Arctic & North Atlantic Oscillations – Watts Up With That?

By Andy May

The Arctic Oscillation (AO) is closely related to the NAO (the North Atlantic Oscillation discussed below) but they are not the same. The NAO is usually measured using the SLP (sea level air pressure) difference between the Azores or the Iberian Peninsula and Iceland and is a North Atlantic regional phenomenon, whereas the Arctic Oscillation is the SLP difference between the northern mid-latitudes and the Arctic, and is evident in all longitudes (Thompson & Wallace, 2001). The AO accounts for more of the variance in Northern Hemisphere surface air temperature than the NAO and is tightly connected to the stratospheric polar vortex (Higgins, et al., 2000) and (Thompson & Wallace, 1998). We will discuss these oscillations together in this post.

The Arctic Oscillation

The Arctic Oscillation (AO) is also called the Northern Annular Mode or NAM. It is analogous to the Southern Annular Mode or SAM discussed in Climate Oscillations 5. However, there is a large difference, whereas SAM is an oscillation over an ocean that surrounds land, NAM is an oscillation over land that surrounds a polar ocean. Thus, they act differently.

When NAM or the AO Index is positive (lower than normal pressure in the Arctic, and/or higher pressure in the mid-latitudes), the high latitude westerly polar jet-stream winds move closer to the pole and storms (which transport heat) move northward. When it is negative (higher pressure in the Arctic), the jet stream weakens, becomes more loopy or wavier, and moves south allowing Arctic air to spill into the middle latitudes causing colder mid-latitude winters. The AO Index is only computed using data from December through February because it only has a significant impact in the winter months (Baldwin & Dunkerton, 1999).

The tropopause is quite low in the Arctic, only about 8 km above the surface, so it is not surprising that the AO is strongly connected to, and influenced by, the stratosphere (Baldwin, et al., 2019), especially in the winter months when the tropospheric and lower stratospheric circulations are coupled in the polar regions (Thompson & Wallace, 2001). A large positive AO Index represents a strong well-organized polar vortex in the stratosphere above the North Pole (Baldwin & Dunkerton, 2001) and (Thompson & Wallace, 1998), just as a positive SAM indicates a strong polar vortex above the South Pole.

The changes in the Sun over the course of the 11-year Schwabe solar cycle affect the stratosphere more than the surface because the shorter wavelength UV (solar ultraviolet radiation) content of sunlight changes more than the longer wavelength visible light that makes it to the surface. The amount of UV absorbed in the stratosphere can increase by 10% or more at the peak of the 11-year Schwabe Cycle. The UV absorbed in the stratosphere both warms it and contributes to stratospheric ozone which also absorbs UV and contributes to further warming. The UV warming affects stratospheric circulation and the strength of the polar vortex which transmits some of the stratospheric changes to the troposphere affecting global weather patterns (Haigh, 2011). We will discuss this later in the series, but in essence most of the ozone is produced in the tropics, which receives the most solar radiation. There is an upward transport of tropospheric air to the stratosphere in the tropics that sets up a transport of stratospheric air toward the poles (the Brewer-Dobson Circulation), where air is taken down from the stratosphere to the troposphere via the polar vortex (Baldwin, et al., 2019). The El Nino/Southern Oscillation (ENSO) process is involved in modulating the tropical transport of tropospheric air into the stratosphere. The AO has been called the “dominant mode of variability in the [northern] extratropics” (Higgins, et al., 2000).

Trends in the AO

As shown in figure 1, the AO is steadily increasing during the 20^th century, but not as strongly as the SAM is ( see figure 3, here). This tells us, that on average, the northern polar vortex is strengthening, which leads to warming in the middle northern latitudes.

Figure 1 shows a slight increase in the full-year average AO and suggests cooling (more negative) from the late 1940s to the 1970s. It also shows warming from the 1970s to the early 1990s. The trend toward a more positive AO has reduced the severity of winter weather in the middle- and high-latitude Northern Hemisphere continental regions (Thompson & Wallace, 2001). The polar vortex is much stronger in the winter storm season in the Arctic, so we show the winter average over the same period in figure 2.

As can be seen in figure 2, the severe winter weather observed in the northern mid-latitudes from the late 1950s to 1970 and from 1976-1985 appears in the AO record. Unusual winter weather in these periods is documented here, here, here, here, here, and here. Mild winter weather was observed in the early 1970s and late 1980s to the early 1990s as shown here. The 1960s were also very cold in Asia, but there has been a warming trend since then (Kim & Choi, 2021).

The North Atlantic Oscillation

The NAO or the North Atlantic Oscillation is a very important oscillation in both climate prediction and weather prediction. However, when researchers compute the NAO indices with CMIP5 and CMIP6 climate model results they look like white noise with almost no serial correlation (Eade, et al., 2022).

Long-term weather observations from across the globe reveal patterns and links between seemingly random events and disconnected places. These long-distance relationships reveal changes in the meridional transport of energy from the tropics to the poles. For example, when you stitch together daily observations of air pressure across the Northern Hemisphere, you see large areas of high and low sea level air pressure (SLP) that flow and shift from place to place. These shifts in surface air pressure represent shifts in atmospheric mass from place to place. There’s a pattern to the shifting that is sort of like water sloshing back and forth in a bowl. The atmosphere sloshes northward; air pressure strengthens over the Arctic and weakens over the midlatitudes (either the Atlantic or Pacific Oceans, or both). Then the atmosphere sloshes southward; air pressure strengthens over the midlatitudes and weakens over the Arctic.

North of the equator, the most significant “long-distance relationship” in the Atlantic is between an area of persistently low pressure in the vicinity of Iceland and an area of persistently high pressure over the Azores Islands or the Gibraltar area. When the pressure is lower than average over Iceland and higher than average over the Azores Islands and Gibraltar, the North Atlantic Oscillation is said to be in its positive mode. When the opposite occurs, the North Atlantic Oscillation is said to be negative.

The Arctic and North Atlantic Oscillations are related to one another, and to the AMO (see here and here). Meteorologists often call these relationships and long-distance oscillations “teleconnections.” Teleconnection is as good a name as any, they are actually components of, and indicators of, changes in meridional transport.

Both the Arctic Oscillation and the North Atlantic Oscillation are defined with sea level air pressure or SLP, and the patterns are well illustrated by Rebecca Lindsey here (Lindsey, 2011). These patterns and the resulting NAO surface temperatures are shown in figure 3.

Past prolonged NAO (see figure 4) trends that last several decades cannot be explained by current climate models. The models clearly do not reflect multidecadal meridional transport regimes. Climate model simulations of NAO indices resemble white noise without serial correlation according to Rosie Eade and her colleagues at the MET office in the UK (Eade et al. 2021). There is a very small chance (1 in 20) that climate models emulate the observed NAO since 1860. Yet, figure 3 suggests that NAO trends are a key indicator of meridional transport (MT) strength. During the negative phase a lot of heat is transported poleward warming the polar region and during the positive phase of the NAO, little heat is transported to the polar region and it stays cold.

If the models cannot simulate meridional transport or the NAO, they cannot explain climate change. As discussed above, the polar vortex is strongest in the winter months and when the AO is positive. A strong winter polar vortex keeps the cold air in the Arctic and keeps warm air from being transported to the pole, thus delaying its expulsion to space and warming the middle latitudes, including the United States and Europe. In figure 4 we plot the AO in winter as a proxy for polar vortex strength and we see that the NAO is generally positive when the AO is positive (Wallace, 2006), we also see that cooler periods in the Northern Hemisphere and globally (1950s, 1960s, and early 1970s) show a declining NAO trend and a negative winter AO.

Figure 5 shows the same thing as figure 4, but only the winter NAO values are averaged. As David Parker and colleagues (Parker, et al., 2007) have noted the increase in the winter NAO from 1965 to 1995 is dramatic. It can be seen in the whole-year average shown in figure 4, but it is much more obvious in figure 5. It also shows a strong correlation to the winter AO and thus the strength of the winter polar vortex.

The NAM and the AO are two names for the same oscillation. The true measure of the strength of the polar vortex is the “PCH” or the composite geopotential height anomaly (“polar cap height”) averaged from 65°N to the pole and normalized by its standard deviation (Kim & Choi, 2021). Except for the PCH, the AO is the strongest proxy for the winter polar vortex strength, but the winter NAO can also be used as illustrated in figure 5. Data quality prior to 1950 is poor, but since then there is a good correspondence between the AO and the NAO in winter.

James Hurrell (Hurrell, 1995) points out the rapid rise in the winter NAO since 1965, and especially from the 1980s to the early 1990s. He adds that past decade-long changes in the NAO, and associated changes in atmospheric circulation, have contributed substantially to regional warming which complicates the interpretation of the effect of greenhouse gases on climate. He adds that the relationship of the NAO to greenhouse gas forcing should be examined. He asks that we investigate how well the climate models simulate the NAO, since it has a large effect on the climate over much of the world. Later Rosie Eade (Eade, Stephenson, & Scaife, 2022) did such a study and could not find the critical NAO in the models at all.

Conclusions

The Arctic and North Atlantic Oscillations are the dominant modes of variability in Northern Hemisphere climate. The observed positive trend in the AO/NAO in recent decades (see figures 2 & 5) is not reproduced in the CMIP5 or CMIP6 climate models, in fact the multi-model multi-member ensemble mean of the trend is zero (IPCC, 2021, p. 490). AR6 adds, on the same page, that the observed NAO trend lies outside the 5^th-95^th percentile range of the CMIP6 climate model distribution and the AO trend lies above the 90^th percentile. It seems very unlikely that the models are useful with results like this.

Download the bibliography here.

Previous posts in this series:

Musings on the AMO

The Bray Cycle and AMO

Climate Oscillations 1: The Regression

Climate Oscillations 2: The Western Hemisphere Warm Pool (WHWP)

Climate Oscillations 3: Northern Hemisphere Sea Ice Area

Climate Oscillations 4: The Length of Day (LOD)

Climate Oscillations 5: SAM

Climate Oscillations 6: Atlantic Meridional Model

Climate Oscillations 7: The Pacific mean SST

Climate Oscillations 8: The NPI and PDO

Related