mass balance can be computed by a climate model (RCM or
AOGCM), but frequently, ablation is still determined from
an empirical formulation that relates the number of
positive-degree-days (integral of positive temperatures) to
the snow and ice melting rates. The main drawbacks of this
approach is that it does not account for albedo changes
and it has been validated on only a few Greenland sites for
the present-day period. The mass balance derived from such
methods is generally more uncertain than that obtained using
regional models. Another difficulty is linked to the difference
of resolution between climate and ice-sheet models. As
temperature, precipitation and surface mass balance are
highly dependent on topography, the development of
appropriate downscaling techniques is required to account
for the effect of altitude.
In addition, researchers still face many challenges in representing some processes related to rapid dynamics, such as
those occurring at the base of the ice sheet or at the ice-ocean
interface, where small disturbances seem capable of triggering
strong instabilities, which can propagate up to a thousand
kilometers upstream, and lead to a destabilization of the entire
ice sheet. However, understanding the impact of small-scale
processes on the large scale is still in its infancy. Over the last
two decades, observations have shown an increase in the flow
of outlet glaciers, suggesting that these could respond much
more quickly than previously expected to variations in
atmospheric and oceanic conditions, making possible a significant retreat of the ice sheet in the more or less distant
future. The physical laws governing ice flow in these glaciers
still remain poorly known. For example, although the marine
ice-sheet instability is increasingly well represented in new
generation ice-sheet models, there is no consensus on
hydro-fracturing yet. Large uncertainties also exist on the
physical laws governing iceberg calving and basal melting
under the ice shelves. A last, but not least, problem is related to
the initial state of the ice sheets. Indeed, starting future
short-term (a few centuries) simulations from a realistic
present-day state of the ice sheet is of primary importance to
avoid as much as possible spurious biases in the results.
Getting an accurate initial state is challenging however
because of the scarcity of observations on vertical velocity and
temperature profiles, and also on basal conditions (e.g. frozen
or thawed bed areas, bedrock topography). Several approaches have been developed ranging from long free-evolving
simulations over one or several glacial-interglacial cycles to
more formal optimization approaches often based on inversion techniques, with various target criteria (either surface
velocities or topography in agreement with observations, or
internal properties accounting for the past history of ice
sheets). Each of these techniques presents some advantages
and drawbacks, but their efficiency should improve in the
future as the number of observations grows. Recent remote
sensing observations made over Greenland and Antarctica are
expected to refine numerical simulations and thus to improve
the relevance of future forecasting models. Finally, part of the
uncertainty in the future ice-sheet responses comes from
uncertainty in the climate forcing itself. This is due to our lack
of knowledge in the future socio-economic pathways and to
the biases of the climate models. Moreover, climate-ice sheet
interactions are still very rarely taken into account in the
models. Accounting for these interactions may have the
potential to strongly modify not only the simulated climate
and ice-sheet responses but also sea-level projections (Bronselaer et al. 2018; Golledge et al. 2019).
Sea Level Change
We have seen, in the first part of this chapter, that at the scale
of a few decades, the main causes of sea-level variations are
due, on the one hand, to the thermal expansion of the ocean
(in the case of a warming climate), and on the other hand, to
changes in the mass of water in the ocean due to exchanges
with continental reservoirs. Based on the results from 21
AOGCMs, the sea-level projections reported in the
Fifth IPCC report (IPCC 2013), all show an increase in
sea-level by the end of the twenty-first century (2081-2100)
compared to the 1986–2005 period, with an average global
mean sea-level rise ranging from 0.28 to 0.61 m for RCP2.6
and from 0.52 to 0.98 m for RCP8.5. For the RCP8.5 scenario, the rate of sea-level rise increases throughout the
twenty-first century, going from 3.2 ± 0.4 mm/yr (for the
1993–2010 period) to 11.2 mm/yr (2081–2100), whereas, in
the RCP2.6 scenario, the rate increases up to 4.4 mm/yr
around 2050 and slightly declines in the second half of the
century.
According to the Fifth IPCC report (IPCC 2013), thermal
expansion remains the main contributor (30 to 55%) to
global mean sea-level rise over the twenty-first century with
median rates of 0.14 ± 0.04 m (RCP2.6) and 0.27 ± 0.04
m (RCP8.5) in 2081–2100. These contributions can be
estimated from changes in ocean heat uptake increasing
roughly linearly with global mean surface air temperatures
simulated by the AOGCMs. It should be noted that AOGCM
simulations do not include volcanic forcing which may
reduce the projected contribution of thermal expansion.
The AOGCM simulations forced by the different RCP
scenarios highlight a spatial variability of sea level
(Fig. 31.16). As they do not take into account the freshwater
fluxes coming from ice sheets, this regional variability,
previously observed in satellite measurements over the 1993–
2010 period, is linked to a variability in the atmosphereocean-sea-ice system. More specifically, regional variability
can be due not only to variations in temperature and salinity
(i.e. precipitation/evaporation ratio) and therefore density, but
also to variations in ocean and atmospheric circulations
472
S. Charbit et al.
AOGCM), but frequently, ablation is still determined from
an empirical formulation that relates the number of
positive-degree-days (integral of positive temperatures) to
the snow and ice melting rates. The main drawbacks of this
approach is that it does not account for albedo changes
and it has been validated on only a few Greenland sites for
the present-day period. The mass balance derived from such
methods is generally more uncertain than that obtained using
regional models. Another difficulty is linked to the difference
of resolution between climate and ice-sheet models. As
temperature, precipitation and surface mass balance are
highly dependent on topography, the development of
appropriate downscaling techniques is required to account
for the effect of altitude.
In addition, researchers still face many challenges in representing some processes related to rapid dynamics, such as
those occurring at the base of the ice sheet or at the ice-ocean
interface, where small disturbances seem capable of triggering
strong instabilities, which can propagate up to a thousand
kilometers upstream, and lead to a destabilization of the entire
ice sheet. However, understanding the impact of small-scale
processes on the large scale is still in its infancy. Over the last
two decades, observations have shown an increase in the flow
of outlet glaciers, suggesting that these could respond much
more quickly than previously expected to variations in
atmospheric and oceanic conditions, making possible a significant retreat of the ice sheet in the more or less distant
future. The physical laws governing ice flow in these glaciers
still remain poorly known. For example, although the marine
ice-sheet instability is increasingly well represented in new
generation ice-sheet models, there is no consensus on
hydro-fracturing yet. Large uncertainties also exist on the
physical laws governing iceberg calving and basal melting
under the ice shelves. A last, but not least, problem is related to
the initial state of the ice sheets. Indeed, starting future
short-term (a few centuries) simulations from a realistic
present-day state of the ice sheet is of primary importance to
avoid as much as possible spurious biases in the results.
Getting an accurate initial state is challenging however
because of the scarcity of observations on vertical velocity and
temperature profiles, and also on basal conditions (e.g. frozen
or thawed bed areas, bedrock topography). Several approaches have been developed ranging from long free-evolving
simulations over one or several glacial-interglacial cycles to
more formal optimization approaches often based on inversion techniques, with various target criteria (either surface
velocities or topography in agreement with observations, or
internal properties accounting for the past history of ice
sheets). Each of these techniques presents some advantages
and drawbacks, but their efficiency should improve in the
future as the number of observations grows. Recent remote
sensing observations made over Greenland and Antarctica are
expected to refine numerical simulations and thus to improve
the relevance of future forecasting models. Finally, part of the
uncertainty in the future ice-sheet responses comes from
uncertainty in the climate forcing itself. This is due to our lack
of knowledge in the future socio-economic pathways and to
the biases of the climate models. Moreover, climate-ice sheet
interactions are still very rarely taken into account in the
models. Accounting for these interactions may have the
potential to strongly modify not only the simulated climate
and ice-sheet responses but also sea-level projections (Bronselaer et al. 2018; Golledge et al. 2019).
Sea Level Change
We have seen, in the first part of this chapter, that at the scale
of a few decades, the main causes of sea-level variations are
due, on the one hand, to the thermal expansion of the ocean
(in the case of a warming climate), and on the other hand, to
changes in the mass of water in the ocean due to exchanges
with continental reservoirs. Based on the results from 21
AOGCMs, the sea-level projections reported in the
Fifth IPCC report (IPCC 2013), all show an increase in
sea-level by the end of the twenty-first century (2081-2100)
compared to the 1986–2005 period, with an average global
mean sea-level rise ranging from 0.28 to 0.61 m for RCP2.6
and from 0.52 to 0.98 m for RCP8.5. For the RCP8.5 scenario, the rate of sea-level rise increases throughout the
twenty-first century, going from 3.2 ± 0.4 mm/yr (for the
1993–2010 period) to 11.2 mm/yr (2081–2100), whereas, in
the RCP2.6 scenario, the rate increases up to 4.4 mm/yr
around 2050 and slightly declines in the second half of the
century.
According to the Fifth IPCC report (IPCC 2013), thermal
expansion remains the main contributor (30 to 55%) to
global mean sea-level rise over the twenty-first century with
median rates of 0.14 ± 0.04 m (RCP2.6) and 0.27 ± 0.04
m (RCP8.5) in 2081–2100. These contributions can be
estimated from changes in ocean heat uptake increasing
roughly linearly with global mean surface air temperatures
simulated by the AOGCMs. It should be noted that AOGCM
simulations do not include volcanic forcing which may
reduce the projected contribution of thermal expansion.
The AOGCM simulations forced by the different RCP
scenarios highlight a spatial variability of sea level
(Fig. 31.16). As they do not take into account the freshwater
fluxes coming from ice sheets, this regional variability,
previously observed in satellite measurements over the 1993–
2010 period, is linked to a variability in the atmosphereocean-sea-ice system. More specifically, regional variability
can be due not only to variations in temperature and salinity
(i.e. precipitation/evaporation ratio) and therefore density, but
also to variations in ocean and atmospheric circulations
472
S. Charbit et al.