Volume 12, Number 7: 18 February 2009
In an intriguing study recently published in Ocean Dynamics, von Storch et al. (2008) write that Rahmstorf (2007) recently proposed "a linear relationship between the rate of global mean sea-level rise and the global mean near-surface air temperature deviations," which "is calibrated with observed data, thus incorporating in a somewhat realistic and condensed manner all known and unknown mechanisms modulating the global sea-level height," which concept sounds quite reasonable. But does it work?
One way of addressing this question -- and which they proceed to employ -- is, in their words, "to test the statistical methods in the virtual reality produced in simulations with state-of-the-art climate models." Following this strategy, they thus examined "several hypotheses concerning the relationship between global mean sea level and other thermal surface variables in a long climate simulation of the past millennium with the climate model ECHO-G driven by estimations of past greenhouse gas, volcanic and solar forcing." So what did they find?
The three researchers report that the linear link between global mean temperature and the rate of change of global mean sea level "turned out to be not reliable over the full time period," noting that "instead, for some periods, even inverse relationships [our italics] were found to describe the simulated data best." Likewise, they say that the second predictor -- the rate of change of temperature -- "did not show markedly better results." And for both predictors, they report "there exist periods in the simulation where the prediction errors are very large."
In discussing their findings, von Storch et al. acknowledge that the type of test they performed in the "virtual reality" produced by climate models "cannot prove whether a certain hypothesis, in this case the different statistical relationships, will hold in the real world." However, as they continue, "they can be used to falsify a particular hypothesis," noting that "if it is not fulfilled in a simple virtual reality, it will probably also fail in a more complex real world."
Al Gore and James Hansen, take note. There is currently no known way to predict -- with any reasonable and demonstrable degree of confidence -- what mean global sea level will do over the 21st century, even if mean global air temperature begins to rise once again (after having remained rather stable for the past decade).
Sherwood, Keith and Craig Idso
Rahmstorf, S. 2007. A semi-empirical approach to projecting future sea-level rise. Science 315: 368-370.
von Storch, H., Zorita, E. and Gonzalez-Rouco, J.F. 2008. Relationship between global mean sea-level and global mean temperature in a climate simulation of the past millennium. Ocean Dynamics 58: 227-236.