Influence of Okhotsk Sea blocking on summer precipitation over South Korea

This study examined the influence of atmospheric blocking on the variability of precipitation over South Korea during summer (June–July–August) by defining the blocking frequency over the Okhotsk Sea (Okhotsk Sea blocking frequency; OK_BF). According to composite analysis for the years of high precipitation over South Korea, blocking occurs more frequently over the Okhotsk Sea (140°E–160°E). Partial correlation and regression analyses were conducted to separate the contribution of OK_BF to precipitation variability from that of the low‐level meridional wind (LLMW) because LLMW over the region is another important aspect of summer precipitation in South Korea. The barotropic structures of positive geopotential height anomalies over the Okhotsk Sea associated with increasing OK_BF can induce negative temperature anomalies over South Korea due mainly to the equatorward advection of cold air masses from higher latitudes. The enhanced meridional temperature gradient can cause increases in baroclinic instability and zonal wind vertical shear according to the thermal wind balance. This instability can induce anomalous cyclonic circulations over South Korea, resulting in positive precipitation anomalies. The partial correlation coefficients (R2 = 0.35–0.40) between the OK_BF and precipitation indices, including mean precipitation, extreme precipitation intensity, wet days, and consecutive wet days, were all statistically significant at the 95% confidence level. Overall, the effects of the increasing OK_BF on both precipitation and potential evapotranspiration can intensify the surface water budget in South Korea.

South Korea has generally been conducted in the frame of EASM (e.g., Ha et al., 2012;Lee and Seo, 2013;Park et al., 2015). The intensity of the EASM is generally modulated by the variability of the western North Pacific subtropical high (WNPSH). According to Chang et al. (2000aChang et al. ( , 2000b), the enhanced low-level jet associated with the westward extension of the WNPSH provides higher moisture transport into East Asia, resulting in increased precipitation over the region. On the other hand, unlike other monsoon systems in the tropics-subtropics, the EASM is a subtropical-mid-latitude synopticscale circulation system. The combined influences from mid-and low-latitudes complicate the understanding of the EASM (Wang et al., 2008;Ha et al., 2012). Therefore, many researchers have investigated the relationship between interannual variability of the EASM and various factors, such as El Niño-Southern Oscillation (e.g., Wang et al., 2000;Yun et al., 2010;Wu et al., 2012;Kim and Kug, 2018), Arctic Oscillation (e.g., Gong and Ho, 2003;Gong et al., 2011), North Atlantic Oscillation (NAO) or North Atlantic sea surface temperature (NASST) (e.g., Wu et al., 2009Wu et al., , 2012Zuo et al., 2013;Zheng et al., 2016), and tropical Atlantic sea surface temperature (TASST) (e.g., Choi and Ahn, 2019).
Blocking is a synoptic-scale quasistationary atmospheric phenomenon that is associated with the formation of strong stationary anticyclones at high latitudes, resulting in weakening (strengthening) zonal (meridional) circulation (Rex, 1950a(Rex, , 1950b. This synoptic phenomenon is usually related to extreme weather events, such as coldwaves (Buehler et al., 2011), heatwaves (Barriopedro et al., 2011), heavy rain (Grams et al., 2014), and drought (García-Herrera et al., 2007), depending on where the blocking is located. Although the frequency and intensity of the summer blocking activities are lower than in winter, the specific role of those systems on the mid-latitude climate has been studied widely. Sousa et al. (2017) reported that the south of the blocked regions (the region of the blocks) exhibits positive (negative) precipitation anomalies. A dipole in the precipitation anomalies under blocking action was confirmed for many regions, including Europe (Sousa et al., 2016) and South America (Rodrigues and Woollings, 2017;Fernandes and Rodrigues, 2018). These studies suggested that the blocking phenomenon is associated directly with the variability of precipitation.
The formation and maintenance of blocking anticyclones are not completely understood because of their nonlinear characteristics, such as internal dynamics, orographic effects, and baroclinic transient eddies (Shutts, 1983;Mullen, 1987;Nakamura and Wallace, 1993;Nakamura and Fukamachi, 2004). Nevertheless, considerable efforts have been made to understand summer blocking. Previous studies reported the relationship between summer blocking and land-sea thermal contrast. Arai and Kimoto (2005) showed that increased meridional surface temperature gradient between the Arctic Sea and Eurasia enhances the polar frontal jet, which is a waveguide of quasi-stationary Rossby waves. Nakamura and Fukamachi (2004) suggested that Rossby wave breaking leads to blocking formation over the Sea of Okhotsk (hereafter, Okhotsk Sea). Several studies reported that the positive phase of summer Northern Hemisphere annular mode could account for much of the anomalous weather associated with the double jet stream structure (polar and subtropical jet streams) and blocking high (Ogi et al., 2005;Tachibana et al., 2010). He et al. (2018) reported that the land-sea thermal contrast and summer blocking form positive feedback, and both show an increasing trend. They suggested that the land-sea thermal forcing can induce a zonal-mean westerly and double jet stream, which are favourable conditions for more frequent blocking.
The Okhotsk high (OKH) is a cold high near the surface that is generally accompanied by a blocking anticyclone in the middle and upper troposphere. The occurrence of the OKH brings cool and wet air masses to East Asia (e.g., Ninomiya and Mizuno, 1985). According to Nakamura and Fukamachi (2004), the formation of the OKH in July is related to the eastward propagation of a quasi-stationary Rossby wave. Many studies reported that the anomalous OKH is related to the Atlantic-Europe-Asia (AEA) teleconnection pattern (e.g., Wu et al., 2009Wu et al., , 2012Zuo et al., 2012Zuo et al., , 2013Zheng et al., 2016). These teleconnection patterns start mainly from the subtropical North Atlantic and propagate eastward to North Europe, the Urals, Central Eurasia, and the Okhotsk Sea. The AEA teleconnection pattern has been associated with the NAO, NASST, and TASST (Zuo et al., 2012(Zuo et al., , 2013Zheng et al., 2016;Choi and Ahn, 2019). Meanwhile, some studies emphasized that the Pacific-Japan (PJ) or East Asia-Pacific (EAP) teleconnection patterns, which have a meridional tripole structure consisting of WNPSH, BCM front, and OKH, are closely related to summer precipitation in East Asia (Chen and Zhai, 2015;Wang et al., 2018). According to previous studies, convective heating in the western North Pacific can induce an anomalous cyclone (anticyclone) over East Asia (Okhotsk Sea) by generating a meridional teleconnection pattern (Nitta, 1987;Yim et al., 2008).
The blocking anticyclone appearing in the Okhotsk Sea (Hereafter, Okhotsk Sea blocking, OKB) can also play an important role in influencing the EASM. Park and Ahn (2014) reported that summertime (June-July-August) blocking is the primary mode in the Okhotsk Sea according to principal component analysis using the 500 hPa geopotential heights. They showed that the OKB frequencies have a distinct negative (positive) correlation with summer temperature (precipitation) in East Asia. Their results suggest that the characteristics of the summer climate over East Asia are closely associated with the occurrence of the OKB. On the other hand, most studies focused on the relationship between the OKB and precipitation over East Asia and specific regions (Japan and China), rather than just South Korea (Wang, 1992;Nakamura and Fukamachi, 2004;Park and Ahn, 2014;Chen and Zhai, 2014b).
Precipitation deficits have a direct impact on the occurrence of drought, which causes significant damage to various aspects of agriculture, society, and the economy. According to previous studies, extreme droughts in South Korea occurred in 1927Korea occurred in -1930Korea occurred in , 1938Korea occurred in -1945Korea occurred in , 1951Korea occurred in -1952Korea occurred in , 1967Korea occurred in -1969Korea occurred in , 1994Korea occurred in -1996Korea occurred in , and 2013Korea occurred in -2015, caused mainly by successive shortages of summer rainfall for multiple years Zhang and Zhou, 2015;Kwon et al., 2016). In an attempt to understand drought in South Korea better, many studies have investigated the characteristics, such as spatiotemporal variation (Min et al., 2003;Azam et al., 2018;Kwon et al., 2019), the relationship with preceding climate conditions (Choi et al., 2009, and future projection (Im et al., 2012;Choi et al., 2016;Lee et al., 2019). Furthermore, potential evapotranspiration (PET), as a function of temperature, is another climate factor impacting drought conditions (Easterling et al., 2007;Im et al., 2012). Therefore, it is necessary to study the effects of blocking on water budget components, including precipitation and PET, which can help understand the flood and drought mechanisms over this region. Focusing on South Korea and the vicinity, this study examined the detailed relationship between the OKB frequencies and both precipitation and PET in the region.
As mentioned earlier, the EASM is closely related to warm and moist air advection at the low-level from the low-latitude regions to the mid-latitude regions. These are favourable conditions for the generation of convective instability. In contrast, the atmospheric blocking activities are accompanied by baroclinic instability (Wang, 1992;Nakamura and Fukamachi, 2004). Therefore, the role of OKB on baroclinic instability is discussed by separating it from convective instability.
Based on the above issues, this paper is organized as follows. Section 2 describes the data, definition of the indices, and statistical method. Section 3 describes the influence of the occurrence of OKB on the water budget components, including precipitation and PET, over South Korea. A discussion and summary are given in Sections 4 and 5, respectively.
2 | DATA AND METHOD

| Data
The data used in this study were the monthly mean enhanced precipitation reanalysis data provided by the Climate Prediction Center Merged Analysis of Precipitation (CMAP) with a horizontal resolution of 2.5 × 2.5 (Xie and Arkin, 1997). The daily and monthly mean atmospheric fields were obtained from the National Centers for Environmental Prediction (NCEP) Department of Energy Reanalysis 2 dataset (hereafter, NCEP2) with a horizontal resolution of 2.5 × 2.5 (but 1.875 × 1.875 for 10-m horizontal wind) (Kanamitsu et al., 2002). The monthly mean temperature and precipitation data from 60 in-situ stations of the Korea Meteorological Administration were also used to support the results derived from the reanalysis data. The period analysed in this study was from 1979 to 2020. Figure 1a represents the topography and location of weather observation stations over South Korea. The topography data used in this study were ASTER Global Digital Elevation Model Version 3 (ASTGTM) (Abrams et al., 2020). The CMAP precipitation data were interpolated onto an NCEP2 grid point to facilitate a comparison with the NCEP2 data. The variability of CMAP precipitation was highly consistent with that of the precipitation averaged over 60 weather stations located in South Korea (Figure 1b). The northeastern Asian summer monsoon index (NEASMI) was defined by the U850 (25 N-35 N, 110 E-150 E) (NEASM_SI) minus U850 (45 N-55 N, 110 E-150 E) (NEASM_NI), where U850 represents the zonal wind at 850 hPa (Park et al., 2018).

| Definition of the Eady growth rate
The Eady growth rate (Eady, 1949;Lindzen and Farrell, 1980;Hoskins and Valdes, 1990) was used to measure the baroclinic instability. This parameter is related directly to the vertical wind shear (or meridional thermal gradient due to thermal wind balance) and static stability. The maximum Eady growth rate is defined as follows: where f is the Coriolis parameter, N is the Brunt-Väisälä frequency, u is the horizontal wind vector, and z is the vertical height.

| Definition of the PET
The surplus and deficit are the basic concepts of the water balance equation. Based on this concept, the standardized precipitation evapotranspiration index (Vicente-Serrano et al., 2010), which was comprised of precipitation and PET, is a widely used index to identify drought conditions over South Korea (Kim et al., , 2013Sohn et al., 2013;Nam et al., 2015). The monthly PET was calculated using the method proposed by Thornthwaite (1948): where PET is in mm Á month − 1 , and T is the monthly mean air temperature ( C). K is the correction coefficient calculated as a function of the latitude and month (Vicente-Serrano et al., 2010): where NDM is the number of days in the month. N is the maximum sunshine duration in hours and was calculated as follows: where ω S is the hourly angle of the sun rising; φ is the latitude; δ is the solar declination; J is the average Julian day of the month. I is the heat index, which was calculated from the monthly heat index (i j ): m is a constant that given by a third-order polynomial with respect to I: m =6:75 ×10 − 7 I 3 − 7:71× 10 − 5 I 2 +1:79× 10 − 2 I +0:4924

| Definition of the blocking index
The blocking events were defined according to Barriopedro et al. (2006). This index, which is based on the index proposed by Tibaldi and Molteni (1990), was used most widely in previous studies (You and Ahn, 2012;Park and Ahn, 2014;Choi and Ahn, 2017;Lee and Ahn, 2017). The northern and southern 500 hPa geopotential height gradient (GHGN and GHGS, respectively) were calculated at each longitude as follows: Topography (shading; unit: m) and location of the observation stations (black dots) over South Korea. (b) Correlation maps between two precipitation datasets during the period of 1979-2020 (JJA); one is CMAP precipitation, and another is precipitation averaged over 60 weather stations located in South Korea. The opened (closed) circles in (b) indicate that the correlation coefficients are significant at the 95% (99%) confidence level [Colour figure can be viewed at wileyonlinelibrary.com] ϕ 0 = 60:0 N + Δ ϕ S =40:0 N +Δ where Δ= −5:0 , −2:5 ,0:0 ,2:5 , or 5:0 : Z(λ,ϕ) is the 500 hPa geopotential height at longitude (λ) and latitude (ϕ). As the five different latitudes were applied, the respective five results of GHGN(λ) and GHGS(λ) were obtained. A given longitude was considered to be blocked when both GHGN(λ) and GHGS(λ) satisfied the following condition in at least one of the five different latitudes.
One persistent blocking event of each longitude was defined as occurring when these conditions lasted for at least five consecutive days. The blocking frequency (BF) was defined as the ratio of the days in which persistent blocking events occur (i.e., blocking days) to the total number of days in summer (92 days).
Because blocking exhibits a significant longitudinal extension, previous efforts to quantify blocking frequency have commonly applied an extension criterion. For example, Barriopedro et al. (2006) applied the following: 1. At least five consecutive blocked longitudes (12.5 ) 2. A nonblocked longitude between the two blocked longitudes is considered a blocked longitude.
However, we did not apply the above extension criterion to the blocking definition for two reasons. First, the BFs decreased when the extension criterion was applied ( Figure S1a). Second, application of the extension criterion had little impact on the interannual variability in blocking frequency, particularly in regions of prominent blocking activity ( Figure S1b).

| Statistical analysis
This study utilized several statistical analyses, including composite, correlation (or equivalently regression), and partial correlation (regression) analysis, to examine the relationship between the OKB frequencies and precipitation over South Korea. The correlation method applied was the linear Pearson correlation (R). The independent associations between the above-mentioned two phenomena were quantified using partial correlation analysis. Partial correlation calculates the correlation between the two variables while eliminating the influence of another variable. Many studies have selected the same analysis to distinguish the impact of one factor from that of the other (e.g., As-syakur et al., 2014;He and Zhu, 2015;Choi and Ahn, 2019). The partial correlation coefficient R xy,z between variable x and y without the effect from z was calculated as: where R xy , R xz , and R yz represent the correlation coefficients between x and y, x and z, and y and z, respectively. The two-sided Student's t test, a parametric test, and Mann-Whitney U test (Mann and Whitney, 1947), a nonparametric test, were used to assess the statistical significance of the differences between two composites (null hypothesis of equal means). The two-sided Student's t test was used to evaluate the confidence level of the correlation (null hypothesis is that the two variables are independent). In this paper, p values <.05 were considered statistically significant.
3 | RESULT 3.1 | Blocking regions associated with summer precipitation over South Korea Figure 2 represents the climatological summertime BF over the Northern Hemisphere along the longitude. The four major regions with high BFs were found in Northern Europe (NE, 0 E-50 E), Ural region (UR, 50 E-80 E), Okhotsk Sea (OK, 120 E-160 E), and Northeastern Pacific (NP, 160 W-130 W). This result is consistent with that reported elsewhere (e.g., Park and Ahn, 2014;Antokhina et al., 2016). Composite analysis was conducted to specify the blocking regions associated with precipitation over South Korea (Figure 3). The PREC_SK was defined by averaging the precipitation obtained from 60 in-situ stations. The high (low) PREC_SK years were defined as values higher (lower) than 1 SD of the time series from 1979 to 2020 (Table 1). The BFs over the region of the UR (50 E-70 E) and OK (140 E-160 E) were higher during the high PREC_SK years than during the low PREC_SK years. In the present study, the UR_BF (OK_BF) was defined as the area-averaged BFs over the region of 50 E-70 E (140 E-160 E). The composite difference of UR_BF was 2.36%, which is not significant at the 95% confidence level (from two-sided Student's t test and Mann-Whitney U test, the p values were .18 and .16, respectively). The composite difference of OK_BF was 5.66%, which is significant at the 95% confidence level (from two-sided Student's t test and Mann-Whitney U test, the p values were .02 and .03, respectively). These results are in line with previous studies reporting that persistent extreme precipitation in Central-Eastern China is closely associated with double-blocking high patterns, which have positive geopotential height anomaly around the Ural Mountains and around the Okhotsk Sea Zhai, 2014a, 2014b). As mentioned in the introduction, a stationary Rossby wave train from northern Europe can induce the development of a Ural high and OKH. The direct effects of blocking on precipitation over South Korea were examined by focusing only on the blocking events in the Okhotsk Sea. This study focused on the relationship between the OK_BF and precipitation in South Korea over interannual timescales. The evolution mechanism of OKB is beyond our scope and this issue will be discussed in more detail in the discussion.
3.2 | The interannual variation of OK_BF and its association with precipitation over South Korea Figure 4 presents the time series of OK_BF. The climatological value of OK_BF was 5.65%. The OK_BF exhibited a long-term trend embedded in the interannual variation. The linear trend was 0.04% year −1 , which is not significant at the 95% confidence level. These results agree with He et al. (2018). They reported that summer blocking increases with increasing summer land-sea thermal contrast, representing the zonal asymmetric thermal forcing changes under global warming.
Composite analysis for the high and low PREC_SK years was also conducted separately to examine the effect of the duration of OK_BF ( Figure 5). The composite differences of OK_BF for five duration groups (i.e., 5, 6, 7, 8, and more than 9 days) were 0.26, 1.24, 1.57, 0.55, and 2.04%, respectively. The BFs were all higher during the high PREC_SK years than during the low PREC_SK years, regardless of the duration, but these are not significant at the 95% confidence level based on both two-sided F I G U R E 3 Composite of the blocking frequency for high and low PREC_SK years based on the ±1 SD of the total time series . The dashed line denotes that the climatological values. The closed (opened) circles indicate the longitudes where the composite differences of the blocking frequency are significant at the 90% confidence level with a two-sided Student's t test (Mann-Whitney U test) ) 1987, 1998, 2002, 2003, 2006, 2011, 2020 1982, 1988, 1992, 1994, 2015, 2016 Student's t test and Mann-Whitney U test. This insignificant difference might be due to the low OK_BF value according to the duration (i.e., few blocking events), indicating that further analysis, such as identical analysis using long-term data or case study, is required. Correlation analysis was conducted to examine the relationship between OK_BF and precipitation over South Korea during summer. The PREC_NEA was defined by the area averaging the precipitation over northeastern Asia (NEA; 30 N-50 N, 110 E-145 E) (Lee et al., 2005). The correlation coefficient between OK_BF and PREC_NEA was 0.40, which is significant at the 95% confidence level. Similarly, NEASM_NI, which is the northern component of NEASMI, was closely related to OK_BF with correlations of −0.41, which is significant at the 99% confidence level (the correlation coefficients between OK_BF and NEASMI and between OK_BF and NEASM_SI were 0.28 and 0.14, respectively). On the other hand, OK_BF was not significantly correlated with precipitation over South Korea ( Figure S2a). This result was also revealed by analysing the in-situ precipitation data. The correlation coefficient between OK_BF and station-averaged precipitation (i.e., PREC_SK) was only 0.27, which is insignificant at the 95% confidence level. Nevertheless, significant positive correlations were observed at 10 stations ( Figure S2b). These results agree with those obtained by Park and Ahn (2014), who reported a low correlation coefficient of 0.15 between OK_BF and PREC_SK (figure 3 and table 6 in Park and Ahn, 2014). They attributed this insignificant relationship to the locality of precipitation over South Korea. The above results suggest that OK_BF is associated with the precipitation over South Korea, but it is unclear if this relationship is robust and how OK_BF contributes to precipitation variability because of their low correlation coefficients in most regions.
We conjectured that other factors strongly modulated the variability of precipitation over South Korea, which are distinct from OK_BF. Figure 6a shows the long-term (1979-2020, JJA) means of the meridional wind in the lower troposphere (850 hPa) and geopotential height in the middle troposphere (500 hPa). In summer, the high-pressure system is located in the subtropical western North Pacific (socalled WNPSH). The poleward wind is dominant in East Asia, which is located on the western flank of the WNPSH (5,880 m isoline in Figure 6a). As mentioned in the introduction, the moisture transported from low latitudes to East Asia is a general consensus on the critical role of the EASM. In line with this issue, many studies on Korean summer precipitation focused on the impact of northward moisture transport across the southern boundaries of South Korea (e.g., Hwang and Lee, 1993;Ha et al., 2003;Park et al., 2003;Shin and Lee, 2005;Baek et al., 2017;Kim et al., 2017). Although a nonlinear relationship between the WNPSH and precipitation over South Korea was reported  (Yeo et al., 2020), the link between the southerly wind and precipitation over South Korea is evident (figure 3 in Yeo et al., 2020). Low-level meridional wind (LLMW) over South Korea has a significant positive correlation with the PREC_SK, which is significant at the 99% confidence level (Figure 6b). The LLMW index was defined by area averaging the 850-hPa meridional wind over South [125][126][127][128][129][130]. The correlation coefficient between the OK_BF and LLMW was −0.13, indicating that both variations are almost independent statistically. These results confirmed that both variations were related to the precipitation variability over South Korea with different mechanisms. Thus, partial correlation and regression analyses were performed to separate the contribution of the OK_BF to precipitation variability from that of the LLMW. F I G U R E 8 Partial regression maps of the (a) 850 hPa air temperature (shading; unit: K), horizontal wind (vectors; unit: m Á s − 1 ), (b) temperature advection (shading; unit: deg Á s −1 ×10 − 6 ), (c) meridional temperature gradient (shading; unit: deg Á m −1 ×10 −6 ), and (d) precipitation (shading; unit: mm Á day − 1 ) anomalies onto the normalized OK_BF after excluding the effects of the LLMW. The opened (closed) circles indicate that the correlation coefficients are significant at the 90% (95%) confidence level. Only the vectors in (a) statistically significant at the 90% confidence level are shown [Colour figure can be viewed at wileyonlinelibrary.com] barotropic positive geopotential height anomalies were observed over the Okhotsk Sea (Figure 7a-c). Blocking is often associated with a high-pressure system on the poleward side and low-pressure systems on the equatorward side (Barriopedro et al., 2006), which exhibits a significant meridional component. Negative geopotential height anomalies over South Korea were evident from the lower to upper troposphere because the trough axis crosses Korean Peninsula (Figure 7a-c). The anomalous anticyclone over the Okhotsk Sea can help maintain the anomalous cyclone over South Korea. Partial regression analyses of air temperature, horizontal temperature advection, and meridional temperature gradient in the lower troposphere (850 hPa) were also conducted to obtain evidence to support this hypothesis (Figure 8a-c). The negative temperature anomalies were confirmed over regions, including Korean Peninsula and Japan (Figure 8a). This is due mainly to the equatorward advection of cold air masses from higher latitudes (Figure 8a, b). In particular, the enhanced meridional temperature gradient, which is one of components of the horizontal temperature advection, was clearly confirmed over the regions mentioned above (Figure 8c). At the same regions, significant positive precipitation anomalies were clearly evident (Figure 8d). The temperature advection of cold air, which is associated with the enhanced meridional temperature gradient, can accelerate the upward motion by the strengthened baroclinic instability. These relationships are discussed in more detail below.

| Influence of frequent occurrence of OKB on precipitation and potential evapotranspiration over South Korea
The vertical structure of air temperature, meridional temperature gradient, and vertical and zonal winds near South Korea associated with increasing OK_BF were F I G U R E 9 Meridional-vertical cross-section averaged along 125 E-130 E for (a) meridional temperature gradient (shading; unit: deg Á m −1 × 10 −6 ), air temperature (contour; unit: K), (b) omega (shading; unit: Pa Á s −1 × 10 −2 ) and zonal wind (contour; unit: m Á s −1 ) anomalies partially regressed upon the normalized OK_BF after excluding the effects of the LLMW. Hatching indicates that the correlation coefficients are significant at the 95% confidence level. The black-line boxes indicate the location of South    (Figure 9). The vertically uniform negative (positive) temperature anomalies were confirmed at the northern regions of South Korea (between 35 N and 55 N), indicating the enhanced meridional temperature gradient over South Korea (Figure 9a). The enhanced meridional temperature gradient can induce baroclinic instability, leading to anomalous cyclonic circulations and ascending motion. This enhanced meridional temperature gradient strengthens the zonal wind vertical shear according to the thermal wind balance (Figure 9b). These results correspond to the increased baroclinicity near the 40 N latitude (Figure 10a). The strong upperlevel subtropical jet is generally located along the belt of 40 N latitude. This subtropical jet was strengthened further by the reinforced meridional temperature gradient induced by the blocking anticyclone. The positive (negative) zonal wind anomalies around latitude 75 N (55 N) were formed directly by the blocking anticyclone. Overall, the enhanced double jet stream, which has a positive zonal wind anomalies around the polar (approximately 75 N) and subtropical (approximately 40 N) jet streams, was present ( Figure 10b). As mentioned in previous studies (Ogi et al., 2005;He et al., 2018), the enhanced double jet structure may inversely lead to a more frequent blocking appearance. These results show that the frequent occurrence of OKB can increase baroclinic instability over South Korea, resulting in cyclonic circulations and increased precipitation over the same regions. Figure 11 presents partial correlation coefficients between OK_BF and precipitation indices after excluding the effects of the LLMW. This study analysed several statistical measures of daily precipitation, such as the mean precipitation (MEAN), extreme precipitation intensity (P90), wet days (WD), and consecutive wet days (CWD). The averaged intensity of the upper-10th percentile daily precipitation per year, P90, was considered. The WD was defined as daily precipitation higher than 1.0 mm, and the CWD was defined as a maximum number of consecutive wet days during summer. The OK_BF was closely related to the variability of the wet days and extreme precipitation as well as mean precipitation over most regions in South Korea (Figure 11). Regarding the mean and extreme precipitation, significant correlations with the OK_BF were found over the northern regions, eastern coastal regions, and southern regions in South Korea (Figures 11a-b). As Park and Ahn (2014) mentioned, the insignificant correlations in the central regions are due mainly to characteristics of regional precipitation. The Taebaek Mountains with high topography (mostly above 1,000 m) are located along the eastern coastal regions in South Korea (Figure 1a). Therefore, precipitation may decrease in the regions located on the leeward side of the mountains because of the effects of Fohn-type winds (easterly wind) induced by the OKB (Figure 12). With wet days, significant values were obtained over the entire regions in South Korea without regional characteristics (Figures 11c-d). The partial correlation coefficients between OK_BF and area-averaged MEAN, P90, WD, and CWD were 0.40, 0.37, 0.35, and 0.37, respectively, which are significant at the 95% confidence level ( Table 2). Despite the characteristics of regional precipitation in South Korea, the frequent occurrence of OKB is closely associated with precipitation in South Korea over interannual timescales. These findings suggest that the Korean summer precipitation also can be modulated by baroclinic instability in association with the OK_BF, regardless of the LLMW.
Although drought is strongly associated with the variability of precipitation, it can also be modulated by the temperature variation. Based on previous analysis, the dominant effects of the increasing OK_BF on climate  over South Korea are decreasing temperature and increasing precipitation. Thus, the change in the water budget can be attributed to the change in the OK_BF. Figure 13 shows the partial regression maps of the 3-month (June to August) accumulated precipitation (P), PET, and difference between precipitation and PET (P-PET), excluding the effects of the LLMW. The spatial patterns of accumulated precipitation were the same as those in Figure 11a (Figure 13a). A decrease in PET was observed over most regions in South Korea due to cold air advection under the influence of the increasing OK_BF (Figure 13b). Both increases in precipitation and decreases in PET helped intensify the hydrological water budget (Figure 13c). The partial correlation coefficients between OK_BF and area-averaged PET and P-PET were −0.39 and 0.41, respectively, which were significant at the 99% confidence level (Table 3). These results suggest that the effects of the increasing OK_BF on both precipitation and PET can reduce the drought risk over South Korea.

| DISCUSSION
As mentioned in the introduction, the formation of the OKH is related to the AEA teleconnection pattern or the PJ (or EAP) teleconnection pattern. In this section, this study investigated the evolution of OKB through composite analysis. Before applying the case composite analysis, we first defined OKB day if blocked longitudes are found at least five grid points (12.5 ) in the Okhotsk Sea (140 E-160 E) on a certain day. One OKB event was defined if the OKB day lasts at least 5 days. In total, 28 OKB events were identified (Table 4). The following composite analyses were based on these identified 28 OKB events. For convenience, day 0 indicated the start date of OKB events and day −n (day +n) represented the nth day before (after) the start date of OKB events. In addition, the horizontal wave activity flux (WAF) described by Takaya and Nakamura (2001) was calculated to represent the propagation of quasi-stationary Rossby waves. Figure 14 represents the temporal variation of OKB events in the middle troposphere (500 hPa). On day −6, the eastward WAFs were detected from the North Atlantic to the Ural Mountains. From day −4 to day 0, the positive geopotential height anomalies near the Ural Mountains continued to be maintained and grew in magnitude. The eastward WAFs from Ural Mountains ramified into two branches. One propagated southeastward to Baikal Lake towards the middle latitudes, and the other propagated northeastward to Siberia towards the high latitudes. The persistent convergence between eastward WAFs from Siberia and northeastward WAFs from Baikal Lake further enhanced the OKH. These eastward WAFs resembled the AEA teleconnection pattern. Consequently, the double-blocking high pattern with the development of two blocking highs near the Ural Mountains and the Okhotsk Sea was predominantly characterized. Meanwhile, after day −2, the western extension of WNPSH (the positions of the 5,880 m isoline), which results in the growth of poleward WAFs from the tropics, was also identified, which was resembled the PJ or EAP teleconnection pattern. After day 0, the OKB became well-established and remained for at least 5 days, which maintained and reinforced the mid-latitude trough over NEA. These results suggested that the evolution of the OKB event was identified as having a more significant relationship with the eastward wave propagation at high latitudes than poleward energy propagation at low latitudes represented by the WAF.
Nevertheless, the impact of OKB events on precipitation over South Korea also should be considered in the individual blocking characteristics, such as duration, intensity, and extension. In addition, the relationship between the OKB events and summer precipitation over South Korea should be examined carefully over intra-seasonal timescales. The precipitation events are controlled mainly by other factors, such as the onset/peak/withdrawal period of BCM front and typhoon activities. According to a recent study, the occurrence of blocking in high latitudes led to stopping the northward march of the BCM front, leading to recordbreaking BCM rainfall in June-July 2020 (Chen et al., 2021). This suggests that further studies are needed to investigate the combined influence of the OKB and BCM front activities on precipitation over South Korea.

| SUMMARY
More than half of the annual precipitation in South Korea occurs in summer (JJA). Therefore, insufficient precipitation during that period can be related directly to the occurrence of droughts. Thus, understanding the variation in water resources is important for research related to drought and flood. OK_BF has a distinct negative (positive) correlation with the summer temperature (precipitation) in East Asia (Park and Ahn, 2014). These results suggest that OK_BF may influence the water budget components, including precipitation and PET, over South Korea. This study examined the detailed relationship between the OK_BF and both precipitation and PET over South Korea during summer using partial correlation and regression analyses.
Composite analysis for high and low PREC_SK years showed that the BFs over the Okhotsk Sea, particularly in the regions of 140 E-160 E, were closely related to the precipitation over South Korea. Thus, this study defined the OK_BF as the area-averaged BFs over the region and examined the relationship between PREC_SK and OK_BF.
Although the correlation coefficient between OK_BF and PREC_SK was 0.27, which is not significant at the 95% confidence level, significant positive correlations were observed at some stations (10 stations) over South Korea. Park and Ahn (2014) attributed this insignificant relationship to the locality of precipitation. Nevertheless, it is still unclear if this relationship is robust and how OK_BF contributes to precipitation variability.
This study performed partial correlation and regression analyses to separate the contribution of the OK_BF to precipitation variability from that of the LLMW. The low-level warm and moist air advection over South Korea strongly modulates the precipitation variability over that region. The correlation coefficient between LLMW and PREC_SK was 0.51, which is significant at the 99% confidence level. The correlation coefficient between the OK_BF and LLMW was −0.13, suggesting that both variations are statistically independent. The partial correlation coefficient between OK_BF and PREC_SK after excluding the effects of the LLMW was 0.40, which is significant at the 99% confidence level. Furthermore, significant positive correlations were observed at most stations (28 stations) over South Korea. The analysis suggests that the variability of PREC_SK can also be modulated by the change in OK_BF, regardless of the change in low-level air advection from low latitudes. Thus, these statistical methods might be even more suitable for examining the impacts of OK_BF on precipitation.
The vertical structures of positive geopotential height anomalies associated with the increasing OK_BF can induce an enhanced meridional temperature gradient over South Korea caused by the equatorward advection of cold air masses from higher latitudes. The enhanced meridional temperature gradient can induce baroclinic instability, resulting in increased precipitation over the region. These results suggest that Korean summer precipitation can also be modulated by the baroclinic instability associated with the OK_BF, regardless of the LLMW. The OK_BF was closely related to the variability of precipitation indices, including mean precipitation, extreme precipitation intensity, wet days, and consecutive wet days, over most regions in South Korea. In addition, the dominant effects of the increasing OK_BF on climate over South Korea are the decreasing PET and increasing precipitation. These results suggest that water resources may be abundant when OKB occurs frequently.