The Water Tower Function of the Tibetan Autonomous Region


Tibet is the water tower of Asia and rivers originating in Tibet flow into various regions in Asia. The Mekong, the Yellow river, the Yangthze, the Yarlung Tsampo (Brahmaputra), the Indus and the Karnali all originate on the Tibetan plateau and support hundreds of millions of people downstream.

Since Asia is monsoon dominated, with precipitation concentrated in just a few months, the perennial flow of the rivers largely relies on the constant flux of the glaciers in Tibet.

It is evident that conservation of Tibet’s water resources is of unprecedented importance and anthropogenic interventions should be carefully assessed on its effects on the water resources both in quantitative and qualitative terms.

However there are many anthropogenic factors already influencing this vulnerable system such as deforestation, hydropower development, irrigation intensification, and increased human pollution of Tibet’s waters. In combination with the retreat of glaciers and reduction of precipitation the downstream effects may become devastating.

Figure 1. DEM Tibet.

Project Objectives

Tibet is a data scarce and extremely inaccessible area and therefore Remote Sensing is an indispensable tool in these areas. Using Remote Sensing to gain insight in distributed vegetation patterns and the interaction between precipitation and vegetation is a first step in identifying areas where food production could be increased.

The objective of this part of the case study therefore is:

  • To understand precipitation patterns and land use interaction on the Tibetan plateau using time-series analysis of Remote Sensing data.

The second scale that has been addressed is the watershed (~100 km2). At this much smaller scale a state-of-the-art hydrological model will be build with the following objective:

  • To quantify the effects of land use change and agricultural intensification on the water resources.


For the first objective time series of SPOT-VEGETATION Normalized Difference Vegetation Index (NDVI) data were analyzed. The time series were processed using the Fast Fourier Transform and different time series derived characteristics were related to precipitation and land use.

For the second objective the public domain hydrological model SWAT was implemented for a small watershed near Tibet’s capital Lhasa. The model was set-up using a variety of local and public domain GIS datasets. In the absence of calibration data a sensitivity analysis was conducted and several land use change scenarios were evaluated.

The different GIS datasets which were used can be viewed using the FutureWater webmapping tool.

FutureWater webmapping tool

Results and Conclusions

An example of one of the Remote Sensing analysis results is shown in Figure 2. The hypothesis is that increases in NDVI are directly related to precipitation. For the Dangxing meteorological station the NDVI increments together with precipitation. The original NDVI increments exhibit a very irregular pattern, while the Fourier transformed NDVI increments show a much better relation with precipitation.

Figure 2. Relation delta NDVI and precipitation in Dangxiong.

Figure 3 shows a result of the scenario analysis with the SWAT model. All analysis are based on the water balance. For this specific scenario water deficits, defined as the difference between actual and potential evaporation, are shown assuming that all areas below 4500 meter in the watershed are cultivated with Barley. Water deficits are large throughout the growing season and improving irrigation systems have a large potential for increasing crop yields.

Figure 3. Water deficits barley below 4500 meter..

Please feel free to contact the Project Leader of this project for more information.