What's in the news?

A news article on the Nature website posted on the 22^nd March 2011 looks at the people living alongside the Colonia River, in the Aysen region of Chilean Patagonia. Their significance is the threat they live under of sudden glacial lake outbursts from the above glacial mountains.

Whilst these lake outbursts are not uncommon globally (see this previous blog post), the region has experienced seven of these events since April 2008. Evidently, this is a much higher rate than seen anywhere else. During each of these seven outbursts, Lake Cachet 2 (of the Colonia glacier), has drained approximately 200 million m³ of water into the Colonia Lake and River in only a few hours. This has created a wave observed as far as 100 km downstream to the Pacific Ocean. The risks to the people of the Colonia River need no further explanation.

The Nature article uses the findings of a paper by Casassa et al. (2010), which has studied the floods to identify a primary cause.  The three year study ultimately found that the main cause was:

“...the repeated opening and closing of a tunnel 8 kilometres beneath the Colonia glacier, connecting Lake Cachet 2 above the glacier and Colonia Lake below it.”

Of course, this alone does not reveal the entire process of glacial lake outburst from the Colonia glacier. As a result, Casassa et al. (2010) found that climate change was the main culprit behind the frequency of these events. This happens as a result of the shrinking and thinning of the glacier in the past few decades, causing a weakening of the natural dam structure that the glacier forms. Thus the water moves between the two lakes with much greater ease. There seems little that can be done about this at present, although the paper offered some hope for the future:

“The researchers conclude that the discharges will continue until the ice has receded or thinned sufficiently to generate a permanent natural drainage channel.”

Figure 1. Ice-laden lake leading up to the Colonia glacier

Questionable impacts of glacier retreat on regional water security: dealing with uncertainties

Following on from the previous blog post, a paper by Archer et al. (2010) looks quite broadly at how the sustainability of water resources in the Indus basin might be altered by future changes in climate, as well as changes in socio-economic conditions. As was noted by the Indus water commissioner in the previous post, climate change effects on glacier shrinkage were suggested to be the primary cause of water scarcity. However, in this paper by Archer et al. it is argued that the impact of glacier retreat may be limited, especially when compared with other physical and socio-economic factors downstream of the Himalayan glaciers.

Agriculture in Pakistan is very much reliant on water which originates in the mountain sources of the upper Indus. As the authors note, these water resources are already highly stressed and are likely to get worse with projected rises in population. The paper considers the impact of climate change on these water resources in terms of three distinct hydrological regimes: a nival regime (dependent on melting winter snow), a glacial regime, and a rainfall regime. The mountainous sources of water are known to be affected by changes in temperature and in precipitation. The authors note that this is due to most of the runoff being derived from the melting of seasonal glacier snow and ice. Thus, any ablation of glaciers could quite easily affect water scarcity downstream.

Figure 1. The Indus basin

However, as the paper importantly notes, there is a great deal of uncertainty regarding how climate change might affect glaciers and river flow in the region. It cites several studies, which show regional conflicts with global patterns. Firstly, that summer temperatures (key for glacial melt) have actually fallen in the Karakoram between 1961 and 2000. Secondly, similar falls in temperature were found for the monsoon and pre-monsoon periods (April to May) in the Karakoram. Thirdly, that there have been significant increases in Upper Indus precipitation (both winter and summer) between 1961 and 1999. Fourthly, that extensive glacial mass balance records do not show shrinking glaciers. And finally, that in the late 1990s, there was widespread evidence of glacier expansion in the Karakoram.

With such conflicting evidence in the upper region of the Indus basin, it is unknown whether climate change will have a positive or negative effect on water resources in the region. The authors state that the hypothesis of reduced water resources relies on two assumptions: firstly that temperature and glacier melt are the primary impact on water resources, and secondly, that temperatures in the Upper Indus will rise in line with global climate predictions. As a result of this, they state that both of these assumptions are questionable. In particular, it is stated that river flow has been shown to not depend uniquely on glacier melt, but also rely on seasonal snowmelt and rainfall. This thus gives us the three hydrological regimes: nival, glacial and rainfall.

In the nival regime, the area of seasonal snow melt gives the largest contribution to downstream flow. This comes largely because the area of seasonal snow is much bigger than the perennial snow and ice. Of course, however, the area greatly reduces during the melt season. For this regime, winter precipitation has been shown likely to have the most significant impact on summer runoff. And unlike the glacial regime, there is a significant negative relationship between runoff and temperature on nival regimes. Archer et al. state that this:

“Can be explained by greater evaporative losses from the snow cover under higher temperatures and thus reduced runoff.”

As a result, the authors estimate that for a 2^oC rise in summer temperature, there would be an 18 percent reduction in runoff. However, the observed Karakoram decline in summer temperature would produce increased summer runoff.

In the glacial regime, the contribution to flow in the very high catchments is significant. However, the combined flow of these high catchments into the Indus represents an average of less than 30 percent. Here, there is a significant positive correlation between summer runoff and temperature. As well as this, winter precipitation doesn’t have such an influence. As such, spring and summer temperatures have the greatest impact on runoff. In this regime, runoff will rise initially with increased global temperature, but reduce sharply with declining glacier mass.

However, the findings of falling summer temperatures in the region mean that with this positive correlation between runoff and summer temperature, there is presently a downward trend in flow. The present and past behaviours of the Karakoram glaciers are noted in the paper:

“...glacier recessions were observed in almost all Karakoram glaciers for most of the 20th century until the mid-1990s. However, at lower elevations glaciers continued to decline. This seems to confirm that glacier loss is reduced in the Karakoram compared both with the neighbouring Himalaya and the Pamir mountains to the west.”

Finally, in the monsoon rainfall regime, the main influence falls over the southern plains and foothills of the Himalaya. Here, seasonal volume of runoff (as a result of rainfall) is lower than in the glacial and nival regimes. However, the monsoon rainfall produces more intense runoff and therefore highest flooding in the region. Therefore, this regime can be very important for water resources in the Indus basin. Yet as the paper states, the IPCC indicate that estimates of precipitation change hold great uncertainty, and that the impact of climate change on monsoon precipitation cannot be safely assumed.

In summary, based on the three regimes identified for the upper Indus basin, the paper shows that there appears to be little evidence for reductions in runoff and thus availability of water resources in the region. It must also be accepted, however, that much of this relies on uncertainty over glacier response to climate change in the Karakoram. Indeed, more recent past climate may not be a reliable guide to future change as well.

Whilst not the focus of this blog, it is important to mention briefly, the overall findings of the paper when physical and socio-economic conditions downstream are considered. Whilst the authors found inconclusive evidence of climate change causing reductions in runoff, they did find that several other factors had much greater influence on water resources further down in the Indus basin. Firstly, that urban growth and industrialisation will increase and demand further shares of the scarce water resources. Thus, economics plays an important role in water resource management, as large investment is needed to provide practical solutions whilst balancing the needs of security, health and education. Secondly, reservoir sedimentation means that water resources will diminish as storage is taken up by sediment. This problem will not reduce unless new reservoirs are built. Finally, the alternative of using groundwater in the spring for agriculture may soon have no practical use when the water tables fall from over pumping.

As can be seen, the issue of water scarcity in Pakistan’s Indus basin is very complex. The presence of and future changes to glaciers upstream is a very important part of the highly stressed water scarcity issues here. Only with greater knowledge of how these regional glaciers will react to global climate change, using past records and future modelling, will a truly certain answer be able to be given for these problems.

Changing glaciers and regional security

Focussing here on the human impacts of potential glacier loss, a 2009 Bloomberg article assesses the security risks associated with water scarcity, through looking at a report by the Asia Society. The article begins by stating that the water supplies of China and India are predicted to decline alongside the shrinking of Himalayan glaciers as a result of global climate change. This in turn is likely to begin or exacerbate regional conflicts.

This conflict, the author states, could come as a result of several factors. Asia contains half of the global population, yet has least water of any continent (with the superfluous exception of Antarctica). There exist problems of waterborne diseases throughout Asia, and these could certainly be amplified by water scarcity. As well as this, lack of fresh water could trigger mass migration and invigorate cross-border conflicts over water control. Evidently, there exist many potential sparks of conflict.

As well as this, the threat to agricultural production in the region is very real. Increased amplitude of dry and wet seasons caused by the influence of climate change on regional atmospheric patterns may overwhelm and destroy a proportion of crops. As well as this, any change in Asian crop yields (particularly China or India) would most certainly affect world food prices.

There is a particular problem in Pakistan, with approximately 77 percent of its water resources coming from outside of its borders. As a result of this situation, Pakistan holds a water treaty (since 1947) with India which guarantees sufficient supply. Of course, since that time, there have been several territorial and other disputes between the two countries. Most recently, Pakistan-based terrorists launched an attack on Mumbai in 2008. This further tension caused those in Pakistan to criticise the failures of India to always adhere to the treaty. Underlying all of this, they suggested, lay the long standing violent disagreement over the Kashmir region. These alleged abuses of the treaty in the past are particularly important when the population figures of 180 million presently (a tripling since 1950) and of a predicted 335 million by 2050 are seen.

The article then looks, albeit briefly, at the rapid melting of some Himalayan glaciers as a consequence of climate change and attempts to link this with the security issues. The author states that:

   

“Melting Himalayan glaciers now account for up to 70 percent of the summer flow of the Ganges River and about 55 percent of Asia’s other major river systems, according to the report. In 30 years, as the glaciers continue to retreat, the Indus and Mekong rivers could be dry during part of the year, the report said.”

Therefore, with the regional effects of climate change on temperature and a seemingly large degree of uncertainty regarding how the Himalayan glaciers will respond (unlikely uniformly), there is a great deal of potential danger associated with glacier retreat and water scarcity in the region.

In a separate article shown on the Policy Research Group website, much the same conclusions are found regarding the Pakistan example. The author, using World Bank estimates, shows that:

“... Pakistan’s Indus River will be negatively impacted by climate change, with incidents of flooding in the Indus basin expected to increase over the next 50 years. In addition, estimates suggest that there will be a 30-40% reduction in river flow over the next 100 years due to natural climatic and environmental changes.”

The article then notes that the Pakistani government, and more specifically the Indus water commissioner, feel the issue has received insufficient media coverage when compared to the potential damage that could take place. And looking at western media outlets, it would be hard to disagree with this assessment. Finally, the Indus water commissioner states belief in climate change as the primary driver of Pakistan’s water scarcity problems both now and in the future. However, as will be discussed in a following blog post, this is certainly not the belief of all scientists. Indeed, several have stated the exact opposite of this opinion.

Figure 1. View of the Indus River, fed from the glacial Himalayas

With regards to this last opinion of the Indus water commissioner; according to a weblog article by the Centre for Strategic and International Studies (CSIS) several officials in India believe that it is governmental mismanagement of water resources - and not climate change - that are responsible for water scarcity. Indeed, the article goes into slightly greater depth over the risk of conflict over water between Pakistan and India. It highlights the frustration of many of the rural poor in Pakistan regarding the observed extended dry spells, and a subsequent temptation to blame India for these droughts. Despite this desperation, there seems to be no evidence of this practice (according to the Indus water commissioner). Whilst cooperation on the water treaty has survived many of the previous disputes between the two countries, several militant Islamic groups have been seen to use this regional frustration for self gain. This threat of water cut-off by India is used by the militants to garner support for their wider agenda as well, which is in all probability aimed at destabilising the region.

It can be seen that water scarcity and regional security are very closely interlinked. Thus, with regional uncertainty of how glaciers may react to climate change, it is of wide ranging importance that greater certainty is achieved. Whilst this blog has looked more at the physical characteristics of past and future glacier change, it has been shown here that there is a inextricably linked human dimension to our knowledge of future glacier dynamics.

Black soot on the Tibetan Plateau

A paper by Xu et al. (2009) examines the impact of black carbon (soot) on the Tibetan Plateau - which contains the largest ice mass outside of the polar regions – and how this affects their survival for the future.

As has often been examined in this blog, the human as well as physical impacts of any changes to the region can be plainly seen. Most importantly, the authors offer clear explanation of the regional significance of water originating from the glaciers of the Tibetan Plateau. Firstly, the meltwaters are key in supplying the Indus, Ganges and Brahmaputra river systems, amongst others. Secondly, this freshwater source sustains over 1 billion people. Thirdly, for 25 percent of the people in western China, the meltwater from the Tibetan Plateau provides the main dry season water source (December – March). It is hard to envisage any clearer evidence of the reliance that so many people have on mountain glaciers.

Indeed, with the loss of headwater glaciers (supplying the source of the rivers), there will inevitably be a rapid decline in the availability of meltwater during the dry season. The increased melting of these glaciers is instead likely to provide spring time flooding. This then brings us onto what is causing this melting. The answer remains mostly obvious, in that glacier retreat is driven primarily by increased air temperature, on which much research exists. However, as Xu et al. (2009) note; areas of over 4000m elevation have warmed by 0.3°C per decade over the last 30 years, which is twice the rate of observed global warming. This high rate of temperature increase and subsequent rapidity of glacial retreat suggests that “additional mechanisms may be involved”. This of course then brings us onto the influence of black carbon on rates of glacier melt.

Whilst black carbon has an atmospheric impact (of warming the troposphere), the focus of this paper remained on the incorporation of black carbon onto snow and ice and the effect that this has for surface melt and albedo. When black soot forms on snow and ice surfaces, it serves to darken them, which as a result increases solar absorption and increases surface melting. As for the input pathways of black carbon, two factors are seen to bring it to the glaciers of the Tibetan Plateau. The first of these is the positioning of the plateau close to SE Asian industry. As the authors state, the region is presently and predicted to continue as the largest source of black carbon globally. Secondly, this positioning allows for the black soot to be carried upwards by winds and attached to snowflakes, prior to their deposition on top of the glacier surfaces. Thus, the glacier surface is darkened.

The study was conducted by taking ice cores from five key locations across the plateau. The aim of this was to establish temporal changes in the amount of black carbon since the 1950s. As well as this, the sources of snowfall for each part of the Plateau were established. For the N and NW parts:

“... associated mainly with the westerly jet stream, which moves southward toward the Himalayas in winter. Thus, black soot deposited on Himalayan glaciers derives primarily from two directions: west and south... so its upwind sources are principally Europe and the Middle East.”

For the S parts of the plateau:

“... [they] receive deposits from the west in winter and from the south in summer.”

Finally, the largest source of snowfall for the entire plateau is shown to be the Indian monsoon, which can move as far as 30-32°N in the summer.

Evidence of these soot and snowfall sources can be seen in the ice core records, which saw decreasing carbon in the NW and central regions of the plateau during the 1970s and 1980s. Such a reduction wasn’t seen in other areas, and reflects the successful implementation of new environmental regulations in Europe at the time. As well as this, the ice core records showed that carbon concentrations have increased in the southern Tibetan Plateau since 1990. This north/south plateau contrast shows both a difference in global attitudes to the emission of black carbon as well as showing the effect of different transport lengths (flow from Europe more easily disrupted/dissipated).

Figure 1. Black and Organic Carbon levels of the 5 ice cores. With annual and 5-year running averages.

The paper showed that the observed amount of black carbon was enough to influence the surface reflectivity of the glaciers. This comes together with the increased industrial activity, rapidly increased soot deposition from the 1990s and the accelerating rate of glacier retreat in the last 30 years. The study also showed that black carbon concentrations of 10 ng g^-1 reduce the visual albedo of thick snow to 0.01 – 0.04. This means that absorption of visible radiation is increased by 10 – 100 percent (dependent on snowflake type and the extent of soot mixing with snow). As well as this, the arrival of soot in spring allows for the melt season to begin earlier. Indeed the deposition of black carbon peaks between November and March, when the snow is at its maximum extent. More worrying, is the presence of black carbon in the accumulation zone of glaciers (where natural snow melt isn’t seen to any great extent). As well as all of these factors, the study has found that the melting process in the ablation zone can serve to increase the concentration of black carbon (and therefore intensifying its effects).

The authors also reference several studies which attempt to model the impacts of black soot on the regional climate of glaciated areas. (Hess et al., 1998) (Hansen & Nazarenko, 2004) (Flanner et al., 2007) (Shindell & Faluvegi, 2009). As Xu et al. note, the findings of these initial studies are very significant:

“These studies suggest that black soot is responsible for a substantial fraction of the regional warming of the past century, comparable to the fraction attributable to carbon dioxide. Assessment of black soot’s impact on glaciers will need to include the contribution that black soot makes to regional climate change, as well as direct effects on the glacier.”

Evidently, the influence of black soot on glacier change has been shown to be quite significant. If loss of the Tibetan Plateau as a fresh water source is to be prevented, then policy aimed at reducing SE Asian black carbon emissions is a necessity. It has been shown not only to directly alter surface albedo, but also to significantly affect regional climate as well. The authors state finally that if black soot levels are reduced and CO₂ levels are stabilised (somewhat unknowingly), then there remains a possibility that Tibetan glaciers may remain beyond this century. It is perhaps better put by stating that the future of the glaciated Tibetan Plateau remains very uncertain indeed.

Further reading: Whilst not looked at directly in this blog, this paper by Kopacz et al. (2011) takes a more technical look at the sources to the Tibetan Plateau and Himalayas, by modelling the global transport and radiative forcing of black carbon.

In the news: This topic was also covered in an article from The Guardian (2009), which looks at black soot originating from developing countries, impacting across the Tibetan Plateau and Himalayas.

What's in the news?

From looking at the changes in subglacial drainage of outlet glaciers before, this news article from August 2010 shows the efforts of French engineers to remove the subglacial lake which had formed from meltwater as a result of increased warming in recent years. The water, which lies under the Tete-Rousse glacier, threatened the safety of people in the underlying Saint Gervais valley with flooding from a sudden outburst event.

Indeed, the danger of this kind of event has been witnessed before. In 1892, 175 people were killed in the St Gervais valley by flooding from a subglacial lake. This occurred on the same glacier, and saw the collapse of the ice wall which had previously held in the subglacial lake. The origins of this original outburst can be seen in the paper by (Vincent et al. 2010). Today, several thousand tourists visit and stay within the valley each year, perhaps justifying the intervention seen here.

Figure 1. Large cavity from 19th century outburst

Therefore, to apply this pre-emptive measure, the engineers aimed to pump the water away by drilling through the ice and thus accessing the 2.3m³ ft subglacial lake. The process of draining took several months and aimed to reduce subglacial pressure by removing one third of the water present.

Figure 2. Drainage hole drilled into the glacier

The level of stress seen under the glacier was proposed to be as a result of both increased temperature causing greater amounts of meltwater, and by a brief anomalous cold period freezing the natural drainage routes for this meltwater.

Melting in the mountains...

Keeping with the topic of ice melt and global sea level rise; a study by Radic & Hock (2011) looks to determine the potential contribution of mountain glaciers to sea level rise. The authors modelled the predicted mass loss of mountain glaciers globally, and from that calculated the impact that this melting could make.

The study, which looked at 40 percent of the world mountain glacier total, was the most comprehensive to date. Unlike the previous Intergovernmental Panel on Climate Change (IPCC) study which conducted the same modelling on much fewer sample sites, Radic & Hock (2011) based their study on 120,299 mountain glacier and 1,638 ice caps. Despite these differences however, the prediction that the majority of small mountain glaciers will have disappeared by the year 2100 is shared by both this study and that done by the IPCC. The paper also shows that by 2100, the melting of mountain glaciers will contribute an equivalent amount to global sea level as the Greenland and Antarctic ice sheets. Evidently then, the concern over rising global sea levels should not just be confined to the large polar ice sheets, when the potential future impact of smaller glacier melting is just as great.

Figure 1. Regional glacier volume predictions for the 21st Century.

Using ten different global climate models to predict the volume changes of the glaciers, the study found that approximately 50 percent of the glaciers with the smallest surface area (< 5 km) will have gone. The method of future volume prediction is described below:

“To quantify future volume changes, we run the calibrated mass balance model ...with downscaled monthly twenty-first-century temperature and precipitation from ten GCMs, based on the widely used mid-range greenhouse emission scenario A1B. As glaciers lose mass owing to temperature increase, they retreat and hence their hypsometry changes. We use volume-area-length scaling to account for these changes and their feedbacks to glacier mass balance ... allowing receding glaciers to approach a new equilibrium in a warming climate.”

Evidently, as discussed many times in this blog, the potential impacts on people around the world could be huge. Particularly in developing countries, the reliance on glacial meltwater for drinking water and irrigation in agriculture can be massive. Looking beyond these regional impacts of glacier shrinking or disappearance, the ten climate models predicted that on average, the small glaciers and ice caps will themselves raise the global sea level approximately 12 cm by 2100. Across all of the models, the contribution of the complete or partially melted glaciers to global sea level was shown to be in the range of 8.7 to 16 cm. As the authors noted:

“All projections for the twenty-first century show substantial mountain glacier and ice cap volume losses.”

Thus, knowing how these mountain glaciers might impact on us globally is important! As the paper finds, these small mountain glaciers are relatively large contributors to sea level rise. Yet as the author has stated, that is important as the small glaciers only represent 1 percent of the Greenland and Antarctic ice sheets, and therefore become easy to overlook.

Additionally worrying is the conclusion by the authors; that their predictions of sea level rise from glacial melt are likely to represent cautious estimates. This comes as a result of modelling only the surface mass balance, which is particularly important for the ice caps that were looked at in this study, as the ocean influence on melting could potentially be greater than surface melting. As seen in previous blog posts, this is particularly true for polar glaciers, where many come into contact with warm ocean waters.

Of course, there are several factors that weren’t consider, quite possibly due to limitations outside of the authors’ control. However, considerations of glacier depth through global measurements could provide a better indication of the differences between individual glaciers based on their thickness. And as discussed in the previous paper by Schaefer et al. (2009), no consideration was made of glacier debris, which could certainly have an effect on glacier melting up to the year 2100.

However, the authors did produce new findings which showed that the Himalayan glaciers (which are affected by debris cover) might potentially grow slightly by 2100. This again disproves the originally contentious IPCC statement that Himalayan glaciers could have disappeared by 2035. The results of this study showed that the ten climate models predicted between a 15 percent decline or possibly net growth by 2100. This slow shrinking or even growth could come as a result of increased snowfall in the future. Of course, the predictions elsewhere are more sobering; with a 50-90 percent loss predicted for the European Alps and a 60-85 percent loss for New Zealand’s Alps.

As one of the lead authors notes: “Most of them will be gone by 2100.”

Increasing melt, decreasing flow?

In contrast to the traditional views on the relationship between ice loss and surface melting, there do exist conflicting opinions on the state of glacier acceleration in Greenland. One of the proponents of such a viewpoint are Sundal et al. (2011) who suggest that the Greenland ice sheet response to melting over time is in fact not increased acceleration, but a reduced flow into the fjords.

Figure 1. Ice velocity flows for the study area

The main premise of the paper involves the adaptation of the Greenland ice sheet’s subglacial drainage system to increased melting in warmer years. Thus the ice sheet, which could raise the global sea level by 7 metres if completely melted, could actually be seeing a reduction in the flow of ice into the fjords and oceans. Evidently, the implications for future predictions of global sea level rise are quite large. Indeed, the Greenland ice sheet has shrunk in the last decade due to rising temperatures, but the question of where the meltwater could be transported to remains unresolved.

The orthodox view of ice loss due to increased melting suggests that warmer temperatures cause ice to melt on the surface of the ice sheet. This meltwater is then transported to the base of the glacier, whereby it acts to increase the rate of ice flow across the bedrock and into the fjords. However the authors of the study, using remotely sensed imagery of 6 glaciers in SW Greenland, produced vastly different indications of how ice flow changes across years of variable melting. They instead found that the expected increases in ice sheet melting over the 21^st century (as a result of increased air temperature) may have no impact on ice loss as a product of flow into the ocean. However, it must be noted that the ice sheet isn’t threatened any less, as the studies of Straneo et al. (2011) Rignot et al. (2010) and Holland et al. (2008) revealed in the uncertainty of ocean-glacier interactions.

The study shows that the initial increase of ice flow was the same for both warmer and colder years. However, it also revealed that the expected slowdown in ice flow (as a result of glacier drainage adaptation) actually came quicker in the warmest years. Sundal et al. propose that as a result of much greater meltwater in warmer years, the internal basal drainage switches more quickly, which creates a drop in pressure and therefore slower ice movement across the bedrock. As the paper notes:

“Abundant melt-water can trigger a switch from inefficient (cavity19) to efficient (channelized20) modes of drainage and, consequently, to a reduction in subglacial water pressure and ice speed. Such events have been observed at High Arctic10 and Alaskan valley glaciers11, where summer speed-up is of shorter duration during years of high melting.”

The authors also conducted a numerical simulation of this switching between modes of drainage, which showed that above a critical meltwater flow rate of 1-2 cm per day, switching and thus glacier slowdown would occur.

Whilst displaying these findings, the authors accept that there are many conflicting studies on the topic and that whilst some studies have shown increased ice sheet acceleration from increased melting (Zwally et al. 2002), other studies have identified a long-term decrease in ice flow from the Greenland ice sheet over such a period of increased melting (van de Wal et al. 2008). The authors believe however, that both findings can be combined. Whilst the data shown in this study does identify an increase in the peak rate of flow during years of high melting; the important finding here is that the subsequent faster transition to more efficient subglacial drainage means that the duration and speed of ice flow is much lower than when compared to years of low melting. In cooler years, where the critical meltwater flow rate of 1-2 cm per day is not exceeded, the more efficient ‘channelized20’ mode of drainage does not occur.

With the predicted increase in Greenland ice sheet contribution to global sea levels over the 21^st century (Mernild et al. 2008), the impact of melt-induced ice flow acceleration must not be ignored. Thus, the findings of this study and others are important for the predictions of future climate change impacts. Particularly if the findings of this study are correct, then the observed and predicted rises in air temperature over the next century may not push the Greenland ice sheet over a suggested ‘tipping point’, as a result of melt induced glacier acceleration. However, the melting will of course still occur (increasing basal storage of meltwater) and as the authors admit, their findings do not cover how the switch to more efficient drainage might be adversely affected by more short term spikes in melting. As well as this, the previously discussed studies of warm ocean influences on the outlet glaciers suggest that a much greater melting influence occurs as a result of subtropical waters. Both ocean and melting influences on glacier acceleration remain non-definitive at present; and so it is important for further research to identify how these processes have acted in the past, so that we might better understand the potentially devastating  impacts that acceleration of the Greenland ice sheet could have in the future.

Ocean-glacier interactions in a changing climate

A recent paper by Straneo et al. (2011), has investigated the role that ocean circulation plays in bringing warm subtropical water into contact with the outlet glaciers of the Greenland ice sheet. As well as identifying the presence of these warm waters, the authors also show the complex factors which influence meltwater circulation and the impact that these warm waters have on the Greenland glaciers.

An initial paper by Straneo et al. (2010) identified the presence of subtropical waters flowing through the large glacial Sermilik Fjord in East Greenland. This initial study involved two extensive surveys of the fjord over the summer months, between July and September. These surveys found temperatures in the fjord as warm as 4^oC. The use of seals tagged with temperature depth recorders also revealed that the fjord warms from July all the way to December, meaning warm temperatures in the winter. This longer temperature and depth record also revealed that the presence of these warmer subtropical waters can also be found throughout the year. This initial study was the first of its kind to identify the presence of these warm subtropical waters in glacial fjords, and also exposed the great unknown regarding interactions between the ocean and glaciers. This is particularly important, as the rate of ice mass loss for the Greenland ice sheet has increased and the contribution it makes to global sea level rise has more than doubled over the last decade. As Rignot et al. (2010) found, there has been a tripling of ice mass loss in Greenland between 1996 and 2007. Additionally, as Mernild et al. (2008) observed, the total amount of Greenland Ice Sheet freshwater input into the North Atlantic Ocean expected from 2071 to 2100 will be more than double what is currently observed as a result of climate change. Of this loss approximately 50-60 percent of ice loss is due to the speeding up (or acceleration) in the flow of outlet glaciers. As shown, this mass exodus of glacial water comes mostly from the acceleration of outlet glaciers into the fjord itself. As Rignot et al. (2010) note, the melt rate of studied glaciers in Greenland is over 100 times larger when looking under the surface. This submarine melting of the glacier front has been shown to contribute to glacier acceleration. Additionally, this acceleration of outlet glaciers coincided with a notable warming trend in the subpolar North Atlantic. Thus, the favoured hypothesis states that the changes in ocean circulation cause ice thinning and ungrounding of the glacier terminus which then leads to acceleration of the ice flow.

A similar paper by Holland et al. (2008) looked at the causes behind the acceleration of several outlet glaciers, in both Greenland and Antarctica. In one particularly strong acceleration, the Jakobshavn Isbrae glacier on the west coast of Greenland saw a doubling in glacier velocity.  This, like the Sermilik Fjord was also linked to the presence of warm subtropical waters. This paper also gives a rather detailed atmospheric link for explaining the presence of these subtropical waters. To best detail this, the explanation made by the author is shown here:

“The warm, subsurface waters off the west Greenland coast are fed from the east by the subpolar gyre of the North Atlantic, via the Irminger current. Since the mid-1990s, observations show a warming of the subpolar gyre and the northern Irminger Basin. A key source of variability in the forcing of the subpolar gyre is the North Atlantic Oscillation (NAO). A major change in the behaviour of the NAO was observed during the winter of 1995-1996, when it switched from a prolonged positive phase with strong westerly winds to a negative phase with weaker winds. The net effect of the change was to weaken the subpolar gyre with the consequence of moving the subpolar frontal system (the boundary between cold polar waters and warm subpolar waters) from an easterly position to a more westerly one. Such a large-scale change in the subpolar gyre allowed warm subpolar waters to spread westward, beneath colder surface polar waters, and consequently on and over the west Greenland continental shelf.”

Evidently, the processes behind the arrival of warm subtropical water are very complex. Indeed, no attempt to link atmospheric phenomena to this observed warming is made by Straneo et al. and the root causes remain largely unknown and unproven. However, it can be seen from these studies that the entry of warm subtropical waters is having an effect on the melting and acceleration of the glaciers.

Moving on to the paper by Straneo et al (2011), further research was conducted one year on from the initial findings of subtropical waters in the fjord. This additional research allowed a winter survey of the fjord, with measurements taken much closer to the glacier. In all, the research in the Sermilik Fjord revealed a complex interaction between the Helheim glacier, freshwater melt and warm saline water. The expected circulation had been for the warm subtropical water to travel towards the glacier as deepwater and subsequently melt it. Then, the cooler mixture of subtropical and freshwater was expected to rise to the surface. However, as the authors have shown, the circulation proved to be much more complex. Whilst the waters do consist of a cold Arctic layer at the top and a warm layer beneath, the behaviour of the glacially modified water was different to what had been expected. Because of the density of the glacially modified water, whilst it was shown to rise, the majority does not come out on the surface and instead spreads horizontally between the polar cold and subtropical warm layers. This has the effect to prevent much greater melting, protecting the upper third of the glacier (acting as a vertical barrier to the transport of heat) and instead creating a floating tongue, below which submarine melting has taken place. The creation of this glacial tongue serves to alter the stability of the glacier and potentially increase acceleration into the fjord.

Figure 1. Survey locations within the Sermilik Fjord

The study shows that the process of quantifying the ocean’s impact on outlet glaciers is much more complex than just estimating temperature and velocity. Instead, more accurate estimates of heat transport to the glaciers will involve flow measurements over long periods of time, measure subglacial discharge and consider the effects of stratification and altering circulations in the fjord. Indeed, changes to the temperature of subtropical waters alone are not adequate indicators of the oceans impact on submarine melting. Thus, more research is needed into exactly how these processes occur and how we can better predict the effect on glaciers in the future. As was mentioned before, the contribution of the Greenland glaciers to global sea level rise has increased in the past decade and as yet we do not fully understand the processes governing this. The potential impact for many countries is great, and knowing how increases in melting and acceleration are occurring is key. Several studies looking at glacial fjords in Greenland have been shown, and the features and the effects of submarine melting are common to many other systems in Greenland, as observed in Holland et al. (2008) and Rignot et al. (2010). A final comment by David Holland shows the significance of these studies for the future and the need for greater research:

"The melting of the ice sheets is the wild card of future sea level, and our results hint that modest changes in atmospheric circulation, possibly driven by anthropogenic influences, could also cause future rapid retreat of the Antarctic Ice Sheet, which holds a far greater potential for sea level rise."

What's in the news?

An article on BBC News from the beginning of April discusses a new study by Glasser et al. (2011) which identified both new ice loss figures and the resulting sea level contribution of melting mountain glaciers in Patagonia. This covered changes for 270 of the largest glaciers in Argentina and Chile, and showed new estimates for the amount of melting that has occurred since the end of the Little Ice Age. The new method, which utilises the spread of glacier debris and vegetation extents to calculate ice loss, has also revealed new findings.

Figure 1. Clear trimline of vegetation used to delimit vertical extent of ice

The study showed that the glaciers, which covered a huge area at the end of the Little Ice Age , have lost approximately 606 km³ since then. As the lead author, Prof. Neil Glasser stated:

"Previous estimates of sea-level contribution from mountain glaciers are based on very short timescales... They cover only the last 30 years or so when satellite images can be used to calculate rates of glacier volume change. We took a different approach by using a new method that allows us to look at longer timescales.”

The paper made an interesting discovery that the rate of melting at the start of the 20^th century was actually lower than previously suggested. However, a much starker discovery showed that in the last 30 years, the rate of glacial melting has become over 100 times that of the previous 320 year average!

Not only does this paper reveal new insights into past melting rates and contributions to global sea level rise, but it also shows how present melting rates are the highest in over 350 years. This has been the first estimate of glacier melt contribution to sea level rise during industrial times and evidently has important points to consider for the future. With local impacts of glacier retreat to consider (such as water supply), it has been shown that the melting of mountain glaciers can also have a global effect both now and certainly in the future.

Reply to Mölg et al. (2006)

In the previous post, the comment of Mölg et al. (2006) gave a number of clear critiques to the original paper of Taylor et al. (2006). As part of this peer review process, the response of Taylor et al. to this criticism was published in the same journal issue. As with before, the reply here will hopefully be stated in a concise manner.

The authors begin by accepting the uncertainty surrounding not only the influence of air temperature on glacial recession in the tropics, but the influence of atmospheric humidity as well. Perhaps most importantly, with respect to the comments of Mölg et al., the authors accept the uncertainty over the relationship between surface and higher elevation temperature trends in the tropical free troposphere. As well as all of this, they agree with the contention that surface energy balance (SEB) models provide the ideal platform to show climate and glacier relationships. However, the authors rebuke the criticism of Mölg et al. by stating that the lack of SEB climatic parameters was made clear in the original paper. Thus, Taylor et al. accept that a definitive understanding of climate-glacier interactions in the Rwenzori Mountains isn’t presently possible. However, the authors do believe that the evidence for air temperature being the primary influence on glacier dynamics is much greater than the available evidence for decreasing humidity (as theorised by Mölg et al).

The main criticism of the original paper was the validity of the observed air temperature trends at meteorological stations being reflective of trends higher up in the mid-troposphere. Whilst Mölg et al. believed that the original paper had simply ignored any inconsistencies, the authors contend that in fact there is mixed consensus on this issue and that the studies cited by Mölg et al. are unfairly selective. Indeed, whilst the critique cited mid-troposphere trends through the 20^th century as showing no warming (as opposed to the lower elevation measurements), the choice of start and end points alone can greatly alter the observed trend. As Taylor et al. state:

“...in the paper by Gaffen et al.  [2000] using MSU data in which at 500 hPa a cooling trend is detected between 1979 and 1997 but an overall warming trend occurs between 1960 and 1997.”

Further to this, more recent studies are mentioned which show the validity of the original paper:

“Nevertheless, recent studies that employ diurnal corrections to MSU observations between 1979 and 2003 [Mears and Wentz , 2005] and homogenized radiosonde data sets (HadAT2) between 1958 and 2002 [Thorne et al. , 2005], show that the middle troposphere warmed at a similar or slightly greater rate to the surface in the tropics [Fu and Johanson , 2005; Santer et al. , 2005]...”

The authors also criticise the use of reanalysed data by Mölg et al. to show a difference between surface air temperature and data in the mid-troposphere. According to Taylor et al. the use of NCEP data is controversial, given the widespread consensus that reanalysis data is unsuitable for trend analysis in climate change studies due to time-dependent errors. The authors use an alternative and perhaps less controversial dataset, the HadAT2 radiosonde data, to show that upper air temperature records closest to the Rwenzori Mountains show consistent warming from 1958 to 2005. As well as this, these upper air trends match the surface air temperature trends over the same period from another homogenised dataset (CRU TS 2.0).

Following this, the authors seek to discredit the observed trends of decreasing specific humidity by Mölg et al. The same NCEP reanalysis data was used to identify this trend in the mid-troposphere through the 20^th century. Apart from the already identified bias in the NCEP data, the reliability of the humidity readings is even more questionable according the Taylor et al. This is mostly due to humidity being a statistically derived parameter, and thus not suitable for trend analyses. The humidity data cited by Mölg et al. was in fact uncorrected, which means that any systemic dry biases remained unresolved. As well as this, a 20^th century decline in humidity is unsupported by the CRU TS 2.0 precipitation / water vapour datasets. Mölg et al. also argues that the observed East African glacial retreat was initiated by a vast reduction in moisture at the end of the 19^th century. This change in moisture is based on historical lake levels in East Africa. Taylor et al. argues that this drop in lake levels was actually the abrupt finish to a brief decade-long high lake stand, and nothing more. Indeed, a good comparison to this event was a 2.3m rise in the level of Lake Victoria between 1961 and 1964 which did briefly increase the regional humidity. However, this increased humidity gave a very small, year-long advance to the Rwenzori Mountain glaciers and does not support the humidity hypothesis of Mölg et al. which suggests continual glacial retreat in the latter half of the 20^th century. A final statement by Taylor et al. identifies another clear flaw in the humidity hypothesis:

“Even ignoring concerns regarding this evidence, the argument that these climate events are responsible for the expected demise of small, fast-responding glaciers that have persisted for at least 5000 years is improbable.”

In summary, the authors accept that both air temperature and air humidity are viable causes of glacier retreat in the tropics and are likely related in some way. There is also wide agreement that the meteorological data surrounding the Rwenzori Mountains is far from ideal. However, in the view of the original authors, there is more robust data supporting changes in air temperature as the primary cause of glacial retreat.

By identifying that glaciers in the tropics may well behave no differently to alpine glaciers elsewhere with regards to rising air temperature, it is important to further research this. As has been mentioned previously, these tropical glaciers often have great cultural significance as well as being key for the success of agriculture and survival in parts of East Africa. A more definitive account of how the Rwenzori Mountain glaciers responded to changes in atmospheric conditions in the past will help us to greater predict how long they may last. However, whilst further research may reveal that some of the glaciers could perhaps last longer than 20 years stated by Taylor et al., there remains little doubt that their days are numbered.

Comment on Taylor et al. (2006)

As mentioned in the previous post covering the Rwenzori Mountains paper (Taylor et al. 2006), there were a number of conflicting theories over the primary forcing factor behind glacier retreat in the region of study. Scientific criticism came from Mölg et al. (2006) in the form of a post-publication comment; highlighting the author’s disagreement over several issues. As part of open peer commentary, Taylor et al. (2006) also replied to these criticisms in the same publication. It is hoped that this exchange can be documented clearly in the next two blog posts.

Mölg et al. begins by accepting the evidence of strong glacier retreat in the Rwenzori Mountains during the 20^th century. However, they state disagreement with the main conclusion of Taylor et al., that there is a strong link between air temperature and this retreat. Part of this is due to the lack of a surface energy balance (SEB) model, as a result of having no meteorological or glaciological records from directly on or very near the glacier. Mölg et al. note that the two glacierised massifs in East Africa that have had SEB modelling performed, Mount Kenya and Kilimanjaro, showed that the parameters relating to atmospheric moisture (humidity, precipitation, surface albedo, cloudiness) are dominant. Thus Mölg et al. state that the focus of the authors on rising air temperature fails to identify these greater intricacies because of the lack of an SEB model and instead, the authors have over simplified the relationship between atmosphere and glaciers in the tropics.

Moving on from this, Mölg et al. addresses the issue of how local meteorological data was misused in this study to show how climate may affect glacier mass balance. The main contention here is how the key rules of atmospheric trend analysis were apparently neglected by Taylor et al. when stating that the measured trends found at much lower elevations, would be identical to those in the higher glacial elevations. Mölg et al. first states a problem with the use of the phrase “thermal homogeneity” by Taylor et al. to describe the same air temperature trends at different tropospheric levels. Instead, Mölg et al. state its use is more appropriately used to refer to seasonally constant temperatures across all tropical regions.

More importantly however, Mölg et al. move on to state that this concept is incorrect in any case. They cite several studies which show through observations and modelling that trends occurring in the lower elevations do not have to appear in the mid-troposphere. Mölg et al. state that the authors ignored these findings and therefore ignored changing trends in the vertical dimension. They then additionally attempt to show the existence of these vertical changes through the use of NCEP/NCAR reanalysis data of past and present atmospheric conditions over the Rwenzori Mountains. Whilst there are known problems with the use of NCEP/NCAR reanalysis data, Mölg et al. state that in these tropical areas, data of surface air temperature and moisture (even on the mountain tops) can be seen as reliable. Firstly, Mölg et al. displays the temperature trends at the elevations of the meteorological stations used by Taylor et al. This confirms a warming in air temperature throughout the 20^th century. However, when they show the temperature trends from higher elevations (equivalent to the height of the glaciers), there apparently exists no significant temperature trend across the same temporal period. Mölg et al. therefore draws the conclusion that air temperature is unlikely to be behind the 20^th century glacier retreat in the Rwenzori Mountains. Instead, using the same reanalysis data, humidity can be shown to have decreased over the same period. Accordingly, this would then increase sublimation, affect the energy flux of incoming longwave radiation and thus alter the net radiation budget. According to Mölg et al. the above mentioned lack of SEB model parameters in the study of Taylor et al. rendered them unable to consider these factors! Following this, a slightly contentious statement discounting the hypothesis of Taylor et al. perhaps puts too much reliance on the use of reanalysis data, which has known problems:

“Thus, the scenario of a Ta -driven humidity increase in East African highlands, and associated accelerated melt... prior to the availability of reanalysis data, seems to be outdated and has not been reconsidered by Taylor et al.  [2006]. However, this is their only causal explanation of how Ta changes would control glacier mass balance on Rwenzori.”

The final criticism from Mölg et al. relates to the use of remotely sensed imagery to show glacier extent. Here, they state that the authors unsatisfactorily discerned between the true glacier surface and transient snow. This issue stems from the use of a false colour composite, which does not apparently allow the differentiation between temporary snow and the perennial snow cover on top of the glacier surface. As a result, the original paper reports incorrect areal extents of each glacier and thus incorrect temporal retreat rates. As a result, Mölg et al. questions the validity of the observed steady rate of decline and perfect linear retreat in time. They instead conduct their own analysis using ASTER satellite imagery, with much clearer delineation of glacier boundary. The results are stated as such:

“The temporal retreat of glacier areas shows clear differences between the three Rwenzori mountains, depending on size and elevation of the glaciers. In particular during the last decades, the larger and higher elevated glaciers on Mount Stanley are retreating much slower...”

In summary, whilst Mölg et al. have critiqued some of the methodology employed in the original paper, their main concern involved the assertion that air temperature was the primary forcing factor in the retreat of the Rwenzori Mountain glaciers. Instead, they believe it to be a much more atmospherically complex issue than this, ultimately related to changes in humidity. Whilst air temperature cannot be denied to have an effect on glacier dynamics globally, it is the belief of Mölg et al. that the glaciers of East Africa may well have acted differently in the past; which is a very important issue if we are to fully understand the rapidly changing glacier conditions of the tropics in the future.

Glacial recession in the Tropics - a complex case

A paper published by Taylor et al. (2006) looks for the climatic causes of recent glacial recession in the Rwenzori Mountains, East Africa. The team, which comprised Richard Taylor of UCL Geography, colleagues from UCL and Makerere University and two others from Ugandan governmental departments used field surveys and analyses of remotely sensed imagery to report the reduction in area of the Rwenzori Mountains glaciers. The study saw an overall reduction in glacier area from 2.01 ± 0.56 km² in 1987 to 0.96 ± 0.34 km² in 2003. At the same time, air temperatures were seen to increase by approximately 0.5^oC each decade; with no significant change in precipitation. By extrapolating trends from area data collected since 1906, it is suggested by the authors that the glaciers of the Rwenzori Mountains could disappear within the next two decades, primarily as a result of rising air temperatures.

Fig 1. The Rwenzori Mountains within Uganda and the Central Rwenzori Massif in East Africa.

The paper indicates the importance of alpine glaciers in the tropics as very sensitive indicators of tropical climate. This is of course very useful for areas such as East Africa, where meteorological records are limited. Many studies are referenced, which investigate the shrinking of glaciers in the highlands of East Africa over the 20^th century; a trend shown by the satellite imagery utilised in this study. The authors indicate that debate exists over the nature of climate change in the highlands, as well as debate over the exact cause of glacial retreat in the tropics of East Africa.

The positions of terminal moraines were measured to show the reduced extent of the glaciers in this study. Two ‘indicator’ glaciers were used, which had been previously monitored, and could thus show any trends in glacial retreat. For the satellite data, LandSat5 and LandSat7 imagery were used to measure snow and ice cover. In slightly more detail, snow and ice were identified using a supervised classification on a false-colour composite, using both visible and infrared bands to do so. The accuracy of the classification was assessed by also comparing it with the results of a Normalised Difference Snow Index (NDSI) algorithm.

The results of this study were found to be in agreement over a clear trend of recent glacial retreat in the Rwenzori Mountains. The retreat of one of the indicator glaciers was shown to be quicker than the other; however, this was posited to be as a result of differences in elevation and bed morphology causing a contrast in the supply of ice to each glacier. Regardless, the authors state their findings clearly:

“Analyses of LandSat imagery... identify a 50% decrease in the total area of glaciers from 1987 (2.01 ± 0.11 km²) to 2003 (0.96 ± 0.34 km²). Broad agreement exists between estimates of glacial cover (<12% difference) derived from each method. The results of the NDSI-classified, Land-Sat image from 1987 are consistent with historical data derived from aerial and terrestrial photography.”

Whilst recent glacial retreat was clearly shown by the results, the paper also notes that when combined with measurements made over the last century, there is a steady rate of decline evident across this longer time period as well. As can be seen, the results shown in this paper are in agreement with many other studies, in stating that the glaciers of the Rwenzori Mountains have decreased in area. However, where much dispute still exists is in the debate over the primary climatic driver for this observed glacial retreat.

Fig 2. Reducing glacial area of the Rwenzori Mountains over time.

The knowledge of exactly what is controlling the retreat of glaciers in the tropics is, as always, very important. Not only do the glaciers in the Rwenzori Mountains have importance for local agriculture and drinking water through seasonal meltwater; but the glaciers also hold great cultural significance for the communities that live around them. If temperature alone is found to be controlling their quick retreat, then we can safely say that the glaciers will not exist for much longer. However, if more complex processes control these glacial dynamics, then perhaps some hope exists for their future – however futile. Putting local impacts aside though, there are many more glaciers in the highlands of the tropics, and thus by improving knowledge of how they react to anthropogenic climate change, we can better respond to protecting them.

As noted before, there are no continuous meteorological observations close to the Rwenzori Mountains. Therefore it is not possible to directly analyse the climatic factors which have been driving regional glacial retreat. The paper notes that that previous studies of glacial dynamics in East Africa, particularly including Mölg et al. (2003), stated that the observed 20^th century retreat came as a result of a sudden decrease in humidity c.1880. As a result of this decrease in humidity, the authors state that cloud cover would decrease and thus increase the exposure of glaciers to solar radiation. On top of this, with reduced cloud cover; precipitation would be much lower and would therefore decrease accumulation, the subsequent lowering of albedo would also serve to increase the absorption of radiation. As a result of all of these effects, the rate of glacier mass loss would also increase over time.

However, to show the complexity of disagreement over this issue, the paper also cites another study which argues for a climatic driving factor that is completely anathema to that of Mölg et al. (2003). A paper by Hastenrath (2001) is described, which states that glacial recession in East Africa during the 20^th century came as a result of a warming trend, which has caused an increase in atmospheric humidity. This increased humidity is then explained as causing a reduction in sublimation and therefore allowing more of the sun’s radiation to melt glacial ice (as a result of a saving of latent heat). The authors note that this process has been well documented in other tropical glaciers, such as those found in Bolivia, where increased humidity of the wet season causes higher melt rates as a result of inhibited sublimation.

After looking at both of these hypotheses, the authors of this paper reject both of them. In particular, the findings of Mölg et al. (2003) are questioned, with the proposed changes in humidity and solar incidence not being accepted as causal factors for the observed spatially uniform retreat of glaciers at lower elevations. Instead, Taylor et al. (2006) find this indicative of increased air temperature as being the primary driving factor for glacial retreat.

This emphasis on air temperature as the primary driving factor is the main conclusion of this paper, and the authors provide evidence of this using long term air temperature records. These records of daily air temperature show consistent trends, with confidence levels of 99% or greater, towards temperature rises of approximately 0.5^oC per decade since the last period of glacial advance in the early 1960s. As well as this, the paper cites other gridded climate data which shows a smaller per decade surface temperature increase of 0.15^oC over the same time period. Evidently, there is a significant regional warming trend identified in this study. One point worth noting is the relevance of this temperature data, which was measured at a lower elevation than that of glacial cover in the tropics. The authors of the paper address this issue however, stating that due to thermal homogeneity of the troposphere, this rise in temperature at lower elevations can be expected to occur in the much higher glacial areas. However, it must be noted that this point is still much debated. As Taylor et al. (2006) note in their reply to Mölg et al. (2003) (both of which will be discussed further in depth):

“Significant uncertainty persists in temperature data for the tropical troposphere whether these derive from satellite-borne Microwave Sounding Unit (MSU) observations or in situ measurements using radiosondes, particularly in data-poor regions like East Africa.”

Regardless of this issue, the use of lower elevation temperature data by Taylor et al. (2006) is taken to be reliable, and is supported by other studies of tropical troposphere temperature data. Rather than discussing these studies here however, an analysis of the scientific disagreements between Taylor et al. (2006) and Mölg et al. (2003) will be looked at in another blog post.

Finally, this paper also looks at the reduced precipitation claim made by Mölg et al. (2003) and finds it unsatisfactory. The Ugandan meteorological data used in this study fails to support any trend in reducing precipitation over the 20^th century, despite there inevitably being much interannual variability present in the records. As well as this, whilst there does exist evidence of an abrupt decrease in humidity around 1880 in East Africa, there is no long term temperature record going back to this point. Therefore, the authors state it is not possible to test the relationship between humidity and temperature in the tropics, as stated by Mölg et al. (2003).

In summary, there are evidently several different proposed theories behind the retreat of glaciers in the Rwenzori Mountains. Whereas the ideas of Taylor et al. (2006) seem to consider fewer variables than the perhaps more complex hypothesis of Mölg et al. (2003), it would seem that the former proves more robust in standing up against the meteorological evidence. Sadly, if this is the case, then the near certain predictions of these mountain glaciers disappearing within a few decades will no doubt have a huge impact for the people of East Africa.

Interhemispheric differences and regional influences on glacier dynamics

A paper by Schaefer et al. (2009) gives an interesting new view of interhemispheric glacier fluctuations during the Holocene, through the use of a high resolution ¹⁰BE chronology and historical moraine records.

Natural climate variability during the interglacial conditions of the Holocene is a fundamental baseline for evaluating the anthropogenic impact on global climate today. As has been explored previously in this blog, by knowing past natural variability, we can better identify the effects of synchronous global climate change at present. Studies in the past few years have challenged the traditional view of a stable Holocene climate. These more recent studies of Northern Hemisphere paleoclimatic data show that cooling temperatures, accompanied by millennial-scale variations (leading to Medieval Warm Period (MWP) / Little Ice Age (LIA) oscillations) occurred through the Holocene. As can be seen, terrestrial paleoclimatic data for the Northern Hemisphere had been widely used. However, sources of this data are much sparser in the Southern Hemisphere and it therefore remains unclear whether the south followed the north in this cooling trend and whether the same MWP/LIA oscillation was observed globally.

The authors of this paper aim to investigate these interhemispheric linkages, by seeing if Holocene glacier advances in New Zealand relate to those seen in the Northern Hemisphere. The terminal moraines deposited by each glacier advance in the Southern Alps of NZ was dated and used to look for any north/south pattern and therefore if global or regional climate signals dominated during the Holocene.

Whilst many previous chronological studies have suggested that the most recent cool interval in New Zealand came at the termination of the northern LIA (c.1850), more recent studies using dendrochronology have instead suggested that the glacial maximum in the southern hemisphere is notably older than the termination of the northern LIA. This problem of obtaining reliable ages has inevitably impacted on the ability of studies to give records of glaciers in a global context. Therefore, this study applies cosmogenic isotope dating, which has a much higher resolution than radiocarbon dating for example, and which can also be applied to more recent deposits. Using ¹⁰Be exposure dating, a total of 74 boulders were sampled on the Mueller Glacier Holocene moraines, which offered the most complete sequence of moraines. These moraine exposure ages were then interpreted as being representative of a terminal moraine and therefore the end of a glacial event. Overall, the authors noted a strong positive correlation between the exposure ages given by the cosmogenic isotopes and the historic ages of the moraines. As well as this, the paper notes that these records also matched well to a radiocarbon chronology using buried wood found within several lateral moraines. With both ¹⁰Be and ¹⁴C records indicating the maximum extents of each glacial advance, a minimum of 15 different pulses of advance and retreat were identified since the mid-Holocene. Using past tree-ring data as well, all of the data showed a clear record of summer temperature changes and accompanying glacier fluctuations during the late Holocene.   

With the widely-supported high resolution record of glacier fluctuations in the Southern Alps of New Zealand, the study compared the Holocene moraine record to Northern Hemisphere records. From this, three very clear and succinct conclusions were drawn. Firstly, noticeable interhemispheric differences were seen in the timing of maximum ice extent. The NZ glaciers were at their greatest extent c.6500 years from present; whilst in the Northern Hemisphere, this point came much more recently during the LIA. The second conclusion was that during the northern warm periods, glacial advances were still being observed in NZ, advancing beyond 19^th century terminus. Thirdly, the greatest similarity between the Northern Hemisphere and NZ records was seen between 300 and 700 C.E. (the Dark Ages), with broad similarities seen during the past 700 years (LIA) as well. The study notes that the similarities during the LIA consisted of multiple glacier advances, followed by a termination c.1850. However, despite these broad similarities; whereas the northern glacier record is dominated by LIA-maximum terminal moraines (under 400 years old), the NZ glaciers of this study proved to have reached their maximum extent earlier than this. For example, the terminal moraine of the Mueller glacier is 570 years old and thus deposited much earlier than the moraines of the Northern Hemisphere. The authors also note that this pattern of broad similarity, with conflicting smaller detail, is true for the past 150 years as well.

Figure 1. Timing of NZ glacial fluctuations with N. Hemisphere glacial records (Schaefer et al., 2009)

The results of this study showed neither interhemispheric synchrony of mid to late-Holocene climate fluctuations, or of any kind of interhemispheric asynchrony. Accordingly, the authors view any kind of global driving mechanism as unsatisfactory in explaining Holocene glacier dynamics. They also consider two other hypotheses which try to deal with regional climate differences. Firstly, that changing strength of deepwater production between the north and the south can explain interhemispheric imbalance. This is rejected for its inability to explain regional differences instead of just hemispheric differences in Holocene climate. A second alternative hypothesis regarding solar forcing of regional temperatures isn’t dismissed, and the authors state that good correlation exists between the solar record and the NZ moraine chronology used in this study. However, as the authors state very clearly:

“...we suggest that regional ocean-atmosphere oscillations may account for the observed glacier fluctuation pattern... the Interdecadal Pacific Oscillation (IPO) has been an important influence on glacier behaviour in New Zealand over the past few decades... These changes are well reflected in New Zealand’s glacier length fluctuations.”

Evidently, the authors propose a number of different hypotheses for the regional glacier dynamics seen in NZ. Whilst one clear regional influence is put forward in the form of the IPO, the suggestion of solar forcing as well perhaps suggests the author’s belief in more than one forcing factor on regional temperature. However the summary of the findings remains to the point:

“...our study shows that mid- to late Holocene glacier fluctuations were neither in phase nor strictly antiphased between the hemispheres, and therefore it is likely that regional driving or amplifying mechanisms have been an important influence on climate.”

This study has shown clearly that through the use of high precision exposure ages for Holocene moraines, a much more accurate chronology has been produced to show interhemispheric differences in glacial extents. Whilst a definitive answer hasn’t yet been given; the issue of what drove these regional Holocene climate variations is much clearer as a result of a greatly improved knowledge of southern hemisphere glacier fluctuations.

All of this helps us to greater understand what influenced glaciers in the past and how for the future, we may be able to apply our more complex knowledge of regional and hemispheric influences to better predict future levels of retreat across the globe.