Correlations Between Sea-Level Components Are Driven by Regional Climate Change

The accurate quantification of uncertainties in regional sea-level projections is essential
for guiding policy makers. As climate models do not currently simulate total sea level, these uncertainties
must be quantified through summation of uncertainties in individual sea-level components. This
summation depends on the correlation between the components, which has previously been prescribed or
derived from each individual component's dependence on global mean surface temperature. In this study,
we quantify, for the first time, regional correlations between sea-level components based on regional
climate change projections. We compute regional sea-level projections consistent with climate projections
from an ensemble of 14 Earth System Models. From the multi-model spread, we estimate the uncertainty
in the regional climate's response to greenhouse forcing. To quantify the total uncertainty, we add the
uncertainty in the response of sea-level components to this regional climate change. This approach
reveals how regional climate processes impose correlations between sea-level components, affecting the
total uncertainty. One example is an anti-correlation between North Atlantic sterodynamic change and
Antarctic dynamic mass loss, suggesting a teleconnection established by the large-scale ocean circulation.
We find that prescribed correlations, applied in the fifth assessment report of the Intergovernmental Panel
on Climate Change, lead to a global overestimation in the uncertainty in regional sea-level projections on
the order of 20%. Regionally, this overestimation exceeds 100%. We conclude that accurate uncertainty
estimates of regional sea-level change must be based on projections of regional climate change and cannot
be derived from global indicators such as global mean surface temperature.