sign in sign up
<<
Home , Ablation zone, Glaciofluvial deposits, Glaciology, Kame

K

KAME AND TOPOGRAPHY the form of low alluvial fans (if deposited on land) or deltas (if deposited in standing ).The glaciofluvial landform include , kame terrace, kettle, kettle holes, Amit Kumar outwash plains, etc. Department of , Centre of Advanced Study in Geology, Panjab University, Chandigarh, India Bibliography Benn, D. I., and Evans, D. J. A., 1998. and Glaciation. Definition : Arnold. A kame is a stratified geomorphologic feature which is Hambrey, M., and Alean, J., 2004. Glaciers, 2nd edn. Cambridge: created by action of , an irreg- Cambridge University Press. ularly shaped or mound composed of , gravel, and , commonly associated with end . A kame may Cross-references occur as an isolated hill but in general each kame is one Glaciofluvial mound in a low-lying terrain of many hummocks, terraces, Glaciogenic Deposits ridges, and hollows. are often associated with ket- tle holes. Kettles are depressions in the outwash plains, which formed due to the melting of large blocks and this is referred to as kame and kettle topography. Kame KATABATIC WIND: IN RELATION WITH SNOW and kettle topography is an indicator of a high-discharge AND GLACIERS supraglacial and englacial drainage system of a glacier in the final stages of melt, and large quantities of glacially Amit Kumar derived debris associated with meltwater. Centre of Advanced Study in Geology, Department of Geology, Panjab University, Chandigarh, India Origin Glacier meltwater that exists in the area of gla- Synonyms ciers during the melting season flows down on the glacier Drainage wind; Fall winds; Piteraq and williwaw surface, in the ice, or on the bedrock, making complex sys- tems of drainage channels. Deformation of ice, movement Definition of glacier, freezing, and melting processes influence on the Katabatic wind (Greek: katabaino – to go down) is the position and form of channels. Meltwater escapes through common name for downslope winds flowing from high numerous small and temporary streams. These streams elevations of mountains, plateaus, and glacier down their carry for longer distances and deposit them in slopes to the valleys or planes underneath. Such winds various forms. Sometimes, these streams also carry some are sometimes also called fall winds. Particularly, it is ice. Thus, the deposition of sediments after the ablation a wind that carries high density air from a higher elevation (melting of glacier) is called and down a slope under the force of gravity. This occurs on the the landforms resulting from such deposits are called largest scale as the outflowing winds from and glaciofluvial landforms. The sediments are deposited in . In Greenland these winds are called Piteraq

Vijay P. Singh, Pratap Singh & Umesh K. Haritashya (eds.), Encyclopedia of Snow, Ice and Glaciers, DOI 10.1007/978-90-481-2642-2, # Springer Science+Business Media B.V. 2011 672 KILIMANJARO and in South America as well as in Alaska, it is wind Geographic setting known as a Williwaw. Kilimanjaro is a massive, dormant volcano in Tanzania, Alpine valleys produce their own local wind systems as built up of both lava flows and pyroclastic material, situ- a result of thermal differences. The cold air slides down the ated roughly equidistant (300 km) south of the Equator slope under gravity during night. The radiative cooling of and west of the Indian Ocean. Three primary volcanic the ground surface under clear and calm conditions during centers are thought to have been active sequentially since night provides colder air near the surface. The nighttime the , which together form the Kilimanjaro downslope movement of the colder air is referred to as massif: Shira (4,005 m), Mawenzi (5,140 m), and Kibo katabatic winds. The anabatic wind is developed prior to (5,895 m). At the apex of Kibo is a relatively flat caldera the daytime, whereas katabatic drainage is developed in measuring 1.9 by 2.7 km (Figure 1); Uhuru Peak is the the night. The katabatic winds usually flow gently down- highest point along the southern scarp, 180 m above slope with low speed, but greater speeds are also experi- the caldera floor. enced when the depth of cold air is large and the slope is higher. History of cryospheric research on Kilimanjaro Bibliography The earliest scientific discussion of snow and ice on Kili- “ ” Benn, D. I., and Evans, D. J. A., 1998. Glaciers and Glaciation. manjaro began with the initial European discovery of London: Arnold. the snowcap by Johannes Rebmann in 1848. English Bennett, M. R., and Glasser, N. F., 1996. Glacial Geology: Ice were incredulous, and dismissed Rebmann’s Sheets and Landforms. Chichester: Wiley. report for more than a decade (Meyer, 1891). Hans Meyer Colbeck, S., 1980. Dynamics of Snow and Ice Masses. London: climbed nearly to the crater rim in 1887, reaching the sum- Academic. mit 2 years later on 6 October 1889 (Meyer, 1891). Addi- Defant, F., 1949. Zur Theorie der Hangwinde, nebst Bemerkungen zur Theorie der Berg- und Talwinde. (On the theory of slope tional European soon reached the area winds, along with remarks on the theory of mountain and and published their qualitative findings. Logistical con- winds.). Archiv für Meteorologie Geophysik und straints rendered ascents and fieldwork considerably more Bioklimatologie, A1, 421–450. difficult than at present, yet virtually every account Singh, P., and Singh, V. P., 2001. Snow and Glacier . describes features and processes not unlike those of today. Dordrecht: Kluwer. Most also discuss the decreasing extent of ice, and many predict disappearance – within decades – of either indi- vidual glaciers or all of the mountain’s ice. KILIMANJARO Mid-twentieth-century perspectives on Kibo’s summit and slope glaciers were provided by Humphries (1959), Douglas R. Hardy Downie and Wilkinson (1972), and Hastenrath (1984). Climate System Research Center and Department Henry Osmaston (1989) then published an analysis of gla- of Geosciences, University of Massachusetts, Morrill cier Moraine (qv) as mapped from aerial photographs, Science Center, Amherst, MA, USA which for the first time quantified the nineteenth century extent of glaciers on the mountain. Hastenrath and Greischar (1997) built upon Osmaston’s work and pro- Definition vided the first cartographic documentation of ice reces- Kilimanjaro is Africa’s highest mountain (5,895 m), sion. Thereafter, a resurgence of research on Kibo began located in northern Tanzania just south of the Kenya bor- 0 0 in February 2000 with (qv) drilling, aerial pho- der (3 4 S; 37 21 E). At the seasonally snow-covered tography, and installation of an automated weather station 2 summit, the extent of glacier ice is now less than 2 km , (AWS) on the Northern Icefield (Hardy, 2002; Thompson roughly half of that remaining on the continent. et al., 2002). A Network of Stakes (qv) has expanded steadily since 2000 to represent most of the glacierized Overview area at the summit, and two additional AWS are now oper- The is sparsely represented in Africa, primar- ating on summit glaciers (Mölg et al., 2008). ily on a small handful of the continent’s highest moun- Today, as during the nineteenth century, snow and ice tains. Among these is Kilimanjaro, the “white roof of on Kilimanjaro are again controversial. A new ice-extent Africa,” whose glaciers have achieved notoriety far out map released in 2001 was accompanied by a prediction of proportion to their size (miniscule), importance as that the glaciers could disappear within 20 years (Irion, a water resource (negligible), or potential contribution to 2001). Kilimanjaro was quickly employed to symbolize sea-level rise (zero). Yet, Kilimanjaro’s summit mantle the impacts of global warming (e.g., Greenpeace, 2001). of Snow (qv) and Ice (qv) is starkly beautiful, and thus However, cautious statements by scientists such as Kaser among the mountain’s most fascinating, distinctive, and et al. (2004, p. 337) that “...mass loss on the summit ... best-known attributes. Thousands of international visitors is little affected by air ,” or Mote and Kaser are attracted annually, bringing valuable tourism revenue (2007,p.325)that“... loss of ice on Mount Kilimanjaro to Tanzania. cannot be used as proof of global warming,” were eagerly KILIMANJARO 673

Kilimanjaro, Figure 1 Kibo peak of Kilimanjaro, with remnants of the (qv) that once encircled the summit. The crater is the area surrounded by ice and labeled “KIBO.” Contours are in meters. Solid circle symbols indicate location of 2000 ice-core drilling sites (Thompson et al., 2002), and the ice extent is shown for five epochs (1912–1989 after Hastenrath and Greischar (1997), 2000 after Thompsen et al. (2002)). NIF, EIF, and SIF are the former Northern, Eastern, and Southern Ice Fields (respectively), FWG Furtwa¨ngler Glacier, UP Uhuru Peak (5,895 m), and LPG Little Penck Glacier. Automated weather stations currently operate near the NIF and SIF drill sites. embraced by those seeking to cast doubts about global (Coutts, 1969; Hemp, 2006). This precipitation pattern warming (e.g., GES, 2004). Resolution of the modern-time accounts in part for the asymmetrical distribution of gla- controversy awaits a comprehensive understanding of how ciers on Kilimanjaro. Kilimanjaro’s summit climate has been impacted by large- Precipitation at the summit annually totals only 10% scale atmospheric circulation changes; this effort is well of that received by the forest below, and snow is the pre- underway (Mölg et al., 2009;Thompsonetal.,2009; dominant form of precipitation at elevations above the Winkler et al., 2010). mean annual freezing-level altitude, roughly 4,700 m (Hastenrath, 1984). Snowfall can occur at any time of year, but is primarily associated with northern Tanzania’s Climate two seasonally-wet periods, the November–December Kilimanjaro rises 5,000 m above the surrounding plains, “short rains,” and the “long rains” of March to May. extending halfway through the tropical atmosphere Summit climate is thus best defined by seasonal humidity to 506 hPa. Climate varies dramatically and sharply with fluctuations, and by strong diurnal cycles driven largely elevation, causing the mountain’s dramatic ecological by the tremendous daily fluctuation in incoming solar zonation. Whereas air temperature drops steadily and uni- radiation; the following synopsis is based on AWS mea- formly with elevation, the annual precipitation amount surements made on the Northern (NIF) since increases and then decreases with elevation. Southern 2000 (Hardy, in prep.). and southwestern slopes reach a maximum annual total Summit climate is most stable through an extended dry at 2,200 m (Hemp, 2006), but northern slopes are drier season centered on July and August. This interval is 674 KILIMANJARO characterized by annual minima of humidity, snowfall, air (see Hastenrath, 1984 for list). These names remain in use, temperature, and wind speed, and by increasing solar irra- despite morphological changes. On the south-facing slope diance after the solstice minima. below the crater’s sharp southern rim, a Southern Ice Field By the middle of September, the beginning of an impor- has encompassed what are now or will soon be separate tant seasonal change is marked by rapidly increasing wind entities, the Heim, Kersten, Decken, and Rebmann speed, and an increase in air temperature. Solar irradiance Glaciers. Within the crater, the Furtwängler Glacier has gradually reaches an annual maximum as the sun moves been the only ice entity for several decades, splitting into into the southern hemisphere. Humidity increases slowly two parts by 2007. Straddling the crater rim’s north side, during September, rapidly into October, and continues the Northern Ice Field (NIF) comprised slightly more than climbing steadily into December. The “short rains” typi- half of Kibo’s total ice area until the 1970s, when an East- cally begin in mid-October, although the timing and mag- ern Ice Field (EIF) became distinct. Although the EIF has nitude vary from year to year – with important since broken into numerous entities, the NIF remains implications for . For example, Kibo’s largest body of ice. Extending down north- and September and October (and even November) can be northwest-facing slopes below the NIF are what remains a time of considerable ablation (e.g., Hardy, 2003), when of the Credner, Drygalski, and Great Penck Glaciers. net solar irradiance remains high in the absence of snow- The Little Penck Glacier, prominently visible today from fall, accompanied by increasing turbulent energy transfer western ascent routes, separated from the NIF during the as wind speed, air temperature, and humidity are 1990s (Figure 1) and has considerably decreased in area increasing. since 2000. By mid-January the short rains are usually ending, with All ice masses on Kibo are here termed glaciers for dis- both humidity and snowfall dropping quickly to early cussion purposes, although most ice is now static and February minima and a brief secondary dry period. The some entities are just tiny fragments. Ice thickness is long rains then typically begin in early March, with poorly known due to limited measurements, with humidity and snowfall continue increasing into April, a probable maximum of 50 m for the Northern Ice Field. and snow accumulation typically continues into May. Little was known of energy and mass balance details on Then, air temperature decreases rapidly, wind direction Kibo’s glaciers until 2000, other than that the total ice area backs slightly from east toward north, and humidity drops had been decreasing for over a century. However, early drastically. By June, another dry season is beginning. observations were astute; Geilinger (1936, p. 9), for exam- Mean annual temperature at the summit is approxi- ple, recognized that glaciers on the outer slopes and those mately –7C, with monthly means ranging only 1.3C. of the crater behaved differently “... with regard to the On the NIF, daily temperature ranges between an average melting process.” Building upon this distinction, Kaser low of 9 and an average maximum of 4C. Thus, air et al. (2004) subdivided summit glaciers into (1) horizontal temperature on the glaciers is consistently below freezing; glacier surfaces, the typical glacier regime studied; extreme radiational cooling at night brings surfaces tem- (2) slope glaciers; (3) near-vertical margins; and (4) basal peratures below 15C (to 27), so a considerable cold surfaces. content must be overcome each day before ablation of the snow or ice surface can occur. Horizontal glacier surfaces Most glaciers on Kibo’s summit have horizontal or nearly Kilimanjaro glaciers horizontal upper surfaces, unbroken by . At Currently, there are roughly eight glaciers on Kibo, some times the surfaces are flat and smooth, but this changes distinct and some in clusters as formerly larger bodies from year to year. On a whole-glacier scale, the surfaces break up (Figure 1). These are all remnants of a once ten- are often comprised of “massive steplike features” fold larger ice cap (Ice Caps qv) that encircled the volca- (Gillman, 1923, p. 17); at smaller scales one sometimes no’s summit in the mid-nineteenth century, filling at sees “fantastic ice-shapes” (Sampson, 1965, p. 121) or least portions of the crater itself while also spilling out- ice “honeycombed in many places to a depth of over 6 ward and down the slopes. The areal extent of this earlier ft, and weathered into countless grooves and ruts and ice cap – likely the maximum Holocene (Holocene Gla- pointed spikes” (Meyer, 1891, p. 147; see also Figure 2). cier Fluctuations qv) extent – was 20 km2, as determined Since 2000 this regime has been the focus of measure- by mapping (Osmaston, 1989). Subsequent ments and modeling (e.g., Mölg and Hardy, 2004; Mölg ice recession through the twentieth century was “dramatic et al., 2008), and is today the best known of the four and monotonic” (Hastenrath and Greischar, 1997, p. 459), regimes. Mass balance here governs glacier thickness, based on four area determinations averaging 26 years and balance fluctuations are largely independent of ice apart and with unknown errors; updated maps show con- area, in the absence of flow. Although details of mass tinuing retreat (Figure 1; Thompson et al., 2002; Cullen exchange at this surface remained unknown until 2000, et al., 2006; Thompson et al., 2009). historical accounts suggest that twentieth century ice By the beginning of the twentieth century, many of the thickness probably never averaged more than about twice ice cap’s broad lobes had been named after early explorers that of today. KILIMANJARO 675

Kilimanjaro, Figure 2 Three views of Kibo summit glaciers, taken: (a) September 28, 2008, (b) January 30, 2009, and (c) October 7, 2007. Note fluting of vertical walls (a, b) and horizontal-surface penitentes (bottom of image a, c). For scale, note people in image (b) and ablation stake at ice surface in image (c) (0.5 m visible; upper right-hand corner of image).

Horizontal and slope glacier surfaces on Kibo are typi- characteristics and processes (Figure 3). Accordingly, cally comprised of hard glacier ice, appearing and behaving simple generalizations based on point-scale modeling as expected for ablation-zone ice of density approaching (e.g., Mote and Kaser, 2007) are questionable. 900 kg/m3. A thin mantle of Seasonal Snow Cover (qv) One important suggestion of recent field measurements blankets the ice during most wet seasons. Snow – or snow is that the rate of mass loss from horizontal surfaces may transitioning to superimposed ice – persists from one wet be increasing (Thompson et al., 2009), even without season to the next when snowfall is above average. a demonstrable change in forcing. Continuing ablation is In the current climate at Kibo’s summit, the horizontal concentrating particulate matter (i.e., dust) on an increas- surface, multiyear mass balance is negative – and the gla- ingly older surface, decreasing albedo and thus increasing ciers are thinning (Figure 3). Indeed, the hypothetical ELA net radiation receipt and accelerating ablation. Indeed, in has probably been above summit-level for some time the vicinity of the NIF weather station and Ice Core (qv) (Humphries, 1959; Kaser and Osmaston, 2002). Nonethe- sites, the glacier surface is almost certainly older than less, speculation continues as to whether the ELA might 57 years, a figure based on 36Cl measurements in the ice be lower on the south slopes (e.g., Kaser and Osmaston, cores (Thompson et al., 2002). More recent 14C-dating 2002; Mölg et al., 2009), despite little evidence. of organic material indicates that the surface could be over Mass loss on horizontal surfaces is resulting from 300 years old (D. Hardy, 2004, 2009), although it is a predominance of negative energy balance, due to net unlikely that the surface age is uniform, given the consid- radiation receipt as controlled by Albedo (qv; Mölg and erable spatial variability in accumulation and ablation Hardy, 2004; Mölg et al., 2008). Field measurements of patterns (e.g., Figure 3). mass balance support these modeling results. Mass is Understanding horizontal-surface processes and mass being lost from surfaces through both sublimation (Subli- balance is important, because these control what is mation from Snow and Ice qv) and melting (Melting Pro- revealed by ice cores penetrating through layers of accu- cesses qv); for a 1-year period March 2001 through mulation, and previously exposed ablation surfaces. Six February 2002, sublimation accounted for 86% of mass cores were drilled on Kibo through three different ice enti- loss at the NIF AWS site. Yet, some areas clearly experi- ties (see Figure 1). In the absence of reliable 14C dates ence a higher proportion of melting, as a high degree of from the core, the resulting chronology was developed – both spatial and temporal variability exists in surface necessarily – by assuming an invariant accumulation rate 676 KILIMANJARO

0

−1

−2

−3

−4

−5 NIF average NIF stake minimum FWG average FWG stake minimum −6 SIF average SIF stake minimum

ICE surface height – relative to datum of 25 February 2000 (m) height – relative ICE surface −7 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Kilimanjaro, Figure 3 Ice surface height change on three Kibo summit glaciers, from measurements at mass balance stakes. In general, only one stake per glacier was monitored through 2004, and 3–19 stakes each subsequently; average values shown. Triangles symbolize minimum height change at individual stakes that ablated out and fell over between measurements (values were used in averages). Punctuating the overall thinning trend are intervals of positive mass balance due to superimposed ice formation. Resolution of height fluctuations between these discrete stake measurements is not possible.

(Thompson et al., 2002). This record, however uncertain To an even greater extent than for horizontal ice sur- (e.g., Gasse, 2002; Kaser et al., 2010), will endure until faces at the summit, some slope glacier surfaces are today an alternative chronology is provided by better dating. quite dirty, due to deposition of wind-blown dust (e.g., Figure 2c). This inorganic matter is being concentrated, due to a prolonged period of negative mass balance, and Slope glaciers is accelerating mass loss by changing the energy balance. Glaciers currently extend down from Kibo’s crater rim in Slope glacier meteorological measurements begun in only a few cases (Figure 1). These include the Kersten 2005 have allowed energy balance modeling for these sur- and Decken Glaciers on the south side, along with the faces. To date, energy and mass balance characteristics Rebmann Glacier to a minor extent, all inclined at appear similar to those on the NIF horizontal surface, with 30–40 and all remnants of the former Southern Ice Field. even a slightly higher proportion of mass loss to sublima- The Credner Glacier still extends down the northwest tion (Mölg et al., 2008). flank. Additional glaciers and small ice bodies remain on the slopes but are no longer connected to ice at the crater rim. These include the Heim and Great Barranco Glaciers Near-vertical ice margins on the southwest flank, and the Little Penck remnant on Vertical or near-vertical margins are a special characteris- the west side. Even by the mid-1930s, Geilinger (1936, tic of Kilimanjaro’s summit glaciers (Figure 2; Winkler p. 12) referred to such ice as “... independent dead gla- et al., 2010). Typically, the vertical surfaces are fluted, ciers of the outer slopes.” and “sometimes slightly undercut at the base” (Downie At their late-nineteenth century maximal extent, and Wilkinson, 1972, p. 40). Such margins have been revealed by moraines, slope glaciers extended down as reported since the earliest observations (e.g., Meyer, low as 4,400 m (Osmaston, 1989), and as recently as 1891; Gillman, 1923). 1971 there were seven tongues reaching below 4,870 m Ablation at the vertical margins is driving the areal (Messerli, 1980). Today, glacier ice is scarce below recession revealed by mapping (e.g., Hastenrath and 5,000 m, likely with an increasing proportion buried under Greischar, 1997). Although ice retreat is probably the a mantle of debris (Humphries, 1959; Downie and best-documented environmental change occurring high Wilkinson, 1972; Young and Hastenrath, 1991; Kaser on the mountain, our understanding of vertical-wall abla- et al., 2004). tion processes remains incomplete, despite the important KILIMANJARO 677 contribution of Mölg et al. (2003) confirming early specu- Few early explorers or mountaineers made snow depth lation that retreat is governed by energy from direct solar measurements, but a seasonal absence of snow has been radiation (e.g., Geilinger, 1936). Melting is the primary noted since the first observations. For example, Meyer mechanism by which vertical walls retreat (lose mass), (1891, p. 316) remarked that in October “...when all the and measurements confirm that the ice temperature often snowfields had disappeared, there was likewise compara- reaches 0C during the day when seasonally irradiated tively little snow to be met with on the ice-cap.” Sampson (Winkler et al., 2010). In addition, collapse features (1965, p. 123) wrote that from August to October “the (calving) can be observed around most summit glacier chances of finding any snow patches... are very poor at margins, and must also be considered a mechanism of heights over 4,000 m.” Additional evidence for the sea- vertical margin ablation, as speculated by Downie and sonal absence of snow in the past comes from historical Wilkinson (1972). photographs, which cannot easily quantify snow depth Vertical ice exposures on Kibo nicely illustrate stratifi- but nicely document times when snow is absent. cation, or the sedimentary banding associated with accu- Snowfall events can be brief, such as those associated mulation over time, and these can often be traced with afternoon convection (e.g., graupel), or can persist laterally for considerable distances. Notably, especially for multiple days. Wet-season events are typically of from the perspective of ice-core interpretation, some longer duration and yield the greatest accumulation. bands appear to represent buried surfaces that suggest Within the past decade, seasonal snow accumulation on a break in snow accumulation (i.e., missing time inter- the glaciers of 1 m has been noted at least twice (early vals), or at which “...there appears to have been marked in 2001, 2007; D. Hardy, 2004, 2009), with lesser accumu- of the ice surface before further accumulations of lations on unglacierized portions of the crater. Anecdotal snow” (Humphries, 1959, p. 477). Such unconformities accounts from guides suggest that snowfall magnitude are sometimes marked by dirt bands, which Downie and has diminished in recent decades, although no reliable Wilkinson (1972, p. 42) described as locally common long-term station data for precipitation exist from above but “almost nonexistent” on a larger scale. Especially at 1,800 m on Kibo. The greatest documented accumula- the upper end of the southern glaciers, unconformities tion is that reported for the 1961–1962 combined wet sometimes illustrate an angular discordance with the over- seasons, involving “snowfall of over six feet,” much of lying , a phenomenon awaiting explanation. which “...was still present more than a year later” (Segal, 1965, p. 126). The extent to which this interval was Basal surfaces unusual remains unknown, as summit visitation is much less frequent during the wet seasons. Kibo glaciers rest primarily on volcanic sand, and to Once deposited, ablation of snow cover from glacier a lesser extent on bedrock. During February 2000 ice-core and crater surfaces involves both melting and sublimation, drilling (Thompson et al., 2002), the NIF basal tempera- based on observations and energy balance modeling ture was 0.4 C, and 0 C was measured within the SIF. (Mölg and Hardy, 2004; Kaser et al., 2004; Mölg et al., The NIF 10-m-depth temperature of 1.2 C suggests that 2008, 2009). Of these two processes, melting is the more meltwater is transporting heat energy into the glacier, readily observable, both on and off the glaciers. Although given the 7 C mean annual air temperature. sometimes difficult to distinguish from ice melt, evidence Little is known about the spatial variability and magni- for snowmelt includes , supraglacial meltwater tude of geothermal heat flux on Kilimanjaro, especially ponds, areas of thin, water-saturated snow, and rarely, relative to that beneath the glaciers. However, fumeroles meltwater runoff. Melting at the summit is not a recent are present within the inner crater, and even prior to their development, for Humphries (1959, p. 477) observed first observation by W.H. Tilman in 1933, Jäger (1909), “melting ice and pools of melt water at the summit during Gillman (1923), and Geilinger (1936) all speculated that the day.” Sublimation of snow is often indicated by the volcanic heat might be influencing ice recession. Local- presence of penitentes (Figure 2a, c), which form over scale glacier impacts, apparently due to steam venting, both ice and crater surfaces in the intense radiation envi- have been observed and reported (e.g., Kaser et al., 2004). ronment on Kibo (cf. Lliboutry, 1954). With fully devel- oped penitentes, melting ice or wet crater sand may be Kilimanjaro snow present between the spikes. Although present recently in Snow on Kibo is ephemeral, meaning that within sufficient density and height to impede glacier travel, a relatively short period of time (i.e., minutes to months) penitentes are transient features, today and in the past. it either ablates and disappears, or is transformed into gla- Gillman (1923, p. 18) for example, did “not come across cier ice. Here, the term snow refers exclusively to solid penitents” on his single trip to the summit, yet their pres- precipitation, or that which has accumulated as snow ence in 1929 was noted by Geilinger (1936). As with other cover on either glacier or crater surfaces. Although gla- aspects of the cryosphere on Kibo, spatial and temporal ciers are a perennial and long-term feature of Kilimanjaro, variability render generalizations difficult. snow cover is not – and the historical literature indicates Lastly, recent research highlights the extreme sensitiv- that snow has always come and gone at high frequency ity of horizontal glacier surfaces to the variability of snow- (annual). fall frequency and amount. By governing albedo 678 KILIMANJARO variability, and thus net receipt of solar radiation, snowfall Hardy, D. R., 2002. Eternal ice and snow? In Salkeld, A. (ed.), Kil- has emerged as the key atmospheric variable controlling imanjaro: To the Roof of Africa. Tampa: National Geographic – ablation and mass balance (Hardy, 2003; Mölg and Hardy, Society, pp. 224 225. Hardy, D. R., 2003. Kilimanjaro snow. In Waple, A. M., and 2004; Mölg et al., 2008, 2009). Figure 3 illustrates that the Lawrimore, J. H. (eds.), State of the Climate in 2002. Boston: system is especially sensitive to the interannual variability Bulletin of the American Meteorological Society, Vol. 84, of snowfall during the short rains (November–December). p. S48. Hastenrath, S., 1984. The Glaciers of Equatorial East Africa. Dordrecht: D. Reidel Publishing Company. Summary Hastenrath, S., and Greischar, L., 1997. Glacier recession on Kili- Currently, there are 8 distinct ice entities on Kibo, manjaro, East Africa, 1912–1989. Journal of , 43, – together covering a total area of less than 2 km2 and all 455 459. Hemp, A., 2006. Vegetation of Kilimanjaro: hidden endemics and remnants of a once larger ice cap. The summit glaciers missing bamboo. African Journal of , 44, 305–328. are relatively flat, with near-vertical margins, and the Humphries, D. W., 1959. Preliminary notes on the glaciology of slope glaciers today are concentrated on the mountain’s Kilimanjaro. Journal of Glaciology, 3, 475–479. southwest- and northwest-facing flanks. Glaciers on Kili- Irion, R., 2001. The melting snows of Kilimanjaro. Science, 291, manjaro are a product of climatic conditions at the summit 1690–1691. that no longer exist, as no area of accumulation has existed Jäger, F., 1909. Forschungen im den Hochregionen des Kilimandscharo. Mitteilungen aus den Deutschen Schutzgebieten, for many decades and perhaps since the current recession 22,113–197. began. Today, as in the past, snow cover on the mountain Kaser, G., and Osmaston, H., 2002. Tropical Glaciers. International is seasonal and subject to considerable interannual vari- Hydrology Series. Cambridge, UK: UNESCO/Cambridge Uni- ability. Measurements and modeling in recent years have versity Press, 207p. demonstrated that the mass and energy balances on hori- Kaser, G., Hardy, D. R., Mölg, T., Bradley, R. S., and Hyera, T. M., zontal ice surfaces are very sensitive to the magnitude 2004. Modern glacier retreat on Kilimanjaro as evidence of cli- and frequency of snowfall events – perhaps increasingly mate change: observations and facts. International Journal of , 24, 329–339, doi:10.1002/joc.1008. so as dirt concentration increases on exposed, ablating Kaser, G., Mölg, T., Cullen, N. J., Hardy, D. R., and Winkler, M., ice surfaces. 2010. Is the decline of ice on Kilimanjaro unprecedented in the Holocene? The Holocene, 20, 1079–1091, doi:10.1177/ 0959683610369498. Acknowledgments Lliboutry, L., 1954. The origin of penitentes. Journal of Glaciology, 2, 331–338. This material is based upon work supported by the Messerli, B., 1980. Mountain glaciers in the Mediterranean and in National Science Foundation (NSF), and NOAA Office Africa. In Proceedings of the Riederalp Workshop, September of Global Programs, Climate Change Data and Detection 1978. IAHS-AISH Publ. no. 126, pp. 197–211. Program, under Grant No. 0402557, and NSF ATM- Meyer, H., 1891. Across East African glaciers. London: George 9909201 (Paleoclimate Program) to the University of Philip & Son. Massachusetts. Additional support was provided by Mölg, T., and Hardy, D. R., 2004. Ablation and associated energy balance on a horizontal glacier surface on Kilimanjaro. Journal NOAA U.S. Global Climate Observing System. of Geophysical Research-Atmospheres, 109, D16104, doi:10.1029/2003JD004338. Mölg, T., Hardy, D. R., and Kaser, G., 2003. Solar-radiation- Bibliography maintained glacier recession on Kilimanjaro drawn from Coutts, H. H., 1969. Rainfall in the Kilimanjaro area. Weather, 24, combined ice-radiation geometry modeling. Journal of Geo- 66–69. physical Research-Atmospheres, 108, 4731, doi:10.1029/ Cullen, N. J., Mölg, T., Kaser, G., Hussein, K., Steffen, K., and 2003JD003546. Hardy, D. R., 2006. Kilimanjaro glaciers: recent areal extent Mölg, T., Cullen, N. J., Hardy, D. R., Kaser, G., and Klok, L., 2008. from satellite data and new interpretation of observed 20th cen- Mass balance of a slope glacier on Kilimanjaro and its sensitivity tury retreat rates. Geophysical Research Letters, 33, L16502, to climate. International Journal of Climatology, 28, 881–892, doi:10.1029/2006GL027084. doi:10.1002/joc.1589. Downie, C., and Wilkinson, P., 1972. The Geology of Kilimanjaro. Mölg, T., Cullen, N. J., Hardy, D. R., Winkler, M., and Kaser, G., Sheffield: Geology Deparment, University of Sheffield. 2009. Quantifying climate change in the tropical mid tropo- Gasse, F., 2002. Kilimanjaro’s secrets revealed. Science, 298, sphere over East Africa from glacier shrinkage on Kilimanjaro. 548–549. Journal of Climate, 22, 4162–4181. Geilinger, W., 1936. The retreat of the Kilimanjaro glaciers. Mote, P. W., and Kaser, G., 2007. The shrinking glaciers of Kiliman- Tanganyika Notes and Records, 2,7–20. jaro: can global warming be blamed? American , 95, Gillman, C., 1923. An ascent of Kilimanjaro. Geographical 318–325. Journal, 61,1–27. Osmaston, H., 1989. Glaciers glaciations and equilibrium line alti- Greening Earth Society, 2004. Snow fooling. Newsletter of 7 March tudes on Kilimanjaro. In Mahaney, W. C. (ed.), 2004. Available from http://www.worldclimatereport.com/ and Environmental Research on East African Mountains. Rotter- index.php/2004/03/07/snow-fooling/, accessed 27 April 2009. dam: Balkema, pp. 7–30. Greenpeace, 2001. Press release: “Kilimanjaro set to lose its ice field Sampson, D. N., 1965. The geology, , and glaciology of by 2015 due to climate change.” Availbale from http://archive. Kilimanjaro. Tanganyika Notes and Records, 64,118–124. greenpeace.org/pressreleases/climate/2001nov6.html, accessed Segal, D., 1965. Expedition ramonage – ten days in the Crater of 27 April 2009. Kibo. Tanganyika Notes and Records, 64, 125–130. KUNLUN MOUNTAINS 679

Thompson, L. G., Mosley-Thompson, E., Davis, M. E., Henderson, accumulation and meltwater, evapotranspiration, K. A., Brecher, H., Zagorodnov, V. S., Lin, P.-N., Mashiotta, T., streamflow, and its recharge to groundwater as well Mikhalenko, V. N., Hardy, D. R., and Beer, J., 2002. Kilimanjaro (IPCC, 2007). ice core records: evidence of Holocene climate change in tropical Africa. Science, 298, 589–593. The landform of the southern Tarim basin is character- Thompson, L. G., Brecher, H. H., Mosley-Thompson, E., Hardy, D. R., ized by the three huge mountains: the Pamirs/China, the and Mark, B. G., 2009. Glacier Loss on Kilimanjaro Continues West Kunlun, and Karakoram Mountains; all rivers origi- Unabated. Proceedings of the National Academy of Science, nate from the mountains and drain into their basins 106, 19770–19775, doi:10.1073/pnas.0906029106. (Figure 1). Typical snow and glacier-fed rivers, the Hotan Winkler, M., Kaser, G., Cullen, N. J., Mölg, T., Hardy, D. R., and and Keriya, are from the glaciated center at the largest Tad Pfeffer, W., 2010. Land-based marginal ice cliffs: focus on Kil- imanjaro. Erdkunde, 64, 179–193, 10.3112/erdkunde.2010.02.05. icecap, and Guliya from the West Kunlun flow into the Young, J. A. T., and Hastenrath, S., 1991. Glaciers of Africa. Satel- Tarim basin. lite image atlas of glaciers of the worlds: glaciers of the Middle The land surface on the upstream of the watersheds is East and Africa. US Professional Paper characterized by snow and glacier cover and a little grass- 1386-G-3. US Geological Survey, : DC, land, and on the downstream by Gobi and desert where pp G49–G70. rainfall seldom produces runoff due to strong evaporation except for storm rainfall. Since their upper streams are Cross-references covered by a great number of glaciers and snowpacks Albedo where there is much precipitation in summer, stream flow Holocene Glacier Fluctuations occurs also in the summer (June–September), and there is Ice a close linear relationship between summer flow and air Ice Caps temperature, as the former contributes 70–90% of stream Ice Core Melting Processes water annually. With the rise in temperature and increase Moraine in rainfall, the maximum monthly runoff in July contrib- Sublimation from Snow and Ice utes about 30% of the annual runoff, and the minimum in February or March contributes only 1.2–1.8%. Eighty percent of their drainage areas are over an altitude of 3,000 m; it means streamflow is closed to the exchange KUNLUN MOUNTAINS between heat and water within the glacier system in the high mountains. Meanwhile, due to hot weather and rainy Jingshi Liu season in July and August annually, their combined role Institute of Research, Chinese Academy causes annual maximum floods. According to hydrologi- of Sciences, Haidian district, Beijing, China cal records, more than 90% of the maximum floods occur during the period of late July to early August. Definition The Hotan River basin is the largest basin in the west Tibet Plateau and includes parts of the Karakoram and Kunlun Mountains extend about 2,500 km from the the south territories of the Tarim basin in the west Pamirs/China in the west to the Mount Yuzhu in Sichuan, China. The basin covers an area of approximately China, in the east, and have an average elevation of 5,000 m. 34,558 km2, which encompasses a wide variety of cli- Their highest mount, Muztag, is 7,723 m. In this region matic conditions, including the periglacial, the alpine there are 6,580 glaciers with a total area of 10,844 km2, 3 permafrost, and desert zones. Glaciers, snow cover, and an ice volume of 1,175 km , which corresponds and patterned ground features associated with continu- to about 72% of the glacier area in the whole Kunlun (Liu, ous and discontinuous permafrost are found in the 2000). The West Kunlun Mountains lies in the western part – south, while agriculture and stock raising are important of the Tibetan Plateau; the largest glaciers mostly 20 30 km economic activities in the northern part. Figure 1 shows long are distributed in the high mounts between Tianshuihai the features of permafrost regions in west China and Pass and the southern Yurunkax River basin; and the alpine in the Hotan River basin. The drainage area is located type glaciers are usually developed on the West Kunlun in the seasonal frozen ground and alpine permafrost where the lands are severely cut and highly shadowed. Ice region. The elevation ranges from 1,860 to 7,167 m with cap (flat-topped glacier), slope glaciers, and the valley gla- an average altitude of 4,200 m estimated by the DEM, ciers are mainly developed on the south slope due to advan- and the low limit of alpine permafrost is about 4,400 m, tageous topographical and climatic conditions. thus approximately 45% of the basin lies within the continuous and discontinuous mountain permafrost zones (Zhou et al., 2000). Introduction The Hotan River consists of two large subbasins: the Researchers have estimated increasing greenhouse gases Yurungkax in the east and the Karakax in west. There which can cause change in regional temperature and pre- are a great number of glaciers in the upstream above cipitation. The change in climate may have affected the 4,800 m, with a total area of 5,127.15 km2 (Yang and hydrological cycle, such as precipitation form, snow An, 1990). Hydrometeorological records indicate a simple 680 KUNLUN MOUNTAINS

65° 70° 75° 80° 85° 90° 95° 100° 105° 110° 115° 120° 125° 130° 135° 140° 145°

45°

45°

40°

40°

35°

35°

30°

30°

25°

Permafrost seasonally frozen ground permafrost limit frozen ground limit

0 180 360 540 km

90° 95° 100° 105° 110°115° 120°125° 110° 120°

Kunlun Mountains, Figure 1 Snow and glacier distribution in the Kunlun Mountains (top graph) and Permafrost regions in the west China (bottom graph). pattern of streamflow under an extreme continental cli- valleys to 350 mm of the glaciated areas up to 5,800 m mate. During mild and rainy summer, weather is (Nakawo et al., 1990). The annual mean air temperature influenced by the westerlies, the streamflow from snow at 1,860 m is 7.9C, the estimated annual air temperature and glacier meltwater is the main supply to the oasis. over the alpine permafrost is below 2.7C, and ranges The precipitation distribution in the mountains is from 5.6Cto14.6C according to a lapse rate of air extremely heterogeneous ranging from 70 mm in the temperature with 0.75C/100 m in cold season from KUNLUN MOUNTAINS 681 observations for 6 years along an eastern neighboring rise in winter. The monthly temperature in December to valley (Abe and Wang, 2004). February has increased about 1.8–1.9C since the 1980s; The ablation period of the glacier is about 120 days whereas, the summer temperature since the 1950s has kept from June to September according to an investigation on consistent, although it is warming since 1990s. the glacier Gozha (5,200 m) in the headwaters, and shorter Precipitation during the 1950s–1970s had been gradu- and shorter eastward due to higher (Cao ally decreasing in the mountains; however, it has started and Aseit, 1992). The seasonal distribution of runoff, increasing since the 1980s. The maximum precipitation therefore, is quite variable as recharged mainly by summer has increased by 10–30% more than that for the previous meltwater and the monsoonal rainfall accounting for about 40 years, a little increase in at Keriya, the maximum before 80% of the annual runoff, and the maximum is 34% of 1980, for example, was 188.5 mm (1963), now once the annual in July, and the minimum is only 0.95% in exceeded 200 mm. Precipitation might have a more dra- February (Liu et al., 1999). matic increase in the high mountains. In long and cold winter from October to March, peri- It is interesting to note that the annual and summer flow odic events of soil freeze and thaw frequently occur. did not respond to the climate warming since the 1980s as Meanwhile snow cover exists from November to May above. River flow only in Keriya, however, indicated and increases snow thickness in high mountains. The river a little increase by 9% after 1985 with a delay of 5–6 years in winter, therefore, drains into the recessive period to the increasing precipitation, but meltwater in Hotan recharged only by groundwater, which is greatly affected River still kept rather stable or decreased a little. It is not by the freeze–thaw cycle of both the seasonally frozen coincidental with favorable water and heat both for rainfall ground and the active layer of permafrost. and meltwater runoff since the 1980s. Glaciers in the West Kunlun develop at the elevation over 5,200 m. Only individual large glaciers can extend Winter streamflow downward at a lower elevation below 4,800 m on a south- The change in winter streamflow is related to permafrost ern slope. The glacier coverage and meltwater contribu- degradation in cold regions (Liu et al., 2003; Qin et al., tion, being mainly hydrological index for such rivers 2005). Graphical and numerical analyses of winter recharged by meltwater in China, can reflect the role of monthly discharges are given in Figure 3, where winter meltwater and floods to streamflow. streamflow during October to March showed a strongly upward trend since 1997, the monthly flow increased by Change in air temperature and precipitation 47.4%, 32.2%, 18.8%, 32.7%, 33.5%, and 31.5%, respec- Climate in the West Kunlun indicates period of both cold tively. The remaining months generally did not exhibit any and warm temperature (Figure 2). The annual and winter significant trend. For example, an insignificant number of temperature has been consistently fluctuating from the trends were found for the summer flow from June to mid-1950s to the mid-1980s, and then it shows an abrupt September. According to the detection of the change point rise after the late 1980s, the annual average in two recent for monthly temperature, there was a statistically signifi- decades since 1987 is 0.8C higher than that of the previ- cant upward year (change point) after 1986. For winter ous 40 years. The highest annual temperature during streamflow, a statistically significant upward year was 53 years was recorded in 2006 and 1994. However, the observed since 1995, but the earliest one was in January rise in monthly temperature is nonuniform – the greatest 1985. Winter streamflow in all 6 months were much more sensitive than the seasonal air temperature; trend changes 2 were found in 4 monthly discharges with a significance

30 1 20 0 10

−1 0

Temperature anomaly (°C) Temperature −10

−2 anomaly winter streaflow 1957 1962 1967 1972 1977 1982 1987 1992 1997 2002 −20 a running mean with a 5 years 1955 1965 1975 1985 1995 2005 running mean of a 5 years Kunlun Mountains, Figure 2 The annual (bold line) and winter (October–March) (fine line) air temperature at Yurungkax Kunlun Mountains, Figure 3 Increase in winter flow station in the west Kunlun. (October–March) of Hotan River in the west Kunlun. 682 KUNLUN MOUNTAINS level at 99%, the most sensitive with the maximum vari- Huang, Y. Y., Liu, J. S., Shang, S. C., Ding, Y. J., and Liu, S. Y., ability was in January discharge (Huang et al., 2008; Liu 2008. Study on monthly runoff in winter, frozen earth and cli- et al., 2009). mate change in the Keriya river basin in the Kunlun mountains. Arid Zone Research, 25(2), 174–178. IPCC, 2007. Summary for policymakers. In Solomon, S., Qin, D., Summary Manning, M., Chen, Z., Marquis, M., Averyt, K. B., Tignor, M., and Miller, H. L. (eds.), Climate Change: The Physical Science In the Kunlun Mountains, snow cover, glacier, and perma- Basis. Contribution of Working Group I to the Fourth Assessment are the various reservoirs of water. The seasonal Report of the Intergovernmental Panel on Climate Change,Cam- meltwater from this region supplies most important bridge University Press, Cambridge, UK and New York. resources for the oasis life along the downstream area in Liu, C. H., 2000. Qaidam inland drainage basin, Glacier inventory the extremely arid desert basin, but the annual flood of China. Beijing: Science Press. can cause loss of human life and property and Liu, J. S., Fukusima, Y., and Hiyama, T., 1999. Hydrological eco-environment as well. From the point of the hydrolog- response of meltwater from glacier covered watersheds to the cli- matic change in the Northwest China. IHAS Publication, 256, ical scale, the response to the climate fluctuations from 193–207. snow and glacier-fed drainage basins with less glacier Liu, J. S., Hayakawa, N., Lu, M. J., Dong, S. H., and Yuan, J. Y., cover and in lower elevation seems to be more sensitive 2003. Winter streamflow, ground temperature and active-layer to the change in precipitation and air temperature than that thickness in Northeast China. Permafrost and Periglacial Pro- with more glacier cover at higher elevations. The hydro- cesses, 14(1), 11–18. logical regime of these rivers in the Kunlun is developing Liu, J. S., Wei, W. S., Huang, Y. Y., and Shang, S. C., 2006. Hydro- logical response of winter stream- flow to climate change and in an advantageous direction in winter to the local econ- permafrost degradation in Manas watershed, Tienshan Mts. omy and eco-environment of the oases. Journal of Glaciology and Geocryology, 28(5), 656–662. Nakawo, M., Ageta, Y., and Han, J. K., 1990. Climatic information from Chongce ice cap, the West Kunlun, China. Annals of Glaci- Bibliography ology, 14, 205–207. Abe, O., and Wang, L., 2004. Meteorological characteristics in Qin, D. H., Chen, Y. Y., and Li, X. Y., 2005. Climate and Environ- upstream regions of the Qira river, West Kunlun Mts., China. mental Changes in China. Beijing: Science Press, pp. 104–181. Bulletin of Glaciological Research, Japanese Society of Snow Yang, and An, 1990. Hotan river basin, Tarim inland drainage and Ice, 21,17–22. basin, Glacier inventory of China. Beijing: Science Press. Cao, Z. T., and Aseit, A., 1992. Runoff characteristics in Gozha gla- Zhou, Y. W., Guo, D. X., Qiu, G. Q., Cheng, G. D., and Li, S. D., cier region on the southern slope of the West Kunlun, China. 2000. Geocryology in China. Beijing: Science Press, Journal of Glaciology and Geocryology, 15(2), 156–161. pp.108–118.