Supporting Information for

Spatiotemporal variations in the volume of closed lakes on the Tibetan Plateau and their climatic responses from 1976 to 2013

Ruimin Yang1,3*, Liping Zhu1,2*, Junbo Wang1,2, Jianting Ju1, Qingfeng Ma1, Falko Turner1,Yun Guo1,3

1 Key Laboratory of Tibetan Environment Changes and Land Surface Processes (TEL), Institute of Tibetan Plateau Research (ITP), Chinese Academy of Sciences, Beijing, China

2 CAS Center for Excellence in Tibetan Plateau Earth System, Beijing, China

3 University of Chinese Academy of Sciences, Beijing 100049, China

Corresponding authors:

Ruimin Yang, E-mail: ; Liping Zhu, E-mail:

Text S1 Data source and their description

SRTM DEM (Shuttle Radar Topography Mission, Digital Elevation Model) is a near-global (from 60°N to 56°S) topographic database generated based on satellite images in February 2000 (Van Zyl, 2001), at a resolution of 90 m (30 m for the United States) (Rabus et al., 2003) and quantified errors (Rodriguez et al., 2006).

LANDSAT images are relatively high-quality images with long and continuous records that are freely available at http://landsat.usgs.gov (Loveland and Dwyer, 2012). A total of 210 images were used in this study to extract the lake areas for 1976, 1990, 2000, 2005 and 2013. When the images acquired in those five years were unavailable, we used those acquired in the near years. The images obtained from October to December were selected to avoid the effects of seasonal changes in different years. The details of the images used in this study are listed in Table S1.

The mean annual precipitation (MAP) and mean annual temperature (MAT) in the lake catchments were extracted from the regional surface meteorological feature dataset for China (Yang et al., 2010; Chen et al., 2011). These data are available for a horizontal spatial resolution of 0.1° and from 1979 to 2012. In this study, MAP and MAT from 1979-1990 and 2005-2012 were used to analyze the changes in lake volume (LV) from 1976-1990 and 2005-2013, because climate data for 1976 to 1978 and 2013 were unavailable.

In this study, the areas of glaciers in the catchments of the studied lakes were obtained from the Second Glacier Inventory Dataset of China (Guo et al., 2014). Most of the studied lakes (83) receive glacier-melt input, whereas 31 lakes do not. The percentages of the glacier areas in the lake catchments are larger for lakes in the western TP and the southern outflow region of the TP than in the other area. The lakes without glaciers cover less than 200 km2 in area and are mainly distributed in the southeastern part of the inner TP.

All data used in this study are summarized in Table S2.

Table S1 List of LandSat images used in this study

No. / WRS-2
Path/Row / ~1976 / ~1990 / ~2000 / ~2005 / 2013
MSS / TM / ETM+ / TM/ETM+ / OLI
1 / 133034 / 1977-2-22 / 1992-11-16 / 2000-8-10 / 2005-11-04 / 2013-10-9
2 / 133035 / 1977-2-22 / 1992-11-16 / 2000-8-10 / 2005-11-04 / 2013-10-9
3 / 138035 / 1976-11-11 / 1991-11-1 / 2000-10-16 / 2006-10-25 / 2014-2-1
4 / 138036 / 1976-11-11 / 1990-11-14 / 2000-10-16 / 2006-11-10 / 2014-2-1
5 / 138037 / 1976-11-11 / 1989-11-11 / 2000-10-16 / 2006-10-25 / 2013-11-19
6 / 138038 / 1976-10-24 / 1990-11-14 / 2000-11-17 / 2005-11-7 / 2013-11-29
7 / 138039 / 1976-11-11 / 1989-11-11 / 2000-11-17 / 2005-11-7 / 2013-11-29
8 / 138040 / 1972-10-18 / 1989-11-11 / 2000-11-17 / 2005-11-7 / 2013-11-13
9 / 139035 / 1976-11-3 / 1989-11-02 / 2001-10-26 / 2005-10-29 / 2013-7-31
10 / 139036 / 1976-11-12 / 1989-10-01 / 2000-11-16 / 2007-10-03 / 2013-12-6
11 / 139037 / 1976-11-3 / 1990-11-05 / 2001-11-3 / 2005-11-14 / 2013-12-6
12 / 139038 / 1976-2-20 / 1990-11-18 / 2000-11-8 / 2005-11-14 / 2013-12-3
13 / 139040 / 1976-11-30 / 1989-11-14 / 2000-11-8 / 2005-11-14 / 2013-12-6
14 / 140034 / 1976-10-27 / 1992-11-17 / 2000-10-30 / 2005-11-5 / 2013-10-26
15 / 140035 / 1976-10-26 / 1992-11-17 / 2000-10-30 / 2005-11-21 / 2013-6-4
16 / 140036 / 1976-10-26 / 1992-11-17 / 2000-10-30 / 2005-11-1 / 2013-12-29
17 / 140037 / 1976-10-11 / 1991-11-15 / 2000-10-30 / 2007-9-24 / 2013-10-26
18 / 140038 / 1976-11-13 / 1991-11-15 / 2004-11-2 / 2009-10-30 / 2013-10-26
19 / 140039 / 1976-12-2 / 1991-11-15 / 2000-10-30 / 2004-11-2 / 2013-10-26
20 / 141034 / 1976-10-27 / 1988-10-12 / 2000-10-29 / 2005-11-12 / 2013-12-04
21 / 141035 / 1976-10-27 / 1992-11-08 / 2000-10-29 / 2005-11-12 / 2013-07-14
22 / 141036 / 1976-10-27 / 1989-01-24 / 2001-11-01 / 2006-10-14 / 2013-12-04
23 / 141037 / 1976-11-5 / 1991-01-07 / 2001-10-24 / 2006-10-30 / 2013-11-26
24 / 141038 / 1976-12-2 / 1989-11-5 / 2000-11-16 / 2005-11-12 / 2013-11-20
25 / 141039 / 1972-12-14 / 1989-1-24 / 2000-11-6 / 2005-11-12 / 2013-11-18
26 / 141040 / 1972-12-15 / 1991-11-30 / 2000-11-22 / 2005-11-12 / 2013-11-2
27 / 142035 / 1976-11-16 / 1991-10-28 / 2000-10-28 / 2006-10-05 / 2013-10-24
28 / 142036 / 1976-10-11 / 1989-3-4 / 2000-10-28 / 2006-10-5 / 2013-10-24
29 / 142037 / 1976-11-16 / 1991-10-28 / 2000-10-28 / 2008-11-11 / 2013-10-24
30 / 142038 / 1976-2-3 / 1991-4-23 / 2000-10-28 / 2009-11-30 / 2013-11-25
31 / 142039 / 1976-12-3 / 1991-11-13 / 2000-9-26 / 2006-11-22 / 2013-12-27
32 / 143035 / 1976-11-16 / 1991-12-6 / 2000-11-4 / 2006-10-12 / 2013-10-31
33 / 143036 / 1976-10-29 / 1991-12-6 / 2000-11-4 / 2006-10-12 / 2013-11-16
34 / 143037 / 1973-11-24 / 1991-12-6 / 2000-10-3 / 2008-11-02 / 2013-12-5
35 / 143038 / 1973-12-12 / 1991-12-6 / 2000-10-3 / 2008-11-02 / 2013-11-17
36 / 143039 / 1972-12-16 / 1991-12-6 / 2000-10-3 / 2005-10-25 / 2013-11-16
37 / 144035 / 1976-10-31 / 1991-12-13 / 1999-9-22 / 2006-10-19 / 2013-12-9
38 / 144036 / 1976-10-31 / 1991-12-13 / 2000-12-13 / 2007-09-20 / 2013-12-25
39 / 144037 / 1976-10-31 / 1991-12-13 / 2000-12-13 / 2007-09-20 / 2013-11-23
40 / 144039 / 1972-9-19 / 1990-10-23 / 2000-10-10 / 2010-12-1 / 2013-11-23
41 / 145036 / 1976-10-31 / 1993-08-19 / 2000-12-4 / 2009-10-7 / 2013-11-27
42 / 146036 / 1977-9-22 / 1991-3-30 / 2000-10-8 / 2010-11-13 / 2013-11-5

Table S2 Data used in this study

Data types / Spatial resolution / Acquisition period / Source
SRTM DEM / 90 m / 2000.2 / Data are freely available and were downloaded under the conditions of the USGS (United States Geological Survey) website: http://www.usgs.gov/
LANDSAT / MSS / 60 m / 1976
TM / 30 m / 1990
ETM+ / 30 m / 2000
OLI / 30 m / 2013
Meteorological data / 0.1° / 1979-2012 / Data are from the China regional surface meteorological feature dataset; they were authorized for use and downloaded from http://www.tpedata.ac.cn.
Glacier area in lake catchment / approximately 2006 / Second Glacier Inventory Dataset of China

Text S2 Building functions between lake volume changes and area

The ArcGIS 10.0 software package was used to build a function relating surface area to volume change using the lake basin topography provided by SRTM DEM. First, the lake area ASRTM extracted from SRTM and the surface area Ai (i=1, 2, 3 …) at an elevation of i m above ASRTM were calculated (Figure S1). Then, using equation S1 (below), the volume Vi' between ASRTM and Ai was computed. Using regression analysis, Vi' is estimated as a function of Ai (equation S2, below). Then, the volume change Vm' when the lake area expands from ASRTM up to Am, can be obtained. Assuming that the slope below ASRTM is similar to that above ASRTM within a certain range, the volume change Vn' when the lake shrinks from ASRTM to An can also be estimated using equation S2. Therefore, the LV change ∆Vm,n from time of tm to tn can be estimated by substituting lake areas Am, An into equation S3. Based on the above method, 114 functions were built for each lake.

Figure S1 Vertical section through a lake basin. The solid black line and dashed black line represent the topography above and beneath the elevation of the lake surface extracted from SRTM DEM (ASRTM), respectively.

Vi'=Vi-1'+(Ai+Ai-1+Ai*Ai-1)/3……………………… (S1)

Vi'=F(Ai)……………………………………………….……… (S2)

Then, ∆Vm,n=Vm'-Vn'=FAm-F(An)…………..…….………….(S3)

Where,

Aiis the area at the elevation of i m above ASRTM, i=1, 2, 3,⋯, A0=ASRTM

Vi'is the volume between ASRTM andAi, V0'=0

∆Vm,nis the LV change from time of tm to tn, Am and An are the lake areas at times tm and tn, respectively.

Text S3 Accuracy assessment

High-accuracy estimates of LV change can be obtained by constructing the underwater topography using bathymetry data (Wu et al., 2014; Zhang et al., 2011). Therefore, we used these more accurate LV change to assess the accuracy of the values computed using equation 3.

Five lakes (Nam Co, Tangra Yumco, Taro Co, Buro Co and Gyado Co) with measured bathymetry data for different regions on the TP (Figure 1) were selected. The volume changes calculated from bathymetry data are compared with the values estimated by the method described in Text S2. The average relative errors of the estimations based on the procedure described in Text S2 in the five lakes are listed in Table S3. The overall average relative error is 4.98%, which indicates the high accuracy of the estimated values.

A regression analysis between bathymetry-based values and estimation values according to Text S2 was performed (Figure S2). The coefficient of determination (R2) was very close to 1, thus indicating a high accuracy of the values together with a non-systematic, Gaussian-distributed error. The accuracy assessment demonstrated that the LV change estimated by SRTM DEM has an acceptable accuracy.

Table S3 The average relative errors of the values for lake volume changes estimated according to the methods described in Text S1 vs. bathymetry-based values.

Lake-name / Lake area (km2) / Periods / Values using SRTM DEM(km3) / Values based on bathymetry (km3) / Average relative error (%)
Name Co / 2020.9 / 1976-1989 / 2.82 / 2.74 / 2.17%
1989-2000 / 3.95 / 3.95
2000-2005 / 4.63 / 4.79
TangraYumco / 846.49 / 1972-1976 / 0.13 / 0.12 / 5.03%
1976-1989 / -1.07 / -1.04
1989-2000 / 0.32 / 0.32
2000-2002 / 0.47 / 0.45
2002-2005 / 0.44 / 0.41
2005-2009 / 0.47 / 0.43
Taro Co / 487.98 / 1972-1976 / -0.38 / -0.37 / 5.72%
1976-1989 / -0.09 / -0.09
1989-2000 / -0.15 / -0.16
2000-2002 / -0.01 / -0.01
2002-2009 / 0.31 / 0.33
2009-2011 / 0.06 / 0.06
Buro Co / 92.71 / 1976-1989 / -0.16 / -0.15 / 8.71%
1989-2000 / 0.14 / 0.13
2000-2006 / 0.29 / 0.27
2006-2013 / 0.16 / 0.15
Gyado Co / 42.72 / 1976-1989 / -0.03 / -0.03 / 3.29%
1989-2000 / -0.01 / -0.01
2000-2006 / 0.05 / 0.06
2006-2013 / 0.04 / 0.05
Overall average relative error / 4.98%

Figure S2 Results of regression analysis between bathymetry-based values for lake volume changes and the values estimated using the method described in Text S2.

References

Chen Y, Yang K, He J, Qin J, Shi J, Du J, He Q (2011) Improving land surface temperature modeling for dry land of China. Journal of Geophysical Research: Atmospheres (1984–2012) 116.

Guo Wanqin, Liu Shiyin, Yao Xiaojun, Xu Junli, SHANGGUAN Donghui, Wu Lizong, ZHAO Jingdong, LIU Qiao, Jiang Zongli, WEI Junfeng, BAO Weijia, YU Pengchun, DING Liangfu, LI Gang, LI Ping, GE Chunmei, WANG Yuan (2014) The Second Glacier Inventory Dataset of China (Version 1.0). Cold and Arid Regions Science Data Center at Lanzhou.

Loveland TR, Dwyer JL (2012) Landsat: Building a strong future. Remote Sensing of Environment 122:22-29.

Rabus B, Eineder M, Roth A, Bamler R (2003) The shuttle radar topography mission—a new class of digital elevation models acquired by spaceborne radar. ISPRS Journal of Photogrammetry and Remote Sensing 57:241-262.

Rodriguez E, Morris CS, Belz JE (2006) A global assessment of the SRTM performance. Photogrammetric engineering and remote sensing 72:249-260.

Van Zyl JJ (2001) The Shuttle Radar Topography Mission (SRTM): a breakthrough in remote sensing of topography. Acta Astronautica 48:559-565.

Wu Y, Zheng H, Zhang B, Chen D, Lei L (2014) Long-Term Changes of Lake Level and Water Budget in the Nam Co Lake Basin, Central Tibetan Plateau. Journal of Hydrometeorology 15:1312-1322.

Zhang B, Wu Y, Zhu L, Wang J, Li J, Chen D (2011) Estimation and trend detection of water storage at Nam Co Lake, central Tibetan Plateau. Journal of Hydrology 405:161-170.

4