A recent conference presentation at AGU (reported here) while confirming that global sea level is indeed rising (in line with other estimates), showed that Arctic sea levels may actually be falling. On the face of it these preliminary results are a little puzzling (though note that this isn’t yet a properly peer reviewed paper, and so may not reflect what ends up in the journal), but it does reveal some of the complexities in analysing sea level in relatively small enclosed basins and so a brief overview of the different factors involved is probably useful.
Firstly, where is this new data from? The ERS-2 satellite has a ‘radar altimeter’ similar to that flown on the TOPEX/ POSEIDON (and now JASON) satellites and it works by beaming down a radar pulse and seeing how long it takes to come back. These techniques are remarkably accurate (down to a few cm’s and can give accuracy in trends to a few tenths of mm per year averaged over large areas). However the Arctic and Antarctic present special problems due to the poorer coverage near the pole (nothing above 82ºN for instance) and, of course, the presence of sea ice. The ‘sea level’ being talked about is that of the water in between the ice floes, and so the data must be very carefully sorted to yield only the open water values (i.e. where there is no floating ice). This clearly removes the majority of the data and must make any Arctic-wide estimates much more uncertain than the global numbers outside of the polar regions). There is also the possibilty that the analysis has still not removed all sources of contamination in the trend – there may be a seasonal bias (since open water has increased more in summer than winter) or the thinning of the ice cover means that any residual ice contamination could make it look incorrectly like ‘sea’ levels were falling. These issues will obviously be considered in greater depth in any actual publication.
But what if the numbers are roughly correct? What would that mean?
Well, sea level in any part of the ocean is a function of the density of the water (which is affected by its salt content and temperature), but also the local currents, pressure and wind patterns. In this analysis, one of the confounding elements in many assessments; the movement of the land due to the recovery from the last ice age or from subsurface geological/volcanic activity – is not a factor since the measurements are taken with respect to the geoid, not any nearby land stations. The Arctic appears to have freshened in recent decades which on its own would lead to a very slight increase in sea level. This is true whether the freshwater comes from land (in which case it is just adding to the total water in the ocean), or whether it is due to sea ice melting (which contrary to popular belief, does – very slightly – add to sea level rise due to the relative freshness of sea ice compared to the (denser) salty ocean water it displaces). Arctic sea water as a whole does not appear to be warming much (since most of the available heat goes into melting ice, rather than raising temperatures), but it is conceivable that changes to the thermal and saline profiles in depth (as a function of changes in, for instance, the Atlantic inflow) might cause some adjustment of sea level. However the bigger effect is most likely to be dynamic, and related to how much water the winds ‘pile up’ in various places. One example of this effect is associated with the dominant westerlies in the mid-latitude regions. There ‘Ekman drift’ causes water to pile up in the sub-tropics rather than the sub-polar gyres leading to about a meter difference in sea level across the Gulf Stream/North Atlantic Drift (higher to the south). Thus a long term increase in the westerlies (as has been seen in recent decades) could cause an increase in that effect, actually lowering the Arctic sea level. Other changes in wind patterns could conceivably have similar effects. Additionally, changes in sea level pressure can have a direct ‘inverse barometer’ effect on the sea level (a 1 mb sustained increase would lead to about a 1 cm decrease in sea level).
All of these mechanisms can (and no doubt will) be assessed for their possible contribution to this result and it may hold up. It is however extremely unlikely to change any results on the global scale – which as I noted above are completely in line with the other satellite and tide gauge estimates.