1. Introduction
The prediction of blocking, commonly regarded as the persistent breakdown of the extratropical westerly zonal flow, characterized by a split jet stream and an upper-level synoptic-scale ridge, remains a significant challenge for numerical weather and climate forecasting models (e.g., Pelly and Hoskins 2003b; Dunn-Sigouin et al. 2013). Because of its potential for triggering catastrophic floods (Galarneau et al. 2012), droughts (Green 1977; Dole et al. 2011), and warm temperature extremes (Pfahl and Wernli 2012), the dynamics and predictability of blocking are important considerations for numerical weather and climate prediction and for the attribution of observed changes in regional extremes of temperature or precipitation. To more fully understand the variety of dynamic and thermodynamic processes that could give rise to blocking events in different regions and seasons, an objective index of blocking must be able to identify representative blocking events across regions and seasons. Unfortunately, no universally accepted index of blocking currently exists, leading to discrepancies in the analysis of the associated dynamics. The lack of a universally accepted blocking index contributes to discrepancies in the climatology of blocking presented by numerous studies.
For example, most studies of blocking that use 500-hPa geopotential height anomalies have tended to identify winter and spring as the active season in the Pacific and Euro-Atlantic regions with reduced frequency in autumn and summer (e.g., Lejenäs and Økland 1983; Tibaldi and Molteni 1990; Barriopedro et al. 2006). Croci-Maspoli et al. (2007), using an index based on vertically integrated potential vorticity (PV), also find that blocking is rare in summer, but suggest that blocking is also frequent in autumn over both regions. Shukla and Mo (1983), who use a seasonally varying definition of the geopotential height anomaly that constitutes a block, find that the seasonal differences in blocking frequency are much smaller than other studies. Pelly and Hoskins (2003a) and Tyrlis and Hoskins (2008b), using an index based on the reversal of the potential temperature gradient on the dynamic tropopause, find that blocking over the central Pacific peaks in summer and winter and is rare in spring and autumn.
The identification of a summer peak in the blocking frequency reported by Pelly and Hoskins (2003a) and Tyrlis and Hoskins (2008b) is important because it clearly demonstrates the sensitivity of the climatology to the choice of an index. For example, individual summer blocking events have been implicated in the occurrence of extreme heat waves, floods, and droughts (e.g., Hawkins 1954; Holland 1954; Winston 1954; Green 1977; Dole et al. 2011; Matsueda 2011; Houze et al. 2011; Hong et al. 2011; Galarneau et al. 2012); yet, most climatological studies report that summer blocking is a rare event. Given the potential for warm season blocking to trigger droughts and floods, this raises the important question of whether the apparent low frequency of summer blocking events in the climatology is because they are indeed a rare occurrence or whether existing indices simply fail to detect them.
No single index can be expected to identify all blocking events given that an assumption must be made somewhere during the process of designing a procedure to detect blocking. However, an index should be able to provide a representative sample of blocking events with different structures (i.e., dipole and omega) and that occur in different seasons and during different synoptic conditions and that are triggered and maintained by different dynamic and thermodynamic processes. They should therefore make no strong assumptions regarding the dynamics producing the blocking that would bias the results toward one type of blocking or one season. For example, the Pelly and Hoskins (2003a) index assumes blocking is synonymous with a breaking wave that reverses the gradient of the potential temperature. It is not surprising that studies using this index find that blocking is primarily associated with cyclonic or anticyclonic wave breaking (Tyrlis and Hoskins 2008b). However, Altenhoff et al. (2008) report that less than half of blocking events are associated with strong wave breaking, suggesting that the assumption about wave breaking is a questionable choice when constructing a blocking index.
This study is motivated by the apparent discrepancy in the seasonal blocking climatology reported by earlier studies. The primary goal of this paper is to propose a blocking index that is dynamically relevant but does not make a strong assumption about the dynamics that would limit the general applicability of the index.
2. Overview and critique of existing blocking indices
One essential ingredient of a block is a large-amplitude equivalent-barotropic anticyclone on the poleward side of anomalous easterlies or reduced westerlies. A blocking pattern is usually identified on synoptic charts by the presence of a persistent surface anticyclone underneath a ridge at 500 hPa, prompting several authors to define blocking indices based on the persistence of midtropospheric geopotential height anomalies (Dole and Gordon 1983; Dole 1986; Mullen 1986; Shukla and Mo 1983).
Blocking is also associated with a reduction in the westerly geostrophic flow, leading many studies to identify blocking events as a persistent, synoptic-scale reversal of the 500-hPa geopotential height gradient (Rex 1950a,b; Austin 1980; Lejenäs and Økland 1983; Tibaldi and Molteni 1990; Tibaldi et al. 1994; Lupo and Smith 1998; Barriopedro et al. 2006; Diao et al. 2006; Barnes and Hartmann 2010). Recent studies have also modeled blocking as a breaking wave that cuts off from the westerly flow and produces a local reversal of the potential gradient on the dynamic tropopause. These studies identify blocking as a reversal of the climatological potential temperature gradient on the dynamic tropopause (e.g., Pelly and Hoskins 2003a; Tyrlis and Hoskins 2008a). Schwierz et al. (2004) suggest that blocking should be defined from the most salient characteristic of the process, the presence of a synoptic-scale, negative potential vorticity anomaly underneath a raised tropopause. They defined a quasi-three-dimensional blocking index based on the evolution of the column-averaged departure of PV from its climatological value from the midtroposphere (500 hPa) to the lower stratosphere (150 hPa).
There are several potential problems with commonly used blocking indices that might bias the results toward the cold season or specific regions. For example, Illari (1984) showed that the circulation anomalies associated with a summer blocking event were strongest at jet level in the upper troposphere near the tropopause and not at 500 hPa. The amplitude of a blocking anticyclone might not always be reflected in the 500-hPa geopotential height field, suggesting that this might not be the most appropriate field for identifying blocking events (Schwierz et al. 2004).
This concern is especially important in summer when the 500-hPa level is likely to be located well underneath jet level in the midtroposphere as the troposphere-spanning eddy-driven jet stream moves north and the midlatitudes become more strongly influenced by the shallower subtropical jet that tends to be confined to the upper troposphere (Woollings et al. 2010). A meridional cross section of the mean daily zonal wind in the central Pacific (180°) demonstrates this is the case (Fig. 1). In winter, the strongest winds are near 200 hPa, but the core of the jet extends much deeper into the midtroposphere than in summer when it is confined well above the 500-hPa level. In section 3, after defining a blocking index, we demonstrate that the circulation anomalies associated with blocking peak at 200 hPa in all seasons.
The meridional cross section of the mean daily zonal wind (m s−1) in the central Pacific (180°) for winter (DJF) and summer [June–August (JJA)]. The thick black line indicates the 500-hPa pressure level.
Citation: Journal of Climate 27, 8; 10.1175/JCLI-D-13-00374.1
Schwierz et al. (2004) also argue that there is no reason to expect that a reversal of an appropriate gradient, whether the 500-hPa geopotential height or tropopause potential temperature, will always signify the presence of an equivalent-barotropic blocking anticyclone. For example, the recent PV–θ blocking index proposed by Pelly and Hoskins (2003a) models blocking as a breaking wave that results in a synoptic-scale reversal of the potential temperature gradient on the dynamic tropopause. Numerous subsequent studies (Tyrlis and Hoskins 2008b; Woollings et al. 2008; Masato et al. 2012) have applied this index to the study of blocking under the assumption that blocking is primarily associated with wave breaking. Altenhoff et al. (2008) tested the validity of this assumption and found that although there is a statistically significant increase in the frequency of wave breaking during blocking episodes compared to climatology, wave breaking is observed in only 36% (42%) of Pacific (Atlantic) blocking cases. This suggests that any blocking index that assumes a gradient reversal might often fail to detect those blocking events that are not triggered by wave breaking that drives a subsequent reversal of the potential temperature or geopotential height gradient.
To demonstrate that blocking events, even those associated with a wave breaking event, are not best characterized by a gradient reversal, we calculated the PV–θ blocking index of Pelly and Hoskins (2003a) from potential temperature interpolated onto the dynamic tropopause between 1948 and November 2012. We used the definition of blocking in Tyrlis and Hoskins (2008a) based on the PV–θ blocking index (PH2003), which defined any 45° of longitude as blocked if it contains at least 15 consecutive degrees of longitude with a reversal of the potential temperature gradient (not shown). We repeated the calculation using the same definition of blocking (i.e., a synoptic-scale reversal of the gradient) using the 500-hPa geopotential height (not shown) instead of potential temperature (Pelly and Hoskins 2003a; Barnes and Hartmann 2010). Figure 2 shows the evolution of the potential temperature and wind fields on the dynamic tropopause during an event in autumn of 2003 that neither index correctly identified as blocking.
The potential temperature (K; shading) and wind [barbs; knots (kt; 1 kt ≈ 0.51 m s−1)] on the dynamic tropopause during a blocking event at selected times between 25 Oct and 3 Nov 2003.
Citation: Journal of Climate 27, 8; 10.1175/JCLI-D-13-00374.1
The blocking ridge over the eastern Pacific (Fig. 2) is highly amplified at 1200 UTC 25 October 2003 (Fig. 2a) with lower potential temperature air to north of very high potential air and a vigorous anticyclonic circulation, demonstrating that a strong reversal of the gradient is not always present. On 28 October 2003 at 1200 UTC (Fig. 1b), a breaking wave is observed in the region of the block along with an irregular reversal of the gradient. Over the following 6 days (Figs. 2c–e), the upper-level anticyclone becomes highly amplified with lower potential temperature air to the north of higher potential temperature air and a marked northward displacement of the jet stream around the quasi-stationary ridge. Although there are small longitude bands where the potential temperature gradient is reversed, the structure of the ridge is better described as a large, relatively uniform air mass that is surrounded by lower potential temperature air. The relatively low potential temperature air in the northern sector of the block is surrounded by even lower potential temperature air, supporting the anticyclonic rotation in the block. The more salient characteristic of a large anticyclone is the characteristic of the air mass relative to its surroundings, not a reversal of the gradient. This example motivates the creation of a potential vorticity-based blocking index defined relative to the zonal mean.
Schwierz et al. (2004) proposed a quasi-three-dimensional blocking index based on the day-to-day overlap of vertically integrated potential vorticity anomalies (APV*) from the midtroposphere to the lower stratosphere to overcome these limitations. It is not clear whether their index identifies negative PV anomalies that are strongest in the troposphere, especially in winter when the 150- and 200-hPa surface are likely to be in the stratosphere. Because the potential vorticity is much larger in the stratosphere than in the troposphere, deviations from the time mean are also likely to be much larger. For example, a 95th percentile PV anomaly in the stratosphere is likely to be much larger than the corresponding value in the troposphere. By defining an anomaly relative to the time mean, their index might be producing a climatology that is biased toward the cold season if stratospheric values are being vertically averaged along with tropospheric anomalies.
In this study, we address the question of how to systematically define a blocking index that is able to identify blocking events in all seasons so that we might better determine how frequently blocking occurs and identify seasonal and regional differences in the duration, scale, and frequency of blocking events.
3. Data and definition of the modified APV* blocking index
Daily averages of the PV between 500 and 150 hPa and potential temperature on the dynamic tropopause are used to identify blocking events in the Northern Hemisphere (poleward of 10°N) from 1 January 1948 through 30 November 2012. The data used in this study are taken from the National Centers for Environmental Predication–National Center for Atmospheric Research (NCEP–NCAR) global reanalysis (Kalnay et al. 1996). The potential temperature was linearly interpolated to the 2 potential vorticity unit (PVU; 1 PVU = 10−6 K m2 kg−1 s−1) surface, assumed to be the level of the dynamic tropopause.
a. Definition of the modified APV* blocking index
The original APV* index of Schwierz et al. (2004) is chosen as a starting point for this study because it is designed to identify blocking events as a persistent, negative PV anomaly underneath a raised tropopause, a dynamically meaningful quantity. The authors defined an anomaly of the vertically integrated PV with respect to time. By vertically integrating the PV from 150 to 500 hPa and taking the anomaly with respect to time it is not guaranteed that the index is actually detecting PV anomalies in the troposphere. For example, when the 150-hPa surface is in the stratosphere, the index likely reflects the large stratospheric values of PV and not smaller tropospheric values. We propose a simple remedy for this issue.
In our modified APV* (mAPV*) index, deviations from the zonal mean (i.e., eddies) are calculated with respect to the 21-day running zonal mean on each pressure level between 150 and 500 hPa and then vertically averaged to produce mAPV values. This is more dynamically consistent than defining the anomaly relative to the time mean because the strength of the circulation induced by a PV anomaly is proportional to the local horizontal PV gradient. Subtracting the zonal mean also helps to ensure that the index is identifying a strong anticyclonic circulation near the tropopause and not in the stratosphere. For example, when a pressure level in the troposphere is underneath a raised tropopause, the surrounding air is of stratospheric origin with much higher potential vorticity, leading to very large negative eddies in the PV field. When vertically averaging the eddy PV field, the resulting mAPV values are largest when the strongest eddy is on a pressure level nearest the raised tropopause. The ability of the mAPV to capture the large negative eddies near the tropopause is verified in the following section (see Fig. 4) after other modifications to the original APV* index of Schwierz et al. (2004) index are discussed.
Blocking is identified as a persistent, negative mAPV value, making the climatology sensitive to the choice of what is considered to be an appropriate anomaly. Because stronger anomalies tend to be smaller in spatial scale and much less persistent than weaker anomalies (Horel 1985; Dole 1986), the identification of persistent anomalies, and therefore blocking events, is highly sensitive to the choice of the spatial scale and cutoff value of what constitutes an anomaly. The choice of a cutoff value effectively chooses the spatial scale and persistence characteristics of anomalies that could be identified as blocking events. Croci-Maspoli et al. (2007) identify blocks as closed contours of low-pass filtered 6-hourly negative APV* anomalies that exceed −1.3 PVU and overlap by 80% from one 6-hourly period to the next. This particular choice of a large cutoff and arbitrary amount of overlap might be one reason why they find that the occurrence of blocking is frequent in the autumn, winter, and spring and rare in summer. To minimize the impact of arbitrarily choosing the magnitude and scale of what constitutes an anomaly, we adopted a systematic approach for identifying blocking.
Before creating an index of blocking, the seasonal cycle is subtracted from the mAPV values to create mAPV*, the quantity from which the modified blocking index will be created. Here, the star refers to an anomaly with respect to time. If the seasonal cycle is not removed, the index identifies unrealistically high frequencies of blocking near the planetary ridges, especially in winter. The seasonal cycle was modeled as the first two harmonics of the annual cycle of daily mAPV estimated with least squares. After removing the seasonal cycle, we identified negative mAPV* anomalies at individual grid points with magnitudes exceeding −0.5, −1.0, −1.5, and −2.0 PVUs across the Northern Hemisphere between 40° and 75°N and calculated how long each anomaly lasted. Most studies of blocking have identified 5 days as the minimum duration of a blocking event, so an appropriate value for the magnitude of a mAPV* anomaly that constitutes a block is one for which an anomaly lasting 5 days or longer is a rare event (i.e., exceeds the 95th percentile). We find that over the Northern Hemisphere, a 5-day mAPV* anomaly of −1.0 PVU is approximately a 95th percentile event (not shown) and is therefore a reasonable choice for the minimum mAPV* of a blocking anticyclone. Blocking is also defined as a synoptic or larger-scale feature, suggesting that the size of the anomalies must also be a criterion in identifying blocks. If we choose −1.0 PVU as a cutoff, we must identify a representative spatial scale for an anomaly of this magnitude during persistent blocking anticyclones.
To do so, we identified closed contours of mAPV* exceeding −1.0 PVU across the Pacific (40°–75°N, 120°E–120°W) and North Atlantic (40°–75°N, 60°W–60°E). Instead of using the closed contours directly, we defined the spatial extent of the anomaly to be the smallest bounding rectangle that will contain the mAPV* contour. The use of a bounding rectangle makes all of the mAPV* anomalies the same shape and helps in the tracking of blocking anticyclones as they develop through time by allowing the anomalies to change size, shape, and orientation during the event without necessarily causing the index to fail. We estimate the representative size of mAPV* anomalies as the area of the smallest rectangle that will contain the closed contour of the −1.0-PVU anomalies. To identify persistent anticyclonic anomalies, we draw bounding rectangles around the closed contours of −1.0-PVU anomalies and estimate the size of the overlap of the rectangles between adjacent days.
To identify a representative size of a persistent anomaly, we examined the overlap characteristics of −1.0-PVU mAPV* anomalies of different sizes (5° × 5°, 10° × 10°, 15° × 15°, etc.). The cumulative distribution functions of the duration of 15° × 15° events in the Atlantic and Pacific are shown in Fig. 3. The duration of events ranged from 1 day (no overlap in time) to 50 days. In the Pacific (Fig. 3a), approximately 72% of all the identified negative mAPV* anomalies last 1 day, with a 5-day event constituting a 95th percentile event for all seasons except winter when 6 days is approximately a 95th percentile event. In the Atlantic (Fig. 3b), an anomaly lasting 5 days is between a 93rd and 95th percentile event in all seasons. A −1.0-PVU mAPV* anomaly that overlaps by 15° × 15° for at least 5 days is a rare event in both regions and across seasons.
Empirical cumulative density function (CDF) for the duration of overlapping mAPV* anomalies (days) across the (a) Pacific and (b) Atlantic basin.
Citation: Journal of Climate 27, 8; 10.1175/JCLI-D-13-00374.1
Blocking events are then identified from persistent, overlapping rectangular bounding rectangles enclosing negative mAPV* anomalies. We identified every closed contour mAPV* anomaly exceeding −1.0 PVU and drew the bounding rectangle. Any rectangle smaller than 15° × 15° and 106 km2 was discarded. The latter criterion was applied to avoid complications caused by grid points near the pole. We then identified events as rectangles that overlap in space by at least 15° × 15° over consecutive days. In this study, only those rectangles that overlap by 15° × 15° for 5 days or longer are considered to be a blocking event. To create the blocking index, all grid points that are contained within a 15° × 15° or larger area where rectangles of APV* anomalies overlap for 5 consecutive days or longer are set to 1 while all other grid points are set equal to 0. The frequency of blocking is then estimated by summing the resulting index at each grid point. Before presenting an example of a blocking event identified by the mAPV* index and a detailed climatology, we demonstrate that the mAPV* identifies anomalies that are near the tropopause and not stratospheric.
One potential weakness of the mAPV* index is the possibility that it might identify features in the stratosphere and not the troposphere as blocks. To demonstrate that the mAPV* primarily captures tropospheric phenomena, we identified the strongest negative PV anomaly (defined with respect to the 21-day running zonal mean) at each pressure level on all days found to be blocked between 1948 and 2012. The stacked histograms for the Pacific and Atlantic (Figs. 4a,c) show that the peak PV anomaly is most often located on the 200-hPa surface in all seasons. The more important consideration is the location of the PV anomaly with respect to the vertical position of the tropopause. We also calculated boxplots of the PV value at the location of the largest negative PV anomaly at each grid point (Figs. 4b,d). The PV value corresponding to the peak of the PV anomaly shows where in the column (relative to the tropopause) that the anomaly is located. The boxplots clearly demonstrate that the maximum negative PV anomaly is below the 2.0-PVU surface approximately 75% of the time in all seasons, below the 3.0-PVU surface 98% of the time, and below the 4.0-PVU surface more than 99.9% of the time. Because the tropopause is often defined as a constant PV surface between 1.5 and 4.0 PVU, the results in Fig. 4 demonstrate that the strongest PV anomalies captured by the mAPV* as blocking anticyclones that reside near or below the dynamic tropopause and not primarily in the stratosphere. The results also suggest that the 200-hPa isobaric surface might be the best pressure level to identify blocking events across seasons.
(a),(c) The stacked histogram showing the pressure level of the maximum PV anomaly in each season [winter (DJF), spring (March–May), summer (JJA), and autumn (September–November)] across the Pacific and Atlantic, respectively. (b),(d) Boxplots showing the distribution of the corresponding PV at the grid point and level of the maximum negative PV anomaly over the Pacific and Atlantic.
Citation: Journal of Climate 27, 8; 10.1175/JCLI-D-13-00374.1
b. An illustrative example of the modified index
We present an example of a blocking event in the Gulf of Alaska that began on 16 September 2012 (day 0) and persisted for 8 days to demonstrate how well the index identifies blocking events. Fig. 5 shows the 200-hPa PV and mAPV* during different stages of the development of the blocking event beginning 2 days before onset (day −2) and continuing until 1 day after (day +1). The red rectangle from onset (day 0) forward is the bounding rectangle enclosing the −1.0-PVU mAPV* anomaly at each time. At day −2 (Fig. 5a), we observe an amplifying, upper-level ridge in the eastern Pacific near Japan. As the ridge continues to build on the following day (day −1; Fig. 5b), the downstream trough deepens and a ridge starts to form over the eastern Pacific. By day 0, the date of onset, a rather substantial ridge, with large negative mAPV* anomalies, has developed over the Gulf of Alaska (Fig. 5c). At this time, the appearance of the red rectangle indicates that the index has identified the start of the blocking event. At day +1 (Fig. 5d), the anticyclone has continued to amplify, and the size of the bounding rectangle has increased.
The daily mean 200-hPa PV (PVU; contours) and mAPV* (PVU; shading) for a Pacific blocking event that began on 16 Sep 2012 (day 0). The red rectangles indicate the region where the mAPV* index identified blocking.
Citation: Journal of Climate 27, 8; 10.1175/JCLI-D-13-00374.1
The continued development of the mature block from day 2 through 8 demonstrates that the proposed blocking index is able to identify blocking anticyclones that change shape and size during their evolution (Fig. 6). For example, on days 2–5 of the event (Figs. 6a–d), the blocking anticyclone exhibits a very clear dipole structure with negative mAPV* anomalies situated to the north of a strong positive anomaly. Between days 6 and 8 (Figs. 6e–g), the scale of the anticyclone increased, and the structure of the event changed from a dipole to more closely resemble an omega block. By day 9 (Fig. 6h; termed lysis), an anticyclone is still clearly evident, but it has moved to the east and no longer overlaps with the mAPV* anomaly on day 0. This example shows that the proposed procedure for identifying blocking events is flexible enough to correctly identify the onset and decay and track the evolution of a blocking event that changes in shape and size over time.
As in Fig. 5, but for 18 Sep 2012 (day 2) through lysis.
Citation: Journal of Climate 27, 8; 10.1175/JCLI-D-13-00374.1
The ability of the index to capture realistic synoptic structures is seen in the 500-hPa geopotential height and mean sea level pressure fields at selected times during the blocking event (Fig. 7). The geopotential height fields indicate an omega block with a highly amplified ridge at 500 hPa and a deep cyclone upstream and surface anticyclone along the eastern edge of the block. An equivalent-barotropic structure can be seen in the high sea level pressure underneath the ridge.
The 500-hPa geopotential height (contours; 60-m contour interval) and sea level pressure (hPa; shading) at selected times during the blocking event presented in Figs. 5 and 6.
Citation: Journal of Climate 27, 8; 10.1175/JCLI-D-13-00374.1
In the following section, a climatology of blocking events from 1948 to 2012 that examines the annual, interannual, seasonal, and monthly patterns of blocking is presented.
4. Blocking climatology
a. Annual blocking frequency
The modified mAPV* blocking index was calculated for the period 1 January 1948 through 30 November 2012 using the procedure described in section 2b. The geographical distribution of the annual blocking frequency (Fig. 8) shows a bimodal frequency structure with peaks in blocking frequency in the Euro-Atlantic and Pacific regions downstream of the Asian and North American continents consistent with most previous studies of blocking (e.g., Rex 1950a,b; Lejenäs and Økland 1983; Tibaldi and Molteni 1990; Croci-Maspoli et al. 2007), along with a third, weaker peak over the European continent. The rest of the study focuses on the seasonal and intraseasonal characteristics of blocking, including differences in the duration, intensity, and locations of blocking events.
The annual blocking frequency (1948–2012) identified using the mAPV* index. The units are fraction of days classified as being blocked.
Citation: Journal of Climate 27, 8; 10.1175/JCLI-D-13-00374.1
b. Seasonal blocking statistics
The statistics of blocking events in the Pacific (40°–75°N, 120°E–120°W) and North Atlantic (40°–75°N, 60°W–60°E) are shown in Fig. 9 to highlight similarities and differences in the seasonal characteristics of blocking events over the two regions. The cumulative distribution function of the blocking event duration (Figs. 9a,d) for the two regions suggests that blocking events in the two basins are of comparable length. In the Pacific, between 25% and 30% of all events last only 5 days, the minimum duration assumed for a blocking events. In the Atlantic, 20% of winter blocking events last 5 days, while nearly 35% of summer events last 5 days. In both basins, winter and autumn events tend to last longer than spring and summer events with a 95th percentile blocking event, for example, increasing from 14 days in summer to 19 days in winter.
Blocking event statistics by season. The empirical CDF of (a) Pacific blocking (120°E–120°W) and (d) Atlantic (60°E–60°W) event duration (days) in different seasons. Boxplots of the (b),(e) size (106 km2) and (c),(f) maximum negative mAPV* (PVU) for all of the days during blocking episodes.
Citation: Journal of Climate 27, 8; 10.1175/JCLI-D-13-00374.1
Winter blocking events in both basins also tend to be much larger in spatial scale than events in other seasons, while summer events tend to be the smallest (Figs. 9b,e). The boxplots show a clear seasonal cycle in both basins, although it is more pronounced in the Atlantic. The median event size peaks with a median size of approximately 4 × 106 km2 in the winter over both basins and decreases in spring before reaching the annual minimum in summer and increasing again in autumn. The intensity of the blocking event is evaluated as the peak negative mAPV* anomaly on each day of blocking events (Figs. 9c,e). In the Pacific, the intensity also exhibits a strong seasonal cycle, peaking in the summer and reaching the annual minimum in the winter. Over the Atlantic, the seasonal differences are less pronounced with a slight shift toward more strongly negative mAPV* anomalies in the autumn. The boxplots of the duration, size, and intensity demonstrate that most blocking events greatly exceed the cutoff values of 5 days, −1.0 PVU, and 106 km2. In general, blocking anticyclones in winter tend to be the largest, longest lasting, and weakest in magnitude, while summer events tend to be the shortest lived, smallest, and most intense. This tendency for summer events to be smaller, shorter lived, and more intense might reflect the importance of diabatic effects for the triggering and maintenance of anticyclones in summer when the temperature gradients and quasigeostrophic forcing are weaker.
The spatial distribution of the blocking frequency in different seasons is shown in Fig. 10. The mAPV* index indicates that winter [December–February (DJF)] blocking peaks strongly in the central Pacific east of the Aleutians and in the western Atlantic southeast of Greenland with between 18% and 20% of days identified as blocked in both regions (Fig. 10a). A secondary area of enhanced blocking frequency is also identified that stretches from western Europe across the Urals into western Asia with little blocking over East Asia or Siberia in winter. The spring pattern of blocking frequency is similar to winter but with reduced frequency in the western Atlantic and central Pacific and enhanced frequency of European continental blocking east of 60°E (Fig. 10b). The frequency of blocking activity decreases substantially across the globe in summer with the primary peak in the western Atlantic remaining in the same location, but with the Pacific peak becoming elongated so that it extends from the central and eastern Pacific westward into eastern Siberia (Fig. 10c). In general, the summer pattern of blocking frequency extends in a narrow band across most of the globe around the 60°N line exhibiting a weak five-peak structure (Atlantic, western Europe, East Asia, central Pacific, and North America) peaking with 8% and 10% of days identified as blocked in the North Atlantic, eastern Europe, and the western and central Pacific.
Seasonal blocking frequency identified using the modified APV* index. The blocking frequency is defined as the fraction of (a) winter, (b) spring, (c) summer, and (d) autumn days identified as being blocked.
Citation: Journal of Climate 27, 8; 10.1175/JCLI-D-13-00374.1
The autumn pattern is generally similar to the winter and spring patterns globally with two large peaks in blocking frequency over the Atlantic and Pacific basins, but with two pronounced differences (Fig. 10d). First, the peak in the Pacific is located in the eastern part of the basin along the west coast of North America and not in the central Pacific east of the Aleutian Islands. The autumn blocking frequency in the Pacific is also much higher than many studies have reported and is only slightly weaker than the winter peak over the basin. Over the Atlantic, blocking frequency reaches its annual peak in the autumn southeast of Greenland in the same general location where it peaks in the other seasons, but also exhibits a strong peak in frequency over Scandinavia and northern Europe that is much higher than in the other seasons.
One interesting feature of the mAPV* climatology in Fig. 10 is the that in the Pacific the blocking pattern migrates with the seasons, peaking in the east-central Pacific in winter and spring, in the west-central Pacific in summer, and the eastern Pacific in autumn. Although not shown, the seasonal migration of the blocking peak tends to follow the seasonal cycle of the jet stream. For example, the blocking peak is located east of the jet exit region in the western Pacific in summer when the jet stream is weak over the basin and in the eastern Gulf of Alaska in autumn after the jet becomes stronger and extends eastward over the Pacific. In the Atlantic, the peak remains in the same location southeast of Greenland in all seasons but with an increase in continental blocking activity that peaks west of 60°E in the spring and east of 30°E in autumn (when it is much stronger). Croci-Maspoli et al. (2007) also reported a large peak in autumn blocking frequency extending eastward into Scandinavia, but otherwise found that the blocking locations are similar across seasons and basins.
The mAPV* index identifies slightly higher blocking frequencies across basins than Pelly and Hoskins (2003a) and Croci-Maspoli et al. (2007) for all seasons. The summer blocking frequency identified over the Pacific is much higher than that found by Croci-Maspoli et al. (2007) and is consistent with that found by Pelly and Hoskins (2003a) and Tyrlis and Hoskins (2008a). In autumn and spring, our analysis agrees well with the analysis of Croci-Maspoli et al. (2007) over both basins, but disagrees substantially with Pelly and Hoskins (2003a) and Tyrlis et al. (2008a), who find that blocking rarely occurs in either season over the Pacific. A possible explanation for this discrepancy is offered in section 4e.
Regional and seasonal differences in the frequency of blocking could be attributable to differences in the number of events occurring over time or because of differences in the duration of the events. The boxplots in Fig. 9 demonstrate that regional differences in the event duration are minimal in all seasons, suggesting that the observed pattern of blocking frequency is attributable to the frequency of events. Maps of the median blocking duration and blocking event density (Fig. 11) show that this is the case. Blocking density is defined as the number of events when individual grid points are blocked for at least 5 consecutive days. The median is used to quantify differences in event duration because the distribution is non-Gaussian. The peaks in the density of blocking events strongly follow the seasonal pattern of blocking frequency (Fig. 10) with two strong peaks over the western Atlantic and central Pacific from autumn through spring and a five-peak pattern in summer. The median event duration shows little seasonal or regional differences except in winter. The median blocking duration is between 6 and 7 days across most of the midlatitude Northern Hemisphere with peaks of 8 or 9 days in the preferred blocking region of the Atlantic and Pacific in winter.
The median duration of blocking events (shading; days) and the local event density (contours) for (a) winter, (b) spring, (c) summer, and (d) autumn. The contours of the event density start at 20 with a contour interval of 20 events.
Citation: Journal of Climate 27, 8; 10.1175/JCLI-D-13-00374.1
c. Interannual variability of blocking frequency
The seasonal blocking climatology was calculated for the 1948–2012 period from the NCEP–NCAR reanalysis and might be subject to the effects of trends or discontinuities in the data. To examine how the blocking frequency varies over time, two measures of the interannual variability in the blocking frequency are presented. First, the seasonal blocking frequency was calculated separately for the pre- and post-1979 periods to identify possible discontinuities because of the introduction of satellite data assimilation in 1979 (Fig. 12). The blocking frequency in the later period is shown as contours while the change from the earlier period is shaded. The difference in blocking frequency shown in Fig. 12 is the percentage of days blocked in the earlier period subtracted from the percentage of blocked days in the latter period. Hovmöller plots of the mean seasonal blocking frequency between 45° and 70°N versus time were also created to visualize any trends or interannual variability (Fig. 13).
The seasonal frequency of blocking from 1979 to 2012 (contours; fraction of blocked days) and the difference in the fraction of blocked days between the periods of 1979–2012 and 1948–1978 (shading). The contour interval is 0.02 beginning at 0.04 with dashed lines representing a decrease in blocking frequency in the later period and solid contours an increase in the latter period. A negative change indicates a decrease in blocking frequency in the later period.
Citation: Journal of Climate 27, 8; 10.1175/JCLI-D-13-00374.1
Hovmöller diagram of the seasonal blocking frequency (fraction of blocked days per season) between 45° and 70°N for (a) winter, (b) spring, (c) summer, and (d) autumn between 1948 and 2012.
Citation: Journal of Climate 27, 8; 10.1175/JCLI-D-13-00374.1
The overall pattern of blocking frequency in the post-1979 period is similar to that calculated from the entire record with a few exceptions. The frequency of blocking in winter and spring exhibit strong peaks in the central Pacific and western Atlantic, although the frequency has increased by 0.07 in the Atlantic and decreased by a similar amount in the central Pacific. The summer pattern is quite different during the later period, when peaks in blocking frequency are only observed over the central Pacific, Hudson Bay, and the western Atlantic. The continental blocking frequency has decreased by as much as 0.08 across most of western Europe and Asia, indicating that the broad peak in blocking frequency across the region has all but disappeared after 1979. The autumn pattern exhibits peaks in the same locations in both periods, although the frequency has decreased over western Europe and to a lesser extent over western North America.
The Hovmöller time series (Fig. 13) helps to determine whether the large decreases in blocking frequency over Siberia in summer and the Aleutian Islands in spring are attributable to changes in the blocking location, a discontinuity in data, a trend, or a few years with extremely high frequencies. The Hovmöller diagram shows that the zonal variability in blocking frequency tends to be much larger in the Pacific than in the Atlantic for all seasons. In winter (Fig. 13a), the blocking frequency peaks near 180° in the Pacific and near 60°E in the Atlantic, but with substantial zonal variability in the Pacific and less in the Atlantic. Across the Pacific, the winter blocking frequency also exhibits strong interannual variability over the period 1948–2012, with 12 yr when blocking frequency has spiked to more than 40% of days. The interannual variability appears to be lower in the Atlantic, although it has not been quantified. The two peaks in blocking frequency are not as consistently observed across the Pacific and Atlantic in the other seasons as they are in winter. In spring, it appears that blocking frequency over the central Pacific peaks in the 1960s and again after 2000, with a quiet period in between, although the earlier spike is much larger (Fig. 13b). The apparent decrease in blocking frequency over the central Pacific in spring is consistent with a large peak in the 1960s over the central Pacific. Over the Atlantic sector, the increase in blocking frequency appears to have been associated with a rapid increase after 1979. In summer (Fig. 13c), the peak in blocking frequency in the central Pacific near 180° is weaker than the other seasons with high interannual variability marked by many years of little or no blocking interspersed with years when 20%–40% of days are blocked (i.e., 1980s and 1990s). The Hovmöller diagram also shows that the large decrease in continental blocking frequency from eastern Europe to Siberia can be explained by a large peak in blocking activity over the region in the 1950s (Fig. 13c) when upward of 45% of all summer days were identified as blocked between 90° and 150°E. In autumn (Fig. 13d), the blocking frequency also exhibits two peaks, one over the eastern Pacific and one over the Atlantic, but with strong zonal and interannual variability such that the distinctness of the two peaks is obscured in many years. The results in sections 4 demonstrate the presence of significant seasonal and interannual variability in the distribution of blocking. In the following section, the question of intraseasonal variability is addressed.
d. Intraseasonal patterns of blocking frequency over the Pacific
We also calculated the monthly climatology of blocking frequency for each of the cold season [October–March (ONDJFM); Fig. 14] and warm season [April–September (AMJJAS); Fig. 15] months. The results show patterns of monthly blocking frequency and location that are markedly different from the seasonal patterns, particularly over the Pacific in August and September and over the Atlantic in September and October.
Cold season monthly blocking frequency identified using the modified APV* index. The blocking frequency is defined as the fraction of (a) October, (b) November, (c) December, (d) January, (e) February, and (f) March days identified as being blocked.
Citation: Journal of Climate 27, 8; 10.1175/JCLI-D-13-00374.1
As in Fig. 14, but for the warm season months of (a) April, (b) May, (c) June, (d) July, (e) August, and (f) September.
Citation: Journal of Climate 27, 8; 10.1175/JCLI-D-13-00374.1
In the Pacific, the cold season (ONDJFM) peak in blocking frequency is located over the eastern Gulf of Alaska and western North America in October before moving westward into the east-central Pacific in November where it remains through March. The October blocking pattern in the Atlantic is also quite different from the other months, strongly resembling the autumn average pattern with a strong peak that extends from the cold season blocking location southeast of Greenland eastward over the United Kingdom, western Europe, and Scandinavia. The continental peak begins to weaken in November with the decrease in blocking frequency over western Europe continuing until March. The bimodal pattern of winter and spring blocking, typical of the cold season, with strong peaks in the central Pacific and western Atlantic southeast of Greenland, appears in November and lasts through March. Only the October peaks in both basins, located well to the east of the cold season maxima differ substantially from the seasonal averages. In terms of blocking frequency, the peak is observed in January and February over the central Pacific and in October over the Atlantic.
During the warm season months (AMJJAS), blocking is much weaker than the cold season in every month except August and September. Over the Pacific, a weak peak in blocking frequency is noted along the east coast of Asia and Japan in July, but otherwise no large peaks are noted over the Pacific. The biggest deviation from the seasonal average blocking frequency over the Pacific is observed in August (Fig. 15). In August, a strong peak in blocking is found in the central Pacific north of the Aleutian Islands along the southwest coast of Alaska. The magnitude of this peak, with as many as 18% of days blocked, is comparable to the winter and early spring peaks in the region. By September, the peak has moved into the eastern part of the Gulf of Alaska and western North America just west of the October position where nearly 20% of days are identified as being blocked. Over the Euro-Atlantic sector, the September blocking maximum is similar to October and the autumn seasonal average pattern with a broad area of high blocking frequency stretching from southeast of Greenland into Scandinavia, although the frequency of blocking over the continent is largest in October.
The results demonstrate that the blocking frequency exhibits a characteristic bimodal cold season blocking pattern between November and March (with a third weak continental peak over western Europe in early spring) and a characteristic warm season pattern between April and July, with no strongly preferred regions of blocking occurrence. Strong intraseasonal variability appears in the Pacific between August and October when the blocking frequency increases to levels comparable to the cold season and the peak moves eastward across the basin from the central Pacific into the eastern Gulf of Alaska and finally well over the western North American continent. In the Euro-Atlantic region, a secondary peak over western Europe and Scandinavia appears in September and October that is very similar to the autumn seasonal average.
The large intraseasonal variability in the location and frequency of blocking over the Pacific during the summer months, particularly the large blocking peak in August near the Aleutians, is masked by a seasonal average. Because most studies use seasonal averages, this is the first study to identify this peak in blocking activity. The results suggest that monthly averages should be used when studying warm season blocking over the Pacific.
e. Difference between the mAPV* and PV–θ blocking indices
Our results found a large peak in blocking frequency in the eastern Pacific that are not detected by studies that use the PV–θ index of Pelly and Hoskins 2003a (i.e., Tyrlis and Hoskins 2008a). To demonstrate the flexibility of the mAPV* index, we identified blocking events in the eastern Pacific using both the PV–θ and mAPV* indices and created composites of the 250-hPa PV and PV anomalies for those events detected by both the PV–θ and mAPV* indices (Fig. 16a) and those that were only detected by the mAPV* index (Fig. 16b). Because there were so few events detected by the PV–θ index in this region, the composites contain events from all four seasons.
Composites of the 250-hPa PV (contours; 0.5-PVU contour interval) and 250-hPa PV anomalies (PVU; shading) for blocking events identified by (a) both blocking indices and (b) only the mAPV* index.
Citation: Journal of Climate 27, 8; 10.1175/JCLI-D-13-00374.1
The composite of the events identified by both indices (Fig. 16a) shows a dipole structure in the block with a pronounced reversal of the PV gradient. The composite of events identified only by the mAPV* index has relatively uniform PV values without a strong reversal of the gradient. The composites suggest that the PV–θ index often fails to detect blocking anticyclones in the eastern Pacific that are uniform low PV air masses or highly amplified ridges that do not break, cut off from the westerly flow, and result in a persistent reversal of the potential temperature gradient on the dynamic tropopause. The proposed mAPV* index detects blocking events that are not associated with persistent reversals of the PV gradient.
5. Concluding discussion
This study has proposed a modification to the APV* blocking index of Schwierz et al. (2004) that identifies quasi-stationary patterns in the vertically averaged potential vorticity anomalies. The vertically averaged PV anomaly (mAPV*) is designed to identify the most physically relevant characteristic of a blocking anticyclone, a negative PV anomaly underneath a raised tropopause. Defining the anomaly relative to the zonal mean PV on each level helps ensure that the selected features are near the tropopause and not in the stratosphere. The results presented in this study clearly demonstrate that the proposed index identifies large PV anomalies during blocking episodes in the troposphere that peak near the 200-hPa pressure surface during all seasons and not at 500 hPa. This suggests that blocking indices using the 500-hPa geopotential height do not identify the peak of the PV anomaly in the column, especially in summer when the 500-hPa level is well beneath the level of the jet.
The overall geographical distribution of the blocking event occurrence shows a bimodal frequency structure with maximum blocking frequency in the Euro-Atlantic and Pacific regions downstream of the Asian and North American continents near the location of planetary ridges and the end of the storm tracks with secondary peaks over the continents in the other seasons. The winter and spring climatology are generally consistent with many previous studies (e.g., Rex 1950a,b; Lejenäs and Økland 1983; Croci-Maspoli et al. 2007) with a peak in Pacific blocking near the Aleutian Islands in the central Pacific (150°W) and a peak in Euro-Atlantic blocking southeast of Greenland (10°E). Our results also agree with Croci-Maspoli et al. (2007), who found a large peak in autumn blocking extending over western Europe and Scandinavia, but disagree over the frequency of summer blocking, with the results in this study suggesting that the summer blocking frequency, while lower than the other seasons, is higher than reported by most other studies. Perhaps the most novel result to emerge from this study is the identification of intraseasonal variability in blocking frequency over the Pacific.
The similarity between patterns of monthly blocking frequency and the seasonal average patterns of blocking frequency in winter and spring show that there is very little intraseasonal variability in those seasons over either basin. The winter pattern of blocking (with two strong peaks in the western Atlantic and central Pacific) is similar to the monthly climatology from November through March. Over the Atlantic, the strong peak over Scandinavia and northern Europe only observed in autumn is much stronger in October than the other months. The blocking pattern of the Pacific in August, September, and October (ASO) are also clearly different from the winter and summer patterns. In August, a large peak in blocking frequency is observed north of the Aleutians in the Bering Sea while September and October blocking peak shifts into the eastern Pacific near the coast of western North America. The large peaks in blocking frequency in ASO in the Pacific suggest a link to extratropical transition (ET) events.
For example, while typhoons in the Pacific basin occur all year around, the frequency of events reaches an annual peak in ASO (Jones et al. 2003). Numerous studies of Pacific basin typhoons that recurve and undergo ET over the western North Pacific demonstrate that recurving tropical cyclones often trigger or amplify Rossby wave trains that propagate along the jet stream waveguide (e.g., Harr and Dea 2009; Hodyss and Hendricks 2010) and trigger ridges over the eastern Pacific 4 days later (Archambault et al. 2013).
We hypothesize that the large peak in blocking frequency in the eastern Pacific in September and October can be explained by the high frequency of ET events occurring in the presence of a strong jet stream is more likely to be extended over the Pacific basin during these months (Archambault et al. 2013). We also hypothesize that in August, without a strong jet stream over the Pacific to enhance Rossby wave propagation, ET events might also contribute to blocking by amplifying a downstream anticyclone that breaks over the central Pacific instead of propagating into the eastern Gulf of Alaska before breaking. These hypotheses are currently being tested and will be addressed in future studies.
We also identified large changes in the blocking frequency over time, especially over Siberia in the summer, the central Pacific in spring, and western Europe in autumn. The frequency of blocking was much lower in the period 1979–2012 (4%–6%) than the period 1948–78 (approximately 10%–12%) over Siberia. Given the scarcity of observations over Siberia and general questions about the quality of the reanalysis data before 1958, the discontinuity could be explained by deficiencies in the data. To verify that the observed peak in blocking frequency is not an artifact of the data, we examined reports on the weather and circulation of individual summer months during the 1950s and looked for mentions of blocking activity over eastern Asia. The summer of 1954 was one of the hottest on record over much of the east-central United States (Westcott 2011; Hawkins 1954; Holland 1954; Winston 1954). During much of that summer, a weak anticyclone persisted over the central United States (Holland 1954) that resulted in 22 days with temperatures over 100°F in many places (Westcott 2011). Hawkins (1954) suggested that the anticyclone over the central United States was part of a five-trough pattern in the midlatitude westerlies that set up in late June and persisted through most of the summer. According to Hawkins, the anomalous circulation over North America was part of a planetary-scale circulation pattern that included persistently amplified, warm continental blocking ridges over western Russia and eastern Asia that were displaced to the north of the climatological positions of the oceanic ridges. The author mentions that a similar pattern was observed during the summer of 1953, another year with a high frequency of blocking over Siberia, consistent with the Hovmöller diagram in Fig. 10. An observational study of Siberian summer blocking events and their potential role in the Great Plains drought of the 1950s is needed to verify that the observed peak in blocking activity in the early 1950s is not an artifact of the data.
Acknowledgments
This research has been supported, in part, by a Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery Grant. We thank the National Center for Atmospheric Research (NCAR) for providing the data. The first author would also like to thank Dr. Charles A. Doswell III for an enjoyable conversation that helped to dramatically improve the quality of this manuscript.
REFERENCES
Altenhoff, A. M., O. Martius, M. Croci-Maspoli, C. Schwierz, and H. C. Davies, 2008: Linkage of atmospheric blocks and synoptic-scale Rossby waves: A climatological analysis. Tellus, 60A, 1053–1063.
Archambault, H. M., L. F. Bosart, D. Keyser, and J. M. Cordeira, 2013: A climatological analysis of the extratropical flow response to recurving western North Pacific tropical cyclones. Mon. Wea. Rev.,141, 2325–2346.
Austin, J. F., 1980: The blocking of middle latitude westerly winds by planetary waves. Quart. J. Roy. Meteor. Soc., 106, 327–350.
Barnes, E. A., and D. L. Hartmann, 2010: Influence of eddy-driven jet latitude on North Atlantic jet persistence and blocking frequency in CMIP3 integrations. Geophys. Res. Lett., 37, L23802, doi:10.1029/2010GL045700.
Barriopedro, D., R. GarcÃa-Herrera, A. R. Lupo, and E. Hernández, 2006: A climatology of Northern Hemisphere blocking. J. Climate, 19, 1042–1063.
Croci-Maspoli, M., C. Schwierz, and H. C. Davies, 2007: A multifaceted climatology of atmospheric blocking and its recent linear trend. J. Climate, 20, 633–649.
Diao, Y., J. Li, and D. Luo, 2006: A new blocking index and its application: Blocking action in the Northern Hemisphere. J. Climate, 19, 4819–4839.
Dole, R. M., 1986: Persistent anomalies of the extratropical Northern Hemisphere wintertime circulation: Structure. Mon. Wea. Rev., 114, 178–207.
Dole, R. M., and N. D. Gordon, 1983: Persistent anomalies of the extratropical Northern Hemisphere wintertime circulation: Geographical distribution and regional persistence characteristics. Mon. Wea. Rev., 111, 1567–1586.
Dole, R. M., and Coauthors, 2011: Was there a basis for anticipating the 2010 Russian heat wave? Geophys. Res. Lett., 38, L06702, doi:10.1029/2010GL046582.
Dunn-Sigouin, E., S.-W. Son, and H. Lin, 2013: Evaluation of Northern Hemisphere blocking climatology in the global environment multiscale model. Mon. Wea. Rev.,141, 707–727.
Galarneau, T. J., T. M. Hamill, R. M. Dole, and J. Perlwitz, 2012: A multiscale analysis of the extreme weather events over western Russia and northern Pakistan during July 2010. Mon. Wea. Rev.,140, 1639–1664.
Green, J., 1977: The weather during July 1976: Some dynamical considerations of the drought. Weather,32, 120–126.
Harr, P. A., and J. M. Dea, 2009: Downstream development associated with the extratropical transition of tropical cyclones over the western North Pacific. Mon. Wea. Rev., 137, 1295–1319.
Hawkins, H. F., 1954: The weather and circulation of July 1954. Mon. Wea. Rev.,82, 209–217.
Hodyss, D., and E. Hendricks, 2010: The resonant excitation of baroclinic waves by the divergent circulation of recurving tropical cyclones. J. Atmos. Sci., 67, 3600–3616.
Holland, J. Z., 1954: The weather and circulation of June 1954. Mon. Wea. Rev.,82, 163–171.
Hong, C.-C., H.-H. Hsu, N.-H. Lin, and H. Chiu, 2011: Roles of European blocking and tropical-extratropical interaction in the 2010 Pakistan flooding. Geophys. Res. Lett., 38, L13806, doi:10.1029/2011GL047583.
Horel, J. D., 1985: Persistence of wintertime 500 mb height anomalies over the central Pacific. Mon. Wea. Rev.,113, 2043–2048.
Houze, R. A., Jr., K. L. Rasmussen, S. Medina, S. R. Brodzik, and U. Romatschke, 2011: Anomalous atmospheric events leading to the summer 2010 floods in Pakistan. Bull. Amer. Meteor. Soc.,92, 291–298.
Illari, L., 1984: A diagnostic study of the potential vorticity in a warm blocking anticyclone. J. Atmos. Sci., 41, 3518–3526.
Jones, S. C., and Coauthors, 2003: The extratropical transition of tropical cyclones: Forecast challenges, current understanding, and future directions. Wea. Forecasting, 18, 1052–1092.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc.,77, 437–471.
Lejenäs, H., and H. Økland, 1983: Characteristics of Northern Hemisphere blocking as determined from a long time series of observational data. Tellus, 35A, 350–362.
Lupo, A. R., and P. J. Smith, 1998: The interactions between a midlatitude blocking anticyclone and synoptic-scale cyclones that occurred during the summer season. Mon. Wea. Rev., 126, 502–515.
Masato, G., B. J. Hoskins, and T. J. Woollings, 2012: Wave-breaking characteristics of midlatitude atmospheric blocking. Quart. J. Roy. Meteor. Soc., 138, 1285–1296.
Matsueda, M., 2011: Predictability of Euro-Russian blocking in summer of 2010. Geophys. Res. Lett., 38, L06801, doi:10.1029/2010GL046557.
Mullen, S. L., 1986: The local balance of vorticity and heat for blocking anticyclones in a spectral general circulation model. J. Atmos. Sci., 43, 1406–1441.
Pelly, J. L., and B. J. Hoskins, 2003a: A new perspective on blocking. J. Atmos. Sci., 60, 743–755.
Pelly, J. L., and B. J. Hoskins, 2003b: How well does the ECMWF Ensemble Prediction System predict blocking? Quart. J. Roy. Meteor. Soc., 129, 1683–1702.
Pfahl, S., and H. Wernli, 2012: Quantifying the relevance of atmospheric blocking for co-located temperature extremes in the Northern Hemisphere on (sub-)daily time scales. Geophys. Res. Lett., 39, L12807, doi:10.1029/2012GL052261.
Rex, D. F., 1950a: Blocking action in the middle troposphere and its effect upon regional climate. I. An aerological study of blocking action. Tellus, 2, 196–211.
Rex, D. F., 1950b: Blocking action in the middle troposphere and its effect upon regional climate. II. The climatology of blocking action. Tellus, 2, 275–301.
Schwierz, C., M. Croci-Maspoli, and H. C. Davies, 2004: Perspicacious indicators of atmospheric blocking. Geophys. Res. Lett., 31, L06125, doi:10.1029/2003GL019341.
Shukla, J., and K. C. Mo, 1983: Seasonal and geographical variation of blocking. Mon. Wea. Rev., 111, 388–402.
Tibaldi, S., and F. Molteni, 1990: On the operational predictability of blocking. Tellus, 42A, 343–365.
Tibaldi, S., E. Tosi, A. Navarra, and L. Pedulli, 1994: Northern and Southern Hemisphere seasonal variability of blocking frequency and predictability. Mon. Wea. Rev., 122, 1971–2003.
Tyrlis, E., and B. J. Hoskins, 2008a: Aspects of Northern Hemisphere atmospheric blocking climatology. J. Atmos. Sci., 65, 1638–1652.
Tyrlis, E., and B. J. Hoskins, 2008b: The morphology of Northern Hemisphere blocking. J. Atmos. Sci., 65, 1653–1665.
Westcott, N. E., 2011: The prolonged 1954 midwestern U.S. heat wave: Impacts and responses. Wea. Climate Soc.,3, 165–176.
Winston, J. S., 1954: The weather and circulation of August 1954. Mon. Wea. Rev.,82, 228–236.
Woollings, T., B. Hoskins, M. Blackburn, and P. Berrisford, 2008: A new Rossby wave–breaking interpretation of the North Atlantic Oscillation. J. Atmos. Sci., 65, 609–626.
Woollings, T., A. Hannachi, and B. J. Hoskins, 2010: Variability of the North Atlantic eddy-driven jet stream. Quart. J. Roy. Meteor. Soc., 136, 856–868.
