Deep Ocean Warming

April 27, 2014 4:31 am7 comments

Note: This post resulted from the claim by Dr. Garth Paltridge that the observed surface warming could have been caused by internal ocean fluctuations. — Andreas Schmittner


 Could the Earth’s observed surface warming of the last 50 years have been caused by ocean circulation?

 
Let’s phrase this question as a hypothesis: Changes in ocean circulation have caused the observed surface warming of the past 50 years. We can test this hypothesis because if it were true there must be regions in the subsurface ocean where it has cooled. This follows from energy conservation. Observations, however, show no evidence for cooling of the subsurface ocean. The upper ocean heat content (top 2 km) has clearly increased as several independent estimates show (here is one). How about the layers below that? Could the heat come from there? Here is a figure estimating heat content changes for the decade from the 1990’s to the 2000’s showing that the deepest layers of the oceans have also warmed. I’m not aware of observations that show that deep ocean layers have cooled over the last 50 years. Thus, existing evidence does not support the hypothesis. Therefore we have to answer the question with “No”.

kouketsu11jgr_fig4

Fig. 4 from Kouketsu et al. (2011) shows ocean heat content changes from the 1990’s to the 2000’s. Note the different scales on the top and bottom panel.

Note that the layers below 2.5 km depth have much smaller changes in heat content compared with the surface. This is because the deepest layers are very old and mix very slowly with the surface. Therefore I think that it is extremely unlikely that ocean circulation has caused the surface warming of the last 50 years.

UPDATE: April 28, 2014:

Mauri: thanks. Here is the figure. This is from an ocean re-analysis, which is essentially a global ocean circulation model driven with (or nudged to) observations. Thus it predicts the changing state of the ocean, including fluid flow, consistent with the existing observations. You could regard this as a reconstruction of past ocean circulation and heat content or a dynamically consistent extrapolation (filling of data gaps) of existing (relatively sparse) observations. Similar to the NOAA Ocean Heat Content figure this estimate also shows an increase over the past 50 years.

Balmaseda Figure 1

Here is another paper with observations of deep ocean heat content changes. Purkey and Johnson 2010 Journal of Climate, vol. 23, p. 6336. These authors use observations from the 1980 to 2010 to estimate the changes from the 1990s to the 2000s. Here is the global figure expressed as heat gain per meter (W/m^3). You can see heat gain at all depths consistent with the Kouketsu et al. figure above.

These authors also show a map of deep (below 4 km) ocean heat flux changes.

Note that bottom waters are warming in areas where sinking from the surface takes place around Antarctica. Particularly in the Atlantic sector along the east coast of South America. This is where Antarctic Bottom Water flows north. So, the pattern of bottom water warming is consistent with heating from the surface.

Note also in the figures above that warming is largest at the surface and small near the sea floor. This distribution is inconsistent with warming from below (the sea floor). It is consistent with warming from above (the atmosphere).

The following figure was added on July 28, 2014.

From Wunsch and Heimbach (2014). Their Fig. 19.

From Wunsch and Heimbach (2014). Their Fig. 19.

 

 

'Deep Blue Dolphin Love' by Patrik Jones via Flickr

Featured image: Patrik Jones via Flickr (link)

THE FORUM'S COMMENT THREAD

  • Well put succinct argument. It could be driven home within another chart. I like the charts from Balmaseda, Trenberth and Källén (2013)in GRL You could add there Figure 1. As a further demonstration, I uploaded it to the media library.

    • John: thanks for forwarding these comments.

      Dear anonymous reader: thank you for the interesting papers. They show the effect of changes in meridional (north-south) heat transport in the ocean on surface climate in idealized model simulations. If heat transport is artificially increased in these models it leads to melting of sea ice at high latitudes and thus warming there, whereas the tropics don’t cool because of changes in atmospheric circulation and clouds there. So, increasing ocean heat transport can lead to global warming.

      I don’t think that these studies are relevant for the discussion of the past 50 years. There is no evidence that I’m aware of that indicates ocean heat transport has increased since the 1960’s. Ocean re-analyses (e.g. http://bit.ly/1iWJfAC) and observations from the RAPID program (http://www.rapid.ac.uk/rapidmoc/) show no trend in northward heat transport in the Atlantic and a reduction of the Atlantic Meridional Overturning Circulation.

      • From my anonymous correspondent:

        Dear Andreas, thank you for your response. I agree there is no evidence that OHT has increased since the 1960s which is near the end of the positive phase of an AMO. There is also no evidence that it hasn’t.

        There is evidence that it has increased since the late 1970s (I note that although you make claims for 1960s your earliest reference begins in 1993) and that would be in the change to the positive phase of the AMO from the negative phase. This would explain why the OHC of the North Atlantic, which is closely tied to OHT (Sonnewald et al 2013) was increasing faster than globally .

        Reconstructions of the AMO do not show a consistent peak to trough and we have no way of knowing, given that we only have measurements of what appears to be the beginning of the decline to the negative phase of the current AMO (Smeed et al 2014) and none of the previous AMO or of the movement to the positive phase of the current AMO, if this oscillation will have a larger peak to trough than the previous one or not. The nice thing about this problem is that I will probably live long enough to find out.

        A more interesting problem due to the noise created by the AMO is what is happening to what would be considered the secular trend in OHT. This is characterized by the Lund et al reconstruction of Gulf stream transport 2006 as shown in this graph from Wanamaker et al 2012 where it is used as a comparison to their results. The Lund reconstruction was still showing an acceleration as of 1950. It is possible it suddenly stopped and it is possible all the feedbacks were complete by 1960, but it doesn’t seem likely that this change in OHT had no influence on the post 1960 temperatures if you assume the reconstruction is correct. I am patiently waiting for more studies and data at this point.

        • Although all of this is fascinating I don’t think it is relevant to the discussion of the warming of global surface temperatures over the past 50 years.

          The Atlantic Meridional Overturning Circulation (AMOC), which is probably associated with AMO (Atlantic Multidecadal Oscillation), transports heat from the southern hemisphere into the northern hemisphere (Crowley 1992, Paleoceanography). We know from modeling studies that AMOC variations do not change global temperatures (or change it only very little). So, while AMOC and AMO variations have large impacts on regional temperatures (stronger AMOC = positive AMO = warm North Atlantic = cool South Atlantic) they do not affect global temperatures much. This is different from the idealized model simulations you mentioned in your first comment. In those models the poleward heat transport was increased artificially everywhere.

  • Adam: thanks for your comment.

    Indeed Wunsch and Heimbach (2014), who are eminent oceanographers, estimate cooling of the deepest layers (below 2000m) during the 1993-2012 period (see their Fig. 19 posted above), which is in contrast to the studies cited above. In their discussion they acknowledge that “these results differ in detail and in numerical values from other estimates, but the determining whether any are “correct” is probably not possible with the existing data sets”.

    Given that different ocean re-analyses disagree it seems to me that the abyssal (below 2km) heat content changes are uncertain (perhaps more uncertain than my original post indicated). However, it is important to note that W&H’s estimates for the upper ocean agree more with Balmaseda. E.g. their 0-700m estimate from 1994 to 2009 is about +5×10^22 J compared to ~+8×10^22 J of Balmaseda. Also the pattern of little change in the 1990s and larger increases in the 2000s agrees. So clearly most of the disagreement is below 700m. Whereas Balmaseda shows a warming of ~5×10^22 J, W&H show essentially no change.

    However, the total (all depths) heat content increases in W&H consistent (qualitatively) with all the other estimates and contrary to the Paltridge’s suggestion that the oceans could have been heating the atmosphere.

    W&H seem to critisize Balmaseda’s error estimates writing “The problem necessitates close examination of claimed abyssal warming accuracies of 0.01 W/m2 based on such methods (e.g., Balmaseda et al., 2013).” I think this is not quite fair since Balmaseda state in their paper already that “The uncertainty is probably underestimated …”.

    The WUWT article is basically OK, but emphasizes uncertainties of results it does not like and certainties of convenient results. E.g. it falsely gives the impression that the deep oceans are in fact cooling, whereas W&H emphasize the uncertainty in their deep ocean estimates as quoted above. On the other hand it criticizes Balmaseda whereas I think both estimates are equally valid. So, as usual with WUWT posts it is biased. It also contains some irrational claims such as geothermal heating could be the cause of heating middle layers of the ocean.

Leave a Reply

You must be logged in to post a comment.

PUBLIC COMMENT THREAD

  • http://ceoas.oregonstate.edu/profile/schmittner/ Andreas Schmittner
  • http://glacierchange.wordpress.com/ Mauri Pelto

    Well put succinct argument. It could be driven home within another chart. I like the charts from Balmaseda, Trenberth and Källén (2013)in GRL You could add there Figure 1. As a further demonstration, I uploaded it to the media library.

  • http://atmo.tamu.edu/profile/JNielsen-Gammon John Nielsen-Gammon
    • http://ceoas.oregonstate.edu/profile/schmittner/ Andreas Schmittner

      John: thanks for forwarding these comments.

      Dear anonymous reader: thank you for the interesting papers. They show the effect of changes in meridional (north-south) heat transport in the ocean on surface climate in idealized model simulations. If heat transport is artificially increased in these models it leads to melting of sea ice at high latitudes and thus warming there, whereas the tropics don’t cool because of changes in atmospheric circulation and clouds there. So, increasing ocean heat transport can lead to global warming.

      I don’t think that these studies are relevant for the discussion of the past 50 years. There is no evidence that I’m aware of that indicates ocean heat transport has increased since the 1960’s. Ocean re-analyses (e.g. http://bit.ly/1iWJfAC) and observations from the RAPID program (http://www.rapid.ac.uk/rapidmoc/) show no trend in northward heat transport in the Atlantic and a reduction of the Atlantic Meridional Overturning Circulation.

      • http://atmo.tamu.edu/profile/JNielsen-Gammon John Nielsen-Gammon

        From my anonymous correspondent:

        Dear Andreas, thank you for your response. I agree there is no evidence that OHT has increased since the 1960s which is near the end of the positive phase of an AMO. There is also no evidence that it hasn’t.

        There is evidence that it has increased since the late 1970s (I note that although you make claims for 1960s your earliest reference begins in 1993) and that would be in the change to the positive phase of the AMO from the negative phase. This would explain why the OHC of the North Atlantic, which is closely tied to OHT (Sonnewald et al 2013) was increasing faster than globally .

        Reconstructions of the AMO do not show a consistent peak to trough and we have no way of knowing, given that we only have measurements of what appears to be the beginning of the decline to the negative phase of the current AMO (Smeed et al 2014) and none of the previous AMO or of the movement to the positive phase of the current AMO, if this oscillation will have a larger peak to trough than the previous one or not. The nice thing about this problem is that I will probably live long enough to find out.

        A more interesting problem due to the noise created by the AMO is what is happening to what would be considered the secular trend in OHT. This is characterized by the Lund et al reconstruction of Gulf stream transport 2006 as shown in this graph from Wanamaker et al 2012 where it is used as a comparison to their results. The Lund reconstruction was still showing an acceleration as of 1950. It is possible it suddenly stopped and it is possible all the feedbacks were complete by 1960, but it doesn’t seem likely that this change in OHT had no influence on the post 1960 temperatures if you assume the reconstruction is correct. I am patiently waiting for more studies and data at this point.

        • http://ceoas.oregonstate.edu/profile/schmittner/ Andreas Schmittner

          Although all of this is fascinating I don’t think it is relevant to the discussion of the warming of global surface temperatures over the past 50 years.

          The Atlantic Meridional Overturning Circulation (AMOC), which is probably associated with AMO (Atlantic Multidecadal Oscillation), transports heat from the southern hemisphere into the northern hemisphere (Crowley 1992, Paleoceanography). We know from modeling studies that AMOC variations do not change global temperatures (or change it only very little). So, while AMOC and AMO variations have large impacts on regional temperatures (stronger AMOC = positive AMO = warm North Atlantic = cool South Atlantic) they do not affect global temperatures much. This is different from the idealized model simulations you mentioned in your first comment. In those models the poleward heat transport was increased artificially everywhere.

  • http://ceoas.oregonstate.edu/profile/schmittner/ Andreas Schmittner

    Adam: thanks for your comment.

    Indeed Wunsch and Heimbach (2014), who are eminent oceanographers, estimate cooling of the deepest layers (below 2000m) during the 1993-2012 period (see their Fig. 19 posted above), which is in contrast to the studies cited above. In their discussion they acknowledge that “these results differ in detail and in numerical values from other estimates, but the determining whether any are “correct” is probably not possible with the existing data sets”.

    Given that different ocean re-analyses disagree it seems to me that the abyssal (below 2km) heat content changes are uncertain (perhaps more uncertain than my original post indicated). However, it is important to note that W&H’s estimates for the upper ocean agree more with Balmaseda. E.g. their 0-700m estimate from 1994 to 2009 is about +5×10^22 J compared to ~+8×10^22 J of Balmaseda. Also the pattern of little change in the 1990s and larger increases in the 2000s agrees. So clearly most of the disagreement is below 700m. Whereas Balmaseda shows a warming of ~5×10^22 J, W&H show essentially no change.

    However, the total (all depths) heat content increases in W&H consistent (qualitatively) with all the other estimates and contrary to the Paltridge’s suggestion that the oceans could have been heating the atmosphere.

    W&H seem to critisize Balmaseda’s error estimates writing “The problem necessitates close examination of claimed abyssal warming accuracies of 0.01 W/m2 based on such methods (e.g., Balmaseda et al., 2013).” I think this is not quite fair since Balmaseda state in their paper already that “The uncertainty is probably underestimated …”.

    The WUWT article is basically OK, but emphasizes uncertainties of results it does not like and certainties of convenient results. E.g. it falsely gives the impression that the deep oceans are in fact cooling, whereas W&H emphasize the uncertainty in their deep ocean estimates as quoted above. On the other hand it criticizes Balmaseda whereas I think both estimates are equally valid. So, as usual with WUWT posts it is biased. It also contains some irrational claims such as geothermal heating could be the cause of heating middle layers of the ocean.

  • Pingback: What Climate Scientists Are Saying About the Global Warming Hiatus