6. Tielidze et al..indd 175Tielidze, L.G. et al. Hungarian Geographical Bulletin 69 (2020) (2) 175–189.DOI: 10.15201/hungeobull.69.2.6 Hungarian Geographical Bulletin 69 2020 (2) 175–189. Introduction Climate change is causing nearly all of the world’s mountain glaciers to lose mass (Vaughan, D.G. et al. 2013) which now ac- counts for about one third of the cryosphere’s total contribution to global sea level rise (Gardner, A.S. et al. 2013). Mountain glaciers also play an important role in the regional hydrological cycle by modulating the storage and release of freshwater, so as they retreat, the availability of runoff for irrigation and hy- dro-power generation is altered (Kaltenborn, B.P. et al. 2010) and there are potential impacts on ecosystem health (Jacobsen, D. et al. 2012). The most important resource provided by glaciers for Georgia is freshwater. Many riv- ers in the mountain regions are fed by the 1 Department of Geomorphology, Vakhushti Bagrationi Institute of Geography, Ivane Javakhishvili Tbilisi State University, 6 Tamarashvili st., 0177, Tbilisi, Georgia. Correspondent author’s e-mail: tielidzelevan@gmail.com 2 Antarctic Research Centre, Victoria University of Wellington, PO Box 600, 6140, Wellington, New Zealand. 3 School of Geography Environment and Earth Sciences, Victoria University of Wellington, PO Box 600, 6140, Wellington, New Zealand. 4 Department of Geography, Faculty of Exact and Natural Sciences, Ivane Javakhishvili Tbilisi State University, 1 Chavchavadze Ave., 0128, Tbilisi, Georgia. 5 Formerly at: Climate Change Institute, University of Maine, Orono, ME 04469 USA. – † deceased. 6 Natural Resources and Environmental Studies, University of Northern British Columbia, 3333 University Way, Prince George, V2N 4Z9, BC, Canada. A 54-year record of changes at Chalaati and Zopkhito glaciers, Georgian Caucasus, observed from archival maps, satellite imagery, drone survey, and ground-based investigation Levan G. TIELIDZE1,2,3, David SVANADZE4, Lela GADRANI1,5, Lasha ASANIDZE 1, Roger D. WHEATE 6 and Gordon S. HAMILTON5† Abstract Individual glacier changes are still poorly documented in the Georgian Caucasus. In this paper, the change of Chalaati and Zopkhito glaciers in Georgian Caucasus has been studied between 1960 and 2014. Glacier geom- etries are reconstructed from archival topographic maps, Corona and Landsat images, along with modern field surveys. For the first time in the Georgian Caucasus aerial photogrammetric survey of both glacier termini was performed (2014) using a drone or Unmanned Aerial Vehicle, where high-resolution orthomosaics and digital elevation models were produced. We show that both glaciers have experienced area loss since 1960: 16.2±4.9 per cent for Chalaati Glacier and 14.6±5.1 per cent for Zopkhito Glacier with corresponding respective terminus retreat by ~675 m and ~720 m. These were accompanied by a rise in the equilibrium line altitudes of ~35 m and ~30 m, respectively. The glacier changes are a response to regional warming in surface air temperature over the last half century. We used a long-term temperature record from the town of Mestia and short-term meteorologi- cal observations at Chalaati and Zopkhito glaciers to estimate a longer-term air temperature record for both glaciers. This analysis suggests an increase in the duration of the melt season over the 54-year period, indicating the importance of summertime air temperature trends in controlling glacier loss in the Georgian Caucasus. We also observed supra-glacial debris cover increase for both glaciers over the last half century: from 6.16±6.9 per cent to 8.01±6.8 per cent for Chalaati Glacier and from 2.80±6.3 per cent to 8.53±5.7 per cent for Zopkhito Glacier. Keywords: glacier change, glacier monitoring, supra-glacial debris cover, climate change, Greater Caucasus, drone survey, Chalaati Glacier, Zopkhito Glacier Received January 2020, Accepted May 2020. Tielidze, L.G. et al. Hungarian Geographical Bulletin 69 (2020) (2) 175–189.176 melting of glaciers and snow. The largest gla- ciers in the Georgian Caucasus (2014) such as Lekhziri (~23 km2), Tsaneri (~12 km2), and Chalaati (~ 10 km2), feed the Enguri River which itself is the most important source of freshwater and hydropower in Georgia. Electricity is generated by a hydroelectric power plant and dam along the Enguri, which is the third highest concrete arch dam in the world with a height of 271.5 metres (Blatter, J. and Ingram, H.M. 2001). A con- tinued retreat of Georgian glaciers could lead to considerable changes in glacier runoff, with implications for regional water resourc- es. Therefore, continued monitoring of gla- cier behaviour across Georgia is necessary. After the dissolution of the Soviet Union, most glaciological monitoring programs stopped in the Caucasus region and the ini- tiation of new monitoring sites was difficult. Even though some important information on recent glacier change (mainly focused on gla- cier mapping) has become available for the Georgian Caucasus (Stokes, C.R. et al. 2006; Lambrecht, A. et al. 2011; Shahgedanova, M. et al. 2014; Tielidze, L.G. 2016; Tielidze, L.G. and Wheate, R.D. 2018; Tielidze, L.G. et al. 2020), the status of individual glaciers is poorly doc- umented. In this paper, we use the same ap- proach as we have applied on glacier change at a regional scale (Tielidze, L.G. and Wheate, R.D. 2018; Tielidze, L.G. et al. 2020) along with limited drone survey and ground-based obser- vations, such as ablation tracking, temperature observations, and terminus surveying. The Chalaati (GLIMS ID – G042713E43130N) and Zopkhito (GLIMS ID – G043422E42884N) glaciers were chosen because modern ground- based measurements are not yet available from any other glaciers in the Georgian Caucasus. The goals of our paper are: i) to reconstruct the dynamics of Chalaati and Zopkhito glaciers over the last half cen- tury, by estimating the length and area chang- es, and to compare the observed changes to those of glaciers from the northern side of the Greater Caucasus and other mountain regions (e.g. European Alps, Middle East, Western Himalaya); ii) to reconstruct the longer-term air temper- ature record for both glaciers in 1960–2014; to observe the length of the melt season (defined as temperatures above 0 °C) during the same time, and to estimate the change of equilib- rium line altitude (ELA); iii) to assess the alteration of the supra-gla- cial debris cover for both glaciers over the last half century. Study area The Greater Caucasus is one of the major mountain systems in Eurasia, stretching ~1,300 km from the Black Sea in the West to the Caspi- an Sea in the East. A recently published inven- tory lists ~2,000 glaciers with ~1,200 km2 total area (Tielidze, L.G. and Wheate, R.D. 2018). The main mountain range exerts a mod- erating influence on the climate of Georgia by protecting it against the penetration of cold air masses from the North. Most mois- ture-bearing weather systems arrive from the West having passed over the Black Sea. Orographic lifting of convergent air masses in western Georgia creates favourable condi- tions for snowfall at any time of the year in the high mountains. In contrast, a secondary pattern of weather systems originates in the drier continental climate to the East of the Caspian Sea. These meteorological condi- tions give rise to a strong West–East gradient in precipitation, reflected in annual snowfalls of several metres in the western parts of the Greater Caucasus and less than a metre in the East (Jincharadze, Z. 2011). An analysis of meteorological observations collected over the period 1957–2006 shows mean annual air temperatures have increased 0.2 °C in western Georgia and 0.3 °C in east- ern Georgia (Jincharadze, Z. 2011), with spring and summer months representing the most rapid warming. There has also been a modest increase in precipitation in western Georgia. Given the observed changes in cli- mate, there is considerable interest in under- standing their net effect on glaciers on the southern slopes of the Greater Caucasus. 177Tielidze, L.G. et al. Hungarian Geographical Bulletin 69 (2020) (2) 175–189. The central Greater Caucasus (Svaneti- Racha section) is the highest part of the main mountain range in morphometry. The relief of the Svaneti-Racha section is mainly construct- ed from Proterozoic and Lower Paleozoic pla- giogranites, plagiogneisses, quartz diorites and crystalline slates. Lower Jurassic clay slates, schists, sandstones, aleurolites, grave- lites, basal conglomerates, and quartzites stretch along the intrusives as a narrow strip in the south. The morphology of the water- shed range of the central Greater Caucasus is formed as a result of modern snow-glaciers influence, intense physical weathering, exca- ration action of Late Pleistocene glaciers and river erosion (Tielidze, L.G. et al. 2019a, b). We focus on two glaciers in the central part of the Greater Caucasus (Figure 1, a). Chalaati Glacier is located at the headwaters of the Enguri River, the main river for hydroelec- tric power generation in Georgia (Svaneti re- gion). Zopkhito Glacier is in the Rioni River basin, approximately 60 km to the south-east of Chalaati Glacier (Racha region). Fig. 1. Study area in the central part of the Greater Caucasus. – a = the location of Chalaati and Zopkhito glaciers and Mestia weather station. Blue colour corresponds to the glaciers, green to the forest zone, and brown to te bedrock. Landsat 8 OLI (03/08/2014) is used as the background; b = Chalaati Glacier; c = Zopkhito Glacier location with surrounding area in close view. Note: GeoEye (2012) are used as the background. Tielidze, L.G. et al. Hungarian Geographical Bulletin 69 (2020) (2) 175–189.178 Chalaati Glacier consists of two tributary glaciers which are fed by snowfall from the 4,000 m peaks: Ushba, Chatini, Cavcasi and Bzhedukhi (Figure 1, b). Its main channel (the eastern tributary) is about 6.8 km long, and the glacier terminus intrudes into the forest zone at an elevation of 1,960 m above sea lev- el (a.s.l.), making it the lowest-elevation ter- minus on the southern slopes of the Greater Caucasus. The glacier has a total surface area of 10.73±0.53 km2 and its lower reaches are covered by ~0.1–0.3 m thick debris (in 2014). Zopkhito Glacier is a simple valley glacier beginning on the south-eastern slopes of Geze peak (4,009 m) (Figure 1, c). It is ~3.6 km long with an area of 2.46±0.12 km2. The ice surface of the cirque sits at an elevation of approximately 3,000 m a.s.l., and the ice tongue ends at 2,605 m. Data and methods Dataset We seek to reconstruct the extents of Cha- laati and Zopkhito glaciers using archival and modern datasets. Baseline data are from 1:50,000 military topographic maps drawn in the 1960s and co-registered by Tielidze, L.G. and Wheate, R.D. (2018). The modern sequence of glacier terminus positions is established using cloud-free Corona (20/09/1971), Landsat 5 TM (6/08/1986), Landsat 7 ETM+ (09/09/2000), and Landsat 8 OLI (03/08/2014) images (Table 1). The images were orthorectified prior to distribution us- ing the ASTER Global Digital Elevation Model (GDEM, 17/11/2011). All images and GDEM were supplied by the US Geological Survey’s Earth Resources Observation and Science (EROS) Center and downloaded using the EarthExplorer tool (http://earthexplorer.usgs. gov/). The images have been co-registered to each other using the August 2014 Landsat im- age as master; registration uncertainties are 1 pixel (30 m). Glacier mapping Landsat images have a pixel resolution of 30 m for the bands used in this study. To facilitate mapping the glacier boundaries, we produced a colour-composite scene for each acquisition date, using the short-wave infrared, near infrared, and blue bands. Each glacier boundary was manually digitized by a single operator. Manual digitiz- ing by an experienced analyst is usually more accurate than automated methods for glaciers with debris cover (Raup, B.H. et al. 2007), such as Chalaati and Zopkhito. Combining the imag- es with topographic maps allows us to estimate the variability of Chalaati and Zopkhito glaciers over four periods corresponding to 1960–1971, 1971–1986, 1986–2000, and 2000–2014. We map the equilibrium line altitude from Landsat 8 OLI image (03/08/2014), towards the end of the ablation season in August, for comparison with equilibrium line altitude in 1960 mapped by Gobejishvili, R.G. (1995). Terminus measurements were conducted by using the glacier outlines for each date, along the ice front – perpendicular to the flow. Table 1. Topographic maps, satellite/ortho images and digital elevation model used in this study Date Map/Sensor Resolution Scene ID 1960 1:50,000 topographic map 5 m k_38_26_v k_38_39_b 20/09/1971 Corona 2 m DS1115-2154DF070_d DS1115-2154DA079 06/08/1986 05/09/2000 03/08/2014 Sept. 2014 17/11/2011 Landsat 5 TM Landsat 7 ETM+ Landsat 8 OLI Orthomosaics ASTER GDEM 30 m 15/30 m 15/30 m 20 cm 30 m LT51710301986218XXX02 LE71710302000249SGS00 LC81710302014215LGN00 UAV_DJI Phantom 4_Chalaati_Zophito ASTGTM2_N43E042/E043_DEM 179Tielidze, L.G. et al. Hungarian Geographical Bulletin 69 (2020) (2) 175–189. Terminus measurement by Unmanned Aerial Vehicle Using the Unmanned Aerial Vehicle (UAV) DJI Phantom 4 pro quadcopter, we performed a limited (terminus only) aerial survey (2014) of the Chalaati and Zopkhito glaciers. Flight planning for the UAV was completed in the office using mission planner software and Google Earth. The UAV was capable of oper- ating at elevation ~3,000 m a.s.l. Total weight including camera and battery was ~1.4 kgs. The maximum flight time for the platform at 3,000 m a.s.l. was around 20 min on a single 5,870 mAh battery. Ground station control was managed by a field tablet running APM Mission Planner for Android. The maximum length of an individual flight line was 0.4 km from the take-off point. The UAV images were processed using the Pix4D software. A dense point cloud was generated from the sparse point cloud model. The DEMs were generated at 20 cm pixel resolution and RGB orthomosaics were created at 20 cm pixel res- olution (Figure 2). The uncertainty between terminus by Landsat image (2014) and drone survey from the same year was ±20 m. Mapping of supra-glacial debris cover and uncertainty assessment Supra-glacial debris cover area clearly visible on the 1960s topographic map and Corona Fig. 2. An example of high resolution (20 cm) orthomosaics and hillshade (generated by 20 cm DEMs) for the Chalaati (a-b) and Zopkhito (c-d) glaciers. Tielidze, L.G. et al. Hungarian Geographical Bulletin 69 (2020) (2) 175–189.180 images from 1971 allowed us to map it manu- ally. For the Landsat images the band ratio segmentation method (RED/SWIR; with a threshold of ≥2.0) was used as the first step in delineating clean-ice outlines (Bolch, T. et al. 2010; Paul, F. et al. 2013), and then intensive manual improvements were performed (re- moval of misclassified areas, e.g. snow, shad- ows), hereafter called the semi-automated method. In the next step, similar to Tielidze, L.G. et al. (2020), supra-glacial debris cover was classified as the residual between semi- automatically derived clean-ice outlines and manually mapped glacier extent outlines. For clean ice uncertainty estimation we used a 15 m (1/2 pixel) buffer (Bolch, T. et al. 2010) and for debris-covered parts 60 m (two pixels) (Frey, H. et al. 2012). This generated an aver- age uncertainty for the clean-ice/debris-cov- ered areas of 4.9%/6.6% for 1960, 5.0%/6.5% for 1971, 4.7%/6.4% for 1986, 4.7%/6.3% for 2000, and 4.9%/6.4% for 2014. The uncer- tainty estimates for all Caucasus glaciers are described in previous studies (Tielidze, L.G. 2016; Tielidze, L.G. and Wheate, R.D. 2018; Tielidze, L.G. et al. 2020). As an independent assessment of the un- certainty estimates, Zopkhito Glacier out- lines from Landsat OLI 8 (03/08/14) (includ- ing clean-ice and debris-covered parts) were imported into Google Earth and manually adjusted using the available high-resolution Quickbird images (19/09/11) superimposed upon the SRTM3 topography (Raup, B.H. et al. 2014). These glacier outlines were then com- pared with original outlines from the Landsat 8 image (03/08/14). The area differences be- tween the two resulting sets of outlines were ±5.9 per cent for supra-glacial debris cover and ±3.8 per cent for clean-ice. We were not able to use Google Earth software for Chalaati Glacier due to lack of cloud-free images, how- ever, we used high-resolution GeoEye image from DigitalGlobe (ArcGlobe 10.6.1 software) as proposed by Paul, F. et al. (2013). We calcu- lated the area uncertainty in a similar way for the Chalaati Glacier. The area differences be- tween the two datasets were ±5.7 per cent for supra-glacial debris cover and ±3.5 per cent for clean-ice which confirms our uncertainty estimate based on the buffer method. An additional uncertainty assessment was performed using GPS (Garmin 62stc) meas- urements of glacier margins (~230 points) obtained during field investigations in 2014. The horizontal accuracy of these measure- ments varied from ±4 to ±10 m. Upper part of Figure 3 (‘a’ and ‘b’) shows the results of comparison between GPS measurements and Landsat based supra-glacial debris cover out- lines. The average accuracy based on both Chalaati and Zopkhito glaciers measure- ments was ±30 m for supra-glacial debris cover, hence again confirming the suitability of the selected buffer method. Ground-based investigation We carried out a limited amount of field work on each glacier. During a three-month period in 2011 (June-August), hourly air temperature observations were made at an elevation of 2,140 m on Chalaati Glacier us- ing a Campbell CR21 data logger. We also tracked ablation by measuring exposed stake heights at several locations across the glacier every 7–10 days (Figure 3, c). During a brief follow-up field visit on August 16, 2014, we surveyed the terminus using global position- ing system (GPS) methods. Similar observations including the Automatic Weather Station (AWS) installa- tion at an elevation of 2,700 m were carried out on Zopkhito Glacier during the months of July and August in the summers of 2007– 2010, and August 2014 (Figure 3, d). Results and discussion Chalaati Glacier Over the study period, Chalaati Glacier area de- creased by ~2.08±0.10 km2, equivalent to ~0.04 km2/yr. Rates of area loss have been variable (Table 2 and 3), with the fastest rate (~0.42% yr-1) occurring between 1960 and 1971. During the 181Tielidze, L.G. et al. Hungarian Geographical Bulletin 69 (2020) (2) 175–189. next 15-year period (1971–1986), the rate of area loss was still higher (~0.39% yr-1). The lowest decrease rate occurred in 2000–2014 (~0.18% yr-1). In contrast to area decrease, supra-glacial debris cover area has increased from 6.16±6.9 per cent to 8.01±6.8 per cent over the study- period, mostly in 2000–2014 (Figure 4). In addition to area changes, we also mapped linear retreat of the terminus since 1960. The fastest rates of retreat occurred dur- ing the first measurement period (1960–1971) while the lowest was measured in 1986–2000 (Table 4). This might be due to a minor 15 m re-advance of the terminus between 1990 and 1993 as evidenced by small terminal moraines (Gobejishvili, R.G. 1995). The equilibrium line altitude on Chalaati Glacier was located at 3,155 m a.s.l. in 2014. The cirque extended to 3,800 m and covered an area of 5.01±0.25 km2. Between 1960 and 2014, the elevation of the equilibrium line rose by ~35 metres, resulting in a decrease in the accumula- tion zone of ~0.44 km2 over the 54-year period. Using the measured areas, we calculated the ac- cumulation-area ratio (AAR) (Table 5). Chalaati Glacier total area has decreased at a faster rate than the accumulation area, so the AAR has ac- tually increased over the study period, although Fig. 3. Example of glacier mapping. – a = Chalaati Glacier – GeoEye (2012) image is used as the background; b = Zopkhito Glacier – Google Earth (19/09/11) image is used as the background; c = Ablation stake installa- tion onto the Chalaati Glacier; d = Automatic Weather Station (AWS) installation onto the Zopkhito Glacier Tielidze, L.G. et al. Hungarian Geographical Bulletin 69 (2020) (2) 175–189.182 the estimated uncertainty is comparable to the obtained relative changes. The comparison of the photos of the glacier terminus in 1970 and in 2011 clearly shows the retreat of the Chalaati Glacier over the last half century (Photo 1). Zopkhito Glacier Over the study period, the area of Zopkhito Glacier decreased from 2.88±0.14 km2 (1960) to 2.46±0.12 km2 (2014), equivalent to a rate of ~0.007 km2/yr. The rate of area loss was the fastest during the most recent part of the re- cord (~0.48% yr-1 between 2000 and 2014) (see Table 2 and 3). Most of the area loss occurred in the ablation area of the glacier, where there was a steady retreat of the Zopkhito Glacier terminus, from 2,435 m a.s.l. (1960) to 2,605 m a.s.l. (2014). This led to a ~720 m reduction in the length of the glacier since 1960 (see Table 4). In contrast to total area decrease, supra- glacial debris covered area increased from 2.80±6.3 per cent to 8.53±5.7 per cent for the Zopkhito Glacier in 1960–2014. The highest increase rate occurred in the period 2000– 2014 (see Figure 4). Fig. 4. Supra-glacial debris covered (SDC) area in percentage of total glacier area for Chalaati and Zopkhito glaciers between 1960 and 2014. Table 2. Terminus position, area, and supra-glacial debris cover change since 1960 for Chalaati and Zopkhito glaciers Year Chalaati Zopkhito Terminus, m a.s.l. Total area, km2 Clean ice area, km2 Debris covered area, km2 Terminus, m a.s.l. Total area, km2 Clean ice area, km2 Debris covered area, km2 1960 1971 1986 2000 2014 1,800 1,860 1,900 1,920 1,960 12.81±0.64 12.31±0.62 11.59±0.56 11.09±0.54 10.73±0.53 12.02±0.59 11.51±0.57 10.81±0.51 10.34±0.49 9.87±0.48 0.79±0.050 0.80±0.050 0.78±0.050 0.75±0.050 0.86±0.050 2,435 2,475 2,525 2,550 2,605 2.88±0.14 2.81±0.14 2.72±0.13 2.64±0.13 2.46±0.12 2.80±0.13 2.71±0.13 2.63±0.12 2.52±0.12 2.25±0.11 0.08±0.005 0.10±0.006 0.10±0.006 0.12±0.007 0.21±0.012 Table 3. Chalaati and Zopkhito glaciers area change between 1960 and 2014 Year Chalaati Zopkhito Decrease, ~ km2 Annual decrease, ~ km2 Annual decrease, ~ % yr-1 Decrease, ~ km2 Annual decrease, ~ km2 Annual decrease, ~ % yr-1 1960–1971 1971–1986 1986–2000 2000–2014 1960–2014 0.60 0.72 0.50 0.36 2.08 0.05 0.05 0.04 0.03 0.04 0.42 0.39 0.30 0.18 0.30 0.07 0.09 0.08 0.18 0.42 0.006 0.006 0.005 0.120 0.007 0.22 0.22 0.21 0.48 0.28 Table 4. Linear retreat rate of Chalaati and Zopkhito glaciers since 1960* Time periods Chalaati Zopkhito Retreat of glacier terminus Retreat of glacier terminus m m yr-1 m m yr-1 1960–1971 1971–1986 1986–2000 2000–2014 1960–2014 270 135 80 190 675 24.5 9.6 5.7 13.6 12.5 190 180 120 230 720 17.3 12.0 8.6 16.4 13.3 *The average uncertainty for length change are ±15 m. 183Tielidze, L.G. et al. Hungarian Geographical Bulletin 69 (2020) (2) 175–189. In 2014, the equilibrium line on Zopkhito Glacier was located at 3,080 m a.s.l. and encompassed cirque basin of an area of 1.69±0.08 km2, extending to 3,800 m eleva- tion a.s.l. There was a ~30 m rise in the el- evation of the equilibrium line between 1960 and 2014, resulting in a decrease in the accu- mulation zone of ~0.31 km2 over the 54-year period. Using the measured areas of snow accumulation and ablation, we calculate the accumulation-area ratio (AAR) which shows that Zopkhito Glacier’s total area decreased at a slightly faster rate than the accumulation area (see Table 5). We note, that this insignifi- cant changes of the AAR remain within un- certainties, which means that AAR remained unchanged during the study-period. The comparison of the photos of the glacier ter- minus shows the change of Zopkhito Glacier over the last half century (Photo 2). Temperature Our study shows that both glaciers are expe- riencing constant retreat, except for a short period of re-advance of the Chalaati Glacier. We examine these changes in the context of regional air temperature conditions. The hourly temperature measurements only ex- ist for occasional short (2–3 months) periods at Chalaati (summer 2011) and Zopkhito glaciers (summers 2008–2009), but detailed records are available for the weather station in the settlement of Mestia, located ~7 km down-valley from the terminus of Chalaati Table 5. Accumulation-area ratio (AAR) change for Chalaati and Zopkhito glaciers between 1960 and 2014 Years Chalaati Zopkhito AAR Accumulation (firn) area, km2 Total area, km2 Accumulation (firn) area, km2 Total area, km2 Chalaati Zopkhito 1960 1971 1986 2000 2014 5.01±0.25 4.95±0.25 4.72±0.22 4.54±0.21 4.57±0.22 12.81±0.64 12.31±0.62 11.59±0.56 11.09±0.54 10.73±0.53 1.69±0.08 1.66±0.08 1.61±0.07 1.60±0.07 1.38±0.06 2.88±0.14 2.81±0.14 2.72±0.13 2.64±0.13 2.46±0.12 0.39±0.44 0.40±0.43 0.40±0.39 0.40±0.38 0.42±0.38 0.58±0.11 0.59±0.11 0.59±0.10 0.60±0.10 0.56±0.09 Photo 1. Chalaati Glacier terminus in 1975 (left) and in 2011 (right). Photos by Gobejishvili, R.G. (1975), and by Tielidze, L.G. (2011). Tielidze, L.G. et al. Hungarian Geographical Bulletin 69 (2020) (2) 175–189.184 Glacier (~60 km from Zopkhito Glacier), at an elevation of 1,440 m a.s.l. (see location on Figure 1). Observations are available for the period between 1960 and 2014. The summertime hourly temperature measurements at both glaciers are in agree- ment with the temperature records at Mestia and enable us to compute the empirical lapse rate for the region (-9.8 °C/km between Mestia and Chalaati, and -7.8 °C/km between Mestia and Zopkhito). They also allow us to establish a transfer function that can extend the record of air temperatures at each glacier back to 1960. For the transfer function, we need to confirm that temperatures at both sites are well correlated. A linear regression yields a correlation coefficient of 0.89 be- tween Mestia and Chalaati, and 0.82 between Mestia and Zopkhito (Figure 5), showing a significant correlation between both glacier records and Mestia weather station. The estimated mean annual temperatures at both glaciers are below to the 0 °C for the entire record. Zopkhito Glacier is colder than Chalaati, as would be expected from its higher eleva- tion terminus (~650 m higher, than Chalaati). In general, the warmest temperatures occur in July (Figure 6), and the melt season (defined as temperatures above 0 °C) lasts an average of 184 days at Chalaati and 145 days at Zopkhito. There has been an increase in the length of the melt season at both glaciers (Figure 7). The in- Photo 2. Zopkhito Glacier terminus in 1966 (a), and in 2010 (b). Photos by Inashvili, Sh. (1966), and by Svanadze, D. (2010). Fig. 5. Correlation between summertime hourly tem- perature observations at Mestia weather station (1,440 m a.s.l.) in 1960–2014, and local hourly temperatures measured at Chalaati (2,140 m a.s.l.), and Zopkhito (2,700 m a.s.l.) glaciers during the same time. Fig. 6. Average summertime monthly air tempera- tures (1960–2014) at Mestia weather station (1,440 m a.s.l.), and estimated values at Chalaati (2,140 m a.s.l.), and Zopkhito (2,700 m a.s.l.) glaciers 185Tielidze, L.G. et al. Hungarian Geographical Bulletin 69 (2020) (2) 175–189. creasing length of the melt season is consistent with a general trend of warming air tempera- tures over the period 1960–2014 (Figure 8). We note that because of the very limited observa- tion time, this has to be considered as a tentative estimate of the air temperatures at the glaciers. Equilibrium Line Altitude For estimating the equilibrium line altitude (ELA) it is appropriate to use at least two var- iables, precipitation, and temperature, which represent the effects of accumulation and ab- lation, respectively (Ohmura, A. et al. 1992). Often, the annual mean 0 °C isotherm is also used as the ELA (Källén, Ε. et al. 1979; Oer- lemans, J. and Van der Veen, C.J. 1984). For this purpose, we use the Mestia temperature record with calculated empirical lapse rate to estimate the height of the summer (June, July, August) 0 °C isotherm for comparison with the observed (by satellite imagery, 2014) ELA. For Chalaati Glacier, the estimated 0 °C isotherm height is 3,052 m a.s.l., which is ~100 m lower than the observed ELA el- evation of 3,155 m a.s.l. For Zopkhito Gla- cier, the estimated 0 °C isotherm occurs at 3,465 m a.s.l., which is ~400 m higher than the observed ELA at 3,080 m a.s.l. Surface air temperatures are not the sole control on snow line altitude, but lacking information on regional precipitation characteristics, we are unable to fully explain the difference in equilibrium line and summertime 0 °C iso- therm altitudes. However, we postulate that the respective aspect of each glacier con- tributes to the difference in offset between the equilibrium line altitude and the height of the 0 °C isotherm. Chalaati Glacier has a predominantly south-facing aspect, which might mean solar radiation is able to drive additional melting and raise the equilibrium line altitude to above the summertime freez- ing isotherm. In contrast, Zopkhito Glacier faces predominantly east and is shaded from the sun by a steep ridge, which might allow snow to survive to an altitude below the re- gional 0 °C isotherm height. Fig. 7. Duration of melt season for Chalaati and Zopkhito glaciers in 1960–2014 and linear trends. Fig. 8. Time series of monthly air temperature anomalies at Mestia weather station with respect to the 1960–2014 average. A 24 month smoothed anomaly is shown by the thick blue line. The red line is the trend showing a modest increase in warm anomalies with time. Tielidze, L.G. et al. Hungarian Geographical Bulletin 69 (2020) (2) 175–189.186 Ablation Field observations at both glaciers of stake heights exposed by the ablation allow us to ex- amine the role of air temperature on ablation. On the basis of the derived lapse rates, we can use the Mestia temperature record to produce a ‘local’ temperature record for each stake loca- tion and compute the cumulative positive de- gree days (PDD) for each site (Hock, R. 1999). The stake observations for Chalaati Glacier were made in summer 2011 at an elevation of 2,040 m a.s.l. The sum of air temperatures and observed ablation in July exceed the same indicators for August. For a 27-day pe- riod from July 4 to July 31, the sum of PDDs is 298.8 °C, and measured ablation was 172.5 cm, yielding an ablation rate of ~0.6 cm/PDD. Over a 31-day period from July 31 until August 31, PDDs summed to 289.7 °C, and observed melting was 129.0 cm, corre- sponding to an ablation rate of ~0.4 cm/PDD. Repeating the analysis for Zopkhito Glacier using observations from 2008 for a stake at 2,700 m a.s.l. yields an ablation rate of ~0.6 cm/PDD in July, and ~0.5 cm/PDD in August. A partial explanation for the difference in de- rived melt rates is solar angle, which in July is farther above the horizon than in August and, thus, supplies a greater amount of in- coming radiation to melt the glacier surface. Comparison with other studies Direct comparisons of glacier change with previous investigations in the Greater Cau- casus are difficult because most of them do not deal with individual glaciers. Therefore, our rates (0.2–0.3% yr-1) are much lower than other regional studies of glacier changes in the Greater Caucasus; e.g. Tielidze, L.G. and Wheate, R.D. (2018) found generally higher rates of glacier shrinkage for south-facing glaciers during the same investigation period (0.69% yr-1). This high rate can be explained by the disappearance of small glaciers (<0.5 km2) from the regional study by Tielidze, L.G. and Wheate, R.D. (2018) in 1960–2014. Comparison to Mount Ararat glaciers (39.70°N, 44.30°E) in the Middle East, show that our rates are significantly higher. The glacier area of Mount Ararat has decreased from 7.98±0.80 km2 to 5.66±0.57 km2, equiva- lent to 29 per cent area loss (or 0.83% yr-1) be- tween 1976 and 2011 (Sarikaya, M.A. 2012). This can partly be explained by the warm and dry climate in the Middle East versus the Greater Caucasus. The continued existence of glaciers, like Chalaati and Zopkhito, at elevations above the summertime 0 °C isotherm altitude is probably due to their topographic setting (Grunewald, K. and Scheithauer. J. 2010) in which surrounding high peaks and steep slopes promote snow accumulation through avalanching and wind-driven processes. Little is known about accumulation and pre- cipitation patterns in Georgian Caucasus, but any future decrease in precipitation might lead to increased rates of glacier loss as re- gional temperatures continue to warm. Since the glacier snout recession is a more sensitive indicator of changes at decadal timescale than area change (Bhambri, R. et al. 2012; Leclercq, P.W. et al. 2014), we compared Chalaati and Zopkhito glaciers cumulative length changes with other similar types of gla- ciers from the northern Greater Caucasus. The comparison shows that both glaciers experi- enced higher retreat rates than the northern counterparts (Figure 9), which is in agreement with other studies suggesting that southern fac- ing glaciers are melting faster than northern ones (Shahgedanova, M. et al. 2014; Tielidze, L.G. and Wheate, R.D. 2018). This might be ex- plained by relatively high radiation input in the southern slopes. Comparison with glaciers from the European Alps shows that Greater Caucasus glaciers are retreating more steadily while the glaciers from the European Alps expe- rience several advancing stages during the same time. Chalaati and Zopkhito glaciers’ retreat was also similar in comparison with Sonapari Glacier from Western Himalaya in 1970–2000 period, while it was different in 2000–2016, when Sonapari Glacier experi- 187Tielidze, L.G. et al. Hungarian Geographical Bulletin 69 (2020) (2) 175–189. enced its highest retreat rate (~37.3 m yr-1) (Majeed, Z. et al. 2020) (see Figure 9). These differences can be attributed to different me- teorological conditions, orographic units, and morphological types of glaciers between these mountain regions. Conclusions We observed a substantial loss in the area of two of the largest glaciers on the southern slope of the central Greater Caucasus, Geor- gia, between 1960 and 2014 based on an analy- sis of archival maps, modern satellite imagery, drone survey, and ground-based measure- ment. The main findings are as follows: i) Chalaati Glacier lost 16.2±4.9 per cent (~0.30% yr-1) of its area since 1960, while Zopkhito Glacier lost 14.6±5.1 per cent (~0.27% yr-1) over the same period. A slight- ly greater change in area of Chalaati Glacier might be due to its terminus extending to lower elevations than Zopkhito Glacier. ii) Chalaati Glacier experienced a termi- nus retreat rate of ~12.5 m yr-1 during the years 1960–2014 with highest retreat rate in 1960–1971 (~24.5 m yr-1), while the Zopkhito Glacier retreat rate was ~13.3 m yr-1 over the last half century, with highest rate in 1960– 1971 (~17.3 m yr-1). iii) The equilibrium line altitude has risen by ~35 metres for Chalaati Glacier (from 3,120 m to 3,155 m a.s.l.), and ~30 metres for Zopkhito Glacier (from 3,050 m to 3,080 m a.s.l.) result- ing in a decrease in the accumulation zone for both glaciers over the 54-year period. iv) There has been an increase in the esti- mated length of the melt season at both gla- ciers (defined as temperatures above 0 °C) lasting an average of 184 days at Chalaati, and 145 days at Zopkhito. v) The observed glacier loss is consistent with a 0.2 °C regional rise in near-surface an- nual air temperature over the last half century period as recorded at a weather station close to both glaciers. An increase in the mean sum- mer temperature (June, July, August) appears to be a particularly important factor in glacier shrinkage, as shown by an increase in the dura- tion of the melt season over the study period. Glaciers on the southern slopes of the Greater Caucasus are expected to continue their retreat as regional air temperatures rise. This deglaciation will have important conse- quences for the management of water resourc- es and hydropower generation in Georgia. Further work should focus on more detailed Fig. 9. Cumulative curves of glacier retreat for Chalaati and Zopkhito compared to glaciers from northern slope of the Greater Caucasus (Bartui, Marukhi Northern, Tsey – Solomina, O. et al. 2016; Tielidze, L.G. and Wheate, R.D. 2018); European Alps (Lower Grindelwald and Mer de Glace – Zumbühl, H.J. et al. 2008); and Western Himalaya (Sonapari – Majeed, Z. et al. 2020). Tielidze, L.G. et al. Hungarian Geographical Bulletin 69 (2020) (2) 175–189.188 field observations of climatic elements, and in- strumental recording of glacier processes such as ablation and supra-glacial debris cover dy- namics as well as ELA definition. Acknowledgements: The authors acknowledge the substantial contributions by co-author Hamilton, G.S. prior to his untimely death in Antarctica in 2016. We also gratefully acknowledge the support of the editor, Varga, G. and three anonymous reviewers, for useful suggestions and comments which clearly enhanced the quality of the paper. REFERENCES Bhambri, R., Bolch, T. and Chaujar, R.K. 2012. Frontal recession of Gangotri Glacier, Garhwal Himalayas, from 1965–2006, measured through high resolution remote sensing data. Current Science 102. 489–494. Blatter, J. and Ingram, H.M. (eds.) 2001. Reflections on Water: New Approaches to Transboundary Conflicts and Cooperation. Cambridge, MA, MIT Press. Bolch, T., Menounos, B. and Wheate, R.D. 2010. Landsat-based inventory of glaciers in western Canada, 1985–2005. Remote Sensing of Environment 114. 127–137. Doi: 10.1016/j.rse.2009.08.015. Frey, H., Paul, F. and Strozzi, T. 2012. Compilation of a glacier inventory for the western Himalayas from satellite data: methods, challenges, and results. Remote Sensing of Environment 124. 832–843. Gardner, A.S., Moholdt, G., Cogley, J.G., Wouters, B., Arendt, A.A., Wahr, J., Berthier, E., Hock, R., Pfeffer, W.T., Kaser, G., Ligtenberg, S.R.M., Bolch, T., Sharp, M.J., Hagen, J.O., van den Broeke, M.R. and Paul, F. 2013. A reconciled estimate of glacier contributions to sea level rise: 2003 to 2009. Science 340. (6134): 852–857. Doi: 10.1126/science.1234532. Gobejishvili, R.G. 1995. Present day glaciers of Georgia and evolution of glaciation in the moun- tains of Eurasia in late Pleistocene and Holocene. Thesis for a Doctor’s degree. Tbilisi. (in Russian) Grunewald, K. and Scheithauer. J. 2010. Europe’s southernmost glaciers: response and adaptation to climate change. Journal of Glaciology, 56. (195): 129–143. Hock, R. 1999. A distributed temperature-index ice- and snowmelt model including potential direct so- lar radiation. Journal of Glaciology 45. (149): 101–111. Jacobsen, D., Milner, A.M., Brown, L.E. and Dangles, O. 2012. Biodiversity under threat in glacier-fed river systems. Nature Climate Change 2. (5): 361–364. Available at http://dx.doi.org/10.1038/nclimate1435 Jincharadze, Z. 2011. Climate Change Scenarios over South Caucasus Region. Climate Change Impact for the South Caucasus. Report to Environment and Security Initiative. United Nations Development Programme, Vienna, UNIS. Kaltenborn, B.P., Nellemann, C. and Vistnes, I.I. (eds.) 2010. High Mountain Glaciers and Climate Change – Challenges to Human Livelihoods and Adaptation. United Nations Environment Programme. Oslo, GRID-Arendal. Available at http://www.grida.no. Källén, Ε., Crafoord, C. and Ghil, M. 1979. Free oscillation in a climate model with ice-sheet dynam- ics. Journal of the Atmospheric Sciences 36. (12): 2292– 2303. Doi.org/10.1175/1520-0469(1979)036<2292:FO IACM>2.0.CO;2. Lambrecht, A., Mayer, C., Hagg, W., Popovnin, V., Rezepkin, A., Lomidze, N. and Svanadze, D. 2011. A comparison of glacier melt on debris-covered glaciers in the northern and southern Caucasus. The Cryosphere 5. 525–538. Available at https://doi. org/10.5194/tc-5-525-2011 Leclercq, P.W., Oerlemans, J., Basagic, H.J., Bushueva, I., Cook, A.J. and Le Bris, R. 2014. A data set of worldwide glacier length fluctuations. The Cryosphere 8. 659–672. Doi:10.5194/tc-8-659-2014, 2014. https://doi.org/10.5194/tc-8-659-2014. Majeed, Z., Mukhtar, M.A., Mir, R.A., Kumar, P. and Krishna, K. 2020. Sonapani Glacier Recession over a Century from 1906–2016, Chandra Basin, Himachal Himalaya. Journal of Geological Society of India 95. 36–44. Doi:10.1007/s12594-020-1384-5. Oerlemans, J. and Van der Veen, C.J. 1984. Ice Sheets and Climate. Dordrecht–Boston–Lancaster, D. Reidel Publishing Company. Doi:10.1007/978-94- 009-6325-2. Ohmura, A., Kasser, P. and Funk, M. 1992. Climate at the equilibrium line of glaciers. Journal of Glaciology 38. (130): 397–411. Doi:10.3189/S0022143000002276. Paul, F., Barrand, N.E., Baumann, S., Berthier, E., Bolch, T., Casey, K., Frey, H., Joshi, S.P., Konovalov, V., Le Bris, R., Mölg, N. Nosenko, G., Nuth, C., Pope, A., Racoviteanu, A., Rastner, P., Raup, B., Scharrer, K., Steffen, S. and Winsvold, S. 2013. On the accuracy of glacier outlines derived from remote-sensing data. Annals of Glaciology 54. 171–182. Doi: 10.3189/2013AoG63A296. Raup, B., Kaeaeb, A., Kargel, J., Bishop, M.P., Hamilton, G.S., Lee, E., Rau, F., Paul, F., Soltesz, D., Singh Khalsa, S.J., Beedle, M. and Helm, C. 2007. Remote sensing and GIS technology in the Global Land Ice Measurements from Space (GLIMS) project. Computers and Geoscience 33. 104–125. Raup, B.H., Singh Khalsa, S.J., Armstrong, R.L., Sneed, W.A., Hamilton, G.S., Paul, F., Cawkwell, F., Beedle, M.J., Menounos, B.P., Wheate, R.D., Rott, H., Shiyin, L., Xin, L., Donghui, S., Guodong, C., Kargel, J.S., Larsen, C.F., Molnia, B.F., Kincaid, J.L., Klein, A. and Konovalov, V. 2014. Quality in the GLIMS glacier database. In Global 189Tielidze, L.G. et al. Hungarian Geographical Bulletin 69 (2020) (2) 175–189. Land Ice Measurements from Space. Eds.: Kargel, J.S., Leonard, G.J., Bishop, M.P., Kääb, A. and Raup, B.H., Berlin–Heidelberg, Springer, 163–182. Doi:10.1007/978-3-54079818-7_7. Sarikaya, M.A. 2012. Recession of the ice cap on Mount Ararat, Turkey, from 1976 to 2011 and its climatic significance. Journal of Asian Earth Sciences 46. 190–194. Doi.org/10.1016/j.jseaes.2011.12.009 Shahgedanova, M., Nosenko, G., Kutuzov, S., Rototaeva, O. and Khromova, T. 2014. Deglaciation of the Caucasus Mountains, Russia/Georgia, in the 21st century observed with ASTER satellite im- agery and aerial photography. The Cryosphere 8. 2367–2379. Doi:10.5194/tc-8-2367-2014. Solomina, O., Bushueva, I., Dolgova, E., Jomelli, V., Alexandrin, M., Mikhalenko, V. and Matskovsky, V. 2016. Glacier variations in the Northern Caucasus compared to climatic reconstructions over the past millennium. Global and Planetary Change 140. (1): 28–58. Doi: 10.1016/j.gloplacha.2016.02.008. Stokes, C.R., Gurney, S.D., Popovnin, V. and Shahgedanova, M. 2006. Late-20th-century changes in glacier extent in the Caucasus Mountains, Russia/ Georgia. Journal of Glaciology 52. (176): 99–109. Tielidze, L.G. 2016. Glacier change over the last cen- tury, Caucasus Mountains, Georgia, observed from old topographical maps, Landsat and ASTER satel- lite imagery. The Cryosphere 10. 713–725. Available at https://doi.org/10.5194/tc-10-713-2016 Tielidze, L.G. and Wheate, R.D. 2018. The Greater Caucasus glacier inventory (Russia, Georgia and Azerbaijan). The Cryosphere 12. 81–94. Available at https://doi.org/10.5194/tc-12-81-2018 Tielidze, L.G., Gobejishvili, R., Maruashvili, L. and Astakhov, N. 2019a. Central Greater Caucasus. In Geomorphology of Georgia. Geography of the Physical Environment. Ed.: Tielidze, L.G., Cham, Springer. Doi.org/10.1007/978-3-319-77764-1_9. Tielidze, L.G., Kumladze, R.M., Wheate, R.D. and Gamkrelidze, M. 2019b. The Devdoraki Glacier catastrophes, Georgian Caucasus. Hungarian Geographical Bulletin 68. (1): 21–36. Doi: 10.15201/ hungeobull.68.1.2. Tielidze, L.G., Bolch, T., Wheate, R.D., Kutuzov, S.S., Lavrentiev, I.I. and Zemp, M. 2020. Supra-glacial debris cover changes in the Greater Caucasus from 1986 to 2014. The Cryosphere 14. 585–598. Available at https://doi.org/10.5194/tc-14-585-2020 Vaughan, D.G., Comiso, J.C., Allison, I., Carrasco, J., Kaser, G., Kwok, R., Mote, P., Murray, T., Paul, F., Ren, J., Rignot, E., Solomina, O., Steffen, K. and Zhang, T. 2013. Observations: Cryosphere. In Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Eds.: Stocker, T.F., Qin, D., Plattner, G.K., Tignor, M., Allen, S.K., Boschung, J., Nauels, A., Xia, Y., Bex, V. and Midgley, P.M., Cambridge, UK, and New York, Cambridge University Press, 317–382. Zumbühl, H.J., Steiner, D. and Nussbaumer, S.U. 2008. 19th century glacier representations and fluctuations in the central and western European Alps: An interdisciplinary approach. Global and Planetary Change 60. (1): 42–57. Doi: 10.1016/j.glo- placha.2006.08.005. Tielidze, L.G. et al. Hungarian Geographical Bulletin 69 (2020) (2) 175–189.190