1. Introduction
The Tibetan Plateau (TP) is the world’s largest reservoir of solid water (glaciers, snow cover, and permafrost) besides the polar region and is therefore called the “Asian Water Tower” (T. Yao et al. 2022). The cryosphere and hydrological cycles over the TP are vulnerable to climate change, especially the warming (Duan and Xiao 2015; L. Yao et al. 2022). The imbalanced hydrological cycles and water resource instability accompanied by considerable spatial heterogeneities in the Asian Water Tower region have become scientific focus (Yao et al. 2012, 2017) and include accelerated glacier retreat (Yao et al. 2012; Hugonnet et al. 2021), permafrost degradation (Yang et al. 2010; Guo and Wang 2013), lake expansion (Lei et al. 2013; Yang et al. 2017; G. Zhang et al. 2020), increased precipitation (You et al. 2008; Jiang et al. 2023), and water storage change (G. Zhang et al. 2017). These imbalanced hydrological cycles may affect the ecosystem (e.g., increased species loss; Klein et al. 2004), resulting in the spatial instability of water resources, and even exert marked impacts on regional climate and global atmospheric circulations (Xu et al. 2008; Chen et al. 2021). Furthermore, the TP serves as the source of several large rivers that supply water to ∼20% of the world’s population (Immerzeel et al. 2010; Wang et al. 2021), meaning that the socioeconomic consequences of such water resource instability may be considerable (Pritchard 2019).
As the primary component of hydrological cycles, precipitation is especially complex over the TP due to its multiple moisture sources (Yao et al. 2013) and varied transportation pathways (Curio et al. 2015; Pan et al. 2019). The Indian summer monsoon (ISM) moisture is humid and bring significant volumes of precipitation, while the westerlies or inner TP moisture sources tend to be dry, providing far fewer precipitation events (Yao et al. 2013). Considerable attention has been paid to the TP’s moisture sources (C. Zhang et al. 2019; Li et al. 2022) and their relationships with precipitation patterns or water resource distributions (Singh and Nakamura 2009; Rajagopalan and Molnar 2013; Gao et al. 2014; Sun et al. 2020). However, how closely these moisture sources’ temporal variabilities correlate with spatially inhomogeneous changes in water resource distributions remains unclear. Furthermore, several large rivers rise on, or flow through, the southeastern TP (TPSE), a region that distributed multiple water vapor transportation pathways. This means that the previous methodologies that taken the entire ISM-affected region as one moisture cluster may be inaccurate. Large-scale atmospheric circulations determine the intensities or proportions of varied moisture sources and their trends, thereby affecting both precipitation patterns and their temporal variabilities. However, few studies have quantitatively analyzed the spatial distributions and temporal variabilities of the TP’s various moisture sources and how they affect precipitation (C. Zhang et al. 2017), an issue that becomes ever more critical against the background of intense climate warming (Immerzeel et al. 2010). The proportional input of various moisture sources and how these proportions have changed in recent decades, and further how closely these are connected to water resource instability over the TP, are questions which have, until now, been left unaddressed.
This study represents an initial attempt to quantitatively map out the spatial distributions and temporal variabilities of various moisture sources using high-temporal resolution modeling (2.5° × 2.5°, 6 h, 1951–2020) and cross compare these with spatially dense precipitation δ18O. The moisture source clustering methods were based on large-scale circulations that considered both moisture sources and transportation pathways, cited from the results of C. Zhang et al. (2019). The quantitative moisture proportions on base of Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) modeling and moisture clustering are first used in this study and over the TP. We aim to 1) qualitatively identify each moisture source proportions and present their spatial distributions in the Asian Water Tower region and 2) reveal the temporal variabilities in various moisture sources in recent decades. These results will improve our understanding of precipitation changes and hydrological cycles and provide insights into water-related studies of the TP.
2. Study area, datasets, and methods
a. Study area
The study area covers the whole TP and its surroundings (Fig. 1). The TP boundary is from the National Tibetan Plateau Data Center (http://data.tpdc.ac.cn) (Zhang et al. 2021). The TP’s climate is mainly influenced by moisture transported by the following moisture sources: the westerlies (West) from mid- and high latitudes of Eurasia (Tian et al. 2001), the ISM (the western arm of ISM, ISM1; the eastern arm of ISM, ISM2) (Fig. 2; Yao et al. 2013), and the TP recycling moisture, southeast (SE) moisture, and northeast (NE) moisture sources. The digital elevation model derived from the Shuttle Radar Topography Mission (https://gisgeography.com/free-global-dem-data-sources/) shows the study area’s general topography.
The TP and usual moisture transportation pathways surrounding the TP. Colored arrows show the various moisture sources: the westerlies (West), the western arm of ISM (ISM1), the right arm of ISM (ISM2), the TP, moisture from the SE, and moisture from the NE. The blue (black) curves show the TP boundary (Zhang et al. 2021); digital elevation model derived from Shuttle Radar Topography Mission shows the study area’s general topography.
Citation: Journal of Climate 38, 1; 10.1175/JCLI-D-24-0093.1
(a)–(d) The 10-day back trajectories (HYSPLIT model) for the days with precipitation events at Nyalam, Lhasa, Naqu, and Tuotuo in summer of 2014–15. Color bars on the right show the height (m; AGL) of moisture trajectories. (e) The locations of moisture sources at the 10-day back trajectories shown in (a)–(d). The bold red lines are the rough boundary that separates the ISM into ISM1 and ISM2.
Citation: Journal of Climate 38, 1; 10.1175/JCLI-D-24-0093.1
b. Datasets and methods
We first traced the back trajectories with the HYSPLIT model, then the moisture sources were clustered into six groups (i.e., West, ISM1, ISM2, TP, SE, and NE) by locations, and finally the proportions and temporal trends for each moisture source during the summer months over 1951–2020 were mapped. The gbl reanalysis (datasets in the format of .gbl) (1951–2020) from NOAA satellite was used to drive the HYSPLIT model. Precipitation isotopes were used to spatiotemporally verify the modeling and clustering results. General steps were as follows. 1) Data preparation: The HYSPLIT codes were accessed and prepared for batch analysis. The gbl reanalysis for the period 1951–2020 and precipitation isotopes was subsequently prepared. 2) HYSPLIT modeling and moisture source clustering: The back trajectories for each 6 h during 1951–2020 were traced for each 2.5° grid in the study area. The locations (longitude and latitude) of the moisture sources were then clustered into six groups (West, ISM1, ISM2, TP, SE, and NE) based on Fig. S1b in the online supplemental material. The proportions of the six moisture sources (mean during 1951–2020) were shown in the background of Fig. 3, Fig. S4, Fig. 4 and Fig. S5. 3) Trends of moisture sources: The linear trends [Eqs. (1) and (2)] for each moisture source during 1951–2020 were provided as Figs. S2 and S3 and shown as ± (increasing/decreasing) in Fig. 3, Fig. S4, Fig. 4, and Fig. S5. For the spatial distributions and temporal variabilities of moisture proportions [steps (2) and (3)], we concentrated on the 10-day results (Fig. 3 and Fig. S4), and we also presented the comparisons for the main moisture sources (West, ISM1, ISM2, and TP) at days 02–04–06–08–10 (Fig. 4 and Fig. S5). 4) Verification of the modeled results: The modeled and clustering results for each moisture sources on both spatial (Fig. 5) and temporal (Fig. 6) scales were verified using precipitation δ18O records. All analyses in the main text were of the summer season, defined as months from June to August. Annually results were provided in Supplements. We presented the abbreviations for items used in this study as Table 1. The detailed datasets and methods were in the following paragraphs.
Overview of items used in this analysis.
1) Gbl reanalysis from NOAA satellite
The gbl reanalysis from NOAA satellite was used to drive the HYSPLIT model and trace the back trajectories of moisture sources for the 1951–2020 period. The NOAA reanalysis was taken from the NCEP/NCAR Reanalysis Project, a joint project between the National Centers for Environmental Prediction and for Atmospheric Research. The quality and utility of the NOAA reanalysis are superior to the original analysis, because of more observations used and better vertical resolutions (https://www.ready.noaa.gov/gbl_reanalysis.php). Two datasets from this reanalysis were available to drive the HYSPLIT model. One is the NCEP reanalysis, whose spatial resolution is as high as 1° × 1°, though the time series is short (2005–22). The gbl reanalysis covers longer time series, i.e., 1948–the present; its spatial resolution is a little coarser, at 2.5° × 2.5°. Both the spatial patterns and temporal variabilities in moisture source proportions are required in order to fully understand the spatially inhomogeneous changes in water resources over the TP. The gbl reanalysis during 1951–2020 was therefore the ultimate choice (NCEP/NCAR Global Reanalysis Data Archive, ftp://arlftp.arlhq.noaa.gov/pub/archives/reanalysis/). It was quality controlled before being public (https://www.cpc.ncep.noaa.gov/products/wesley/reanalysis.html).
2) Precipitation isotopes (δ18O)
Precipitation isotopes were used to verify moisture source modeling and clustering results. As many precipitation δ18O datasets as possible were recovered from four research programs, including Isoscape (C-Isoscape, Northwest Normal University, http://cisoscape.wz.hwdlszywz.net) (S. Wang et al. 2022), the Global Network of Isotopes in Precipitation (GNIP, from the International Atomic Energy Agency, https://www.iaea.org/services/networks/gnip) (Dansgaard 1964), the Tibetan Plateau Network of Isotopes in Precipitation (TNIP, from the National Tibetan Plateau Data Center, http://data.tpdc.ac.cn) (Yao et al. 2013; Gao 2020), and isotope values collected by our research group (Lhasa). The Isoscape archive represents monthly precipitation isotope datasets, compiled from modern precipitation measurements at 223 sites across China and 48 sites in China’s surrounding countries. Using a regional fuzzy clustering method, the Isoscape archive provides an accurate and high-resolution (10′ × 10′) mapping of Chinese precipitation isotopes (S. Wang et al. 2022). The GNIP and TNIP datasets are open access precipitation isotope archives from around the world and for the TP, respectively. Precipitation samples in our research group were collected using a specifically designed container (Gröning et al. 2012). We sampled the rainfall at 20 h each day it occurred; snowfall was collected immediately and then melted in sealed plastic bags at room temperature. All samples were kept in freezers and then analyzed in the State Key Laboratory of Tibetan Plateau Earth System, Environment and Resources, Institute of Tibetan Plateau Research. The test results were obtained using Picarro-2130i Liquid Water Isotope Analyzers, with reference to the Vienna Standard Mean Ocean Water (VSMOW) standard. The analytical precision for δ18O was ±0.1‰. Detailed information about precipitation δ18O datasets (Isoscape, GNIP, TNIP, and our research group) was given in Table S1, which included the collection sites and study periods. When verifying the results of moisture modeling spatially, all the four kinds of precipitation δ18O were used; precipitation δ18O values at one site (Lhasa) were used to temporally verify the moisture modeling results. Precipitation isotopes at Lhasa for the 1986–92 (1993–2012) period were taken from the GNIP dataset (our research group).
3) Moisture source proportions (%) and their trends
The HYSPLIT model was used to trace the back trajectories of moisture sources. HYSPLIT is a complete system for computing simple air parcel trajectories, as well as complex transport, dispersion, chemical transformation, and deposition simulations (https://www.arl.noaa.gov/hysplit/). The model calculation method is a hybrid between the Lagrangian approach, using a moving frame of reference for the advection and diffusion calculations as the trajectories or air parcels move from their initial locations, and the Eulerian methodology, which uses a fixed three-dimensional grid as a frame of reference to compute pollutant air concentrations (Stein et al. 2015). HYSPLIT has evolved over more than 30 years, from estimating simplified single trajectories based on radiosonde observations to a system accounting for multiple interacting pollutants transported, dispersed, and deposited over local to global scales. A common application is a back trajectory analysis to determine the origin of air masses and establish source–receptor relationships. As one of the most extensively used atmospheric transport and dispersion models in the atmospheric sciences community, HYSPLIT has also been used in a variety of simulations describing the atmospheric transport, dispersion, and deposition of pollutants and hazardous materials.
The calculation steps can be summarized as below. 1) Data and HYSPLIT model preparation: The gbl reanalysis from NOAA satellite during 1951–2020 was downloaded. The HYSPLIT model codes can be accessed at http://www.arl.noaa.gov/HYSPLIT_info.php, and this was taken to be an effective register. 2) HYSPLIT modeling: The 10-day back trajectories at 1000 m AGL (above ground level) for each 6 h and for each 2.5° grid were traced for the 1951–2020 period. The precise locations (latitude and longitude, in °) at 01–10 day (24, 48, 72, 96, 120, 144, 168, 192, 216, and 240 h) for each trajectory were then extracted. A total of 135 (15 × 9) grids (Fig. 3) were used in this analysis, which covers the regions over and surrounding the TP. 3) Clustering of moisture sources: For the modeled locations of moisture sources [step (2)], we clustered them into six groups. The moisture clustering was based on large-scale circulations, cited from the results of Zhang et al. (2019) (Fig. S1). The six moisture sources were the West, ISM1, ISM2, TP, SE, and NE. In the work of C. Zhang et al. (2017, 2019), the modeled moisture sources were clustered into four groups, i.e., the West (the westerlies), the SW (ISM), the TP (the central TP), and the SE (southeast) (Fig. S1a); moistures from these four groups accounted for >90% of all moistures. In this study, these four clusters were extended, with moistures from the northeastern TP defined as NE and the SW specially classified into ISM1 and ISM2. Significant spatially inhomogeneous patterns were evident in ISM activity over the southern TP, i.e., the western part of the southern TP experienced earlier ISM onset (the dates that ISM arrives or the dates that ISM affects certain regions or grids), delayed retreat, longer duration, and greater ISM intensity and strength; the inverse was true in the eastern sector of the study area (Guo et al. 2023a). This inhomogeneous ISM activity may affect moisture sources and contributed to the spatially inhomogeneous precipitation patterns and water resource changes over the TP, topics of considerable concern that have been addressed by previous studies (Yao et al. 2013; Guo and Tian 2022). When tracing the 10-day back trajectories at four typical sites along a southwest–northeast interface over the TP (Nyalam, Lhasa, Naqu, and Tuotuo), the summer moisture trajectories were clearly distributed along a boundary (Fig. 2): moisture over the western part was directly transported to the TP over the Himalaya; moisture over the eastern part was transported upstream from the Hengduan Mountains before turning west and being transported to the central TP. The factors controlling spatially inhomogeneous changes in precipitation/water resources cannot be easily captured if the ISM is analyzed as one cluster. The boundary in Fig. 2 shows, therefore, the dividing line between ISM1 and ISM2. The mean (1951–2020) proportions for each of the six moisture sources are presented in Fig. 3 (colored background). 4) The trends in each moisture source’s proportionality: The linear trends for each moisture source during 1951–2020 were computed (Fig. 3), potentially providing new insights into precipitation and water resource instability over the TP. The linear trends were based on the 5-yr mean values from 1951 to 2020. Note that, in HYSPLIT modeling, there are four trajectories per day; when calculating the proportions of various moisture sources, it was assumed that the amounts of moisture in each of the four trajectories were the same. Thus, the moisture clustering and their temporal trends were based on the trajectories’ numbers, calculations that did not consider the water vapor flux.
(a)–(f) The proportions (%) of various moisture sources at 10-day trajectories averaged over 1951–2020, for the summer months (colored background) and their positive (crosses) or negative (vertical strings) trends during the same 1951–2020 period. The trends significant at p < 0.1 were marked with blue circles. (g) The areas where different moisture sources predominate (colored background), with arrows showing their general trends over the last 70 years (blue upward arrows, increasing; red downward arrows, decreasing).
Citation: Journal of Climate 38, 1; 10.1175/JCLI-D-24-0093.1
4) Verification using precipitation isotopes
Higher ISM moisture contributions relate to more intense convective activity (Singh et al. 2021), and more depleted precipitation isotopes, vice versa for higher West or TP moisture contributions. The modeling and clustering results of various moistures were verified by comparing them with precipitation δ18O values. The SE and NE moistures were not seriously considered due to their low proportions in the two typical interfaces (Fig. 5a). The spatial relationship between precipitation δ18O and moisture proportions (Fig. 5) was to verify the multiyear (1951–2020) modeling and clustering results; the temporal variabilities shown in Fig. 6 verified the temporal trends in the proportionality of moisture contributions. The steps taken in these verification processes were as below. 1) Two typical interfaces that both represent moisture transportation pathways (southwest–northeast and south–north) over the TP (Tian et al. 2001; Yao et al. 2013) and have precipitation isotope records were selected. The spatial variabilities in precipitation δ18O and moisture contributions for these two typical interfaces [Fig. 5(middle)] were then determined using precipitation isotopes extracted from Isoscape, GNIP, TNIP archives, and our research group. All six moisture sources were presented, but only the West, ISM1, and ISM2 moistures showed significant variabilities. 2) Precipitation δ18O values from the Isoscape archive were resampled to 2.5° × 2.5° resolution and regressed with moisture sources for regions of China in our study area [Fig. 5(down)]. The West, ISM1, ISM2, and ISM1 + ISM2 moistures were considered. 3) Lhasa is affected by various moistures, and it has long-term (1986–2012) precipitation isotope archive, so this was used to show temporal variabilities in the proportions of these different moisture sources as well as build linear regressions between precipitation δ18O and said moisture sources (Fig. 6). The West, ISM1, ISM2, ISM1 + ISM2, and TP moistures were considered; the corresponding linear regressions were provided in Fig. 6f.
3. Results and discussion
a. Moisture source proportions (%) and their temporal trends
Figure 3 shows the moisture source proportions (%) in summer (colored background) and the positive/negative trends (±) during 1951–2020. The West moisture dominated the northwestern sector of the study area, the proportions being as high as >70%. For inner TP, the western and northern regions were affected by West and gradually decreasing from the northwestern to southeastern TP (Fig. 3a). The ISM1 and ISM2 moistures affected the southwestern and southeastern regions of the study area, with highest proportions reaching ∼40% and >80%, respectively (Figs. 3b,c). The ISM1 moisture affected the southwestern margin of TP region alongside the Himalaya Mountains. Especially high ISM2 proportions were observed in southern and eastern TP; the spatial distributions of ISM2 (gradually decreasing alongside a southeast–northwest band) indicated the ISM2 transportation routes when entering the TP. Figure 3d shows that moisture from the TP was <20% for most regions. The western of inner TP had the highest TP moisture values, indicating intensive local recycling in these dry areas with rare external moistures. The SE moisture affected the eastern margins of the study area, principally beyond the TP’s eastern boundary. The NE moisture affected the northeastern corner of the study area, with contributions of ∼15%. The SE and NE moistures can rarely affect the regions within TP boundary and its western parts, where spatially inhomogeneous precipitation changes were widely concerned (T. Yao et al. 2022; Jiang et al. 2023). The proportions of varied moisture sources evinced significant spatial variabilities. Each moisture had a region where it predominates, balkanizing the study area into several sectors: West-northwest sector, ISM1-southwest sector, ISM2-southeast sector, TP-central TP region, SE-east sector, and NE-northeast sector (Fig. 3g).
Trends in proportions of different moisture sources during 1951–2020 were then mapped, with those significant at p < 0.1 marked with blue circles (Fig. 3, Fig. S4). The TP, SE, and NE moistures increased in most sectors of the study area. For each moisture, they had a certain domain where its proportions are especially high (Fig. 3f) and then its positive or negative trends will play important roles in their domains. The western parts of the TP (west of 90°E) experienced significantly increasing TP moisture; in eastern parts of the TP, TP moisture decreased insignificantly. The trends shown by the three major moisture sources (West, ISM1, and ISM2) also exhibited distinct variabilities across the study area. The West moisture generally increased in western and northern TP (Fig. 3a) (regions with especially high West proportions), defined as the causes of water resource patterns in northern TP (T. Yao et al. 2022). The observed decreasing West in central TP is uniform to the northward shift of M-W boundary (the boundary between the ISM- and westerlies-controlled zones) (Guo et al. 2023b), while it may be nonsignificant for precipitation changes because of its low proportions there. Increasing contributions from the ISM1 moisture were principally distributed in the southwestern sector of the study area, significant at p < 0.1. Contributions from the ISM2 moisture decreased in most sectors of the study area, but exceptions to this were found in eastern TP. Figure S2 shows the specific changing rates of moisture proportions. The ISM2 moisture did decrease in surrounding the TP region, while more ISM2 was transported to inner TP (increase) (Fig. S2c). Combining the generally inverse patterns for trends of ISM1 and ISM2, the spatial patterns of ISM2 may indicate the intertwined competition between the western and eastern arm of ISM (ISM1 and ISM2). Besides that, with the general strengthening ISM (T. Yao et al. 2022; Guo et al. 2023a; Fig. S6), the moisture transported by ISM may show varied distributions in its western and eastern arms (ISM1 and ISM2). The southeastern TP region had the highest ISM2 values and negative trends, and thus its affections to precipitation changes there were significant; these are in consistent with precipitation decreases in southeastern TP (Guo et al. 2024), verifying the reliable results in ISM2. Trends in the proportional contributions made by various moisture sources that are predominant in different sectors of the TP are critical to isolate, as these contributions are likely to be the primary controls that affect future water resource distributions. The trends of different moistures in summer can be summarized as increasing West moisture in the northwestern sector of the study area; significant increases in ISM1 contributions in the southwestern sector of the study area; decreases in ISM2 moisture in the southeastern sector of the study area; significant increases in TP moisture in the western part of central TP; increases in SE in the region to the east of TP boundary; and increasing NE in the northeastern sector of the study area. The strengthening or weakening trends for various moistures in Fig. 3f are only applicable for specific regions where a certain moisture dominated, but not for the entire TP. In conclusion, each moisture source has its own sector where it is the predominant source, and each shows specific temporal variabilities within the area in which it predominates (Fig. 3g). The moisture proportions and their trends showed similar spatial distributions in summer and annually (Fig. S4). The combination of the proportions of these six moisture sources and their temporal trends determines spatiotemporal water resource distributions over the TP.
The former moisture source proportions and trends were based on the 10-day back trajectories. How may the spatial patterns change for moisture sources at varied days back? We then compared the summer results at back-tracing days 02–04–06–08–10 (Fig. 4), as presentation; the four main moistures (West, ISM1, ISM2, and TP) were included. The annual comparison is supplied in Fig. S5. Though varied days of back trajectories, the moisture proportions showed similar spatial patterns; true for West, ISM1, ISM2, and TP. While, the specific values of moisture proportions changed with varied back-tracing days, especially for ISM1, and TP. For example, the longer days of moisture tracing, the lower moisture proportions. The ISM1 in southeastern sector reached >70% at 02 day, and it decreased to ∼40% at 10 day. For TP moisture, it was ∼50% at 02 day and down to <20% at 10 day. For ISM2, the spatial patterns and scopes of specific proportions were similar, while the number of grids with high proportions was decreasing from 02 to 10 day. The West moisture at varied tracing days showed little differences, perhaps because of the especially strong dominance of West to our study area. The annual results were similar (Fig. S5) to summer; the low ISM1 and ISM2 proportions may be due to dilution in spring, autumn, and winter with rare ISM moisture. For the trends of varied moistures, they had similar spatial patterns, being more significant for longer tracing days. There were blank trends for grids with extremely low proportions for certain moisture and at shorter tracing days, e.g., the northeastern sector for ISM at 02 day and the northwestern sector for ISM2 at 02 day. At the region where one certain moisture predominates, their specific trends were the same at varied tracing days: West-northwest sector-increase, ISM1-southwest sector-increase, ISM2-southeast sector-decrease, and TP-central TP region-increase. Similar characteristics could be found for trends annually (Fig. S5). The moisture proportions in varied back-tracing days provided us the insights into moisture transportation processes over the TP. The gradually lower ISM1 and ISM2 with longer tracing days indicated long moisture transportation pathways; vise the inverse for TP moisture (which was traced out of the TP boundary at longer tracing days). The West moisture was rarely varied with tracing days change, due to its persistent dominance to our study area. The apparent differences observed for ISM1 and ISM2 moistures in varied tracing days between annually and summer were result from the seasonal variations of ISM, prevailing in summer. The moistures of West and TP did not vary much in different seasons, and therefore the patterns in summer and annually were generally the same.
As in (a)–(f) in Fig. 3, but for moisture sources at back-tracing days 2 (d-02), 4 (d-04), 6 (d-06), 8 (d-08), and 10 (d-10) in summer. Moistures including West, ISM1, ISM2, and TP.
Citation: Journal of Climate 38, 1; 10.1175/JCLI-D-24-0093.1
Our findings were compared with previous moisture source detections, ISM activity, and precipitation or water resource patterns over the TP. The moisture sources present over the TP and its surroundings were identified using various datasets and analytical methods (Chen et al. 2012; Feng and Zhou 2012; Ma et al. 2018). Curio et al. (2015) showed that moisture from outside the TP (local recycling) accounted for ∼37% (∼63%) of the total. C. Zhang et al. (2017) estimated that >69% of TP moisture had a terrestrial source, with ∼21% coming from the ocean. The modeled results from Pan et al. (2019) demonstrated that moisture from the tropical Indian Ocean (central TP) was the dominant supplier to the southern (northern) TP in terms of summer precipitation events. Qi et al. (2016) explained the spatial distributions of TP precipitation, and results showed that moisture from the Bay of Bengal, the east, the south, and west contributed 32.56%, 23.59%, 23.48%, and 20.37% to the total. Li et al. (2022) tracked and quantified the precipitation moisture sources of different drainage basins over the TP; terrestrial moisture dominated in all the subbasins, whereas oceanic moisture contributions in Yarlung Zangbo River Basin (southern TP) were 38%–41%. Although no consistent conclusions have reached, mainstream viewed that moisture from inland sources/local recycling (or ISM) dominates the northern (or southern) TP. The spatial patterns in various moisture proportions in this study (Fig. 3) are consistent with previous research; the extraordinarily low proportions of TP moisture (<20%, Fig. 3d) at 10 day may result from long back trajectories, during which the TP moisture has been traced out of the TP region; the TP values reached to ∼50% at shorter tracing days (Fig. 4). The moisture proportions are controlled by large-scale circulations (e.g., the ISM and the westerlies), the circulations themselves evinced spatiotemporal variabilities (Guo et al. 2023a). With the zonal index showing the strength of midlatitude westerlies, the deuterium excess records in ice core implicated generally strengthening westerlies in recent decades (Zhao et al. 2012; Fig. S7); these were further verified by different glacier status over the TP (Yao et al. 2012). The strengthening westerlies along their northern axis noted by Sha et al. (2020) is consistent with the increasing contributions from the West moisture shown in Fig. 3a (the northern and western TP) (O’Connor et al. 2021; T. Yao et al. 2022; Bridges et al. 2023). There also exists inconsistent activity of westerlies, which may be due to varied study period and different regions it denotes (Jiang et al. 2023). Guo et al. (2023a) observed ISM activity over the southern TP; the overall strengthening and spatially inhomogeneous variabilities of ISM activity (strengthening in the western part and weakening in eastern part) are the same as the positive/negative ISM1/ISM2 trends shown in Figs. 3b and 3c. Indeed, this can be further verified by the generally northward and inversely southward shifts in the M-W boundary over the TP (Guo et al. 2023b).
Moisture sources and their proportions will affect precipitation and water resource distributions. The spatially inhomogeneous changes in precipitation and water resource instability over the TP are widely concern (G. Zhang et al. 2017; Jiang et al. 2023). The south–north dipole precipitation changes (south drying-north wetting) over the TP were reported (Jiang et al. 2023; Ma et al. 2023), which can be attributed to significantly decreased vertical moisture advection over the TP’s southern slope (Z. Wang et al. 2022) and human activity or internal climate variability (Jiang et al. 2023). In general, precipitation is increasing in central and northern TP, whereas significant precipitation decreases have been observed in regions surrounding the TP (Yao et al. 2013; Yin et al. 2013; Tong et al. 2014; Deng et al. 2017; Guo and Tian 2022). Yang et al. (2011) reviewed the spatial patterns in precipitation, evaporation, river runoff, and soil moisture changes; the most significant decreases were found in TPSE, the region influenced by a weakening ISM2 (Fig. 3c). This is also the case for trends in total water storage (G. Zhang et al. 2017). The consistency between changes in water-related resources (previous studies) and the proportions of various moistures and their trends (this study) confirm the credibility of this study’s quantitative moisture source detections.
b. Verification using precipitation isotopes
The various moisture sources identified in this study display distinct characteristics. For example, the West and TP moistures are dry because of the dry climate conditions of their provenances, long transportation pathways, or intense recycling. Moisture associated with the ISM is humid due to humid climate conditions prevalent above the Indian Ocean and along the moisture transportation pathways to the TP (Yao et al. 2013). The proportions of different moisture sources affect not only precipitation amounts/intensities but also the elements within precipitation. Precipitation δ18O is an ideal tracer that can be used to reflect moisture sources or transportation pathways, providing insights into the proportions of moistures that have been sourced from humid or dry provenances. Usually, moisture from dry continental areas renders enriched precipitation δ18O; conversely, humid ocean climate sources give rise to intensive precipitation and depleted precipitation δ18O values. However, although the SE moisture has a humid provenance, its proportional contributions are extremely low (Fig. 3e); this is also true for the dry NE moisture (Fig. 3f). Precipitation δ18O values are therefore valuable to capturing the characteristics of the principal moisture sources in the study area. Spatial comparisons between moisture contributions and precipitation δ18O values were conducted to verify the moisture source clustering rendered by the HYSPLIT model (Fig. 5). Temporal comparisons at one typical site (Lhasa, affected by ISM, West, and TP moistures) were made to verify the trends in proportional moisture contributions (Fig. 6).
(a) The locations of precipitation isotope sites (Isoscape, GNIP, and TNIP), set against a background of ISM1 + ISM2 moisture proportions (%); (b),(c) the spatial distributions of precipitation δ18O values (‰; gray columns or blue squares) and moisture sources (%; colored curves) along the left/right interfaces in (a); (d)–(g) scatterplots and linear regressions between precipitation δ18O (from Isoscape) and moisture proportions, for (d) the West, (e) the ISM1, (f) the ISM2, and (g) the ISM1 + ISM2. The positive/negative r values at p < 0.01 confidence level (***) are provided.
Citation: Journal of Climate 38, 1; 10.1175/JCLI-D-24-0093.1
(a)–(e) Temporal variabilities in precipitation δ18O (black mattes) and various moisture sources (circles) during 1986–2012 at Lhasa (green triangle in Fig. 5a); (f) the corresponding linear regressions, with r values and confidence levels (*/**/*** at p < 0.1, p < 0.05, and p < 0.01) provided. The symbol ± in (a)–(e) correlates with the positive/negative r values in (f).
Citation: Journal of Climate 38, 1; 10.1175/JCLI-D-24-0093.1
In Fig. 5 (middle), precipitation δ18O exhibits covariabilities with West moisture proportions and antivariabilities with ISM proportions; this is true for each of the two interfaces. For the left interface, the points were located both outside and within the TP boundary. For locations 1–7 (outside the TP), precipitation δ18O was more depleted, with increased contributions from ISM moisture and decreased West moisture contributions. For locations 8–23 (in the TP), precipitation δ18O was more enriched, with increased contributions from West moisture and decreased ISM moisture contributions. The wave-shaped fluctuations in moisture proportions may be due to the geographically locations of precipitation δ18O datasets (along the Left interface) not being identical with the moisture source measurement sites (2.5° grids). However, this minor difference does not change the generally same (contrary) variabilities in precipitation δ18O and the West (ISM) moisture. At the right interface, precipitation δ18O appears generally enriched from south to north, with the ISM2 moisture contributions decreasing and the West contributions increasing, proportionally. Precipitation δ18O became more depleted from locations 1 to 4. As the ISM meets the steep Hengduan Mountains, intensive orogenic precipitation events occur; heavier isotopes are left behind, and lighter isotopes remain in the moisture sources. Thus, the rainfall effect on precipitation δ18O plays a more important role than the moisture proportions when passing through the Hengduan Mountains. In Fig. 5 (below), the moisture proportions and precipitation δ18O values for the whole study area are compared. It can be observed that precipitation δ18O correlates positively with West moisture proportions; the higher the moisture from the westerlies and the drier the moisture, the more enriched precipitation δ18O values. Negative linear relations were found between precipitation δ18O and ISM moisture (i.e., ISM1, ISM2, and ISM1 + ISM2). More intense moisture contributions from the ISM are likely to increase the proportions of humid moisture sources and result in lower (more depleted) precipitation δ18O values. All the regressions were significant at p < 0.01 confidence level.
The spatial verification provided by precipitation isotope analysis for both the study area and two typical interfaces along moisture transportation pathways (Fig. 5) indicates the feasibility of HYSPLIT modeling and moisture clustering methods. Then, what is the relationship between precipitation isotopes and moisture proportions in recent decades? Could these methods capture the variabilities inherent in moisture contributions against a background of sustained climate change? If the answer to these questions is affirmative, this research has the potential to aid studies of modern-day water resources, their past variabilities, and possible future changes. Figure 6 shows the temporal variabilities in precipitation δ18O and moisture proportions during 1986–2012 at Lhasa; the corresponding linear regressions are also provided. Precipitation δ18O appears consistent with the proportions of West and TP moistures, i.e., a year with high (low) West/TP moisture proportions corresponds with high (low) precipitation δ18O values; significant positive linear regressions were also found (p < 0.1). For the humid ISM1, ISM2, and ISM1 + ISM2 moistures, opposite correlations with precipitation δ18O values were observed; these were further verified by their negative linear regressions. These temporal comparisons between precipitation isotopes and moisture sources (Fig. 6) would appear consistent with the spatial verification outlined in Fig. 5. The grids/years with higher ISM proportions indicate more humid moisture, further triggering low precipitation δ18O values, and vice versa for grids/years with higher West moisture. The gentler (and some insignificant) linear regressions in Fig. 6 may reflect the short (27 years) precipitation isotope time series at Lhasa.
Our findings are well in accordance with previous research into precipitation patterns and water resource distributions over the TP. These results were consistent with previous moisture source detections, and the comparisons with precipitation isotopes verified the reasonableness of the results obtained using the HYSPLIT model and the moisture source clustering methods employed in this study. The variable proportions of moisture sources and the trends in these proportions may help determine both present-day climate conditions and potential future variabilities in, or the spatial patterns of, water resource distributions over the TP. An understanding of such variabilities is therefore vitally important. There are caveats that deserve further consideration, e.g., the coarse resolution (2.5°) of moisture modeling affects the credibility especially around huge mountains (complicated geographical conditions); water vapor flux was not considered when defining the proportionality of various moisture sources; only the summer period was analyzed. The verification processes using exactly geographically correlated precipitation isotope datasets could fill the gaps in the extant analysis. Further prospects including a finer resolution modeling especially surrounding the water vapor channels in southern TP, comprehensive studies considering water vapor flux, and seasonal variabilities will be addressed in our future research.
4. Conclusions
Set against a background of intensive climate change, the TP has experienced inhomogeneous changes in precipitation and hydrological cycles. These have led to water resource instability and triggered threats to water use security for millions of people who live in the “Asian Water Tower” region. Moisture sources are the primary controls of these changes and are therefore a topic of wide concern. Relatively few quantitative detections of moisture source proportions or decadal trends in their proportionality have been reported. In this study, the back trajectories (at days 01–10) were traced using the HYSPLIT model for each of the 2.5° grids over the TP. The moisture sources were then clustered into six groups (i.e., the West, ISM1, ISM2, TP, SE, and NE) based on their locations and the work of C. Zhang et al. (2019). The variable proportions of different moisture sources and temporal their trends during 1951–2020 were spatiotemporally delineated and verified versus precipitation isotope (δ18O) records. It was established that 1) each moisture source has its own area where it predominates and where its moisture proportions are especially high (West-northwest sector, ISM1-southwest sector, ISM2-southeast sector, TP-central TP region, SE-the eastern region outside the TP, and NE-northeast sector) and that the West, ISM1, ISM2, and TP moistures are the TP’s principal sources of moisture; 2) the moisture sources have experienced spatially inhomogeneous changes during 1951–2020, with each moisture evincing a particular trend in the area in which it predominates (i.e., the West, ISM1, TP, SE, and NE moistures significantly increases proportionally and the ISM2 decreases); 3) the moisture proportions and temporal trends change with varied days of back trajectories (days 02–04–06–08–10), while their spatial patterns are similar (e.g., the longer days of moisture tracing, the lower the moisture proportions for ISM1 and TP; the trends of moisture proportions are generally more significant for longer tracing days); and 4) when verifying the modeled moisture proportions, precipitation δ18O values correlate positively or covary with dry sources such as West or TP moisture, and inverse is true for humid ISM (ISM1 or ISM2) moisture. The modeled results were well in accordance with previously detected moistures, and these temporal variabilities of moisture sources correspond spatially with precipitation patterns and water resource changes over the TP. These confirmed the credibility of this study’s use of the HYSPLIT model and moisture clustering. There exists an issue; for example, the water vapor flux along with the trajectories is not considered; this may increase the uncertainty, while the spatiotemporal verification with precipitation isotopes may refine the gap to some extent. Further prospects including a finer resolution modeling especially surrounding the water vapor channels in southern TP will be addressed in our future research. It is the primary trail for quantitative moisture source proportions and further their spatial distributions or the temporal variabilities in the recent 70 years for the whole TP and its surrounding regions. This work will highly improve water-related studies and provide further insights into water resource instability in the Asian Water Tower region.
Acknowledgments.
This research is funded by the National Natural Science Foundation of China (42071090), Xizang science and technology planning base and talent project (XZ202401JD0025), the Second Tibetan Plateau Scientific Exploration and Research Program (STEP) (2019QZKK020604), and the National Natural Science Foundation of China (42271143, 42201141, and 41988101). Sincerely thank our senior fellow apprentices, for continuous collections of precipitation isotopes; thank Dongmei Qu for her help when testing the water samples; thank Edward A Derbyshire, who helps us improve the English.
Data availability statement.
The TP boundary is from the National Tibetan Plateau Data Center (http://data.tpdc.ac.cn) with Title “Integration dataset of Tibet Plateau boundary.” The digital elevation model is derived from Shuttle Radar Topography Mission (https://gisgeography.com/free-global-dem-data-sources/). HYSPLIT model codes can be accessed at http://www.arl.noaa.gov/HYSPLIT_info.php with Keyworks “Get/Run HYSPLIT” in the right. The gbl reanalysis to drive the HYSPLIT model during 1951–2020 is from the NCEP/NCAR Global Reanalysis Data Archive (ftp://arlftp.arlhq.noaa.gov/pub/archives/reanalysis/); this web describes how to use the FTP server: https://www.ready.noaa.gov/gbl_reanalysis.php. Precipitation δ18O datasets include Isoscape (http://cisoscape.wz.hwdlszywz.net, S. Wang et al. 2022), the GNIP (https://www.iaea.org/services/networks/gnip), and the TNIP (http://data.tpdc.ac.cn) with Title “Dataset of δ18O stable Isotopes in Precipitation from Tibetan Network for Isotopes (1991–2008).”
REFERENCES
Bridges, J. D., J. A. Tarduno, R. D. Cottrell, and T. D. Herbert, 2023: Rapid strengthening of westerlies accompanied intensification of Northern Hemisphere glaciation. Nat. Commun., 14, 3905, https://doi.org/10.1038/s41467-023-39557-4.
Chen, B., X.-D. Xu, S. Yang, and W. Zhang, 2012: On the origin and destination of atmospheric moisture and air mass over the Tibetan Plateau. Theor. Appl. Climatol., 110, 423–435, https://doi.org/10.1007/s00704-012-0641-y.
Chen, F., L. Ding, S. Piao, T. Zhou, B. Xu, T. Yao, and X. Li, 2021: The Tibetan Plateau as the engine for Asian environmental change: The Tibetan Plateau Earth system research into a new era. Sci. Bull., 66, 1263–1266, https://doi.org/10.1016/j.scib.2021.04.017.
Curio, J., F. Maussion, and D. Scherer, 2015: A 12-year high-resolution climatology of atmospheric water transport over the Tibetan Plateau. Earth Syst. Dyn., 6, 109–124, https://doi.org/10.5194/esd-6-109-2015.
Dansgaard, W., 1964: Stable isotopes in precipitation. Tellus, 16 (4), 436–468, https://doi.org/10.1111/j.2153-3490.1964.tb00181.x.
Deng, H., N. C. Pepin, and Y. Chen, 2017: Changes of snowfall under warming in the Tibetan Plateau. J. Geophys. Res. Atmos., 122, 7323–7341, https://doi.org/10.1002/2017JD026524.
Duan, A., and Z. Xiao, 2015: Does the climate warming hiatus exist over the Tibetan Plateau? Sci. Rep., 5, 13711, https://doi.org/10.1038/srep13711.
Feng, L., and T. Zhou, 2012: Water vapor transport for summer precipitation over the Tibetan Plateau: Multidata set analysis. J. Geophys. Res., 117, D20114, https://doi.org/10.1029/2011JD017012.
Gao, J., 2020: Data set of δ18O stable isotopes in precipitation from Tibetan Network for Isotopes (1991–2008). National Tibetan Plateau Data Center, accessed 19 April 2021, https://doi.org/10.11888/Geogra.tpdc.270940.
Gao, Y., L. Cuo, and Y. Zhang, 2014: Changes in moisture flux over the Tibetan Plateau during 1979–2011 and possible mechanisms. J. Climate, 27, 1876–1893, https://doi.org/10.1175/JCLI-D-13-00321.1.
Gröning, M., H. O. Lutz, Z. Roller-Lutz, M. Kralik, L. Gourcy, and L. Pöltenstein, 2012: A simple rain collector preventing water re-evaporation dedicated for δ18O and δ2H analysis of cumulative precipitation samples. J. Hydrol., 448–449, 195–200, https://doi.org/10.1016/j.jhydrol.2012.04.041.
Guo, D., and H. Wang, 2013: Simulation of permafrost and seasonally frozen ground conditions on the Tibetan Plateau, 1981–2010. J. Geophys. Res. Atmos., 118, 5216–5230, https://doi.org/10.1002/jgrd.50457.
Guo, X., and L. Tian, 2022: Spatial patterns and possible mechanisms of precipitation changes in recent decades over and around the Tibetan Plateau in the context of intense warming and weakening winds. Climate Dyn., 59, 2081–2102, https://doi.org/10.1007/s00382-022-06197-1.
Guo, X., L. Tian, L. Wang, and L. Zhang, 2023a: Spatiotemporal variabilities of the recent Indian summer monsoon activities in the Tibetan Plateau: A reanalysis of outgoing longwave radiation datasets. J. Climate, 36, 3955–3970, https://doi.org/10.1175/JCLI-D-22-0518.1.
Guo, X., L. Wang, and L. Tian, 2023b: Spatial distributions and temporal variabilities of the recent Indian Summer Monsoon northern boundaries in Tibetan Plateau: Analysis of outgoing longwave radiation dataset and precipitation isotopes. Climatic Change, 176, 43, https://doi.org/10.1007/s10584-023-03522-3.
Guo, X., L. Tian, L. Wang, Y. Wang, and J. Zhou, 2024: Controls of the recent precipitation anomalies in the southeastern Tibetan Plateau: From the perspective of Indian summer monsoon activities and moisture sources. Climate Dyn., 62, 399–412, https://doi.org/10.1007/s00382-023-06919-z.
Hugonnet, R., and Coauthors, 2021: Accelerated global glacier mass loss in the early twenty-first century. Nature, 592, 726–731, https://doi.org/10.1038/s41586-021-03436-z.
Immerzeel, W. W., L. P. H. Van Beek, and M. F. P. Bierkens, 2010: Climate change will affect the Asian water towers. Science, 328, 1382–1385, https://doi.org/10.1126/science.1183188.
Jiang, J., and Coauthors, 2023: Precipitation regime changes in High Mountain Asia driven by cleaner air. Nature, 623, 544–549, https://doi.org/10.1038/s41586-023-06619-y.
Klein, J. A., J. Harte, and X.-Q. Zhao, 2004: Experimental warming causes large and rapid species loss, dampened by simulated grazing, on the Tibetan Plateau. Ecol. Lett., 7, 1170–1179, https://doi.org/10.1111/j.1461-0248.2004.00677.x.
Lei, Y., T. Yao, B. W. Bird, K. Yang, J. Zhai, and Y. Sheng, 2013: Coherent lake growth on the central Tibetan Plateau since the 1970s: Characterization and attribution. J. Hydrol., 483, 61–67, https://doi.org/10.1016/j.jhydrol.2013.01.003.
Li, Y., F. Su, Q. Tang, H. Gao, D. Yan, H. Peng, and S. Xiao, 2022: Contributions of moisture sources to precipitation in the major drainage basins in the Tibetan Plateau. Sci. China Earth Sci., 65, 1088–1103, https://doi.org/10.1007/s11430-021-9890-6.
Ma, J., H.-L. Ren, M. Cai, and J. Huang, 2023: Seasonally evolving trends explain the north‐south dipole pattern observed in Tibetan Plateau precipitation. Geophys. Res. Lett., 50, e2023GL104891, https://doi.org/10.1029/2023GL104891.
Ma, Z., Y. Xu, J. Peng, Q. Chen, D. Wan, K. He, Z. Shi, and H. Li, 2018: Spatial and temporal precipitation patterns characterized by TRMM TMPA over the Qinghai-Tibetan Plateau and surroundings. Int. J. Remote Sens., 39, 3891–3907, https://doi.org/10.1080/01431161.2018.1441565.
O’Connor, G. K., E. J. Steig, and G. J. Hakim, 2021: Strengthening Southern Hemisphere westerlies and Amundsen Sea Low deepening over the 20th century revealed by proxy‐data assimilation. Geophys. Res. Lett., 48, e2021GL095999, https://doi.org/10.1029/2021GL095999.
Pan, C., B. Zhu, J. Gao, H. Kang, and T. Zhu, 2019: Quantitative identification of moisture sources over the Tibetan Plateau and the relationship between thermal forcing and moisture transport. Climate Dyn., 52, 181–196, https://doi.org/10.1007/s00382-018-4130-6.
Pritchard, H. D., 2019: Asia’s shrinking glaciers protect large populations from drought stress. Nature, 569, 649–654, https://doi.org/10.1038/s41586-019-1240-1.
Qi, W.-w., B.-p. Zhang, Y.-h. Yao, F. Zhao, S. Zhang, and W.-h. He, 2016: A topographical model for precipitation pattern in the Tibetan Plateau. J. Mt. Sci., 13, 763–773, https://doi.org/10.1007/s11629-015-3522-x.
Rajagopalan, B., and P. Molnar, 2013: Signatures of Tibetan Plateau heating on Indian summer monsoon rainfall variability. J. Geophys. Res. Atmos., 118, 1170–1178, https://doi.org/10.1002/jgrd.50124.
Sha, Y., X. Ren, Z. Shi, P. Zhou, X. Li, and X. Liu, 2020: Influence of the Tibetan Plateau and its northern margins on the mid-latitude Westerly Jet over Central Asia in summer. Palaeogeogr. Palaeoclimatol. Palaeoecol., 544, 109611, https://doi.org/10.1016/j.palaeo.2020.109611.
Singh, B. B., and Coauthors, 2021: Linkage of water vapor distribution in the lower stratosphere to organized Asian summer monsoon convection. Climate Dyn., 57, 1709–1731, https://doi.org/10.1007/s00382-021-05772-2.
Singh, P., and K. Nakamura, 2009: Diurnal variation in summer precipitation over the central Tibetan Plateau. J. Geophys. Res., 114, D20107, https://doi.org/10.1029/2009JD011788.
Stein, A. F., R. R. Draxler, G. D. Rolph, B. J. B. Stunder, M. D. Cohen, and F. Ngan, 2015: NOAA’s HYSPLIT atmospheric transport and dispersion modeling system. Bull. Amer. Meteor. Soc., 96, 2059–2077, https://doi.org/10.1175/BAMS-D-14-00110.1.
Sun, J., K. Yang, W. Guo, Y. Wang, J. He, and H. Lu, 2020: Why has the inner Tibetan Plateau become wetter since the mid-1990s? J. Climate, 33, 8507–8522, https://doi.org/10.1175/JCLI-D-19-0471.1.
Tian, L., V. Masson‐Delmotte, M. Stievenard, T. Yao, and J. Jouzel, 2001: Tibetan Plateau summer monsoon northward extent revealed by measurements of water stable isotopes. J. Geophys. Res., 106, 28 081–28 088, https://doi.org/10.1029/2001JD900186.
Tong, K., F. Su, D. Yang, L. Zhang, and Z. Hao, 2014: Tibetan Plateau precipitation as depicted by gauge observations, reanalyses and satellite retrievals. Int. J. Climatol., 34, 265–285, https://doi.org/10.1002/joc.3682.
Wang, L., and Coauthors, 2021: TP-River: Monitoring and quantifying total river runoff from the Third Pole. Bull. Amer. Meteor. Soc., 102, E948–E965, https://doi.org/10.1175/BAMS-D-20-0207.1.
Wang, S., S. Lei, M. Zhang, C. Hughes, J. Crawford, Z. Liu, and D. Qu, 2022: Spatial and seasonal isotope variability in precipitation across China: Monthly isoscapes based on regionalized fuzzy clustering. J. Climate, 35, 3411–3425, https://doi.org/10.1175/JCLI-D-21-0451.1.
Wang, Z., S. Yang, H. Luo, and J. Li, 2022: Drying tendency over the southern slope of the Tibetan Plateau in recent decades: Role of a CGT-like atmospheric change. Climate Dyn., 59, 2801–2813, https://doi.org/10.1007/s00382-022-06262-9.
Xu, X., C. Lu, X. Shi, and S. Gao, 2008: World water tower: An atmospheric perspective. Geophys. Res. Lett., 35, L20815, https://doi.org/10.1029/2008GL035867.
Yang, K., B. Ye, D. Zhou, B. Wu, T. Foken, J. Qin, and Z. Zhou, 2011: Response of hydrological cycle to recent climate changes in the Tibetan Plateau. Climatic Change, 109, 517–534, https://doi.org/10.1007/s10584-011-0099-4.
Yang, M., F. E. Nelson, N. I. Shiklomanov, D. Guo, and G. Wan, 2010: Permafrost degradation and its environmental effects on the Tibetan Plateau: A review of recent research. Earth-Sci. Rev., 103, 31–44, https://doi.org/10.1016/j.earscirev.2010.07.002.
Yang, R., L. Zhu, J. Wang, J. Ju, Q. Ma, F. Turner, and Y. Guo, 2017: Spatiotemporal variations in volume of closed lakes on the Tibetan Plateau and their climatic responses from 1976 to 2013. Climatic Change, 140, 621–633, https://doi.org/10.1007/s10584-016-1877-9.
Yao, L., J. Lu, W. Zhang, J. Qin, C. Zhou, N. N. Tran, and E. R. Pinagé, 2022: Spatiotemporal analysis of extreme temperature change on the Tibetan Plateau based on quantile regression. Earth and Space Sci., 9, e2022EA002571, https://doi.org/10.1029/2022EA002571.
Yao, T., and Coauthors, 2012: Different glacier status with atmospheric circulations in Tibetan Plateau and surroundings. Nat. Climate Change, 2, 663–667, https://doi.org/10.1038/nclimate1580.
Yao, T., and Coauthors, 2013: A review of climatic controls on δ18O in precipitation over the Tibetan Plateau: Observations and simulations. Rev. Geophys., 51, 525–548, https://doi.org/10.1002/rog.20023.
Yao, T., and Coauthors, 2017: Chained impacts on modern environment of interaction between westerlies and Indian monsoon on Tibetan Plateau (in Chinese). Bull. Chin. Acad. Sci., 32, 976–984.
Yao, T., and Coauthors, 2022: The imbalance of the Asian water tower. Nat. Rev. Earth Environ., 3, 618–632, https://doi.org/10.1038/s43017-022-00299-4.
Yin, Y., S. Wu, and D. Zhao, 2013: Past and future spatiotemporal changes in evapotranspiration and effective moisture on the Tibetan Plateau. J. Geophys. Res. Atmos., 118, 10 850–10 860, https://doi.org/10.1002/jgrd.50858.
You, Q., S. Kang, E. Aguilar, and Y. Yan, 2008: Changes in daily climate extremes in the eastern and central Tibetan Plateau during 1961–2005. J. Geophys. Res., 113, D07101, https://doi.org/10.1029/2007JD009389.
Zhang, C., Q. Tang, and D. Chen, 2017: Recent changes in the moisture source of precipitation over the Tibetan Plateau. J. Climate, 30, 1807–1819, https://doi.org/10.1175/JCLI-D-15-0842.1.
Zhang, C., Q. Tang, D. Chen, R. J. van der Ent, X. Liu, W. Li, and G. G. Haile, 2019: Moisture source changes contributed to different precipitation changes over the northern and southern Tibetan Plateau. J. Hydrometeor., 20, 217–229, https://doi.org/10.1175/JHM-D-18-0094.1.
Zhang, G., and Coauthors, 2017: Lake volume and groundwater storage variations in Tibetan Plateau’s endorheic basin. Geophys. Res. Lett., 44, 5550–5560, https://doi.org/10.1002/2017GL073773.
Zhang, G., and Coauthors, 2020: Response of Tibetan Plateau lakes to climate change: Trends, patterns, and mechanisms. Earth-Sci. Rev., 208, 103269, https://doi.org/10.1016/j.earscirev.2020.103269.
Zhang, Y., B. Li, L. Liu, and D. Zheng, 2021: Redetermine the region and boundaries of Tibetan Plateau (in Chinese). Geogr. Res., 40, 1543–1553, https://doi.org/10.11821/dlyj020210138.
Zhao, H., B. Xu, T. Yao, G. Wu, S. Lin, J. Gao, and M. Wang, 2012: Deuterium excess record in a southern Tibetan ice core and its potential climatic implications. Climate Dyn., 38, 1791–1803, https://doi.org/10.1007/s00382-011-1161-7.
