• Bliss, C. I., 1958: Periodic regression in biology and climatology. Connecticut Agricultural Experiment Station Rep., 55 pp.

  • Boomer, I., , N. Aladin, , I. Plotnikov, , and R. Whatley, 2000: The palaeolimnology of the Aral Sea: A review. Quat. Sci. Rev., 19, 12591278.

    • Search Google Scholar
    • Export Citation
  • Bortnik, V., 1996: Changes in the water-level and hydrological balance of the Aral Sea. The Aral Sea Basin, P. P. Micklin and W. D. Williams, Eds., NATO ASI Series, Vol. 12, Springer, 2532.

    • Search Google Scholar
    • Export Citation
  • Cai, X., , D. C. McKinney, , and M. W. Rosegrant, 2003: Sustainability analysis for irrigation water management in the Aral Sea region. Agric. Syst., 76, 10431066.

    • Search Google Scholar
    • Export Citation
  • Cheng, M. K., , and B. D. Tapley, 2004: Variations in the Earth's oblateness during the past 28 years. J. Geophys. Res., 109, B09402, doi:10.1029/2004JB003028.

    • Search Google Scholar
    • Export Citation
  • CIA, 2012: CIA World Factbook 2012. [Available online at https://www.cia.gov/library/publications/download/download-2012/index.html.]

  • Crowley, J. W., , J. X. Mitrovica, , R. C. Bailey, , M. E. Tamisiea, , and J. L. Davis, 2006: Land water storage within the Congo Basin inferred from GRACE satellite gravity data. Geophys. Res. Lett., 33, L19402, doi:10.1029/2006GL027070.

    • Search Google Scholar
    • Export Citation
  • De Beurs, K. M., , and G. M. Henebry, 2005: Land surface phenology and temperature variation in the International Geosphere–Biosphere Program high-latitude transects. Global Change Biol., 11, 779790.

    • Search Google Scholar
    • Export Citation
  • FAO-UNESCO, cited 2013: AquaStat: Aral Sea basin. [Available online at http://www.fao.org/nr/water/aquastat/main/index.stm.]

  • Forkutsa, I., , R. Sommer, , Y. Shirokova, , J. Lamers, , K. Kienzler, , B. Tischbein, , C. Martius, , and P. Vlek, 2009a: Modeling irrigated cotton with shallow groundwater in the Aral Sea basin of Uzbekistan: II. Soil salinity dynamics. Irrig. Sci., 27, 319330.

    • Search Google Scholar
    • Export Citation
  • Forkutsa, I., , R. Sommer, , Y. Shirokova, , J. Lamers, , K. Kienzler, , B. Tischbein, , C. Martius, , and P. Vlek, 2009b: Modeling irrigated cotton with shallow groundwater in the Aral Sea basin of Uzbekistan: I. Water dynamics. Irrig. Sci., 27, 331346.

    • Search Google Scholar
    • Export Citation
  • Geruo, A., , J. Wahr, , and S. Zhong, 2013: Computations of the viscoelastic response of a 3-D compressible Earth to surface loading: An application to glacial isostatic adjustment in Antarctica and Canada. Geophys. J. Int., 192, 557572.

    • Search Google Scholar
    • Export Citation
  • Ibragimov, N., , S. R. Evett, , Y. Esanbekov, , B. S. Kamilov, , L. Mirzaev, , and J. P. A. Lamers, 2007: Water use efficiency of irrigated cotton in Uzbekistan under drip and furrow irrigation. Agric. Water Manage., 90 (1–2), 112120.

    • Search Google Scholar
    • Export Citation
  • Karimov, A., and Coauthors, 2013: Managed aquifer recharge: The solution for water shortages in the Fergana valley. International Water Management Institute Rep., 51 pp.

  • Kummerow, C., , W. Barnes, , T. Kozu, , J. Shiue, , and J. Simpson, 1998: The Tropical Rainfall Measuring Mission (TRMM) sensor package. J. Atmos. Oceanic Technol., 15, 808816.

    • Search Google Scholar
    • Export Citation
  • Landerer, F. W., , and S. C. Swenson, 2012: Accuracy of scaled GRACE terrestrial water storage estimates. Water Resour. Res., 48, W04531, doi:10.1029/2011WR011453.

    • Search Google Scholar
    • Export Citation
  • Micklin, P. P., 1988: Desiccation of the Aral Sea: A water management disaster in the Soviet Union. Science, 241, 11701176.

  • Micklin, P. P., 2007: The Aral Sea disaster. Annu. Rev. Earth Planet. Sci., 35, 4772.

  • Paulson, A., , S. Zhong, , and J. Wahr, 2005: Modelling post-glacial rebound with lateral viscosity variations. Geophys. J. Int., 163, 357371.

    • Search Google Scholar
    • Export Citation
  • Precoda, N., 1991: Requiem for the Aral Sea. Ambio, 20,109114.

  • Rabus, B., , M. Eineder, , A. Roth, , and R. Bamler, 2003: The shuttle radar topography mission—A new class of digital elevation models acquired by spaceborne radar. ISPRS J. Photogramm. Remote Sens., 57, 241262.

    • Search Google Scholar
    • Export Citation
  • Rodell, M., , and J. S. Famiglietti, 1999: Detectability of variations in continental water storage from satellite observations of the time dependent gravity field. Water Resour. Res., 35, 27052723.

    • Search Google Scholar
    • Export Citation
  • Rodell, M., and Coauthors, 2004: The Global Land Data Assimilation System. Bull. Amer. Meteor. Soc., 85, 381394.

  • Schneider, U., , A. Becker, , P. Finger, , A. Meyer-Christoffer, , M. Ziese, , and B. Rudolf, 2014: GPCC’s new land surface precipitation climatology based on quality-controlled in situ data and its role in quantifying the global water cycle. Theor. Appl. Climatol., 115, 1540, doi:10.1007/s00704-013-0860-x.

    • Search Google Scholar
    • Export Citation
  • Seo, K. W., , C. R. Wilson, , J. S. Famiglietti, , J. L. Chen, , and M. Rodell, 2006: Terrestrial water mass load changes from Gravity Recovery and Climate Experiment (GRACE). Water Resour. Res., 42, W05417, doi:10.1029/2005WR004255.

    • Search Google Scholar
    • Export Citation
  • Singh, A., , F. Seitz, , and C. Schwatke, 2012: Inter-annual water storage changes in the Aral Sea from multi-mission satellite altimetry, optical remote sensing, and GRACE satellite gravimetry. Remote Sens. Environ., 123, 187195.

    • Search Google Scholar
    • Export Citation
  • Small, E. E., , L. C. Sloan, , and D. Nychka, 2001: Changes in surface air temperature caused by desiccation of the Aral Sea. J. Climate, 14, 284299.

    • Search Google Scholar
    • Export Citation
  • Small, I., , J. Van der Meer, , and R. Upshur, 2001: Acting on an environmental health disaster: The case of the Aral Sea. Environ. Health Perspect., 109, 547549.

    • Search Google Scholar
    • Export Citation
  • Spoor, M., , and A. Krutov, 2003: XI. The power of water in a divided Central Asia. Perspect. Global Dev. Technol., 2 (3–4), 593614.

  • Swenson, S., , and J. Wahr, 2006: Post-processing removal of correlated errors in GRACE data. Geophys. Res. Lett., 33, L08402, doi:10.1029/2005GL025285.

    • Search Google Scholar
    • Export Citation
  • Swenson, S., , D. Chambers, , and J. Wahr, 2008: Estimating geocenter variations from a combination of GRACE and ocean model output. J. Geophys. Res., 113, 113, B08410, doi:10.1029/2007JB005338.

    • Search Google Scholar
    • Export Citation
  • Vinogradov, S., , and V. P. Langford, 2001: Managing transboundary water resources in the Aral Sea basin: In search of a solution. Int. J. Global Environ. Issues, 1, 345362.

    • Search Google Scholar
    • Export Citation
  • Vörösmarty, C. J., , P. Green, , J. Salisbury & , and R. B. Lammers 2000: Global water resources: Vulnerability from climate change and population growth. Science,289, 284–288.

  • Wahr, J., , S. Swenson, , V. Zlotnicki, , and I. Velicogna, 2004: Time-variable gravity from GRACE: First results. Geophys. Res. Lett., 31, L11501, doi:10.1029/2004GL01977.

    • Search Google Scholar
    • Export Citation
  • Waltham, T., , and I. Sholji, 2001: The demise of the Aral Sea—An environmental disaster. Geol. Today, 17, 218228.

  • Zavialov, P., , A. Kostianoy, , S. Emelianov, , A. Ni, , D. Ishniyazov, , V. Khan, , and T. Kudyshkin, 2003: Hydrographic survey in the dying Aral Sea. Geophys. Res. Lett., 30, 1659, doi:10.1029/2003GL017427.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Location map: The location and extent of the Aral Sea watershed. Lakes and reservoirs are shown in gray outline and the reservoir and dam locations are denoted by triangles.

  • View in gallery

    Seasonal precipitation averages: The average season precipitation received per month within the Aral Sea watershed is shown. Spring values in the southeast part of the basin reach upward to 175 mm month−1. The desert areas closer to the Aral have little year-round precipitation. Areas with no data are shown as gray and no interpretation was done in those areas.

  • View in gallery

    Average gravity of basin: The average gravity deviation for (left) the entire watershed and (right) the region of the Aral Sea is shown. Each of the three solutions is shown. Annual cycles can be seen in both and an overall negative trend is evident within all three solutions. A slight recovery can be seen in 2010–12 because of large amounts of snow/rain.

  • View in gallery

    GPCC trend map: Trend of precipitation in the Aral watershed using GPCC data is shown for trends (left) from 1980 to 2010 and (right) from 2000 to 2010. Areas with a significant trend are denoted by a dot in each pixel. Gray colors show no trend, warm colors show a decrease in precipitation, and cool colors show increased precipitation.

  • View in gallery

    TRMM trend map: Trend of precipitation in the Aral watershed using TRMM data is shown. Areas with a significant trend are denoted by a dot in each pixel. Gray colors show no trend, warm colors show a decrease in precipitation, and cool colors show increased precipitation.

  • View in gallery

    Overall GRACE trends: A trend map of each of the GRACE solutions is shown. Significant trends are denoted with a black dot in each pixel. The JPL and CSR solutions show very similar trends with GFZ.

  • View in gallery

    Groundwater estimates are shown after the GLDAS models of soil moisture and of ice and snow cover were removed from the total signal. Some areas have no data where the GLDAS model is not valid because of surface water or significant ice cover.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 107 107 12
PDF Downloads 76 76 12

Estimating the Effects of Anthropogenic Modification on Water Balance in the Aral Sea Watershed Using GRACE: 2003–12

View More View Less
  • 1 University of Toledo, Toledo, Ohio
© Get Permissions
Full access

Abstract

The decrease in size the Aral Sea in central Asia, seen as both lower water levels and reduction in areal extent, has been one of the greatest examples of anthropogenic modification of a natural system in recent history. Many studies have monitored the extent and rate of this water loss and provided estimates on the expected life span of the remaining water. However, with little data for groundwater monitoring in the post-Soviet era, it is unclear what the water balance currently is in the remainder of the watershed. Redistribution of water upstream in the watershed including damming to create reservoirs and groundwater recharge from irrigation has not only deprived the sea of water but also increased evapotranspiration and altered local climate patterns. Using Tropical Rainfall Measurement Mission (TRMM) and Global Precipitation Climatology Centre (GPCC) data, rainfall trends for the Aral Sea watershed were analyzed over 10- and 30-yr periods and only minimal changes in rainfall were detected. Using Gravity Recovery and Climate Experiment (GRACE) gravity data from 2003 to 2012, trends in equivalent water mass were determined for the entire watershed. Estimates show up to 14 km3 of equivalent water mass has been lost from the watershed annually from 2002 to 2013. The mass loss throughout the basin is most likely attributable to increased evapotranspiration due to the inefficient irrigation systems and other human modification increasing the need for international cooperation and conservation programs to minimize negative impacts throughout the region.

Corresponding author address: Richard Becker, University of Toledo, 2801 Bancroft St., Toledo, OH 43606. E-mail address: richard.becker@utoledo.edu

Abstract

The decrease in size the Aral Sea in central Asia, seen as both lower water levels and reduction in areal extent, has been one of the greatest examples of anthropogenic modification of a natural system in recent history. Many studies have monitored the extent and rate of this water loss and provided estimates on the expected life span of the remaining water. However, with little data for groundwater monitoring in the post-Soviet era, it is unclear what the water balance currently is in the remainder of the watershed. Redistribution of water upstream in the watershed including damming to create reservoirs and groundwater recharge from irrigation has not only deprived the sea of water but also increased evapotranspiration and altered local climate patterns. Using Tropical Rainfall Measurement Mission (TRMM) and Global Precipitation Climatology Centre (GPCC) data, rainfall trends for the Aral Sea watershed were analyzed over 10- and 30-yr periods and only minimal changes in rainfall were detected. Using Gravity Recovery and Climate Experiment (GRACE) gravity data from 2003 to 2012, trends in equivalent water mass were determined for the entire watershed. Estimates show up to 14 km3 of equivalent water mass has been lost from the watershed annually from 2002 to 2013. The mass loss throughout the basin is most likely attributable to increased evapotranspiration due to the inefficient irrigation systems and other human modification increasing the need for international cooperation and conservation programs to minimize negative impacts throughout the region.

Corresponding author address: Richard Becker, University of Toledo, 2801 Bancroft St., Toledo, OH 43606. E-mail address: richard.becker@utoledo.edu

1. Background and introduction

Extensive research has been done to monitor the extent of the Aral Sea shrinkage (Micklin 2007; Micklin 1988; Spoor and Krutov 2003; Zavialov et al. 2003). Previous research includes Gravity Recovery and Climate Experiment (GRACE) data, which have been used in conjunction with satellite altimetry data to obtain information about reductions in water mass and depth of the sea itself over the past decade using gravity measurements, but less effort has been spent looking at the remainder of the watershed as a whole (Singh et al. 2012). Five former Soviet states share water resources in the region: Kazakhstan, Uzbekistan, Tajikistan, Turkmenistan, and Kyrgyzstan. Much of the water upstream has been impounded by reservoirs in Tajikistan, Kyrgyzstan, and eastern Kazakhstan (Vinogradov and Langford 2001). These reservoirs have increased groundwater recharge in some areas but also have prevented runoff from reaching the sea (FAO-UNESCO 2013). Relocating this much water in the drainage system has implications on water resources throughout the basin both politically and economically because these countries depend on the water for irrigation of crops. Thousands of people have been forced to relocate as water refugees into areas that were already stressed for water resources (I. Small et al. 2001).

The Aral Sea is a closed basin and up until the 1960s net evapotranspiration was roughly equal to precipitation. Boomer et al. (Boomer et al. 2000) summarized extensive studies on exposed and submerged terraces, which record past shorelines, sediment core analyses, micropalaeontological studies, and archaeological investigations. These show that the Aral Sea level has fluctuated in the past as a result of climate change and natural variation in the outlet of the Amu Darya 3000 years before present, when the sea was 73 m above mean sea level (MSL), 20 m higher than the 1960 level of 53 m MSL, and had merged with the Sarykamysh Lake to the south and then dried up again roughly 1600 years before present (Boomer et al. 2000). Before 1960, the Aral Sea level had remained stable for over 150 years, only varying by less than 4 m (Boomer et al. 2000). Soviet programs to grow cotton were implemented in the mid-twentieth century in the region. There have been significant modifications made to the natural system via canals, irrigation ditches, and reservoirs over the past half century. Soviet scientists knew shrinkage of the Aral Sea was a potential repercussion of this modification but proceeded with agricultural expansion anyway (Waltham and Sholji 2001). To better understand what has happened in the watershed as a whole, a few questions must be answered: How much of the water is lost from the watershed through increased evapotranspiration? How much water is stored as groundwater? What are the combined effects of desertification in the areas formerly under the sea and from local climate change? What has the net effect been on water in the basin?

With no central monitoring network it is difficult to quantify the change in water balance within the watershed. In this study, we determined the trends in total water storage in the Aral Sea watershed and discuss the implications. Precipitation trends over the past 30 years were analyzed over the watershed to determine if a significant change in rainfall has contributed to a change in water balance. Specific modifications to the watershed included the diversion of the Amu Darya and Syr Darya for irrigation, such as the Karakum Canal in Turkmenistan, and also for storage in reservoirs, such as the Shardara Reservoir in southern Kazakhstan. These modifications have reduced much of the input into the Aral Sea (Spoor and Krutov 2003; Waltham and Sholji 2001). GRACE data were used to monitor large-scale mass changes and then used to estimate the overall water balance within the Aral Sea watershed.

1.1. Geography

The Aral Sea, located on the border between Kazakhstan and Uzbekistan, was the world’s fourth largest inland water body prior to 1960 at over 68 000 km2, as shown in Figure 1. In just over 50 years, the sea has lost 74% of its surface area and an estimated 90% of its volume (Micklin 2007). This water loss has split the Aral into two distinct basins: the North Aral Sea and the South Aral Sea. The cause of this shrinkage can be attributed to Soviet-era expansion of canals and the irrigation network, most notably the Karakum Canal in Turkmenistan. The diverted water now supports large irrigated agricultural regions and also Uzbekistan’s cotton industry, which is the fifth largest exporter in the world (CIA 2012). Additionally, human population since 1960 in the Aral Sea watershed has grown from 15 million to over 65 million increasing the demand for both food and water (Vörösmarty et al. 2000).

Figure 1.
Figure 1.

Location map: The location and extent of the Aral Sea watershed. Lakes and reservoirs are shown in gray outline and the reservoir and dam locations are denoted by triangles.

Citation: Earth Interactions 18, 3; 10.1175/2013EI000537.1

The Aral Sea is supplied with water by two main rivers and has no natural outlets. The two main rivers in the Aral Sea watershed, the Amu Darya and Syr Darya, both flow from southeast to northwest, eventually emptying into the Aral. As of the late 1990s the Amu Darya only reaches the sea as groundwater as a result of the continued shrinkage and receding shoreline (Precoda 1991). The Amu Darya and Syr Darya are mostly glacial and snow fed by the Pamir Mountains and the Tian Shen Mountains, respectively, in the southeastern part of the watershed. The Amu Darya average annual flow in 2001 was roughly 80 km3 and the Syr Darya flow was roughly 40 km3, which after evaporative losses contributes roughly 55 km3 to the Aral Sea annually (Waltham and Sholji 2001).

The majority of the precipitation falls over the southeastern portion of the basin where annual precipitation can be in excess of 1000 mm annually, shown by Tropical Rainfall Measurement Mission (TRMM) data (Kummerow et al. 1998). As the rivers travel through the basin, the amount of precipitation decreases to the driest areas in the northwest, which receive less than 50 mm annually. Figure 2 shows seasonal precipitation within the watershed. Over the past 10–15 thousand years, fluctuations in the Aral region’s arid climate, periodic shrinkage, and significant sea level changes of up to 40 m in elevation have been recorded and attributed to climate change and anthropogenic factors (Boomer et al. 2000). However, between the mid-eighteenth century and 1960, lake level varied by no more than 4.5 m (Bortnik 1996).

Figure 2.
Figure 2.

Seasonal precipitation averages: The average season precipitation received per month within the Aral Sea watershed is shown. Spring values in the southeast part of the basin reach upward to 175 mm month−1. The desert areas closer to the Aral have little year-round precipitation. Areas with no data are shown as gray and no interpretation was done in those areas.

Citation: Earth Interactions 18, 3; 10.1175/2013EI000537.1

1.2. Anthropogenic modification

Contemporary water use has modified the Aral Sea watershed by construction of engineering structures including over 80 dams and reservoirs along the length of both rivers. The majority of dams were built during the Soviet era and are now located in Tajikistan and Kyrgyzstan the upstream countries in the watershed. This has impounded over 60 km3 of water upstream of the sea. Up to 70% of the water used from both rivers is diverted for agricultural needs, but the irrigation systems are outdated and inefficient, which increases water losses to evapotranspiration (Cai et al. 2003). Because of these outdated irrigation systems, up to 40% of the water allocated to some farmers is lost to evapotranspiration and infiltration on transmission (Forkutsa et al. 2009b). Population expansion has also increased stress on water resources for drinking water and animals. In addition, constant irrigation and evapotranspiration cycles in the region have led to increased soil salinity reducing overall crop yields, which is a symptom of perpetual water mismanagement (Forkutsa et al. 2009a).

During the summer months, rainfall across the entire watershed is minimal and groundwater is much more heavily used in conjunction with withdrawal from reservoirs. Water scarcity during the dry season and reduced flow in the river due to irrigation has increased the reliance on groundwater for agriculture. Irrigation has also locally increased groundwater recharge rates, which causes secondary salinization as the water remains near the surface (Forkutsa et al. 2009b). Programs have been implemented in specific areas to artificially increase winter and spring infiltration into aquifers by diverting water during these wet months, an example of which is found in the Fergana valley in Kyrgystan (Karimov et al. 2013).

1.3. Consequences of water loss

Consequences of the shrinking of the Aral Sea have been severe. A fishing industry that once supported over 50 000 tons of annual harvests completely crashed because salinity rates increased to almost 3 times that of ocean water in the South Aral Sea, in which only brine shrimp, a type of highly saline tolerant shrimp, could survive (Precoda 1991). As the seashore retreated, fishing vessels that once were docked in harbors now are abandoned on dry land. The Aral was heavily polluted prior to 1960 and the chemicals left behind now mix with saltwater to form toxic dust storms that have increased cancer rates and other health problems in the areas surrounding the south and east of the sea (I. Small et al. 2001). Local climate, which was formerly heavily influenced by the Aral Sea, has been altered as a result of the shrinkage, reducing rainfall and average humidity rates and also causing an increase in local air surface temperatures (E. E. Small et al. 2001). Soil salinization from both overirrigation and dust storms from the former sea basin have reduced crop yields in some areas and increased desertification across the region (Micklin 2007). A region that formerly supported a large fishing industry has created many water refugees. These displaced individuals move to other areas in the region, which are also water stressed, exacerbating conditions for locals. In the future, current trends may cause problems for neighboring countries as water supplies continue to decline and crop yields diminish (Spoor and Krutov 2003).

1.4. International water issues

In addition to the many environmental issues, many political issues have arisen since the disbanding of the Soviet Union around 1990. The source of over 70% of the flow within the basin lies in Kyrgyzstan and Tajikistan (Spoor and Krutov 2003). Both countries have developed extensive hydroelectric power, which is an important energy resource for these countries that lack many hydrocarbon resources (Spoor and Krutov 2003). Excess runoff during spring and winter months is stored in the reservoirs for irrigation, drinking water, and power generation. However, the downstream countries of Uzbekistan, Kazakhstan, and Turkmenistan depend on water from the rivers reaching their countries. The Interstate Commission for Water Coordination (ICWC) was established in the region in 1992 between the five countries (Uzbekistan, Kazakhstan, Turkmenistan, Tajikistan, and Krygyzstan) to establish regional cooperation, but most agreements have been nonbinding and little change has been implemented (Spoor and Krutov 2003).

Groundwater and weather monitoring in the region was fairly well documented through 1990, but the consistency of these records has decreased as water management once overseen by the Ministry of Land Reclamation and Water Resources in Moscow through the Ministries of Water ended following the breakup of the Soviet Union. The newly independent governments ended many of these programs in order to dedicate scarce resources to other programs (Spoor and Krutov 2003). GRACE and other remotely sensed products allow environmental assessment of the region at a large scale, which otherwise would not be possible.

2. Data and methods

Gravity Recovery and Climate Experiment (GRACE) measures Earth’s gravity field using twin satellites along the same orbit. Repeat measurements allow monitoring of large-scale mass changes on a month-to-month basis. GRACE was launched in 2002 and continues to obtain monthly gravity anomalies. GRACE data can be converted to equivalent water thickness to show total terrestrial water storage (Wahr et al. 2004). GRACE measures all terrestrial water storage changes as shown by the following equation (Rodell and Famiglietti 1999):
eq1

2.1. Data acquisition

GRACE data in the form of monthly mass grids for the land were obtained from the Jet Propulsion Laboratory (JPL) Tellus website (http://grace.jpl.nasa.gov/data/gracemonthlymassgridsland/) (Landerer and Swenson 2012; Swenson and Wahr 2006). The raw data are spatially averaged to remove errors from noise interference in the gravity field. These data are then filtered to remove correlation errors and reduce data striping. The spherical harmonic coefficients are produced from these data and converted to mass of equivalent water thickness providing level-2 solutions. Three different solutions are produced by the GRACE team including one each from the following: the JPL at the California Institute of Technology, the Center for Space Research (CSR) at the University of Texas at Austin, and the German Research Center for Geosciences (GFZ) at Helmholz Centre Potsdam. The data are based on the RL05 spherical harmonics, using degree-2, order-0 coefficients derived from satellite laser ranging by Cheng and Tapley (Cheng and Tapley 2004), degree-1 coefficients derived by Swenson et al. (Swenson et al. 2008), and removal of postglacial rebound using the model by Paulson et al. (Paulson et al. 2005) as revised by Geruo et al. (Geruo et al. 2013). Mass grids were further processed using a destriping filter and smoothed using a 200-km Gaussian filter (Rodell and Famiglietti 1999). The smoothed data were then segmented into a 1° raster to allow easier comparisons with other datasets. The gridded data were scaled by the land scaling grid (CLM4.SCALE_FACTOR.DS.G200KM) obtained from the same site to restore energy removed destriping, Gaussian, and degree-60 filters for each 1° bin.

2.2. Data processing

The GRACE data were first subtracted from a 5-yr average from January 2003 to December 2008 for comparison across multiple years. All three solutions of GRACE data were georeferenced and stacked into a chronological sequence using Interactive Data Language (IDL). Data for each pixel was then extracted to ASCII format for data analysis. A linear model was applied to determine the annual trend within the data using R software. To minimize the effects of interannual variability within the data, an annual periodic (sine) function was simultaneously fit to the data (Crowley et al. 2006; Wahr et al. 2004).

2.2.1. Precipitation data processing

For precipitation data analysis two datasets were used: the Global Precipitation and Climatology Center version 6.0 (GPCC) and Tropical Rainfall Measuring Mission (TRMM) version 7 (Kummerow et al. 1998; Schneider et al. 2013). GPCC is a monthly precipitation dataset based on ground observations at weather stations from across the globe from 1900 through the present. The dataset used was a 1° × 1° grid. Pixels that contain many stations are highly accurate, whereas pixels that have little or no weather stations use interpolation, which limits accuracy in remote areas, as can be found in parts of our study area. To complement GPCC data, we use the remotely sensed TRMM 3B43 monthly rainfall dataset from 1998 to the present, with coverage from 50°N to 50°S with a 0.25° × 0.25° resolution. TRMM uses microwave and radar sensors to monitor clouds, which augment the ground station data used in the GPCC product. Figure 2 shows average seasonal rainfall from 1900 to 2000 using an interpolation of 180 weather stations.

To analyze trends within precipitation data, a periodic annual function was first fit to the rainfall and a linear trend was established (Bliss 1958). For the GPCC data, both 10-yr (2000–10) and 30-yr (1980–2010) trend maps were produced, and a 10-yr trend map was produced for the TRMM dataset. The findings of the analysis are discussed in the results section of this paper.

2.2.2. Groundwater estimates

To isolate the groundwater component of GRACE, it is necessary to subtract soil moisture, snow and ice, and surface water. Because of the small size of many of the reservoirs, it is difficult to obtain altimetry data, so for purposes of this paper groundwater and surface water are assumed to be connected. The Global Land Data Assimilation System (GLDAS) is a model that provides estimates of land surface climatology using ground observations and interpolation methods. A monthly estimate of soil moisture and snow/ice cover was obtained using the GLDAS–Noah model (Rodell et al. 2004). The values were subtracted from a 5-yr average from 2003 to 2008 and converted to equivalent water thickness for direct comparison with the GRACE dataset. Irrigation and evapotranspiration estimates are provided as part of the GLDAS model produced by Noah. Groundwater estimates were obtained using the following equation:
eq2

2.3. Watershed delineation

Delineation of watersheds in arid climates can prove challenging because many local streams simply flow into the desert and disappear. There is not enough rainfall to sustain small rivers and only the Amu Darya and Syr Darya eventually reach the Aral. Groundwater flow into and out of the basin can be difficult to determine. Some streams may contribute groundwater to the Aral, but little groundwater data are available for the region. Additionally, manmade channels and canals alter watershed boundaries. For the purposes of this study, a combination of surface topography from the Surface Radar Topography Mission (SRTM) and annual precipitation maps from the TRMM dataset were used to delineate the outline of the watershed region to ensure maximum inclusion of runoff areas (Kummerow et al. 1998; Rabus et al. 2003). Figure 1 shows the extent of the watershed used for total water balance analysis with GRACE.

2.4. Water balance

All three GRACE solutions were used to estimate trends in the water balance for the entire watershed. Each 1° × 1° pixel with more than 50% area within the watershed boundary was included within the analysis. Leakage between each pixel was included in the analysis to minimize the effects of duplicate signals. The annual variation for the entire watershed was averaged and multiplied by the total area of the watershed. Time series of JPL, GFZ, and CSR anomalies averaged across the watershed is shown in Figure 3.

Figure 3.
Figure 3.

Average gravity of basin: The average gravity deviation for (left) the entire watershed and (right) the region of the Aral Sea is shown. Each of the three solutions is shown. Annual cycles can be seen in both and an overall negative trend is evident within all three solutions. A slight recovery can be seen in 2010–12 because of large amounts of snow/rain.

Citation: Earth Interactions 18, 3; 10.1175/2013EI000537.1

3. Results and discussion

3.1. Precipitation

Precipitation trends within the region were analyzed over two time periods from 1980 to 2010 and also from 2000 to 2010. Both TRMM and GPCC were used for the 2000–10 analysis, but only the GPCC dataset was used for the 30-yr trends. Figure 4 shows the trend map for the GPCC data. From 1980 to 2010 there was no change in precipitation for the majority of the watershed. From 2000 to 2010 there was a slight increase in precipitation in the southern part of the watershed of a 2–5 mm yr−1 increase in precipitation, within error of a zero trend. Annual precipitation in these southern areas can reach 175 mm month−1 during the wet season, making a 2–5 mm yr−1 increase insignificant on such a scale. Figure 5 shows the trend analysis for the TRMM dataset from 2000 to 2010. Overall, the TRMM dataset is in agreement with the GPCC data, suggesting that there is overall no measurable change in precipitation in the watershed. The TRMM data directly over the Aral Sea surface show an increased rate of rainfall during the period of the TRMM record, which may be a result of evaporative processes forming clouds above this region.

Figure 4.
Figure 4.

GPCC trend map: Trend of precipitation in the Aral watershed using GPCC data is shown for trends (left) from 1980 to 2010 and (right) from 2000 to 2010. Areas with a significant trend are denoted by a dot in each pixel. Gray colors show no trend, warm colors show a decrease in precipitation, and cool colors show increased precipitation.

Citation: Earth Interactions 18, 3; 10.1175/2013EI000537.1

Figure 5.
Figure 5.

TRMM trend map: Trend of precipitation in the Aral watershed using TRMM data is shown. Areas with a significant trend are denoted by a dot in each pixel. Gray colors show no trend, warm colors show a decrease in precipitation, and cool colors show increased precipitation.

Citation: Earth Interactions 18, 3; 10.1175/2013EI000537.1

3.2. Water equivalent depth changes from gravity

Figure 6 shows the trends analysis of the GRACE water equivalent depth data from 2003 to 2012 for each of the three solutions. Both JPL and CSR show very similar patterns of water loss throughout the entire basin. The GFZ solution shows a slight increase in the central part of the basin. In this area, the Shardara Reservoir, which was constructed in the 1960s, provides irrigation for the region. Overspill from the dam fills up the Aydar Kul series of artificial lakes in the region. This slight increase in mass in this central portion of the basin is consistent with infiltration of agricultural irrigation into groundwater. However, the JPL and CSR solutions show continued shrinkage, which would be consistent with inefficient irrigation losing water to evapotranspiration.

Figure 6.
Figure 6.

Overall GRACE trends: A trend map of each of the GRACE solutions is shown. Significant trends are denoted with a black dot in each pixel. The JPL and CSR solutions show very similar trends with GFZ.

Citation: Earth Interactions 18, 3; 10.1175/2013EI000537.1

3.3. Groundwater trend

Figure 7 shows the estimates of groundwater trends within the region. Locations of dams and reservoirs are shown for comparison. Groundwater recharge has increased in the central parts of basin and near agricultural areas most likely as a result of losses irrigation losses to groundwater. Many of the irrigation systems are unlined, which allows infiltration of some of the water through the bottom of ditches and furrows. Reservoirs may also increase recharge depending on the underlying geology. Some areas have no data because the GLDAS model is not valid in areas where there is a significant ice component, such as alpine glaciers. Underestimates in evapotranspiration will lead to overestimates in potential groundwater recharge, which must be considered when interpreting the results.

Figure 7.
Figure 7.

Groundwater estimates are shown after the GLDAS models of soil moisture and of ice and snow cover were removed from the total signal. Some areas have no data where the GLDAS model is not valid because of surface water or significant ice cover.

Citation: Earth Interactions 18, 3; 10.1175/2013EI000537.1

3.4. Irrigation and agricultural transpiration

Within the Aral Sea watershed up to 9% of the total land area consists of agricultural lands (De Beurs and Henebry 2005). The remainder of the land use consists of mostly barren lands and grasslands. Increased cotton production after 1960 increased evapotranspiration rates in these irrigated regions, where irrigated lands lose between 430 and 739 mm to evapotranspiration (ET) from April to October, equivalent to the amount of irrigation applied (Forkutsa et al. 2009b; Ibragimov et al. 2007). Forkutsa et al. (Forkutsa et al. 2009a) found that shallow groundwater partially replenished through irrigation contributed to a majority of the actual evapotranspiration in irrigated cotton fields from 54% to 88%. After evapotranspiration occurs the water mass that is lost becomes atmospheric moisture and the processed GRACE solutions already have atmospheric mass removed (Seo et al. 2006). Any remaining shallow groundwater contributes to the total GRACE mass signal that is the focus of this paper.

3.5. Water balance

Using each of the three GRACE solutions, estimates of the total water balance change within the Aral Sea watershed were produced. Figure 3 shows the average gravity for each of the datasets used. All three processing methods show that the region reached a minimum mass in 2008 and recovery in 2009, 2010, and 2012. The JPL and CSR estimates showed average annual losses for the entire watershed of 14 and 12 km3, respectively, from 2003 to 2012. The GFZ estimate showed a much smaller amount at only 0.5 km3. These both agree in direction but not in magnitude of this shift. Micklin (Micklin 2007) estimated the total losses for the Aral Sea surface at roughly 12 km3 annually from 2001 to 2005. Singh et al. (Singh et al. 2012) used GRACE to predict overall mass loss in just the Aral Sea and estimated a mass loss from mid-2005 through 2008 of roughly 60 or ~24 km3 annually. Other parts of the basin increased in mass slightly presumably in the form of groundwater or reservoir storage, which offsets some of the mass lost within the Aral Sea, resulting in overall mass loss, which is less than other predicted studies. The differences in the GRACE solutions arise from discrepancies in the background models used in the GRACE processing. If errors of commission exist in the GFZ background model, it would overpredict the increase in mass, providing an underestimate in total water loss. Conversely, if errors of omission exist in the JPL and CSR background models, there would be an underestimate of mass, which would overestimate the total mass loss. Overall, interpretation of the GRACE solution, however, shows a general trend of mass loss throughout the majority of the basin and the interpretation of the magnitude of this loss would need to be adjusted according to the corresponding omission or commission errors.

4. Conclusions

The relocation of up to 80 km3 of water upstream in the Aral Sea watershed has provided the basis of the world’s fifth largest cotton exporter and many other agricultural cash crops (Micklin 2007). However, diversion of water combined with inefficient irrigation transport systems has caused the Aral Sea to lose 90% of its volume, as calculated by old bathymetry maps and satellite altimetry, over the past 50 years. Using 10 years of available GRACE data, we estimated current trends in water balance within the Aral Sea watershed and also offer an estimate of groundwater trends by subtracting other known components from the overall gravity signal. The GLDAS model was used to predict soil moisture, evapotranspiration, and snow and ice within the region and subtracted from the total gravity. When using GRACE combined with GLDAS to estimate groundwater trends, it is important to consider the lack of a significant irrigation component in the GLDAS model. This most likely leads to an underestimation in total evapotranspiration and decreased potential mass of infiltration. Additionally, rainfall trends for the region were calculated using both GPCC and TRMM datasets to determine if changes in gravity were due to changes in precipitation patterns.

Over the past 30 years, rainfall in the region has changed very little: water changes in the system can be attributed to anthropogenic modification and increased evapotranspiration. The increasing trends in the central portion of the basin may be attributed to increased infiltration due to surface water storage in the form of reservoirs or loss of water from many of the unlined irrigation channels in the region. While storage of water in the form of groundwater may reduce evapotranspiration, this water does not reach the sea and cannot aid in any recovery. Without any observable changes in precipitation trends over the study time period, the loss of water mass must be attributed to increased evapotranspiration.

As water is removed from the watershed instead of simple relocation into groundwater and reservoirs, agricultural prospects in the region decline and political problems inevitably result. Each country in the region has a different situation financially, but they must work together on the shared resource of water. Investments in irrigation upgrades have decreased total water losses and may continue to improve in the future. With many people’s sustenance and economic livelihood dependent on cotton and grain production in the region, it is important to provide a clearer picture of available water resources in this dry region. Water refugees from the area immediately surrounding the Aral Sea have caused greater stress on other infrastructure and continually increasing populations may stress this further. It is important to continue to monitor the water situation in the Aral Sea watershed, and using GRACE allows predictions about how long water will last in the region. It is hopeful that the situation has improved since the early 2000s and, with savvy water policy, further improvements can be made in the future for each of the countries within the watershed.

Acknowledgments

GRACE land data were processed by Sean Swenson, are supported by the NASA MEaSUREs program, and are available online (at http://grace.jpl.nasa.gov).

References

  • Bliss, C. I., 1958: Periodic regression in biology and climatology. Connecticut Agricultural Experiment Station Rep., 55 pp.

  • Boomer, I., , N. Aladin, , I. Plotnikov, , and R. Whatley, 2000: The palaeolimnology of the Aral Sea: A review. Quat. Sci. Rev., 19, 12591278.

    • Search Google Scholar
    • Export Citation
  • Bortnik, V., 1996: Changes in the water-level and hydrological balance of the Aral Sea. The Aral Sea Basin, P. P. Micklin and W. D. Williams, Eds., NATO ASI Series, Vol. 12, Springer, 2532.

    • Search Google Scholar
    • Export Citation
  • Cai, X., , D. C. McKinney, , and M. W. Rosegrant, 2003: Sustainability analysis for irrigation water management in the Aral Sea region. Agric. Syst., 76, 10431066.

    • Search Google Scholar
    • Export Citation
  • Cheng, M. K., , and B. D. Tapley, 2004: Variations in the Earth's oblateness during the past 28 years. J. Geophys. Res., 109, B09402, doi:10.1029/2004JB003028.

    • Search Google Scholar
    • Export Citation
  • CIA, 2012: CIA World Factbook 2012. [Available online at https://www.cia.gov/library/publications/download/download-2012/index.html.]

  • Crowley, J. W., , J. X. Mitrovica, , R. C. Bailey, , M. E. Tamisiea, , and J. L. Davis, 2006: Land water storage within the Congo Basin inferred from GRACE satellite gravity data. Geophys. Res. Lett., 33, L19402, doi:10.1029/2006GL027070.

    • Search Google Scholar
    • Export Citation
  • De Beurs, K. M., , and G. M. Henebry, 2005: Land surface phenology and temperature variation in the International Geosphere–Biosphere Program high-latitude transects. Global Change Biol., 11, 779790.

    • Search Google Scholar
    • Export Citation
  • FAO-UNESCO, cited 2013: AquaStat: Aral Sea basin. [Available online at http://www.fao.org/nr/water/aquastat/main/index.stm.]

  • Forkutsa, I., , R. Sommer, , Y. Shirokova, , J. Lamers, , K. Kienzler, , B. Tischbein, , C. Martius, , and P. Vlek, 2009a: Modeling irrigated cotton with shallow groundwater in the Aral Sea basin of Uzbekistan: II. Soil salinity dynamics. Irrig. Sci., 27, 319330.

    • Search Google Scholar
    • Export Citation
  • Forkutsa, I., , R. Sommer, , Y. Shirokova, , J. Lamers, , K. Kienzler, , B. Tischbein, , C. Martius, , and P. Vlek, 2009b: Modeling irrigated cotton with shallow groundwater in the Aral Sea basin of Uzbekistan: I. Water dynamics. Irrig. Sci., 27, 331346.

    • Search Google Scholar
    • Export Citation
  • Geruo, A., , J. Wahr, , and S. Zhong, 2013: Computations of the viscoelastic response of a 3-D compressible Earth to surface loading: An application to glacial isostatic adjustment in Antarctica and Canada. Geophys. J. Int., 192, 557572.

    • Search Google Scholar
    • Export Citation
  • Ibragimov, N., , S. R. Evett, , Y. Esanbekov, , B. S. Kamilov, , L. Mirzaev, , and J. P. A. Lamers, 2007: Water use efficiency of irrigated cotton in Uzbekistan under drip and furrow irrigation. Agric. Water Manage., 90 (1–2), 112120.

    • Search Google Scholar
    • Export Citation
  • Karimov, A., and Coauthors, 2013: Managed aquifer recharge: The solution for water shortages in the Fergana valley. International Water Management Institute Rep., 51 pp.

  • Kummerow, C., , W. Barnes, , T. Kozu, , J. Shiue, , and J. Simpson, 1998: The Tropical Rainfall Measuring Mission (TRMM) sensor package. J. Atmos. Oceanic Technol., 15, 808816.

    • Search Google Scholar
    • Export Citation
  • Landerer, F. W., , and S. C. Swenson, 2012: Accuracy of scaled GRACE terrestrial water storage estimates. Water Resour. Res., 48, W04531, doi:10.1029/2011WR011453.

    • Search Google Scholar
    • Export Citation
  • Micklin, P. P., 1988: Desiccation of the Aral Sea: A water management disaster in the Soviet Union. Science, 241, 11701176.

  • Micklin, P. P., 2007: The Aral Sea disaster. Annu. Rev. Earth Planet. Sci., 35, 4772.

  • Paulson, A., , S. Zhong, , and J. Wahr, 2005: Modelling post-glacial rebound with lateral viscosity variations. Geophys. J. Int., 163, 357371.

    • Search Google Scholar
    • Export Citation
  • Precoda, N., 1991: Requiem for the Aral Sea. Ambio, 20,109114.

  • Rabus, B., , M. Eineder, , A. Roth, , and R. Bamler, 2003: The shuttle radar topography mission—A new class of digital elevation models acquired by spaceborne radar. ISPRS J. Photogramm. Remote Sens., 57, 241262.

    • Search Google Scholar
    • Export Citation
  • Rodell, M., , and J. S. Famiglietti, 1999: Detectability of variations in continental water storage from satellite observations of the time dependent gravity field. Water Resour. Res., 35, 27052723.

    • Search Google Scholar
    • Export Citation
  • Rodell, M., and Coauthors, 2004: The Global Land Data Assimilation System. Bull. Amer. Meteor. Soc., 85, 381394.

  • Schneider, U., , A. Becker, , P. Finger, , A. Meyer-Christoffer, , M. Ziese, , and B. Rudolf, 2014: GPCC’s new land surface precipitation climatology based on quality-controlled in situ data and its role in quantifying the global water cycle. Theor. Appl. Climatol., 115, 1540, doi:10.1007/s00704-013-0860-x.

    • Search Google Scholar
    • Export Citation
  • Seo, K. W., , C. R. Wilson, , J. S. Famiglietti, , J. L. Chen, , and M. Rodell, 2006: Terrestrial water mass load changes from Gravity Recovery and Climate Experiment (GRACE). Water Resour. Res., 42, W05417, doi:10.1029/2005WR004255.

    • Search Google Scholar
    • Export Citation
  • Singh, A., , F. Seitz, , and C. Schwatke, 2012: Inter-annual water storage changes in the Aral Sea from multi-mission satellite altimetry, optical remote sensing, and GRACE satellite gravimetry. Remote Sens. Environ., 123, 187195.

    • Search Google Scholar
    • Export Citation
  • Small, E. E., , L. C. Sloan, , and D. Nychka, 2001: Changes in surface air temperature caused by desiccation of the Aral Sea. J. Climate, 14, 284299.

    • Search Google Scholar
    • Export Citation
  • Small, I., , J. Van der Meer, , and R. Upshur, 2001: Acting on an environmental health disaster: The case of the Aral Sea. Environ. Health Perspect., 109, 547549.

    • Search Google Scholar
    • Export Citation
  • Spoor, M., , and A. Krutov, 2003: XI. The power of water in a divided Central Asia. Perspect. Global Dev. Technol., 2 (3–4), 593614.

  • Swenson, S., , and J. Wahr, 2006: Post-processing removal of correlated errors in GRACE data. Geophys. Res. Lett., 33, L08402, doi:10.1029/2005GL025285.

    • Search Google Scholar
    • Export Citation
  • Swenson, S., , D. Chambers, , and J. Wahr, 2008: Estimating geocenter variations from a combination of GRACE and ocean model output. J. Geophys. Res., 113, 113, B08410, doi:10.1029/2007JB005338.

    • Search Google Scholar
    • Export Citation
  • Vinogradov, S., , and V. P. Langford, 2001: Managing transboundary water resources in the Aral Sea basin: In search of a solution. Int. J. Global Environ. Issues, 1, 345362.

    • Search Google Scholar
    • Export Citation
  • Vörösmarty, C. J., , P. Green, , J. Salisbury & , and R. B. Lammers 2000: Global water resources: Vulnerability from climate change and population growth. Science,289, 284–288.

  • Wahr, J., , S. Swenson, , V. Zlotnicki, , and I. Velicogna, 2004: Time-variable gravity from GRACE: First results. Geophys. Res. Lett., 31, L11501, doi:10.1029/2004GL01977.

    • Search Google Scholar
    • Export Citation
  • Waltham, T., , and I. Sholji, 2001: The demise of the Aral Sea—An environmental disaster. Geol. Today, 17, 218228.

  • Zavialov, P., , A. Kostianoy, , S. Emelianov, , A. Ni, , D. Ishniyazov, , V. Khan, , and T. Kudyshkin, 2003: Hydrographic survey in the dying Aral Sea. Geophys. Res. Lett., 30, 1659, doi:10.1029/2003GL017427.

    • Search Google Scholar
    • Export Citation
Save