Larsen C Ice Shelf Calving

This is a quick final post to summarise the recent developments of the Larsen C Ice Shelf on the AntarcticPeninsula, which is monitored by the British Antarctic Survey. Many media outlets are reporting the imminent calving of an iceberg up to 5,000 km^2,  (Figure 1) e.g. the BBC.  It is only retained by a final 20 km of ice. It is also reported in the Guardian, which includes a helpful video. As it so recent, there is no peer-reviewed literature on this development.

Figure 1: Map of the Antarctic Peninsula, showing the current position of the rift.

Source: BBC 

Although this will not directly contribute to sea level rise (SLR), floating ice shelves such as Larsen C are important for buttressing marine-terminating glaciers, which transport land-based ice to the ocean. Therefore, it is likely the eventual collapse of Larsen C will indirectly contribute to SLR, on a longer time scale.

Images of the rift taken from the air show the enormous scale of the ice berg:

 

Figure 2

Source: British Antarctic Survey 

Figure 3

Source: BBC 

Guardian Editorial

This is a link to a relevant article in The Guardian on the 27th of December.

Although it is only brief, the article highlights an important, often under-represented environmental impact of warming Arctic temperatures - the plight of the Arctic reindeer population. This piece discusses the inherent fragility of the interdependence of human and reindeer populations in the Arctic. Whilst rising Arctic temperatures can be accurately monitored, the intangible cultural ramifications mentioned here are often ignored. Anomalously warm Arctic temperatures this year were highlighted in a previous post on this blog.

Figure 1: Reindeer and human populations in northern Sweden.

Source: The Guardian

The Climate Change section of The Guardian website is a useful resource, with articles on a diverse range of topics. These are discussed in a broad context in a less academic style. A particularly interesting recent focus of this site is a series of wider environmental, economic and cultural articles in light of The Warming Arctic.

Above and Below the Pine Island Glacier

A key technique to calculate mass loss, and consequential Sea Level Rise (SLR) contribution, from Antarctica is to measure the ice flow from individual outlet glaciers. The Antarctic glacier which has received the greatest scientific (and media e.g. BBC) attention is a marine-terminating glacier named the Pine Island Glacier (PIG), which drains into the Amundsen Sea (AS). PIG is located on the coast of the West Antarctic Ice Sheet (WAIS) (shown in Figure 1), in a region called the Amundsen Sea Embayment (ASE).

Figure 1: Map of Antarctica with the location of PIG labelled.

Source: RealClimate 

PIG has been studied frequently because observation suggests it is flowing at a much greater rate than other Antarctic glaciers. As it drains a large portion of the WAIS, PIG has a significant potential contribution to SLR. The exceptional retreat of the PIG can be explained by its physical characteristics. PIG is grounded below sea level, on a bedrock topography which retreats inland. This means that the PIG is especially sensitive to changes in ocean temperature.

This post will discuss two papers from 2016 – (1) Smith et al. and (2) Konrad et al. – both of which aim to reconstruct a long term historical record of the behaviour of PIG, although they use very different methods.

Smith et al. 2016

This study addresses an important weakness of the satellite observational record. Satellite records only began in 1973, hence it is insufficient to analyse long-term temporal variation of PIG .

Smith et al. took sediment cores from three bedrock locations below the PIG (labelled as A, B, and C on Figure 2). A combination of biological proxies and isotope dating are used to reconstruct ice extent over a much longer period.

Results from this analysis indicate three separate states (Figure 2) of the PIG. In its earliest state, ice from the PIG is consistently in contact with the bedrock up to the maximum of the ridge (Figure 2, panel a). However, in 1945 (± 12 years) an ocean cavity developed between the ice and the underlying bedrock (Figure 2, panel b). The authors attribute the growth of this cavity to warm ocean temperatures caused by El Niño activity. By 1970 (± 4 years) the PIG was completely detached from the bedrock ridge, in a state of continual retreat (Figure 1, panel c).

Key findings from Smith et al. (2016), enabled by greater time-period of analysis:

1.  Significant retreat of the PIG began approximately in the 1940s.
2.    This retreat has continued up to present, despite the removal of the initial forcing (El Niño).

Figure 2: ‘Processes and sedimentation beneath the PIG ice shelf’. Caption quoted from source.

Source: Smith et al. 2016, Figure 3, top 3 panels.

Konrad et al. (accepted 2016, yet to be published):

This study observes the evaluation of the PIG over a shorter time-period, from 1992 to the present. In contrast to the previous study, Konrad and colleagues combine five satellite altimetry datasets to analyse the ASE (including PIG) from above. The high spatial resolution of data in this study enables in-depth comparison between glaciers on the ASE.

Altimetry data is used specifically to measure changes in glacier elevation. In turn, this can be used to determine the net amount of ice discharged by each glacier into the ocean (equivalent to SLR contribution). The term ‘ice-dynamical imbalance’ is used to describe rate of thinning and consequent mass loss.

Results from Konrad et al. indicate that PIG has thinned continually throughout the period of observations (Figure 3). Figure 3 shows an important trend in the PIG over the satellite observational period. The rate of surface lowering has increased up the glacier over time, at an approximately linear rate. This means that the point of a given rate of surface lowering has moved higher up the glacier over time (i.e. the red and blue lines on Figure 3). In addition, the rate of surface lowering close to the grounding line has increased over time. Thinning rate has spread up the PIG at a rate of approximately 13 km yr^-1 (red and blue lines), which is approximately double the equivalent rates for the nearby Pope, Smith, and Kohler glaciers.

Key findings from Konrad et al. (2016):

1.  PIG surface-lowering rate has increased continually (linearly) from 1992 to the present.
2.  PIG thinning rate is notably greater than surrounding glaciers on the ASE.

Figure 3: ‘Temporal evolution of surface-lowering rates. Distance is taken from the grounding line’. Caption quoted from the source.

Source: Konrad et al. 2016, Figure 2, panel A.

To summarise, both studies analyse the temporal evolution of the PIG. However, they use highly contrasting forms of data – and therefore study the glacier at different time scales. Despite these varying methodologies, the pattern shown in both studies is similar – an increasing PIG contribution to SLR. The current diversity of research in the polar regions is undoubtedly beneficial to improving estimates of SLR from Antarctica. 

2016 Update to GIS Mass Loss: McMillan et al. 2016

A recent study by McMillan et al. (2016) extends the comprehensive, inter-methodology comparison used to calculate SLR from the GIS (as discussed in the previous post). Focussing on CryoSat-2 altimetry and GRACE gravimetric measurements, this paper finds a strong correlation (R = 0.96) between mass loss from both techniques (Figure 1). Mean linear trend of these datasets is 269 ± 51 Gt yr^-1, corresponding to 0.74 ± 0.14 mm yr^-1 SLR between 2011 and 2014. This is almost identical to the previously discussed 2005-2010 estimate from Sheperd et al. (2012).

Figure 1: ‘Greenland mass evolution. Monthly evolution in ice sheet mass since 2003 from GRACE gravimetry (green) and since 2011 from CryoSat-2 altimetry and firn modelling (blue). The CryoSat-2 time series has been referenced to the GRACE data at the start of 2011. The inset shows the correspondence between GRACE and CryoSat-2 monthly estimates of mass evolution since 2011 (solid blue dots), together with a linear regression (solid blue line), the regression slope, and the Pearson correlation coefficient, R. The dashed line indicates equivalence, although the GRACE results include, additionally, mass changes of peripheral ice caps and unglaciated regions’. Caption quoted from source.

Source: McMillan et al. (2016), Figure 2.

A strength of this study is the high spatial (5 km) and temporal (monthly) resolution of the data. This enables the regions of the GIS which contribute the greatest amount to SLR to be identified with a higher degree of certainty than before. Highest mass loss is found at low latitudes, in coastal areas containing outlets glaciers. In particular, the authors observe that the South West region of the GIS is the region with the greatest contribution (approximately 41% of total ice mass loss between 2011 and 2014). The spatial pattern of mass loss measured by RA is visualised in the video below:

Source: European Space Agency 

Credit: Planetary Visions

To conclude, building on a previously developed approach, updated data corroborates IPCC estimates of SLR from the GIS. This study by McMillan et al. (2016) indicates that the linear trend of mass loss from the GIS observed  throughout the 21^st century has continued. 

Greenland Ice Sheet Contribution to Sea Level Rise

A critically important area of research in the polar regions is the contribution of ice sheets to Sea Level Rise (SLR) (IPCC 2013 Chapter 13). Accurate calculations of ice mass loss from large geographical areas of the climate system such as the Greenland Ice Sheet (GIS), West Antarctic Ice Sheet (WAIS) or the East Antarctic Ice Sheet (EAIS) requires substantial international cooperation, hence studies of this magnitude can only be undertaken every few years.

In light of the upcoming IPCC 6th Assessment Report, a second phase of the programme designed to update our understanding of this issue is being undertaken, named the Ice sheet Mass Balance Inter-comparison Exercise (IMBIE).

The most update research indicates that if the GIS were to melt entirely, sea level would rise by 7.36m. This post will summarise our contemporary knowledge of one of the most challenging research areas in polar science – specifically ice mass loss from the GIS (SLR from Antarctica will not be discussed in this post).

Greenland Ice Sheet contribution to Sea Level Rise in IPCC AR5

Observed global mean SLR from 1900 to 2013 is shown in Figure 1. The IPCC estimate for SLR between 1901 and 2010 is 0.19 (0.17 to 0.21) m (IPCC 2013 pp. 11). Rate of SLR is believed to have increased over this time, culminating in an estimated rate of 3.2 (2.8 to 3.6) mm yr^-1 between 1993 and 2010. This is also the period with the greatest certainty in observations of SLR due to increased accuracy and coverage of satellite observations. For the same period, GIS contribution is estimated as 0.33 (0.25 to 0.41) mm yr^-1 – approximately 10%.

Figure 1: ‘Global mean sea level relative to the 1900-1905 mean of the longest running dataset, and with all datasets aligned to have the same value in 1993, the first year of satellite altimetry data. All time-series (coloured lines indicating different data sets) show annual values, and where assessed, uncertainties are indicated by coloured shading’. Caption quoted from the source.

Source: IPCC 2013, Figure SPM.3 Panel (d).

Whilst the loss of sea ice does not contribute to SLR, the loss of land based ice does contribute directly to SLR (Figure 2). Variations in the mass of the entire GIS are used to keep track of ice exchange between the ocean and land.  Figure 2 is taken from a comprehensive study by Sheperd et al. (2012), which shows the conversion between mass loss and SLR from the GIS (Figure 2 also show the contribution of Antarctica). Sheperd et al. (2012) estimate that the GIS lost 2700 ± 930 Gt of ice between 1992 and 2011, although as the time series in Figure 2 shows, the rate of this change is not linear and increases substantially over time. Linear mass loss between 1992 and 2000 is calculated to be 51 ± 65 Gt yr^-1, compared to 263 ± 30 Gt yr^-1 between 2005 and 2010 (Sheperd et al. 2012).

Figure 2: ‘Cumulative changes in the mass of (left axis)… …GrIS and AIS and the combined change of the AIS and GrIS (bottom), determined from a reconciliation of measurements acquired by satellite RA, the IOM, satellite gravimetry, and satellite LA (Lidar Altimetry). Also shown is the equivalent global sea-level contribution (right axis), calculated assuming that 360 Gt of ice corresponds to 1mm of sea-level rise’. Caption quoted from the source.

Source: Sheperd et al. 2012, Figure 5, bottom panel.

Methodology for Calculating Mass Loss

There are a variety of methods used to calculate the mass loss of an ice sheet on the scale of the GIS. Only by synthesising these methods can a robust estimate of total ice loss be attained. Here is a quick overview of each method:

• Input-Output Method (IOM): Increases and decreases in ice are calculated separately – allowing them to be analysed discretely. Ice mass loss is calculated at the catchment basin scale using direct observations of sublimation, meltwater and glacier outflow  (). Snowfall (accumulation – ice mass increase) is derived from regional climate models.

• Radar Altimetry (RA) and Lidar Altimetry (LA): Satellites (specifically the European Space Agency’s (ESA) CryoSat-2) and aeroplanes are used to measure variations in the surface elevation of the GIS. Changes in ice sheet volume are converted to changes in ice sheet mass using models of ice density.

• Gravimetry (GRACE satellite): The Gravity Recovery And Climate Experiment (GRACE) satellite measures the gravitational force from the GIS, therefore directly measuring changes in ice mass. A significant source of uncertainty for this method is correcting for Glacial Isostatic Adjustment (GIA) of the underlying crust.

Sheperd et al. (2012) compares these methods in Figure 3. A key weakness of the combined dataset is inconsistency between the time periods of each methodology, which can be seen in Figure 3. The IOM has much greater uncertainty bounds than the alternative techniques, and estimates a greater rate of ice mass loss. Each method calculates a different value for ice mass loss, and the methods are all subject to considerable uncertainty. However, using multiple methods is critical to distinguishing robust trends in mass loss over the Anthropocene – and estimating GIS contribution to SLR.

Figure 3: Rate of mass change of the GIS for three of the methods described, including uncertainty. Rates of mass balance derived from ICESat LA were computed as time varying trends. The gravimetry and RA mass trends were computed after applying a 13-monthmoving average to the relative mass time series. Caption adapted from source.

Source: Sheperd et al. 2012, Figure 4, top-right panel.

To summarise, SLR from the GIS is one of the greatest challenges facing science in the polar regions. IPCC observational ensemble estimates show a substantial and increasing annual SLR contribution from the GIS throughout the Anthropocene. 

Arctic Sea Ice Update

This is a quick update on the current state of Arctic sea ice. It comes in light of widespread reports, including The Guardian, The Financial Times and The Independent, of significantly anomalous climate patterns for this time of year. Climate blogs, such as Arctic News and Union of Concerned Scientists have offered an immediate analyse of this data, which is summarised below.

November 2016 Arctic Sea Ice Anomaly: Notably, current temperatures across the Arctic are exceedingly high compared to previous years (Figures 1 and 2). Temperatures anomalies in some parts of the Arctic are as great as 20°C (Figure 1). However, the most recent figures (Figure 2) shows that peak warming in the Arctic for November appears to have been passed after a rapid decline. Nevertheless, this means current Arctic temperatures are over 10°C greater than 1958-2002 averages (Figure 2).

Figure 1: Temperature anomalies across the Arctic region for 25 November 2016, relative to 1979-2000. Source: Climate Reanalyzer. Accessed 25 November 2016.

Figure 2: Annual time series of daily Arctic (greater than 80° latitude North) temperature for the year 2016 (red line) and the 1958-2002 mean (green line). Temperature of freezing (i.e. 0°C) is shown for reference (blue line). Source: Danish Meteorological Institute. Accessed 25 November 2016.

The most up to date explanation indicates that these temperatures are caused by perturbed atmospheric patterns. An increased strength jet stream has caused a greater atmospheric transport of warm air to high northern latitudes – creating these large temperature anomalies.

Significantly, increased November Arctic temperatures have triggered a pattern previously unobserved at this time of year – declining Arctic sea ice. As mentioned in a previous post, the observed Arctic sea ice extent annual cycle typically reaches a minimum in September, before growing throughout the Northern Hemisphere autumn and winter. Most recent data (Figure 3) indicate a greatly reduced growth rate, resulting in the lowest ever recorded Arctic sea ice extent for this time of year. Between 16 November 2016 and 20 November 2016, Arctic sea ice extent decreased from 8.674 million km² to 8.625 million km² (Figure 3). Since then, sea ice extent has increased, although it remains significantly below both the 1981-2010 average and 2012, the year with the lowest sea ice minimum extent, for the end of November.

Figure 3: Daily Arctic sea ice extent (area of ocean with at least 15% sea ice) for 2016 (solid blue line) and 1992-2010 mean (solid black line) shown with ±2 Standard Deviations. Sea ice extent for 2012, the year with the lowest minimum sea ice extent is also shown for reference (dashed green line). Source: National Snow and Ice Data Center, updated as of 27 November 2016. Accessed: 28 November 2016.

Academic Context: As these observations are so recent, there is no peer reviewed literature on the extreme warming in November 2016 (although as mentioned earlier this topic has received wide coverage in newspapers and blogs). Hence it is important to put these findings in context of published academic research.

Slow sea ice growth during Northern Hemisphere Autumn is likely to be caused by the culmination of a multidecadal sea ice decline. Climate theory and model experiments indicate that a positive ice-albedo feedback mechanism exacerbates a prolonged period of decline. Hansen et al. (1984) produced the earliest comprehensive analyse of the strength of the sea ice albedo feedback. In this paper, the sea ice feedback is attributed a feedback factor of approximately 1.1 (see Hansen et al. 1984, Equation 4). This essentially means that for a change in mean global temperature, this feedback will contribute a further 10% temperature change in the same direction.  This therefore identifies the sea ice-albedo feedback as one of the most sensitive feedbacks in the climate system.

A renowned paper by Lenton et al.in 2007 highlights the behaviour of sea ice as a threshold response – or ‘tipping point’. The authors define a tipping point as thresholds in the climate system, which when surpassed, trigger a rapid change to a distinct qualitative state. Lenton et al. estimate that Arctic sea ice as part of the climate system which could be subject to a tipping point within 0.5°C to 2.0°C of global warming, and could operate on an approximately 10-year timescale, making it one of the most sensitive parts of the climate system. Recent observed nonlinear behaviour if Arctic sea ice supports this theory and indicates that this particular tipping point may have already been passed.

Finally, sea ice thickness is a factor that has recently received increased attention considering its importance relating to changes in sea ice extent. The 2013 IPCC report includes analysis of sea ice thickness changes, acknowledging that the loss of older, thicker ice contributes to a significantly reduced sea ice minimum extent. This report (pp. 319) indicates a likely average decline in thickness of between 1.3m and 2.3m (between 1980 and 2008). Importance of sea ice thickness is also emphasised in Video 2 of the previous post on this blog. It is important to note that changes in thickness do not directly contribute to the sea ice-albedo feedback, which is dependent on surface area coverage. However, the critical importance of sea ice thickness is being recognised, exemplified by a recent paper by Xieet al. (2016). In the study, sea ice thickness is observed and incorporated into an Arctic sea ice forecasting model for the first time. This is an important advance in sea ice forecasting, as understanding of the processes contributing to sea ice decline improves.

Arctic Sea Ice Visuals

In light of the Arctic sea ice data release discussed in a previous post, several interesting visual tools on the topic have been produced. I have compiled several here, which demonstrate the wide range of ways in which geographical data can be presented:

(1) This first video is taken from the ThinkProgress blog. It shows minimum sea ice volume at an annual resolution (1979 to 2016). The video effectively shows one of them most critical variables associated with the polar regions (it is important to note this shows sea ice volume, not sea ice extent).

(2) This satellite imagery is from the NSIDC and shows monthly averaged sea ice extent for every September (1979 to 2014). Although it is the closest of these visualisations to raw data (showing the images used to calculate the data in Figure 1 of my initial post on the Arctic sea ice minimum) it is limited in the sense that it only shows annual minimum sea ice extent.

(3) A much more detailed approach is this detailed 3D visualization from NASA. This includes the age of sea ice – a factor representative of thickness – as well as a helpful audio explanation. It presents data at a monthly resolution (1984 to 2016) hence showing the seasonal cycle of sea ice. For more detail see the source.

(4) Finally, a much less conventional but none the less informative visualisation is this climate spiral produced by Ed Hawkins at the University of Reading. It captures both the seasonal cycle and the overall (1979 to 2016) decline of Arctic sea ice volume throughout the Anthropocene.