water

Article Morphometric Controls on Glacier Mass Balance of the Puruogangri Ice Field, Central Tibetan Plateau

Lin Liu 1,2, Liming Jiang 1,2,*, Yafei Sun 1,2, Hansheng Wang 1, Chaolu Yi 3,4 and Houtse Hsu 1

1 State Key Laboratory of Geodesy and Earth’s Dynamics, Institute of Geodesy and Geophysics, Chinese Academy of Sciences, 340 XuDong Rd, Wuhan 430077, China; [email protected] (L.L.); [email protected] (Y.S.); [email protected] (H.W.); [email protected] (H.H.) 2 University of Chinese Academy of Sciences, 19A Yuquan Rd, Beijing 100049, China 3 Institute of Tibetan Plateau Research, Chinese Academy of Sciences, 16 Lincui Rd, Beijing 100101, China; [email protected] 4 Center for Excellence in Tibetan Plateau Earth Sciences, Chinese Academy of Sciences, 16 Lincui Rd, Beijing 100101, China * Correspondence: [email protected]; Tel.: +86-27-8677-8612

Academic Editor: Daene C. McKinney Received: 30 July 2016; Accepted: 25 October 2016; Published: 1 November 2016

Abstract: Evaluating the impacts of climatic changes and morphometric features on glacier mass balance is crucial to providing insight into glacier changes and their effects on regional water resources and ecosystems. Here, we presented an evaluation of morphometric effects on the glacier mass balances of the Puruogangri ice field (PIF) on the Tibetan Plateau. A clear spatial variability of glacier mass balances, ranging from −0.035 to +0.019 m·w.e.·year−1, was estimated by comparing the TanDEM-X DEM (2012) with the SRTM-X DEM (2000). In general, the observed glacier mass changes were consistent with our fieldwork investigations. Furthermore, by applying the method of linear regression analysis, we found that the mass changes of individual glaciers on the PIF were mainly dominated by the mean altitude (R = 0.84, p < 0.001), however, they were statistically independent of glacier size, aspect, and surface velocity. At a local scale (grid size of 10 × 10 pixels), apart from the factor of altitude, surface velocity was correlated with glacier mass change.

Keywords: morphometric controls; glacier mass balance; Puruogangri ice field; Tibetan Plateau

1. Introduction Mountain glaciers on the Tibetan Plateau (TP) and its surroundings play a key role in the water and food security of people living in eastern and southern Asia because these glaciers are located at the headwaters of many prominent Asian rivers [1,2]. Therefore, it is necessary to identify the factors controlling glacier mass balance and provide insight into glacier changes and their effects on regional water resources and ecosystems. Basically, the mass balance of mountain glaciers is controlled by regional climate variability and local morphometric factors [3]. In recent decades, our knowledge of mountain glacier responses to climate changes has been substantially improved with advances in regional climate models (RCMs) and climatic observations [4–7]. Moreover, the spatially distributed glacier mass balances, which were mainly derived with geodetic method [8–13] and glacier surface energy and mass balance model [14–18], provide opportunities for investigating the effect of morphometric factors on glacier mass changes. In general, previous research suggested that there are several effective morphometric factors of glacier mass balances on the TP and its surroundings, such as topographic characteristics (i.e., glacier size, altitude, surface slope, curvature and aspect) [19–21], ice surface dynamics [22,23], and debris covers [24,25]. Racoviteanu et al. [21] reported that glacier changes in the Kanchenjunga-Sikkim area

Water 2016, 8, 496; doi:10.3390/w8110496 www.mdpi.com/journal/water Water 2016, 8, 496 2 of 18

(eastern Himalaya) were primarily controlled by glacier sizes and altitudes. Berthier et al. [19] found that glacier mass balance in the Himachal Pradesh (western Himalaya) was mainly affected by glacier altitudes and not correlated with size and aspect. In terms of glacier dynamics, Bolch et al. [22] and Gardelle et al. [23] agreed that low glacier surface velocity may result in a clear mass loss; however, a relatively stable front was observed over the stagnant tongue regions of glaciers covered with heavy surface debris in the central Himalayas [24]. Furthermore, Gardelle et al. [23] found variable effects of debris cover on glacier mass balance over nine study sites distributed in the Pamir, Karakoram, and Himalayas, which might be dependent on the fraction and thickness of debris, as well as the glacier velocity [24]. Overall, the local-scale behavior of glacier mass changes is influenced by multiple morphometric factors and is not fully understood on the Tibetan Plateau and its surroundings [2], particularly on the central TP, where few investigations have been conducted. In this study, we aim to evaluate the morphometric effects on glacier mass balances of the Puruogangri ice field (PIF) on the central TP, which is the largest modern ice field on the Tibetan Plateau. The glacier elevation changes and mass balances of the PIF between 2000 and 2012 were estimated by comparing the SRTM DEM (2000) with the TanDEM-X DEM (2012). Importantly, the observed glacier elevation changes were validated with our fieldwork investigations. We calculated glacier size from a well-delineated glacier outline derived from a Landsat TM image and extracted topographic terrain features (altitude, surface slope, curvature, and aspect) from the high-resolution TanDEM-X DEM. Ice flow velocity was derived according to the repeat-pass differential SAR interferometry method with a pair of ALOS PALSAR data. Furthermore, we employed linear regression analysis to evaluate the effects of topographic features and surface ice dynamics on mass balances of individual glaciers on the PIF. Additionally, the entire ice field was gridded with a size of 10 × 10 pixels to further estimate the morphometric effect on glacier mass changes at a more local scale.

2. Geographical Setting of the Study Area The PIF is located on the northern central Tibetan Plateau and the western portion of the Tanggula Mountains, with central coordinates of 33◦550 N and 89◦100 E[26] (Figure1). The glaciated area covers more than 420 km2 and a volume of approximately 52 km3 [27]. The equilibrium-line altitude (ELA) of the PIF lay at 5748 m·a.s.l. between 2001 and 2012 [28]. The glacier system is composed of approximately 50 glaciers with variable lengths radiating from the center of the ice field and extending halfway down-valley to the piedmont terminus. Glacier melt-water contributes to three endorheic lakes, named Linggo Co, Dogai Coring, and Chibuzhang Co [29]. According to the lake catchments, the ice field is divided into three parts—i.e., the northern, eastern and western slopes. The climate of the PIF is mainly affected by continental climatic processes, although it lies close to the boundary between the monsoon-dominated south and the westerly dominated north [30]. The records of the two ice cores drilled in 2000 (locations are shown in Figure1) indicated an abrupt onset of the recent warming in the early 20th century [30]. Following the Little Ice Age, the northern slope glaciers demonstrated an unambiguous and abrupt retreat, whereas the terminus changes of glaciers on the eastern and western slopes were relatively more stable [29]. The geodetic measurements of surface elevation changes on the PIF exhibited a mean value of approximately −0.238 m·year−1 from 1974 to 2000 [31] and −0.049 m·year−1 between 2000 and 2012 [12]. Huintjes et al. estimated glacier elevation change of −0.58 m between 2000 and 2011 by using the ‘COupled Snowpack and Ice surface energy and MAss balance model’ (COSIMA) [14]. The latest study revealed a much more rapid glacier thinning (−0.317 m·year−1) for the period 2012–2016 than in the periods 1974–2000 and 2000–2012 [13]. This finding indicates an accelerating trend of glacier surface thinning over the PIF since the beginning of the 2010s. Water 2016, 8, 496 3 of 18 Water 2016, 8, 496 3 of 18

Figure 1. Geographic map of the Puruogangri ice fieldfield (white) and its surroundings. The background isis thethe Landsat TM imageimage acquiredacquired onon 2424 SeptemberSeptember 2007.2007. ContoursContours (m(m·∙a.s.l.) are drawn at intervals of 250 m,m, andand the the thick thick line line (5750 (5750 m) m) is is close close to to the the ELA ELA for for the the period period 2001–2012. 2001–2012. The The locations locations of the of two the icetwo cores ice cores are indicated are indicated by black by triangles.black triangles. The coverage The coverage of two pairs of two of TSX/TDXpairs of TSX/TDX SAR data SAR is shown data asis greyshown and as white grey and lines. white The heightlines. The GCPs height of the GCPs 1:50,000 of the topographic 1:50,000 topographic map, ICES map, at footprints ICES at footprints and GPS pointsand GPS are points indicated are indicated as blue, brown as blue, and brown red dots, and respectively.red dots, respectively.

3. Data Data and and Methods Methods

3.1. Glacier Mass Balance Estimation

3.1.1. TanDEM-XTanDEM‐X DEMDEM andand GlacierGlacier MassMass BalanceBalance We employedemployed the the DEM DEM differencing differencing method method to calculate to calculate the elevation the elevation changes changes and mass and balances mass ofbalances all glaciers of all over glaciers the over PIF by the comparing PIF by comparing the TanDEM-X the TanDEM DEM‐X (2012) DEM and (2012) the and SRTM-X the SRTM DEM‐X (2002). DEM The(2002). TanDEM-X The TanDEM DEM‐ ofX theDEM entire of the ice fieldentire was ice derived field was from derived two sequential from two pairs sequential of TerraSAR-X/ pairs of TanDEM-XTerraSAR‐X/TanDEM SAR imagery‐X SAR (their imagery coverage (their is shown coverage in Figure is shown1), which in Figure were 1), acquired which onwere 26 acquired January 2012.on 26 The January TanDEM-X 2012. The bistatic TanDEM SAR data‐X bistatic were received SAR data as co-registered were received Single as co Look‐registered Complex Single (SLC) Look data inComplex the CoSSC (SLC) format, data with in the 3.3 CoSSC m and 2.5 format, m resolution with 3.3 pixel m and spacing 2.5 inm the resolution azimuth pixel and range spacing directions, in the respectively.azimuth and Theserange directions, original SLC respectively. data were These processed original according SLC data to thewere bistatic processed SAR according interferometric to the methodbistatic SAR [13] tointerferometric generate a glacier method surface [13] DEMto generate of the PIFa glacier with asurface spatial DEM resolution of the of PIF 10 with m. It a is spatial noted thatresolution absolute of and10 m. relative It is noted height that errors absolute in the and bistatic relative interferometric height errors processing in the bistatic were interferometric corrected with independentprocessing were height corrected reference datasetswith independent derived from height 1:50,000 reference topographic datasets maps, derived ICESat from laser altimeter1:50,000 data,topographic and GPS maps, measurements. ICESat laser The altimeter locations data, of these and height GPS referencemeasurements. data are The shown locations in Figure of 1.these heightThe reference SRTM-X data DEM are only shown covers in Figure a large 1. part of the ice field (~90%) because the X-band radar was operatedThe toSRTM derive‐X dataDEM along only discretecovers a 50 large km widepart of swaths the ice [32 field]. The (~90%) remaining because area the of X the‐band SRTM-X radar DEM was wasoperated filled to by derive the 1 arc-second data along SRTM-C discrete DEM 50 km (30 wide m spatial swaths resolution) [32]. The released remaining in January area of 2015, the SRTM and the‐X penetrationDEM was filled bias by of the the 1 C-band arc‐second and SRTM X-band‐C radarDEM (30 wave m spatial for ice resolution) and snow wasreleased corrected in January following 2015, and the penetration bias of the C‐band and X‐band radar wave for ice and snow was corrected following the method described by Gardelle et al. [23]. Furthermore, the 30 m resolution SRTM‐X

Water 2016, 8, 496 4 of 18 the method described by Gardelle et al. [23]. Furthermore, the 30 m resolution SRTM-X DEM was resampled using the bi-linear algorithm to 10 m resolution pixels to match with the TanDEM-X DEM. Before the process of DEM differencing, the co-registration method of horizontal shift, which was suggested by Berthier et al. [19], was used to eliminate the influence of geometric errors induced by the horizontal difference between the two DEMs used [33]. In this proposed method [19], the shift values between the TanDEM-X DEM and SRTM-X DEM were determined by the minimum in the standard deviation of the two DEM elevation differences of the off-glacier regions. Once accurately co-registered, the DEMs were differenced to map elevation changes for individual glaciers over the PIF, in which the outlines of individual glaciers were extracted from Landsat TM images, as described in Section 3.2. The resulting elevation changes were multiplied with the density of lost or gained glacial material between two periods of investigation to estimate the mass balances of all glaciers on the PIF. Because there are no measurements of the glacier ice density available in our study area, we employed an average glacial density of 850 kg·m−3. This value of glacial density recommended by Huss (2013) was estimated by applying an empirical firn-compaction model, and was proposed for geodetic observations with a time interval of more than five years [34].

3.1.2. Uncertainty Analysis In this study, we applied a common method for accuracy assessment of elevation changes obtained by the DEM differencing [35] due to the lack of in situ observations available. This method is generally used to calculate the statistical error (σ) of the elevation differences over off-glacier regions according to the following equation: s STDV 2 σ = m2 + ( √ ) , (1) N where m and STDV are the mean value and standard deviation of elevation differences over off-glacier areas, respectively. N is the number of independent measurements and can be calculated with the Equation (2): N × PS N = total , (2) 2D where Ntotal is the total number of measurements, PS is the pixel size of the DEMs used, and D is the distance of spatial autocorrelation. According to Bolch et al. [22], the decorrelation distance (D value) is generally equal to 20 pixels. Here, we utilized the 30 m resolution DEM and a decorrelation distance of 600 m. Furthermore, we added an uncertainty of 7% when the volume change was converted into mass balance due to the lack of in situ measurements of the glacial density [36]. Thus, an error of ±60 kg·m−3 was used as the density conversion factor.

3.2. Glacier Boundary Delimitation A well-delineated glacier outline is necessary to accurately estimate glacier mass balance and size. Glacier outlines can be downloaded from the Global Land Ice Measurements from Space (GLIMS) database of glacier inventory [37]. However, these boundaries do not always match our requirements of accuracy [23]. Satellite optical images—such as Landsat TM/ETM+, ASTER, and SPOT—have been demonstrated to be valuable data sources for delineating glacier outlines [23,24]. In this study, we applied the Normalized Difference Snow Index (NDSI), which has been used by Silverio and Jaquet [38], for a Landsat TM image acquired on 24 September 2007. Following the principle of the NDSI, glaciated areas show high values due to the snow and ice coverage, whereas areas of no snow coverage in the surrounding regions show low values. According to the histogram of NDSI images, a threshold of 0.55 was used to delineate the glacier outlines. Moreover, the glacier boundaries were manually corrected Water 2016, 8, 496 5 of 18 for several glaciers to pursue high accuracy and precision. Finally, the sizes of all glaciers on the PIF were calculated from the well-delineated glacier outlines. In order to validate the surging event of glacier 5Z213E0012, we detected changes in glacier terminus positions from 1976 to 2016, which were retrieved by one Landsat Multi-Spectral Scanner (MSS) image in 1976, two Landsat Thematic Mapper (TM) images in 1998 and 2010, two Landsat Enhanced Thematic Mapper Plus (ETM+) images in 2000 and 2002, and one Landsat Operational Land Imager (OLI) image in 2016.

3.3. Extraction of Topographic Parameters The topographic terrain features (altitude, surface slope, curvature, and aspect) were extracted from the resulting TanDEM-X DEM with 10-m resolution, as described in Section 3.1. With regard to geometry and geography, the altitude is defined as the height above sea level, the slope is the feature used to describe the steepness of a terrain surface, curvature is the indicator of a surface’s roughness, and aspect is the bearing or azimuth of the slope direction. The altitude is equal to the elevation of the glacier surface DEM, whereas the other three topographic parameters require retrieval with GIS (geographic information system) software. In this study, we employed the spatial analyst toolbox in the ArcGIS software program (version 10.0) (http://www.esri.com/software/arcgis) to estimate the parameters of slope, curvature, and aspect. Moreover, mean values of these topographic parameters for individual glaciers over the PIF were estimated with the glacier outlines described in Section 3.2.

3.4. Ice Flow Velocity Determination Speckle/feature tracking and repeat-pass differential SAR interferometry (D-InSAR) are two well-established methods for estimating glacier velocity [39,40]. Here, we applied the D-InSAR method to estimate the ice flow velocities of glaciers on the ice field using the DIFF module in the GAMMA software (http://www.gamma-rs.ch/). This method is prone to decorrelation due to ice melting and movement, and the L-band radar of the longer wavelength has a better capacity to preserve coherence over glacier surfaces than C-band and X-band radars [41]. Therefore, we used a pair of L-band ALOS PALSAR data over the PIF, acquired on 20 January 2009 and 7 March 2009, to minimize the decorrelation influence on the glacier velocity determination. Due to a lack of in situ observations of glacier surface velocity in this study area, accuracy assessment was based on the assumption that the ice-free areas are stable (no movement) [42,43]. The detailed calculations are similar with that of elevation changes, as described in Section 3.1.2. However, the m and STDV in the Equation (1) were replaced by the mean value and standard deviation of surface displacements over ice-free areas, respectively.

3.5. Statistical Analysis The linear regression analysis was conducted to evaluate the dependence of the glacier mass balance over the PIF on the selected topographic factors (size, altitude, surface slope, curvature, and aspect) and the ice flow velocity. Specifically, we employed simple linear regression (SLR) analysis to calculate the correlation coefficient (R), which describes the linear relationship between glacier mass balance and single independent morphometric factors. In addition, multiple linear regression (MLR) analysis, as an extension of the SLR, was employed to examine the relationship between the glacier mass balance and two or more morphometric factors by fitting a linear equation to observed data. In this study, we used the SLR and MLR tools in the SPSS (Statistical Product and Service Solutions) version 19.0 software package.

3.6. Field Works Since 2000, we carried out five field-working campaigns in order to study glacier dynamics and glacial moraines of the PIF, although it is very difficult to reach. During these in situ investigations, the surface morphologies in glacier termini regions were monitored with photographic records. Water 2016, 8, 496 6 of 18

Water 2016, 8, 496 6 of 18 Furthermore, we performed GPS surveying on the front of glacier 5Z611A0006 with RTK technique in in October 2012 to provide a high‐accuracy elevation reference for bistatic SAR interferometric October 2012 to provide a high-accuracy elevation reference for bistatic SAR interferometric processing. processing. The reference station of RTK was installed on a rock outcrop and surveyed with the The reference station of RTK was installed on a rock outcrop and surveyed with the Unistrong E650 Unistrong E650 GPS receiver for more than five hours. Moreover, the records of base GPS receiver GPS receiver for more than five hours. Moreover, the records of base GPS receiver were post-processed were post‐processed using differential GPS (DGPS) technique with the Lhasa station of CMONOC using differential GPS (DGPS) technique with the Lhasa station of CMONOC (Crustal Movement (Crustal Movement Observation Network of China) to obtain a result of height accuracy up to Observation Network of China) to obtain a result of height accuracy up to ±0.1 m. Another GPS 0.1 m. Another GPS receiver of the same type with base receiver was simultaneously employed for receiver of the same type with base receiver was simultaneously employed for roving at 45 points that roving at 45 points that are apart from the reference station less than 3 km. In particular, the error of are apart from the reference station less than 3 km. In particular, the error of relative range between the relative range between the floating and reference station was required to be better than 1 cm, in floating and reference station was required to be better than 1 cm, in order to keep the high elevation order to keep the high elevation accuracy of these GPS points. The locations of GPS measurements accuracy of these GPS points. The locations of GPS measurements are shown in Figure1. are shown in Figure 1. 4. Results 4. Results 4.1. Glacier Elevation Changes 4.1. Glacier Elevation Changes Figure2 illustrates detailed elevation changes over the entire PIF generated by comparing the SRTM-X/CFigure 2 illustrates DEM indetailed February elevation 2000 changes with the over resulting the entire TanDEM-X PIF generated DEM by in comparing January 2012. the TheSRTM glacier-wide‐X/C DEM in meanFebruary elevation 2000 with difference the resulting between TanDEM 2000‐X DEM and in 2012 January is − 2012.0.27 The± glacier0.18 m‐wide (or −mean0.023 elevation± 0.015 mdifference·year−1), between which compares 2000 and to2012−0.51 is −±0.273.71 ± 0.18 m (DEM m (or differencing −0.023 ± 0.015 method) m∙year observed−1), which bycompares Neckel to et −0.51 al. [ 12± 3.71] in m the (DEM same differencing period. This method) discrepancy observed is by presumably Neckel et al. due [12] to in difference the same period. in the coverageThis discrepancy of used is TanDEM-X presumably bistatic due to SARdifference data andin the the coverage vertical of accuracy used TanDEM of generated‐X bistatic TanDEM-X SAR data DEMs.and the Invertical this study, accuracy the of entire generated PIF was TanDEM covered‐X byDEMs. two In pairs this of study, TanDEM-X the entire images, PIF was whereas covered only by onetwo pairpairs ofof TanDEM-XTanDEM‐X images, data (covers whereas about only 90% one ofpair the of ice TanDEM field)‐ wasX data applied (covers by about Neckel 90% et of al. the [12 ice]. Moreover,field) was applied our DEM by result Neckel has et aal. better [12]. Moreover, vertical accuracy our DEM of 1.91result± has0.76 a mbetter (see vertical Liu et al. accuracy 2016) than of 1.91 that ± of0.76 3.24 m ±(see1.11 Liu m et reported al. 2016) by than Neckel that etof al.3.24 [12 ±], 1.11 because m reported we have by corrected Neckel et the al. [12], absolute because and we relative have heightcorrected errors the inabsolute the bistatic and relative interferometric height errors processing in the bybistatic using interferometric the well-distributed processing height by referenceusing the datasetswell‐distributed (their locations height reference are indicated datasets in Figure(their locations1). are indicated in Figure 1).

Figure 2.2. ElevationElevation changeschanges (unit:(unit: m) overover thethe PuruogangriPuruogangri iceice fieldfield betweenbetween 20002000 andand 2012.2012. GlacierGlacier outlines werewere delineated delineated from from a Landsata Landsat TM TM image. image. The The last fivelast codesfive codes of the of WGMS the WGMS IDs of the IDs glaciers of the withglaciers a size with exceeding a size exceeding 10 km2 are 10 presented. km2 are presented. Glacier E0012 Glacier (WGMS E0012 ID: (WGMS 5Z213E0012) ID: 5Z213E0012) is a surge-type is a glacier.surge‐type The glacier. glacier centralThe glacier flow central lines are flow indicated lines are in indicated dashed black. in dashed black.

Additionally, Huintjes et al. estimated glacier elevation change of −0.58 m between 2000 and 2011 with the ‘COupled Snowpack and Ice surface energy and MAss balance model’ (COSIMA) by

Water 2016, 8, 496 7 of 18

Water Additionally,2016, 8, 496 Huintjes et al. estimated glacier elevation change of −0.58 m between 20007 andof 18 2011 with the ‘COupled Snowpack and Ice surface energy and MAss balance model’ (COSIMA) by using the meteorological dataset from the High Asia RefinedRefined analysis (HAR) [[14].14]. This difference of −0.310.31 m m between between our our observed observed results results and and modeled modeled glacier elevation changes is possibly attributed to two reasons.reasons. Firstly,Firstly, there there are are different different time time periods periods of of investigation investigation (2000–2012 (2000–2012 and and 2000–2011) 2000–2011) in inthe the two two studies. studies. Secondly, Secondly, precipitation precipitation scaling scaling factor factor determined determined at at Zhadang Zhadang glacier glacier (30 (30°28.57◦28.570′ N,N, 90°38.7190◦38.71′ 0 E)E) was used for the PIF (33(33°55◦55′ 0 N, 89°10 89◦10′ 0E)E) by by Huintjes Huintjes et et al. [14]. [14]. However, according to Yao etet al.al. [[2],2], precipitationprecipitation variations at Zhadang glacier and the PIF are not similar due to different latitudes. Importantly, Importantly, the the modeled glacier elevation changes are highly sensitive to precipitation scaling factor, as exhibited inin TableTable 33 ofof HuintjesHuintjes etet al.al. [[14].14]. As shown in Figure2 2,, clear clear surface surface thinning thinning was was observed observed in in most most glacier glacier tongue tongue regions, regions, with with a maximum elevation decrease of upup toto approximatelyapproximately −−4040 m, whereas a slight thickening was detected in in the the broad broad and and flat flat interior interior regions regions (mostly (mostly in in accumulation accumulation areas) areas) above above 5800 5800 m∙a.s.l. m·a.s.l. In general,In general, this this spatial spatial pattern pattern of of glacier glacier elevation elevation changes changes over over the the PIF PIF was was similar similar with with that that revealed revealed by Neckel et al.al. [[12].12]. Particularly, it is worth noting that the surface elevation changes varied for individual glaciers.glaciers. Most Most glaciers glaciers experienced experienced predominant predominant surface surface lowering lowering of approximately of approximately 20–40 20– m 40at glacierm at glacier terminal terminal regions, regions, e.g., glacier e.g., glacier No. 5Z611A0008 No. 5Z611A0008 and No. and 5Z513B0018; No. 5Z513B0018; however, however, several glaciersseveral (e.g.,glaciers No. (e.g., 5Z611A0006) No. 5Z611A0006) exhibited exhibited no obvious no obvious changes changes in elevation, in elevation, except forexcept one for thickening one thickening glacier (No.glacier 5Z213E0012) (No. 5Z213E0012) in the eastern in the PIF. eastern Probably PIF. dueProbably to surging due to activities, surging this activities, anomalous this glacieranomalous was glacierthickened was by thickened up to approximately by up to approximately 80 m at the terminus, 80 m at the whereas terminus, negative whereas elevation negative changes elevation were changesobserved were in the observed upper ablation in the areas.upper Figureablation3 highlights areas. Figure the detailed3 highlights features the ofdetailed surface features elevation of surfacechanges elevation and altitudes changes along and the altitudes central flowalong line the of central these fourflow selectedline of these glaciers. four selected glaciers.

Figure 3. Surface elevation changes changes vs. altitude altitude along along the the central central flow flow lines lines for for glacier 5Z611A0006 5Z611A0006 ( (aa);); 5Z611A0008 (b); 5Z513B0018 (c); and 5Z213E0012 ((d).). The profileprofile locationslocations are shownshown inin FigureFigure2 2..

The DEM‐derived elevation changes at glacier tongues were generally consistent with our fieldwork investigations. For example, the contradictory behavior at the tongue regions of two typical glaciers (No. 5Z611A0006 and No. 5Z611A0008) could be interpreted by their photographic records from 2000 to 2012, as shown in Figures 4 and 5. The relatively stable glacier 5Z611A0006, as evidenced by elevation change results (see Figures 2 and 3a), experienced almost no

Water 2016, 8, 496 8 of 18

The DEM-derived elevation changes at glacier tongues were generally consistent with our fieldwork investigations. For example, the contradictory behavior at the tongue regions of two

Watertypical 2016 glaciers, 8, 496 (No. 5Z611A0006 and No. 5Z611A0008) could be interpreted by their photographic8 of 18 Waterrecords 2016 from, 8, 496 2000 to 2012, as shown in Figures4 and5. The relatively stable glacier 5Z611A0006,8 of as 18 geomorphologicalevidenced by elevation changes change (Figure results 4). (seeHowever, Figures for2 and glacier3a), experienced 5Z611A0008, almost intensive no geomorphological surface melting at geomorphological changes (Figure 4). However, for glacier 5Z611A0008, intensive surface melting at thechanges terminal (Figure areas4). However,was found for in glacier our field 5Z611A0008, campaign intensive in 2011 surface(Figure melting 5). The at melt the‐ terminalwater and areas ice the terminal areas was found in our field campaign in 2011 (Figure 5). The melt‐water and ice wascracking found occurring in our field in the campaign snout indicated in 2011 (Figure that glacier5). The 5Z611A0008 melt-water exhibited and ice cracking significant occurring mass loss in the or cracking occurring in the snout indicated that glacier 5Z611A0008 exhibited significant mass loss or instability.snout indicated that glacier 5Z611A0008 exhibited significant mass loss or instability. instability.

Figure 4. Comparative positions of the glacier front (No. 5Z611A0006) over the past decade: (a) 2000; Figure 4. Comparative positions of the glacier front (No. 5Z611A0006) over the past decade: ( (aa)) 2000; 2000; (b) 2004; (c) 2006; and (d) 2012. (b) 2004; (c) 2006; and (d) 2012.

Figure 5. Melt‐water and ice cracking at the glacier snout (No. 5Z611A0008) recorded in 2011. (a) Figure 5.5. Melt‐water and ice cracking at the glacier snout (No. 5Z611A0008) recorded in 2011.a (a) View fromMelt-water the eastern and side; ice ( crackingb) view from at the the glacier western snout side. (No. 5Z611A0008) recorded in 2011. ( ) View Viewfrom thefrom eastern the eastern side; ( side;b) view (b) fromview thefrom western the western side. side. 4.2. Glacier Surface Velocity 4.2. Glacier Surface Velocity Figure 6 shows the InSAR‐derived ice flow velocities of the Puruogangri ice field in 2009. The Figure 6 shows the InSAR‐derived ice flow velocities of the Puruogangri ice field in 2009. The glacierFigure‐wide6 meanshows surface the InSAR-derived velocity is 2.25 ice ± 0.61 flow m velocities∙year−1, and of the the maximum Puruogangri velocity ice fieldis up into 2009.about glacier‐wide mean surface velocity is 2.25 ± 0.61 m∙year−1, and− 1the maximum velocity is up to about The17 m glacier-wide∙year−1. In general, mean surfaceour results velocity are similar is 2.25 ± with0.61 the m· yearobserved, and glacier the maximum surface velocities velocity isin up1996 to 17 m∙year−1. In general,−1 our results are similar with the observed glacier surface velocities in 1996 (Figureabout 17 1 m in·year Huintjes. In et general, al. [14]). our For results example, are similar obvious with surface the observed movements glacier were surface detected velocities by two in (Figure 1 in Huintjes et al. [14]). For example, obvious surface movements were detected by two 1996studies (Figure in the 1 glacier in Huintjes tongue et al.areas [14 ]).of Forglacier example, 5Z611A0003 obvious and surface 5Z611A0006, movements and were in the detected upper regions by two studies in the glacier tongue areas of glacier 5Z611A0003 and 5Z611A0006, and in the upper regions ofstudies glacier in the5Z611A0008, glacier tongue 5Z611A0009, areas of glacier5Z513B0007, 5Z611A0003 and 5Z513B0014. and 5Z611A0006, However, and inthe the rates upper of regionssurface of glacier 5Z611A0008, 5Z611A0009, 5Z513B0007, and 5Z513B0014. However, the rates of surface velocitiesof glacier were 5Z611A0008, significantly 5Z611A0009, changed 5Z513B0007,over glacier 5Z213E0010 and 5Z513B0014. and 5Z513B0010 However, the between rates of1996 surface and velocities were significantly changed over glacier 5Z213E0010 and 5Z513B0010 between 1996 and velocities2009. In particular, were significantly for glacier changed 5Z213E0012, over glacier we 5Z213E0010found that andice flow 5Z513B0010 velocities between over upper 1996 and regions 2009. 2009. In particular, for glacier 5Z213E0012, we found that ice flow velocities over upper regions wereIn particular, much higher for glacier than 5Z213E0012, that of glacier we tongue found that areas ice in flow 2009, velocities whereas over relatively upper regionsfaster movements were much were much higher than that of glacier tongue areas in 2009, whereas relatively faster movements werehigher observed than that over of glacier glacier tongue tongue areas areas in in 2009, 1996 whereas[14]. relatively faster movements were observed wereover glacierobserved tongue over areasglacier in tongue 1996 [14 areas]. in 1996 [14].

Water 2016, 8, 496 9 of 18 Water 2016, 8, 496 9 of 18 Water 2016, 8, 496 9 of 18

FigureFigure 6. 6.Glacier Glacier surface surface velocity velocity in in line-of-sight line‐of‐sight direction direction of of the the Puruogangri Puruogangri ice ice field field in 2009.in 2009. The The letter andletterFigure numbers and 6. Glaciernumbers are the surface are last the five velocity last codes five in ofcodes line the ‐of glacier ‐thesight glacier WGMS direction WGMS IDs. of the TheIDs. Puruogangri glacierThe glacier outlines outlines ice field are are shown in shown2009. in The solidin black,solidletter and black, and the numbers and dashed the dashedare black the line lastblack five is line the codes centralis the of central the flow glacier lineflow ofWGMS line glacier of glacierIDs. 5Z213E0012. The 5Z213E0012. glacier outlines are shown in solid black, and the dashed black line is the central flow line of glacier 5Z213E0012. 4.3.4.3. Spatial Spatial Variability Variability of of Glacier Glacier Mass Mass Balances Balances 4.3. Spatial Variability of Glacier Mass Balances A slightly negative glacier mass balance of −0.019 ± 0.014 m∙w.e.∙year−1 was measured over the A slightly negative glacier mass balance of −0.019 ± 0.014 m·w.e.·year−1 was measured over the entireA ice slightly field negativefrom 2000 glacier to 2012. mass We balance observed of − 0.019clear ±spatial 0.014 mvariability∙w.e.∙year of−1 wasglacier measured mass balances over the entire ice field from 2000 to 2012. We observed clear spatial variability of glacier mass balances ranging rangingentire ice from field −0.035 from to 2000 +0.019 to m2012.∙w.e. We∙year observed−1 in the studyclear area.spatial Table variability 1 lists the of massglacier balances, mass balances glacier from −0.035 to +0.019 m·w.e.·year−1 in the study area. Table1 lists the mass balances, glacier sizes, sizes,ranging and from mean −0.035 values to of+0.019 other m morphometric∙w.e.∙year−1 in factorsthe study for area. 30 glaciers Table 1on lists the the PIF mass with balances,a size of greater glacier and mean values of other morphometric factors for 30 glaciers on the PIF with a size of greater than thansizes, 1 andkm2 mean (except values for the of othersurge ‐morphometrictype glacier). A factors total offor 18 30 glaciers glaciers experienced on the PIF with mass a loss,size ofbut greater mass 2 1 kmgainthan(except occurred1 km2 (except for in the 12 for surge-type glaciers. the surge Furthermore, glacier).‐type glacier). A total a Aheterogeneous oftotal 18 of glaciers 18 glaciers pattern experienced experienced of glacier mass massmass loss, loss, balance but but mass masswas gain occurredobservedgain occurred in within 12 glaciers. in three 12 glaciers. sub Furthermore,‐regions Furthermore, corresponding a heterogeneous a heterogeneous to three pattern topographic pattern of glacier unitsof massglacier of the balance massnorthern, balance was western observed was withinandobserved threeeastern within sub-regions parts three in the sub corresponding PIF‐regions (see Figurecorresponding to three 7). On topographic the to threewestern topographic units slope, of thethere units northern, were of the three westernnorthern, glaciers and western easternwith partsnegativeand in eastern the mass PIF parts balance (see in Figure the and PIF7 seven). On(see glaciers the Figure western with 7). On positive slope, the western there mass werebalance. slope, three Similarly,there glaciers were 10 withthree mass negative‐glaciersloss glaciers with mass balanceandnegative 5 andmass mass seven‐gain balance glaciersglaciers and withwere seven positive observed glaciers mass on with the balance. positive northern Similarly, mass slope. balance. 10However, mass-loss Similarly, on glaciers the 10 easternmass and‐loss slope, 5 mass-gainglaciers the glaciersbehaviorand 5 were mass of observed‐allgain five glaciers glaciers on thewere was northern observedrelatively slope. onconsistent, the However, northern demonstrating on slope. the eastern However, distinct slope, on mass the loss.eastern behavior slope, of all the five glaciersbehavior was of relatively all five glaciers consistent, was relatively demonstrating consistent, distinct demonstrating mass loss. distinct mass loss.

Figure 7. Scatterplot of mass changes of glaciers in the western, eastern, and northern parts of the PIF. FigureFigure 7. Scatterplot 7. Scatterplot of of mass mass changes changes of of glaciers glaciers inin the western, eastern, eastern, and and northern northern parts parts of the of the PIF. PIF.

Water 2016, 8, 496 10 of 18

Table 1. Mass balances of glaciers with a size of greater than 1 km2 on the PIF (except for the surge-type glacier) and glacier size and mean values of other morphometric factors.

WGMS Mass Balance Altitude Slope Aspect Velocity Size (km2) Curvature ID (m·w.e.·Year−1) (m) (Degree) (Degree) (m·Year−1) 5Z611A0002 0.017 3.51 5849.5 17.7 0.004 210.0 1.36 5Z611A0003 0.000 6.60 5750.3 10.8 0.008 234.9 3.40 5Z611A0005 0.009 5.13 5878.3 16.2 0.034 217.5 2.06 5Z611A0006 0.004 19.86 5833.4 8.7 0.010 216.9 3.83 5Z611A0008 −0.021 46.28 5758.6 5.5 0.000 214.2 2.72 5Z611A0009 −0.006 11.14 5810.4 9.8 0.005 237.6 2.90 5Z611A0010 0.003 50.18 5770.5 8.5 0.001 207.0 2.40 5Z611A0013 0.005 3.99 5824.2 17.0 −0.040 225.0 1.00 5Z611A0014 0.018 1.34 5990.6 13.7 0.026 167.8 1.21 5Z611A0015 0.006 14.85 5831.0 9.1 0.008 189.6 2.55 5Z213E0001 −0.018 6.08 5668.3 12.3 −0.011 152.7 0.99 5Z213E0004 −0.019 1.86 5622.3 10.7 −0.008 199.3 2.03 5Z213E0008 −0.017 1.79 5645.0 14.1 −0.038 104.1 0.98 5Z213E0009 −0.035 7.86 5627.2 10.7 −0.023 130.5 1.78 5Z213E0010 −0.008 32.27 5787.3 12.4 −0.004 140.3 2.23 5Z513B0005 0.002 8.33 5841.6 15.5 −0.032 134.1 1.42 5Z513B0007 0.019 14.34 5917.1 19.9 0.016 119.7 1.85 5Z513B0009 0.013 4.38 5830.6 21.0 0.056 80.7 0.86 5Z513B0010 0.008 13.31 5773.8 12.2 0.003 126.2 4.50 5Z513B0011 −0.025 1.19 5552.6 15.1 −0.029 266.7 0.74 5Z513B0013 −0.012 2.34 5685.5 17.1 −0.051 231.1 0.70 5Z513B0014 −0.004 28.85 5791.0 15.7 −0.009 173.2 1.90 5Z513B0015 −0.006 1.62 5780.1 25.0 0.069 87.5 0.93 5Z513B0016 −0.012 4.31 5720.5 14.0 −0.006 109.7 1.45 5Z513B0022 −0.013 6.17 5665.9 13.5 −0.019 210.0 0.76 5Z513B0020 −0.002 5.23 5767.3 18.5 −0.001 98.3 1.17 5Z513B0018 −0.007 53.58 5775.6 9.4 −0.007 145.9 1.68 5Z513C0001 −0.011 10.42 5657.3 10.6 −0.008 194.4 1.37 5Z513C0002 0.004 3.45 5701.3 17.4 0.034 266.8 0.52 5Z513B0024 −0.003 32.34 5689.2 16.1 −0.011 165.8 1.42

4.4. The Effect of Morphometric Factors on Glacier Mass Changes The results of both SLR and MLR analysis revealed variable morphometric effects on the mass balance of individual glaciers on the PIF (see Figure8 and Table2). In general, the relationships between glacier mass balance and the morphometric controls modeled by SLR analysis are consistent with those revealed by MLR analysis. The results showed a strong positive correlation (R = 0.84 > 0.75 in Figure8f) between glacier mass balance and the mean altitude at the 99% confidence interval (p < 0.001 in Table2). Glacier surface slope might be slightly correlated with glacier mass change (R = 0.35 > 0.25) in Figure8d at the 95% confidence interval ( p = 0.025 < 0.05). Furthermore, the SLR analysis suggested that glacier size, aspect, and ice flow velocity were not statistically significant controls on glacier mass balance (see Figure8a–c). These findings were generally consistent with the p-values of the above three factors (p > 0.05, 95% confidence interval) in Table2. However, the possibly positive correlation between curvature and glacier mass change (R = 0.49 > 0.25 in Figure8e) was not consistent with the results estimated by MLR analysis (p = 0.775 > 0.05, 95% confidence interval).

Table 2. Summary of the MLR analysis in the SPSS software package (T-value (t) and p-value are two parameters obtained from the t-test).

Variable Parameter Estimate Standard Error t p-Value Size 0.002 0.010 0.174 0.863 Altitude 0.010 0.002 6.538 <0.001 Slope 0.118 0.049 2.393 0.025 Curvature 1.661 5.743 0.289 0.775 Aspect 0.005 0.003 1.822 0.082 Velocity 0.215 0.181 1.191 0.246 Water 2016, 8, 496 11 of 18

Table 2. Summary of the MLR analysis in the SPSS software package (T‐value (t) and p‐value are two parameters obtained from the t‐test).

Variable Parameter Estimate Standard Error t p‐Value Size 0.002 0.010 0.174 0.863 Altitude 0.010 0.002 6.538 <0.001 Slope 0.118 0.049 2.393 0.025 Curvature 1.661 5.743 0.289 0.775

Water 2016, 8, 496 Aspect 0.005 0.003 1.822 0.082 11 of 18 Velocity 0.215 0.181 1.191 0.246

Figure 8.8. Correlations between glacier mass balance and individualindividual morphometric factors from the SLR analysis: (a) size; (b) aspect; (c) ice flowflow velocity; ((d)) slope;slope; ((ee)) curvature;curvature; andand ((ff)) altitude.altitude.

In order to further estimate the effect of morphometric factors on glacier mass changes at a In order to further estimate the effect of morphometric factors on glacier mass changes at a more more local scale, the whole ice field was divided into the grids with a size of 10 × 10 pixels. Note local scale, the whole ice field was divided into the grids with a size of 10 × 10 pixels. Note that the that the factor of glacier area was not included in the analysis due to the same size of all meshes. factor of glacier area was not included in the analysis due to the same size of all meshes. Figure9 Figure 9 exhibits the correlation between glacier elevation changes and five morphometric factors exhibits the correlation between glacier elevation changes and five morphometric factors over entire over entire meshes of the PIF. Obviously, the factor of aspect was not a statistically significant meshes of the PIF. Obviously, the factor of aspect was not a statistically significant control on glacier control on glacier elevation changes. As shown in Figure 9a, there was not any apparent difference elevation changes. As shown in Figure9a, there was not any apparent difference of glacier elevation of glacier elevation changes in different aspects. The glacier mass changes might be slightly changes in different aspects. The glacier mass changes might be slightly correlated with surface slope correlated with surface slope and curvature. Figure 9c,d illustrates that the grid with smaller slope and curvature. Figure9c,d illustrates that the grid with smaller slope and curvature might be more and curvature might be more possible to experience a larger value of glacier thinning or thickening. possible to experience a larger value of glacier thinning or thickening. Figure9e shows that the factor Figure 9e shows that the factor of altitude should be an effective control of glacier elevation changes. of altitude should be an effective control of glacier elevation changes. In particular, the relationships In particular, the relationships between altitude and glacier elevation changes are different between between altitude and glacier elevation changes are different between the glacier tongue and upper the glacier tongue and upper regions. Specifically, when altitude increases, surface thinning over regions. Specifically, when altitude increases, surface thinning over glacier tongue areas (altitudes glacier tongue areas (altitudes of 5400–5700 m) decreases quickly, but surface thickening increases of 5400–5700 m) decreases quickly, but surface thickening increases slowly around the ELA and its slowly around the ELA and its upper areas (altitudes of 5700–6200 m). It is noteworthy that glacier upper areas (altitudes of 5700–6200 m). It is noteworthy that glacier grids with higher ice flow rates grids with higher ice flow rates possibly experienced a relatively stable glacier surface elevation, possibly experienced a relatively stable glacier surface elevation, whereas obvious glacier thinning whereas obvious glacier thinning and thickening were usually found over quasi‐stagnant grids and thickening were usually found over quasi-stagnant grids (lower ice flow rate), as demonstrated in (lower ice flow rate), as demonstrated in Figure 9b. It infers that surface velocities might affect Figure9b. It infers that surface velocities might affect glacier elevation changes at local scales. glacier elevation changes at local scales. Generally, the relationships between glacier elevation changes and morphometric factors over individual glaciers of the PIF were consistent with that at more local scale. However, obvious correlation between ice flow velocity and glacier mass change which was masked by averaging over individual glaciers was found for entire meshes on the ice field. Overall, according to the results of statistical analysis at different scales, we found that the observed spatially variable glacier mass changes might be related with the morphometric factors of altitude and velocity. Interestingly, the altitude-dependent glacier mass changes were basically supported by the significant correlations between altitude and main climatic parameters such as air temperature, relative humidity, air pressure, and precipitation (see Table 1 in Huintjes et al. [14]). Water 2016, 8, 496 12 of 18 Water 2016, 8, 496 12 of 18

Figure 9.9. Correlations between glacier elevation changes and morphometric factors at local scalescale (grid(grid sizesize ofof 1010 ×× 1010 pixels). pixels). ( (aa)) aspect; aspect; ( (bb)) ice ice flow flow velocity; velocity; ( (cc)) slope; slope; ( (dd)) curvature; curvature; and ( (e)) altitude.

5. DiscussionGenerally, the relationships between glacier elevation changes and morphometric factors over individual glaciers of the PIF were consistent with that at more local scale. However, obvious 5.1.correlation Morphometric between Effects ice onflow Glacier velocity Mass and Balances glacier in the mass TP andchange Its Surroundings which was masked by averaging over individual glaciers was found for entire meshes on the ice field. Overall, according to the In this study, our results revealed that the glacier mass balance of the Puruogangri ice field results of statistical analysis at different scales, we found that the observed spatially variable glacier is primarily affected by mean altitude and ice flow velocity, and may be slightly affected by slope. mass changes might be related with the morphometric factors of altitude and velocity. Interestingly, Note that the debris cover is not included in our analysis because it does not cover glaciers of the the altitude‐dependent glacier mass changes were basically supported by the significant PIF [29]. Generally, the altitude dependency of glacier mass balance over the PIF was supported correlations between altitude and main climatic parameters such as air temperature, relative by the stronger surface melting at lower elevations, which was found for several glaciers in the humidity, air pressure, and precipitation (see Table 1 in Huintjes et al. [14]). Himalaya and Karakoram [1,2]. However, Berthier et al. [19] reported that glaciers in the Spiti/Lahaul (western5. Discussion Himalaya) experienced thickening or limited thinning below 4400 m. Moreover, the velocity-dependent glacier elevation changes confirm the effect of velocity on glacier mass changes in Bhutan,5.1. Morphometric Himalaya Effects [23]. Additionally, on Glacier Mass although Balances surface in the slopeTP and was Its slightlySurroundings correlated with glacier mass change on the PIF, it was a major topographic control of glacier mass balance in eastern Himalaya [21]. GlacierIn this study, mass balancesour results of therevealed PIF were that not the relatedglacier withmass the balance factors of ofthe size Puruogangri and aspect, ice which field isis generallyprimarily in affected agreement by mean with the altitude results and reported ice flow by Berthier velocity, et and al. [19 may] and be Scherler slightly et affected al. [24]. However,by slope. theNote glacier that the mass debris balances cover is of not some included other glaciersin our analysis in the adjacentbecause it regions does not were cover dependent glaciers of on the these PIF factors.[29]. Generally, Specifically, the altitude the mass dependency balances of of glaciers glacier inmass eastern balance Himalaya over the were PIF generallywas supported determined by the bystronger the factor surface of size melting [21]. at An lower aspect-dependent elevations, which glacier was mass found change for several was alsoglaciers observed in the inHimalaya eastern Himalayaand Karakoram [23] and [1,2]. West However, Kunlun Berthier [20]. Consequently, et al. [19] reported for other that glaciers glaciers in thein the TP Spiti/Lahaul and its surroundings, (western theHimalaya) effects of morphometricexperienced thickening factors on glacier or limited mass balances thinning are notbelow always 4400 consistent m. withMoreover, our results. the Furthermore,velocity‐dependent morphometric glacier elevation effects on changes glacier massconfirm balance the effect are far of from velocity homogeneous on glacier overmass different changes glaciatedin Bhutan, regions. Himalaya Additionally, [23]. Additionally, in one glaciated although region, surface glacier slope mass was changes slightly arecorrelated commonly with affected glacier bymass several change morphometric on the PIF, factors.it was a For major example, topographic both aspect control and of velocity glacier aremass important balance controlsin eastern of glacierHimalaya mass [21]. balances in the Bhutan, Himalaya [23]. Overall,Glacier mass glaciers balances in the of vicinity the PIF dowere not not always related experience with the factors the same of size mass and change aspect, behaviors which is duegenerally to these in agreement local morphometric with the effectsresults evenreported in the by same Berthier climatic et al. setting. [19] and Consequently, Scherler et al. a high[24]. uncertaintyHowever, the might glacier exist mass in balances region-wide of some glacier other mass glaciers change in the results adjacent by using regions the were area-weighted dependent extrapolationon these factors. method Specifically, from limited the observationsmass balances over of several glaciers study in eastern sites in theHimalaya Tibetan were Plateau generally and its surroundingsdetermined by [4 the,23 ,factor44]. of size [21]. An aspect‐dependent glacier mass change was also observed in eastern Himalaya [23] and West Kunlun [20]. Consequently, for other glaciers in the TP and its surroundings, the effects of morphometric factors on glacier mass balances are not always consistent

Water 2016, 8, 496 13 of 18

with our results. Furthermore, morphometric effects on glacier mass balance are far from homogeneous over different glaciated regions. Additionally, in one glaciated region, glacier mass changes are commonly affected by several morphometric factors. For example, both aspect and velocity are important controls of glacier mass balances in the Bhutan, Himalaya [23]. Overall, glaciers in the vicinity do not always experience the same mass change behaviors due to these local morphometric effects even in the same climatic setting. Consequently, a high uncertainty might exist in region‐wide glacier mass change results by using the area‐weighted extrapolation method from limited observations over several study sites in the Tibetan Plateau and its surroundings [4,23,44].

5.2. Climatic Influences on Glacier Mass Balance of the PIF As shown in Figure 2, there was a slight mass gain in the interior areas, whereas obvious mass loss occurred in most glacier termini regions. To evaluate the effects of climate changes on the observed patterns of glacier mass balance over the PIF, we analyzed the precipitation and air temperature records between 1980 and 2012 from the Tuotuohe (34°13′ N, 92°26′ E), Bange (31°22′ N, Water90°012016′ ,E),8, 496 and Anduo (32°21′ N, 91°06′ E) meteorological stations, located approximately 300 13km of 18 from the study area. These meteorological data were downloaded from the China Meteorological Data Sharing Service System (http://data.cma.cn/). 5.2. Climatic Influences on Glacier Mass Balance of the PIF Changes in glacier mass balance should be mainly related to spatial and temporal variations in summerAs shown temperature in Figure (June,2, there July, was and a slight August) mass and gain annual in the precipitation interior areas, in whereas the Tibetan obvious Plateau mass and loss occurredits surroundings, in most glacier because termini glacier regions. melting To usually evaluate dominates the effects on of the climate summer changes ablation on season the observed and patternssnowfall of occurs glacier during mass all balance seasons over of the the year PIF, [2,45]. we analyzed Figure 10a the exhibits precipitation that summer and temperature air temperature at recordsTuotuohe between station 1980 is andlower 2012 than from two the other Tuotuohe stations (34 due◦13 0toN, higher 92◦26 0latitude.E), Bange Moreover, (31◦220 N, for 90 ◦temporal010 E), and Anduochanges (32 ◦in21 0summerN, 91◦06 temperature,0 E) meteorological a clearly stations, positive located trend approximately was found over 300 kmthese from three the studystations area. Thesebetween meteorological 1980 and 2012, data which were was downloaded generally consistent from the Chinawith continuous Meteorological warming Data trend Sharing since early Service System20th century (http://data.cma.cn/ [30]. Therefore,). significant mass losses at the glacier terminus regions of the PIF were probablyChanges due in to glacier continuous mass increase balance in should summer be mainlytemperature. related to spatial and temporal variations in summerAnnual temperature precipitation (June, over July, the and PIF August) (central and TP) annual was much precipitation more stable in the than Tibetan that in Plateau the Pamir, and its Karakoram, and Himalaya, because the climate of this study area is dominated by continental climatic surroundings, because glacier melting usually dominates on the summer ablation season and snowfall conditions [2]. Furthermore, the ice field lies close to the boundary between the monsoon‐dominated occurs during all seasons of the year [2,45]. Figure 10a exhibits that summer temperature at Tuotuohe south and the westerly dominated north. Thus the decrease in precipitation caused by weakened station is lower than two other stations due to higher latitude. Moreover, for temporal changes in Indian monsoon might be compensated by the strengthened westerlies. Figure 10b shows that the summer temperature, a clearly positive trend was found over these three stations between 1980 and annual precipitations of these three stations were relatively stable during the period of 1980–2012. 2012,This which result wasis basically generally agreement consistent with withthe precipitation continuous trend warming from 1979 trend to since 2010 early(Figure 20th 4a in century Yao et al. [30 ]. Therefore,[2]). Particularly, significant it is massnoted lossesthat there at the was glacier a slight terminus increase of regions annual of precipitation the PIF were since probably 2000 (Figure due to continuous10b), which increase might contribute in summer to temperature. the glacier mass gain over the interior areas of the ice field.

Water 2016, 8, 496 14 of 18

Figure 10. The records of mean summer temperature (a) and annual precipitation (b) from 1980 to Figure 10. The records of mean summer temperature (a) and annual precipitation (b) from 1980 to 2012, 2012, which were estimated from climatic observations of the Tuotuohe, Bange, and Anduo which were estimated from climatic observations of the Tuotuohe, Bange, and Anduo meteorological meteorological stations. The five‐year moving averages of mean summer temperature and annual stations. The five-year moving averages of mean summer temperature and annual precipitation are precipitation are shown in red lines. The changes of linear fitting in temperature and precipitation shown in red lines. The changes of linear fitting in temperature and precipitation are indicated as black are indicated as black lines. Two vertical dashed lines indicate the year of 2000 and 2012, which are lines. Two vertical dashed lines indicate the year of 2000 and 2012, which are the start and end time of the start and end time of the period of investigation in this study. the period of investigation in this study. 5.3. Surge‐Type Glacier Surge‐type glacier is characterized by the life cycle of an active phase and a quiescent phase [46]. During the active phase, the glacier experiences a rapid transfer of mass from the upper “reservoir” area to the lower “receiving” area, and high ice flow rates that increase by a factor of 10 to 100, whereas ice flow rates decline substantially and the reservoir area builds up again during the quiescent phase [47]. According to the changes in glacier terminus location and surface velocity, some surge‐type glaciers have been identified in the Karakoram [48,49], Pamir plateau [50], and Kunlun Shan [40,51]. Here, elevation changes probably induced by a surging event were clearly detected over the glacier 5Z213E0012 of the PIF, in the central Tibetan Plateau. Over the 12‐year period (2000–2012), this glacier was thickened by up to 80 m at the terminus while thinned by approximately 20 m along its central flow line, as shown in Figure 3d. Moreover, Figure 11 demonstrates that glacier terminus advanced in the almost same time period (2000–2011) as that of the elevation change estimation, and continued in the following period of 2011–2016. These changes in glacier surface elevation and terminus locations imply that a surging event might be fully developed between 2000 and 2016. In addition, an obvious glacier retreat was observed over this glacier from 1976 to 2000 (Figure 11), which is a typical feature of the quiescent phase. Consequently, the surging event of glacier 5Z213E0012 might have started in 2000 and lasted for about 16 years.

Figure 11. Outlines of glacier 5Z213E0012 in 1976, 1998, 2000, 2002, 2011, and 2016. The background is the Landsat‐8 OLI image acquired in 2016.

Water 2016, 8, 496 14 of 18

Water 2016, 8, 496 14 of 18

Annual precipitation over the PIF (central TP) was much more stable than that in the Pamir, Karakoram, and Himalaya, because the climate of this study area is dominated by continental climatic conditions [2]. Furthermore, the ice field lies close to the boundary between the monsoon-dominated southFigure and the 10. westerly The records dominated of mean north. summer Thus temperature the decrease (a) and in precipitation annual precipitation caused by(b) weakenedfrom 1980 to Indian monsoon2012, mightwhich bewere compensated estimated from by the climatic strengthened observations westerlies. of the FigureTuotuohe, 10b Bange, shows and that Anduo the annual precipitationsmeteorological of these stations. three The stations five‐year were moving relatively averages stable of during mean summer the period temperature of 1980–2012. and annual This result is basicallyprecipitation agreement are shown with in the red precipitationlines. The changes trend of fromlinear 1979fitting to in 2010 temperature (Figure and 4a inprecipitation Yao et al. [2]). Particularly,are indicated it is noted as black that lines. there Two was vertical a slight dashed increase lines ofindicate annual the precipitation year of 2000 and since 2012, 2000 which (Figure are 10b), whichthe might start and contribute end time to of the the glacier period of mass investigation gain over in the this interior study. areas of the ice field.

5.3. Surge Surge-Type‐Type GlacierGlacier Surge-typeSurge‐type glacier is characterizedcharacterized by the life cycle of an active phase and a quiescent phase phase [46]. [46]. During the active phase, the glacier experiences a a rapid transfer of of mass from the upper “reservoir” area to to the the lower lower “receiving” “receiving” area, area, and and high high ice flow ice flow rates ratesthat increase that increase by a factor by a of factor 10 to of100, 10 whereas to 100, whereasice flow rates ice flow decline rates substantially decline substantially and the andreservoir the reservoir area builds area up builds again up during again during the quiescent the quiescent phase phase[47]. According [47]. According to the to changes the changes in glacier in glacier terminus terminus location location and and surface surface velocity, velocity, some some surge-typesurge‐type glaciers have been identifiedidentified in the Karakoram [48,49], [48,49], Pamir plateau [50], [50], and Kunlun Shan [40,51]. [40,51]. Here, elevation changes probably induced by a surging event were clearlyclearly detecteddetected over the glacierglacier 5Z213E0012 of the PIF, inin thethe centralcentral TibetanTibetan Plateau. Plateau. OverOver thethe 12-year12‐year periodperiod (2000–2012),(2000–2012), this glacier was thickened by up to 80 mm atat thethe terminusterminus whilewhile thinnedthinned by approximately 20 m along its central flowflow line,line, asas shownshown inin Figure Figure3 d.3d. Moreover, Moreover, Figure Figure 11 11 demonstrates demonstrates that that glacier glacier terminus terminus advanced advanced in thein the almost almost same same time periodtime period (2000–2011) (2000–2011) as that ofas the that elevation of the changeelevation estimation, change andestimation, continued and in thecontinued following in periodthe following of 2011–2016. period These of 2011–2016. changes in These glacier changes surface in elevation glacier andsurface terminus elevation locations and implyterminus that locations a surging imply event that might a surging be fully event developed might between be fully 2000 developed and 2016. between In addition, 2000 and an obvious2016. In glacieraddition, retreat an obvious was observed glacier over retreat this was glacier observed from 1976 over to this 2000 glacier (Figure from 11), 1976 which to is 2000 a typical (Figure feature 11), ofwhich the quiescent is a typical phase. feature Consequently, of the quiescent the surging phase. event Consequently, of glacier 5Z213E0012 the surging might event have of started glacier in 20005Z213E0012 and lasted might for have about started 16 years. in 2000 and lasted for about 16 years.

Figure 11. Outlines ofof glacierglacier 5Z213E00125Z213E0012 in in 1976, 1976, 1998, 1998, 2000, 2000, 2002, 2002, 2011, 2011, and and 2016. 2016. The The background background is theis the Landsat-8 Landsat‐ OLI8 OLI image image acquired acquired in in 2016. 2016.

Figure 12 illustrates the surface velocities in 2009 along the central flow line of glacier 5Z213E0012. In general, the magnitudes (0.001–0.045 m·day−1) of our observed glacier velocities are similar with that (0.005–0.052 m·day−1) in 1996 [14]. This finding is not consistent with that of most identified surge-type glaciers, which the velocity over active phase was 10–100 times more than that during quiescent phase [47], such as Karayaylak glacier [50] and central Yulinchuan glacier [51]. However, it is noteworthy that a significant change in ice flow pattern was found on glacier 5Z213E0012 between 1996 and 2009. As shown in Figure 12, a relatively higher ice flow rate was occurred on glacier upper Water 2016, 8, 496 15 of 18

Figure 12 illustrates the surface velocities in 2009 along the central flow line of glacier 5Z213E0012. In general, the magnitudes (0.001–0.045 m∙day−1) of our observed glacier velocities are similar with that (0.005–0.052 m∙day−1) in 1996 [14]. This finding is not consistent with that of most identified surge‐type glaciers, which the velocity over active phase was 10–100 times more than that during quiescent phase [47], such as Karayaylak glacier [50] and central Yulinchuan glacier [51]. However, it is noteworthy that a significant change in ice flow pattern was found on glacier Water 2016, 8, 496 15 of 18 5Z213E0012 between 1996 and 2009. As shown in Figure 12, a relatively higher ice flow rate was occurred on glacier upper portion (3000–6000 m from terminus) in 2009, whereas glacier tongue areasportion (0–3000 (3000–6000 m from m from terminus) terminus) experienced in 2009, whereas a more glacier rapid tongue movement areas in (0–3000 1996 m(see from Figure terminus) 3 in Huintjesexperienced et al. a more [14]). rapid This movementchange in inglacier 1996 (seesurface Figure velocity 3 in Huintjes might be et al.caused [14]). by This the change surging in glacierevent; however,surface velocity further might work be is caused required by theto understand surging event; it, however,such as more further detailed work isinvestigations required to understand of glacier dynamicit, such as mechanisms more detailed and investigations bedrock characteristics. of glacier dynamic mechanisms and bedrock characteristics.

Figure 12. IceIce flow flow velocities (in 2009) vs. altitude along the central flow flow line line of of glacier glacier 5Z213E0012. 5Z213E0012. The profile profile location is is shown shown in in Figure Figure 66.. In order to comparing with the resultsresults inin 1996,1996, thethe unitunit ofof glacier surface velocity was converted from m·∙year−1 1toto m m∙day·day−1−. 1.

6. Conclusions Conclusions and and Future Future Work Work Clear spatialspatial variability variability of of mass mass balances balances for individualfor individual glaciers, glaciers, ranging ranging from − from0.0351 − to0.0351 +0.0194 to −1 +0.0194m·w.e.· yearm∙w.e.−1,∙year was observed, was observed on the on Puruogangri the Puruogangri ice field ice on field the on central the central TP from TP 2000 from to 2000 2012 to based 2012 basedon a comparison on a comparison of the of TanDEM-X the TanDEM DEM‐X (2012)DEM (2012) with the with SRTM-X the SRTM DEM‐X (2000).DEM (2000). The method The method of linear of linearregression regression analysis analysis was employed was employed to investigate to investigate the impacts the impacts of local of morphometric local morphometric factors factors on glacier on glaciermass balances. mass balances. The results The results revealed revealed a significant a significant correlation correlation (R = 0.84, (R p= <0.84, 0.001) p < between 0.001) between the glacier the glaciermass balance mass balance and mean and altitude, mean altitude, whereas glacierwhereas size, glacier aspect, size, and aspect, ice flow and velocity ice flow were velocity not statistically were not statisticallysignificant controls. significant Furthermore, controls. inFurthermore, order to further in estimateorder to the further effect of morphometricestimate the factorseffect onof morphometricglacier mass changes factors at on a moreglacier local mass scale, changes the whole at a more ice field local was scale, divided the whole into the ice grids field withwas adivided size of 10into× the10 pixels.grids with Importantly, a size of apart 10 × from 10 pixels. the factor Importantly, of altitude, apart surface from velocity the factor is correlated of altitude, with surface glacier velocityelevation is change correlated at a with local glacier scale. elevation change at a local scale. In addition, for the entire ice field, field, an evident pattern of mass loss was observed in most glacier tongue regions, whereas a slight mass gain was detected in thethe interiorinterior areas.areas. This overall spatial behavior of glacier mass changes might be attributed to the regional climatic variations associated with the continuous increases in in summer summer temperature temperature between between 1980 1980 and and 2012 2012 and slight increase of annual precipitation since since 2000. Future Future work work will be conducted for comprehensively evaluating the impacts of climatic changes and morphometric factors factors on on glacier glacier mass mass balance, balance, in in order to provide greater insight into glacier changes and their effects on regional waterwater resourcesresources andand ecosystems.ecosystems.

Acknowledgments: The workwork in in this this study study was was supported supported by theby Nationalthe National Natural Natural Science Science Foundation Foundation of China of China(grant numbers(grant numbers 41590854, 41590854, 41230523, 41230523, 41431070, 41431070, 41274024, 41274024, and 41321063), and 41321063), the Key the Research Key Research Program Program of Frontier of FrontierSciences, Sciences, CAS (grant CAS number (grant number QYZDB-SSWDQC027), QYZDB‐SSWDQC027), and the and National the National Key Basic Key Research Basic Research Program Program of China of (grant number 2012CB957702). The authors acknowledge the German Aerospace Center (DLR) for providing the TSX/TDX bistatic SAR datasets and the SRTM-X DEM and the National Aeronautics and Space Administration (NASA) for providing the SRTM-C DEM and the Landsat TM images. Author Contributions: Liming Jiang and Houtse Hsu conceived and designed the experiments; Lin Liu and Yafei Sun performed the experiments; Chaolu Yi and Lin Liu conducted all the field works; Lin Liu carried out the data analysis with the significant contributions of Liming Jiang and Hansheng Wang. Lin Liu and Liming Jiang wrote the paper. Water 2016, 8, 496 16 of 18

Conflicts of Interest: The authors declare no conflict of interest.

References

1. Bolch, T.; Kulkarni, A.; Kääb, A.; Huggel, C.; Paul, F.; Cogley, J.; Frey, H.; Kargel, J.; Fujita, K.; Scheel, M. The state and fate of Himalayan glaciers. Science 2012, 336, 310–314. [CrossRef][PubMed] 2. Yao, T.; Thompson, L.; Yang, W.; Yu, W.; Gao, Y.; Guo, X.; Yang, X.; Duan, K.; Zhao, H.; Xu, B. Different glacier status with atmospheric circulations in Tibetan Plateau and surroundings. Nat. Clim. Chang. 2012, 2, 663–667. [CrossRef] 3. Oerlemans, J. Extracting a climate signal from 169 glacier records. Science 2005, 308, 675–677. [CrossRef] [PubMed] 4. Gardelle, J.; Berthier, E.; Arnaud, Y. Slight mass gain of Karakoram glaciers in the early twenty-first century. Nat. Geosci. 2012, 5, 322–325. [CrossRef] 5. Kapnick, S.; Delworth, T.; Ashfaq, M.; Malyshev, S.; Milly, P. Snowfall less sensitive to warming in Karakoram than in Himalayas due to a unique seasonal cycle. Nat. Geosci. 2014, 7, 834–840. [CrossRef] 6. Maussion, F.; Scherer, D.; Mölg, T.; Collier, E.; Curio, J.; Finkelnburg, R. Precipitation seasonality and variability over the Tibetan Plateau as resolved by the High Asia Reanalysis. J. Clim. 2014, 27, 1910–1927. [CrossRef] 7. Curio, J.; Maussion, F.; Scherer, D. A 12-year high-resolution climatology of atmospheric water transport over the Tibetan Plateau. Earth Syst. Dyn. 2015, 6, 109–124. [CrossRef] 8. Bolch, T.; Buchroithner, M.; Pieczonka, T.; Kunert, A. Planimetric and volumetric glacier changes in the Khumbu Himal, Nepal, since 1962 using Corona, Landsat TM and ASTER data. J. Glaciol. 2008, 54, 592–600. [CrossRef] 9. Nuimura, T.; Fujita, K.; Yamaguchi, S.; Sharma, R. Elevation changes of glaciers revealed by multitemporal digital elevation models calibrated by GPS survey in the Khumbu region, Nepal Himalaya, 1992–2008. J. Glaciol. 2012, 58, 648–656. [CrossRef] 10. Hagg, W.; Braun, L.; Uvarov, V.; Makarevich, K. A comparison of three methods of mass-balance determination in the Tuyuksu glacier region, Tien Shan, Central Asia. J. Glaciol. 2004, 50, 505–510. [CrossRef] 11. Farinotti, D.; Longuevergne, L.; Moholdt, G.; Duethmann, D.; Mölg, T.; Bolch, T.; Vorogushyn, S.; Güntner, A. Substantial glacier mass loss in the Tien Shan over the past 50 years. Nat. Geosci. 2015, 8, 716–722. [CrossRef] 12. Neckel, N.; Braun, A.; Kropáˇcek,J.; Hochschild, V. Recent mass balance of the Purogangri Ice Cap, central Tibetan Plateau, by means of differential X-band SAR interferometry. Cryosphere 2013, 7, 1623–1633. [CrossRef] 13. Liu, L.; Jiang, L.; Sun, Y.; Yi, C.; Wang, H.; Hsu, H. Glacier elevation changes (2012–2016) of the Puruogangri ice field on the Tibetan Plateau derived from bi-temporal TanDEM-X InSAR data. Int. J. Remote Sens. 2016, 37, 5687–5707. [CrossRef] 14. Huintjes, E.; Neckel, N.; Hochschild, V.; Schneider, C. Surface energy and mass balance at the Purogangri Ice Cap, central Tibetan Plateau, 2001–2011. J. Glaciol. 2015, 61, 1048–1060. [CrossRef] 15. Li, B.; Yu, Z.; Liang, Z.; Acharya, K. Hydrologic response of a high altitude glacierized basin in the central Tibetan Plateau. Glob. Planet. Chang. 2014, 118, 69–84. [CrossRef] 16. Fujita, K.; Sakai, A. Modelling runoff from a Himalayan debris-covered glacier. Hydrol. Earth Syst. Sci. 2014, 11, 2441–2482. [CrossRef] 17. Gao, H.; He, X.; Ye, B.; Pu, J. Modeling the runoff and glacier mass balance in a small watershed on the central Tibetan Plateau, China, from 1955 to 2008. Hydrol. Process. 2012, 26, 1593–1603. [CrossRef] 18. Zhao, L.; Tian, L.; Zwinger, T.; Ding, R.; Zong, J.; Ye, Q.; Moore, J.C. Numerical simulations of Gurenhekou Glacier on the Tibetan Plateau. J. Glaciol. 2014, 60, 71–82. [CrossRef] 19. Berthier, E.; Arnaud, Y.; Kumar, R.; Ahmad, S.; Wagnon, P.; Chevallier, P. Remote sensing estimates of glacier mass balances in the Himachal Pradesh (Western Himalaya, India). Remote Sens. Environ. 2007, 108, 327–338. [CrossRef] 20. Phan, V.; Lindenbergh, R.; Menenti, M. Orientation dependent glacial changes at the Tibetan Plateau derived from 2003–2009 ICESat laser altimetry. Cryosphere 2014, 8, 2425–2463. [CrossRef] Water 2016, 8, 496 17 of 18

21. Racoviteanu, A.; Arnaud, Y.; Williams, M.; Manley, W. Spatial patterns in glacier characteristics and area changes from 1962 to 2006 in the Kanchenjunga-Sikkim area, eastern Himalaya. Cryosphere 2014, 9, 505–523. [CrossRef] 22. Bolch, T.; Pieczonka, T.; Benn, D. Multi-decadal mass loss of glaciers in the Everest area (Nepal Himalaya) derived from stereo imagery. Cryosphere 2011, 5, 349–358. [CrossRef] 23. Gardelle, J.; Berthier, E.; Arnaud, Y.; Kääb, A. Region-wide glacier mass balances over the Pamir-Karakoram- Himalaya during 1999–2011. Cryosphere 2013, 7, 1263–1286. [CrossRef] 24. Scherler, D.; Bookhagen, B.; Strecker, M. Spatially variable response of Himalayan glaciers to climate change affected by debris cover. Nat. Geosci. 2011, 4, 156–159. [CrossRef] 25. Lejeune, Y.; Bertrand, J.; Wagnon, P.; Morin, S. A physically based model of the year-round surface energy and mass balance of debris-covered glaciers. J. Glaciol. 2013, 59, 327–344. [CrossRef] 26. Li, X.; Yi, C.; Chen, F.; Yao, T.; Li, X. Formation of proglacial dunes in front of the Puruogangri Icefield in the central Qinghai-Tibet Plateau: Implications for reconstructing paleoenvironmental changes since the Lateglacial. Quat. Int. 2006, 154, 122–127. [CrossRef] 27. Yi, C.; Li, X.; Qu, J. Quaternary glaciation of Puruogangri—The largest modern ice field in Tibet. Quat. Int. 2002, 97–98, 111–121. [CrossRef] 28. Spiess, M.; Maussion, F.; Möller, M.; Scherer, D.; Schneider, C. MODIS Derived Equilibrium Line Altitude Estimates for Purogangri Ice Cap, Tibetan Plateau, and their Relation to Climatic Predictors (2001–2012). Geogr. Ann. Ser. A Phys. Geogr. 2015, 97, 599–614. [CrossRef] 29. Pu, J.; Yao, T.; Wang, N.; Ding, L.; Zhang, Q. Puruogangri ice field and its variations since the Little Ice Age of the Northern Tibetan Plateau. J. Glaciol. Geocryol. 2002, 24, 87–92. (In Chinese) 30. Thompson, L.; Yao, T.; Davis, M.; Mosley-Thompson, E.; Mashiotta, T.; Lin, P.; Mikhalenko, V.; Zagorodnov, V. Holocene climate variability archived in the Puruogangri ice cap on the central Tibetan Plateau. Ann. Glaciol. 2006, 43, 61–69. [CrossRef] 31. Lei, Y.; Yao, T.; Yi, C.; Wang, W.; Sheng, Y.; Li, J.; Joswiak, D. Glacier mass loss induced the rapid growth of Linggo Co on the central Tibetan Plateau. J. Glaciol. 2012, 58, 177–184. [CrossRef] 32. Rabus, B.; Eineder, M.; Roth, A.; Bamler, R. The shuttle radar topography mission-a new class of digital elevation models acquired by spaceborne radar. ISPRS J. Photogramm. Remote Sens. 2003, 57, 241–262. [CrossRef] 33. Paul, F.; Bolch, T.; Kääb, A.; Nagler, T.; Nuth, C.; Scharrer, K.; Shepherd, A.; Strozzi, T.; Ticconi, F.; Bhambri, R. The glaciers climate change initiative: Methods for creating glacier area, elevation change and velocity products. Remote Sens. Environ. 2013, 162, 408–426. [CrossRef] 34. Huss, M. Density assumptions for converting geodetic glacier volume change to mass change. Cryosphere 2013, 7, 877–887. [CrossRef] 35. Koblet, T.; Gärtner-Roer, I.; Zemp, M.; Jansson, P.; Thee, P.; Haeberli, W.; Holmlund, P. Reanalysis of multi-temporal aerial images of Storglaciären, Sweden (1959–99)-Part 1: Determination of length, area, and volume changes. Cryosphere 2010, 4, 333–343. [CrossRef] 36. Zemp, M.; Jansson, P.; Holmlund, P.; Gärtner-Roer, I.; Koblet, T.; Thee, P.; Haeberli, W. Reanalysis of multi-temporal aerial images of Storglaciären, Sweden (1959–99)-Part 2: Comparison of glaciological and volumetric mass balances. Cryosphere 2010, 4, 345–357. [CrossRef] 37. Raup, B.; Kääb, A.; Kargel, J.; Bishop, M.; Hamilton, G.; Lee, E.; Paul, F.; Rau, F.; Soltesz, D.; Khalsa, S. Remote sensing and GIS technology in the Global Land Ice Measurements from Space (GLIMS) project. Comput. Geosci. 2007, 33, 104–125. [CrossRef] 38. Silverio, W.; Jaquet, J. Glacial cover mapping (1987–1996) of the Cordillera Blanca (Peru) using satellite imagery. Remote Sens. Environ. 2005, 95, 342–350. [CrossRef] 39. Liu, L.; Jiang, L.; Wang, H. Extraction of glacier surface elevation and velocity in high Asia with ERS-1/2 Tandem SAR data: Application to Puruogangri ice field, Tibetan Plateau. In Proceedings of the IEEE International Geoscience and Remote Sensing Symposium, Munich, Germany, 22–27 July 2012; pp. 4442–4445. 40. Yasuda, T.; Furuya, M. Short-term glacier velocity changes at West Kunlun Shan, Northwest Tibet, detected by synthetic aperture radar data. Remote Sens. Environ. 2013, 128, 87–106. [CrossRef] 41. Strozzi, T.; Kouraev, A.; Wiesmann, A.; Wegmüller, U.; Sharov, A.; Werner, C. Estimation of Arctic glacier motion with satellite L-band SAR data. Remote Sens. Environ. 2008, 112, 636–645. [CrossRef] Water 2016, 8, 496 18 of 18

42. Quincey, D.J.; Copland, L.; Mayer, C.; Bishop, M.; Luckman, A.; Belo, M. Ice velocity and climate variations for Baltoro Glacier, Pakistan. J. Glaciol. 2009, 55, 1061–1071. [CrossRef] 43. Zhou, J.; Li, Z.; Li, X.; Liu, S.; Chen, Q.; Xie, C.; Tian, B. Movement estimate of the Dongkemadi Glacier on the Qinghai-Tibetan Plateau using L-band and C-band spaceborne SAR data. Int. J. Remote Sens. 2011, 32, 6911–6928. [CrossRef] 44. Kääb, A.; Nuth, C.; Treichler, D.; Berthier, E. Brief Communication: Contending estimates of 2003–2008 glacier mass balance over the Pamir-Karakoram-Himalaya. Cryosphere 2015, 9, 557–564. [CrossRef] 45. Wu, K.; Liu, S.; Guo, W.; Wei, J.; Xu, J.; Bao, W.; Yao, X. Glacier change in the western Nyainqentanglha Range, Tibetan Plateau using historical maps and Landsat imagery: 1970–2014. J. Mt. Sci. 2016, 13, 1358–1374. [CrossRef] 46. Meier, M.F.; Post, A. What are glacier surges? Can. J. Earth Sci. 1969, 6, 807–817. [CrossRef] 47. Copland, L.; Sylvestre, T.; Bishop, M.; Shroder, J.; Seong, Y.; Owen, L.; Bush, A.; Kamp, U. Expanded and recently increased glacier surging in the Karakoram. Arct. Antarct. Alp. Res. 2011, 43, 503–516. [CrossRef] 48. Quincey, D.J.; Braun, M.; Glasser, N.F.; Bishop, M.P.; Hewitt, K.; Luckman, A. Karakoram glacier surge dynamics. Geophys. Res. Lett. 2011, 38, L18504. [CrossRef] 49. Quincey, D.J.; Glasser, N.F.; Cook, S.J.; Luckman, A. Heterogeneity in Karakoram glacier surges. J. Geophys. Res. Earth Surf. 2015, 120, 1288–1300. [CrossRef] 50. Shangguan, D.; Liu, S.; Ding, Y.; Guo, W.; Xu, J.; Jiang, Z. Characterizing the May 2015 Karayaylak Glacier surge in the eastern Pamir Plateau using remote sensing. J. Glaciol. 2016, 62, 1–10. [CrossRef] 51. Guo, W.; Liu, S.; Wei, J.; Bao, W. The 2008/09 surge of central Yulinchuan glacier, northern Tibetan Plateau, as monitored by remote sensing. Ann. Glaciol. 2013, 54, 299–310. [CrossRef]

© 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).