The IPCC sea level numbers

Sea level rise as observed (from Church and White 2006) shown in red up to the year 2001, together with the IPCC (2001) scenarios for 1990-2100. See second figure below for a zoom into the period of overlap.

The TAR showed sea level rise curves for a range of emission scenarios (shown in the Figure above together with the new observational record of Church and White 2006). The range was based on simulations with a simple model (the MAGICC model) tuned to mimic the behaviour of a range of different complex climate models (e.g. in terms of different climate sensitivities ranging from 1.7 to 4.2 ºC), combined with simple equations for the glacier and ice sheet mass balances (“degree-days scheme”). This model-based range is shown as the grey band (labelled “Several models all SRES envelope” in the original Figure 5 of the TAR SPM) and ranged from 21 to 70 cm, while the central estimate for each emission scenario is shown as a coloured dashed line. The largest central estimate of sea level rise is for the A1FI scenario (purple, 49 cm).

In addition, the dashed grey lines indicate additional uncertainty in ice sheet behaviour. These lines were labelled “All SRES envelope including land ice uncertainty” in the TAR SPM and extended the range up to 88 cm, adding 18 cm at the top end. One has to delve deeply into the appendix of Chapter 11 of the TAR to find out what these extra 18 cm entail: they include a “mass balance uncertainty” and an “ice dynamic uncertainty”, where the latter is simply assumed to be 10% of the total computed mass loss of the Greenland ice sheet. Note that such an ice dynamic uncertainty was only included for Greenland but not for Antarctica; instability of the West Antarctic Ice Sheet, a scenario considered “very unlikely” in the TAR, was explicitly not included in the upper limit of 88 cm.

As we mentioned in our post on the release of the SPM, it is apples and oranges to say that IPCC reduced the upper sea level limit from 88 cm to 59 cm, as the former included “ice dynamic uncertainty” (albeit only for Greenland, as rapid ice flow changes in Antarctica were considered too unlikely to bother at the time), while the latter discusses this ice flow uncertainty separately in the text, stating it could add 10 cm, 20 cm or even more to the 59 cm in the table.

So is it better to compare the model-based range 21 – 70 cm from the TAR to the 18 – 59 cm from the AR4? Even that is apples and oranges. For one, TAR cites the rise up to the year 2100, the AR4 up to the period 2090-2099, thus missing the last 5 years (or 5.5 years, but let’s not get too pedantic) of sea level rise. For 2095, the TAR projection reduces from 70 cm to 65 cm (the central estimate for A1FI reduces from 49 cm to 46 cm). Also, the TAR range is a 95% confidence interval, the AR4 range a narrower 90% confidence interval. Giving the TAR numbers also as 90% ranges shaves another 3 cm off the top end.

Sounds complicated? There are some more technical differences… but I will spare you those. The Paris IPCC meeting actually discussed the request from some delegates to provide a direct comparison of the AR4 and TAR numbers, but declined to do this in detail for being too complicated. The result was the two statements:

The TAR would have had similar ranges to those in Table SPM-3 if it had treated the uncertainties in the same way.


For each scenario, the midpoint of the range in Table SPM-3 is within 10% of the TAR model average for 2090-2099.

(In fact delegates were told by the IPCC authors in Paris that with the new AR4 models, the central estimate for each scenario is slightly higher that with the old models, if numbers are reported in a comparable manner.)

The bottom line is thus that the methods have significantly improved (which is the reason behind all those methodological changes), but the expectation of how much sea level will rise in the coming century has not significantly changed. The biggest change is that ice sheet dynamics look more uncertain now than at the time of the TAR, which is why this uncertainty is not included any more in the cited range but discussed separately in the text.

Critique – Could these numbers underestimate future sea level rise?

There’s a number of issues worth discussing about these sea level numbers.

Page 3 of 6 | Previous page | Next page