Introduction

Vegetation plays an important regulatory role in the water, material, and energy cycles of terrestrial ecosystems1,2,3. The amount of fixed CO2 or synthesized dry matter consumed per unit of water in vegetation photosynthesis is called water use efficiency (WUE)4. WUE can characterize the strength of carbon water coupling and also reflect the impact of climate change on vegetation water use strategies5. At the ecosystem scale, WUE is usually represented by the ratio of total primary productivity (GPP) to evapotranspiration (ET)6, therefore, GPP and ET are key factors affecting WUE. Previous studies have shown that WUE is significantly influenced by climate factors such as temperature, precipitation, and carbon dioxide concentration, which can lead to changes of vegetation community outcomes and distribution patterns7. The differences and specificity of the regional environment are the key factors that contribute to the spatial heterogeneity of vegetation productivity and ET8. WUE is an important indicator that connects plant water consumption with the fixation of organic carbon. The patterns of WUE across different regions can illustrate the balance between water use and carbon fixation in various ecosystems. Additionally, these patterns can reveal how regional environmental conditions and climate change influence this relationship.Therefore, studying the spatial distribution patterns of WUE across different regional vegetation ecosystems is essential for enhancing our understanding of water dissipation and organic carbon fixation in different vegetation ecosystems.

Precipitation and temperature are the two most studied climate factors that affect vegetation change and are crucial for plant photosynthesis and respiration9,10,11. Within a certain temperature range, the increase in temperature leads to a greater increase in photosynthetic carbon fixation than the increase in ET rate, resulting in an improvement in the WUE of the ecosystem. If the temperature exceeds a certain level, it will cause a decrease in the WUE of the ecosystem12. Precipitation change is another important aspect of climate change, and current research indicates that the response of ecosystem WUE to precipitation change depends on the water conditions of the ecosystem. The WUE of wetland and farmland ecosystems decreases with increasing precipitation, while the WUE of forest and grassland ecosystems increases with increasing precipitation13. Due to differences in regional climate conditions, vegetation growth status, and other factors, there are differences in the spatiotemporal changes of WUE and the response to temperature and precipitation14. Global warming leads to frequent extreme climate events15, and extreme climate variables are more sensitive to climate change and significantly correlated with vegetation growth16,17,18. However, the response mechanism of WUE to atmospheric temperature and precipitation in different regions is not yet clear, especially on the long-term scale. Further research is needed on the adaptive strategies and driving directions of WUE to global extreme weather change. Therefore, investigating the response relationship of WUE in different vegetation ecosystems to meteorological factors will provide a reference for rational planning of regional water resources and sustainable development of the ecological environment.

The time effect is an undeniable phenomenon in the interaction between climate and vegetation, including time delay and accumulation19,20,21. Climate indices, including precipitation and temperature, have a time marked effect on vegetation change22,23, and the time lag effect of extreme climate variables on vegetation has also been confirmed24,25. Wu et al. (2015)26 found that when considering time-delay effects, climate factors can explain 64% of global vegetation changes, which is 11% higher than when ignoring time-delay effects. In terms of the cumulative effect of climate on vegetation growth, Zhang et al. (2018)27 pointed out that the accumulated temperature can significantly promote the growth and development of plant leaves. Evans and Geerken (2004)28 found that precipitation has a maximum cumulative effect on vegetation NDVI for 7 months when quantifying the impact of climate and human activities on grassland degradation in arid areas. However, the time lag or cumulative effects of climate do not exist independently, and their combined effects on vegetation growth need to be considered more comprehensively to explain the temporal effects of climate on vegetation growth29,30,31.

The Shiyang River Basin (SRB) is situated in the mid-latitude Eurasian hinterland in northwestern China. The environment is fragile, with a dry climate and severe interference from human activities. It is a typical ecologically sensitive area, and the severe habitat characteristics exacerbate the contradiction between the ecological environment and socio-economic development32,33. In recent years, the climate in the SRB has significantly warmed and humidified34, and the atmospheric demand for water has increased, promoting the ET process of the ecosystem and leading to a decrease in WUE. The SRB is an inland river basin, and there are significant differences in the ecological environment formation foundations among different regions. Due to the influence of different climates, terrains, and water resources, different ecosystems are formed within the basin, which are strongly intervened by human activities, leading to the transformation of some natural ecosystems into artificial ecosystems. Therefore, the division of functional zones is conducive to the effective management and protection of the regional ecological environment, the utilization of regional resource advantages, the stability of basin ecosystems, and the improvement of ecological environment quality. However, the temporal and cumulative effects of WUE spatial distribution characteristics of different vegetation ecosystems in functional areas under the background of climate warming and humidification on climate factors have not been elaborated in detail. Therefore, the present study will attempt to address the following two questions by using satellite and ground-based observation data from 2001 to 2020: (1) investigate the spatiotemporal dynamic changes of WUE in the SRB; (2) clarify the response mechanism of WUE in different ecological functional areas to temperature and precipitation time effects during the process of basin humidification.

Materials and methods

Study area

The SRB is situated in the transition zone where the Loess Plateau, the Tibetan Plateau, and the Mongolian-Xinjiang Plateau converge. This region is characterized by diverse landform types and significant relative elevation differences. Deep within the Eurasian continent, the basin is influenced by the East Asian monsoon, the South Asian monsoon, and the westerlies. It experiences perennial drought and little rainfall, abundant sunshine, significant diurnal temperature differences, and intense evaporation. The winters are cold, summers are hot, and the area is prone to frequent winds, dry hot winds, and severe wind-sand hazards, resulting in a highly complex climate system. Geographically, the basin is bordered by the northern foothills of the Qilian Mountains to the south and adjoins the Tengger Desert and the Badain Jaran Desert to the north. Within the basin, desert areas account for 68%, while the total vegetation cover, including forests, grasslands, and crops, is only 32%. It is a typical climate-sensitive area and an ecologically fragile zone8,33.

The SRB has a large geographical span, complex and diverse landforms, and significant spatial heterogeneity of the ecosystem, resulting in significant differences in vegetation of the same type in the upper, middle, and lower reaches of the basin. (Fig. 1) Therefore, it is necessary to divide the study area into ecological functional zones35,36. Zone I (Qilian Mountains Water Conservation Area): This zone covers the Qilian Mountains area, including parts of Tianzhu, Gulang, Liangzhou, Yongchang County, Sunan, and Shandan Military Horse Farm. This region will be a key area for future farmland reversion to forests and grasslands. It will implement ecological migration in the core area of water conservation forests, reduce the population size in buffer zones and their surrounding areas, and enhance water conservation capabilities.Zone II (Hexi Corridor Oasis Functional Zone): This zone encompasses the Hexi Corridor plain between the Qilian Mountains and the Bei Mountains, involving most parts of Gulang, Liangzhou District, Yongchang, and the area south of the Hongyanshan Reservoir in Minqin. This zone aims to establish a sustainable water resource allocation system that conforms to the characteristics of basin resources and the laws of the market economy. It also aims to build an urban oasis ecological forest network and an ecological protection forest network around the oasis to ensure ecological water use in downstream areas.Zone III (Desert-Oasis Interaction Functional Zone): This zone covers the oasis basin north of the Bei Mountains in the Hexi Corridor, including Jinli District, northern Yongchang, and the Minqin Basin. It is distributed discontinuously and is located in the transitional zone between the oasis and the desert downstream of the basin, directly facing two major deserts.Zone IV (Desert Area): This zone mainly refers to the desert areas surrounding the Hexi Corridor Oasis, Minqin Oasis, Changning Oasis, and the Jinchuan Industrial and Mining Urban Area, including the Bei Mountains of the Hexi Corridor. The main land types are deserts, Gobi, and strongly eroded low mountains. Precipitation in this zone is sparse, and the arid desert ecosystem is composed of xerophytic and extremely xerophytic vegetation.Vegetation in different ecosystems within the basin is significantly affected by natural resources such as climate, topography, and water resources. There are clear differences in carbon accumulation and cycling processes. Tian et al. (2010)13, especially in the significant geographical differentiation characteristics of vegetation in water use strategies and dry matter transformation processes in response to climate change30.

Fig. 1
Fig. 1
Full size image

Land use types and meteorological stations in the SRB.

Data acquisition and processing

Remote sensing data

This study utilized the National Aeronautics and Space Administration (NASA) of the United States (http://earthdata.nasa.gov). The ET (MOD16A2GF) and GPP (MOD17A2HGF) products of the Medium Resolution Imaging Spectrometer (MODIS) have a spatial resolution of 500 m, a temporal resolution of 8 days, and a period of 2000–2020. The GPP data algorithm is based on light utilization efficiency, and detailed calculations can be found in the literature31. The ET data algorithm is derived from the Penman-Monteith equation, which considers three processes: soil surface evaporation, canopy interception of precipitation, and plant transpiration. For more detailed information, please refer to Mu et al. (2011)37.The MODIS data were batch mosaicked and reprojected using the MRT (MODIS Reprojection Tool) software. The reprojected data were then composited monthly and clipped to the SRB. MODIS GPP and ET product data have been validated through flux tower site data in multiple regions around the world, and their accuracy has been confirmed in multiple studies.

Meteorological data

The annual data of temperature and precipitation in the SRB and the surrounding 8 meteorological stations from 2000 to 2020 is sourced from the Gansu Provincial Meteorological Bureau. Use ANUSPLIN meteorological data interpolation software to interpolate and obtain meteorological grid data with the same pixel size and projection as WUE data. Previous studies have shown that ANUSPLIN software has high accuracy and credibility in simulating climate data, and has been widely used in China’s climate spatial simulation38.

WUE calculation

WUE refers to the amount of dry matter fixed per unit mass of water consumed by an ecosystem. This study used the commonly used method for calculating WUE in ecosystem research39, which is the ratio of GPP to total ET40.

$$WUE=\frac{{GPP}}{{ET}}.$$
(1)

In the equation, WUE represents the water use efficiency of the vegetation ecosystem over a unit of time (gC·kg− 1 H2O). GPP denotes the gross primary productivity of the vegetation ecosystem over a unit of time (gC·m− 2), and ET indicates the total evapotranspiration of the vegetation ecosystem over a unit of time (kg H2O·m− 2).

Statistical and analysis

Trend analysis

This study used univariate linear regression analysis to simulate the monthly WUE from 2000 to 2020 based on each pixel and obtained the change trend. The calculation formula is as follows:

$${\theta _{slope}}=\frac{{n \times \left( {\sum\nolimits_{{i=1}}^{n} {i \times {c_i}} } \right) - \left( {\sum\nolimits_{{i=1}}^{n} i } \right)\left( {\sum\nolimits_{{i=1}}^{n} {{c_i}} } \right)}}{{n \times \sum\nolimits_{{i=1}}^{n} {{i^2} - {{\left( {\sum\nolimits_{{i=1}}^{n} i } \right)}^2}} }}.$$
(2)

Where n represents the number of years in the period (n = 21); \(\:{\theta\:}_{slope}\) is the slope of the trend; \(\:{C}_{i}\) for section I WUE for each month of the year. \(\:{\theta\:}_{slope}\) a negative value indicates a downward trend in WUE, \(\:{\theta\:}_{slope}\) a positive value indicates an upward trend in WUE.

Correlation analysis

Pearson correlation analysis was used to calculate the correlation coefficients between monthly WUE, temperature, and precipitation on a pixel-by-pixel basis, using years as time units. The calculation formula is as follows:

$${R_{xy}}=\frac{{\sum\nolimits_{{i=1}}^{n} {\left[ {\left( {{x_i} - \overline {x} } \right)\left( {{y_i} - \overline {y} } \right)} \right]} }}{{\sqrt {\sum\nolimits_{{i=1}}^{n} {\left( {{x_i} - \overline {x} } \right)} } \sum\nolimits_{{n=i}}^{n} {\left( {{y_i} - \overline {y} } \right)} }}.$$
(3)

Where \(\:{R}_{xy}\) represents the correlation coefficient between two variables; \(\:{x}_{i}\) represents the annual GPP/ET/WUE of each month; \(\:{y}_{i}\) represents the temperature and precipitation in the I year of each month; \(\:\overline{x}\) represents the average annual GPP/ET/WUE for each month; \(\:\overline{y}\) represents the average annual temperature and precipitation for each month; I represents the number of samples. \(\:{R}_{xy}\)is the correlation between GPP/ET/WUE and temperature and precipitation; When \(\:{R}_{xy}\) is positive, it indicates a positive correlation between GPP/ET/WUE and temperature and precipitation, and when negative, it indicates a negative correlation between GPP/ET/WUE and temperature and precipitation.

Significance test

The significance represents the level of confidence in the trend of change and the correlation coefficient, and the calculation formula is as follows:

$$F=U \times \frac{{n - 2}}{Q}.$$
(4)

In the equation, \(U=\sum\limits_{{i=1}}^{n} {{{\left( {{y_i} - \overline {y} } \right)}^2}}\) is the regression sum of squares; \(Q=\sum\limits_{{i=1}}^{n} {{{\left( {{y_i} - \overline {y} } \right)}^2}}\) is the residual sum of squares. The variable \({y_i}\) represents the WUE value for each month in the\(\:\:i\) year. \(\overline {y}\) is the average WUE value over multiple years for each month, with n = 20 representing the number of years. To analyze the temporal and spatial variation trends of WUE in each month, six change levels are established based on the results of the F-test: extremely significant decrease, significant decrease, non-significant decrease, non-significant increase, significant increase, and extremely significant increase. Similarly, for the correlation between WUE and temperature/precipitation, six change levels are also defined according to the F-test results: extremely significant negative correlation, significant negative correlation, non-significant negative correlation, non-significant positive correlation, significant positive correlation, and extremely significant positive correlation.

Lag effect of temperature and precipitation on WUE

The Pearson correlation coefficient (R) is utilized to analyze the strength and lag time of the effects of temperature and precipitation on WUE in the SRB41,42. The coefficient ranges from − 1 to 1, reflecting the correlation’s intensity, from negative to positive.Firstly, to ensure the comparability of the correlation results, this study selected WUE data along with monthly temperature and precipitation data from the SRB, covering the period from 2001 to 2020. The monthly WUE was analyzed in conjunction with the temperature and precipitation data from the previous i months before the current month, where 0 ≤ i ≤ 12. This method produced a series of data sets, and the Pearson correlation coefficient (R) was calculated for each of these lagged time scales, resulting in 12 correlation coefficients for each pixel, as illustrated in Eq. (5).For example, if we consider a lag time of 3 months, we can conduct a correlation analysis between temperature and precipitation data from January to July for the years 2001 to 2020, and the monthly WUE data from April to October for the same period. This analysis can be repeated, extending the lag time up to 12 months.

The maximum correlation coefficient, denoted as Rmax_lag, is selected, and the corresponding lag month is identified as the lag effect size and lag time for the pixel, as illustrated in Eq. (6). For instance, if the Rmax_lag between monthly WUE and the temperature or precipitation from one month prior occurs in March, then the lag time for WUE in response to temperature or precipitation is recorded as 3 months. This indicates that the temperature or precipitation from three months ago significantly influences changes in WUE.

$${R_i}=corr\left( {WUE,prep/temp} \right)1 \leq i \leq 12$$
(5)
$${R_{\hbox{max} \_lag}}=\hbox{max} \left( {{R_i}} \right)1 \leq i \leq 3.$$
(6)

In this equation, WUE refers to the monthly WUE time series spanning from 2001 to 2020, with a lag of i months. The terms for temperature and precipitation are the respective time series, also lagged by i months. Additionally, Ri represents the Pearson correlation coefficient calculated with a lag of i months.

Cumulative effect of temperature and precipitation on WUE

To evaluate the cumulative impact of temperature/precipitation on WUE, the Pearson correlation coefficient (R) between monthly WUE and cumulative temperature/precipitation was calculated to obtain the corresponding maximum cumulative temperature/precipitation correlation41,42.Firstly, the WUE time series was correlated with the temperature and precipitation time series over m months (1 ≤ m ≤ 12) to calculate R, as shown in Eq. (7). The cumulative months of temperature and precipitation that correspond to the maximum correlation Rmax_cum WUE are identified as the time scale of the cumulative effect, as shown in Eq. (8). Rmax_cum represents the magnitude of this cumulative effect. For instance, if the correlation between monthly WUE and the temperature/precipitation over three months is highest, the time scale of the cumulative effect will be recorded as three months. This indicates that the cumulative temperature and precipitation from the three months before the current month have the greatest impact on vegetation photosynthesis.

$${R_m}=corr\left( {WUE,mprep/mtemp} \right)1 \leq m \leq 12$$
(7)
$${R_{\hbox{max} \_cum}}=\hbox{max} \left( {{R_m}} \right)1 \leq m \leq 3.$$
(8)

In the equation, m denotes the cumulative time scale for temperature and precipitation data, while mprep/mtemp represents the time series of precipitation and temperature over m cumulative months. Additionally, Rm is the Pearson correlation coefficient between WUE and temperature/precipitation.

Results

Temporal–spatial of WUE in SRB

Spatial patterns of WUE in SRB

Over the past 20 years, the average annual WUE in the SRB has been measured at 0.99gC·kg− 1 H2O. The spatial distribution of WUE typically reveals lower values in the upstream and downstream areas, while the midstream region exhibits higher values (Fig. 2a). The regions with an annual WUE higher than 1.2 gC·kg− 1 H2O are mainly distributed in the southeastern part of Zone I, the central part of Zone II, the Minqin Oasis in Zone III, and the Jinchuan and Hexibao Oasis areas, accounting for 27.87% of the total vegetation area in the basin. The regions with an annual WUE lower than 0.4 gC·kg− 1 H2O are mainly located near the cold desert area of the Qilian Mountains in the southwestern part of Zone I, accounting for 3.87% of the total vegetation area in the basin. From the perspective of each ecological functional zone, Zone I has an average annual WUE of 0.90 gC·kg− 1 H2O; Zone II has the highest WUE in the basin, at 1.23 gC·kg− 1 H2O; Zone III has a WUE of 1.06 gC·kg− 1 H2O. Since most of Zone IV is located in desert areas with low vegetation coverage, the overall average annual WUE in this region is the lowest, at 0.80 gC·kg− 1 H2O.

Temporal–spatial trends of WUE in SRB

From 2001 to 2020, the annual WUE in the SRB exhibited a downward trend. The area with a WUE greater than 1.2 gC·kg− 1 H2O gradually declined, with regions showing this decrease accounting for 79.05% of the total vegetation area in the basin. The annual mean WUE reached its lowest point in nearly 20 years in 2011(Fig. 2d), measuring 0.9 gC·kg− 1 H2O, while in 2002, it peaked at 1.17 gC·kg− 1 H2O. Over this period, the average rate of change in WUE was − 0.0056 gC·kg− 1 H2O per year. The reduction in WUE was primarily characterized by a nonsignificant decrease, which covered 35.43% of the basin’s total area. In contrast, a very significant decrease in WUE affected 29.08%(Fig. 2c) of the area. Across different ecological functional zones in the SRB, WUE consistently showed a downward trend, with Zone III exhibiting the most pronounced decline—89.85% of its total vegetation area experienced a decrease(Fig. 2b). Meanwhile, Zone II had the highest proportion of areas with increasing annual WUE among the four zones, at 25.32% (Fig. 2e).

Based on the multi-year monthly changes of WUE, the WUE in the SRB showed an increasing trend in February, November, and December (Fig. 2f). The area with an increasing WUE trend in February was the largest, accounting for 88.13% of the total area, but the increase was nonsignificant. In contrast, the WUE showed a decreasing trend in other months, with the largest area of decreasing WUE occurring in April, accounting for 98.31% of the total area. Among this, the area with a significant decrease accounted for 27.29% of the total area, and the area with a highly significant decrease accounted for 12.50%.

Fig. 2
Fig. 2
Full size image

Spatial and temporal patterns of WUE in the SRB from 2001 to 2020. (a) Multi-year average WUE pattern; (b) Univariate linear change of WUE over the past 20 years; (c) WUE change trend over the past 20 years (EDT: Extremely Significant Decline; SDT: Significant Decline; NDT: Not Significantly Decline; NUT: Not Significantly Increase; SUT: Significantly Increase; EUT: Extremely Significant Increase); (d) Yearly change of WUE (the left Y-axis represents the cumulative percentage of areas in different WUE ranges; the right Y-axis represents the annual average value of WUE); (e) WUE change trend in different ecological functional areas (the left Y-axis represents the percentage of areas with different WUE change trends in each ecological functional zone over the past 20 years, where the negative Y-axis indicates the area with a declining WUE trend; the right Y-axis represents the average rate of WUE change over 20 years in each ecological functional zone); (f) Monthly change trend of WUE (the left Y-axis represents the percentage of areas with different WUE change trends in each month over the past 20 years, where the negative Y-axis indicates the area with a declining WUE trend; the right Y-axis represents the average rate of WUE change over 20 years for each month).

Correlation between temperature/precipitation and WUE

Correlation between temperature and WUE

From 2001 to 2020, the annual WUE in the SRB generally showed a negative correlation with the annual average temperature (Fig. 3), with the area showing a negative correlation accounting for 96% of the total basin area. Among this, the area with a nonsignificant negative correlation accounted for 58.13% and was mainly distributed in Zone I (31.56%) and Zone II (17.80%). The regions with significant and extremely significant negative correlations were primarily located in the northeastern part of Zone I (18.50%). The area with a positive correlation between WUE and temperature accounted for 3.18% and was mainly distributed in the southwestern part of Zone I and the central part of Zone II. The regions with nonsignificant and significant positive correlations were mainly found in Zone I, accounting for 1.57% of the total basin area. The correlation between WUE and temperature in Zone III (R = -0.5191) was stronger than that in Zone I (R = -0.5163), Zone II (R = -0.4170), and Zone IV (R = -0.5185), indicating that the WUE in Zone III responded more strongly to temperature.

Fig. 3
Fig. 3
Full size image

Spatial distribution of significant correlations between annual WUE and TEMP in the SRB (a) (ENC: Extremely significant negative correlation, SNC: Significant negative correlation; UNC: Unsignificant negative correlation, UPC: Unsignificant positive correlation, SPC: Significant positive correlation, EPC: Extremely significant positive correlation; the same applies to the following explanations) and the heatmap of correlations between WUE and TEMP in different ecological functional areas (b) (WUE 1/2/3/4 represent the WUE values of Zones I/II/III/IV, respectively, and the same applies to the following explanations; Temp 1/2/3/4 represent the annual average temperatures of Zones I/II/III/IV, respectively).

Over the past 20 years, the WUE in the SRB showed a positive correlation with temperature in January, February, March, May, September, October, November, and December. In contrast, a negative correlation was observed in April, June, July, and August, with the negative correlation in June and July mainly distributed in Zone I (Fig. 4f, g). From September to March of the following year, the monthly WUE in the SRB generally showed a positive correlation with temperature. The area with a significant positive correlation was relatively large from December to March of the following year, with the area showing an extremely significant positive correlation in December being the highest among all months, accounting for 98.58% of the vegetation-covered area in the basin. In April, the WUE mainly showed a nonsignificant negative correlation with temperature, with the nonsignificant negative correlation area accounting for 51.93% of the vegetation-covered area in the basin. From May to August, the correlation between WUE and temperature gradually shifted from positive to negative over time. The negative correlation between WUE and temperature in August was the strongest among all months, with the negative correlation area accounting for 92.66% of the vegetation-covered area in the basin, of which the nonsignificant negative correlation area accounted for 83.54%. Comparisons of the correlation between monthly WUE and temperature in different ecological functional zones of the SRB revealed that Zones I–IV all showed a positive correlation from September to March of the following year and gradually shifted from positive to negative from May to August. In April, due to the nonsignificant negative correlation between WUE and temperature in Zone I, the overall WUE in the basin showed a negative correlation with temperature. Additionally, due to the low temperature from December to February of the following year, the vegetation cover in the southeastern part of Zone I was low, resulting in an almost zero correlation between WUE and temperature in this area.

Fig. 4
Fig. 4
Full size image

Trend of monthly correlation between temperature and WUE in the SRB from 2001 to 2020.

Correlation between precipitation and WUE

From 2001 to 2020, the annual WUE in the SRB showed an overall negative correlation with precipitation (Fig. 5). The area exhibiting a negative correlation accounted for 93.65% of the total basin area. Specifically, 70.84% of the area showed a nonsignificant negative correlation, which was mainly distributed in Zone I (31.02%) and Zone II (17.62%). The regions with significant and extremely significant negative correlations were primarily located in Zone I, accounting for 15.07% of the total basin area. The area where annual WUE was positively correlated with precipitation was 6.35%, mainly found in Zone I and Zone II. Among these, the areas with nonsignificant and significant positive correlations were predominantly in the southwest of Zone I, accounting for 5.50% of the total basin area. The correlation between annual WUE and precipitation was stronger in Zone II (R = -0.5199) than in Zone I (R = -0.0350), Zone III (R = -0.4435), and Zone IV (R = -0.1579), indicating that WUE in Zone II was more significantly influenced by precipitation compared to other zones.

Fig. 5
Fig. 5
Full size image

Spatial distribution of significant correlations between annual WUE and PREP in the SRB (a) and the heatmap of correlations between WUE and PREP in different ecological functional areas (b) (WUE 1/2/3/4 are the same as in Fig. 4; Prep 1/2/3/4 represent the annual precipitation in Zones I/II/III/IV, respectively).

Unlike temperature, the correlation between monthly WUE and precipitation in the SRB is negative across all months (Fig. 6), with June showing the strongest degree of negative correlation. The area exhibiting significant and extremely significant negative correlation accounts for 67.91% of the total vegetation-covered area in the basin. From June to August, WUE in all ecological functional zones showed a negative correlation with precipitation (Prep). However, over time, except for Zone I, the other three zones exhibited large areas of positive correlation. In June, a large area of positive correlation between WUE and Prep was observed in the central part of Zone II, accounting for 13.75% of the total vegetation area in Zone II. In July, this positive correlation appeared in the northwest of Zone IV, accounting for 32.42%. In August, a large area of positive correlation was observed in the northern part of the Minqin Oasis in Zone III, accounting for 17.50% of the total vegetation area in Zone III. In September, the proportion of the area with a positive correlation between WUE and Prep was 18.75% in Zone I and 11.74% in Zone II. From September to May of the following year, WUE in all ecological functional zones showed a negative correlation with Prep. However, in November, both Zone I and Zone II exhibited positive correlation, with the area proportions being 19.63% in Zone I and 9.78% in Zone II.

Fig. 6
Fig. 6
Full size image

Trend of monthly correlation between precipitation and WUE in the SRB from 2001 to 2020.

Effect of lag and cumulative on WUE

Figure 7(a) illustrates the spatial distribution of the lagged and cumulative response times of Water Use Efficiency (WUE) to temperature in the SRB. Meanwhile, Fig. 7(c, d) presents the proportion of lagged and cumulative response areas across different ecological functional zones. The primary response of monthly WUE to temperature is characterized as TLA0-0 (TLA0-0) (TLA i-j means that the lag and accumulation months are i and j months, respectively.), indicating that WUE in most areas of the basin reacts instantaneously to temperature fluctuations. This instantaneous response accounts for 91.59% of the vegetation-covered area within the basin. The average lagged and cumulative response time of WUE to temperature in Zones I-IV aligns with that of the overall basin, all demonstrating the TLA0-0 pattern. Additionally, the proportion of TLA0-0 in Zones II-IV exceeds 99%. However, within the cold desert region near the Qilian Mountains in the southwestern part of Zone I, the lagged and cumulative response of WUE to temperature reflects a 0 months lag and 3 months accumulation (TLA0-3 ). This finding suggests that while WUE in the SRB is generally sensitive to temperature changes, the cumulative effect of temperature on WUE is notably significant in the cold desert area of the Qilian Mountains in Zone I.

The lagged and cumulative response of monthly WUE to precipitation in the SRB generally shows a pattern of 0 months lag and 3 months accumulation (TLA0-3) (Fig. 7b). The average lagged and cumulative response time of WUE to precipitation in Zones I–IV is consistent with the overall basin, both showing TLA0-3, and the proportion of TLA0-3 in Zones II–IV is higher than 95%. However, in the mid- and high-altitude areas of the Qilian Mountains in the southwestern part of Zone I, the lagged and cumulative response of WUE to precipitation is characterized by 2 months lag and 1 month accumulation (TLA2-1). This indicates that the WUE in the SRB is mainly influenced by the accumulated precipitation over 3 months and shows an instantaneous response to non-accumulated precipitation. The WUE in the southwestern part of Zone I is affected by the combined effects of lag and accumulation of precipitation.

Fig. 7
Fig. 7
Full size image

The lag and cumulative response of WUE in the SRB to temperature and precipitation. (a) The lag and cumulative response of WUE to average temperature; (b) The lag and cumulative response of WUE to precipitation; (c) The proportion of different lag and cumulative response areas of WUE in each ecological functional area to average temperature; (d) The proportion of different lag and cumulative response areas of WUE in different ecological functional zones to precipitation. Note: (TLA i-j means that the lag and accumulation months are i and j months, respectively.) TLA0-0 (0 months lag 0 months cumulative), TLA0-1 (0 months lag 1 month cumulative), TLA0-2 (0 months lag 2 months cumulative), TLA0-3 (0 months lag 3 months cumulative), TLA1-0 (1-month lag 0 months cumulative), TLA1-1 (1-month lag 1 month cumulative), TLA1-2 (1-month lag 2 months cumulative), TLA1-3 (1-month lag 3 months cumulative), TLA2-0 (2 months lag 0 months cumulative), TLA2-1 (2 months lag 1 month cumulative), TLA2-2 (2 months lag 2 months cumulatively, TLA2-3 (2 months lag 3 months cumulative), TLA3-0 (3 months lag 0 months cumulative), TLA3-1 (3 months lag 1 month cumulative), TLA3-2 (3 months lag 2 months cumulative), TLA3-3 (3 months lag 3 months cumulative).

Discussion

Characteristic of ecological function area WUE from 2001 to 2020

The spatial distribution characteristics of WUE represent the characteristics of water use in different regions and vegetation types of ecosystems, as well as the adaptation results of human intervention processes43. The SRB is located in the northwest arid region, with fragile habitats, a dry climate, and significant human disturbance. It belongs to a typical ecologically sensitive area44,45. Zone I can gather natural precipitation and conserve water sources, providing a habitat for high-altitude organisms in mid-latitudes. The southern part of the area is mainly composed of high-coverage grasslands and tall trees, with a balanced and stable distribution of rivers. The soil moisture content is high, providing a foundation for high carbon sequestration of vegetation. The diverse vegetation types and abundant water sources are important reasons for the high annual WUE in the area. Providing a foundation for the high carbon sequestration value of vegetation, the diverse vegetation types and abundant water sources are important reasons for the high annual WUE in the region. However, in the cold desert region, the lowest value of basin WUE (<0.4 gC·kg− 1 H2O) occurs. This is because, although precipitation and glacier meltwater can replenish the water supply in a timely manner, the low temperature inhibits the photosynthetic efficiency of plants. This, in turn, becomes a significant factor restricting vegetation growth, thereby reducing the carbon absorption per unit of water, resulting in the low WUE in the cold desert region. Zone II is the main oasis in the SRB, with abundant groundwater resources that can provide water for vegetation growth. After 2000, large-scale ecological governance projects were carried out in the Hexi Corridor region. After 20 years of continuous efforts, progress has been made in areas such as sand prevention and control, watershed ecological protection, and vegetation restoration23. The increase in vegetation coverage leads to a higher average annual WUE in this area. Zone III is located in the transitional zone between oasis and desert in the lower reaches of the watershed, directly facing the two major deserts. Within the region, there are Minqin Oasis, Changning Oasis, Jinchuan Oasis, and Hexibao Oasis. The area lacks water and is affected by drought, making it difficult for vegetation to sequester carbon. At the same time, the desert terrain makes it difficult to retain water, resulting in increased water evaporation. The overall average annual WUE value is relatively low, but the oasis area is higher than other areas in the same area. The precipitation in Zone IV is sparse, and the arid desert ecosystem is composed of arid and ultra-arid vegetation. The vegetation is mostly low and sparse, and the groundwater system is relatively underdeveloped. Therefore, the overall average annual WUE value of vegetation is the lowest in the SRB.

The spatial distribution of WUE characterizes the differences in WUE among different ecological functional zones, while the temporal variation of WUE well presents the results of vegetation’s changes in water consumption and productivity in the face of environmental factors and human intervention. From 2001 to 2020, the overall WUE in the SRB showed a decreasing trend year by year, mainly due to the trend of warming and humidifying the overall climate in the northwest arid region in recent years, as well as the recent sand fixation and afforestation plans (ecological governance plans) in the SRB. This has led to an increase in vegetation coverage, enhanced carbon sequestration capacity, and enhanced water retention capacity in the SRB, especially in Zone II. The ecosystem environment has been greatly improved, increasing WUE. The vegetation in Zone III has shown the most significant decline in WUE year by year, possibly due to urban expansion in Liangzhou District, where the land cover type has changed from arable land to urban land, resulting in a decrease in vegetation coverage in this area; After 2010, the Minqin Oasis made significant adjustments to its agricultural industrial structure. In certain areas, traditional crops were replaced with economic crops, and the construction of agricultural greenhouses was increased. While the cultivation of these economic crops and the development of greenhouses can boost agricultural productivity, they may also lead to adverse effects such as soil structure degradation, reduced soil fertility, and the destruction of natural vegetation, ultimately resulting in vegetation degradation. This degradation reduces vegetation cover and biomass, which diminishes the plants’ ability to capture carbon dioxide and weakens the process of ET. One limitation of this study is its focus primarily on the dynamics of vegetation within ecological functional areas. In the future, we plan to conduct a more in-depth analysis using multi-scale coupling to examine the differentiated effects of changes in various vegetation types on WUE. Our goal is to establish an ecological governance framework based on the coupling and feedback mechanisms of carbon and water, providing decision-making support for the sustainable management of oases in arid regions.

Interactions between temperature/precipitation and WUE

There are significant spatial differences and seasonal variations in the response of WUE to water and heat factors in different ecological functional zones. From 2001 to 2020, the WUE in the SRB showed a negative correlation with temperature and precipitation year by year. The sensitivity of WUE to precipitation in Zone I and IV was lower than that of temperature. The correlation between vegetation growth and temperature was higher in the high-altitude grassland area of Zone I; In Zone IV, the vegetation is dominated by xerophytic species, which have already adapted to the arid environment. Their growth and photosynthesis have relatively low water demands. Therefore, the impact of precipitation on WUE is weak, while temperature changes have a greater influence on the physiological activities of vegetation. Hence, on the existing vegetation basis, water is no longer a limiting factor, and temperature has become an important factor restricting vegetation46. However, in Zone II, the vegetation’s WUE is more sensitive to precipitation on an annual basis. This may be due to the fact that the land use type in this area is mainly cropland. Cropland usually requires frequent human activities, such as irrigation, fertilization, and cultivation. These activities can alter the soil structure and the growth environment of vegetation, making vegetation more sensitive to changes in precipitation. During the cultivation process, land management is adjusted according to needs, especially the use of irrigation techniques, which directly disturbs the correlation between WUE and climate. Therefore, the annual WUE in different ecosystems in the SRB is limited by temperature but is influenced by factors such as human activities, land use patterns, and management measures.

The WUE in the SRB has seasonal characteristics. From December to February of the following year, the overall climate of the SRB is cold. Due to the replenishment of groundwater for vegetation growth, heat becomes the main limiting factor for vegetation growth. Therefore, the monthly WUE is positively correlated with temperature, but not significantly negatively correlated with precipitation. During the spring flood period from March to May, Zone I has a high altitude and most areas are located at high altitudes. The melting of ice and snow increases the replenishment of soil moisture, while the rise in temperature increases the stomatal conductance of vegetation, which is beneficial for vegetation growth. However, vegetation is in the developmental stage and has weak carbon sequestration ability. Therefore, there is a large negative correlation between WUE and temperature in April. Zone II and Zone III are densely populated areas in the basin, and vegetation is greatly affected by human activities. Because some of the above-ground runoff is used for irrigation in the central agricultural area, while the underground water infiltrated is extracted for domestic use, there are very few water resources available for ecological use. It can be seen that vegetation located in desert areas and oasis desert transition zones has a low utilization rate of above-ground runoff, and monthly WUE shows a significant negative correlation with precipitation. From June to August, as temperatures rise, stomatal conductance and transpiration in Zones II and III decrease39. Despite this decline, vegetation is in its growth stage during this period. Consequently, the amount of organic matter produced per unit of water consumed increases, resulting in a positive correlation between monthly WUE and temperature. In Zone I, between June and July, excessively high temperatures can lead to rapid melting and loss of snow and ice, as well as increased evaporation of soil moisture. This situation causes vegetation to experience greater water stress. As a result, the limited water availability prevents plants from fully utilizing the higher temperatures to enhance photosynthetic efficiency. Consequently, in this zone, WUE is negatively correlated with temperature. As the temperature rises in Zone I, the southern ice, and snow melt water supply is sufficient, the soil moisture content increases and vegetation growth is not affected by water stress. The monthly WUE is significantly negatively correlated with precipitation. From September to November, vegetation is in the late growing season, and a decrease in temperature can lead to a decrease in the stomatal conductance of vegetation leaves. Plants can use less water to exchange for an equal amount of carbon, and the physiological effect of CO2 enrichment has a greater alleviating effect on vegetation water stress than the exacerbation of atmospheric water deficit, thereby significantly reducing the water demand of the ecosystem and improving plant WUE. Therefore, monthly WUE is positively correlated with temperature46,47. The altitude of Zone I is mostly above 3000 m. Due to sparse human habitation, WUE has a higher response to climate factors every month. As the temperature begins to decrease, melting water from ice and snow decreases, and soil water supply decreases. Therefore, monthly WUE is positively correlated with precipitation.

The temporal effect of climate on vegetation growth varies with climate factors and regions, and these research results are consistent with Ding et al. (2020)22 findings at the global scale. The lagged response of monthly WUE in the SRB to temperature and precipitation is not significant, but the cumulative response is significant. The main combination of temperature and precipitation lag and accumulation months is TLA0-0 (0 months lag 0 months accumulation) and TLA0-3 (0 months lag 3 months accumulation), which reflects the rapid response of vegetation to climate factors. The response of vegetation in Zone I to temperature every month is mainly characterized by a lag of 1 months and a cumulative effect of 2 months. This may be due to the high sea level and year-round cold temperatures in the Qilian Mountains. At the same time, most areas are mainly composed of high mountain wasteland, distributed with alpine meadows, with a relatively single vegetation type and strong adaptability to low-temperature environments. The response of vegetation to the lag and cumulative effects of precipitation on a monthly basis is relatively strong at 0 month lag and 3 months accumulation. The study area is located in the arid region of northwest China, with year-round drought little rainfall, and a poor climate environment. Precipitation is one of the main influencing factors of the vegetation ecosystem in the SRB, so vegetation responds quickly to precipitation. The occurrence of a 3-month accumulation may be due to the need for plant growth and development to accumulate a certain amount of temperature and water to differentiate, germinate, and bloom22,43. There has been significant progress in the ecological environment management on the SRB in Zone II, III, and IV, including the artificial water conveyance of Qingtu Lake and the continuous management of the Minqin Oasis area. These have provided better water conditions for plants, stabilized the ecosystem, and enhanced their resistance to short-term drought. Therefore, the cumulative effect of WUE on precipitation every month is manifested as a 3-month accumulation.

Conclusions

The spatiotemporal distribution and variation characteristics of WUE in the SRB reveal significant ecological functional zone heterogeneity. From 2001 to 2020, the spatial pattern of WUE was characterized by high values in the middle reaches and low values in the upper and lower reaches. The highest WUE was observed in Zone II, an oasis area with optimal vegetation cover and resource conditions, while the lowest WUE was found in the desertified Zone IV due to sparse vegetation. Overall, WUE in the SYRB showed a decreasing trend from 2001 to 2020, with the most significant decline in Zone III, attributed to its ecological fragility. Notably, local increases in WUE were observed in parts of Zone II, suggesting potential regulatory effects of regional management measures, including human interventions and water resource allocation.

The impact of climate factors on WUE exhibited complex temporal variations. Annually, WUE was significantly negatively correlated with both air temperature and precipitation. However, monthly responses differed markedly: from September to March, increasing air temperature may mitigate low-temperature stress and enhance vegetation photosynthesis, resulting in a positive correlation with WUE. In contrast, from April to August, high temperatures exacerbate water stress, leading to a negative correlation with WUE. Precipitation was predominantly negatively correlated with WUE across all months, with the strongest negative correlation occurring in June, likely due to intensified water supply-demand conflicts at the onset of the rainy season. Different ecological functional zones exhibited distinct sensitivities to climatic factors: WUE in Zone III was more responsive to air temperature, while WUE in Zone II was more significantly regulated by precipitation, reflecting regional differences in the interaction between vegetation types and hydrothermal conditions.

The response of monthly WUE to air temperature and precipitation showed significant spatial variations across the SYRB. The WUE response to air temperature was predominantly immediate (TLA0-0) throughout most of the basin, accounting for 91.59% of the vegetated area. This pattern was consistent across Zones I-IV. However, in the southwestern part of Zone I, within the Qilian Mountain cold desert region, WUE exhibited a response with 0 months lag and 3 months cumulative effect to air temperature (TLA0-3), indicating a notable cumulative effect of air temperature on WUE in this area. The WUE response to precipitation basin-wide was generally characterized by 0 months lag and 3 months cumulative effect (TLA0-3), suggesting that WUE was primarily influenced by the cumulative precipitation over three months. Nevertheless, in the medium and high-altitude areas of the Qilian Mountains in the southwestern part of Zone I, WUE responded to precipitation with 2 months lag and 1 month cumulative effect (TLA2-1), indicating a combined effect of lagged and cumulative precipitation on WUE in this region.