Modern Sea Level Change
Changes in the global average of the Earth’s sea level (GMSL) are a sensitive indicator of the climate system. Prior to the advent of satellite altimetry observations in the 1990s, measurements of sea level were only made by tide gauges located around the coastlines of the continents. The sparseness of these observations made estimates of the global average challenging to obtain. Using various approaches to accommodate the temporal and spatial sparseness of the records, previous studies have estimated the mean GMSL rate over the 20th century as 1.6 - 1.9 mm/yr. Carling Hay, Bob Kopp, Jerry Mitrovica, and I have developed an approach to reconstructing the global sea level field that Bayesian inference to estimate the underlying contributions to sea-level change which are then summed to capture GMSL changes. We actually performed two different analyses, one based upon a Kalman smoother the other on the machine-learning approach of Gaussian process regression. The benefits of such probabilistic approaches are that they naturally accommodate sparse sampling of a global field, they provide a defined framework for uncertainty estimation and they allow for a correction of a distribution of isostatic adjustment and ocean dynamics models. Using these approaches we estimate that the rate of GMSL rise from 1901 through to 1990 was 1.2 mm/yr which is lower than past estimates over that time frame. However, we also estimate GMSL rise from 1993 - 2010 as 3.0 mm/yr, which is consistent with prior estimates. Consequently, we conclude that the acceleration to the present-day rates of GMSL rise are higher than previously thought.