Climate Change and Hurricane Sandy

Disclaimer: This draft is an evolving research assessment and not a final report. The analyses presented have not yet been peer reviewed and do not represent official positions of ESRL, NOAA, or DOC. Comments are welcome. For more information, contact Dr. Martin Hoerling (martin.hoerling@noaa.gov)

Afterword

Other aspects of climate change, beyond the surmise of the possible impact of sea ice depletion, have also been speculated to have been key to the dynamics of Hurricane Sandy. These speculations can also be largely rejected based on current scientific understanding. To reiterate, NOAA's State of Science Fact Sheet concerning Atlantic hurricanes and global warming (released May 2012), based on assessment by numerous researchers and NRC review, states that there has been no change in Atlantic tropical cyclones since the late 1800s. The IPCC (2012) SREX report further indicates that there is low confidence concerning the effects of global warming on tropical cyclones...either their intensity, frequency, o their tracks. Regarding extratropical cyclones, a feature that strongly affected the path and eventual dynamics of Sandy along the Mid-Atlantic coast, the IPCC SREX report indicates a poleward shift of the tracks in such storms due in part to global warming, contrary to the southward plunge of this particular storm that merged with Hurricane Sandy in late October 2012. The case of the unusual merger of two weather systems into a single potent and destructive force along the eastern seaboard in late October 2012 thus is most likely an example of a great event having little underlying cause.

Nonetheless, the particular synoptic development itself was highly predictable a week in advance, as a consequence of major advances in NOAA weather forecasting capabilities. This, despite the fact that it was not an event to have been anticipated either specifically or statistically from consideration of underlying climate forcings. The combination of the two weather systems, one a late-season tropical cyclone and the other an early winter-season extratropical cyclone, is a rare occurrence along the eastern seaboard. In this sense, the combined storm was a "surprise" from a climatological perspective. It does not follow thereby, however, that the resulting super-storm must have been a consequence of climate change, as some hypothesized, or that the event could not have occurred in the absence of climate change (either related to sea ice changes or other manifestations of a warming climate including warming oceans).

The physical processes leading to the event, though uncommon over the west Atlantic, are more typical in other regions. Such combinations are especially climatologically more common in the far western Pacific. There, a greater frequency of northward moving typhoons interact with the polar jet steam of east Asia, which owing to the vast expanse of the Asian land mass, begins to acquire winter-like characteristics earlier than it's Atlantic counterpart. The resulting Pacific "super storms" are less notorious since they mature over the open waters of the North Pacific and pose little threat to major metropolitan areas. Yet, the physical processes of their formation are likely an excellent analogue for the "super storm" that was initially Hurricane Sandy and then became post-tropical in the shadow of the eastern seaboard during late October 2012.

Executive Summary (See sections below for references and citations)

Concerning the issue of whether storms (Sandy and the nor’easter) are signals of the impact of climate change on extreme weather:

  • No significant increase in Atlantic hurricanes since the late 1880s has been observed.
  • The number of hurricanes that make U.S. landfall has not significantly increased or decreased.
  • In the future, it is likely that global frequencies of tropical cyclones will either decrease or remain essentially the same as a result of projected greenhouse gas induced climate change, though rainfall rates related to such storms are likely to increase.

  • There is low scientific confidence that overall storminess has changed, however, it is likely that there has been a human-induced increase in coastal extreme sea level events due to overall sea level rise.

  • Near Sandy’s landfall, sea level has risen over one foot since the mid-19 Century, mostly (but not solely) due to the increase in volume of the ocean attributable to its warming resulting from climate change.
  • It is very likely that further sea level rise will contribute to increased coastal high water levels in the future, conditions that led to Sandy’s primary impacts on coastal New York and New Jersey.

Concerning the issue of whether storms (Sandy and the nor’easter) are signals of the impact of sea ice on extreme weather:

  • Arctic sea ice extent has significantly diminished in a manner inconsistent with internal variability alone and anthropogenic forcing has likely been a contributing factor.

  • There is low confidence on changes in either the number or the intensity of mid-latitude storms, and there is also low confidence on the role played by sea ice forcing.

  • Scientific understanding remains controversial whether there is an appreciable or detectable impact of Arctic sea ice loss on subarctic weather during Fall and early winter.

  • The immediate cause for the severe U.S. impacts induced by Hurricane Sandy is the fatal, albeit random, merger of two transitory weather systems. It is very unlikely that either of these weather systems individually was appreciably affected by Arctic sea ice loss. The case of the unusual merger of two weather systems into a single potent and destructive force along the eastern seaboard in late October 2012 thus is most likely an example of a great event having little underlying cause.
  • Although there was evidence of local atmospheric changes due to record sea ice loss in 2012, the large-scale pattern of atmospheric circulation over the North Atlantic, which was dominated with rapidly progressing weather systems rather than by stationary climate anomalies, was unlikely a detectable signal of depleted Arctic sea ice.

I. Overarching Assessment of Climate Change Impacts on Sandy

The connection between storms such as Sandy and global climate change including Arctic sea ice loss continues to be a matter of active research, though there is currently no strong evidence for appreciable linkages.

No significant increase in Atlantic hurricanes since the late 1880s has been observed. Nor have the number of hurricanes that make U.S. landfall significantly increased or decreased. In the future, it is likely that global frequencies of tropical cyclones will either decrease or remain essentially the same as a result of projected greenhouse gas induced climate change, though rainfall rates related to such storms are likely to increase.

There is low scientific confidence that overall storminess has changed; however, it is likely that there has been a human-induced increase in coastal extreme sea level events due to overall sea level rise. Near Sandy’s landfall, sea level has risen over one foot since the mid-19 Century, mostly (but not solely) due to the increase in volume of the ocean attributable to its warming resulting from climate change. It is very likely that further sea level rise will contribute to increased coastal high water levels in the future, conditions that are exacerbated along the eastern seaboard by tropical cyclones and Nor’easters.

Sources:

# NOAA's State of the Science Fact Sheet: " Atlantic Hurricanes, Climate Variability and Global Warming" (May 2012)

  1. IPCC, 2012: Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation. A Special Report of Working Groups I and II of the Intergovernmental Panel on Climate Change (Field, C.B., V. Barros, T.F. Stocker,?D. Qin, D.J. Dokken, K.L. Ebi, M.D. Mastrandrea, K.J. Mach, G.-K. Plattner, S.K. Allen, M. Tignor, and P.M. Midgley (eds.)). Cambridge University Press, Cambridge, UK, and New York, NY, USA, 582 pp.

II. Review on Sea Ice, Storminess, and the North Atlantic Oscillation

A recent review article by Bader et al. (2011) examines the state of the science regarding the relationship between Arctic sea-ice, storminess and the North Atlantic Oscillation (NAO). It assesses the evidence drawn from observational and modeling studies drawing on about 250 peer-reviewed articles spanning nearly a half-century of scholarship through 2010 (several studies published subsequently are discussed in Sections III and IV). The main conclusions can be summarized as follows.

Arctic sea ice extent has significantly diminished in a manner inconsistent with internal variability alone and anthropogenic forcing has likely been a contributing factor. A poleward shift of mid-latitude storms has been observed during winter. There is low confidence on changes in either the number or intensity of mid-latitude storms, and also low confidence on the role played by sea ice forcing. A statistical relation between Arctic sea ice loss and occurrences of negative (blocked) phase of the North Atlantic Oscillation (NAO) have been observed in recent decades. Possible causal effects of the sea ice on the atmosphere are difficult to deduce from the observations because (i) to zero order the atmosphere forces sea ice, and (ii) there is strong internal variability of the NAO. Atmospheric model simulations designed to test climate sensitivity to specified Arctic sea ice change tend to reveal a wintertime blocked NAO response, but the seasonality and intensity of responses vary considerably among the different models used. Most coupled ocean–atmosphere models of the last IPCC AR4 project a significant future reduction in Arctic sea-ice with a moderate tendency to a positive (zonal) phase on the NAO. This shows that the forcing associated with the sea ice reduction is not the dominant forcing on the NAO in the future, though it does not negate the possibility that it may be a more influential factor (relative to GHG and aerosols forcing) in current climate.

Source:

Bader, J., M. Mesquita, K. Hodges, N. Keenlyside, s. Osterhus, and M. Miles, 2011: A review on Northern Hemisphere sea-ice, storminess and the North Atlantic Oscillation: Observations and projected changes. Atmospheric Research, 101, 809-834.

III. Hypothesis on Arctic Sea Ice-Sandy Linkages

Conjecture regarding a strong link between Arctic sea ice loss and Hurricane Sandy emerged in the storm’s immediate aftermath. This was based on observations of weather and climate conditions observed during the time of Sandy, which are described further below. Assumptions of linkages to sea ice were drawn by inferring how those observations fit into a recent proposed theory for how Arctic warming may affect middle latitude weather (Francis and Vavrus 2012, subsequently FV2012).

According to their hypothesis, depleted Arctic sea ice induces a deep tropospheric warming at high latitudes. An attending reduction of the pole-to-equator temperature gradient is then argued to weaken westerly winds at the jetstream level. FV2012 apply linear, steady-state barotropic Rossby wave principles to propose that large-scale atmospheric waves would slow their normal rate of eastward progression as a consequence of weakened westerlies. Further, the authors argue, the slower moving waves would acquire greater meridional amplitude.
Source:

Francis, J., and S. Varvus, 2012: Evidence linking Arctic amplification to extreme weather in mid-latitudes. Geophys. Res. Lett., 39, L06801, doi:10.1029/2012GL051000

IV. Assessment of Arctic sea ice impact on Sandy

Arctic sea ice depletion since 1979 has caused little or no warming of Arctic temperatures outside the low troposphere

Several new and independent studies do not support nor confirm a hypothesis for either an appreciable or a detectable impact of Arctic sea ice loss on mid-latitude weather during Fall. Supporting material is given below, which are also relevant in order to clarify what can, and cannot be said, concerning Arctic sea ice-Sandy linkages.

First, the FV2012 empirical study made the assumption that observed trends in Arctic temperatures throughout the deep tropospheric column during recent decades were solely caused by Arctic sea ice loss. This was key to their conjecture for the existence of strong evidence that Arctic sea ice affected mid-latitude weather patterns. Subsequent studies by Screen et al. (2012a,b) have tested such an assumption, and find contrary indications. Their work has found significant differences between the trends in recent decades and those related to an Arctic sea ice impact. Most notably, the new findings of Screen et al. reveal that reduction in Arctic sea ice does not induce a deep troposphere warming, consistent with results from other previous studies also reviewed in Bader et al. (2011). This new finding undermines the notion that a deep tropospheric Arctic amplification is a consequence of sea ice loss, a feature invoked under the FV2012 hypothesis to argue for an sea ice induced reduction in mid-latitude westerlies via weakened pole-to-equator temperature gradients.

Second, in a separate study, Screen et al. (2012) conducted one of the more realistic climate simulations to address the possible effects of sea ice loss on climate to date. This used realistic time-evolving sea ice from 1979-2009, specified as a forcing for 2 different climate models. These possessed high vertical resolution to better represent the Arctic boundary layer, and also high horizontal resolution to better model storms and the jetstream. In addition, the period was simulated 13 times so as to better identify the intrinsic variability of 30-yr trends, and thereby more clearly extract the sea ice-induce signal. Their results reveal an appreciable un-forced intrinsic variability in trends of Arctic and mid-latitude climate during this 31-year period. Indicated hereby is the difficulty and the peril of attributing the observed trend patterns during this period to Arctic sea ice loss alone, as in FV2012. The multi-model ensemble average also identified a weak sensitivity toward a trend in the negative phase of the early winter NAO, consistent with the preponderance of evidence reviewed in Bader et al. (2011) The authors state, however, that such a signal is “easily exceeded by intrinsic atmospheric variability. One implication of this result is that such a circulation response to Arctic sea ice loss may be difficult or impossible to detect in observations”. Another recent study, but focused on the 2007 extreme sea ice year (Orsolini et al. 2012), used an advanced weather forecast system. They also found little evidence for an appreciable mid-latitude climate response to sea ice loss in October or November, though an weak NAO-like signal occurred in model runs for December, yet whose phase was opposite in sign to that in Screen et al. (2012).

Figure 10 from Screen et al. (2012) showing vertical cross sections of simulated 31-yr trends during November-December for 1979-2009 of heights (shading), and zonal winds (contours, 1 m/s). Each plot is a individual run ; the super-ensemble mean is shown in lower right. Shaded ensemble mean height departures are about 10m
Figure 10 from Screen et al. (2012) showing vertical cross sections of simulated 31-yr trends during November-December for 1979-2009 of heights (shading), and zonal winds (contours, 1 m/s). Each plot is a individual run ; the super-ensemble mean is shown in lower right. Shaded ensemble mean height departures are about 10m

In sum, whereas there is compelling evidence that Arctic sea ice decline has been a major cause of Arctic amplification in near-surface warming trends over the last 3 decades, strong evidence indicates that Arctic sea ice decline has had little or no impact on temperature trends outside the low troposphere. The possible signal of remote impacts on weather and climate due to the loss in Arctic sea ice is thus currently hard to confirm and remains uncertain.

Sources:

Screen, J. A., C. Deser, and I. Simmonds, 2012a: Local and remote controls on observed Arctic warming, Geophys. Res. Lett., 39, L10709, doi:10.1029/2012GL051598.

Kumar, A., J. Perlwitz, J. Eischeid, X. Quan, T. Xu, T. Zhang, M. Hoerling, B. Jha, and W. Wang , 2010: Contribution of sea ice loss to Arctic amplification, Geophys. Res. Lett., 37, L21701, doi:10.1029/2010GL045022.

Screen, J., C. Deser, I. Simmonds, and R. Tomas, 2012: The atmospheric response to three decades of observed Arctic sea ice loss. J. Climate. doi:10.1175/JCLI-D-12-00063.1, in press.

Orsolini, Y., R. Senan, R. Benestad, and A. Melsom (2012), Autumn atmospheric response to the 2007 low Arctic sea ice extent in coupled ocean–atmosphere hindcasts, Clim. Dyn., 38, 2437-2448. doi:10.1007/s00382-011-1169-z

V. Diagnosis of Particular Conditions during October 2012

The immediate cause for the severe U.S. impacts induced by Hurricane Sandy is the fatal, albeit random, merger of two transitory weather systems. Though neither was extreme individually, they became potent and deadly upon combination. The broader circulation of a developing cold storm ensnared Sandy, redirected it's path toward the coastal margins of New Jersey, and created a favorable environment for the sustenance and additional strengthening of intense cyclonic winds as this hybrid and combined system struck the eastern seaboard on 29 October. It is very unlikely that either of these weather systems individually was appreciably affected by Arctic sea ice loss. Nor is there strong evidence that the large-scale pattern of atmospheric circulation over the North Atlantic, which was dominated with rapidly progressing weather systems rather than by stationary climate anomalies, was a detectable signal of depleted Arctic sea ice.

a) Arctic sea ice conditions

Observational estimates indicate that Artic sea ice extent achieved a new record low (relative to a 1979-2011 reference) during mid-September 2012, the normal period of the seasonal cycle minimum in Arctic sea ice extent . At its minimum in 2012, sea ice extent was roughly 50% of its climatological mean. Most of the sea ice loss took place, as in prior years, over the Arctic Ocean north of Asia, north of western North America, and within the archipelago region of northeastern Canada. From the October 1 through October 31, Arctic sea ice coverage doubled with the normal march of the seasons toward colder conditions. The monthly mean October 2012 Arctic sea ice extent, though still anomalously depleted (about 20% below normal), was no longer of record proportion.

1)
2)
3)

b) Atmospheric conditions and the track of Sandy

A schematic of the circulation pattern at the time of Sandy, drawn for a period when Sandy was a minimal hurricane upon exiting the Caribbean Sea on 26-27 October, reveals a key catalyst to the ensuing meteorological events--- a strong late Fall storm in the upper troposphere with an attending meander of the jetstream. As this storm evolved through the typical process of baroclinic development, this initially open wave transformed into a closed circulation along the Mid-Atlantic seaboard by 29 October, sequestering Sandy and drawing it westward as the baroclinic wave completed its cyclogenesis. It is owing to this large-scale process of cyclogenesis, a phenomenon quite common along the eastern seaboard, that numerical weather prediction models were so successful in predicting the precise path, timing, and location of Sandy’s landfall by up to a week in advance.

An animation of daily 250 mb heights, beginning on 19 October some 10 days before Sandy’s landfall, and continuing through 4 November illustrates the highly transitory character of conditions over the eastern U.S. and the west Atlantic, entirely consistent with weather driving. The notion that Sandy assumed an unusual path (westward to the eastern seaboard) due to the influence of a persistent North Atlantic block is inconsistent with this synoptic evidence. The conditions in late October 2012 were characterized by highly transient (not steady state) evolving weather patterns., and the re-direction of Sandy westward on 29 October was a direct consequence of the lifecycle of baroclinic development attending the swiftly moving extratropical storm, not from the actions of a stationary blocking anticyclone. NOAA’s monitoring of a blocking-index revealed a sequence of discrete blocking anticyclone events over the North Atlantic, and do not confirm the presence of stationary feature as some hypothesized based on prescence of steady boundary forcing. The blocking events retrograde (move westward) decaying and reforming in concert with individual migrating extratropical storms. The combination of Hurricane Sandy and the large- scale extratropical storm that absorbed and transformed it into a post-tropical cold-core system by 29 October acted to transport significant quantities of tropical/subtropical warm air poleward leading to intensification of the blocking index over the North Atlantic. But this too was transitory, and blocking in the North Atlantic was all but absent by the beginning of November.