History of US sea level rise scenarios

The range was set in 2012 and justification became untethered

Photo by Alex Diaz on Unsplash

This post lays out the history of the federal global mean sea level (GMSL) rise scenarios to 2100.

It is written in preparation for taking a look at the recently released North Carolina Science Panel sea level draft report recently released.

TLDR: The scenarios have not actually changed much since 2012. Rather, what has changed is their justification which has become increasingly untethered. The 2.0m scenario harks back to 2008 and required assumptions of rapid acceleration to extremely high limits.

The 2022 Sea Level Rise Technical Report is the most recent iteration and builds on three previous works and I look at these individually. These are:

1. Global Sea Level Rise Scenarios for the US National Climate Assessment (Parris et al 2012)

2. Regional Sea Level Scenarios for Coastal Risk Management (Hall et al 2016)

3. Global and Regional Sea Level Rise Scenarios for the US (Sweet et al 2017)

All of the reports except for #2 were developed to feed into the National Climate Assessment (NCA) process.

Report #1 and #2 were multi-agency reports but it was not developed by a “Task Force” per se. The main audience for Report #2 was the Department of Defense.

Report #3 and the 2022 Report were developed by White House choreographed Task Forces.

Here is a summary chart of the scenarios from 2012, 2017, and 2022. The 2016 report (#2) established the 0.5m delineations but the justification for the range did not carry over to 2017.

#1, The 2012 Report

The 2012 Report created four global mean SLR scenarios for a range of 0.2m- 2m.

The report was clear that there were no probabilities associated with any of their SLR scenarios: “Probabilistic projections of future conditions are another form of scenarios not used in this report.”

Each SLR scenario represents different degrees of ocean warming and ice sheet loss. However, the SLR scenarios assume different socio-economic assumptions from which to develop emissions projections, too. The expectation was that users would approach the scenarios based on their individual risk tolerance.

The four scenarios are shown above and there descriptions are as follows:

1.Highest: For situation with little risk tolerance such as long life cycle infrastructure

Derived from a combination of estimated ocean warming from the IPCC AR4 global SLR projections and a calculation of the maximum possible glacier and ice sheet loss by the end of the century. Specifically, the scenario is drawn from Pfeffer et al 2008, which finds 2.0m by 2100 “but only if all variables are quickly accelerated to extremely high limits.”

2.Intermediate-High: For situations where assessment of risk from limited ice sheet loss is of interest

Based on an average of the high end of semi-empirical, global SLR projections¹. The studies used for this scenario are all based on the A2 storylines described as below.

3.Intermediate- Low: For situations where assessment of risk from ocean warming is of interest

Based on the upper end of IPCC AR4 global SLR projections resulting from climate models using the B1 emissions scenario as described below.

4.Lowest: For use when there is a great tolerance for risk

Based on a linear extrapolation of the historical SLR rate derived from tide gauge records beginning in 1900

The authors of the 2012 Report concluded that they have:

very high confidence that global mean sea level will rise at least 0.2 meters (8 inches) and no more than 2.0 meters (6.6 feet) by 2100.

The authors are saying that global mean SLR by 2100 falls somewhere between a projection based on observations and the highest SLR estimate available in the climate change literature which is only possible assuming extreme acceleration.

Hence their felt certainty.

#2, The 2016 Report

The 2016 Report created five SLR scenarios that were “bounded” on the low end and high end by those produced in 2012.

The five scenarios are: Lowest 0.2m; Low 0.5m; Medium 1.0; High 1.5m; and Highest 2.0. The trajectories of each are shown below.

The 2016 report spent a great deal of time emphasizing the lack of probability that could be assigned to any of the scenarios it provided arguing at the start that,

…a variety of uncertainties, including the uncertainties associated with human behaviors (i.e., emissions futures), limit the predictive capabilities of climate-related sciences…The appropriate application of the scenarios in a context-dependent manner is often more important than the quantitative scenario.

Because of this, the authors explain, instead of implying accuracy to intermediate scenarios they simply create equally proportional 0.5 m increments between the low and the high.

Then, in “acknowledg[ment] that some decision-makers will still want to see some correspondence” with the IPCC reports, the Task Force authors align the intermediate scenarios with the RCP emission pathways used at the time. Interestingly, the authors go through a fair amount of discussion about the history of RCP development and what they mean.

The authors suggest that SLR scenario users can “constrain their tolerance for risk based on their view of society’s decisions regarding future emissions.”

This shifts the risk aversion advice from a focus on uncertainty in geophysical processes to uncertainty about socioeconomic processes.

The result is Table 3.3 shown below.

• The “lowest” scenario- derived from a linear extrapolation of the observed record- also happens to be in the range of those estimates produced using RCP 2.6.

• The low scenario- now the first 0.5m increment- happens to be in the range of estimates produced using: RCP 2.6, RCP4.5, and RCP8.5

  • In 2012 the 0.5m scenario was derived from the former B1 scenario which is akin to RCP 4.5

• The next two increments (1m and 1.5m) happen to be in the range of RCP8.5 estimates.

  • In 2012 the 1.2m scenario was derived from the A2 scenario which is akin to RCP8.5.

• Finally, the highest scenario, the upper bound, originally derived from Pfeffer et al is “associated” with RCP 8.5².

#3, The 2017 Task Force Report

In 2017, the Task Force shifted the full range higher producing 6 scenarios for a range of 0.3m-2.5m. These scenarios are: Low = 0.3m; Intermediate- Low = 0.5m; Intermediate = 1.0m; Intermediate-High = 1.5m; High = 2.0 m; and Extreme = 2.5m.

The minor shift on the low end reflects observed annual rates of SLR from satellite altimetry.

The authors justified the shift with a focus on the literature producing the most extreme SLR results:

we assessed the most up-to-date scientific literature on scientifically supported upper-end GMSL projections, including recent observational and modeling literature related to the potential for rapid ice melt in Greenland and Antarctica.

Accelerated ice loss from Antarctica and Greenland was, at the time, a hot topic, fierce debated, and deeply uncertain. It remains deeply uncertain.

The scenarios took on a probabilistic approach partially derived from emissions driven climate models to produce “emissions based, conditional probabilistic storylines and global model projections.”

Presenting the extreme 2.5m as a worst case was a bit of a stretch- even then. Was it a worst case appropriate for an advisory document or was it a statistical artifact produced from the effort to turn the exercise into a probabilistic approach using methodological judgements that massaged uncertainty towards its most extreme ends?

The Task Force supported the 2.5m scenario by referencing Jackson and Jevrejeva [JJ16] finding ~2.2 m (99th percentile), and Kopp et al. [K14] finding ~2.5 m (99.9th percentile)³. Both using the outdated and extreme RCP8.5 emission scenario.

2022 Task Force Report

The 2022 Task Force Report walks back the zeal from 2017 and goes back to 5 scenarios by lopping off the extreme. The whole thing is a bit more complicated but but the outcomes are pretty much the same. In their words,

In short, while the scenarios continue to be defined by projected values of GMSL increase in 2100, it is important to note that the paths to get to these target values have changed in this report compared to the previous one.

The five scenarios are: Low = 0.3m; Intermediate- Low = 0.5m; Intermediate = 1.0m; Intermediate-High = 1.5m; and High = 2.0.

Now there are two projection methods.

The first is a near term trajectory to 2050 based only on observation. The second is a projection using IPCC methods and referenced to temperature increases above preindustrial levels often referred to as “global warming levels”

The spread between the five GMSL rise scenarios is intended to reflect the range of potential future emissions pathways and associated warming levels that depends highly on global socioeconomic factors that have yet to unfold.

The effort was to align with IPCC practices and climate policy focused on temperature targets.

But because the IPCC does not project GMSL rise as high as 2 meters by 2100, the US scenarios do not translate well to the IPCC context. The table below attempts to align IPCC SLR projections to US scenarios and in the process align US scenarios with emission scenarios and global warming levels.

Note that none of the “bounding” scenarios are higher than the US Task Force Intermediate scenario. To resolve this the 2022 Task Force states that their

Intermediate-High and High scenarios represent potential futures in which these deeply uncertain ice-sheet processes play important roles in the late 21st century and beyond

But it is these low confidence uncertainties represented in the last two columns and even those two columns do not align with more than the US intermediate scenario.

In other words, the US scenarios Intermediate High and High fall well outside the views of the IPCC. They are seemingly untethered from any logical justification.

In the figure below, the Task Force demonstrates the extent to which their scenarios are dependent on high emission scenarios.

The authors explain,

As an example interpretation of this figure, the Low scenario generally requires a low emissions pathway, while the Intermediate-Low scenario arises from low, intermediate, and high emissions pathways. Pathways consistent with the Intermediate scenario include low emissions trajectories but are mostly related to high emissions scenarios. In fact, the Intermediate, Intermediate-High, and High scenarios are all heavily driven by high emissions scenarios, and differences between these scenarios are associated predominantly with the possible role and contributions of the low-confidence ice-sheet processes

Important to note is that the high emission scenarios- SSP3-7.0 and SSP5-8.5, are not plausible. The latter has been wholesale discarded and will not be a focus in the next IPCC report. There is also indication that global emissions are falling below the SSP2-4.5 trajectory.

Moreover, it is not until 2100 that the intermediate-low scenario significantly deviates from the observation based trajectory of GMSL⁴.

Final Thoughts…

Recall that in 2012 GMSL rise was bounded on the low end by the observed trajectory and on the high end by the highest at the time, 2m, which was characterized by substantial uncertainty about ice sheet melt. The author team was confident that future global mean SLR would be within this range. And why not? This was the range of observation and the prevailing highest estimate.

Despite the complicated modeling and shift to probabilities the underlying range and upper end scenarios persisted while becoming increasingly detached from a plausible (and logical) justification.

To be sure, the sea is rising globally and regionally, with regional sea level more significant for near term on the ground coastal management. A 0.5m of SLR is substantial for coastal communities even if it is not the dire projection of 2m.

Even in regions where sea level is rising more quickly then elsewhere to put that in the scenario context is misleading because of the way the scenarios are justified.

Stay tuned for a look at the NC Draft Report.

1. This breaks down to the use of the following studies:
Highest: (Pfeffer et al 2008); Intermediate-High (Grinsted et al. 2009, Jevrejeva et al. 2010, Vermeer and Rahmstorf 2009, Horton et al. 2008); Lowest: Meehl et al (2007)

2. To make this association the report uses Jevrejeva et al 2014 and Kopp 2014.

3. Both of these referenced worked used to support the extreme scenario were associated with advocacy research efforts.

JJ16 was supported by Climate Econometrics, at the time initiated by the Robertson Foundation, and partnered with the Environmental Defense Fund. At least today, the advisory board demonstrates plenty of cross over with K14.

Kopp and coauthor Horton were also authors of the 2017 Task Force report; they were also authors to NCA4. K14 had been, at least in part, a product of the politically oriented Risky Business Project and K14 was co-authored by several engaged with Climate Central, the media focused climate change advocacy group.

4. Perhaps the divergence may be different on regional scales.??