The Tibetan Plateau (TP) in central Asia has an average elevation of more than 4000m and contains over 36,800 glaciers and 1500 lakes over an area of ∼2.5 million square kilometers. Since it is the largest ice mass on the planet aside from the poles, it is referred to as the ‘Third Pole’. As a whole, the TP has warmed over the past three decades, with its temperature increase of 0.3°C per decade being twice the global average. We hypothesized that this warming has accelerated glacier melting in the region. If so, this would contribute to widespread lake level increases, particularly in salty lakes in closed basins. However, because of the TP's remoteness, high altitude, thin atmosphere, and harsh weather conditions, quantitative glacier melting and lake water levels are difficult to monitor. Here, we use satellite altimetry data to quantify lake water levels as an indication of glacier melting.
We used the Ice, Cloud, and land Elevation Satellite (ICESat) altimetry data to quantify precise lake elevations and their changes in the TP over the period 2003–2009, the life span of the satellite. ICESat was a NASA-run laser altimeter that primarily mapped the polar ice caps. Use of satellite elevation data was necessary because many of the TP lakes have no recorded data (i.e., no physical gauges in place). We found that of the examined 74 largest lakes—56 of which are saltwater—84% show trends of increasing levels (see Figure 1).1 The proportion was higher (89%) when only considering the saltwater lakes. We calculated the mean lake level increase rate to be 0.27m/year for these. We found, for the first time, that Lake Cedo Caka had the greatest lake level increase (0.80m/year) of all the TP lakes examined.
Figure 1. Spatial distribution of the 74 lakes of the Tibetan Plateau in four subareas (I, II, III, and IV), with lake level rate of change (meters/year) for each lake indicated.
We verified the accuracy of the ICESat measurements by comparing our results to in situ lake level measurements made for Lake Qinghai.2 This lake has had continuous gauge measurements since 1959 and is the only lake in the TP with such records (see Figures 2 and 3). The increase rate of 0.11m/year from ICESat data matched well with the Station Xiashe gauge measurements (r2=0.90 and mean absolute error of 0.06m). This confirms that the absolute error of the ICESat elevation is less than 10cm. Thus, we can accurately use ICESat data to study lake level changes.
We next used the data to determine the long-term general trends for Lake Qinghai. The annual mean level of Lake Qinghai—based on the Station Xiashe gauge data over 50 years—showed a general decreasing trend (see Figure 3). However, there were three large interannual peaks in 1968, 1990, and 2009. For an enclosed lake, the major contributor to water loss is evaporation (E). In contrast, precipitation (P)—direct water entry by rainfall—and runoff from the basin (F), as well as runoff caused by snow/glacier melting (M), are the major sources of lake level increases. Thus, lake level (L) can be calculated by P+F+M−E. When L is positive, the lake levels rise. When it is negative, the reverse is true. However, we can approximate this by calculating P−E, which increases as lake levels rise and vice versa.
Figure 2.Lake level time series of ICESat and gauge from 2003 to 2009 for Lake Qinghai.
Figure 3. The annual mean lake level change from the Station Xiashe gauge at Lake Qinghai, 1959–2009.
The three peaks of increased lake levels for Lake Qinghai matched well with three positive peaks of P−E,2 as well as the three peaks (in 1967, 1989, and 2005) of increased runoff into the lake (see Figure 4). The most recent and longest interval of increased lake level was over 2004–2009, which was also observed from the ICESat altimetry data (see Figure 2). Interestingly, this interval was not accompanied by a large or long P−E peak. Instead, the continuous increase of the average runoff into the lake since 1996—see Figure 4—is consistent with the trend of increased temperature. This suggests that 1996, when the five-year running average temperature trend changed from negative to positive, marked the tipping point that could have accelerated the glacier/perennial snow cover melt in the region. It is expected that the current trend of lake level increase will continue for a certain period to another tipping point, when lake levels will begin to decrease again. This is likely to occur because of reduced glacier/perennial snow melting, which will be unable to offset the runoff loss into the lake because of other factors, such as precipitation decrease and/or evaporation increase.
Figure 4. The annual mean river runoffs at Stations Buha and Gangcha at Lake Qinghai, 1956–2009. The Buha River contributes almost half of the total runoff into Lake Qinghai.
Finally, we grouped the 74 lakes into four subareas based on their geographical location and trends in lake level change (see Figure 1). Three of the subareas show increased lake levels. The mean rate of change of lake level for subareas I, II, III, IV, and the entire TP were 0.12, 0.26, 0.19, −0.11, and 0.20m/year, respectively. Notably, subarea IV showed a decreasing water level, despite the overwhelming trend of increasing lake levels in the other subareas. Although the Himalayan glaciers have generally been melting rapidly, the small number of lakes sampled (total four lakes in subarea IV) does not provide sufficient data to draw a definitive conclusion about subarea IV.
In summary, we mapped increases in TP lake levels using ICESat data. The increases, particularly for a high percentage of salty lakes, supports the hypothesis that the TP glaciers are melting rapidly because of global warming. However, inaccessibility, complex terrain, and few meteorological stations currently make it difficult, if not impossible, to separate glacier/perennial snow cover melt water from precipitation runoff into a given lake. Ongoing measurements of precipitation, evaporation, and runoff into each lake, as well as high-resolution satellite images to quantify changes in glacier volume and lake area, will enable us to eventually evaluate the real contribution of glacier/perennial snow cover melt to lake level increase.
We acknowledge support from NASA (NNX08AQ87G) and the Chinese Scholarship Council for funding Guoqing Zhang's two years of study at the University of Texas at San Antonio.
Hongjie Xie, Guoqing Zhang
Laboratory for Remote Sensing and Geoinformatics
Department of Geological Sciences
University of Texas at San Antonio
San Antonio, TX
Hongjie Xie is an associate professor whose research is primarily focused on remote sensing of cryosphere, surface hydrology, and Martian surface processes.
1. G. Zhang, H. Xie, S. Kang, D. Yi, S. F. Ackley, Monitoring lake level changes on the Tibetan Plateau using ICESat altimetry data (2003–2009), Remote Sens. Environ
. 115, pp. 1733-1742, 2011. doi:10.1016/j.rse.2011.03.005
2. G. Zhang, H. Xie, S. Duan, M. Tian, D. Yi, Water level variation of Lake Qinghai from satellite and in situ measurements under climate change, J. Appl. Remote Sens.
5, pp. 053532, 2011. doi:10.1117/1.3601363