• <tr id="yyy80"></tr>
  • <sup id="yyy80"></sup>
  • <tfoot id="yyy80"><noscript id="yyy80"></noscript></tfoot>
  • 99热精品在线国产_美女午夜性视频免费_国产精品国产高清国产av_av欧美777_自拍偷自拍亚洲精品老妇_亚洲熟女精品中文字幕_www日本黄色视频网_国产精品野战在线观看 ?

    Global Freshwater Storage Capability across Time Scales in the GRACE Satellite Era

    2021-06-04 08:37:56EndaZHUandXingYUAN
    Advances in Atmospheric Sciences 2021年6期

    Enda ZHU and Xing YUAN

    1Key Laboratory of Regional Climate-Environment for Temperate East Asia (RCE-TEA), Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing 100029, China

    2School of Hydrology and Water Resources, Nanjing University of Information Science and Technology, Nanjing 210044, China

    3College of Earth and Planetary Sciences, University of Chinese Academy of Sciences, Beijing 100049, China

    ABSTRACT

    Key words:freshwater, GRACE, land surface model, soil moisture, climate variability, storage capability

    1.Introduction

    Freshwater is widely viewed as a fundamental natural resource, yet it is threatened by human activities (Meybeck,2003). Over the last 50 years, water consumption has tripled due to global warming, population increase, and urbanization (Carbon Disclosure Project, 2010). Currently, nearly 80% of the global population and 65% of continents suffer from water scarcity (V?r?smarty et al., 2010). Water stress has become an inevitable obstacle to sustainable development, and freshwater security has been listed as one of the grand challenges in the coming decades by the World Climate Research Programme (Trenberth and Asrar, 2014). Terrestrial water storage (TWS) is the most ubiquitous source for high-quality freshwater which not only supports food and livestock production but also influences various aspects of the natural environment, such as affecting sea level(Pokhrel et al., 2012) and the rotation of the Earth (Kuehne and Wilson, 1991).

    The TWS can be divided into surface water, soil water,snow, and groundwater. Recently, due to the vital impacts of water on sustaining human society and ecosystems, its changes over land (e.g., streamflow and TWS) have been extensively investigated. For instance, the Yellow River streamflow displays a persistent decline, and climate factors can explain about 65% of the trend (Piao et al., 2010). For TWS, similar declining trends have been observed in southern and eastern Europe (Stahl et al., 2010), northwestern North America, and the Gulf of Mexico (Kalra et al., 2008).In contrast, a slight increase in streamflow is found over the Yangtze River (Piao et al., 2010) and the Amazon River(Scanlon et al., 2018). In the future, the river discharge is projected to increase over high northern latitudes, India, and Africa, and is expected to decrease in the Mediterranean region, Australia, and parts of North and South America under a high greenhouse gas emission scenario (Schewe et al., 2014). However, the ability of the global land to retain freshwater, which directly influences the freshwater fluxes,receives less attention due to the lack of appropriate methods and global-scale observations.

    McColl et al. (2017) introduced a new metric called,stored precipitation fraction, to quantify the ability of surface soil to retain a positive freshwater anomaly after hours to days during 2016, based on soil moisture observations from NASA’s Soil Moisture Active Passive (SMAP) mission. It can be viewed as a measure of surface soil memory.They found that surface soils (depth of 0-5 cm) accounted for less than 0.001% of the global freshwater storage, but retained 14% of precipitation after three days and that this fraction decreases rapidly as time increases. Hence, for longer prediction, deeper soil moisture and other terrestrial water components should be considered as indicators of TWS. Land-atmosphere coupled modeling experiments have shown land surface conditions, including snow mass and deeper soil, are crucial sources of predictability on seasonal timescales (Koster et al., 2011). The TWS involving various kinds of water can be applied to improve seasonal fire forecasts (Chen et al., 2013), and the memory of TWS can provide additional information for long-term hydrological prediction (Yuan and Zhu, 2018; Zhu et al., 2019).Therefore, the storage capability of the land and its components, such as deeper soil layers, snow, river, lake, and canopy need further investigation, especially on longer time scales. Here, we modify the method proposed by McColl et al. (2017) to quantify the freshwater storage capability(FSC) for land and various TWS components and provide further analysis regarding the land surface storage capability and its hydrological dynamics at different time scales over global major river basins.

    The FSC is jointly controlled by complex factors such as land cover, precipitation, and potential evapotranspiration (PET). Investigating the spatiotemporal distribution of FSC provides implications for the global water balance in a changing climate. Based on the Gravity Recovery and Climate Experiment (GRACE) satellites (Tapley et al., 2004),large-scale terrestrial water storage change (TWSC) can be measured efficiently (Scanlon et al., 2012). In addition,TWSC can also be reasonably simulated by advanced land surface models (LSMs) (Lawrence et al., 2019). The spaceborne observations and state-of-the-art LSMs provide an opportunity to revisit the FSC of global land.

    Considering that GRACE satellites only provide TWS anomaly (TWSA) observations on monthly time scale, we use GRACE TWSA and global precipitation observations to quantify the monthly FSC of the land (including storage within surface water bodies, soil, snow, and aquifers). Then,a state-of-the-art community land model (CLM5)(Lawrence et al., 2019) was applied to provide finer simulation for further analysis regarding the FSC features for TWS, snow, and different soil layers across multiple time scales. Lastly, the study investigates the decadal variability of FSC by connecting it with major modes of internal climate variability.

    2.Data and methods

    In an effort to reduce uncertainty, we apply three global precipitation datasets. These include products sourced from the Global Precipitation Climatology Centre (GPCC) (Udo et al., 2011), the Climatic Research Unit-National Centers for Environmental Prediction (CRU-NCEP) (Viovy, 2018),and the Precipitation Reconstruction over Land (PREC)(Chen et al., 2002). In addition, we use the PET dataset provided by the Global Land Evaporation Amsterdam Model (GLEAM), which maximizes the recovery of evaporation information contained in current satellite observations(Miralles et al., 2011; Martens et al., 2017), and the leaf area index (LAI) dataset which is based on Global Land Surface Satellite (GLASS) (Xiao et al., 2016) from 2001 to 2014.

    The mean of GRACE mascon products provided by the Jet Propulsion Laboratory (JPL) (Swenson and Wahr, 2006;Landerer and Swenson, 2012; Swenson, 2012; Wiese et al.,2016), Center for Space Research (CSR) (Save et al., 2016),and German Research Centre For Geosciences (GFZ)(Dahle et al., 2012) is viewed as the global TWSA observation, which spans 14 years from 2003 to 2016. Furthermore,the SMAP soil moisture from 2015 to 2016 (Entekhabi et al., 2010) is used to validate the model simulation. Besides that, the CLM5 is applied to provide credible, long-term simulations of the TWS and its components, further noting that good model performance has been shown in previous studies (Niu et al., 2007; Xia et al., 2017). In this model, the TWS involves soil moisture, canopy water, snow, and surface water for rivers, lakes, and wetlands. It resolves the water and energy balance from the ground surface down to 8.6 m by dividing it into 20 hydrologically active layers(Lawrence et al., 2019). Compared with the previous models (i.e., CLM4 and CLM4.5), the CLM5 increased the vertical resolution of the soil layer, especially within the top 3 m, and removed unconfined aquifers (Lawrence et al.,2019). Meanwhile, the model is updated to account for spatial variations in soil thickness from a spatially explicit bedrock dataset (Brunke et al., 2016; Pelletier et al., 2016).When soil is below the bedrock, the soil moisture is close to or equal to zero. In this study, we used the biogeophysical part of CLM5, noting that the carbon-nitrogen processes were switched off. Two continuous centurial cycles (i.e., the first cycle ending acts as the initial condition of the second cycle) of CLM5 simulation, with a spatial resolution of onedegree, are carried out during 1901-2016 and were driven by the CRUNCEP observed meteorological forcing datasets (Piao et al., 2012). The last 66-year simulations during 1951-2016 in the second cycle are analyzed and the first 166-year simulations are regarded as land surface model spin-up, noting that the aquifer model needs a long time to reach the equilibrium (Yang et al., 1995).

    To quantify the ability of the land to store freshwater,we use a newly proposed metric called stored precipitation fraction (McColl et al., 2017). The FSC(t) represents the increasing water in land normalized by precipitation as follows,

    where

    where Pis the sum of precipitation in the ith time step,ΔVstands for the change of water for different components of TWS (or the change of TWS) in the ith time step, Δ Vrepresents the increase of water for TWS components (or the increase of TWS) at given time step, ΔTWSis the change of TWS in the ith time step, t is the analysis time step (from one day to one month in this study), and T represents the total time span (e.g., 2003-16). Considering that the change of soil moisture affects infiltration and capillarity, we choose ΔTWSas the only threshold to study the contributions of different TWS components to the total FSC.Because the GRACE data only provides monthly TWSA,the central difference method is applied to calculate the average change in TWS from one month to the other (Zhang et al., 2018),

    where Vis the volume of the ith TWS observation. For CLM5 simulation, we use the TWS value on the last day of the prescribed time step minus the one on the first day of the time step to get the TWSC. Note that McColl et al.(2017) used this precipitation fraction to analyze the surface soil moisture response to precipitation at three-day time scale, while here we use it to quantify the FSC for TWS and soil at different depths from daily to monthly time scales.

    In the water balance equation [d(TWSA)/dt = P - ET -R, where P is precipitation, ET is evapotranspiration, and R is runoff] the impact of runoff, including lateral flow, on the increase of TWS cannot be ignored especially for longer time scales and deeper soil at a given point. However, the precipitation is the only input for a river basin that normally retains water and allows no outflow to other basins (Fig. 1),and the positive change of TWS stands for the volume of precipitation that is retained in a basin. Hence, this study analyzes the FSC of land and soil columns at different depths over global major river basins (Yuan et al., 2015) rather than grids. The selected 194 river basins, based on the Global Energy and Water EXchanges (GEWEX) project,cover most of the land surface representing a broad range of climate and land cover conditions. In this regard, the FSC describes the proportion of precipitation falling on land that can be retained in the basin after a given time, which is closely linked with water security.

    The impacts of the uncertainty of the observations (e.g.,TWS and precipitation) are nontrivial. To estimate the influences of uncertainty in precipitation datasets, we calculate the standard deviations of three groups of FSC based on mean values of TWSA from different centers and the three precipitation datasets, respectively. Conversely, to estimate the influence of GRACE uncertainty, the standard deviations of three groups of FSC based on the mean precipitation and TWSA from different centers are calculated.

    Fig. 1. A schematic diagram of water exchange between different soil columns. TWS includes all the forms of storage,such as river, lake, canopy water, snow, ice, soil moisture, and groundwater.

    Lastly, to investigate the decadal variability of total land water-based FSC, 57 groups of 10-years of CLM5 simulation (i.e., 1951-60, 1952-61, 1953-62, …) are used to estimate FSC, respectively. We also calculate the anomaly correlation coefficient (ACC) between the decadal FSC and the climate indices (i.e., PDO, IPO, and AMO). The monthly climate indices are obtained at https://www.esrl.noaa.gov/psd/data/climateindices/list, and they are aggregated into 10-year running mean values. In this study, the significances of ACC are determined by the student’s T-test.

    3.Results

    Figure 2 shows the FSC based on the average of GRACE TWS from different centers (i.e., CSR, JPL, and GFZ) and the average of precipitation datasets (i.e., GPCC,CRU-NCEP, and PREC). The weighted average of FSC across the global basins is 0.28, which means that 28% of precipitation can be retained by the land after one month.However, the distribution of FSC is spatially uneven over the globe. Generally, the lower FSC is mainly located in monsoonal regions, including southern and northeastern China,eastern North America, and parts of South America and Africa while FSC is large in arid basins, such as the Middle-East, parts of Africa, and the west coast of America. The larger value is indicative of a greater ability to retain a positive anomaly for a river basin, noting that a reliable estimate of FSC relies on the accuracy of observation. Regional differences do exist among different precipitation and TWSA datasets. Therefore, we show the standard deviation of FSC that is due to the uncertainty of TWS and precipitation (Fig. 3ab). We can see that larger standard deviations are mainly located over the arid or semi-arid regions, such as high-latitudes of the Northern Hemisphere, the middle of Asia, parts of the Sahara, and the Arabian Peninsula (Fig. 3a), where a small observation error would exert a great influence on the estimates of FSC. Similarly, the influence of precipitation uncertainty is nontrivial over arid regions (Fig. 3b).However, the impact of precipitation uncertainty is relatively small, except in the Amazon, Congo, Middle East and high latitudes (Fig. 3c), where the in-situ observations are insufficient.

    To explore factors that influence the FSC, we analyze the distributions of the observed FSC which are conditional on the magnitude of the aridity index and LAI (Fig. 4). Previous works (McColl et al., 2017) found that the surface soil FSC is smaller in wet regions due to the significant increases in drainage which occurs when soil moisture increases. Here, we arrive at a similar conclusion that total land water-based FSC significantly increases with the increase of aridity (r = 0.92, p < 0.05) (Fig. 4a). Meanwhile,the impact of LAI on total land water-based FSC is significant and should be considered as well. It is found that FSC decreases with increasing LAI, which means that a greater LAI reduces the ability for land to retain water due to larger evapotranspiration. The correlation between them is?0.36 (p < 0.05) (Fig.4b).

    Due to the availability of GRACE, the simulation with CLM5 is a good choice to analyze FSC for land and its different components, such as soil at different depths and snow,at finer temporal resolution. Following Yuan and Zhu(2018), we utilize the CRU-NCEP dataset, in which precipitation is consistent with CRU at monthly time scale, as the atmospheric forcing for running the CLM5 simulation. We compared the seasonal TWSC from the GRACE and CLM5 simulations spanning the period from 2003 to 2016. The CLM5 simulation presents reasonable distributions of seasonal TWSC over the major global river basins (Fig. 5), and the ACCs between them are 0.66 (MAM), 0.92 (JJA), 0.77(SON), and 0.85 (DJF) (p < 0.05), respectively. Additionally, compared to the SMAP observation from 2015 to 2016, the CLM5 can simulate the climatological distribution of surface soil moisture (5 cm) reasonably well except over high-latitude and rainforest regions where both simulations and observations may have large uncertainties (Fig. 6).The ACC over 194 basins between the SMAP observations and CLM5 simulation is 0.69 (p < 0.01). Therefore, CLM5-simulated TWS and soil moisture are used for the FSC analysis in this study. Compared with the observation (Fig. 2),CLM5 reasonably captures the climatology of the FSC distribution of the land surface (Fig. 7a) in most river basins. The basin weighted average simulated FSC (0.26) is smaller than the observation (0.28), and the ACC between them is 0.45 (p < 0.05). Moreover, the simulated total land waterbased FSC shows obvious seasonal variations, especially at high-latitudes (not shown). It is large in fall (SON) and winter (DJF) (weighted means are 0.24 and 0.5, respectively) and small in spring (MAM) and summer (JJA) (0.10 and 0.07). The phenomena are closely related to freezing and melting processes because winter precipitation at northern latitudes will mostly remain on land as snow and ice until spring. However, it is noteworthy that the CLM5 simulation in this study does not include the impact of human activity, such as pumping, irrigation, and land cover change that are implicitly presented in the GRACE observation. So the differences between them, which are mainly located in North China and Central America, might be attributed to anthropogenic influences.

    Fig. 2. Global distribution of total land water-based freshwater storage capability (FSC). FSC is estimated by an average of GRACE observed terrestrial water storage (TWS) from three centers (i.e., CSR, JPL, and GFZ) and the ensemble means of precipitation observation datasets from GPCC, PREC, and CRUNCEP during 2003-16.

    Fig. 3. The uncertainty of FSC due to TWS observation (standard deviation σTWS) (a),precipitation observation (standard deviation σpr ) (b), and the ratio between the uncertainties( σpr/σ TWS) (c). (a) The uncertainty (standard deviation σTWS) of FSC due to different TWS observations is based on mean values of different precipitation datasets (PREC, GPCC, and CRU-NCEP) and GRACE TWSA from different centers (CSR, JPL, and GFZ) during 2003-16. (b) The uncertainty (standard deviation σpr ) of FSC due to different precipitation observations is based on mean values of TWS from the three centers and different precipitation datasets (PREC, GPCC, and CRU-NCEP) during 2003-16. (c) The standard deviationofFSCdueto precipitationobservations ( σpr) is divided by the standard deviation ofFSCdueto TWSobservations(σ TWS).

    Fig. 4. (a-b) The relationship between FSC and aridity index/LAI over 194 major river basins. In (a) and (b), the anomaly correlation coefficients are 0.92 and ?0.36 (p < 0.05),respectively.

    Land surface water is stored as various components including snow, surface water, canopy water, and soil moisture at different depths. Therefore, the total land waterbased FSC is controlled by a series of complicated land hydrological dynamic processes (e.g., internal drainage, capillary effect). Due to the lack of large-scale and long-term observations of these components, LSMs are widely used to provide spatiotemporal continuous estimations in practical applications. Although more water is stored in deep soil(below 3 m), the snow and soil at the top 3 m play nontrivial roles in the land FSC. Here, we investigate the temporal and spatial characteristics of FSC for various depths of soil and snow based on the CLM5 simulation. To analyze the FSC for surface and deep layer soil, we divide the soil column into three layers including depths of 0-0.1 m (surface soil), 0.1-1 m (middle soil), and 1-3 m (deep soil). It is worth noting that the bedrock depth is less than 3 m, or even 1 m in some areas, and the soil moisture most often is equal to 0 when the soil is below the bedrock, while the averaged depth in most basins is over 3 m. Therefore, in this study,the soil moisture in the 1-3 m (0.1-1 m) column over basins represents the total soil water from 1 m (0.1 m) to 3 m (1 m)or bedrock. In the simulation, the global basins weighted averaged FSC for them are 0.04, 0.11, and 0.03 at monthly time scale, respectively (Figs. 7b?7d). Although a large proportion of TWS is stored in deep soil layers over most river basins, the middle soil column (0.1-1 m) contributes to about 40% of the land FSC, especially in arid or semi-arid regions (Fig. 7c). In addition, the impact of snow should be emphasized in high latitudes where more precipitation is stored as snow instead of soil moisture in the middle soil layer, which is closely related to the large land FSC over these regions (Fig. 7e). Therefore, besides the surface soil water that can be measured by microwave remote sensing(McColl et al., 2017), the middle soil layer and snow should also be considered in the FSC analysis, especially at longer time scales.

    For the global basins, using the CLM5 simulation, we calculate the weighted mean FSC for land, the three layers of soil columns, and snow from day 1 to day 30 (Fig. 8a-b),respectively. In general, total land water-based FSC has decreased by about 38% in 30 days (Fig. 8a). The results based on the CLM5 simulation show that the surface soil FSC rapidly decreases during the first week (Fig. 8a) due to the combined effects of evaporation and internal drainage.The middle soil FSC also decreases from 0.16 to 0.11, but a slight increase in deep soil FSC is noted from 1 day to 30 days, due to the slow water movement. In addition, the decreasing trend of snow FSC is small as well, which contributes more than 20% to the land FSC after three days (Fig. 8b).Despite the large decrease of middle soil FSC across time scales, this layer contributes about 40% to the global total land water-based FSC at various time scales (Fig. 8b). We further note that the contribution by the surface soil is nontrivial at short intervals, but the storage of freshwater by snow plays a more important role than the surface soil after five days (Fig. 8b).

    Fig. 5. Validation of CLM5-simulated (left) seasonal mean terrestrial water storage change (TWSC) against GRACE satellite retrievals (right) averaged over 194 river basins during 2003-16.

    With the time step increasing, the decrease of FSC can be observed globally, especially in monsoonal regions such as western America, southern China, India, and parts of Africa (Fig. 9a-9o). Compared with the work of McColl et al. (2017), the distribution of the surface soil FSC at 3 days(Fig. 9b) is similar, with higher FSC mainly located in arid and semiarid regions, such as parts of the Middle East, central Asia, and northwestern China, where the water cycle at the interface of the atmosphere and the land surface overturns at a slower rate. For the global basins in this study, the weighted averaged FSC for surface soil at three days is 0.09, which is smaller than the value of 0.14 found by McColl et al. (2017), noting that our surface column is thicker than the SMAP observation (5 cm). The middle soil column contributes most to the total land water-based FSC,especially in semi-arid regions such as central Asia, western America, and the Middle East (Figs. 9c, h, m), and the change of deep soil FSC is not obvious. Despite the major impacts of the middle soil layer (Fig. 8b), the impact of snow is comparable to the middle soil in some basins, especially in high latitudes (Fig. 9e), even at short time scales.

    To investigate the decadal variability of global FSC,Figs. 10a-c display the spatial distributions of the ACC between climate indexes (i.e., PDO, AMO, and IPO) and the 10-year moving average of FSC for land from 1951-2016, based on a CLM5 simulation. The total land waterbased FSC of Eastern Australia and parts of the mid-and high-latitudes of Asia presents a positive correlation with the PDO, while the ACCs for most basins in Africa and North America are negative (Fig. 10a). The AMO plays a major role in the Amazon and Sahara where the FSC is large when the AMO is positive, while its impact on Australia is the opposite (Fig. 10b). In addition, the spatial pattern of the ACC between IPO and FSC is similar to that of the PDO, except in the Amazon and Mississippi River basins (Fig. 10c).

    Fig. 6. The climatological distribution of surface soil moisture (5 cm) based on (a)SMAP observation and (b) CLM5 simulation from June, 2015 to December, 2016.

    Fig. 7. The FSC of land (a), soil columns at different depths (b-d), and snow (e). The TWS, soil moisture at different depths and snow are simulated by CLM5 over river basins during 2003-16.

    Fig. 8. (a) The weighted averaged FSC for land, snow, and different soil columns at various depths across time scales for global river basins. (b) The contributions of snow and soil moisture at different depths to the total land water-based FSC. All statistics are based on the data during 2003-16.

    Fig. 9. Distribution of FSC for land, different soil columns, and snow over global river basins for 3-day (a-e), 10-day (f-j),and 20-day (k-o) time scales based on CLM5 simulations during 2003-16.

    Fig. 10. (a-c) Anomaly correlation coefficients between CLM5-simulated total land waterbased FSC and the climate indices (PDO, IPO, and AMO) during 1951-2016.

    4.Conclusions

    This study applies a state-of-the-art method to measure FSC over global major river basins and investigates the associated hydrological dynamics through observations and land model simulations. Based on the retrievals of GRACE satellites and multisource precipitation observations, the estimated land FSC on a monthly time scale is over 0.35 for 25%of the river basins, especially in parts of the Middle East, central Asia, and high-latitude regions. In terms of weighted average, over one-fourth of the rainfall can be retained in basins after one month. Although the simulated total land waterbased FSC is smaller than observations, the correlation between them is significant. In addition, the climate condition and land cover exert great influences on the FSC, with significant positive (negative) correlation between FSC and the aridity index (LAI).

    The GRACE satellites provide valuable measurements for an integrated estimation of global FSC on monthly time scale, and the CLM5 land model simulations can be used to separate the contributions of different TWS components to the FSC across multiple time scales. Although the TWS can be divided into different components, the globally averaged TWS changes can be mainly partitioned between the change of soil moisture and snow (Syed et al., 2008). Similarly, we find that a large proportion of water is stored in the deeper soil, but the middle soil column (0.1-1 m) has the largest impact on the storage capability, especially in some basins,like Western Europe. Although the aquifer is another crucial component of TWS, we were not able to analyze its impacts on FSC because the CLM5 removed the unconfined aquifer module. The aquifer is recharged mainly by the water infiltration through the lower boundary, but we think the impact of aquifer on basin-scale land FSC is relatively small on monthly time scale due to the long residence times of groundwater. Nevertheless, the impacts of groundwater on longer time scales at specific regions need to be further analyzed through appropriate observations and quantification methods. Surface soil contributes more than 25% to the FSC at short time scales especially in arid regions, but it declines quickly over time, noting that snow gradually plays a more important role, especially in high-latitude regions.Therefore, middle soil moisture and snow should be considered in the analysis of the global FSC, especially at longer time scales.

    The GRACE data is able to represent the impact of human activities, but the CLM5 simulations in this study ignored anthropogenic influences. Although the impact of human intervention does not contribute much to the total water storage change in large parts of the world, it should be significant and thus accounted for in areas where human activities are intensive (Haddeland et al., 2014). The difference between FSC based on observation and simulation could partly be attributed to the influence of humans, and the issue deserves more attention in some specific regions,such as North China, India, and North America. Besides that, inter-basin water transfers also need more attention,which is common in many countries such as China, America, India, and Australia (Allison and Meselhe, 2010).

    Due to the long residence time, the variation in TWS containing previous climate information can affect the subsequent weather and climate through a series of complicated land-atmosphere feedback processes. Therefore, the memory from TWS, which can be measured with FSC, is a non-negligible source of climate predictability (Reager and Famiglietti, 2009). Though the memory is always viewed as a stationary feature of TWS or other components, we find that the internal climate variability or external climate forcings can alter the land memory time scale. Here, we show the response of the FSC to decadal climate variability. Our work represented the variability of FSC to climate indexes,such as the PDO, AMO, and IPO. Besides the internal climate variability, human interventions, such as land use/land cover change, and the management of water resources (e.g.,reservoir regulation, irrigation, and groundwater exploitation), are also critical factors directly affecting regional or local FSC, where the analysis of LAI provides a good example. The capability to comprehend such anthropogenic pathways would in turn influence many aspects of hydrology and agriculture such as the water cycle, crop yield, and so on. Therefore, separating the influence of human activities on the changes in FSC is worthy of comprehensive research.

    Acknowledgements. This work was supported by the National Key R&D Program of China (2018YFA0606002), the National Natural Science Foundation of China (41875105), and the Startup Foundation for Introducing Talent of NUIST(2018r078). The CRUNCEP forcing data are available on the UCAR website (https://svn-ccsm-inputdata.cgd.ucar.edu/trunk/inputdata/atm/datm7/). The GPCC precipitation datasets are available at https://www.dwd.de/EN/ourservices/gpcc/gpcc.html and the PREC datasets are available at ftp://ftp.cpc.ncep.noaa.gov/precip/50yr/land_ocean/. The CLM5 is available at CESM website (http://www.cesm.ucar.edu/models/cesm2/). The GRACE data can be downloaded from NASA website (https://grace.jpl.nasa.gov/).

    成人欧美大片| 全区人妻精品视频| 在线播放国产精品三级| 国产av一区在线观看免费| 亚洲精品456在线播放app | 国产国拍精品亚洲av在线观看| 国产一区二区激情短视频| 亚洲狠狠婷婷综合久久图片| 一级a爱片免费观看的视频| 99热只有精品国产| av在线天堂中文字幕| 国产精品女同一区二区软件 | 在线免费观看的www视频| 丝袜美腿在线中文| 亚洲avbb在线观看| 在线a可以看的网站| 亚洲国产精品999在线| 欧美性猛交╳xxx乱大交人| 精品人妻熟女av久视频| 精品久久久久久久人妻蜜臀av| 不卡一级毛片| 国产一区二区在线av高清观看| 桃红色精品国产亚洲av| 精品人妻一区二区三区麻豆 | 精品人妻一区二区三区麻豆 | 观看免费一级毛片| 小说图片视频综合网站| 亚州av有码| 中国美女看黄片| 亚洲精品乱码久久久v下载方式| 国内精品一区二区在线观看| 久久精品国产亚洲av香蕉五月| 女人十人毛片免费观看3o分钟| 色视频www国产| 国产精品不卡视频一区二区 | 亚洲男人的天堂狠狠| 亚洲欧美日韩东京热| 亚洲美女黄片视频| 亚洲欧美日韩卡通动漫| 精品人妻一区二区三区麻豆 | 久久久久久久亚洲中文字幕 | x7x7x7水蜜桃| 国产伦在线观看视频一区| 亚洲精品在线观看二区| 国内毛片毛片毛片毛片毛片| 午夜福利成人在线免费观看| 欧美最新免费一区二区三区 | 性欧美人与动物交配| 在线播放无遮挡| 婷婷精品国产亚洲av在线| 欧美性猛交黑人性爽| 日韩有码中文字幕| 欧美日本视频| 国产一区二区在线观看日韩| 久久6这里有精品| 老司机福利观看| 一个人看视频在线观看www免费| 亚洲美女黄片视频| 夜夜爽天天搞| 亚洲avbb在线观看| 亚洲第一区二区三区不卡| 黄片小视频在线播放| 日韩高清综合在线| 国产色婷婷99| 波多野结衣高清无吗| 成人特级av手机在线观看| 日韩 亚洲 欧美在线| 国产精品电影一区二区三区| 亚洲精品在线观看二区| 欧美激情在线99| 精品久久久久久成人av| 动漫黄色视频在线观看| 少妇的逼好多水| 午夜精品久久久久久毛片777| 日韩大尺度精品在线看网址| 亚洲专区中文字幕在线| 极品教师在线免费播放| 色综合欧美亚洲国产小说| 欧美精品啪啪一区二区三区| 色吧在线观看| bbb黄色大片| 免费人成在线观看视频色| 欧美日韩瑟瑟在线播放| 又爽又黄无遮挡网站| 有码 亚洲区| 国产激情偷乱视频一区二区| 在线免费观看不下载黄p国产 | 亚洲性夜色夜夜综合| 女生性感内裤真人,穿戴方法视频| 午夜免费成人在线视频| 国产男靠女视频免费网站| 国产男靠女视频免费网站| 久久国产精品影院| 老司机深夜福利视频在线观看| 国产一区二区三区视频了| 国产精品久久久久久精品电影| 日韩欧美免费精品| 欧美丝袜亚洲另类 | 嫩草影视91久久| 综合色av麻豆| 久久久国产成人精品二区| 亚洲欧美日韩东京热| 久久这里只有精品中国| 一区二区三区免费毛片| 看片在线看免费视频| 99热只有精品国产| 麻豆国产av国片精品| 国产欧美日韩精品一区二区| 亚洲成人精品中文字幕电影| 中国美女看黄片| 成人美女网站在线观看视频| 精品人妻1区二区| 日韩欧美免费精品| 午夜久久久久精精品| 欧美三级亚洲精品| 久久久久久久午夜电影| 亚洲av不卡在线观看| 国产白丝娇喘喷水9色精品| 日韩欧美一区二区三区在线观看| 精品人妻熟女av久视频| 欧美+日韩+精品| 色噜噜av男人的天堂激情| 国模一区二区三区四区视频| 中文字幕高清在线视频| 少妇被粗大猛烈的视频| 成年人黄色毛片网站| 国产av麻豆久久久久久久| 此物有八面人人有两片| 亚洲av成人av| 日韩欧美免费精品| 日韩欧美国产一区二区入口| 日本精品一区二区三区蜜桃| 亚洲真实伦在线观看| 少妇高潮的动态图| av在线蜜桃| 亚洲自偷自拍三级| 观看美女的网站| 久久午夜福利片| 中国美女看黄片| 免费黄网站久久成人精品 | 1000部很黄的大片| 国产高清视频在线观看网站| 极品教师在线免费播放| 亚洲av第一区精品v没综合| 成年人黄色毛片网站| 性欧美人与动物交配| 欧美高清成人免费视频www| 国产一区二区三区视频了| 亚洲精品成人久久久久久| 国产精品日韩av在线免费观看| 国内毛片毛片毛片毛片毛片| 99在线视频只有这里精品首页| 日韩欧美国产在线观看| 色综合婷婷激情| 99视频精品全部免费 在线| 亚洲国产精品999在线| a级毛片免费高清观看在线播放| 高清毛片免费观看视频网站| 麻豆成人午夜福利视频| 午夜福利成人在线免费观看| 中文资源天堂在线| 亚洲成人久久性| 国产蜜桃级精品一区二区三区| 99国产综合亚洲精品| 性色avwww在线观看| 亚洲久久久久久中文字幕| 搡老熟女国产l中国老女人| 日本免费一区二区三区高清不卡| 黄片小视频在线播放| 在线观看午夜福利视频| 久久草成人影院| 91狼人影院| 午夜福利在线在线| 淫秽高清视频在线观看| 国产成人欧美在线观看| 国产午夜福利久久久久久| 淫秽高清视频在线观看| 亚洲第一电影网av| 人人妻人人澡欧美一区二区| 桃红色精品国产亚洲av| 伊人久久精品亚洲午夜| 亚洲国产欧洲综合997久久,| 色视频www国产| 国产精品1区2区在线观看.| 哪里可以看免费的av片| 精品99又大又爽又粗少妇毛片 | 五月伊人婷婷丁香| 日韩亚洲欧美综合| 波野结衣二区三区在线| 免费看美女性在线毛片视频| 免费在线观看日本一区| 国产精品av视频在线免费观看| 欧美日韩福利视频一区二区| 欧美最新免费一区二区三区 | 亚洲欧美日韩高清在线视频| avwww免费| 能在线免费观看的黄片| 欧美日韩综合久久久久久 | 91久久精品国产一区二区成人| 看免费av毛片| 搡老岳熟女国产| 国产精品嫩草影院av在线观看 | 极品教师在线视频| 亚洲第一欧美日韩一区二区三区| 天美传媒精品一区二区| 免费观看的影片在线观看| 国产黄a三级三级三级人| 亚洲电影在线观看av| 成人国产一区最新在线观看| 97碰自拍视频| 久久久久久久午夜电影| 狂野欧美白嫩少妇大欣赏| 真人做人爱边吃奶动态| 午夜福利高清视频| 女生性感内裤真人,穿戴方法视频| 日本一二三区视频观看| 99久久成人亚洲精品观看| 99国产精品一区二区三区| 变态另类成人亚洲欧美熟女| 国产精品一区二区三区四区久久| 一边摸一边抽搐一进一小说| 丰满人妻熟妇乱又伦精品不卡| 国产精品1区2区在线观看.| 免费看a级黄色片| 免费电影在线观看免费观看| 中文字幕熟女人妻在线| 99久久精品国产亚洲精品| 日日摸夜夜添夜夜添av毛片 | 桃红色精品国产亚洲av| 国产单亲对白刺激| 搞女人的毛片| 久久这里只有精品中国| 午夜老司机福利剧场| 欧美一级a爱片免费观看看| 国产精品综合久久久久久久免费| 草草在线视频免费看| 久久久久国内视频| 男女那种视频在线观看| 午夜福利在线观看免费完整高清在 | 国产在视频线在精品| 精品久久久久久久末码| 男女床上黄色一级片免费看| 亚洲欧美清纯卡通| 成熟少妇高潮喷水视频| 看十八女毛片水多多多| 丰满的人妻完整版| 欧美乱妇无乱码| 久久精品国产亚洲av天美| 亚洲精品一卡2卡三卡4卡5卡| 亚洲熟妇熟女久久| 午夜激情欧美在线| 天堂av国产一区二区熟女人妻| av在线观看视频网站免费| 在线免费观看的www视频| 久久午夜福利片| 亚洲人成网站在线播| 9191精品国产免费久久| 少妇人妻一区二区三区视频| 久久亚洲真实| 国内精品久久久久精免费| 此物有八面人人有两片| 色吧在线观看| 亚洲自偷自拍三级| 一个人看的www免费观看视频| 亚洲中文字幕一区二区三区有码在线看| 亚洲乱码一区二区免费版| 嫩草影院精品99| 成人鲁丝片一二三区免费| 人妻制服诱惑在线中文字幕| 亚洲精品成人久久久久久| 亚洲精品久久国产高清桃花| 免费观看精品视频网站| 大型黄色视频在线免费观看| 国产精品女同一区二区软件 | 午夜激情福利司机影院| 在线观看舔阴道视频| 97人妻精品一区二区三区麻豆| 性色av乱码一区二区三区2| 高潮久久久久久久久久久不卡| 亚洲片人在线观看| 97超级碰碰碰精品色视频在线观看| 日韩中文字幕欧美一区二区| www.www免费av| 伊人久久精品亚洲午夜| 国产精品自产拍在线观看55亚洲| 中文字幕久久专区| 精品国产三级普通话版| 亚洲av电影不卡..在线观看| 九九在线视频观看精品| 丰满乱子伦码专区| 99热精品在线国产| 国产精品,欧美在线| 婷婷丁香在线五月| 深夜a级毛片| bbb黄色大片| 色哟哟哟哟哟哟| 欧美在线一区亚洲| 99久久九九国产精品国产免费| 露出奶头的视频| 夜夜躁狠狠躁天天躁| 精品一区二区免费观看| 亚洲男人的天堂狠狠| 国产真实伦视频高清在线观看 | 色综合婷婷激情| 欧美黄色片欧美黄色片| 久久久久久久亚洲中文字幕 | 国产亚洲欧美98| 赤兔流量卡办理| 嫩草影院精品99| 天天躁日日操中文字幕| 欧美高清性xxxxhd video| 国产精品一区二区三区四区免费观看 | 18禁黄网站禁片午夜丰满| 757午夜福利合集在线观看| 亚州av有码| 国产成人欧美在线观看| 亚洲电影在线观看av| 超碰av人人做人人爽久久| 丰满人妻熟妇乱又伦精品不卡| 精品人妻偷拍中文字幕| 狠狠狠狠99中文字幕| 国内精品久久久久久久电影| 91在线观看av| 亚洲 国产 在线| 色播亚洲综合网| 亚洲久久久久久中文字幕| 欧美日韩黄片免| 亚洲精品在线观看二区| 免费看光身美女| 国产一级毛片七仙女欲春2| 哪里可以看免费的av片| 亚洲精品456在线播放app | 极品教师在线视频| 欧美激情在线99| 亚洲精品乱码久久久v下载方式| 亚洲精品一卡2卡三卡4卡5卡| 亚洲av成人不卡在线观看播放网| 亚洲美女黄片视频| 免费在线观看影片大全网站| 哪里可以看免费的av片| 老司机午夜十八禁免费视频| 在线观看免费视频日本深夜| 亚洲国产欧洲综合997久久,| 成年免费大片在线观看| 熟女电影av网| 午夜福利在线在线| 免费搜索国产男女视频| 免费av观看视频| avwww免费| 热99re8久久精品国产| 免费无遮挡裸体视频| 亚洲内射少妇av| 欧美色欧美亚洲另类二区| 成人高潮视频无遮挡免费网站| 成人永久免费在线观看视频| 日韩中文字幕欧美一区二区| 高清日韩中文字幕在线| 国产淫片久久久久久久久 | 亚洲国产精品成人综合色| 夜夜夜夜夜久久久久| 国产白丝娇喘喷水9色精品| 久久精品影院6| 欧美日韩亚洲国产一区二区在线观看| 成年女人看的毛片在线观看| or卡值多少钱| 最近最新中文字幕大全电影3| 此物有八面人人有两片| 欧美成人a在线观看| 18+在线观看网站| 少妇丰满av| 国产精品嫩草影院av在线观看 | 精品一区二区三区av网在线观看| 午夜影院日韩av| 日韩欧美精品免费久久 | 国产精品久久视频播放| 波野结衣二区三区在线| 俺也久久电影网| 国产精品亚洲一级av第二区| 午夜福利视频1000在线观看| 亚洲中文字幕一区二区三区有码在线看| 给我免费播放毛片高清在线观看| 永久网站在线| 一进一出抽搐动态| 日韩精品青青久久久久久| 热99re8久久精品国产| 大型黄色视频在线免费观看| 日本黄色片子视频| 五月伊人婷婷丁香| 婷婷丁香在线五月| 草草在线视频免费看| 高清在线国产一区| 久久久久久久久大av| 国产69精品久久久久777片| 精品国产亚洲在线| 亚洲成av人片在线播放无| 村上凉子中文字幕在线| 俺也久久电影网| av专区在线播放| 午夜福利欧美成人| 免费黄网站久久成人精品 | 精品午夜福利视频在线观看一区| 亚洲欧美日韩无卡精品| 欧美高清性xxxxhd video| 国产男靠女视频免费网站| 久久久国产成人精品二区| 美女大奶头视频| 成人亚洲精品av一区二区| 3wmmmm亚洲av在线观看| 色尼玛亚洲综合影院| 久久这里只有精品中国| 99热这里只有是精品50| 精品久久久久久久久av| 色5月婷婷丁香| 日本免费a在线| 欧美激情久久久久久爽电影| 性插视频无遮挡在线免费观看| 国产精品亚洲一级av第二区| 好男人在线观看高清免费视频| 中文字幕av成人在线电影| 热99re8久久精品国产| 亚洲美女黄片视频| 亚洲男人的天堂狠狠| 久久精品国产清高在天天线| 99热这里只有是精品50| 综合色av麻豆| 天天一区二区日本电影三级| 久久亚洲真实| 日本黄色视频三级网站网址| 精品人妻熟女av久视频| 国产野战对白在线观看| 日本三级黄在线观看| 一进一出好大好爽视频| 99精品在免费线老司机午夜| 国产精品自产拍在线观看55亚洲| 少妇熟女aⅴ在线视频| 中亚洲国语对白在线视频| 中文亚洲av片在线观看爽| 亚洲国产色片| av中文乱码字幕在线| 亚洲av不卡在线观看| 久久这里只有精品中国| 日韩大尺度精品在线看网址| 久久久精品欧美日韩精品| 757午夜福利合集在线观看| 性插视频无遮挡在线免费观看| 精品国产亚洲在线| 美女 人体艺术 gogo| 一级a爱片免费观看的视频| 欧洲精品卡2卡3卡4卡5卡区| 亚洲中文字幕日韩| 日韩高清综合在线| 不卡一级毛片| 免费在线观看亚洲国产| eeuss影院久久| 又爽又黄a免费视频| 免费高清视频大片| 免费av观看视频| 欧美激情在线99| 亚洲 欧美 日韩 在线 免费| 国产成人aa在线观看| 最好的美女福利视频网| 欧美高清性xxxxhd video| 日本精品一区二区三区蜜桃| 中文字幕高清在线视频| 精品免费久久久久久久清纯| 国产老妇女一区| 国产淫片久久久久久久久 | 麻豆国产av国片精品| 此物有八面人人有两片| 亚洲男人的天堂狠狠| 日韩欧美三级三区| 国产精品不卡视频一区二区 | 亚洲一区二区三区色噜噜| 51国产日韩欧美| 日韩欧美精品v在线| 成熟少妇高潮喷水视频| 天堂动漫精品| 国产亚洲精品综合一区在线观看| 天天躁日日操中文字幕| 免费在线观看影片大全网站| 一级av片app| 宅男免费午夜| 午夜免费男女啪啪视频观看 | 偷拍熟女少妇极品色| 久久久久国产精品人妻aⅴ院| 18禁裸乳无遮挡免费网站照片| 日韩人妻高清精品专区| 又爽又黄无遮挡网站| 国产私拍福利视频在线观看| 欧美高清成人免费视频www| 日日干狠狠操夜夜爽| 99视频精品全部免费 在线| 欧美日韩福利视频一区二区| 少妇被粗大猛烈的视频| 成人美女网站在线观看视频| 欧美性感艳星| 一个人看的www免费观看视频| 久久久国产成人免费| av中文乱码字幕在线| 直男gayav资源| 变态另类丝袜制服| 亚洲国产精品合色在线| 国产亚洲av嫩草精品影院| av天堂在线播放| 老司机福利观看| 国产精品一区二区三区四区久久| 国产精品人妻久久久久久| 老鸭窝网址在线观看| 悠悠久久av| 久久久国产成人精品二区| 美女黄网站色视频| 亚洲国产日韩欧美精品在线观看| 色精品久久人妻99蜜桃| 亚洲 国产 在线| 亚洲美女搞黄在线观看 | 亚洲激情在线av| 成人美女网站在线观看视频| 亚洲av一区综合| 亚洲 欧美 日韩 在线 免费| 桃色一区二区三区在线观看| 亚洲激情在线av| 日韩欧美在线乱码| 国模一区二区三区四区视频| 亚洲精品亚洲一区二区| 黄色丝袜av网址大全| 女生性感内裤真人,穿戴方法视频| 欧美丝袜亚洲另类 | 国产一区二区亚洲精品在线观看| 成人高潮视频无遮挡免费网站| 久久九九热精品免费| 午夜福利18| 男女那种视频在线观看| 偷拍熟女少妇极品色| 一区二区三区激情视频| 日本撒尿小便嘘嘘汇集6| 高清日韩中文字幕在线| 韩国av一区二区三区四区| 午夜免费激情av| 亚洲专区中文字幕在线| 丰满人妻熟妇乱又伦精品不卡| 一个人免费在线观看的高清视频| 国产黄片美女视频| 国内精品久久久久久久电影| 丁香欧美五月| 亚洲三级黄色毛片| 男女之事视频高清在线观看| 亚洲一区二区三区色噜噜| 国产成人欧美在线观看| 天堂√8在线中文| 99久久九九国产精品国产免费| 午夜精品久久久久久毛片777| 中文字幕人成人乱码亚洲影| 高潮久久久久久久久久久不卡| av视频在线观看入口| 国产精品一区二区性色av| 成人亚洲精品av一区二区| 国产精品日韩av在线免费观看| 又黄又爽又免费观看的视频| 精品熟女少妇八av免费久了| 一区福利在线观看| 亚洲中文日韩欧美视频| 夜夜爽天天搞| 波多野结衣高清无吗| 国产又黄又爽又无遮挡在线| 99国产综合亚洲精品| 成人特级黄色片久久久久久久| 免费在线观看成人毛片| 亚洲,欧美精品.| 熟妇人妻久久中文字幕3abv| 欧美成狂野欧美在线观看| 91av网一区二区| 国内精品久久久久精免费| 91久久精品国产一区二区成人| 亚洲专区中文字幕在线| 日韩欧美国产在线观看| 天美传媒精品一区二区| 悠悠久久av| 丰满人妻熟妇乱又伦精品不卡| or卡值多少钱| 亚洲精品在线美女| 中文字幕免费在线视频6| 琪琪午夜伦伦电影理论片6080| 国产精品1区2区在线观看.| 免费在线观看成人毛片| 日本精品一区二区三区蜜桃| 欧美不卡视频在线免费观看| 简卡轻食公司| 国产私拍福利视频在线观看| 久久久久久久久久成人| 能在线免费观看的黄片| 午夜福利欧美成人| 精品久久久久久成人av| 深夜a级毛片| 欧洲精品卡2卡3卡4卡5卡区| 男人舔女人下体高潮全视频| 国产美女午夜福利| 日韩人妻高清精品专区| 无人区码免费观看不卡| 美女免费视频网站| 色尼玛亚洲综合影院| 国产精品久久久久久精品电影| 直男gayav资源| 精品久久国产蜜桃| 久久久久国产精品人妻aⅴ院| 亚洲精华国产精华精| 成人av一区二区三区在线看| 亚洲人成伊人成综合网2020| 精品久久久久久久久亚洲 | 亚洲在线观看片| 国产精品日韩av在线免费观看| 在线十欧美十亚洲十日本专区| 亚洲精品影视一区二区三区av| 麻豆久久精品国产亚洲av| 18禁黄网站禁片午夜丰满| 日韩中字成人|