Spatial Hydrology of the Aral Sea
Using Geographic Information Systems (GIS)

Sandra Akmansoy and Daene C. McKinney, PhD
Center for Research in Water Resources
University of Texas at Austin
December 1997

These materials may be used for study, research, and education, but please credit the authors and the Center for Research in Water Resources, The University of Texas at Austin. All commercial rights reserved. Copyright 1997 Center for Research in Water Resources.

Table of Contents


The Center for Research in Water Resources's involvement in the Aral Sea Project is headed by Dr. Daene McKinney. CRWR researchers are involved in a multi-national water allocation research project with scientists from the Central Asian Republics of Uzbekistan, Kazakstan, Kyrgistan, Tajikistan, and Turkmenestan, which depend on the Aral Sea for survival.

Due to large-scale diversion of the Amudarya and Syrdarya Rivers for crop irrigation, the Aral has lost half its surface area. The scene is Central Asia's Aral Sea, once larger than any of the Great Lakes except Superior, now dried up to one-half its 1960s surface area and less than one-third its volume. Fed since ancient times by two large rivers that were largely diverted since the 1950s to irrigate Soviet cotton fields, the lake has been choked off and has become an environmental disaster. Water quality continues to decrease as the lake level drops, and wind storms on the former sea bottom carry salty grit to area residents, who suffer from respiratory ailments and throat cancer.

The purpose of this page is to give the reader a feeling of the potentiality of available-for-whole-world data, such as the 30 arc-second digital elevation model (DEM) and the Digital Chart of the World, when used in combination with the PC version of ArcView 3.0 and its extension Spatial Analyst. 30 arc-second DEM's (approximately 1 Km² square cells) for the entire world have been developed by the Earth Resources Observation Systems (EROS) Data Center of the United States Geological Survey (USGS), and can be downloaded from their internet site. Instructions for downloading, unzipping, and setting these DEM grids are available at How to Download a 30" DEM?. The Digital Chart of the World is a database of geographic features of the world, developed by the Environmental System Research Institute (ESRI).

Overview of the Study Site

Uzbekistan, Kazakstan, Kyristan, Tajikistan, and Turkeminstan are dependent upon the Aral Sea for irrigation and power generation. Here is a view of the Aral Sea in an Orthographic Projection with central meridian 60°E and reference parallel 42°N.
Figure 1: The globe seen from above the Aral Sea.

A topographic map (i.e., a ramp-shaded presentation of the DEM) of the area is presented in Figure 2 in an Albers projection (Figures2 though 15 will also be in Albers Projection).

Figure 2: Topographic map of Aral Sea, according to the USGS 30" DEM.

A hill shade - 3-D like - representation of the Aral Sea territory is shown in Figure 3.

Figure 3: "Hillshade" map of the Aral Sea, according to the USGS 30" DEM.

A combination of the "hillshade" and the DEM is shown below.

Figure 4: Hillshade + DEM


1. Stream and Watershed Delineation

Stream and Watershed Delineation

The standard methodology for delineating streams and watersheds from a raster digital elevation model (DEM) is based on the eight pour-point algorithm. This algorithm identifies the grid cell, out of the eight surrounding cells, towards which water will flow if driven by gravity. This methodology consists of: To do the stream and watershed delineation, I used an Delineation AML which could be run "in the background" from ARC/INFO without my being there. This is conveninent for large areas which usually require several hours to be processed.
It should be noted that watershed and stream delineation can now be done in the Spatial Analyst extension of ArcView 3.0. However, this also requires you to be present.

Inland Catchment

When the stream and network delination were made, shown here below, all the streams seemed to be flowing towards the Caspian Sea rather than the Aral Sea.
Figure 5: Original Stream Delination

Therefore, an Inland Catchment AML written by Dr. Francisco Olivera which addresses the problem of inland catchments with Digital Elevation Models, during watershed delineation was used. Inland catchments constitute closed hydrologic systems that do not drain to the ocean, as most watersheds. In inland catchments, water drains towards an inland pour point located within the basin, and not towards the basin border, which usually occurs.

Digital Chart of the World

The next step was to use the Digital Chart of the World (DCW) CD Rom to get the streams of the region. The blue lines indicate the thousands of streams which were included in the DCW. The red lines show the streams which were selected from the DCW to be burned in the DEM.
Figure 6: Streams from the Digital Chart of the World

Edited this coverage took a very long time, but the coverage was compared to the European Union TACIS Program coverage. The EU TACIS Program coverage, shown below, was originally in a Transverse projection, and therefore had to converted from Transverse to Albers .

Figure7: European Union TACIS Program Coverage of Digitized Streams

The final coverage of streams to be burned in looked like this:

Figure8: Streams to be Burned In

Burning in the DEM

An extra and prior process has been added to this methodology, and it consists of burning-in the digitized streams that have been observed in the field. This burning-in process consists of raising the elevation of all the cells but those that coincide with the digitized streams. By doing this, water is forced to remain in the streams once it gets there; however, it is not forced to flow towards them. Extensive experience at the CRWR has shown that the streams delineated using this improved methodology represent much better the real stream network. This process consists of: The burned-DEM is shown here below.
Figure 9: Burned DEM

Watershed and Stream Delineation of Burned DEM

In order that water can "flow" across the landscape, any spurious pits have to be filled in. Then, the flowdirection grid can be detrmined. The flowdirection function in grid assigns to each cell a number corresponding to which of the 8 neighbouring cells lies on the path of steepest descent. The cells flow to their nearest neighbour along 1 of 8 compass directions labelled as East = 1, SE = 2, S = 4, SW=8, W=16, NW=32, N=64, NE=128.
Figure 11: Flowdirection Grid

The histogram shown below demonstrates that most of the flow is west bound and north bound.

Figure 12: Histogram of Flow Direction

The flowaccumulation grid, shown here, counts all the cells upstream of a given cell.

Figure 13: Flowaccumulation

Next a grid of the streams was created. Different results are obtained depending on the threshold value of flow accumulation used. In this case, a threshold of 3000 was used. Then the watershed delineation was created by locating the outlet cell at the bottom end of each watershed. This figure shows the delineated streams and watersheds in an Albers Projection.

Figure 14: Stream and Watershed Delineation

As shown below, the delineated streams coincide almost perfectly with the streams from the EU TACIS Program.

Figure 15: Comparison of the Delineated and EU TACIS Streams

Watershed Division

To determine the extend of the watershed within each country, the intersect command was used to intersect the watershed coverage with the political country coverage.
The results were interesting. Most of the basin lies Uzbekistan, Turkmenistan, Afganistan and Kazakhstan. The pie chart below illustrates the divison of the basin more clearly.
Figure 16: Percent Basin in Each Country

The calculated values were then compared to documented values. The area of the Aral Sea basin referenced to January 1987 was found in the Diagnostic Study for the Development of an Action Plan for the Conservation of the Aral Sea. The figure below shows the comparison between the documented and calculated values.

Figure 17: Comparison of the Computed and Documented Basin

One should note that the calculated values for the two upstream countries, Kyrgystan and Tajikistan, are lower than the documented values. However, the calculated values for the downstream countries, Uzebkistan, Tajikistan, and Turkmenistan, are higher than the documented values.

2. Digital Atlas of the World Water Balance

In order to calculate a soil water budget of the area, I obtained data from the CD ROM of The Digital Atlas of the World Water Balance published by the Center for Research in Water Resources. The CD ROM contains 30 by 30 arcsecond cells for precipitation, temperatures, water holding capcity, and radaition for the whole world. The figure below show the water holding cpacity in millimeters for the area. Notice how the land on which the rivers closest to the delta have a much higher water holding capacity.
Figure 18: Soil Water Holding Capacity (mm)

The figure below shows the average temperature in Celsius degrees for the area. (Click on the figure to see an animated version of the temperatures)

Figure 19: Average Temperatures (C)

The figure belows show the monthly net average radiation in watts per square meter.(Click on the figure to see an animated version of the radiation)

Figure 20: Mean Monthly Net Radiation (W/m2)

The figure belows show the annual precipitation in millimeters. (Click on the figure to see an animated version of the precipitation)

Figure 21: Annual Precipitation (mm)

3. Soil Water Balance

Soil water balancing refers to the partitioning of precipitation into evaporation, and runoff or recharge by using a water balance calculation applied to a volume of soil. Seann M. Reed has developped a series of Avenue scripts and exercies to determine the soil water balance.
The input data needed for the soil water balance were the temperature, radiation, water holding capacity and precipitation grids obtained from the Digital Chart of the World, and the extent of the watershed. The Avenue scripts use the Priestley-Taylor equations to calculate the potential evaporation and then the soil moisture budget. However, it should be noted that the script used here do not take snow into consideration. Therefore, when the potential evaporation was calculated some cells contained negative values. These negative values were changed to zero in order to run the soil moisture budget.
The final evaporation values for the area can be seen below. As it can be expected, the evaporation upstream in mountaineous areas are low, while the evaporation values are high in the low plains.
Figure 22: Annual Evaporation (mm/yr)

The first calculation of water surplus was calculuated using Seann Reed's program. This program did not include snowmelt. The figure below show the results of that program. Surplus is defined as water which does not evaporate or remain in soil storage and includes both surface and subsurface runoff.

surplus = precipitation - evaporation - (change in storage)
Figure 23: Annual Water Surplus Without Snow Melt (mm/yr)

 Snow Melt Hydrology

Because of the high altitudes in Central Asia, it was important to incorporate snow melt into the Fortran program. Snow Melt was defined as follows: The snow storage at any time is equal to the snow storage at time before plus the fallen snow within that time minus the snowmelt. The next step was to make sure that the snow storage is never below zero.  The program was run twice: once for  temp >= 2 then rain and for  temp >= 0 then rain.  The results show that there is no difference between using these 2 temperatures.  However, there is a great difference in water surplus when snowmelt isn't calulated.
Figure 24: Water Surplus for Snow at 2C, 0C, and Without Snow Melt (mm/yr)

Figure 25 further illustrates the precipitation versus snowmelt for each country. It also shows how much of the precipitation becomes surplus depending on whether or not snowmelt is calculated.

Figure 25: Annual Precipitation & Water Surplus per each country (mm/yr)
Figure 26 shows the ArcView layout of the water surplus in the Aral Sea basin. The brown cells have extremely high surpluses because those cells have a negative soil water holding capacity (that area is rock or ice).
Figure 26: Annual Water Surplus With Snow Melt as 2C(mm/yr)

To find out how much water surplus each country contributes to the total annual water surplus, the ArcInfo intersect command was used. After Afganistan, the upstream countries of Tajikistan and Kyrgyzstan produce the most waters surplus.

Figure 27: Share of the Annual Water Surplus for each Country

Results and Analysis of the Helsinki Rules

The Helsinki Rules written in 1966, and redrafted in May 1997 state that hydrologic factors, such as the extent of the drainage area in the territory of each basin State and the contribution of water by each basin State are important factors in determining water rights.

The results of this GIS study show that Tajikistan and Kyrgyzstan, the two upstream countries, produce the greatest amount of water surplus, while Kazakstan, Uzebkiatan and Turkmenistan, the downstream countries, have the greatest watershed extent.

Figure 28 shows the water produced, consumed, and allocated by the Interstate Commission for Water Coordination. Uzbekistan is by far the country which consumes the most water according to production and allocation.

Figure 28: Production, Allocation, and Consumption of Water by each Country

The five central Asian countries have agreements to share the water. They exchange water releases for natural gas, coal, oil or their monetary equivalent.

Figure 28: Exchange of Resources within Central Asia

To find out more about my analysis of the Helsinki Rules when applied to the Aral Sea basin, read my Master's Thesis, Water Rights in the Aral Sea Basin.

Last updated December1, 1997
email me for any comments or suggestions

These materials may be used for study, research, and education, but please credit the authors and the Center for Research in Water Resources, The University of Texas at Austin. All commercial rights reserved. Copyright 1997 Center for Research in Water Resources.