Effect of mid-latitude blocking anticyclones on the weather of the Arabian Peninsula

Authors

  • H. Athar,

    Corresponding author
    1. Center of Excellence for Climate Change Research/Department of Meteorology, King Abdulaziz University, P. O. Box 80208, Jeddah 21589, Saudi Arabia
    • Center of Excellence for Climate Change Research/Department of Meteorology, King Abdulaziz University, P. O. Box 80208, Jeddah 21589, Saudi Arabia.
    Search for more papers by this author
  • Mansour Almazroui,

    1. Center of Excellence for Climate Change Research/Department of Meteorology, King Abdulaziz University, P. O. Box 80208, Jeddah 21589, Saudi Arabia
    Search for more papers by this author
  • M. Nazrul Islam,

    1. Center of Excellence for Climate Change Research/Department of Meteorology, King Abdulaziz University, P. O. Box 80208, Jeddah 21589, Saudi Arabia
    Search for more papers by this author
  • M. Adnan Abid,

    1. Center of Excellence for Climate Change Research/Department of Meteorology, King Abdulaziz University, P. O. Box 80208, Jeddah 21589, Saudi Arabia
    Search for more papers by this author
  • M. Azhar Ehsan

    1. Center of Excellence for Climate Change Research/Department of Meteorology, King Abdulaziz University, P. O. Box 80208, Jeddah 21589, Saudi Arabia
    Search for more papers by this author

Abstract

The statistical relationships among the various 10°–70°E mid-latitude blocking anticyclone parameters and the weather of the Arabian Peninsula (AP) (35°–60°E, 12°–32°N) over a 40-year period (1968-2007), on seasonal, interannual, decadal and long-term scales, are studied. The studied parameters include the number of blocking anticyclone events, the duration, the intensity, and the longitude at the blocking anticyclone onset. It is found that 31% of the Northern Hemisphere mid-latitude blocking anticyclone events occurred over the 10°–70°E longitudes, and out of these, the maximum number of mid-latitude blocking anticyclone event onsets are at 30°E (24%). On the seasonal basis, the annual and decadal relationships of the 10°–70° blocking anticyclones with the El-Niño Southern Oscillation (ENSO), the North Atlantic Oscillation (NAO) and the Arctic Oscillation (AO) indices are presented. The results show that the number of days the blocking anticyclones persists is sensitive to the ENSO phase. The mid-latitude blocking anticyclone occurrence over the 10°–70°E longitudes is indicative of the reduced surface temperature variance, both upstream and downstream, during the blocking anticyclone period, over the AP. A shift in the mean surface temperature distribution occurs, in all seasons, during the blocking anticyclone period. The blocking anticyclones initiate a surface temperature change (both positive and negative) that persists even after the blocking anticyclone's decay. The AP surface weather during the months of October, November, and December is affected more by the occurrence of mid-latitude blocking anticyclones over the 10°–70°E longitudes in the presence the of El-Niño phase. Copyright © 2012 Royal Meteorological Society

1. Introduction

Climatological relationships provide an assessment of the long-term behavior of various atmospheric variables, such as conditions at the 500 hPa geo-potential height, and temperature at the surface (IPCC, 2007). These relationships serve as guidelines for the weather and for climate projection studies for any given region. A regional climatological relationship study not only contributes towards a better understanding of, ultimately, the global climate, but also provides an informed and analysed view of the local climate for the regional policy makers (IPCC, 2007).

The development of a predominantly mid-tropospheric, meridional circulation pattern within a sector of the Northern Hemisphere (NH) or the Southern Hemisphere is commonly referred to as a mid-latitude blocking anticyclone (e.g., Rex, 1950a, 1950b). This stagnation in the zonal flow gives rise to difficulties in operational weather forecasting for regions within (and near) the blocked region (e.g., Benzi et al., 1986; Watson and Colucci, 2002). Developing an overview of the processes that lead to the formation of such circulation patterns, and the study of their associated implications for surface weather, are thus of significant interest (e.g., Athar and Lupo, 2010).

In the recent past, numerous global and hemispheric climatological studies of atmospheric mid-latitude blocking anticyclone events have been performed using various blocking anticyclone parameters in order to summarize the characteristics of mid-latitude blocking anticyclones (Wiedenmann et al., 2002; Barriopedro et al., 2006). A recent brief review of the mid-latitude blocking anticyclone climatological studies may be found in Barriopedro et al. (2010). Cumulatively, these studies provide a description of the hemispheric/global characteristics of the parameters of mid-latitude blocking anticyclone events, such as the occurrence frequencies (over seasonal to decadal scales), the blocking intensity, and the latitudinal/longitudinal distributions of the blocking anticyclone occurrences, as well as their relationships with large-scale circulation indices, such as the El-Niño Southern Oscillation (ENSO) index.

Other recent atmospheric mid-latitude blocking anticyclone studies based on dynamical concepts, such as using the wave breaking index as a block identification index, are presented by Tyrlis and Hoskins (2008a, 2008b). The mid-latitude blocking anticyclone climatologies based on the extent of negative potential vorticity considerations are also available (Schwierz et al., 2004; Scherrer et al., 2006; Croci-Maspoli et al., 2007a, 2007b).

Existence of a broad frequency peak for mid-latitude blocking anticyclones at 0°–30°E longitudes is indicated in several of the above climatological studies. However, no detailed study has been performed to analyse the surface weather implications associated with this frequency peak. In this article, we consider the mid-latitude blocking anticyclone events occurring over the wide longitudinal window of 10°–70°E in order to study the surface weather implications of these for the Arabian Peninsula (AP), over a 40-year period (1968–2007). The 10°–70°E longitudinal window was also used earlier to study the relationship between atmospheric circulation patterns and surface weather elements in Saudi Arabia (Almazroui, 2006). The latitudinal boundary of the AP is taken as 12°–32°N and longitudinal boundary of the AP is taken as 35°–60°E (Figure 1). The blocking anticyclone parameters are obtained based on the measure of the amplitude of the westerly flow (Tibaldi and Molteni, 1990). The longitudinal boundaries for blocking anticyclone detection are taken as being between 10°E and 70°E (Figure 1) in order to account for the accompanying upstream and downstream impacts of the blocking anticyclone events (Carrera et al., 2004). Thus, this study is an impact study performed to document the details and affects of the mid-latitude blocking anticyclone events on the weather and climate of the AP. In addition, in this article, we present the details of the statistical relationships between the various mid-latitude blocking anticyclone parameters, including the duration, the intensity, the onset longitude, and the frequency, occurring between 10° and 70°E. The intention here is to provide a reference document to build the basis for weather and long-range forecasting, as well as for climate projection studies for the AP. Another aspect of this study is the analysis of the teleconnections of the mid-latitude blocking anticyclones with subtropical and tropical weather and climate.

Figure 1.

Map of the study area. The Arabian Peninsula region (35°–60°E and 12°–32°N) is marked with a solid-line rectangle

The weather of the AP is dominated by arid to semi-arid climatic conditions (Al-Jerash, 1985; Abdullah and Al-Mazroui, 1998; Almazroui, 2011a). Prolonged spells of extreme weather conditions, such as high-intensity rainfall events, droughts and heatwaves, have socio-economic implications for the AP (Almazroui, 2011b). The blocked anticyclonic upper-air flows cause stagnant weather patterns over the AP, and such stagnation may be quantified in terms of reduced variance in surface temperature. The stagnation or reduced variance in the surface atmospheric variables, such as surface temperature, affects several socio-economic factors, particularly agriculture, water resources and power generation.

Thus, the subject of each section is as follows. In Section 2, the data and methodology used to detect the upper-air mid-latitude blocking anticyclones is presented. In Section 3, the details of the climatological relationships of the detected mid-latitude blocking anticyclone parameters are presented, including the decadal trend analysis, for the mid-latitude blocking anticyclone events occurring between 10° and 70°E. In Section 4, the relationship between the number of the mid-latitude blocking anticyclones, the blocked days, and the Niño 3.4, the North Atlantic Oscillation (NAO) and the Arctic Oscillation (AO) indices are presented. In Section 5, the details of some possible statistical linkages between the upper-air mid-latitude blocking anticyclones and the AP surface weather are presented. Section 6 summarizes our results and presents the conclusions.

2. Data and methodology

2.1. Data

The National Center for Environmental Protection/National Center for Atmospheric Research (NCEP/NCAR) gridded re-analysis dataset with a horizontal resolution of 2.5° × 2.5° is used (Kalnay et al., 1996; Kistler et al., 2001). This re-analysis gridded dataset at 500 hPa and at the surface level is used on 6 hourly and daily bases. This dataset was also used by Barriopedro et al., (2006, 2010) to discuss the climatological features of the mid-latitude blocking anticyclones.

2.2. Blocking anticyclone detection definition

The 500 hPa geo-potential height at 1200 UTC is used as a diagnostic atmospheric variable. The Lupo and Smith (1995) and the Wiedenmann et al. (2002) criteria are used for the detection of the mid-latitude blocking anticyclones, and are briefly summarized below.

In the pioneering work of Rex (1950a, 1950b), the climatology of blocking was examined and used the following criteria to detect blocking anticyclones: (1) the basic westerly current must split into two branches, (2) each branch current must transport an appreciable mass, (3) the double-jet system must exceed over at least 45° of longitude, (4) a sharp transition from zonal type flow upstream to meridional type downstream must be observed across current split, and (5) the pattern must persist with recognizable continuity for at least 10 d.

Briefly, the 10°–70°E mid-latitude blocking anticyclones detection criteria used in this study include: (1) satisfying the above-mentioned Rex (1950a, 1950b) criteria for the blocking anticyclone with the minimum duration of the blocking anticyclone as 5 d, (2) a negative or small positive LO 83 index (Lejenäs and Øakland, 1983), must be present on a time-longitude or Hovmöller diagram, where

equation image(1)

The Z represents the 500 hPa geo-potential height at the respective latitude, (3) conditions (1) and (2) must be satisfied for 24 hours after onset to 24 h before termination, (4) the blocking anticyclone should be poleward of 35°N, and the ridge should have an amplitude of greater than 5° latitude [the amplitude is defined as the farthest northward extent of the contour representative of the wave in which the mid-latitude blocking anticyclone is embedded, and a line connecting the upstream and downstream inflection points on the wave (Lupo and Smith, 1995)], (5) the mid-latitude blocking anticyclone onset is described as occurring when condition (4) and either conditions (1) or (2) are satisfied, and (6) the termination is designated as the time that the event fails conditions (1) or (2) for a 24-h period or longer. This procedure is used to detect the mid-latitude blocking anticyclone events at 500 hPa, and defines the blocking anticyclone duration with start- and end-dates (for more details, see, http://weather.missouri.edu/gcc).

The typical latitudinal and longitudinal scales of mid-latitude blocking anticyclones are 40° by 60° (Barriopedro et al., 2010 and references cited therein). Therefore, even though the detected and studied blocking anticyclones are poleward of 35°N, they impact the AP weather, which is south of 35°N (see Section 5 for some details). Importantly, as mentioned in Section 1, the detected blocking anticyclone events upstream of the AP include the all season broad frequency peak located approximately at 0°–30°E (see, for instance, Wiedenmann et al., 2002; Barriopedro et al., 2006).

2.3. Blocking anticyclone intensity definition

The blocking anticyclone intensity (BI) is defined as:

equation image(2)

In Equation 2, Zmax is the maximum 500 hPa geo-potential height in the closed anticyclone region or on a line associated with the ridge, and Z is the subjectively chosen 500 hPa geo-potential height contour encompassing the upstream and downstream troughs. The dimensionless BI measures the amplitude of the flow around the block. Subtraction of 1 from the ratio Zmax/Z, and multiplication by 100 yields BI typically in the range of 1 to 10. For further details and examples, the reader is referred to Lupo and Smith (1995) and Wiedenmann et al. (2002).

The above procedure provides the BI and the longitudinal value of the blocking anticyclone onset for all the detected mid-latitude blocking anticyclone events over the NH/10°–70°E. The blocking anticyclone characteristic parameters thus obtained are the following: the number of blocking anticyclone events, the duration of the blocking anticyclone events, the BI, and the blocking anticyclone onset longitude. In the next section, the spatio-temporal distributions and statistical relationships among these mid-latitude blocking anticyclone parameters are presented and compared with the blocking anticyclone parameters of the entire NH.

3. Climatological characteristics of the mid-latitude blocking anticyclones affecting the weather of the AP

Table I displays the ten most persistent mid-latitude blocking events occurring between 10° and 70°E. The table indicates that the longest events occur during all seasons. The longitude distribution of the events indicates their occurrence both upstream and downstream of the AP, as well as within the AP longitudes. A search for ten longest mid-latitude blocking anticyclone events over the entire NH is also performed, and it is noted that the first five longest mid-latitude blocking anticyclones between 10° and 70°E fall within the ten longest events over the NH (not shown).

Table 1. The top ten most persistent mid-latitude blocking anticyclones occurring between 10° and 70°E
RankEventInitial DateEnd DateLifetime (days)Longitude (°E)
1December 2002262832.510
2July 20039103250
3February 200418183130
4January 1995512720
5October 19871362430
6September 196827182230
7July 197224152140
8January 19831552160
9January 19723232030
10June 19855252040

Table II displays the ten strongest mid-latitude blocking events occurring between 10° and 70°E. Table II indicates that eight out of ten strongest events occurred during the winter season (January–February–March or JFM) and upstream of the AP region. These events fall within the broad frequency peak located at 0°–30°E. The ten strongest events in the NH are also investigated, and it is noted that Event 1 is the tenth strongest event in the NH (not shown). The above comparisons indicate that the longer but less-intense mid-latitude blocking events dominate the 10°–70°E event distribution being considered for the AP weather and climate impacts.

Table 2. The top ten strongest mid-latitude blocking anticyclones occurring between 10° and 70°E
RankEventBILongitude (°E)
1January 20075.9910
2January 19705.8830
3March 19755.4710
4January 19905.4010
5February 19935.3610
6December 19975.3530
7January 19695.3415
8January 19885.2320
9February 19745.0915
10November 19914.9715

3.1. Spatial and temporal distributions

The longitudinal distribution of the 10°–70°E blocking anticyclone events is displayed in Figure 2(a). The maximum number (24%) of the blocking anticyclone events occurred at 30°E, followed by 18% at 60°E. This behavior is in general consistent with previous findings (Wiedenmann et al., 2002). The minimum number of the 10°–70°E blocking anticyclone events occurred at 35°E, 55°E, and 65°E (1%). The longitudinal values of the maximal and the next-to-maximal occurrence lie close to the western and eastern geographical borders of the Kingdom of Saudi Arabia (in the AP), respectively. These two longitudinal values also define the approximate geographical longitudinal boundaries of the AP. Given the longitude dependence of the mid-latitude blocking anticyclone frequency between 10° and 70°E, on average, the flow is more zonal between 30°E and 60°E, compared to between 10°E and 30°E over the AP.

Figure 2.

(a) The longitudinal distribution of the mid-latitude blocking anticyclone event occurrences over 10°–70°E, during the 40-year period under study (1968–2007). (b) The seasonal variability of the relative frequency of the 10°–70°E and the NH blocking anticyclone events for the 40-year period

A total of 341 mid-latitude blocking anticyclone events occurred over the 40-year period, with an annual frequency of 8.5 events, between 10° and 70°E. The total number of blocked days was 3009, which is approximately 21% of the total number of days during the 40-year period, with 365 daily values per year. The average BI per blocking anticyclone event was 2.95. For comparison, in the NH (Lupo et al., 2008), over the same 40-year period (1968–2007), there were 1105 mid-latitude blocking anticyclone events, with 9583 d of blocking anticyclone persistence. This is approximately 65% of the total number of days, with an average blocking anticyclone persistence period of 8.67 d. The annual frequency was 28 events. These numbers are somewhat higher than those reported in some other studies (Barriopedro et al., 2006), reflecting the differences in the methodologies used to identify the duration of the mid-latitude blocking anticyclone persistence, including the relatively higher number of detected mid-latitude blocking anticyclones. The maximum (minimum) BI was 6.42 (0.73) with an average BI of 3.02 per blocking anticyclone event. Thus, the blocking anticyclone events occurring between 10° and 70°E were 31% of the entire mid-latitude NH blocking anticyclones.

Figure 2(b) displays the relative seasonal variability within 10°–70°E and the NH blocking anticyclone events, for the entire study period. During the summer (July–August–September or JAS) and the fall (October–November–December or OND) seasons, the relative occurrence of the 10°–70°E blocking anticyclone events is higher than that of the NH blocking anticyclone events (25.51 and 24.93% relative to 21.36 and 21.90%, respectively). Over 10°–70°E (NH), during the JAS season, there were 87 (236) blocking anticyclone events, with a total of 341 (1105) events. In the same season, over the NH, the total number of days with blocking anticyclones was 2008. During the OND season, over the 10°–70°E (NH), there were 85 (242) blocking anticyclone events. Over the NH, in the OND season, the total number of days with blocking anticyclones was 2025.5. Over the NH, during the winter season, there were 303 blocking anticyclone events, with a total number of days with blocking anticyclones at 2765; during this season, the 10°–70°E blocking anticyclone occurrence is approximately 5% smaller relative to that over the NH. Over the NH, during the spring season (April–May–June or AMJ), there were 324 blocking anticyclone events, with a total number of days with blocking anticyclones at 2784.5; during this season, the 10°–70°E blocking anticyclone occurrence is approximately 2% smaller relative to that over the NH.

Figure 3(a) displays the BI distribution of the 10°–70°E blocking anticyclone events for the entire period of study. The distribution peak falls within the 2 to 3.5 range; values with 0.86 < BI < 5.99. On a scale of 1 to 10 in units of BI, this implies a dominance of moderate (2.0 < BI < 4.3) strength blocking anticyclone events (Lupo and Smith, 1995; Wiedenmann et al., 2002). The distribution is positively skewed, as mean BI is 2.93 and the mode is 2.37.

Figure 3.

(a) The BI distribution of the 10°–70°E blocking anticyclone event occurrences during the 40-year period under study (1968–2007). (b) The 10°–70°E blocking anticyclone lifetime distribution during the 40-year period under study (1968–2007)

Figure 3(b) displays the lifetime distribution of the 10°–70°E blocking anticyclone events, over the 40-year period. As pointed out earlier, the maximum lifetime of the mid-latitude blocking anticyclones over 10°–70°E is 32.5 d. A sharp drop occurs after lifetime duration of 5 d, followed by a long exponential tail. Qualitatively, this distribution compares well with the earlier studies conducted for the NH (Wiedenmann et al., 2002).

Figure 4(a) displays the interannual variability of the mid-latitude blocking anticyclones at 30°E and 60°E. The figure displays a relatively negative trend in the blocking anticyclone events at 30°E in recent years, compared to that at 60°E. The figure also indicates that during 1988, 1993, and 1996, no blocking anticyclone events occurred at 30°E. During the years 1971, 1986, 1987, 1995, 1999, and 2001, no blocking anticyclone events occurred at 60°E. The maximum number of blocking anticyclone events occurred at 30°E (60°E) during 1970 (1980) and 1974 (1984).

Figure 4.

(a) The time series for the 10°–70°E blocking anticyclone events at 30E and 60°E for the period 1968–2007. The 10°–70°E blocking anticyclone occurrences peak at 30°E and 60°E (Figure 2(a)). (b) The seasonal variability of the 10°–70°E blocking anticyclone events at 30°E and 60°E for the period 1968–2007

Figure 4(b) displays the seasonal variability of the 10°–70°E blocking anticyclone events at 30°E and 60°E. In each season, the occurrence of the 10°–70°E blocking anticyclone events is predominantly at 30°E, with the largest relative difference in the JAS season.

3.2. Long-term and decadal trends

Figure 5 displays the 40-year linear trend of the 10°–70°E and the NH blocking anticyclone events. The standard deviation value for the 10°–70°E blocking events is 2.32, whereas for the NH, it is 6.80. The figure indicates that the coefficient of determination (R2) value is higher for the NH. Overall, the NH blocking anticyclone events display a more positive trend compared to the 10°–70°E blocking anticyclone events (3.3 blocking events per decade, as compared to 0.50 blocking events per decade, although the latter value is statistically less robust). The linear trends for both the 10°–70°E and the NH were calculated on seasonal basis also, for the entire period of study (not shown). Statistically, the R2 value for the NH is higher compared to that for the 10°–70°E.

Figure 5.

The trend analysis of the 10°–70°E and the NH blocking anticyclone event occurrences during the 40-year period under study (1968–2007). The linear trend is shown by the continuous solid line

Variability in the occurrence of the blocking anticyclone events may be present on the decadal scale, and Table III displays the decadal trend of the 10°–70°E and the NH blocking anticyclone event parameters during four consecutive decades (1968–1977, 1978–1987, 1988–1997, and 1998–2007). Table III is suggestive of the observation that over both the 10°–70°E and the NH, while the mean BI per decade has a decreasing trend, the mean lifetime of the blocking anticyclones (per decade) has an increasing trend over both the 10°–70°E and the NH. Note also the occurrence of relatively less intense but prolonged 10°E–70°E blocking anticyclone events, relative to the NH (for the period of study on a decadal basis).

Table 3. The decadal trend analysis of the 10°–70°E and the NH blocking anticyclone occurrences over the 40-year period (1968–2007). The first column gives the decadal period of study. The pair in the second column, labeled 10–70 and NH, gives the percentage of blocked days during the corresponding decade, for 10°–70°E and NH, relative to the total number of days the blocking anticyclone persisted during the 40-year period, respectively. The pair in the third column gives the mean BI per decade. The fourth column pair entries give the mean blocking anticyclone lifetime (in days) per decade. Note the relative opposite trend in the last two pairs of columns between the 10° and 70°E entries
 Total blocking days in decades (%)Mean BI per decade (days)Mean lifetime per decade
Years10–70NH10–70NH10–70NH
1968–197722.3721.503.243.207.717.91
1978–198724.6320.333.033.129.008.45
1988–199720.3121.862.913.138.468.13
1998–200732.7036.312.632.759.919.55

Over the 10°–70°E, for the decade 1968–1977, the maximum duration of the blocking anticyclone event is 22 d, with a total of 673 d of blocking anticyclones. For the decade 1978–1987, the maximum duration of the blocking anticyclone events is 24 d, with 741 d of blocking anticyclones. For the decade 1988–1997, the maximum duration of the blocking anticyclone events is 27 d, with 611 d of blocking anticyclones. For the decade 1998–2007, the maximum duration of the blocking anticyclone events is 32.5 d, with 984 d of blocking anticyclones.

4. Relationship of blocking anticyclones with the Niño 3.4, NAO, and AO indices

In this section, first the interannual and decadal variability of the Niño 3.4, NAO, and AO indices vis-à-vis the number of blocking anticyclones occurring over the 10°–70°E is assessed on the seasonal basis. Following this, the seasonal relationship between the blocked days and the Niño 3.4 index is discussed. The Niño 3.4 monthly index dataset was obtained from the University Corporation for Atmospheric Research (UCAR) website (http://www.cgd.ucar.edu/cas/catalog/climind/). For more details, see Reynolds and Smith (1994), Trenberth (1997), and Hurrell and Trenberth (1999). The NAO monthly index dataset was obtained from the Climate Prediction Center (CPC) at the NOAA website (http://www.cpc.noaa.gov/products/precip/CWlink/pna/nao.shtml). The AO monthly index dataset was obtained from the CPC at the NOAA website (http://www.cpc.noaa.gov/products/precip/CWlink/daily_ao_index/ao.shtml).

Several recent studies have analysed the climatological relationships between the hemispheric/global blocking anticyclones and the large-scale circulation indices based on different blocking anticyclone detection methods and various statistical measures to quantify these relationships (Wiedenmann et al., 2002; Barriopedro et al., 2006; Croci-Maspoli et al., 2007a, 2007b).

First, the given monthly indices were converted into the seasonal means, and then the correlation coefficients (CC) between these means and the number of blocking anticyclone events per season occurring over the 10°–70°E were calculated. The CC values of the indices with the blocking anticyclone events over the 10°–70°E are displayed in Figure 6(a) on the seasonal basis for the period of study. The seasonal analysis indicates that during the AMJ season, the AO index has a positive and statistically significant correlation (at 95% CL). This implies that during the AMJ season, the occurrence of mid-latitude blocking anticyclones over 10°–70°E is more frequent during the positive phase of AO.

Figure 6.

(a) The seasonal CC of the Niño 3.4, the NAO, and the AO indices with the 10°–70°E blocking anticyclones for the 40-year period. (b) The relative number of blocked days, over 10°–70°E, in the four seasons during the La-Niña, the El-Niño and the neutral events, for the 40-year period

The decadal correlations were calculated for the 10°–70°E blocking anticyclone events with the Niño 3.4, NAO, and AO indices on the seasonal basis, and are displayed in Table IV. During the 1968–1977 decade, in the OND season, the positive phase CC value of the 10°–70°E blocking anticyclone events with the Niño 3.4 is statistically significant (at 95% CL). The mid-latitude blocking anticyclone occurrence over 10°–70°E is more frequent during the warm phase of El-Niño. The suggested mechanisms for this correlation may include (1) Rossby wave propagation, (2) instability of the horizontally and vertically sheared zonal flows, and (3) modification of the Hadley Circulation (Shukla, 1986). There is also a statistically significant (at 95% CL) positive relationship between the 10°–70°E blocking anticyclone events and the NAO index in the JAS season during the 1988–1997 decade. During the 1988–1997 (1998–2007) decades, for the AMJ (JAS) season, a statistically significant positive correlation (at 95% CL) exists between the AO index and number of 10°–70°E mid-latitude blocking anticyclones. These observations may have some implications for the inter-seasonal forecasting.

Table 4. The seasonal CC of the 10°–70°E blocking anticyclone events with the Niño 3.4/NAO/AO indices on the decadal scale. The asterisk (*) indicates that the quoted CC numbers are significant at 95% CL
Year/seasonJFMAMJJASOND
1968–1977− 0.09/− 0.09/− 0.020.22/0.47/0.350.43/− 0.38/− 0.530.70* /− 0.10/− 0.20
1978–19870.17/− 0.03/− 0.440.34/0.42/0.04− 0.35/− 0.32/− 0.230.28/0.11/0.22
1988–19970.43/0.51/0.580.26/0.32/0.67*0.07/0.64* /0.290.41/− 0.19/− 0.46
1998–20070.22/0.22/0.48− 0.08/− 0.12/0.550.39/0.38/0.67*0.06/− 0.08/0.01

To assess the relationship between the number of blocked days and the phases of the ENSO cycle, the seasons were classified into the La-Niña, the El-Niño, and the neutral phases, based on the definition provided at the Niño 3.4 website used earlier. The seasonal mean from the monthly ENSO anomaly data was constructed for the 40-year period. If the anomaly value is greater (less) than 0.5 °C (−0.5 °C) in a given season, then the season is classified as an El-Niño (La-Niña) season, otherwise it is classified as a neutral season. During the entire period of study, the El-Niño years are: 1968, 1969, 1972, 1976, 1977, 1982, 1986, 1987, 1991, 1994, 1997, 2002, 2004, and 2007, whereas the La-Niña years are: 1970, 1971, 1973, 1974, 1975, 1984, 1988, 1995, 1998, 1999, 2000, and 2007; the remaining years being the neutral ones.

The ratio of number of blocked days during each classified season and the total number of blocked days belonging to that season/phase is calculated and is displayed in Figure 6(b). A relatively dominant role of the warm El-Niño phase favoring the maintenance of the blocking anticyclones over 10°–70°E is evident, particularly during the OND season. These results are in general agreement with those reported and summarized by Wiedenmann et al. 2002.

5. Effect of mid-latitude blocking anticyclones on the weather of the AP

5.1. Characteristics of blocking anticyclone sample

A sample of 30 events was selected from the constructed catalogue of the mid-latitude blocking anticyclones occurring over 10°–70°E in order to exemplify the effects of mid-latitude blocking anticyclones throughout all seasons and all decades over the AP. The selected events occurred both upstream and downstream as well as within the AP longitudes (Table V). The lifetime (BI) of these events ranged from 16 d (1.77) to 32.5 d (4.79), with an average lifetime of 19.75 d (3.15). No overlap of the blocking anticyclone events was found in this sample (event 23 and 24 have different onset longitudes; see Table V). For each of the 30 events, the weather impact before, during, and after blocking anticyclone occurrence is studied to assess the relative impact of mid-latitude blocking anticyclone occurrences on the weather of the AP.

Table 5. Description of the selected mid-latitude blocking anticyclone events between 10°E and 70°E used to study their effects on the weather of the AP. For each event, the lifetime (days), the initial and end date, the longitude at onset (°E), the BI, the month of onset along with the year are displayed
Event no.Lifetime (days)Initial dateEnd dateLongitude (°E)BIMonthYear
116824303.5221968
2222718302.7391968
317217303.48121968
4162715353.4531969
516217602.4591969
620323304.7311972
7212415403.0771972
817239453.55121979
9161430602.4961980
1017.5124452.4061982
1121155603.2411983
12161026654.05121984
1319827303.8031985
1420525402.9261985
15171431203.5251986
1617184502.9931987
1724136303.78101987
1817421602.1371988
19172411253.2251992
202751203.9011995
21171128104.7922002
2232.52628104.30122002
23172512602.0162003
2432910501.7772003
25311818303.4622004
2615.51025602.9652004
2718.52715302.9182004
28183018101.8772006
29182715202.6952007
3019.52413202.30102007

An all-season analysis was performed to obtain the 500 hPa geo-potential height and the surface temperature distributions, before, during, and after the blocking anticyclone events. The results are displayed only for the OND season, because of the more prominent impact of the mid-latitude blocking anticyclone events during this season, on AP. An inter-seasonal comparison is also presented, where applicable.

Figure 7 displays the horizontal distribution of the 500 hPa geo-potential height (m) for a representative event (event no. 16) from Table V during the three phases (before, during, and after) of the blocking. The blocking onset longitude is at 50°E on 18th March 1987 (defined as t). During the first 6 d (defined as t and t + 6) of the blocking, the gradual formation of the 500 hPa split flow is noticeable with the center at 50°E, acquiring maximal amplitude during the mature phase of the blocking (defined as t + 12). This is followed by the subsequent decay of amplitude, simultaneously accompanied by the formation of an upstream low, during the final phase of the blocking (defined as tt). The mid-latitude 500 hPa flow distribution is more zonal after the blocking (defined as tt + 4).

Figure 7.

The horizontal distribution of the 500 hPa geo-potential height (m) during the various phases of a representative event (event no. 16) from Table V. Row-wise, from top to bottom, the displayed distributions are for: 4 d before the blocking (t − 4), during the blocking (t, t + 6, t + 12, tt), and after the blocking (tt + 4)

5.2. Composite geo-potential height distribution

The composite 500 hPa geo-potential height anomalies are displayed in Figure 8, for the OND season, as an example. The three panels (from left to right) correspond to the before-blocking, during-blocking, and after-blocking periods. A 40-year (1968–2007) climatological seasonal mean was subtracted from the geo-potential height value of each time-averaged blocking anticyclone event. The before- and after-blocking periods are averaged over 19 d, whereas the during-blocking period is averaged over the blocking duration. The period of 19 d is taken as the period prior to and following the occurrence of the blocking anticyclone, since this time period is approximately the average lifetime of the selected sample of the blocking anticyclone events (Carrera et al., 2004).

Figure 8.

The composite of 500 hPa geo-potential height anomalies (m) for the 10°–70°E sample blocking anticyclone events (a) before blocking, (b) during blocking, and (c) after blocking in the OND season. The AP region is also marked

The OND season is characterized by relatively more negative 500 hPa geo-potential height anomalies over the AP during and after the occurrence of the blocking anticyclone events, as compared to before the blocking anticyclone events. For comparison, during the JFM season, the same height anomaly of ± 5 m persists over most of the AP region, before and during the blocking period (not shown). During the AMJ and the OND seasons, the relative variation in the anomalies is more prominent than in the JFM and the JAS seasons (not shown). These are also reflected in statistically significant shifts in the surface temperature distribution during the three durations for these two later seasons (see below).

5.3. Composite surface temperature distribution

In order to quantify the relationship between the mid-latitude blocking anticyclone occurrences over the 10°–70°E and the AP surface weather, several aspects of the surface temperature distributions over the AP are examined.

Figure 9 displays the composite surface temperature anomalies ( °C) during the three periods in the OND season, as an example. The before-blocking and after-blocking differences are statistically the largest in this season. For all the seasons, the absolute AP area-averaged temperature differences among the three periods were calculated. On average, during the JFM and AMJ seasons, the absolute AP area-averaged temperature rise is 1.61 °C once the blocking sets in, relative to the before-blocking period (and is statistically significant at 19% CL), whereas relative to the during-blocking period, the after-blocking period temperature rise is 1.24 °C; statistically significant at 24% CL (not shown). The before-blocking and after-blocking results are significant at 99% CL. Overall, this comparison is indicative of an area-averaged absolute AP surface temperature rise of almost a 1 °C, once the blocking occurs, during the JFM and the AMJ seasons (not shown). A rise in the AP temperature associated with the occurrence of mid-latitude blocking anticyclones may have several socio-economic implications (such as those mentioned briefly in Section 1).

Figure 9.

The OND season composite of surface temperature anomalies ( °C) during the same three time periods for the selected blocking anticyclones sample, as in Figure 8

In particular, for the OND season, during the before- and during-blocking periods, on average, the maximum absolute AP area-averaged temperature change occurs. The AP area-averaged absolute temperature in the before-blocking (during-blocking) period was 23.87 °C (21.26 °C), for the OND season. Thus, the AP area-averaged absolute temperature drop is about 2.61 °C, once the blocking sets in, during the OND season. Similarly, the absolute AP area-averaged temperature changes were estimated for other seasons.

The reduced surface temperature variance is quantified in terms of the mean square difference (ΔTsfc) in the AP area-averaged consecutive-day surface temperature. The lower, middle and upper quantiles of the ΔTsfc were calculated for each of the three durations to determine if extremes in surface temperature are more likely to occur during the blocking anticyclone event in any particular quantile range. Another recent approach to analyse the extremes in weather associated with mid-latitude blocking anticyclone events can be found in Buehler et al. (2010).

Figure 10 displays the log–log relationship among the three ΔTsfc quantiles for the three periods (before, during, and after the blocking anticyclone occurrence) for all the seasons. The sample ΔTsfc lower and middle quantile values tend to have somewhat less spread in the before-during and the during-after periods (the first two columns in Figure 10) around the mid-line, as compared to the upper quantile values of the ΔTsfc in the before-to-after period of blocking.

Figure 10.

The log-log scatter plots of ΔTsfc quartiles for the selected sample of the blocking anticyclone events occurring over 10°–70°E. The diagonal line in each panel indicates 1:1 correspondence. For each quartile, there are three panels, from left to right, row-wise

The student's t-test was used to assess the statistical significance of the all-season mean of each of ΔTsfc quantile for the three durations. When averaged over all seasons and all longitudes between 10°–70°E, the ΔTsfc lower quantile mean difference between the before- and the during-blocking anticyclone occurrence is statistically significant only at 26% CL, whereas the same CL is at 10% for the during and after durations. The before to after difference is statistically significant at 76% CL. The stagnation in surface temperature settles in once the blocking is initiated. Similar analysis was carried out for the ΔTsfc middle and upper quantiles. Progressively, statistically less robust differences exist for the ΔTsfc middle and upper quantiles during the three durations, respectively (≥12% CL), indicating that the ΔTsfc tend to fall dominantly in the lower quantile range. Depending upon the CL, in each instance, a difference in the AP area-averaged surface temperature exists once the mid-latitude blocking anticyclone sets in over the 10°–70°E.

Seasonal stratification indicates that during the OND season, the mean for ΔTsfc in the middle quantile is relatively statistically more significant, as compared to that in other seasons. Also, once the stagnation in the surface temperature is established, due to the 10°–70°E mid-latitude blocking anticyclone occurrence, it persists even after the blocking anticyclone decays away, at least for the duration we have investigated. The variation of number of days before and after the blocking was also studied for the above analysis. Similar results for the comparison of statistically significant quantiles for ΔTsfc were obtained.

6. Summary and conclusions

The obtained results may be summarized in two parts. First, we present the climatological features of the mid-latitude blocking anticyclone events occurring over 10°–70°E, and then present the summary of the effects of these mid-latitude blocking anticyclone events on the weather of the AP.

The statistical relationships among the following parameters for the mid-latitude atmospheric blocking anticyclones, over 10°–70°E, for a 40-year period (1968–2007), are presented in terms of: (1) the number of the blocking anticyclone events, (2) the duration of the blocking anticyclone events, (3) the longitudinal value of the blocking anticyclone at onset, and (4) the blocking anticyclone intensity. Comparison with the similar 40-year-long blocking anticyclone climatology for the NH is discussed to provide the context for this present study.

Based on the blocking anticyclone detection criteria used in this study, the main climatological characteristics of the 10°–70°E mid-latitude blocking anticyclones during 1968–2007 include:

  • (1) The maximal blocking anticyclone occurrence at 30° is followed by occurrence at 60°E, (2) there is a positively skewed and moderate BI distribution, with an average BI per event of 2.95, and (3) the total number of days for which blocking anticyclones persisted over the 10°–70°E corresponds to 31% of the total number of blocking anticyclone days in the NH.
  • There is a simultaneous decreasing (increasing) trend in BI per decade (lifetime span per decade), in the 10°–70°E blocking anticyclone events, during the study period, similar to that in NH. On average, on the decadal scale, there is higher life cycle duration in the 10°–70°E blocking anticyclone events, compared to the ones in the NH.
  • There is a statistically significant CC value (at 95% CL) of 0.70 between the Niño 3.4 index and the number of mid-latitude blocking anticyclones over the 10°–70°E during the OND season in the 1968–1977 decade. During the 1988–1997 decade, in the JAS season, the CC is found to be significant (at 95% CL) with a value of 0.64 between the 10°–70°E blocking anticyclones and the NAO index. During the 1988–1997 (1998–2007) decade, and in the AMJ (JAS) season, a statistically significant correlation (at 95% CL) was found between the 10°–70°E blocking anticyclone events and the AO index, with the same value of 0.67.
  • The number of days for which the blocking anticyclones persisted over the 10°–70°E is sensitive to the ENSO phase. In all seasons, the relative contribution of the El-Niño phase biases the number of days that the blocking anticyclones persist. This contribution is more prominent during the OND season.

In summary, using the NCEP-NCAR re-analysis gridded data, the climatological relationships of the 10°–70°E mid-latitude blocking anticyclone parameters are presented (during the latter half of the 20th Century) to provide an analysed view of the characteristics of the AP weather impacts resulting from the mid-latitude blocking anticyclones (lasting for 5 d or more). The results are presented on seasonal, annual, and decadal bases. The seasonal analysis reveals some differences in the characteristics of the 10°–70°E blocking anticyclones compared to the NH ones. These include the relatively high occurrence of the blocking anticyclones during the summer (JAS) (25.51%, relative to 21.36%), and during the fall (OND) (24.93%, relative to 21.90%) seasons.

A longer lifetime span per decade of the 10°–70°E blocking anticyclones, relative to that in the NH, is noticed during the last three decades (1978–2007). During the first decade (1968–1977), the lifetime span per decade of the 10°–70°E blocking anticyclones is smaller by ⩽3%, relative to the NH ones. The first three longest duration mid-latitude blocking anticyclone events occurred during the 1998–2007 decade.

An all-season sample of 30 blocking events was selected to represent the mid-latitude blocking anticyclones occurring over the 10°–70°E longitudes for the AP weather impact study. The weather impacts of the mid-latitude blocking anticyclone events include a shift in the AP area-averaged surface temperature. The mean AP area-averaged absolute surface temperature rises by 1.61 °C during the JFM and the AMJ seasons, once the blocking sets in, over the 10°–70°E; after the blocking dismantles, the absolute temperature further rises by 1.24 °C. On the other hand, during the JAS and the OND seasons, a mean absolute surface temperature drop of 1.47 °C per season occurs; the AP area-averaged absolute temperature further drops by 1.04 °C per season once the blocking anticyclone dissipates. These findings are supported by the positive and negative 500 hPa anomalies in the three periods, over the AP, respectively.

The mid-latitude blocking anticyclones initiate stagnation in the surface temperature throughout the blocking period. Averaged over all seasons, the consecutive-day surface temperature difference over the AP tends to fall in the lower quantile range during the mid-latitude blocking anticyclone occurrences.

Ancillary