Comment on “Response of the global ocean to Greenland and Antarctic ice melting” by D. Stammer


1. Introduction

[1] Stammer [2008] comes to the conclusion that sea level rise from melting Greenland ice will have little effect on sea levels in the Pacific and other distant areas for over 50 years. This result has important ramifications and has been reported in the press. He derives a similar result for melting of Antarctica. However, these are conclusions that the paper is unable to support. A retraction needs to be printed.

[2] Stammer's [2008] study is on the redistribution of water properties (including sea level) in the oceans of the world in response to injection of fresh water near Greenland and Antarctica. Stammer's Figure 1 shows locations and amounts of fresh water entering the oceans in a graphical, semiquantitative form. Stammer's Figures 2–11 show maps and plots of the response of the global ocean after time steps of up to 50 years. Stammer's Figure 6 shows that in 50 years, sea level will have increased in the Baffin Bay/Labrador Sea area by about 30 mm, but in the rest of the Atlantic by less than 10 mm, and in the Pacific by only 1.5 to 2 mm. Similarly, Stammer's Figure 10 shows sea level rise after 50 years due to melting of Antarctica, of 10 cm close to Antarctica, but less than 0.5 mm for areas north of 60°S.

[3] Stammer's [2008] paragraph 47 in section 5 states, “The corollary of our findings is that melt water dumped into the North Atlantic from Greenland will reside first of all in the Atlantic and will only slowly propagate into the other basins. In particular, it will take a significant length of time until the Pacific Ocean will “feel” this extra volume, for example, in form of sea level rise. This is an important result since it implies that melting of Greenland's ice cap is much less of a threat to tropical islands in the Pacific than it is for the coasts of North America and Europe.”

2. Discussion

[4] While it may be true that the actual water will reside in the Atlantic, the paper cannot demonstrate that it will take 50 years and longer until the Pacific Ocean will “feel” it, since the actual addition of fresh water is explicitly omitted from the model. Actual addition of water into the model would lead to a much larger rise in sea level which would swamp the results presented.

[5] In the model, the effect of fresh water injection is simulated by removing salt [Stammer, 2008, paragraph 15]. This changes the structure of the water column in a way consistent with Greenland ice melting, but omits any actual addition of mass, thereby removing the main cause of sea level rise. Stammer's paragraph 36 states “We followed the approach suggested by Greatbatch [1994], by assuring that the globally averaged bottom pressure remains constant over the entire model run. This assumption is equivalent to the model's mass remaining unchanged.”

[6] Given that no net mass is added to the model, it is at first surprising that Stammer finds any sea level rise at all. However, when salt is removed, water must be added to maintain the model's mass, causing a relatively small increase in model ocean volume. This increase, with the values shown in Figure 6 of Stammer [2008], is equivalent to about 3.5% of the volume increase that would occur if fresh water were really added. This fraction represents the weight of salt compared to the weight of water in the reference seawater used, see Stammer's paragraph 15. “Real” addition of water would of course significantly increase the model's mass.

[7] An important piece of information that should have been provided by Stammer [2008] is the total fresh water flux considered in the model. Figure 1 of Stammer's paper shows the locations of the “additions,” with amounts indicated by colored squares. Visual integration of squares near Greenland is difficult, but suggests a total freshwater flux on the order of 20,000 m3 s−1, equivalent to about 600 Gt/yr, or about 3 times the present (2003–2009) melt rate indicated by the GRACE satellite [Velicogna, 2009]; 600 Gt/yr is equivalent to a global rise rate of 1.5 mm/yr, or 75 mm in 50 years, roughly 30 times the global average value in Figure 6 of Stammer. This is as expected, given that Stammer's model will see only 3.5% of the true rise.

3. Conclusions

[8] The actual time response with which the global oceans feel this huge increase in mass from Greenland cannot be studied by Stammer's model. It may be the same as implied by Stammer's [2008] Figure 6, but the actual rate of addition of fresh water implies a sea level increase of about 10 m near Greenland, which could not be balanced by any reasonable geostrophic flows. A much faster response of global sea level is more likely.

[9] Stammer [2008] supports his conclusions by noting that they are consistent with earlier studies by Bryan [1996] and Hsieh and Bryan [1996]. These studies are also of the redistributions of ocean properties, including sea level. However, the changes are in response to warming of the oceans, in which no mass addition occurs, so they are not subject to Stammer's error. Greatbatch [1994], referred to by Stammer as noted above, also discussed steric sea level rise, and so was also not concerned with the addition of mass. In view of the relevance of this result in planning for impacts of global climate change, it is important that errors like this are discussed and corrected.


[10] This work was supported by Fisheries and Oceans Canada and by the Canadian Space Agency (CSA) under the GRIP (Government Related Initiative Program). I thank colleagues at IOS for discussions and suggestions.