Ural Blocking and the Amplitude of Wintertime Cold Surges over North China Detected by a Cooling Algorithm

Abstract

A new algorithm is proposed to estimate the cooling amplitude over North China induced by each Ural blocking event. Taking full account of potential transient temperature rises during the cooling process and the lag time of cooling relative to the blocking, this algorithm provides more detailed analysis which should not be possible by using former methods. The amplitude of the Ural blocking-related cooling events is found to have a broad distribution. Further, although most Ural blocking events lead to severe cold surges over North China, the number of Ural blocking events that do not induce significant cooling over North China cannot be ignored. The possible reasons for the wide range in cooling amplitude are explored in terms of the lifetimes and geographical centers of the blocking highs, the circulation patterns preceding the onset of the cooling events, and the snowfall associated with cooling events. Larger amplitude cooling events occur in Ural blocking highs that have longer lifetimes and northwestward displacements of their geographical centers. The northward displacement of a Ural blocking center favors the advection of extremely cold air from the Arctic regions, which accumulates in West Siberia and subsequently gives rise to the most severe cold surges over North China. The lack of activities of cold surges before the blocking-related cooling events not only amplifies the cooling amplitude directly, but also increases the occurrence probabilities of snowfalls through its modulation on the local specific humidity. The increased albedo and subsequent snow-melt induced cooling further amplify the cooling amplitude.

1. Introduction

East Asia is a region subject to frequent wintertime cold surges [1,2,3,4,5,6,7,8]. The atmospheric blocking events over the Ural Mountains (Ural blocking), the occurrence frequency of which is relatively high, was identified as one of the primary causes for severe cold surges over North China [9,10,11,12,13,14,15,16,17]. The mechanism by which Ural blocking induces strong temperature anomalies over North China is primarily through the enhancement of the cold advection downstream of the blocking itself [18,19]. Park et al. [20] compared the impacts of blocking-activated and wave-train-activated cold surges over East Asia (including North China) during the winter months from 1954 to 2005, and found that the most severe cold surges over East Asia were associated with blocking events. Huang et al. [15] identified the six most severe cold surges over East Asia (including North China) during the winter months of 1979–2014, and revealed that these cold surges were all accompanied by blocking events over the Ural Mountains and surrounding regions.

However, the occurrence of Ural blocking is not an effective precursor signal for the remarkable cooling over North China for the following two reasons: First, the primary analysis methods of the above two studies first identified the severe cold surges and then composited the circulation patterns before and/or during these cold surges. Second, the cooling amplitude over North China associated with each specific Ural blocking event (and, therefore, the distribution of the cooling amplitude for all the Ural blocking events) has not been fully defined. Knowledge of the fraction of the Ural blocking events that lead to significant cooling over North China is essential for providing early warnings for probabilities of severe cold surges. This study is devoted to exploring the relationship between Ural blocking events and their related cooling amplitude over North China.

Therefore, it is a key issue to derive the cooling amplitude over North China associated with Ural blocking events. In general, the magnitude of composited temperature anomaly over a region is regarded as the cooling amplitude. Meanwhile, some research calculated the difference in temperature between the end and start dates of events [21]. However, it cannot be ignored that there could exist a warming process among the composited events or during the cooling events. Given this situation, this study proposes an algorithm to identify the cooling amplitude associated with the Ural blocking. The algorithm takes into account not only the warming process, but also the lag time of a cooling process relative to the Ural blocking event. This algorithm can be easily applied to other regions, even to derive warming amplitude based on a similar framework. However, it cannot be applied directly, but needs to be adjusted in parameters (e.g., lag time and the criteria of warming process; see below for details) to suit different situations.

This paper is organized as follows: sec2 introduces the two reanalysis datasets used and an algorithm for estimating the Ural blocking-related cooling amplitude over North China. sec3 presents the cooling amplitude associated with all observed blocking events during the reanalysis period and identifies the conditions responsible for the different cooling amplitude including both the circulation patterns related to the blocking high itself and the meteorological conditions over North China before and during the cooling events. The major findings of this study are summarized in sec4.

2. Materials and Methods

2.1. Data

This study is focused on the winter months of North China. The winter months of a specified year refer to the December of that year and the January, February, and March of the next year (DJFM). The primary data used for analysis is the Japanese 55-Year Reanalysis product (JRA55) [22]. The JRA55 operational system used a reduced Gaussian T319 grid with a horizontal resolution of about 60 km and a 4D-variational (4D-Var) data assimilation scheme. In particular, the reanalysis assimilated the snow depth over Russia and Mongolia from 1958 onward and over China from 1970 onward [23], which improves the accuracy of the circulation and temperature anomalies over North China and the surrounding regions. The variables taken from JRA55 for this paper’s analysis include 6-hourly geopotential heights, surface temperatures, surface-specific humidities, and snowfall rates (water equivalent) for the period from December 1958 to March 2022. The daily means for these four variables are calculated for the analysis in this study. To support the JRA55-based analysis, we also employ the reanalysis dataset from the National Centers for Environmental Prediction and the National Center for Atmospheric Research (NCEP-NCAR) running from December 1948 to March 2022 [24]. The variables from this reanalysis dataset include the daily geopotential height and surface temperature.

2.2. A Method for Estimating the Cooling Amplitude for a Blocking Event

Following Barriopedro et al. [25], the wintertime blocking events are tracked from the two different reanalysis datasets over their respective time periods. The Ural blocking events are identified as those blocking events having a center within 40–80Ŷ E. Once identified, the start date and end date for each blocking event are recorded.

The method for calculating the cooling amplitude associated with each blocking event is based on the principle of finding all pairs of neighboring local maximum and minimum. The algorithm includes the following steps:

1.

For each spatial grid point, the surface temperature anomaly for each calendar day is estimated by removing the long-term average of the surface temperature on that calendar date during 1958–2022 (for JRA55). The area-averaged land surface temperature anomalies (LSTA) over North China (90–150Ŷ E and 30–50Ŷ N) are estimated for each calendar day during 1958–2022 (for JRA55);

2.

For each blocking event, the start and end dates for the analysis period are determined based on the start and end dates of that blocking event and the start date of the next blocking event (Figure 1a). The start date for the analysis period is the start date of the particular blocking event. When another blocking event occurs within ten days of the end date of a particular blocking event, the end of the analysis period of the event is the end date of the blocking event under analysis, otherwise the analysis period ends five days after the end date of the event;

3.

All local maximum and minimum values of the LSTA time series within the analysis period of a blocking event are identified (Figure 1b). Note that, in this step, the LSTA on the start (end) date of the analysis period cannot be considered a minimum (maximum). This exclusion is to ensure that all identified extrema can be partitioned into chronologically ordered pairs, beginning with local maximum and followed by local minimum. The end dates of each pair of extrema can be viewed as the dates remarking cooling events;

4.

The cooling amplitude for each cooling event is defined as the difference between the LSTA extremum (Figure 1c);

5.

Two neighboring cooling events are merged together when the increase in LSTA is less than 2 K (Figure 1d). As a result of this step, several major cooling events are recognized;

6.

The cooling amplitude of each major event is calculated as the difference between the maximum and minimum LSTA during the corresponding analysis period (Figure 1d). The dates for the corresponding maximum and minimum are recorded;

7.

The major event with maximum cooling amplitude of all the major events is viewed as the cooling event associated with a specific blocking event (Figure 1d). The start and end dates of the associated cooling event are the dates for the corresponding major cooling events saved during the sixth step. Note that when the LSTA monotonically increases with time over the analysis period, the cooling amplitude is assigned to 0 K.

3. Results and Discussion

3.1. Distributions of Cooling Amplitude Associated with Blocking Events

Figure 2 displays the interannual variation in the number of blocking events and blocking days during the winter months for the two reanalysis datasets. In total, the number of blocking events for JRA55 and NCEP-NCAR are 76 (1958–2021) and 98 (1948–2021), with the accumulated number of blocking days being 643 and 843 days, respectively. Considering just the overlap period (1958–2021) of these two reanalyses, the total number of blocking events become 76 and 84, with the associated number of blocking days over this period being 643 and 736 days. Both the number of blocking events and the accumulated number of blocking days constructed based on JRA55 are highly correlated with those based on NCEP-NCAR (R = 0.87, R = 0.95, respectively) over the overlap period. The similar number of blocking events and blocking days and high correlations during the overlap period confirm the applicability of the JRA55 product as the major reanalysis data for identifying blocking events and the use of NCEP-NCAR reanalysis as supporting data.

Using the algorithm proposed in Section 2.2, the cooling amplitude over North China associated with each blocking event is estimated for the two reanalysis data products and presented in Figure 3. It is clear that the estimated cooling amplitude has a rather wide range in all two reanalysis products. Specifically, the cooling ranges for JRA55 and NCEP-NCAR are 0.22∼12.57 and 0.15∼13.74 K, respectively. The number of blocking events with cooling amplitude less than 2 K for JRA55 and NCEP-NCAR are 19 (accounting for 25%) and 18 (18.37%), respectively, indicating that the occurrence of Ural blocking does not necessarily lead to significant cooling over North China. In addition to having more Ural blocking events, the analysis results from NCEP-NCAR tend to have a larger cooling amplitude over North China associated with the blocking events it identifies. As reported by Liu et al. [26] and Tang et al. [27], the recent increased incidence of wintertime cold extremes over the Northern Hemisphere can be linked to the rapid retreat of Arctic sea ice through the altered mid-latitude circulation patterns and the associated increased occurrence of blocking highs. Here, we estimate the average cooling amplitude for two periods; 1958–1999, prior to the retreat of sea ice, and 2000–2021, while the rapid retreat in Arctic sea ice is occurring. The average cooling amplitudes during 1958–1999 are 3.63 K for JRA55, and 4.10 K for NCEP-NCAR, while those during the rapid retreat period are 4.04 K for JRA55, and 4.79 K for NCEP-NCAR. Therefore, the amplitude of the cooling events has slightly increased during the period with rapid reduction in sea ice relative to the period prior to sea ice retreat. This suggests that the melting of sea ice may be also having an impact on the circulation patterns associated with the Ural blocking events, which leads to an increase in the intensity of the cold extremes over North China.

For each reanalysis, we divide the blocking events into three groups according to their cooling amplitude. The `strong’ and `weak’ groups include blocking events with cooling amplitude in the upper and lower quartile of all events, respectively, with the remaining half of the events assigned to the `medium’ group. The lower (upper) quartile of cooling amplitude for blocking events based on the JRA55 product is 1.83 K (5.12 K), while that based on the NCEP-NCAR product is 2.30 K (5.51 K). According to these criteria, the number of blocking events belonging to the weak, medium, and strong groups, respectively, are 19, 38, and 19 for the JRA55 product, and 24, 50, and 24 for the NCEP-NCAR product. Within each group, there are significant anticyclonic blocking anomalies anchored over the Ural Mountains and cyclonic blocking anomalies located over the region south of Lake Baikal (Figure S1). The average lifetimes of the blocking events in the JRA55 product for the weak, medium, and strong groups are 6.79, 8.39 and 10.26 days, respectively (Figure 4a). The longer duration of the blocking events in the strong group is more conducive to persistent cold air intrusion from high latitudes over North China, which explains the larger amplitude of the cooling events within this group. This result is also seen in the NCEP-NCAR (Figure 4b). In order to see if the location of the blocking events plays a role in the cold intrusion over North China, the average geographical center for each blocking event is also calculated following the algorithm in Barriopedro et al. [25]. In addition, the composite geographical center for each group is computed by averaging the geometric centers of the individual events within each group. In both reanalysis products, the weak group blocking events tend to be located to the southeast of the strong group blocking events (Figure 4c,d). Since the Ural blocking is embedded within the westerlies, a westward displacement of its blocking center favors a longer residence over the Ural Mountains, leading to the longer lifetimes of the strong group blocking events. As cold air intrusion occurs downstream of the blocking center, a northward displacement of the Ural blocking center brings colder air originating from higher latitudes into North China, which explains the largest amplitude of the strong group cooling events.

3.2. Early Signals for Blocking Events with Different Cooling Amplitude

We composite the geopotential height anomalies on the 500 hPa surface for the three different groups three days prior to the onset of the associated cooling event. In all three groups, `Omega-type’ blocking high-related circulation anomalies, characterized by an anticyclonic anomaly over the Ural Mountains surrounded by two cyclonic anomalies to its east and west, are evident (Figure 5a–c). Further, wave train-like circulation can be found around the mid-high latitudes over the Northern Hemisphere (Figure S2). The center of the Ural anticyclonic geopotential height anomaly for the strong group is farther north than that for the weak group, which is consistent with the composited geographical centers for the blocking events belonging to each group (Figure 4c,d). Moreover, the horizontal distributions of the cyclonic geopotential height anomalies downstream of the blocking highs for the three groups are rather different. Within the weak group, the cyclonic geopotential height anomalies are mainly located to the south of Lake Baikal, extending from east of the Caspian Sea to Japan. Further, the cyclonic geopotential height anomalies over East Asia indicate that the East Asian trough is enhanced during weak group events, which is conducive to the activation of cold surges over North China before the onset of the blocking-related cooling events [28,29,30]. By contrast, during strong group events, the cyclonic geopotential height anomalies are located in central Siberia and have much stronger amplitude. According to the thermal wind relationship, both the northward displacement and the amplification of the cyclonic anomaly strengthen the northerly winds downstream of the anticyclonic anomaly. In addition, during strong group events, no significant cyclonic geopotential height anomaly is observed over East Asia, indicating that there are no cold surges over North China prior to the onset of the blocking-related cooling events.

We further composite the surface temperature anomalies three days prior to the onset of the subsequent cooling event. For all three cooling event groups, there are significant warm temperature anomalies over Arctic Ocean, which is centered at the Barents and Kara Seas. Meanwhile, there are significant cold temperature anomalies over Siberia. The temperature contrast is known as the Warm Arctic–Cold Siberia (WACS) or Warm Arctic–Cold Eurasia (WACE) pattern [31], which is closely related to the upstream warm and downstream cold temperature advection of the Ural blocking high [32,33,34,35,36]. For weak and medium groups, the warm temperature anomalies can extend into the coastal areas of the Eurasian continent. By contrast, in the strong group, due to the stronger northerly winds associated with the colder advection downstream of the blocking high, these warm anomalies are confined to the Arctic areas. As a consequence, for strong group events, there are significant cold surface temperature anomalies with amplitude exceeding 7 K centered over West Siberia. While there are also cold temperature anomalies downstream of the blocking high for weak group, these cold anomalies are relatively small (less than 4 K), and located farther south over parts of South Siberia, Mongolia, and North China. It should be noted that there are significant warm anomalies over North China under the strong group. Therefore, prior to the onset of blocking-related cooling events, there is an accumulation of cold air in the northwest of North China for all three groups, with the accumulated cold air in the strong group being the coldest. Further, the warm temperature anomaly over North China prior to the onset of blocking-related cooling events tends to amplify the associated cooling amplitude for the strong group events. It should be noted that the behavior of geopotential height anomalies and surface temperature anomalies for the three event groups is also present in the NCEP-NCAR reanalysis (Figure S3).

The time sequences of the composite geopotential height and its anomaly over the Eurasian continent for all three groups for a period 12 days prior to the onset of the cooling event to 10 days after onset are presented in Figures S4–S6. The precursor signals for the anticyclonic anomaly in the three groups can be traced back prior to the onset of the cooling event associated with a blocking event to 10–12 days for the weak group and 4–6 days for the medium and strong groups. Therefore, it is speculated that the larger amplitude blocking-related cooling events tend to have lower potential predictabilities.

In addition, the time sequences of the composite geopotential height and its anomaly for blocking events associated with all three groups are also analyzed (Figures S7–S9). For all three groups, there are evident pressure ridges and blocking-related anticyclonic anomalies centered over the Ural Mountains. The anticyclonic anomalies begin to dissipate about 6, 10 and 14 days after the start date of the blocking events in weak, medium, and strong groups, respectively, which is consistent with the difference in residence time revealed in Section 3.1 (Figure 4a,b). The precursor signals for the anticyclonic anomaly can be observed about 8 days prior to the onset of the blocking events for the weak group and 4 days for the medium and strong groups. Moreover, the precursor signals are located further north in the medium and strong groups, coinciding with the northward displacement of the Ural blocking center documented in Section 3.1 (Figure 4c,d). As mentioned above, it can be speculated that long-lived Ural blocking event, which is associated with larger amplitude cooling events over North China, tends to have a further north original region, but a lower potential predictability.

3.3. Relation of Snowfall to the Amplitude of Cooling Events

Figure 6 displays the horizontal distributions of the composite mean snowfall rate observed during blocking-related cooling events for the three different groups. The cooling related to the melting of snow combined with the high albedo of snow cover causes considerable ground cooling [37,38,39,40], which further amplifies the cooling of the blocking-related cooling event. Because there is a remarkable increase in the snowfall rate over North China during strong group cooling events relative to weak group cooling events, the effect of snowfall is to further amplify the cooling during strong group cooling events. The increase in specific humidity that leads to increased snowfall during the strong group events can be explained as follows: First, according to the Clausius–Clapeyron equation [41,42], the warm temperature anomalies (Figure 5f) over North China lead to a significant increase in specific humidity there (Figure S10). Second, the lack of cold surge activity over North China prior to the onset of the cooling event during a strong group event corresponds to weakened northerly winds over the coastal areas over East Asia, which is conducive to the import of water vapor from the coastal oceans into North China. The increase in the specific humidity prior to the onset of the blocking-related cooling event and the following cooling event together cause the considerable snowfall in strong group events.

4. Conclusions

A new algorithm for calculating the cooling amplitude of the Ural blocking-related cooling events over North China is proposed to examine the relationship between Ural blocking and the amplitude of the subsequent cooling events over North China. This algorithm is superior to former methods in giving a more detailed analysis. According to the proposed algorithm, and using the JRA55 reanalysis product, about 25% of the Ural blocking events are associated with cooling amplitude of less than 2 K. This result is consistent with the result from NCEP-NCAR (18.37%). To examine the reasons for the different cooling amplitude, the blocking events are divided into three groups depending on the strength of the cooling. It is determined that strong group events (upper quartile) are associated with blocking centers located more to the northwest of the blocking centers belonging to the weak group events. The northwestward displacement of blocking centers favors longer maintenance of the blocking highs and the advection of colder air from higher latitudes that accumulates over the Eurasian continent. The downstream circulation patterns prior to the onset of the blocking-related cooling events are rather different between the weak and strong groups. During weak group, negative geopotential height anomalies at 500 hPa, associated with early cold surge activity, can be observed over North China. By contrast, during strong group there are warm temperature anomalies over North China. The severe contrast between warm temperature anomalies over North China and cold temperature anomalies accumulating in the northwest of North China substantially contributes to the maximum cooling amplitude associated with strong group events. The lack of cold surge activity preceding the cooling event is conducive to an increase in the specific humidity, which leads to more snowfall during strong group events. The subsequent snow melt cooling process and increased albedo feedback to further intensify the cooling amplitude of the strong group blocking-related cooling events. The proposed algorithm for estimating the cooling amplitude associated with Ural blocking events is not specific to this region and can easily be applied to calculating the cooling amplitude associated with blocking events over other regions. This makes the algorithm not only beneficial for estimating the climate impacts of atmospheric blocking events, but also the triggering mechanisms for blocking-related cooling events.