Winter extremes and natural variability

So far I’ve looked at patterns of precipitation and temperature, and also related drought, determined by natural modes of variability. Is there also a connection with snow anomalies and extreme winters? Yes, of course there is.

The winter of 2009/10 had anomalously large snowfall in the central parts of the US and in northwestern Europe, which some of you may remember. December’s snowcover in the contiguous US was the greatest ever for that month and Washington D.C. had to be closed down for a weak.

Washington D.C., December 19, 2009. Here

As mentioned in the beginning of this blog, the NAO exerts a strong influence on wintertime climate across the North Atlantic basin, and a negative index NAO give cold temperature anomalies in the eastern US and northern Europe. ENSO, with its warm phase El Niño, also influences wintertime climate over the US by a southward displaced stormtrack. A paper by Seageret al. (2010) takes a closer look at the winter of 2009/10 and finds that a combination of a negative NAO and an El Niño event was causing the cold temperatures and the anomalous amounts of snow experienced in parts of the US and Europe that winter. The changes in storminess associated with El Niño explained snow anomalies in western, central and southern US, while the negative NAO provided sufficient cold air for the precipitation to fall as snow in the northeastern US and northern Europe.

In a study published in Science in 2001, Thomson and Wallace find that the different phases of the NAO (NAM, see link) are marked by distinct differences in the frequency distribution of significant weather events in the Northern Hemisphere that impact human activities. They report that cold events (daily minimum temperatures dropping below a specified threshold, frozen precipitation) occur with much greater frequency over North America and Europe (and other places, see table below) during negative index days, increasing the risk of frost damage and the frequency of snowfall.

Significant weather events associated with high and low NAM index days. As you can see most of the events occur with much higher frequency during NAM- (negative index). 

So heavy snowfall events and extreme winters, like the winter of 2009/10, are also influenced, to a big extent, by natural modes of variability. But is there a relation to climate change? If one is thinking about climate change in general, I think most people will relate it to warming. The winter of 2009/10 made people really question climate change. I’ll be discussing whether or not winter extremes are related to anthropogenic climate change in my next post.

Drought in the context of anthropogenic climate change

In the previous post I talked about European droughts association to natural modes of variability. Since this is a blog about natural versus anthropogenic climate change I thought it would be natural to follow up with a question about how anthropogenic influence on climate is affecting droughts. In particular, considering the severe drought in the US the last summers, this question is of high relevance.

A relatively newly released study in Nature by Sheffield et al. look closer at the historical record of global-scale drought trends and actually find that they have most likely been overestimated. They report high uncertainties in these trends over the past 60 years and little evidence of an increase in the total area affected by drought. One might think that it is the opposite way around in these times of a warming climate, but the climate system is highly complex and we don’t necessarily get the results that seem most logic. There exists a hypothesis in the scientific world saying that “wet is getting wetter and dry is getting drier”, meaning that the areas which normally receive a lot of precipitation will get more intense rainfall and flooding, and that areas that are already suffering from a precipitation deficit and drought will get more severe droughts. This will have huge consequences for people living in such areas.

Another study produced by a big group of scientist for the American Meteorological Society looks at six extreme events during the year of 2011 and tries to explain them from a climate perspective. Among these, they look at the severe 2011 Texas drought and ask: “Was the likelihood of either the heat wave or the drought altered by human influence on global climate?” Now considering drought over the North American continent, it is important to point out that ENSO with its cold phase La Niña, is considered to be a key driver of drought conditions in the central US (Atlas et al. 1993). In the study they use the La Niña year of 2008 as a proxy for 2011, because simulations under 2011 forcing conditions were not available, and compare to earlier decades. They find “that extreme heat events were roughly 20 times more likely in 2008 than other La Niña years in the 1960s and indications of an increase in frequency of low seasonal precipitation totals.” These findings suggest that drought is more probable now than for 40-50 years ago.

Picture from Texas drought 2011, Google

This also contribute to strengthen the “dry getting drier” hypothesis by saying that the probability of the occurrence of drought in an already dry area like Texas, is more likely with global warming.

European drought’s relationship to global SST

So far on this blog I’ve described natural modes of variability across the North Atlantic basin like the NAO, the AMO and AMOC and their related climatic impacts. They are all associated with specific climatic patterns of temperature and precipitation across large areas, and I’ve mentioned and showed illustrations of these in my previous posts. In this post I will only focus on the drought impact and try to give a summary of the factors controlling droughts across Europe.

A study by Ionita et al. (2012) looked at variability of European summer drought and its relation to global sea surface temperature (SST) by using the Palmer drought severity index averaged over the European region (see figure below).

 

This time series show the strong interannual and decadal variability of European drought. The authors show that winter SST has a strong impact in determining drought variability over Europe in the upcoming summer through different large-scale teleconnection patterns. By the use of correlation analysis they reveal the existence of three coupled modes of summer drought pattern and winter SST anomalies with different timescales:

1.    The first coupled mode represents the long-term warming trend in global SST caused by anthropogenic greenhouse gasses, in addition to a tripole-like pattern in SST resembling the positive phase of the NAO.

2.    The second coupled mode is associated with an inter-annual SST pattern in the Pacific which resembles the cold phase of ENSO (La Niña) together with the decadal fluctuation in extratropical SST resembling the Pacific DecadalOscillation (PDO).

3.    The last coupled mode is associated with strong multidecadal variability in SST across the Atlantic basin which corresponds with the AMO, also for the interannual variability. In a previous post I described how the AMO exerts a strong influence on European summertime climate, including drought. As they write in the paper: “According to Briffa et al. (2009) the summers of 1921, 1976, and 1990 were among the driest in the last 250 years, all these dry summers occurring during a cold North Atlantic phase of the AMO.”

I’ve previously described the NAO- and the AMO’s impact of heat and drought on the European climate. It now turns out that drought across Europe is associated with four different modes of variability! (NAO, ENSO, PDO and AMO). Drought is not an easy thing to define given the complexity of the phenomenon, and there also exists several types of drought. So to say that drought across Europe is determined exactly by these four modes of variability is of course just looking at the big picture.

AMO vs global warming

I ended my last blog post by asking a few questions related to the future of the North Atlantic Ocean and the AMO. What can we expect looking forward? As I also mentioned in the end of the last post, the AMO is very hard to model and therefore also to predict. Almost all models have a difficulty in simulating the AMO variability. So we really don’t know what to expect into the future…

But we do have knowledge and numerous studies which we can draw theories from. We do know that the AMO is a dominant mode of variability in the North Atlantic SST with a duration of 55-80 years (Wei & Lohmann 2012). But the question we are asking these days, and the same question I asked in my previous post, is to what extent the North Atlantic Ocean is influenced by global warming relative to the internal variability (AMO), and what consequences that may have for the future.

A study based on observational data by Wang and Dong in 2010 finds that both global warming and AMO variability make a contribution to the recent warming in the North Atlantic basin, and that their relative contribution is approximately equal. They also find (after removing a linear trend and the seasonal cycle found in observational records) that atmospheric CO₂ anomalies show a multidecadal variation approximately coinciding with the cold and warm phases of the AMO (see figure below).

They discuss that this may be related to the ocean’s CO₂ uptake through ocean circulation and the strength of the AMOC. As mentioned in my previous post, the phases of the AMO are determined by AMOC variability. There also exists a relationship between solubility of CO₂ and ocean temperature, which says that a warmer ocean leads to a release of CO₂. So a warm (cold) phase of the AMO will lead to a release (uptake) of CO₂ to (from) the atmosphere.  Summarized the warming of the North Atlantic due to AMO variability may influence global SST via the increase of atmospheric CO₂.

So based on findings from this study it seems like there exists a two-way relationship between North Atlantic Ocean temperatures and CO₂ concentrations in the atmosphere. Increased amounts of atmospheric CO₂ contributes to a warming of the North Atlantic basin. The other way around, the warm phase of the AMO with anomalously warm SSTs contributes to an increase in atmospheric CO₂, which further increases global warming. With future predictions of further CO₂ increases this does not look good. But then again the AMO is a mode of natural variability internal to the climate system which likely will change to a negative phase sometime in the future, which will reverse the picture. It is pretty clear, considering the huge climatic impacts of the AMO, that further research is needed in this field.

A shift in the European climate in the 1990 linked to the AMO

During the 1990s there was a substantial shift in the European climate. We experienced more wet summers in northern Europe and more hot and dry summers in southern Europe relative to earlier summers. What caused this anomalous shift in European climate? Was it related to anthropogenic climate change or is it just a case of natural variability in the climate system? 

A recently published paper in Nature by Sutton & Dong explains that it is the North Atlantic Ocean and a pattern of variability known as the Atlantic Multidecadal Oscillation (AMO) that is responsible for this shift in European climate. The AMO is a multidecadal variation in North Atlantic sea surface temperature (SST) which fluctuates between anomalously warm and anomalously cool phases, each lasting several decades at a time. Its instrumental record is shown below, where you can see the shifts about the mean state over the years.

As you can see there was a shift towards a positive phase in the mid 1990s, and Sutton & Dong finds evidence that this has caused the climate shift in European summers from 1996-2010. AMO-like variations in SST are closely related to the variations in the Atlantic Meridional Overturning Circulation (AMOC) (an Atlantic circulation feature part of the global ocean circulation which transports heat from lower- to higher latitudes, where the water gets so dense that it sinks and deepwater is formed, thereby overturning circulation). Evidence suggests that the 1990s warming of the Atlantic Ocean was largely caused by an acceleration of the AMOC in response to the persistent positive phase of the winter NAO, which I’ve talked about in previous posts.

As you also can see from the above figure, there was another warm period between about 1930 and 1960, with very similar pattern of North Atlantic SST as for the recent warm period. Previous research has shown that this warm state, relative to the cold period 1960-1990, forced a certain pattern of sea level pressure (SLP), surface air temperatures (SATs) and precipitation over Europe. So the similarities in North Atlantic SST anomalies between the two warm periods suggest that similar climate impacts may have been excited in the recent warm period. And that is true – very similar patterns in SLP, SATs and precipitation does exist for the two warm periods for spring (MAM), summer (JJA) and autumn (SON). This is shown in the model simulation figures below (which show anomalies relative to the intervening cool phases), which also show the typical climate pattern associated with a warm phase of the AMO:

SLP

SAT

Precipitation

This is in close agreement with the observed recent wet northern Europe summers and hot and dry southern Europe summers. It is also consistent with the observed conditions for recent springs and autumns. As Sutton & Dong explains;

“The consistency between the two warm North Atlantic periods in the patterns of anomalies in SAT, precipitation and SLP is strong circumstantial evidence that the North Atlantic Ocean was an important driver of these decadal changes in European climate.”

So what can we expect into the future? How long will the AMO stay in its warm phase, keeping the European climate locked in the same pattern? And are there also external forcings contributing to the recent warming, like anthropogenic greenhouse gasses? These are hard questions to answer given that the AMO is a natural mode of variability internal to the climate system, varying with no particular pattern, which makes it hard to predict in the future. I will try to dig deeper into this for my next blog post and further discuss these questions.

Atlantic hurricane activity and climate change

I think it is about time that I mention hurricane Sandy here on my blog. The “Superstorm” is the largest Atlantic hurricane on record and we’ve all heard about the devastating effects it had on parts of the Caribbean and the US east coast. It is easy to conclude that this unusually intense storm was caused by climate change and global warming, as a relatively newly released article in the Guardian Environment discusses.

I decided to do my own little research in the field and read a paper published in Nature by Knutson et al. (2010) on tropical cyclones and climate change. In this study they try to answer whether past changes in tropical cyclone activity have exceeded the variability expected from natural causes. First they state that there is a relationship between tropical Atlantic SSTs and the upward trend in the Atlantic hurricane activity. They then go on by looking at cyclone- (1) frequency, (2) intensity, (3) rainfall, and (4) genesis, tracks, duration and surge flooding separately. It is hard to model tropical cyclone activity because it depends on so many constantly changing factors (such as tropical SSTs), but with the improvements in models and analyzing techniques, the authors could raise their confidence level concerning cyclone-activity projections and conclude the following:

1.       It remains uncertain whether past changes in tropical cyclone frequency have exceeded the variability expected through natural causes. In the future it is likely that global mean tropical-cyclone-frequency will either decrease or remain essentially unchanged owing to greenhouse warming. Among the proposed mechanisms is the weakening of the tropical circulation.

2.       Future surface warming and changes in the mean thermodynamic state of the atmosphere (as projected by climate models) will lead to an increase in tropical cyclone intensity – both in the mean intensities and in the frequency of cyclones at higher intensity levels.

3.       Atmospheric moisture content has increased in recent decades in many regions, and will continue to increase as the atmosphere warms. This should increase rainfall rates in systems such as tropical cyclones, but that has not been established by existing studies. Tropical-cyclone-related rainfall rates are likely to increase with greenhouse warming though.

4.       There is no conclusive evidence that any observed changes in tropical cyclone genesis, tracks, duration and surge flooding exceeded the variability expected from natural causes. However, with highly confident predictions of future sea-level rise, costal environments are more vulnerable to storm-surge flooding.

      Time series of late summer tropical Atlantic sea surface temperature (blue) and the Power Dissipation Index (green), a measure of hurricane activity which depends on the frequency, duration, and intensity of hurricanes over a season. From Emanuel (2007). As you can see there exists a very close relationship.

So it seems like climate change and global warming is not directly affecting hurricane activity in the Atlantic Ocean, but more indirectly through ocean warming, sea-level rise and circulation changes. Concerning hurricane Sandy, anthropogenic activity has definitely played at least a supporting role in its intensity and destructiveness.

It was like Sandy was a sign from above to the American people, right before the election, that they and their political leaders should open their eyes and realize that climate change is happening and needs to be integrated in their politics. If hurricane Sandy had a saying in who won the election, we can literally thank God Obama won. That way there will at least be some effort to slow down the rate of anthropogenic climate change, easing future damage by superstorms like Sandy.

Tropical SST relation to the strong NAO phase

In my last post I discussed and reviewed a paper about the reason for the strong positive phase in the NAO index and the associated warming over Europe, and concluded that it is not just internal variability of NAO – forcing factors such as anthropogenic greenhouse gasses do play a central role.

I would like to point out a couple of other studies done by Hoerling et al. (2001) and Hurrell et al. (2004) that investigates another forcing factor on winter North Atlantic climate; low-latitude sea surface temperatures (SSTs). They investigate tropical SST forcing on the NAO winter index by using atmospheric general circulation models (AGCMs) forced with the observed evolution of global SSTs since 1950. They find that variance in tropical SSTs in their models simulates a positive trend in the NAO index, and the special pattern of the simulated trend agrees with that observed. They present evidence that tropic-wide changes in the atmospheric circulation associated with warming surface waters over the tropical Indian and western Pacific Oceans and associated rainfall over the tropical Indian Ocean produce a North Atlantic anomaly pattern very much like the positive index phase of the NAO.

Hurrell et al. published a part two paper (Hoerling et al. 2004) following up their first study, where they explore their previous findings. They support their earlier arguments and state that SST warming over the Indian Ocean sector is the key forcing mechanism for the observed trend in the NAO, pointing out the striking similarity in the time series of Indian Ocean and North Atlantic climate variations (see figure below).

They also discuss whether this warming contains an anthropogenic component and conclude that it most likely contains a signature of anomalous greenhouse gas forcing. They’re then using this as an argument to why models are not correctly simulating the magnitude of the NAO phase (as I described in the previous post); “Our theory for an anthropogenic, dynamical oceanic forcing of North Atlantic climate change by Indian Ocean SST anomalies might clarify why neither unforced AGCMs nor unforced coupled models are able to produce winter NAO index trends of the magnitude observed since 1950”.

So again it does look like we’re to some extent causing climate change and warming over the Euro-Atlantic part of the hemisphere with our increasing emissions of greenhouse gasses, though through a different mechanism from these papers point of view.