Improving the communication of uncertainties for future marine flooding as sea-level rises using advanced statistical methods
Résumé
Among all adverse consequences of sea-level rise (SLR), a most immediate one should be more frequent marine flooding of low lying coastal areas. However, a large number of different uncertainty sources affect the impact assessment related to SLR. Part of these originate from global sea-level rise SLR scenarios, their regional variability, and their interactions with solid earth deformation processes. Other are related to oceanographic processes underpinning extreme events and coastal geomorphic changes. In addition, local vertical ground motions may aggravate or moderate sea-level changes at the coast. This raises the following questions: what is the relative importance of each source of uncertainty in the final flooding projections? Which sources of uncertainty need to be considered? how to account for the lack of knowledge in these different sources? Hence, getting better insights in the role played by these uncertainties enables to ease their communication and to structure the message on future coastal impacts. In this view, we propose to explore the fesability of two advanced staistical methods.
The first one relies on probabilistic projections and variance-based global sensitivity (revisit of Le Cozannet et al., 2015). The method is applied on an urban low-lying coastal site located in the north-western Mediterranean, where the yearly probability of damaging flooding could grow drastically after 2050 if sea-level rise follows IPCC projections. The analysis shows that 1) coastal processes and particularly the wave set-up are dominant factors during the first part of the 21st century; 2) Sea-level rise and climate change scenarios dominate by 2080. Sea-level rise variability has its maximum contribution to the uncertainties in-between these two periods 3) The uncertainties on the upper limit of sea-level rise projections is important a few decades later.
The second method relies on new theories of uncertainty (here possibility theory) for representing our partial knowledge on projections of future SLR. These projections remain highly uncertain, especially due to large unknowns in the melting processes affecting the ice-sheets in Greenland and Antarctica. Based on climate-models outcomes and the expertise of scientists concerned with these issues, the IPCC provided constraints to the quantiles of sea-level projections. Moreover, additional physical limits to future sea-level rise have been established, although approximately. However, many probability functions can comply with this imprecise knowledge. We show how possibility theory can provide a more flexible and less ambiguous framework than classical probabilities to represent in a concise representation of uncertainties in future sea-level rise and of their intrinsically imprecise nature, including a maximum bound of the total uncertainty (Le Cozannet et al., 2017).
We suggest that the probabilistic and possibilistic theories are complementary frameworks to support the communication of uncertainties on future impacts of sea-level rise: while the probabilistic framework captures average evolutions, the possibilistic approach appears appropriate to explore the fuzzy high-end (or low-end) sea-level scenarios and their impacts. The issue is of particular importance to coastal managers responsible for adapting to the adverse effects of SLR.
References
Le Cozannet, G., Rohmer, J., Cazenave, A., Idier, D., van De Wal, R., De Winter, R. Pedreros, Y. Balouin, C. Vinchon, C. Oliveros (2015). Evaluating uncertainties of future marine flooding occurrence as sea-level rises. Environmental Modelling & Software, 73, 44-56.
Le Cozannet, G., Jean-Charles Manceau and Jeremy Rohmer (2017). Bounding probabilistic sea-level projections within the framework of the possibility theory. Environmental Research Letters, 12, 1748-9326.