Glacier variations of Hielo Patagónico Norte, Chile, over 70 years from 1945 to 2015

Glacier variations of Hielo Patagónico Norte, Chile, over 70 years from 1945 to 2015 Masamu ANIYA 1 Abstract 1. Introduction Background

Glacier variations of Hielo Patagónico Norte, Chile, over 70 years from 1945 to 2015 Glaciers are known that in addition to inter-annual variations, they often fluctuate seasonally,., advance in winter and retreat in summer. Such annual moraines (washboard or De Geer moraines) are recognized at Gl. Soler (Aniya, 1985). So I made three winter aerial surveys in 2004 (July), 2005 (Aug.), and 2011 (July) to check if this is true for the HPN glaciers. I found that there were no substantial seasonal changes. Last time I reported the HPN variations was 2011 (Aniya 2011) in which the latest information was 2009. Since then, I made several aerial surveys and updated the variations to 2015. The purposes of this paper are to: (1) document in detail the variation of each glacier for 1945 2015 (70 years); (2) discuss their characteristics; and (3) consider possible factors for variations. 2. Study Area Hielo Patagónico Norte (HPN) Hielo Patagónico Norte is located in Chile near the tip of South America between 46 30 and 47 35 S along about 73 30 W (Fig. 1). It is about 100 km long from north to south and about 40 60 km wide. The area was ca. 4200 km2, with 28 major outlet glaciers when I made a comprehensive inventory of the outlet glacier (Aniya, 1988). About 20 years later, Rivera (2007) made another inventory using Landsat ETM+ data of 2001, in which measured area was ca. 3950 km2, with 48 outlet glaciers including minor ones. In order to maintain the consistency, Aniya s inventory data are used in this study (Fig. 2). The highest mountain in Patagonia, Monte San Valentin (3910 m), is located at the northeast corner of the HPN. In the middle of the icefield near the eastern edge, Cerro (Co.) Arenales (3365 m) looms above the surrounding icefield whose elevations range 1000 1500 m. To the south of Co. Arenales there are a few mountains with elevations 2500 3000 m, which are limiting the east side of the icefield. Of the 28 major outlet glaciers, 21 have been monitored for variation by Aniya. Now 17 of these 21 outlet glaciers are calving glaciers in proglacial lakes, except one tidewater glacier, Glaciar (Gl.) San Rafael that is located at the lowest latitude in the world at 46 33 S. Gl. San Rafael and San Quintin, which are located next to each other in the north-south direction, are two largest glaciers in the HPN with ca. 760 km2 each (Aniya, 1988). The HPN glaciers are temperate (Matsuoka and Naruse, 1999), characterized by large amounts of accumulation and ablation. The annual precipitation is estimated at ca. 6000 7000 mm (Escobar., 1992) on the icefield, while 3000 4000 mm are reported from ground recordings at Laguna San Rafael and Gl. Exploradores (Aniya., 2011). The ice thickness is estimated at 1460 ± 500 m around the center part of the icefield from the gravity anomalies (Casassa, 1987). No mass balance study in the field of any glacier has yet been carried out in the HPN. Fig. 1. Location of the Hielo Patagónico Norte (HPN or Northern Patagonia Icefield) and major outlet glaciers. Fig. 2. Landsat image of the HPN (March 11, 2011), with names of outlet glaciers and GLOF lakes (Laguna Arco is located in front of Gl. Arco).

3. Data and Method 4. Results Grosse Reicher Gualas

Glacier variations of Hielo Patagónico Norte, Chile, over 70 years from 1945 to 2015 snouts, each with a small proglacial lake. Between 1945 and 1975, retreating rates were very small; 3 m a 1 and 8 m a 1 for the north and the south snout, respectively. After 1975 retreating rates picked up to ca. 20 to 30 m a 1 until 1996. Then, between 1996 and 1999 the south snout advanced ca. 310 m, while the north snout retreated ca. 280 m. What happened to the south snout next was snout disintegration with a retreat of 700 m in one year by 2000. A change at the south snout between 2000 and 2002 was an advance of ca. 450 m, while the north snout retreated ca. 550 m. After 2002 retreats of both the snouts were extensive by snout disintegration at each and by 2004, proglacial lakes of the two snouts coalesced into one, thereby resulting in one snout. The retreat in this twoyear period was ca. 1400 m. Shortly after, the proglacial lake became drained only by the north outlet stream. These two episodes of an advance and an ensuing huge retreat in the following year were the results of an apparent advance due to crevasse stretching and subsequent snout disintegration. Interestingly after 2004, retreats became very slow with only ca. 200 m until 2013. Particularly the retreat between 2009 and 2013 was nil, suggesting that the glacier had reached equilibrium. During this period, we could see bedrock at some parts of the snout, indicating that the glacier was leaving the proglacial lake. Then between 2013 and 2015 this glacier advanced ca. 80 m. San Rafael (Fig. 4): The total retreat in the 70 years was ca. 4150 m (59 m a 1). Although Heusser (1960) reported an advance in 1958/59 from field observation, the overall trend was retreat until 1975 at slow rates of 13 30 m a 1. Retreat accelerated to 200 m a 1 for 1975 86 and during 1986 91 the retreating rate was one of the fastest in the HP at up to 300 m a 1. Then all of sudden, the right part advanced ca. 50 m while the left part retreated 60 m by 1994, with a small net gain of area. After stagnation during 1994 96, it advanced again ca. 320 m by 1999. Between 1999 and 2002 it retreated fast at rates 150 450 m a 1. The period 2002 04 saw another advance of ca. 150 m. After 2004, this glacier seems to have stabilized, although during 2013 15 a small part of the right side advanced ca. 50 m. So this glacier has made advances, although small, four times since the 1990 s. San Quintin (Fig. 5): Today this big glacier terminates in a huge proglacial lake with a width of ca. 6 km. Variation measurements were taken at three sides,., front (west side), north and south sides. The total retreat in the 70 years at the front is up to 4500 m (64 m a 1), while those at the N- and S-sides are ca. 1500 m (21 m a 1) and ca. 3100 m (44 m a 1), respectively. The front part retreated slowly until 1991 at rates up to 45 m a 1, but thinning had been considerable. During 1991 99 all sides retreated and the front part detached completely from the damming moraine so that those proglacial lakes located on the northern, western and southern margins coalesced together and the snout became circumscribed by a huge proglacial lake. Snout disintegration of the front commenced in 1994 96 at the right part, and a large retreat started during 1999 2002, when it retreated at rates 250 500 m a 1. The period 2002 04 saw an accelerated retreat of ca. 2300 m in two years at the right part of the front. Large retreat continued during 2004 05 Fig. 4. Variations of Gl. San Rafael from 1945 (44/45) to 2011 (July 17). A, trimetrogon; B, Chilean IGM. An arrow indicates the same spot for comparison. The glacier was protruding into Laguna San Rafael until the beginning of the 1980 s. Since 2004 the snout has been more or less stationary. Note: in D (winter scene) the valley-confined portion (ca. 3 km wide) of the glacier is not covered with snow. Monte San Lorenzo (3598 m) is a separate mountain area located on the Argentinean border.

ANIYA Fig. 5. Variations of Gl. San Quintin from 1945 (44/45) to 2011 (July 17). A, trimetrogon; B, Chilean IGM. Arrows a to g indicate the same, corresponding spots. The width of the lake is ca. 6 km. On 1974 photograph, a proglacial lake can be recognized only at the north side (bottom of frame) of the snout. The front still abutted the damming moraines. Retreat between 1974 and 1986 was slow, but thinning was considerable. There were three large outlet streams (d, e and f) in 1974; but on D (2011), outlet stream at f was already dry and by 2014 outlet stream at e was also dry, due to lowering of the lake level. with ca. 1200 m and during 2005 07 with ca. 1250 m. Thus the snout disintegration lasted for about 13 years, because it occurred at different parts of the snout due to its large size. Then, retreat slowed down to rates 150 300 m a 1 until 2013. During 2013 15, the left part of the front advanced ca. 1000 m, while the right part retreated a maximum of 500 m, with a net loss of area, though. Benito: retreated slowly and steadily from 1945 to 2004 at rates 17 100 m a 1. After 2004 the snout retreated at much faster rates of more than 200 m a 1 until 2011, though with stagnation at 2007 08. The period 2011 12 saw a maximum retreat of 500 m, which was followed by 250 m retreat at all across the front during 2012 13. This rapid retreat continued to 2015 with a rate of 125 m a 1. HPN 1: was land terminating glacier until around 2007 when a proglacial lake, though small and shallow without damming terminal moraine, was formed, in which icebergs have rarely been observed. The total retreat in the 70 years is 3700 m (53 m a 1). Between 1945 and 1986, it slowly retreated at rates 27 47 m a 1 for a total of 1700 m. During 1986 91, it most actively retreated with rates ranging from 80 to 220 m a 1 at different parts of the snout. The period 1991 99 saw slow retreat with rates up to 17 m a 1. The following period 1999 2004 was moderate retreat at 50 100 m a 1. Then for 2004 09 it became nearly stationary with only 100 m retreat in 5 years. After 2009, it retreated with rates 50 175 m a 1 at the center part of snout until 2015. HPN 2: From 1945 to 2002, retreat was modest with rates up to 88 m a 1. Retreat rates were accelerated after 2002, to a maximum of 200 m a 1 for 2002 04, 300 m a 1 or 2004 05, and 375 m a 1 for 2005 07. From 2007 to 2015, retreat rates were 100 125 m a 1 with a stagnant period 2008 09. HPN 3: retreated rather slowly until 2000 at rates of 20 83 m a 1. Between 2000 and 2002 this glacier had the first snout disintegration with ca. 1150 m retreat. In the following period of 2002 04, it advanced a maximum of ca. 150 m. Then, it retreated ca. 200 m by 2005. During 2005 12 retreats were moderate with rates of 25 150 m a 1, at different parts of the snout. Between 2012 and 2013, this glacier underwent the second snout disintegration, thereby losing 1250 m. After 2013, it again advanced ca. 200 m by 2015. The total retreat in the 70 years is 3900 m. It is very interesting to note that this glacier made an advance after snout disintegration. Steffen (Fig. 6): Variation data for this glacier were taken for the front (south) and the east side of the snout until 2011, when it became unnecessary to take data for the E-side due to the front retreat. The front retreated a total of 5100 m (75 m a 1) in the 70 years, which is the second largest in the HPN, next only to Gl. Reicher, while the E-side retreated 1700 m (28 m a 1) in 66 years. During the 70 years, the front retreated at rates ranging from 23 m a 1 in the beginning to a maximum of 450 m a 1 after 2000. Meantime there were one period of stagnation in 1996 99 and two periods of advance; a maximum of 300 m in 2007 08 and ca. 200 m in 2012 13. Each advance was followed by a large retreat,., 400 m a 1 (2008 09) and 250 m a 1 (2013 15). Calving has been very active and many large icebergs are normally observed in the proglacial lake, implying that the snout is floating. Gl. Steffen have two side lakes on the right side,

Glacier variations of Hielo Patagónico Norte, Chile, over 70 years from 1945 to 2015 Fig. 6. Variations of Gl. Steffen. A, trimetrogon; B, Chilean IGM. Arrows a and b indicate the same spot. From 1945 to 1974, it retreated rather slowly, but after that retreat was rapid with very active calving, thereby producing many large icebergs. C is a winter scene and the lake was frozen with large icebergs. The width of the proglacial lake in 1974 photo is ca. 1.8 km. Laguna de los Témpanos (lower) and Laguna Bolson (upper). It was found out recently that Laguna de los Témpanos had had GLOFs (Glacial-Lake Outburst Flood) since at least the mid 1980 s (Aniya, 2014), which will be discussed later. Piscis: retreated between 1945 and 1975 at an average rate of ca. 25 m a 1 for a total of 760 m in 30 years. The rate slowed down until 1994, retreating only ca. 140 m. Then, this glacier advanced slightly ca. 30 m by 1996 at the right part. After 1996, this glacier had retreated ca. 450 m by 2008, with a stagnant period of 2004 05. Since 2008, it has been stagnant. Pared Sur: The snout area is now extensively covered with debris, which protects glacier ice underneath, resulting in very slow wasting. However, the snout had not been as extensively covered as today until around 1986 and it made rapid retreat during 1945 75 with 1000 m (33 m a 1). By 1986 retreat amounted to 1250 m. After 1986, however, there has been no discernible retreat, although thinning. Pared Norte: Since 1945, this glacier has been slowly retreating, with a few stagnant periods, at rates of ca. 25 80 m a 1 until 2013 for a total retreat of ca. 2250 m. For 2013 15, retreat accelerated to 250 m. Arco: used to be a calving glacier (Figs. 7 A and 7B) and during the 1950 s the proglacial lake, Laguna Arco, had a GLOF every year or two (Tanaka, 1980), which drained under the right side of Gl. Colonia. It had become land terminating by 1975 due to lowering of the lake level (Aniya and Enomoto, 1986). Its retreat is the least in the HPN in terms of the area with 0.46 km 2. Retreat was recognized in two periods of 1986 91 (70 m a 1) and 1994 96 (450 m a 1). This is probably because the snout is extensively covered with debris, and terminates in a narrow valley with high walls where sunshine is limited. Colonia (Fig. 7): Retreat had been slow and steady from 1945 to 2000 with rates of less than 10 m to 50 m a 1, with two periods of stagnation at 1994 96 and 1999 2000. After 2000, retreat started accelerating to 75 m a 1 to 2002, 125 m a 1 to 2004, 350 m a 1 to 2005, 100 m a 1 to 2007, and 350 m a 1 to 2008. The year 2008 saw the first GLOF of Lago Cachet Dos located on the left side of and dammed by this glacier, which will be discussed later. Eight GLOFs recognized by 2013 may have contributed to rapid retreats. Cachet: The snout used to terminate as one in the proglacial lake, Cachet Uno, until 2002 when the snout separated due to extensive retreat of ca. 2900 m since 1945. Therefore, after 2002 statistics were taken as Cachet north and Cachet west. Soon after separating, Cachet N became land terminating and retreated fast over a cliff until 2005. Then it has stagnated just above the cliff top. Cachet W retreated about 150 m until 2005 and then stagnated for 2005 09, before resuming slow retreat until 2013. Between 2013 and 2015, the center part of the snout advanced about 50 m. Nef: had a long snout in the proglacial lake in 1945 and in 30 years to 1975, the snout only narrowed by 400 700 m without retreat. Until 1991 main wasting of the snout was by narrowing of 400 600 m than front retreat of ca. 350 m. Then, during 1991 94, the snout advanced ca. 300 m with conspicuous bending in the proglacial lake, suggesting snout floating. This is probably the first time that the snout of a temperate glacier was suspected to be floating

ANIYA Fig. 7. Variations of Gl. Arco and Colonia. A, trimetrogon; B, Chilean IGM. An arrow indicates the same spot. Gl. Arco was not debriscovered in 1945 (44/45) and snout was emptying into the proglacial lake. In 1974, debris had accumulated at the snout and the snout was no longer emptying in the lake due to lowering of the lake level. In A, B and C, snout margin of Gl. Colonia is dark with debris that was brought to surface by thrusting, which actively continued until around 1998. Rapid retreat between 2011 and 2014, probably influenced by GLOFs of Lago Cachet Dos that started in 2008. By 2014, proglacial lakes on either side of the snout have coalesced, thereby becoming a large single lake. in Patagonia. Following this advance, the snout was found disintegrated between 1994 and 1996, thereby retreating 2800 m. This was the second snout disintegration caught in the HPN, after Gl. Reicher. Strong retreat continued until 1999 with a maximum of 1200 m. Rapid retreat slowed down considerably between 1999 and 2002, with a total of only 150 m. The period 2002 04 saw an advance of ca. 200 m, which was followed by slow yet steady retreat at rates 50 100 m a 1 until 2015. Soler (Figs. 8 and 9): The total retreat is ca. 850 m between 1945 and 2015 with an average rate of 12 m a 1, which is the third smallest in the HPN, next only to Gl. León and Exploradores. This glacier had been slowly, steadily retreating and a small proglacial lake was formed on either side of the snout tip around 1990. Due to continued slow retreat, during 2002 04 the snout tip broke up from the damming moraine, thereby two small proglacial lakes coalesced into one, resulting in 300 m retreat, which was by far the largest rate for this glacier. After 2004, it was more or less stagnant until 2009. Then, it started advancing, with a gain of 150 m by 2011, when many other HPN glaciers were retreating. With the continued advance it gained another ca. 50 m for 2011 12. While 2012 13 did not see any change, it again advanced ca. 80 m between 2013 and 2015, and the snout tip reached the damming moraine, thereby dividing the proglacial lake into two. These advances are interpreted to be the result of crevasse stretching, which is the phenomenon preceding the imminent snout disintegration. León: consists of three branches, León Norte, Central and Sur that coalesce just above the snout front in Lago Leones. Although the snout is without debris cover, this glacier has varied very little in the 70 years with retreat of only ca. 350 m. It retreated for 1945 94, but after 1994 it has been stationary more or less with occasional small advances at 2002 04 and 2013 15 and retreats. Fiero: is almost completely covered with debris from the southeast side of Monte San Valentin. It had retreated slowly between 1945 and 2011 in the proglacial lake at a rate of ca. 15 m a 1. However, after 2011, it suddenly

Exploradores 5. Discussion

ANIYA Fig. 9. Detailed variations of Gl. Soler using (semi) vertical photographs of the snout area from 1974 (Nov. 8) to 2014 (Dec. 2). A and D are taken by Chilean IGM, while the rest are mosaics of a few tens of aerial survey photographs. Since they are uncontrolled, scales vary greatly. The width of the glacier is ca. 2 km. G is a winter scene; although there was no snow at all (elevation 350 m), the proglacial lake was frozen. Note: On C, a line of thrust moraines (white strip parallel to the front shape) appeared. On D, a proglacial lake is recognized. On F, a proglacial lake is clearly recognized on either side of the snout. Gl. Soler started advancing in 2011 and the 2014 photo shows that protruding snout was split after reaching the damming moraine, indicating that advance was temporal and snout disintegration was imminent.

Glacier variations of Hielo Patagónico Norte, Chile, over 70 years from 1945 to 2015 Fig. 10. Variations of Gl. Exploradores. A and B are taken by Chilean IGM, while the rest are uncontrolled mosaics of aerial survey photographs. C and D are incomplete coverage. The width of the glacier is ca. 2.5 3 km. D is a winter scene, but without a trace of snow at all (elevation, less than 200 m). Although the front edge does not show conspicuous retreat, surface of the snout area has become more pitted, indicating thinning. Note: upstream of the debris-covered snout, there are many lumps of debris that are brought up to surface by thrusting. Since the HPN is located at the southern Westerly zone, the east west contrast in meteorology, climate and glacier conditions is notable (Nakajima, 1985; Warren and Sugden, 1992). Table 2 shows statistics of those glaciers located on the west side (Reicher to Steffen, 9 glaciers) and on the east side (Grosse to Piscis, 12 glaciers). The west side lost a total area of 96.85 km2 in the 70 years, which translates to 1.38 km2 a 1 or 0.154 km2 a 1 glacier 1, while the east side lost a total area of 29.88 km2, which translates to 0.43 km2 a 1 or 0.036 km2 a 1 glacier 1. Thus, the total area lost on the west side is more than three times that on the east side. It is argued that a larger glacier can lose more area than a smaller glacier and this was checked using the glacier surface area of 1974/75 (Aniya, 1988) as the base, because there is no data for 1944/45. Table 2 shows that the area loss per glacier surface ranges widely from 1.12 % of Gl. Arco to 13.12 % of Gl. Cachet, both located on the east side, and there is no definite relationship between the glacier surface area and the area lost. Large ones include Gl. Guals with 10.70 % on the northwest side, and Gl. Pared Sur with 7.13 % and Gl. Piscis with 7.08 %, both on the southeast corner of the HPN. The total glacier area of those on the west side is 2836.1 km2, while that on the east side is 1067.5 km2 (Table 2), making the area loss per glacier area 3.41 % and 2.80 %, respectively. So, also in terms of the area loss per glacier surface, there is substantial difference between those on the west and the east side. Since Gl. San Rafael and San Quintin are two exceptionally large ones in the HPN, statistics excluding these two from the west side are considered. Then, the total glacier area of seven glaciers on the west side becomes 1318.6 km 2 and the area lost by these seven glaciers is 43.52 km 2, which translates to 3.33 %, smaller than the inclusive figure. Individual figures of Gl. San Rafael and San Quintin are 1.68 % and 5.31 %, respectively; widely different figures despite the similar topoclimatic situations. For a comparison, Gl. Steffen, the third largest, has 2.42 % and Gl. Colonia, the fourth largest, has 1.35 %. Gl. Arco that lost the smallest area has 1.12 %. Accumulation-Area Ratio (AAR) also has no particular relationship with the area loss as shown in Table 2. Next, the north-south contrast was also examined. Considering the location of the accumulation area, those considered to be located on the north side include the following 10 glaciers: San Quintin, San Rafael, Reicher, Gualas, Grosse, Exploradores, Fiero, León, Soler and Nef, with a total glacier surface area of 2193.1 km2. Those on

Gl. Reicher:

Gl. Gualas: Gl. San Quintin Gl. Benito: Gl. HPN3: Gl. Steffen: Gl. Colonia: Gl. Nef

Gl. Reicher SW snout Gl. Gualas Gl. San Rafael Gl. HPN3

Gl. León Gl. Soler Gl. Nef Gl. Piscis Gl. Grosse: Gl. Fiero: Gl. Exploradores: Gl. Arco: Gl. Pared Sur:

ANIYA Fig. 12. Three GLOF lakes. A and B are Laguna del Co. Largo, head of Soler Valley; C E are Lago Cachet Dos on the left side of Gl. Colonia; and F and G are Laguna de los Témpanos on the right side of Gl. Steffen. Arrows in the same area indicate the same spot. In A, lake labeled no name had a GLOF probably in the 17th century. In B, strandline of old water level is ca. 97 m higher. In E, the distinctive strandline is clearly recognized and white ice cliff suggests a recent GLOF. In F, there are many stranded icebergs left after discharge near the far end of the lake. In G, lake was filled in with water again. SR retreated between 1975 and 1986 at a very rapid rate of 200 m a 1, thereby having lost an area of 4.83 km2 (0.44 km2 a 1), while the front of SQ retreated at a rate of less than 45 m a 1 with a loss of 0.87 km2 (0.08 km2 a 1). The behavior of SQ had changed after 1986, when it commenced a very rapid retreat that had lasted until around 2012 (see Fig. 3), during which time (26 years), it lost an area of 30.85 km2 (1.19 km2 a 1), by far the largest in the HPN. On the other hand, during 1986 91 SR accelerated the previous rapid rate to 300 m a 1; however, it abruptly stopped retreating after 1991. It even advanced four times in periods 1991 94, 1996 99, 2002 04 and 2013 15, although temporarily. Therefore, SR can be regarded as having been more or less stationary for 24 years since 1991, when the neighboring SQ had been retreating at a very fast rate of 1.25 km 2 a 1 until 2012. After 2012, retreat of SQ slowed down to a loss of 0.69 km2 to 2013, and in 2013 15 it all of sudden slowed down considerably to 0.01 km 2 a 1, because a part of the snout made an advance., Fig. 12 One of the severe consequences of global warming related to glaciers that threaten local inhabitants is GLOF. This is a phenomenon in which a lake associated with glacier drains a huge quantity of lake water in a short period of time. There are in general two types of GLOFs, depending upon the location of a lake relative to glacier; proglaciallake outburst and side-lake discharge. Proglacial-lake outburst mainly occurs when the damming moraine is breached, for example by excessive, strong overflow accompanying severe erosion. Another type, sudden side-lake discharge may occur when a glacial lake located on the side of a glacier drains underneath the blocking glacier body quickly. This typically occurs when the blocking glacier cannot withstand the hydrostatic pressure of the side lake, which is caused by either the increased volume of lake water or

Laguna Arco Laguna del Co. Largo Lago Cachet Dos Laguna de los Témpanos 6. Factors for Variations

7. Concluding Remarks Acknowledgments

References 20 10 31 18 24 18 10 15 29 29 4 50 31 59 13 75 A 25 78 A