng, J., Wang, X., & Li, W. (2025). Spatio-Temporal Variation Characteristics of Grassland Water Use Efficiency and Its Response to Drought in China. Water, 17(8), 1134. https://doi.org/10.3390/w17081134

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop