Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, CLIMATE SCIENCE (climatescience.oxfordre.com). (c) Oxford University Press USA, 2016. All Rights Reserved. Personal use only; commercial use is strictly prohibited. Please see applicable Privacy Policy and Legal Notice (for details see Privacy Policy).

date: 22 August 2017

North Atlantic Oscillation

Summary and Keywords

Many variations in the weather in the European and North Atlantic regions are linked with changes in the North Atlantic Oscillation (NAO). The NAO is measured using a south-minus-north index of atmospheric surface pressure variation across the North Atlantic and is closely connected with changes in the North Atlantic atmospheric polar jet stream and wider changes in atmospheric circulation. The physical, human, and biological impacts of NAO changes extend well beyond weather and climate, with major economic, social, and environmental effects. The NAO index based on barometric pressure records now extends as far back as 1850, based on recent work. Although there are few significant overall trends in monthly or seasonal NAO (i.e., for the whole record), there are many shorter-term multidecadal variations. A prominent increase in the NAO between the 1960s and 1990s was widely noted in previous work and was thought to be related to human-induced greenhouse gas forcing. However, since then this trend has reversed, with a significant decrease in the summer NAO since the 1990s and a striking increase in variability of the winter—especially December—NAO that has resulted in four of the six highest and two of the five lowest NAO Decembers occurring during 2004–2015 in the 116-year record, with accompanying more variable year-to-year winter weather conditions over the United Kingdom. These NAO changes are related to an increasing trend in the Greenland Blocking Index (GBI; equals high pressure over Greenland) in summer and a significantly more variable GBI in December. Such NAO and related jet stream and blocking changes are not generally present in the current generation of global climate models, although recent process studies offer insights into their possible causes. Several plausible climate forcings and feedbacks, including changes in the sun’s energy output and the Arctic amplification of global warming with accompanying reductions in sea ice, may help explain the recent NAO changes. Recent research also suggests significant skill in being able to make seasonal NAO predictions and therefore long-range weather forecasts for up to several months ahead for northwest Europe. However, global climate models remain unclear on longer-term NAO predictions for the remainder of the 21st century.

Keywords: climate change, extreme events, Greenland blocking, North Atlantic Oscillation, predictability

Introduction

The North Atlantic Oscillation (NAO) is a prominent “seesaw” of atmospheric surface pressure fluctuation between the Azores and Iceland that has been meteorologically well defined since at least the late 19th century (e.g., Hurrell, Kushnir, Ottersen, & Visbeck, 2003). It is defined using the NAO index, which is typically a normalized mean sea-level pressure (SLP) index between a southern station located in the Azores or continental Iberia and a northern station in western Iceland (Cropper, Hanna, Valente & Jónsson, 2015; Hurrell, 1995; Jones, Jónsson, & Wheeler, 1997; van Loon & Rogers, 1978). The NAO has historically been recognized since at least the time of the Vikings; pioneering work based on early instrumental meteorological records was undertaken by Hildebrandsson (1897), who using surface air pressure data discovered the inverse relation between Iceland and Azores pressure, and by Sir Gilbert Walker who in works published in 1924 and 1932 (the latter with Bliss) undertook correlation analysis and constructed a robust multivariate NAO index based on surface air pressure and surface air temperature data from several European stations (Stephenson, Wanner, Brönnimann, & Luterbacher, 2003).

The strength of the pressure difference between the high- and low-SLP centers of action exerts a strong control over the strength and direction of the mid-latitude westerly storm tracks. As such, the NAO has been linked to a variety of climatological, biological, hydrological, and ecological variables across several locations (Ottersen et al., 2001; Westgarth-Smith, Roy, Scholze, Tucker, & Sumpter, 2012) but is most frequently recognized as directly affecting the west of Europe (from Iberia to Scandinavia) and North America. A greater than normal pressure difference between the Azores and Iceland is a positive NAO, and a weaker than normal pressure difference is a negative NAO. During the winter months, a positive NAO is associated with warmer and wetter conditions across northwest Europe and cooler and drier conditions across southern Europe as the stronger pressure gradient between the Azores and Iceland drives the storm tracks poleward (Fig. 1). The opposite is generally true for negative NAO conditions as the weaker pressure gradient generally results in southward-shifted storm tracks, and a SLP reversal will typically result in more easterly conditions. As such, the NAO index is strongly related to favored positions of the North Atlantic atmospheric polar jet stream (Hall, Erdélyi, Hanna, Jones & Scaife, 2015; Overland et al., 2015; Woollings et al., 2015).

North Atlantic OscillationClick to view larger

Figure 1. Correlation coefficient between the Hurrell PC NAO and the ERA-Interim reanalysis (1979–2015; Dee et al., 2011) for winter/summer (DJF/JJA): (a/b) sea-level pressure; (c/d) 2-meter near-surface air temperature; and (e/f) surface precipitation. The solid black lines denote areas of 95% significance. The green vectors in plots (a) and (b) denote the regression relationship between 500 hPa-level (mid-troposphere) winds and the NAO.

Another way to define the NAO, the principal component (PC)–based NAO index (e.g., Hurrell, 2015), uses the first empirical orthogonal function (EOF) of atmospheric pressure variability across the North Atlantic region and is strongly correlated with the station-based index. NAO indices are closely related to the Arctic Oscillation (AO) index—the latter being the first EOF of variability of atmospheric surface pressure across the whole Northern Hemisphere north of 20°N (Thompson & Wallace, 1998)—but there are subtle and notable differences in NAO and AO variations (Hanna, Cropper, Jones, Scaife, & Allan, 2015), and the NAO can perhaps best be seen as the regional Atlantic-wide manifestation of the AO.

Striking positive trends in the NAO and AO between the 1960s and 1990s were noted around the late 1990s and early 2000s (e.g., Gillett, Graf, & Osborn, 2003). However, later this trend reversed, with some extreme negative excursions from the mid-2000s (e.g., Hanna et al., 2015).

For this article we provide a broad discussion of the general state of research regarding the NAO and the importance of the Oscillation in its effect on European climate and wider-scale impacts. We then present some of the latest research in the analysis of seasonal and monthly NAO variability for the available period of record at the time of writing (November 2015) and comment on observed trends and extreme events. We interpret recent NAO changes, since about 1990, in the context of the extended record. Finally, we summarize some recent research findings on NAO seasonal predictability and potential NAO changes under future climate change until 2100.

NAO Physical Structure

The NAO and AO are preferred modes of variability of atmospheric circulation in the Northern Hemisphere (Fig. 1). The alternative designation Northern Annular Mode (NAM, which is the same as the AO) is not as widely used as its southern counterpart the Southern Annular Mode (Abram et al., 2014). The NAO can be regarded as an Atlantic sector regional expression of the hemisphere-wide AO (or NAM). The NAO reflects changes in the position and strength of the North Atlantic polar front jet stream and has associated effects on the weather and climate of mid-to-high latitudes within and around the Atlantic (Fig. 1). A more positive (negative) NAO/AO index represents stronger (weaker) airflow around the Northern Hemisphere and a jet stream that is shifted further north (south) over the North Atlantic.

The NAO/NAM pattern is a result of the eddy-driven extratropical atmospheric circulation: specifically, the transport of heat and momentum by stationary eddies (longwaves or planetary waves in the northern polar jet stream) and transient eddies (cyclones and anticyclones forming within or along the jet stream) (e.g., Kaspi & Schneider, 2013). The polar jet stream is directly related to NAO changes and has a mean latitude somewhere between 50°N and 60°N over the eastern North Atlantic. The strongest westerly winds (of up to about 200 km/hr in the core of the jet near the tropopause) are typically experienced at these latitudes, and there is a clear clustering of extratropical storm tracks along the polar jet stream. The prevailing direction is westerly due to the Coriolis Effect of earth’s rotation, which deflects air masses to the right of their direction of motion in the Northern Hemisphere. Longwaves develop in the jet stream because of orographic obstacles (e.g., the Rocky Mountains over North America) or east–west heating contrasts between land and sea, or variations in latent heating due to condensation and rainfall. Low- and high-pressure systems form due to strong horizontal contrasts in temperature, typically where cold polar air meets relatively warm tropical air masses. These transient eddies are very important in providing energy for maintaining the polar jet stream flow and mid-latitude westerlies, otherwise friction with the surface would slow and eventually halt the winds. However, a significant contribution to maintaining the westerlies—greater than in the Southern Hemisphere—comes from the stationary eddies: this is due to the much stronger land–ocean contrast effects in northern mid-latitudes.

Being linked with the jet stream, there is a deep and pronounced vertical structure to the AO and NAO, which extends up into the stratosphere; this is most notable for the AO, which lies further north and is more directly linked with the polar vortex. What happens in the stratosphere in polar winter can also have a big bearing on conditions in the troposphere: for example, stratospheric sudden warmings are associated with a weakening and sometimes reversal of the polar vortex and development of negative NAO/AO that sometimes occurs in mid- to late winter (e.g., Cohen et al., 2014; Marshall & Scaife, 2010). Stratosphere–troposphere interaction and coupling is not very well understood, yet is important for NAO dynamics (Kidston et al., 2015). It appears from theory and observations that planetary-scale Rossby waves can propagate upwards from the troposphere into the stratosphere under conditions of moderate westerly flow during boreal winter; the stratosphere is effectively decoupled from the troposphere in other seasons. If the wintertime polar vortex is weak (strong), the upward-propagating waves can (cannot readily) interact with and slow the upper-level westerly flow. There is also a kind of reverse effect where airflow anomalies in the stratosphere can propagate down to affect the near-surface circulation (Baldwin & Dunkerton, 2001). The time of operation of these changes is typically 2–3 weeks, although dynamical couplings range over timescales from daily to multidecadal (Kidston et al., 2015).

The NAO/AO exist in atmosphere-only computer models of the global climate system, so the NAO does not depend on the ocean for its existence. However, it is thought that low-frequency variations of oceanic circulation in the Atlantic—which comprise the Atlantic Multidecadal Oscillation (AMO)—are driven at least partly by the NAO (Clement et al., 2015). The AMO may in turn drive the NAO in a distinctly seasonal response, with a warm AMO phase promoting a negative NAO in winter (Gastineau & Frankignoul, 2015).

The AO and NAO tend to be strongest in winter because this is when one of the key factors driving the jet stream—the equator-pole temperature gradient—is greatest, due to less seasonal cooling in the tropics than the polar regions. The jet stream and NAO correspondingly tend to weaken in summer. The summertime NAO is shifted northwards, having its southern node over northwest Europe and a smaller accompanying pressure pattern. A detailed discussion of NAO patterns and changes during this season is provided by Folland et al. (2009), who develop a summer NAO index that they compare with changes in European summertime climate. These authors also note a small but significant link of interannual variation in SNAO with La Niña sea-surface temperatures in the eastern tropical Pacific, as well as a persistent link with the AMO.

Due to the seasonal migration of the jet stream, which generally lies further south (north) in winter (summer), the association between NAO and, for example, British weather conditions varies markedly depending on the season. Thus, in winter, a positive (negative) NAO is generally associated with mild, wet (cold, dry) weather over the United Kingdom but a positive (negative) NAO in summer is often linked with dry and sunny (wet and cool) conditions. Key examples are the exceptionally cold 2009/2010 U.K. winter, with a record low (Hurrell PC) NAO value of -2.93 and the exceptionally wet 2007 and 2012 U.K. summers with low NAO values of -1.15 and -1.59. However, the NAO-U.K. weather relation in summer is less clear than in winter: for example, summer 2015 had a similarly low NAO value of -1.61 but had only slightly above average rainfall, e.g., 113% of the 1981–2010 summer average for U.K. precipitation (http://www.metoffice.gov.uk/climate/uk/summaries/2015/summer). While the NAO is the single most important factor determining changes in weather and climate over the North Atlantic region, it does not explain everything, and most notably the East Atlantic and Scandinavian atmospheric circulation patterns—with respective main centers to the west of Ireland and around Bergen, Norway—also need to be considered (Moore, Renfrew, & Pickart, 2013).

NAO Impacts

Large excursions in the NAO index, both positive and negative, are often associated with extreme/unusual events. To illustrate how the hemispheric flow can alter with markedly different NAO regimes, Figure 2 compares two winter seasons, which represent some of the most negative (2009/2010) and positive (2013/2014) recorded excursions in the NAO index. Both animations start on November 1 of their respective seasons and display a typical, westerly (from the west) flow regime throughout November. During 2009/2010 (Fig. 2a) during the last few days of November and through the first half of December, westerly storms continue to propagate across the Atlantic but do not manage to track across the United Kingdom. This allows northerly and easterly flow to intrude and begin to dominate the wind patterns above the United Kingdom throughout the rest of winter season (although the occasional westerly system does transit across the United Kingdom). The winter season of 2013/2014 (Fig. 2b) displayed opposite conditions, with a high frequency of storms buffeting the United Kingdom, advected rapidly across the Atlantic by a strong jet stream (Huntingford et al., 2014). The impacts of the different NAO regimes on the United Kingdom contrasted greatly; 2009–2010 was a cold, snowy winter, with travel delays, frost/ice-related accidents, power disruptions, school closures, and avalanches (Osborn, 2011; Prior & Kendon, 2011) whereas the succession of storms throughout the 2013–2014 winter was the wettest winter in the 248-year England and Wales Precipitation record, which caused flooding in over 7,000 U.K. homes and businesses (Kendon & McCarthy, 2015). The contrasting 2010 and 2014 NAO seasons were analyzed by Rivière and Drouard (2015), who suggested that changes in the Pacific jet stream drove the variability in these two seasons.

Figure 2a. Animation of near-surface wind speed and direction from the NCEP/NCAR reanalysis (Kalnay et al., 1996) for November–March 2009/2010. The 2009/2010 season represents a strongly negative NAO winter.

Figure 2b. Animation of near-surface wind speed and direction from the NCEP/NCAR reanalysis (Kalnay et al., 1996) for November–March 2013/2014. The 2013/2014 season is an example of a strongly positive season.

NAO impacts on “extremes,” typically defined as those set out according to the Expert Team on Climate Change and Indices have also been studied. Scaife, Folland, Alexander, Moberg, and Knight (2008) used a general circulation model to examine the effects of the positive NAO trend during the 1960s–1990s and found that the NAO was likely responsible for an increase in the frequency of winter precipitation above the 90th percentile and the numbers of frost days and temperatures below the 10th percentile across Europe. Using daily data from central England, Kenyon and Hegerl (2008) found that the temperature distribution narrows under positive NAO conditions, leading to fewer days with maximum and minimum temperatures below the 10th percentile than under negative NAO conditions. As fluctuations in the NAO correspond to changes in the large-scale atmospheric circulation, correlations between extreme weather and the NAO are to be expected (and align with the spatial pattern of correlations presented in Fig. 1). Logically, stronger relationships in winter might be expected, when atmospheric circulation features are most dominant, whereas in summer cloud cover has been shown to exert a stronger control on temperatures than circulation metrics (Moberg, Alexandersson, Bergström, & Jones, 2003). However, strong NAO-extreme temperature and precipitation relationships have been found in both winter (Burić et al., 2014 Diao, Xie, & Luo, 2015; Donat et al., 2014) and summer (Casanueva, Rodríguez-Puebla, Frias, & González Reviriego, 2014).

The NAO has been linked with a variety of meteorological and non-meteorological effects across a wide spatial and multiple temporal scales, and only a selection of these impacts can be mentioned here. For example, Nesje, Lie, and Dahl (2000) showed a strong relationship between the mass balance of Scandinavian glaciers and the NAO due to the controlling influence of the storm tracks by the NAO, which influenced precipitation amounts, and glacier mass balance as a result. Coincidentally, the NAO has been shown to explain a large amount of the variance in Norwegian streamflow (55%) and hydropower output (30%), influencing electricity consumption and prices (Cherry et al., 2005). Baltic sea-ice extent is also strongly related to NAO changes (Karpechko, Peterson, Scaife, Vainiko, & Gregow, 2015). Cropper, Hanna, and Bigg (2014) found an influence of the NAO as far south as 20°N in coastal upwelling-inducing winds along the northwest African coastline. The great-circle distance between northwest Africa and Scandinavia is ~5,700 km, indicating the great spatial extent of the NAO influence. Recent NAO–climate linkages literature includes a strong signal of the (non-summer) NAO on precipitation in Iraq (Khidher & Pilesjö, 2015), an influence on sea-ice breakup date in south-central Ontario (Fu & Yao, 2015) and even a Southern Hemisphere influence, via a decadal-scale mechanism, on subtropical eastern Australian rainfall (Sun, Feng, & Xie, 2015a).

Ottersen et al. (2001) provide a comprehensive overview of the ecological effects of the NAO. They found a huge variety of NAO impacts of various environmental and biological parameters and summarized the proposed mechanisms of the NAO linkage. Recent examples of NAO-ecological linkages include Dolphin strandings on the eastern U.S. coastline (Harry, 2015), mangrove extent along the coastline of French Guiana due to the NAO low-frequency modulation of waves (Walcker et al., 2015), and a change in the environmental location of marine top predators in the eastern North Atlantic Ocean (Ramos et al., 2015). The detection of an NAO signal in all of the above studies highlights the important and prominence of the oscillation as the major mode of variability across the North Atlantic sector, with its potential impacts extending even further than the direct zone of spatial influence.

The NAO has also been shown to directly influence energy-generating capabilities. Colantuono, Wang, Hanna, and Erdélyi (2014) identify a negative relationship between the NAO and solar radiation availability across the United Kingdom, also showing a clear and intriguing zonal contrast between west and east regions, which they attribute to a topographic rain-shadow effect (more clouds and rain in the west of the United Kingdom under a positive NAO can sometimes be linked with cloud breakup and clear, sunnier weather in eastern England). Jerez et al. (2013) identify that for southern Europe, negative NAO conditions enhance hydropower resources and wind power by up to 30% while diminishing solar potential by 10–20% (the contrasting influence on solar availability in these studies is a function of the spatial locations analyzed, e.g., Fig. 1 shows they are regions which correlate differently with the NAO). Curtis, Lynch, and Zubiate (2016) show that the NAO-induced variability in the Irish electrical grid could cause detectable signals in the total carbon dioxide emissions from the system. Ely, Brayshaw, Methven, Cox, and Pearce (2013) call for improved understanding of the potential effects of the NAO on European power generation, as they surmise that under negative NAO conditions, lower temperatures, and less wind power generation across the United Kingdom/Scandinavia lead to an increased demand and lowered supply. On this note, while negative NAO conditions may be less favorable for the U.K.-Scandinavian renewable energy system, for regions like Iberia, the European Alps, or the Middle East, negative NAO conditions may be more favorable (Beniston, 2012; Sowers, Vengosh, & Weinthal, 2011; Trigo et al., 2004). The energy industry is already very weather-dependent due not only to demand fluctuations with temperature and hence the NAO but also because Germany and other countries rely so heavily on wind power. Therefore, NAO predictability will be an even more valuable asset in the medium- to long-term future, when renewable energy sources are expected to contribute much more significantly toward total power generation.

It is assumed that the NAO has been a significant influence on much longer temporal scales than we can directly observe (i.e., back to the mid-19th century). Knowledge of the NAO phase and variability from further back in time is derived from climate “proxy” records. A proxy record is an environmental archive from which usable information can be extracted: examples include tree rings, sediments, and ice cores, of which Jones et al. (2009) provide a thorough review. Trouet et al. (2009) created an NAO reconstruction using a drought index from Morocco using tree rings from the Atlas Mountains and a precipitation record from a stalagmite from a cave in Scotland, which served as a good proxy-NAO index, as these locations correspond to the hydrological centers of action of the NAO index (Fig. 1). A problem with proxy reconstructions, however, is that while the location of the proxy environment stays constant in time and space, the variability of the atmospheric teleconnection may not, i.e., the NAO centers of action can change (Lehner, Raible, & Stocker (2012) advance the work of Trouet et al. (2009) by assessing this issue). It was previously assumed that a persistent positive NAO state was responsible for the Medieval Warm Period—a localized North Atlantic warm period between 950 and 1200 ce; however, Ortega et al. (2015) have recently shown that this was not necessarily the case. Noise inherent in the proxy series, which arises due to non-NAO variance, further conflates the assessment of centennial NAO reconstruction signals. On longer-than-centennial timescale reconstructions, the NAO often gets cited as either a driver of, a leading response to, or part of a feedback cycle involved in major climatic shifts (Fletcher et al., 2013; Morley, Rosenthal, & Demenocal, 2014; Muschitiello, Zhang, Sundqvist, Davies, & Renssen, 2015; Schulz & Paul, 2002).

NAO Indices and Observed Changes

Although the station-based NAO indices have been very widely applied, their limited two-station nature implies that they do not completely capture shifting seasonal, interannual, and multidecadal spatial variability of the North Atlantic pressure patterns. A second issue is that they could suffer from data homogeneity issues of the climate stations, especially the southern/Iberian node (Jones, Osborn, & Briffa, 2003; Jónsson & Hanna, 2007). One of the most widely used NAO datasets is the PC-based NAO index of Hurrell, which is the monthly time series of the leading EOF of SLP anomalies over the Atlantic sector, 20–80°N, 40°E–90°W, which we hereafter abbreviate to Hurrell PC NAO. This does not extend as far back as the station-based NAO indices but still covers the period since 1899 and, importantly, should avoid the above problems. PC-based indices depend on the reliability of the gridded data source; however, for the NAO this is considered to be relatively high for the northern North Atlantic with its comprehensive ship-based observational coverage since well back into the 19th century.

In order to compare NAO variations from climate models with observed NAO, with a view toward future changes (until 2100), Hanna et al. (2015) compiled an NAO index for each of the 77 climate-model runs that were based on Representative Concentration Pathway (RCP) scenario 8.5 in the Coupled Model Intercomparison Project Phase 5 (CMIP5) (Taylor, Stouffer, & Meehl, 2012). CMIP5 is a standard suite of climate model experiments carried out in preparation for the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Based on the CMIP5/RCP8.5 monthly model output, downloaded from the Climate Explorer website, Hanna et al. (2015) constructed a CMIP5 “station NAO” index as the normalized difference (using 1865–2010 as the baseline period) between the Azores and Iceland model grid cells.

Hanna et al. (2015) used simple descriptive statistics, including means and standard deviations and linear least-squares regression trend-line analysis, to define changing characteristics of the NAO (and AO and Greenland Blocking Index—see below) datasets over time. They also used 5-year running means and 5- and 11-year running standard deviations to analyze changes in variability over time in these datasets. Hanna et al. (2015) analyzed the Hurrell PC NAO as their primary NAO dataset up to February 2014. Here we update this analysis based on the latest available Hurrell PC NAO data, which extended through February 2016 at the time of writing.

Hanna et al. (2015) identified that 5 of the 10 most extreme (i.e., the five most positive and five most negative values) values in the Hurrell PC December NAO index occur in the latest 10-year period available at the time of publication (2004–2013). To determine the statistical likelihood of observing 5 out of 10 extreme NAO values in any 10-year consecutive period (window) during 1899–2013, Hanna et al. (2015) ran several tests, including two statistical probability tests and an analysis of the pseudo-NAO indices from the CMIP5 models, the results of which are discussed later in this section.

Figure 3 shows the seasonal Hurrell PC NAO index plotted yearly from 1899 to summer 2015/2016, updated from Hanna et al. (2015). Short-term/annual-to-decadal variability in the NAO is evident, but as these variations largely cancel out, there are no significant seasonal trends for the overall record. However, visual inspection shows strong multidecadal variations: for example, decreases in winter and especially summer NAO since about 1990. These changes are confirmed by the trend analysis in Table 1, where statistically significant trends are highlighted.

Table 1. Seasonal NAO variability (expressed through sigma, = standard deviation expressed through standard deviation of the yearly values for the particular season) and trends for standard climatological normal periods, based on the Hurrell PC NAO index and the January 1899–February 2016 period, updated from Hanna et al. (2015). Trends that are statistically significant at the 95% confidence level are highlighted in bold.

Period

Parameter

Winter (DJF)

Extended winter (DJFM)

Spring (MAM)

Summer (JJA)

Autumn (SON)

Annual

1900–2015/16

sigma

1.00

1.00

1.00

1.00

1.00

1.00

1900–2015/16

trend

0.23

0.24

0.18

0.54

0.36

-0.06

1991–2015/16

sigma

1.15

1.17

1.09

1.17

0.89

1.09

1991–2015/16

trend

-0.71

-0.87

0.21

-1.86

0.64

-0.63

1981–2010

sigma

1.16

1.09

1.04

1.01

0.92

1.20

1981–2010

trend

-0.72

-0.74

-0.48

-1.40

-0.75

-1.44

1971–2000

sigma

1.15

1.03

0.89

0.89

0.92

1.00

1971–2000

trend

1.01

1.02

0.30

-0.18

-0.52

0.36

1961–1990

sigma

1.17

1.10

0.91

0.86

1.06

1.15

1961–1990

trend

1.29

1.49

1.05

0.18

0.37

1.61

North Atlantic OscillationClick to view larger

Figure 3. Seasonal NAO values from 1899 to winter 2016, plotted both annually (faint red lines) and with a 5-year running mean applied (bolder blue lines). NAO index data are from Hurrell (2015).

North Atlantic OscillationClick to view larger

Figure 4. Running standard deviation (sigma) of seasonal NAO values from 1899 to winter 2016. Faint red lines and bold blue lines show the data plotted using a 5-year and 11-year running sigma, respectively. NAO data are from Hurrell (2015).

There are no obviously preferred timescales of variability of the NAO, although Hertig, Beck, Wanner, and Jacobeit (2015) and Menary, Hodson, Robson, Sutton, and Wood (2015) note weak spectral (cyclic) peaks at around 6–10 and 17 years: the latter may be linked to a similar-timescale variation in advection of heat anomalies in the North Atlantic ocean subpolar gyre (Menary et al., 2015). Woollings et al. (2015 report that short interannual timescale (less than 30 years) variations in the NAO are mainly related to changes in atmospheric jet-stream latitude, while longer-term multidecadal changes are predominantly linked to changes in the strength of the jet. They argue that this is due to the different dynamics arising from ocean circulation changes acting on the longer timescale, and therefore this has implications for predictability of seasonal and decadal NAO changes.

Tables 1 and 2 summarize changes in the mean and variance of the NAO for the seasons and calendar months, respectively, for several standard normal climatological time periods and the whole Hurrell PC NAO index period. They confirm large decreases in NAO in summer and annually for the latest (since 1981) time periods. This summer decline is sustained (significant for both the 1981–2010 and 1991–2015 periods), which contributes to the annual NAO trends being significantly negative for 1981–2010 (Table 1). Significant positive trends are seen in the extended winter (DJFM) and spring seasons/months for sub-periods between 1951 and 1990 (Tables 1 and 2). There was a peak in storminess in northwest Europe around 1990 that is related to the winter NAO positive peak (Hanna et al., 2008). The last few decades of significant summer decline in the NAO is not seen for the AO (Hanna et al., 2015).

We also note a remarkable increase in variability in winter NAO in the latter half of the record, shown in the running standard deviation plots in Figure 4. The figure confirms that the enhanced variability is restricted to winter, when NAO winter (DJF) variability has almost doubled over the whole record. The standard deviation of winter (DJF) NAO for successive and overlapping 30-year climatological normal periods has increased steadily and significantly from 0.67 in 1901–1930 to 1.16 in 1981–2010, while other seasons’ trends are less dramatic (Table 1). In the monthly NAO index time series, this enhanced variability is strongest and most significant (at the 99% confidence level) in December (Fig. 5). The standard deviation of December NAO was 1.27 for the overall 1900–2014 period but increased systematically from 1.06 in 1941–1970 to 1.17 in 1951–1980 to 1.28 in 1961–1990, 1.39 in 1971–2000, 1.65 in 1981–2010, 1.79 in 1991–2014, and 1.99 in 2001–2014 (Table 2): this is therefore a sustained increase, not just a recent phenomenon. Enhanced winter variability is also seen for the AO; although it is not quite as marked as for the NAO, it is still very noticeable, and has recently been highlighted by Overland and Wang (2015). Table 3 shows the five lowest and five highest NAO values for each season and calendar month for 1899–2015/16. The years 2009 and 2011 are prominent in the five lowest monthly and seasonal summer values, although 2009 also has the third highest September NAO value.

Table 2. Monthly NAO variability (expressed through sigma, = standard deviation of the yearly values for the particular month) and trends for standard climatological normal periods, based on the Hurrell PC NAO index and the January 1899–February 2016 period, updated from Hanna et al. (2015). Trends that are statistically significant at the 95% confidence level are highlighted in bold.

Period

Parameter

Jan.

Feb.

Mar.

Apr.

May

June

July

Aug.

Sept.

Oct.

Nov.

Dec.

1900–2015/16

sigma

1.44

1.59

1.50

0.81

0.67

0.61

0.47

0.48

0.55

0.79

0.90

1.28

1900–2015/16

trend

-0.07

0.30

0.37

-0.07

0.14

-0.05

0.04

0.08

-0.03

-0.38

0.33

-0.13

2001–2015/16

sigma

1.13

1.49

1.81

0.83

0.75

0.67

0.48

0.48

0.60

0.91

0.81

1.99

2001–2015/16

trend

-0.11

0.72

0.95

0.30

0.11

-0.49

-0.75

-0.39

0.35

0.15

0.33

1.00

1991–2015/16

sigma

1.18

1.50

1.61

0.76

0.75

0.74

0.45

0.51

0.64

0.82

0.91

1.78

1991–2015/16

trend

-0.76

-1.23

-0.60

0.57

0.42

-0.62

-0.52

-0.90

0.61

0.08

-0.02

0.30

1981–2010

sigma

1.43

1.72

1.52

0.75

0.68

0.68

0.39

0.50

0.61

0.94

0.95

1.65

1981–2010

trend

-0.51

-0.96

-0.79

-0.15

0.01

-0.27

-0.51

-0.67

0.10

-1.16

-0.28

-1.22

North Atlantic OscillationClick to view larger

Figure 5. Eleven-year running standard deviation (sigma) of monthly (January, February, March, and December) NAO values from 1899 to winter 2016. NAO data are from Hurrell (2015).

Interestingly, 5 out of the 10 most extreme December values are in the last 12 years from 2004 to 2015: 2010 and 2009 have the two lowest December NAO values, whereas 2011 and 2006 have the two highest December NAO values and 2004 has the fourth highest December NAO value (2013 was the sixth highest December NAO value in the series).

Table 3. Five lowest and five highest NAO years for each calendar month and season, based on the Hurrell PC NAO index and the January 1899–February 2016 period, updated from Hanna et al. (2015). Years since 2000 are highlighted in bold.

Month

5 lowest

5 highest

Jan

1966, 1969, 1940, 1963, 1945

1993, 1989, 1983, 1928, 1990

Feb

1947, 2010, 1978, 1942, 1960

1990, 1989, 1997, 2000, 1959

Mar

2013, 1962, 1958, 1931, 1952

1986, 1990, 1913, 1920, 1994

Apr

1966, 1978, 1988, 1979, 2008

1947, 2011, 1943, 1990, 1904

May

1993, 2008, 1954, 1952, 1909

1956, 1963, 2009, 1914, 2015

Jun

1902, 1903, 2009, 1982, 2011/2012

1994, 1961, 1967, 1922, 1919

Jul

2015, 1907, 1962, 2009, 1918,

1964, 1920, 1946, 1935, 1975

Aug

1943, 1964, 1958, 2011, 1966

1991, 1971, 1983, 1961, 2013

Sep

1998, 1930, 1968, 1939, 1915

1975, 1947, 2009, 1917, 1950

Oct

2006, 1960, 2012, 1966, 1968

1986, 1957, 1983, 1938, 1935

Nov

1910, 1947, 1955, 1915, 1965

1978, 1982, 1992, 1953, 1913

Dec

2010, 2009, 1961, 1995, 1978

2011, 2006, 1951/1982, 2004

DJF

2010, 1969, 1979, 1936, 1966

1989, 1993, 2000, 2015, 1992

MAM

1924, 1951, 1909, 1962, 2013

1990, 2015, 1913, 1943, 1921

JJA

1958, 1902, 1903, 2011, 2009

1983, 1994, 1955, 2013, 1967

SON

1915, 1968, 1910, 2002, 1941

1986, 1938, 1978, 1954, 1982

Two separate probability analyses conducted in Hanna et al. (2015) determined the likelihood of obtaining 5 of the top 10 extreme Hurrell PC NAO index December values in a 10-year period to be between 0.033 and 0.037% (Hanna et al., 2015; at the time of analysis the authors were analyzing the last 10 Decembers rather than the last 12, but this makes little difference to the results). From this we might deduce that the observed variability in the December NAO is statistically highly significant; however, one model (out of 77) in the CMIP5 archive-based analysis was also found to have 5 of its top 10 December NAO values occurring in a 10-year period. Indeed, two other models displayed the same variability over 11- and 14-year windows, with 5 models having 5 out of the most 10 extreme values in a 15-year window (Hanna et al., 2015). Although these variations in the model world are extreme given the CMIP5 model range, the clustering of extreme NAO values in short time periods in these models—arguably a better sample of the December NAO than randomly generated series—indicates that the observed clustering of extreme values in the December NAO since 2004 is perhaps more likely than standard probability tests would suggest (Hanna et al., 2015).

Hanna et al. (2015) found similar results when analyzing several different station-based NAO time series in addition to the primary Hurrell PC-based NAO dataset. These similar results confirmed that changes in weather-station coverage, which might affect spatial variance patterns used in calculating PCs, cannot explain the more recently enhanced winter NAO variability (Hanna et al., 2015). Analysis of the constituent southwest Iceland and Iberian December pressure series for 1899–2014 (not shown here) shows no significant trend for southwest Iceland but a systematic significant increase in interannual pressure variability at the southern NAO node: this increase is greatest for Gibraltar and Lisbon but still significant for the Azores. This increased variability was supported by gridded maps of differences in standard deviation of December mean sea-level pressure (MSLP) for climatological normal and more recent periods since 1948 relative to the whole 1948–2014 period. Hanna et al. (2015) showed reduced (enhanced) variability around the locations of the NAO nodes for earlier (later) periods, based on both NCEP/NCAR and 20CR meteorological reanalyses. They therefore concluded that the increased variability in December NAO is linked with more variable year-to-year pressure around the NAO nodes.

Hanna et al. also analyzed changes in the CMIP5 model mean winter (DJF) NAO index for 1862–2100, and found no model evidence of the increased variance seen in the observed winter NAO over the past century (2015, Fig. 10a). Hanna et al. (2015) used the most extreme RCP8.5 scenario to best highlight any changes. It should be emphasized, however, that the fact that CMIP5 model mean output does not reproduce observed NAO variations over the last century (Hanna et al., 2015, Fig. 10b) may simply mean that these changes are due to internal atmospheric variability. Also, while the ensemble mean of the CMIP5 runs reflects “forced” variability, it is possible that any single model member may have similar if not greater variance/values in the NAO. CMIP5 simulations overall indicate a weakly increasing mean winter NAO index since 1862 and projected forward to 2100 (Gillett & Fyfe, 2013; Hanna et al., 2015).

Relationship between NAO and Greenland Blocking

The NAO is strongly related to the occurrence and strength of blocking high pressure over Greenland, using a measure called the Greenland Blocking Index (GBI; Fang, 2004; Hanna et al., 2013; Hanna, Cropper, Hall, & Cappelen, 2016). GBI is defined as the 500 hPa geopotential height monthly anomaly over the 60–80°N, 20–80°W region, and is based on NCEP/NCAR Reanalysis 500 hPa geopotential height data (Fang, 2004; Hanna et al., 2013). An extended, fully homogenized GBI index utilising 20CR data (Compo et al. 2011) and spanning 1851–2015 has recently been constructed (Hanna et al., 2016).

Correlation coefficients between the Hurrell PC NAO index and GBI are strongly negative, ranging from -0.75 in July to -0.94 in December/February, for the overall 1851–2015 period of common data, with the correlations being relatively stable through time (Hanna et al., 2016). These highly negative correlations depict the close correspondence between Greenland blocking and the wider North Atlantic atmospheric circulation changes, mainly in winter that is defined by the NAO (Woollings, Hannachi, Hoskins, & Turner, 2010).

Variations and trends in GBI generally show opposite trends to NAO, for example, a recent increase in GBI since 1981 (Hanna et al., 2015). However, for the 1981–2010 and 1991–2013 time periods, this rising GBI trend has been found to be significant across more seasons and months than the contemporaneous NAO decrease (Hanna et al., 2015). Specifically, GBI shows a significantly increasing trend for most seasons, whereas the significant NAO decrease is restricted to summer. Based on their initial GBI series spanning 1948–2013/2014, Hanna et al. (2015) found that four of the five highest June GBI values were in years since 2007 inclusive, while three of the five highest December GBI values were in the 2000s. The most prominent high GBI year is 2010, which featured in the five highest GBI year list (1948–2013) in six calendar months (Hanna et al., 2015). Hanna et al. (2016) compiled and analyzed an extended 1851–2015 GBI monthly and seasonal time series, corroborating and extending the above results. They found a clustering of high GBI values in summer, when 7 of the top 11 values in the 165-year record—including the 2 last years 2014 and 2015—occurred since 2007. There are clear links between some recent extreme high GBI and low NAO monthly/seasonal values (Hanna et al., 2015, 2016), but based on regression analysis, Hanna et al. (2015) concluded that the recent GBI and NAO summer trends are mostly statistically distinct and cannot be exclusively related.

However, GBI has also become significantly more variable in the last few decades in December, in line with increased NAO and AO variability in that month (Hanna et al., 2015, 2016; Overland & Wang, 2015).

Interpretation of Recent (2000–2015) NAO Changes

We attempt to interpret the more negative summer NAO and more volatile winter NAO in the context of recent and ongoing climate change. Until fairly recently the main argument, following the strong positive NAO trend observed between the 1960s and 1990s, was that we could reasonably expect a slightly more positive NAO due to human-induced greenhouse gas forcing (e.g., Gillett et al., 2003); however, it has also been realized that there is high internal variability in the NAO and related parts of the climate system (e.g., Osborn, 2004; Stephenson, Pavan, & Bojariu, 2000). Although projected trends are small when compared with the internal variability (Stephenson et al., 2000), the latest CMIP5 climate-model projections do show an overall slight NAO increase in all seasons for the remainder of this century (Gillett & Fyfe, 2013).

The significant recent decline in summer NAO is not repeated in the AO index (Hanna et al., 2015), even though—as we have seen—the two indices are strongly related. This suggests that it may be caused by a regional mechanism in the North Atlantic and northeastern North America sector. This appears to be related to an increase in anticyclonic blocking over Greenland (Hanna et al., 2013, 2016), with effects on the North Atlantic polar jet stream circulation “downstream.” Overland, Francis, Hanna, and Wang (2012) observed that enhanced early summer high-pressure blocking over Greenland and North America between 2007 and 2012 was linked with a more north–south waving North Atlantic jet stream flow, and a more negative NAO tends to accompany this weaker summer jet-stream configuration. Moreover, according to Woollings et al. (2010), more negative winter NAO events represent a higher frequency of Greenland blocking episodes, and if this hypothesis is correct, the same may well be true in summer. The ultimate cause of the increased Greenland blocking still has to be determined but may be related to earlier seasonal snowmelt and loss of Arctic sea-ice cover (Overland et al., 2012); however, once it occurs it is likely to be enhanced by warmer Greenland summers raising geopotential heights (Fettweis et al., 2013; Hanna et al., 2013, 2014). Enhanced north–south airflow in summer over eastern North America may also be related to Greenland blocking (Francis & Vavrus, 2012; Hanna et al., 2014; Neff, Compo, Ralph, & Shupe, 2014; Overland et al., 2012, 2015). It is therefore possible that anthropogenic greenhouse warming and resulting cryospheric changes may lead to negative summer seasonal trends in the NAO, which is opposite to previous conclusions and climate-model simulations of long-term trends (Bladé, Liebmann, Fortuny, & Oldenborgh, 2011; Folland et al., 2009; Gillett & Fyfe, 2013). Belleflamme, Fettweis, Lang, and Erpicum (2013) showed that the negative excursion of the summer NAO from about 2000, based on four different indices, is comfortably outside the ±1 standard deviation spread of historical and projected CMIP5 variability. Again, these authors suggest that increased anticyclonic conditions over Greenland are a primary cause of the observed negative summer NAO trend.

Other regional climatic forcing factors such as unusually high North Atlantic sea-surface temperatures may have been active in destabilizing atmospheric circulation over the mid-latitude North Atlantic and causing a more negative summer NAO (Sutton & Dong, 2012). Similarly, recent extreme negative values of the winter NAO have been partly linked to the unusually deep solar minimum observed in 2009 (Ineson et al., 2011) and to other coincident but transient effects due to the El Niño Southern Oscillation (forcing from the tropical Pacific) and the Quasi Biennial Oscillation (forcing from the stratosphere) (Fereday, Maidens, Arribas, Scaife, & Knight, 2012). Nakamura et al. (2015) present observational and model evidence of the relation between the recent negative shift in the winter NAO/AO and reduced sea-ice in early winter in the Barents Sea, which they argue is dynamically caused by changes in the polar vortex, stratosphere–troposphere coupling, and atmospheric planetary waves propagation affecting the near-surface atmospheric circulation.

Predicting the Seasonal NAO

Scaife et al. (2014) report results using a new Met Office Hadley Centre forecast system called GloSea5 to assess the seasonal predictability of the winter NAO. They did this based on running the model and producing forecasts for a previous 20-year time period (1993–2012) and comparing the results with the observed NAO values. They found a correlation coefficient of 0.62 between the average of their range (ensemble) of 24 model simulations and actual winter NAO, a result that is highly statistically significant. Independent studies (Cohen & Jones, 2011; Folland, Scaife, Lindesay, & Stephenson, 2012; Hall, 2016) support the principle that the NAO is predictable between one to four months ahead, rather than just being random noise. Scaife et al. (2014) reported four sources of predictability for the winter NAO:

  1. 1. Tropical Pacific sea temperatures through El Niño (La Niña) ocean warming (cooling) events and a teleconnection (i.e., remote climate forcing) via changes in the stratosphere—El Niño (La Niña) conditions in November tend to promote a negative (positive) winter NAO.

  2. 2. North Atlantic sea-surface temperatures being colder (warmer) than normal in the subpolar gyre in November were linked to more positive (negative) NAO conditions.

  3. 3. Low (high) Kara Sea sea-ice cover anomalies in November were linked to a more negative (positive) NAO.

  4. 4. The direction of motion of the quasi-biennial oscillation (QBO)—a jet stream in the equatorial stratosphere. If the QBO is in its westerly phase (winds blowing from west to east), this is linked to a stronger extratropical jet stream and more positive NAO.

There was accompanying skillful prediction of Northern Hemisphere wintertime blocking frequency, which was supported by an independent seasonal prediction system (Athanasiadis et al., 2014). Scaife et al. (2014) also noted that it should be possible to improve the GloSea5 model predictions through use of a greater number of model runs (a larger ensemble) to achieve an improved correlation coefficient (model skill) of around 0.8, and this kind of predictability may also be extended to include other seasons. Although the early results reported here are certainly promising, achieving more skillful seasonal NAO forecasts will no doubt be highly beneficial from social, economic, and environmental perspectives across Europe. In analogous work, Riddle, Butler, Furtado, Cohen, and Kumar (2013) found modest but significant skill in using the operational Climate Forecast System (CFSv2) seasonal hindcast dataset to forecast the interannual variability of the winter AO two months ahead and noted that more accurately representing Eurasian snow cover extent in October resulted in a slightly enhanced predictive skill. They suggested that this was enabled through skillful prediction of sea-surface-temperature anomalies, since realistic stratosphere–troposphere coupling in the model—a widespread challenge—was somewhat lacking. Skillful boreal winter AO forecasts were achieved for a wider range of dynamical ensemble prediction systems by Kang et al. (2014).

NAO and Future Climate Change

It is still unclear what will happen to the NAO with ongoing anthropogenic climate change, even discounting other external forcing factors and natural (internal) variability. The tropopause is highest at the equator (about 15 km above the surface) and slopes down toward the poles (about 10 km altitude). With increasing greenhouse gas levels, temperatures warm at the surface and in the lower troposphere while the stratosphere cools. This effect has been well observed in recent decades and is due to a denser blanket of greenhouse gases trapping infrared radiation in the lower atmosphere. At the same time, the surface has been warming most rapidly at high latitudes: called polar (or here Arctic) amplification of global warming (Overland et al., 2014). This latter change has the effect of reducing the meridional (north–south) temperature gradient, which might be expected to reduce the amount of energy available for driving the polar jet stream, all other factors being equal (Francis & Vavrus 2015; Overland et al., 2015). But this is just a (near-)surface expression of global warming. Meanwhile, in the upper troposphere at low latitudes, there is a higher specific humidity under global warming, and this raises the tropopause and increases upper troposphere temperatures near the equator (about 15 km up) while the same altitude near the poles (i.e., well within the stratosphere at these high latitudes) significantly cools with global warming. Therefore there is a significantly enhanced meridional temperature gradient at this higher altitude just at the same time that the north–south temperature gradient reduces near the surface (Harvey, Shaffrey, & Woollings, 2015). Thus there are two competing influences that can result in changes in mid-latitude (i.e., polar) jet-stream dynamics under conditions of global warming. One recent model-based study (Harvey et al., 2015) suggests that the near-surface meridional temperature gradient change is most important for determining changes in the wintertime North Atlantic storm track, but overall this is far from certain and more work is clearly needed (Barnes & Screen, 2015).

Gillett and Fyfe (2013) used the CMIP5 suite of climate-model simulations—based on 37 different models using the RCP4.5 emissions scenario—to show weak trends toward a more positive NAM and NAO in all seasons by 2100. No CMIP5 model showed a significant decrease in any season. Gillett and Fyfe (2013) also found no significant response of the modeled NAM and NAO to changes in solar and volcanic climate forcings, although this may simply reflect limitations with the CMIP5 models (e.g., Hall et al., 2015) rather than proving that such effects do not exist. Lastly, Gillett and Fyfe (2013) also noted that the more negative NAM since 2000 is not shown in the models (when run retrospectively)—this difference reflects either model deficiency or internal variability in the NAM/NAO climate system. However, Hall et al. (2015) also point out that other studies of model projections do show, or don’t discount, a negative NAO response to anthropogenic global warming. One such study was made by Sun, Deser, and Tomas (2015b) who modeled the atmospheric response of 21st-century Arctic sea-ice decline, finding a negative NAM/NAO effect in winter.

Hall et al. (2015) discuss a range of external forcing mechanisms—including Arctic sea-ice retreat, Northern Hemisphere snow-cover loss, Atlantic sea temperatures, El Niño Southern Oscillation, stratospheric circulation changes, solar variability, and volcanic eruptions—that may cause the North Atlantic polar jet stream and NAO to vary during the next few decades. Many of these changes seem to predispose the North Atlantic atmosphere system toward a more negative NAO (e.g., snow and ice losses, warmer Atlantic sea waters, El Niño episodes, reduced solar activity and related changes in stratospheric winds and heating), and Hall et al. (2015) also discuss limitations with the current generation of climate models in replicating these forcings. Overall, our level of understanding of the physical mechanisms remains insufficient to be able to make a reliable NAO trend projection to 2100, especially as we are unable to reliably predict future changes in several of the key potential driving factors listed above. Moreover, there are major problems with many climate models in adequately capturing coupling between the troposphere and stratosphere, or with the surface and troposphere, which are crucial for driving and understanding NAO changes (Hall et al., 2015).

Conclusions

The NAO is the primary variation in barometric pressure variation over the North Atlantic that affects the weather and climate of much of Europe. It is subject to internal variability or chaos in the climate system but is also influenced by slowly varying climatic forcing factors including anthropogenic greenhouse warming and solar and volcanic variability, which makes the NAO inherently predictable—at least in part—on a timescale of up to at least several months. Between the 1960s and 1990s the NAO was becoming more positive, but since then this trend has tended to reverse. Recently updated observational records and reanalyses show increasing variability of winter NAO and AO, which is a feature not just of the 2000s and early 2010s but has been ongoing during the 20th century. We have also noted during the last 20–30 years a statistically significant decline in the NAO in summer—and to a lesser extent winter (with a recent record negative December value, although the winter negative trend appears strongest during the period 1989–2011 but has returned to more positive/neutral values in the last 5–6 years)—whereas no significant change has occurred in spring and autumn. This asymmetric seasonal response of the NAO, and its increased winter variability, was not foreseen in previous general circulation climate model predictions analyzed here but may have resulted from several climatic forcing factors and feedbacks conspiring together: these include enhanced blocking arising from cryosphere–atmosphere couplings, solar variability, and/or changes in North Atlantic sea temperatures. The increasingly more variable winter NAO that we have detected based on the last century or so of observations appears to be a seasonally uneven change and does not show up as a forced response—i.e., responding to an external climatic driving factor—in state-of-the-art (CMIP5) climate models. Although the winter increase in NAO variance was sustained over the 20th century and there may be errors in climate models, we still cannot discount the possibility that this feature is due to internal variability (random noise and chaos generation in the climate system), particularly as there was no prior reason to expect this change to occur in winter as opposed to some other season. It is currently uncertain how the NAO will change during the rest of the current century, as both the climate models and our understanding of the physical processes causing NAO variations need to be improved. Whatever the reason(s) behind the observed seasonal NAO changes, there are also clearly important consequences for the heavily populated circum–North Atlantic land masses if these changes continue.

Acknowledgments

NAO data were provided by the Climate Analysis Section, NCAR, Boulder, Colorado (Hurrell, 2003), updated regularly; we thank Adam Phillips (UCAR) for providing data updates. We acknowledge the meteorological reanalysis data providers at the European Centre for Medium-Range Weather Forecasts and the U.S. National Center for Environmental Prediction/National Center for Atmospheric Research. We thank David McCutcheon for help with drawing figures. EH appreciates useful discussions with Richard Hall and James E. Overland among other colleagues and acknowledges Avalon Hanna for support and encouragement. NCEP Reanalysis data provided by the NOAA/OAR/ESRL PSD, Boulder, Colorado, from their website.

References

Abram, N. J., Mulvaney, R., Vimeux, F., Phipps, S. J., Turner, J., & England, M. H. (2014). Evolution of the Southern Annular Mode during the past millennium. Nature Climate Change, 4, 564–569.Find this resource:

Athanasiadis, P. J., Bellucci, A., Hermanson, L., Scaife, A. A., MacLachlan, C., Arribas, A., et al. (2014). The representation of atmospheric blocking and the associated low-frequency variability in two seasonal prediction systems. Journal of Climate, 27, 9082–9100.Find this resource:

Baldwin, M. P., & Dunkerton, T. J. (2001). Stratospheric harbingers of anomalous weather regimes. Science, 294, 581–584.Find this resource:

Barnes, E. A., & Screen, J. A. (2015). The impact of Arctic warming on the midlatitude jet-stream: Can it? Has it? Will it? WIREs Climate Change, 6, 277–2286.Find this resource:

Belleflamme, A., Fettweis, X., Lang, C., & Erpicum, M. (2013). Current and future atmospheric circulation at 500 hPa over Greenland simulated by the CMIP3 and CMIP5 global models. Climate Dynamics, 41, 2061–2080.Find this resource:

Beniston, M. (2012). Is snow in the Alps receding or disappearing?. WIRES Climate Change, 3, 349–358.Find this resource:

Bladé, I., Liebmann, B., Fortuny, D., & Oldenborgh, G. (2011). Observed and simulated impacts of the summer NAO in Europe: Implications for projected drying in the Mediterranean region. Climate Dynamics, 39, 709–727.Find this resource:

Blessing, S., Fraedrich, K., Junge, M., Kunz, T., & Lunkeit, F. (2005). Daily North-Atlantic Oscillation (NAO) index: Statistics and its stratospheric polar vortex dependence. Meteorologische Zeitschrift, 14, 763–769.Find this resource:

Burić, D., Lukovic, J., Ducic, V., Dragojlovic, J., & Doderovic, M. (2014). Recent trends in daily temperature extremes over southern Montenegro (1951–2010). Natural Hazards and Earth System Science, 14, 67–72.Find this resource:

Casanueva, A., Rodríguez-Puebla, C., Frías, M. D., & González Reviriego, N. (2014). Variability of extreme precipitation over Europe and its relationships with teleconnection patterns. Hydrology and Earth System Sciences, 18, 709–725.Find this resource:

Cherry, J., Cullen, H., Visbeck, M., Small, A., & Uvo, C. (2005). Impacts of the North Atlantic Oscillation on Scandinavian hydropower production and energy markets. Water Resources Management, 19, 673–691.Find this resource:

Clement, A., Bellomo, K., Murphy, L. N., Cane, M. A., Mauritsen, T., Radel, G., et al. (2015). The Atlantic multidecadal oscillation without a role for ocean circulation. Science, 350, 320–324.Find this resource:

Cohen, J., & Jones, J. (2011). A new index for more accurate winter predictions. Geophysical Research Letters, 38, L21701.Find this resource:

Cohen, J., Screen, J. A., Furtado, J. C., Barlow, M., Whittleston, D., Coumou, D., et al. (2014). Recent Arctic amplification and extreme mid-latitude weather. Nature Geoscience, 7, 627–637.Find this resource:

Colantuono, G., Wang, Y., Hanna, E., & Erdélyi, R. (2014). Signature of the North Atlantic Oscillation on British solar radiation availability and PV potential: The winter zonal seesaw. Solar Energy, 107, 210–219.Find this resource:

Compo, G. P., Whitaker, J. S., Sardeshmukh, P. D., Matsui, N., Allan, R. J., Yin, X., et al. (2011). The Twentieth Century Reanalysis Project. Quarterly Journal of the Royal Meteorological Society, 137, 1–28.Find this resource:

Cropper, T. E., Hanna, E., & Bigg, G. R. (2014). Spatial and temporal seasonal trends in coastal upwelling off northwest Africa, 1981–2012. Deep Sea Research I, 86, 94–111.Find this resource:

Cropper, T., Hanna, E., Valente, M. A., & Jónsson, T. (2015). A daily Azores–Iceland North Atlantic Oscillation index back to 1850. Geoscience Data Journal, 2, 12–24.Find this resource:

Curtis, J., Lynch, M. Á., & Zubiate, L. (2016). Carbon dioxide (CO2) emissions from electricity: The influence of the North Atlantic Oscillation. Applied Energy, 161, 487–496.Find this resource:

Dee, D. P., Uppala, S. M., Simmons, A. J., Berrisford, P., Poli, P., Kobayashi, S., et al. (2011). The ERA-interim reanalysis: Configuration and performance of the data assimilation system. Quarterly Journal of the Royal Meteorological Society, 137, 553–597.Find this resource:

Diao, Y., Xie, S.-P., & Luo, D. (2015). Asymmetry of winter European surface air temperature extremes and the North Atlantic Oscillation. Journal of Climate, 28, 517–530.Find this resource:

Donat, M. G., Peterson, T. C., Brunet, M., King, A. D., Almazroui, M., Kolli, R. K., et al. (2014). Changes in extreme temperature and precipitation in the Arab region: long-term trends and variability related to ENSO and NAO. International Journal of Climatology, 34, 581–592.Find this resource:

Ely, C. R., Brayshaw, D. J., Methven, J., Cox, J., & Pearce, P. (2013). Implications of the North Atlantic Oscillation for a UK–Norway renewable power system. Energy Policy, 62, 1420–1427.Find this resource:

Fang, Z.-F. (2004). Statistical relationship between the Northern Hemisphere sea ice and atmospheric circulation during wintertime. In X. Zhu (Ed.), Observation, theory and modeling of atmospheric variability (pp. 131–1441). World Scientific Publishing Company, Singapore.Find this resource:

Fereday, D. R., Maidens, A., Arribas, A., Scaife, A. A., & Knight, J. R. (2012). Seasonal forecasts of northern hemisphere winter 2009/10. Environmental Research Letters, 7, 034031.Find this resource:

Fettweis, X., Hanna, E., Lang, C., Belleflamme, A., Erpicum, M., & Gallée, H. (2013). Brief communication: Important role of the mid-tropospheric atmospheric circulation in the recent surface melt increase over the Greenland ice sheet. The Cryosphere, 7, 241–248.Find this resource:

Fletcher, W. J., Debret, M., & Sanchez Goñi, M. F. (2013). Mid-Holocene emergence of a low-frequency millennial oscillation in western Mediterranean climate: Implications for past dynamics of the North Atlantic atmospheric westerlies. The Holocene, 23, 151–164.Find this resource:

Folland, C. K., Knight, J., Linderholm H. W., Fereday, D., Ineson, S., & Hurrell, J. W. (2009). The summer North Atlantic Oscillation: Past, present, and future. Journal of Climate, 22, 1082–1103.Find this resource:

Folland, C. K., Scaife, A. A., Lindesay, J., & Stephenson, D. (2012). How potentially predictable is northern European winter climate a season ahead?. International Journal of Climatology, 32, 801–818.Find this resource:

Francis, J. A., & Vavrus, S. J. (2012). Evidence linking Arctic amplification to extreme weather in mid-latitudes. Geophysical Research Letters, 39, L06801.Find this resource:

Francis, J. A., & Vavrus, S. A. (2015). Evidence for a wavier jet stream in response to rapid Arctic warming. Environmental Research Letters, 10.Find this resource:

Fu, C., & Yao, H. (2015). Trends of ice breakup date in south-central Ontario. Journal of Geophysical Research Atmospheres, 120, 9220–9236.Find this resource:

Gastineau, G., & Frankignoul, C. (2015). Influence of the North Atlantic SST variability on the atmospheric circulation during the twentieth century. Journal of Climate, 28, 1396–1416.Find this resource:

Gillett, N. P., & Fyfe, J. C. (2013). Annular mode changes in the CMIP5 simulations. Geophysical Research Letters, 40, 1189–1193.Find this resource:

Gillett, N. P., Graf, H. F., & Osborn, T. J. (2003). Climate change and the North Atlantic Oscillation. In J. W. Hurrell, Y. Kushnir, G. Ottersen, & M. Visbeck (Eds.), The North Atlantic Oscillation: Climatic significance and environmental impact (pp. 193–209). Washington, DC: American Geophysical Union.Find this resource:

Hall, R. J. (2016). The North Atlantic Polar Front Jet Stream: Variability and Predictability, 1871–2014. Chapter 6: Simple Probabilistic Forecasts of the NAO. Unpublished PhD thesis, pp. 244–291.Find this resource:

Hall, R., Erdélyi, R., Hanna, E., Jones, J. M., & Scaife, A. A. (2015). Drivers of North Atlantic polar front jet stream variability. International Journal of Climatology, 35, 1697–1720.Find this resource:

Hanna, E., Cappelen, J., Allan, R., Jónsson, T., le Blancq, F., Lillington, T., et al. (2008). New insights into North European and North Atlantic surface pressure variability, storminess and related climatic change since 1830. Journal of Climate, 21, 6739–6766.Find this resource:

Hanna, E., Cropper, T. E., Hall, R., & Cappelen, J. (2016). Greenland Blocking Index 1851–2015: A regional climate change signal. International Journal of Climatology.Find this resource:

Hanna, E., Cropper, T. E., Jones, P. D., Scaife, A. A., & Allan, R. (2015). Recent seasonal asymmetric changes in the NAO (a marked summer decline and increased winter variability) and associated changes in the AO and Greenland Blocking Index. International Journal of Climatology International Journal of Climatology, 35, 2540–2554.Find this resource:

Hanna, E., Fettweis, X., Mernild, S. H., Cappelen, J., Ribergaard, M. H., Shuman, C. A., et al. (2014). Atmospheric and oceanic climate forcing of the exceptional Greenland ice sheet surface melt in summer 2012. International Journal of Climatology, 34, 1022–1037.Find this resource:

Hanna, E., Jones, J. M., Cappelen, J., Mernild, S. H., Wood, L., Steffen, K., et al. (2013). The influence of North Atlantic atmospheric and oceanic forcing effects on 1900–2010 Greenland summer climate and ice melt/runoff. International Journal of Climatology, 33, 862–880.Find this resource:

Harry, C. T., II. (2015). Linkages between New England dolphin stranding frequency and North Atlantic Oscillation variability. Open Access Master’s Thesis, Paper 543. Retrieved from http://digitalcommons.uri.edu/theses/543.Find this resource:

Harvey, B. J., Shaffrey, L. C., & Woollings, T. J. (2015). Deconstructing the climate change response of the Northern Hemisphere wintertime storm tracks. Climate Dynamics, 45, 2847–2860.Find this resource:

Hegerl, G. C., & Kenyon, J. (2008). Influence of modes of climate variability on global temperature extremes. Journal of Climate, 21, 3872–3889.Find this resource:

Hertig, E., Beck, C., Wanner, H., & Jacobeit, J. (2015). A review of non-stationarities in climate variability of the last century with focus on the North Atlantic-European sector. Earth Science Reviews, 147, 1–17.Find this resource:

Hildebrandsson, H. H. (1897). Quelques recherches sur les centres d’action de l’atmosphère. Kongl. Svenska Vetenskaps-okad. Handl., 29(3), 36.Find this resource:

Huntingford, C., Marsh, T., Scaife, A. A., Kendon, E. J., Hannaford, J., Kay, A. L., et al. (2014). Potential influences on the United Kingdom’s floods of winter 2013/14. Nature Climate Change, 4, 769–777.Find this resource:

Hurrell, J. W. (1995). Decadal trends in the North Atlantic oscillation: Regional temperatures and precipitation. Science, 269, 676–679.Find this resource:

Hurrell, J. W. (2015). NAO data. Retrieved from https://climatedataguide.ucar.edu/climate-data/hurrell-north-atlantic-oscillation-nao-index-pc-based

Hurrell, J. W., Kushnir, Y., Ottersen, G., & Visbeck, M. (Eds.). (2003). The North Atlantic Oscillation: Climatic significance and environmental impact. Washington, DC: American Geophysical Union.Find this resource:

Ineson, S., Scaife, A. A., Knight, J. R., Manners, J. C., Dunstone, N. J., Gray, L. J., et al. (2011). Solar forcing of winter climate variability in the Northern Hemisphere. Nature Geoscience, 4, 753–757.Find this resource:

Jerez, S., Trigo, R. M., Vicente-Serrano, S. M., Pozo-Vázquez, D., Lorente-Plazas, R., Lorenzo-Lacruz, J., et al. (2013). The impact of the North Atlantic Oscillation on renewable energy resources in southwestern Europe. Journal of Applied Meteorology and Climatology, 52, 2204–2225.Find this resource:

Jones, P. D., Jónsson, T., & Wheeler, D. (1997). Extension to the North Atlantic Oscillation using early instrumental pressure observations from Gibraltar and south-west Iceland. International Journal of Climatology, 17, 1433–1450.Find this resource:

Jones, P. D., Osborn, T. J., & Briffa, K. E. (2003). Pressure based measures of the North Atlantic Oscillation (NAO): A comparison and an assessment of changes in the strength of the NAO and in its influence on surface climate parameters. In J. W. Hurrell, Y. Kushnir, G. Ottersen, & M. Visbeck (Eds.), The North Atlantic Oscillation: Climatic significance and environmental impact (pp. 51–62). American Geophysical Union Washington, DC.Find this resource:

Jones, P. D., Briffa, K. R., Osborn, T. J., Lough, J. M., van Ommen, T. D., Vinther, B.M., et al. (2009). High-resolution paleoclimatology of the last millennium: A review of current status and future prospects. Holocene, 19, 3–49.Find this resource:

Jónsson, T., & Hanna, E. (2007). A new day-to-day pressure variability index as a proxy of Icelandic storminess and complement to the North Atlantic Oscillation Index 1823–2005. Meteorologische Zeitschrift, 16, 25–36.Find this resource:

Kalnay, E., Kanamitsu, M., Kistler, R., Collins, W., Deaven, D., Gandin, L. et al. (1996). The NCEP/NCAR 40-year reanalysis project. Bulletin of the American Meteorological Society, 77, 437–471.Find this resource:

Kang, D., Lee, M.-I., Im, J., Kim, D., Kim, H.-M., Kang, H.-S., et al. (2014). Prediction of the Arctic Oscillation in boreal winter by dynamical seasonal forecasting systems. Geophysical Research Letters, 41, 3577–3585.Find this resource:

Karpechko, A. Y., Peterson, K. A., Scaife, A. A., Vainio, J., & Gregow, H. (2015). Skilful seasonal predictions of Baltic Sea ice cover. Environmental Research Letters, 10, 044007.Find this resource:

Kaspi, Y., & Scheider, T. (2013). The role of stationary eddies in shaping midlatitude storm tracks. Journal of the Atmospheric Sciences, 70, 2596–2613.Find this resource:

Kendon, M., & McCarthy, M. (2015). The UK’s wet and stormy winter of 2013/2014. Weather, 70, 40–47.Find this resource:

Kenyon, J., & Hegerl, G. C. (2008). Influence of modes of climate variability on global temperature extremes. Journal of Climate, 21, 3872–3889.Find this resource:

Khidher, S. A., & Pilesjo, P. (2015). The effect of the North Atlantic Oscillation on the Iraqi climate 1982–2000. Theoretical and Applied Climatology, 122, 771–782.Find this resource:

Kidston, J., Scaife, A. A., Hardiman, S. C., Mitchell, D. M., Butchart, N., Baldwin, M. P., et al. (2015). Stratospheric influence on tropospheric jet streams, storm tracks and surface weather. Nature Geoscience, 8, 433–440.Find this resource:

Lehner, F., Raible, C. C., & Stocker, T. F. (2012). Testing the robustness of a precipitation proxy-based North Atlantic Oscillation reconstruction. Quaternary Science Reviews, 45, 85–94.Find this resource:

L’Heureux, M., Butler, A., Jha, B., Kumar, A., & Wang, W. (2010). Unusual extremes in the negative phase of the Arctic Oscillation during 2009. Geophysical Research Letters, 37, L10704.Find this resource:

Marshall, A. G., & Scaife, A. A. (2010). Improved predictability of stratospheric sudden warming events in an atmospheric general circulation model with enhanced stratospheric resolution. Journal of Geophysical Research, 115, D16.Find this resource:

Menary, M. B., Hodson, D. L. R., Robson, J. I., Sutton, R. T., & Wood, R. A. (2015). A mechanism of internal decadal Atlantic Ocean variability in a high-resolution coupled climate model. Journal of Climate, 28, 7764–7785.Find this resource:

Moberg, A., Alexandersson, H., Bergström, H., & Jones, P. D. (2003). Were southern Swedish summer temperatures before 1860 as warm as measured? International Journal of Climatology, 23, 1495–1521.Find this resource:

Moore, G. W. K., Renfrew, I. A., & Pickart, R. S. (2013). Multidecadal mobility of the North Atlantic Oscillation. Journal of Climate, 26, 2453–2466.Find this resource:

Morley, A., Rosenthal, Y., & Demenocal, P. (2014). Ocean-atmosphere climate shift during the mid-to-late Holocene. Earth and Planetary Science Letters, 388, 18–26.Find this resource:

Muschitiello, F., Zhang, Q., Sundqvist, H. S., Davies, F. J., & Renssen, H. (2015). Arctic climate response to the termination of the African Humid Period. Quaternary Science Reviews, 125, 91–97.Find this resource:

Nakamura, T., Yamazaki, K., Iwamoto, K., Honda, M., Miyoshi, Y., Ogawa, Y., et al. (2015). A negative phase shift of the winter AO/NAO due to the recent Arctic sea-ice reduction in late autumn. Journal of Geophysical Research, 120, 3209–3227.Find this resource:

Neff, W., Compo, G. P., Ralph, F. M., & Shupe, M. D. (2014). Continental heat anomalies and the extreme melting of the Greenland ice surface in 2012 and 1889. Journal of Geophysical Research Atmospheres, 119, 6520–6536.Find this resource:

Nesje, A., Lie, Ø., & Dahl, S. O. (2000). Is the North Atlantic Oscillation reflected in Scandinavian glacier mass balance records? Journal of Quaternary Science, 15, 587–601.Find this resource:

Ortega, P., Lehner, F., Swingedouw, D., Masson-Delmotte, V., Raible, C., Casado, M., et al. (2015). A model-tested North Atlantic Oscillation reconstruction for the past millennium. Nature, 523, 71–74.Find this resource:

Osborn, T. J. (2004). Simulating the winter North Atlantic Oscillation: The roles of internal variability and greenhouse gas forcing. Climate Dynamics, 22, 605–623.Find this resource:

Osborn T. J. (2011). Winter 2009/10 temperatures and a record-breaking North Atlantic Oscillation index. Weather, 66, 19–21.Find this resource:

Ottersen, G., Planque, B., Belgrano, A., Post, E., Reid, P. C., & Stenseth, N. C. (2001). Ecological effects of the North Atlantic Oscillation, Oecologia, 128, 1–14.Find this resource:

Overland, J., Francis, J. A., Hall, R., Hanna, E., Kim, S.-J., & Vihma, T. (2015). The Melting Arctic and midlatitude weather patterns: Are they connected? Journal of Climate, 28, 7917–7932.Find this resource:

Overland, J., Hanna, E., Hanssen-Bauer, I., Kim, S.-J., Walsh, J., Wang, M., et al. (2014). Air temperature. Arctic Report Card. Retrieved from http://www.arctic.noaa.gov/reportcard.Find this resource:

Overland, J. E., Francis, J., Hanna, E., & Wang, M. (2012). The recent shift in early summer arctic atmospheric circulation. Geophysical Research Letters, 39, L19804.Find this resource:

Overland, J. E., & Wang, M. (2015). Increased variability in the early winter subarctic North American atmospheric circulation. Journal of Climate, 28, 7297–7305.Find this resource:

Prior, J., & Kendon, M. (2011). The UK winter of 2009/2010 compared with severe winters of the last 100 years. Weather, 66, 4–10.Find this resource:

Ramos J. A., Fagundes, A. I., Xaviera, J. C., Fidalgoa, V., Ceiaa, F. R., Medeirosa, R., et al. (2015). A switch in the Atlantic Oscillation correlates with inter-annual changes in foraging location and food habits of Macronesian shearwaters (Puffinus baroli) nesting on two islands of the sub-tropical Atlantic Ocean. Deep Sea Research I, 104, 60–71.Find this resource:

Riddle, E. E., Butler, A. H., Furtado, J. C., Cohen, J. L., & Kumar, A. (2013). CFSv2 ensemble prediction of the wintertime Arctic Oscillation. Climate Dynamics, 41, 1099–1116.Find this resource:

Rivière, G., & Drouard, M. (2015). Understanding the contrasting North Atlantic Oscillation anomalies of the winters of 2010 and 2014. Geophysical Research Letters, 42, 6868–6875.Find this resource:

Scaife, A. A., Folland, C. K., Alexander, L. V., Moberg, A., & Knight, J. R. (2008). European climate extremes and the North Atlantic Oscillation. Journal of Climate, 21, 72–83.Find this resource:

Scaife, A. A., Arribas, A., Blockley, E., Brookshaw, A., Clark, R.T., Dunstone, N., et al. (2014). Skillful long-range prediction of European and North American winters. Geophysical Research Letters, 41, 2514–2519.Find this resource:

Schulz, M., & Paul, A. (2002). Holocene climate variability on centennial-to-millennial time scales: 1. Climate records from the North-Atlantic realm. In G. Wefer, W. Berger, K.-E. Behre, & E. Jansen (Eds.), Climate development and history of the North Atlantic realm (pp. 41–54). Berlin: Springer.Find this resource:

Sowers, J., Vengosh, A., & Weinthal, E. (2011). Climate change, water resources, and the politics of adaptation in the Middle East and North Africa. Climatic Change, 104, 599–627.Find this resource:

Stephenson, D. B., Pavan, V., & Bojariu, R. (2000). Is the North Atlantic Oscillation a random walk? International Journal of Climatology, 20, 1–18.Find this resource:

Stephenson, D. B., Wanner, H., Brӧnnimann, S., & Luterbacher, J. (2003). The history of scientific research on the North Atlantic Oscillation. In J. W. Hurrell, Y. Kushnir, G. Ottersen, & M. Visbeck (Eds.), The North Atlantic Oscillation: Climatic significance and environmental impact (pp. 37–50). Washington, DC: American Geophysical Union.Find this resource:

Sun C., Li, J., Feng, J., & Xie, F. (2015a). A decadal-scale teleconnection between the North Atlantic Oscillation and subtropical eastern Australian rainfall. Journal of Climate, 28, 1074–1092.Find this resource:

Sun, L. T., Deser, C., & Tomas, R. A. (2015b). Mechanisms of stratospheric and tropospheric circulation response to projected Arctic sea ice loss. Journal of Climate, 28, 7824–7845.Find this resource:

Sutton, R. T., & Dong, B. (2012). Atlantic Ocean influence on a shift in European climate in the 1990s. Nature Geoscience, 5, 788–792.Find this resource:

Taylor, K. E., Stouffer, R. J., & Meehl, G. A. (2012). An overview of CMIP5 and the experiment design. Bulletin of the American Meteorological Society, 93, 485–498.Find this resource:

Thompson, D. W. J., & Li Y. (2015). Baroclinic and barotropic annular viability in the Northern Hemisphere. Journal of the Atmospheric Sciences, 72, 1117–1136.Find this resource:

Thompson, D. W. J., & Wallace, J. M. (1998). The Arctic Oscillation signature in the wintertime geopotential height and temperature fields. Geophysical Research Letters, 25, 1297–1300.Find this resource:

Trigo, R. M., Pozo-Vazquez, D., Osborn, T. J., Castro-Dıez, Y., Gamiz-Fortis, S., & Esteban-Parra, M. J. (2004). North Atlantic oscillation influence on precipitation, river flow and water resources in the Iberian Peninsula. International Journal of Climatology, 24, 925–944.Find this resource:

Trouet, V., Esper, J., Graham, N. E., Baker, A., Scourse, J. D., & Frank, D. C. (2009). Persistent positive North Atlantic Oscillation mode dominated the Medieval Climate Anomaly. Science, 324, 78–80.Find this resource:

van Loon, H., & Rogers, J. C. (1978). The seesaw in winter temperatures between Greenland and northern Europe. Part I: General description. Monthly Weather Review, 106, 296–310.Find this resource:

Walcker, R., Anthony, E. J., Cassou, C., Aller, R. C., Gardel, A., Proisy, C., et al. (2015). Fluctuations in the extent of mangroves driven by multi-decadal changes in North Atlantic waves. Journal of Biogeography, 42, 2209–2219.Find this resource:

Westgarth-Smith, A. R., Roy, D. B., Scholze, M., Tucker, A., & Sumpter, J. P. (2012). The role of the North Atlantic Oscillation in controlling U.K. butterfly population size and phenology. Ecological Entomology, 37, 221–232.Find this resource:

Wigley, T. M. L., & Jones, P. D. (1987). England and Wales precipitation: A discussion of recent changes in variability and an update to 1985. Journal of Climatology, 7, 231–246.Find this resource:

Woollings, T., Franzke, F., Hodson, D. L. R., Dong, B., Barnes, E. A., Raible, C., et al. (2015). Contrasting interannual and multidecadal NAO variability. Climate Dynamics, 49, 539–556.Find this resource:

Woollings, T., Hannachi, A., Hoskins, B., & Turner, A. (2010). A regime view of the North Atlantic Oscillation and its response to anthropogenic forcing, Journal of Climate, 23, 1291–1307.Find this resource:

Wrzesiński, D., Choiński, A., & Ptak, M. (2015). Effect of the North Atlantic Oscillation on the thermal characteristics of lakes in Poland. Acta Geophysica, 63, 863–883.Find this resource: