March 2019

Last update: 23/04/2019

This month I focus on a new paper about the user needs for sea level rise information, the high-end sea level projection from the New York City Panel on Climate Change report, the response of ice sheet models to basal melt and a mechanism for marine ice cliff instability.

In a new paper “Meeting User Needs for Sea Level Rise Information: A Decision Analysis Perspective” a group of scientists lead by Jochen Hinkel investigated the interface between sea level science and decision making for adaptation to sea level rise. The authors first assess which information is needed for different decision analysis methods, and then which information can be provided by current sea level science. This is extremely valuable information to both sea level scientists who want to provide relevant information and decision makers who need to choose a decision analysis method that is compatible with the information that sea level science can provide. The paper is very pedagogical and easy to read . A good summary of the paper is shown in their Figure 1:

The choice of the decision analysis method depends on two main criteria: the time horizon of the decision and the uncertainty tolerance of users. For short term projects and medium to high uncertainty tolerance the decision analysis method suggested by the authors is the expected utility maximisation, for example cost-benefit analysis. The sea level information that is required is a probabilistic prediction. That is a difficult product to provide because it should include all sources of uncertainties, even the ones that are generally not included in sea level scenarios, like natural climate variability and greenhouse gazes emission. Still the authors consider that the field of sea level is able to provide such a product for some locations. For long-term decisions and/or for low uncertainty tolerance the authors propose the robust decision-making approach. For this approach the necessary information about future sea level is high-end and low-end scenarios and ideally also an upper bound. Unfortunately the sea level community, and in particular the ice sheet community, is not able to provide a useful upper bound for sea level rise. The safest upper-bound would be 65 meters (melting of all the available land ice) but it is not very useful for decision making at century time scale. On the other hand high-end and low-end scenarios can be developed, even though this is not straight-forward, as the authors put it: “In summary, a dilemma remains. While uncertainty intolerant decision making requires SLR information beyond the likely range, efforts to go beyond this range have serious caveats and hence provide lower confidence information.” I recall here that the IPCC definition for the a “likely range” is a range that has a probability of 66% or more. Meaning that, assuming the projection is correct, there is a probability of 33% or less that future sea level will fall outside this range. The paper proposes two strategies to develop such upper-bound. The first one is “not to combine results of different studies, model runs, or expert opinions but rather to report disaggregated results and document the reasons why these differ”. And the second one is to use extra-probabilistic theories when probabilities are poorly defined. These two approaches can be combined and this is what is done in the paper.

One thing we learn from this study is that sea level projections produced by expert panels (e.g. IPCC, KNMI scenarios) only cover a small part of the user needs. They only cover the needs of users interested in long term planning with low risk aversion. Because for short term planning one needs a full probabilistic prediction while for long term planning with high risk aversion users need information about possible extreme futures with a low probability of occurrence.

In the conclusion, I find the thoughts of the authors about co-production confusing. On the one hand they write: “These results illustrate that the classical division of work of physical science producing SLR information and others converting and communicating this information is fraught because the production of suitable information depends on the kinds of decisions users are facing. This reiterates the importance of coproduction of sea level information between producers and users of this information” which is a very clear praise for co-production.

On the other hand in following sections they propose a way to organise the work that leaves little room for co-production. First, an authoritative panel assesses (subjectively) the level of confidence of scientific studies and then: “It is the role of decision makers, who are the users of the information from physical scientists, to judge how much risk they are willing to take, informed by the experts' confidence judgments on available studies. This requires well‐informed users but is probably the only robust way forward at this time.” That means that users can just decide which confidence level they are interested in and then follow the recommendation from the authoritative panel. So this brings us back to "the classical division of work of physical science producing SLR information and others converting and communicating this information".

Co-production could already start when deciding about the level of confidence of different methods. This is subjective so there is room for discussion with users to help make these categories. If a user trusts numerical models then they can be given higher confidence than paleo-proxies for example. To facilitate this way forward expert panels should make clear what is the level of subjectivity in their decisions. Discussing strength and weaknesses of different papers can be reasonably objective. Saying one numerical model performs better than another or one paleo proxy is more accurate than another can also be objective. Saying that a structured expert judgment has less confidence than the IPCC expert judgement, as done in AR5 (Fifth Assessment Report from IPCC) and repeated in this paper, is subjective. Especially when this judgement is made by the authors of the AR5 themselves. Where the subjectivity has been clearly flagged there is room for co-production. The downside is that it would make the IPCC even more difficult to read, after the estimation of likelihood and confidence here comes the subjectivity index.

On the topic of high-end scenario, a new report of the New York City Panel on Climate Change was recently published. Chapter 3 is about sea level. The high-end scenario in this report is build from Kopp et al. (2017) that used the DeConto and Pollard (2016) projections for Antarctica. The sea level rise is 2.9m in 2100 compared to 2000-2004. What is interesting is that they do not only provide a physically consistent possible future, they also provide an estimate of the probability of this high-end scenario to be exceeded based on a structured expert judgment. They asked 22 ice sheet experts to evaluate the future contribution of Antarctica and Greenland to sea level rise under a low global warming scenario (2ºC in 2100) and a high scenario (5º in 2100). They find that in the low scenario the probability to exceed 2.9m of sea level rise in 2100 is close to zero and for the high scenario it is around 3%.

Another important study that caught my attention is a manuscript from the ISMIP6 team. ISMIP6 is the Ice Sheet Model Intercomparison Project for CMIP6. It is a very important effort to compare ice sheet models, understand why they differ and find ways to improve them. There are a few sources of uncertainties when projecting the future of Antarctica. First, how much is the ocean going to warm up around the Antarctic ice shelves and how much the ocean circulation is going to change and bring warm water masses (Circumpolar Deep Water) in contact with the shelves. Second, how much is this going to melt the bottom of the ice shelves (e.g. basal melt) and finally how sensitive the ice sheet is to basal melt. It is this last source of uncertainty that is the focus of an experiment where all the groups apply the same additional rate of basal melt to their ice sheet models. This rate is roughly equivalent to present day basal melt rates. The resulting sea level contribution of the models in one century covers almost two orders of magnitude, from 1.3 to 42.7 cm (have a look at all the models in their Figure 4).

One uncertainty that is not considered in ISMIP6 is Marine Ice Cliff Instability (MICI). The debate goes on at the moment about whether there will be enough surface melt to start hydrofracturing of the Antarctic ice shelves outside of the Antarctic Peninsula and, what would happen if large ice cliffs emerge (~1000m high including ~100m above the water level). Previous studies suggested that such large ice cliffs would collapse under their own weight (e.g. brittle failure, Bassis and Walker (2012) ). A new study by Pariziek and colleagues provides some insights about a mechanism that is currently responsible for calving of the Helheim glacier in Greenland (see pictures of location and calving front bellow). First a “retrogressive slumping” (e.g. think of an avalanche) of the ice above the water level occurs. Then in the absence of surface load on the ice that is under water a strong buoyancy force pushes it upward, if a crack is present in the ice this force opens it further and an iceberg is created.

References:

Bassis, J. N., & Walker, C. C. (2012). Upper and lower limits on the stability of calving glaciers from the yield strength envelope of ice. Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 468(2140), 913–931. http://doi.org/10.1098/rspa.2011.0422

Deconto, R. M., & Pollard, D. (2016). Contribution of Antarctica to past and future sea-level rise. Nature, 531(7596), 591–597. http://doi.org/10.1038/nature17145

Gornitz, V., Oppenheimer, M., Kopp, R., Orton, P., Buchanan, M., Lin, N., … Bader, D. (2019). New York City Panel on Climate Change 2019 Report Chapter 3: Sea Level Rise. Annals of the New York Academy of Sciences, 1439(1), 71–94. http://doi.org/10.1111/nyas.14006

Hinkel, J., Church, J. A., Gregory, J. M., Lambert, E., Le Cozannet, G., Lowe, J., … Wal, R. (2019). Meeting User Needs for Sea Level Rise Information: A Decision Analysis Perspective. Earth’s Future, 7(3), 320–337. http://doi.org/10.1029/2018EF001071

Kopp, R. E., DeConto, R. M., Bader, D. A., Hay, C. C., Horton, R. M., Kulp, S., … Strauss, B. H. (2017). Evolving Understanding of Antarctic Ice-Sheet Physics and Ambiguity in Probabilistic Sea-Level Projections. Earth’s Future. http://doi.org/10.1002/2017EF000663

Parizek, B. R., Christianson, K., Alley, R. B., Voytenko, D., Vaňková, I., Dixon, T. H., … Holland, D. M. (2019). Ice-cliff failure via retrogressive slumping. Geology, 47(5), 1–4. http://doi.org/10.1130/G45880.1