Projecting Sea Level Rise

To project future sea levels at each of the 55 NOAA water level stations in this study, we added projections of global sea level rise to separate local sea level change components.

For the global component, we used projections from a 2009 paper by Vermeer and Rahmstorf. Their approach, based on the recent historic relationship between global sea level and temperature, has successfully hindcasted sea level rise over the last century and millennium with great fidelity. If the ongoing increase in global temperatures leads ice sheets to unravel in ways not experienced during the model’s twentieth century calibration period, then this approach may understate the problem. 

Use of Vermeer and Rahmstorf’s approach allowed this analysis to take into account a wide range of possible futures, from ones where humanity continues to send great amounts of heat-trapping gasses into the atmosphere, to ones where we sharply reduce these emissions. Through Vermeer and Rahmstorf’s method we were also able to incorporate a range of possible relationships between emissions and global temperature increases, and a range of possible relationships between temperature and sea level. Our analysis rolled all of these factors together to produce one set of best estimates, and a range of potential outcomes around them.

Changes in local sea level come not only from changes in global sea level, but also from local effects such as the slow rising or sinking of coastal land, driven largely by the ancient retreat of massive ice sheets across North America. To determine local effects, we removed global rise from the total observed local sea level increase over a 50-year period at each of the 55 stations analyzed. In our projections, we then assumed that each local component will continue at a constant rate into the future. A detailed analysis using multiyear data from high-precision continuous GPS stations showed that vertical land motion can explain most or all of these local components; the forces behind such motion generally stay constant for thousands of years, justifying our assumption. (We made special adjustments for Louisiana and Texas sites, where human extraction of oil, gas and water has led to localized short-term rapid sinking of land. We used the slowest-sinking thirty year period, instead of the full 50-year record, leading to lower estimates of future sea level rise at these sites.)

We did not take into account a widely-anticipated slowing of the Gulf Stream later this century due to climate change. This slowing may add several inches of rise along the Northeast corridor by 2050 and more by the end of the century. We also did not take into account “gravitational fingerprint” effects, likely smaller than, and partly counteracting, potential Gulf Stream effects this century.

Our projections should not be interpreted as precise predictions for specific years, but rather best estimates that indicate overall trends, because of all of the factors that could lead to a range of outcomes (for example, different emissions futures) and because of natural year-to-year and decade-to-decade variability. For this reason, we present projections at the decade level only.

All these methods, together with these and these, underlie the sea level rise, flood level probability and flood timing statistics presented in the Surging Seas interactive map, the Surging Seas report, and their associated state factsheets and data downloads. For additional description and explanation, see the Methods section of this peer-reviewed paper.