Influence of meridional circulation on extreme high temperature and weakened rainfall over the Yangtze River Valley in August 2022

The monthly average surface air temperature at 2m, rainfall, geopotential height, specific humidity, horizontal winds, vertical wind, surface pressure, and SST datasets for August from 1979–2022 were derived from the ERA5 reanalysis data (Hersbach et al 2020) and used in this study. The datasets were interpolated into the horizontal resolution of $0.5^\circ \times 0.5^\circ .$ In addition to the ERA5 reanalysis data, this study also adopts rainfall data from the GPCP data with a $2^\circ \times 2^\circ$ horizontal resolution (Adler et al 2018). The GSMaP_NRT rainfall data with a $0.1^\circ \times 0.1^\circ$ horizontal resolution was also used. This is important to note that the temporal period of the GSMaP_NRT data is 2000–2022. The anomalies were calculated by subtracting the climatological mean values of 1991–2020, while those from the GSMaP_NRT data were based on the climatological mean values of 2001–2020.

A rectangular area (105°E–122°E, 25°N–33°N) was selected to represent the YRV region (purple boxes in figures 1(a) and (b)). As discussed in section 1, the SAT and rainfall anomalies over the YRV were the hottest and driest in August 2022 after the linear trends were removed. Therefore, the influence of global warming was eliminated by removing the linear trends of the variables in the remainder of this study.

2.2.1. Three-pattern decomposition of the global atmospheric circulation

2.2.2. Novel moisture budget equation of the three-pattern circulations

In recent studies, a novel moisture budget equation of the three-pattern circulations (Han et al 2021, Cheng et al 2022, 2023) was established by combining the 3P-DGAC method with the moisture budget equation (Seager et al 2010).

$$\begin{eqnarray}&&\delta P=\delta P\_H+\delta P\_M+\delta P\_Z+\delta R,\end{eqnarray} \tag{ 1 }$$

where $\delta P$ is the abnormal precipitation, and the four terms from left to right on the right-hand side of equation (1) are the abnormal precipitation triggered by horizontal circulation (HC), MC, ZC, and the residual term. The difference between the value for 2022 and the climatological mean value for 1991–2020 is written as $\delta .$

$\delta P\_M$ can be further decomposed as follows:

$$\begin{eqnarray}&&\delta P\_M=\delta MCDA\_M+\delta MCDD\_M+\delta THA\_M+\delta THD\_M\end{eqnarray} \tag{ 2 }$$

where $\delta MCDA\_M$ and $\delta MCDD\_M$ ($\delta THA\_M$ and $\delta THD\_M$) are the abnormal rainfall related to dynamic term (thermodynamic term), which are triggered by the anomalous advection and divergence of MC. The terms on the right of equation (2) can be gotten using the following formulas:

$$\begin{eqnarray}&&\left\{\begin{array}{c}\delta MCDA\_M=\displaystyle \frac{1}{{\rho }_{W}g}\displaystyle {\int }_{{P}_{s}}^{0}(\delta {\vec{V}}_{M}\cdot {\rm{\nabla }}{q}_{0})dp\\ \delta MCDD\_M=\displaystyle \frac{1}{{\rho }_{W}g}\displaystyle {\int }_{{P}_{s}}^{0}({q}_{0}{\rm{\nabla }}\cdot \delta {\vec{V}}_{M})dp\\ \delta THA\_M=\displaystyle \frac{1}{{\rho }_{W}g}\displaystyle {\int }_{{P}_{s}}^{0}({\vec{V}}_{M0}\cdot {\rm{\nabla }}\delta q)dp\\ \delta THD\_M=\displaystyle \frac{1}{{\rho }_{W}g}\displaystyle {\int }_{{P}_{s}}^{0}(\delta q{\rm{\nabla }}\cdot {\vec{V}}_{M0})dp\end{array}\right.\end{eqnarray} \tag{ 3 }$$

where ${\rho }_{W},$ $g,$ ${\vec{V}}_{M},$ and $q$ are water density, gravitational acceleration, horizontal wind, and specific humidity, respectively. The climatological mean variables for 1991–2020 are represented by the subscript 0. Similar decomposition can also be obtained for $\delta P\_H$ and $\delta P\_Z.$

In the present study, the novel moisture budget equation was used to explore the quantitative influence of three large circulations (i.e., HC, MC, and ZC) on the weakened rainfall over the YRV in August 2022. In addition to the two main methods proposed above, linear fitting and regression analysis were performed. Additionally, the empirical orthogonal function (EOF) decomposition was used to obtain the dominant modes of the anomalous circulations.

Figure 2(a) illustrates the 500-hPa geopotential height anomaly in August 2022. It is evident that the WPSH has considerably strengthened and shifted westward comparing the 5880-gpm isopleth in 2022 and the climatological mean. Moreover, the strengthening and westward movements persisted throughout the August 2022 (Figure S2). The abnormal geopotential height at middle to high latitudes exhibited a 'positive-negative-positive' triple distribution, resulting in a low-pressure center in the northern part of the YRV. This further hindered the northward shift of WPSH and contributed to the persistent of the WPSH in the region. The persistent sinking motion over the YRV was triggered by the anomalous WPSH (figure 3(a)), leading to extreme high temperature and weakened rainfall events over the region. This finding was verified by the regressions of the regionally averaged SAT and rainfall over the YRV onto the geopotential height (vertical velocity). The regressions correspond to the spatial distribution of the abnormal geopotential height (vertical velocity) as shown in figures 2(a)–(c) (figures 3(a)–(c)). The scatter plots in figure 4 illustrates the association between the area mean SAT (rainfall) anomalies and the vertical velocity anomalies over the YRV. Figures 4(a) and (d) show that the vertical velocity anomaly is closely related to the SAT and rainfall anomalies, with explained variances of 0.46 and 0.87, respectively.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

To investigate the origin of the abnormal subsidence induced by the anomalous WPSH over the YRV, the vertical velocity was decomposed into two components using the 3P-DGAC method. Figures 3(a), (d), and (e) depict the abnormal vertical velocity and those of the MC and ZC at 500 hPa in August 2022. Figures 3(a), (d), and (e) show that MC is the primary contributor to the abnormal vertical velocity over the YRV, while the contribution of ZC was negative. Specifically, the quantitative contribution of the MC is 114%, while that of the ZC is –14%. Comparing figures 4(b) and (c) (figures 4(e) and (f)), the abnormal SAT (rainfall) is significantly related to the abnormal vertical velocity of MC, while the relationship between the anomalous SAT (rainfall) and anomalous ZC is marginal and nonsignificant, suggesting a leading role of MC in affecting the extreme high temperature and weakened rainfall events over the YRV through the excitation of the anomalous sinking motion.

Figure 5(a) displays the rainfall anomaly caused by the joint effects of the HC, MC, and ZC in August 2022. Figures 5(a) and 1(b) show that the spatial pattern of the precipitation anomaly and that triggered by the joint effect of the three-pattern circulations are rather similar. Furthermore, figure 6 shows that the area mean precipitation anomaly over the YRV triggered by the three-pattern circulations was –2.65 mm d^–1, comprising approximately 85% of the actual anomalous rainfall (–3.1 mm d^–1). This implies that the joint effect of the three large circulations can generally explain the weakened rainfall over the YRV in August 2022.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Further decomposition of the precipitation anomaly triggered by the three-pattern circulations is displayed in figures 5(b)–(d). The precipitation anomaly triggered by the MC resembles that triggered by the three large circulations (figures 1(b) and 5(a)). Moreover, the regionally averaged precipitation anomaly caused by MC over the YRV was –2.38 mm d^–1, approximately 77% of the factual precipitation anomaly (figure 6). However, the regionally averaged rainfall anomalies caused by HC and ZC over the YRV were merely –0.05 and –0.22 mm d^–1, which can explain 1% and 7% of the factual precipitation anomaly (figure 6). Therefore, the analysis concludes that the MC played a leading role in the extreme weakened rainfall event in August 2022.

The contribution of the MC to the rainfall anomaly can be decomposed into four components using equation (2). $\delta MCDD\_M$ was the main contributor to the rainfall anomaly triggered by the MC when comparing figures 5(c) and (f), while the contributions of the other three components were small, implying that the weakened rainfall over the YRV in August 2022 is mainly controlled by the divergence of the abnormal MC, and the role of moisture change is negligible (see 'Physical explanation of $\delta MCDD\_M$' in the Supplementary Material). Specifically, the quantitative rainfall anomaly over the YRV caused by $\delta MCDD\_M$ was –2.58 mm d^–1, which can explain 83% of the factual precipitation anomaly.

As the moisture is generally located below 500 hPa (Cheng et al 2023), integrating equation (3) up to 500 hPa does not generally alter the outcome of the analysis (comparing figures 5, 6(a) and S3, 6(b)). Therefore, the result obtained from the novel moisture budget equation theoretically support the conclusion that the MC dominated the extreme high temperature and weakened rainfall events in August 2022.

As the MC dominated the extreme high temperature and weakened rainfall events in August 2022, the characteristics and possible causes of the anomalous MC in August 2022 should be investigated. Figure 7(a) displays the zonal mean vertical velocity anomaly of MC between 105°E–122°E. The abnormal vertical velocity of MC was characterized as a 'negative-positive-negative-positive-negative' quintuple distribution with sinking motion over the YRV (figure 7(a)). The quintuple distribution of the MC can also be found in other high temperature and weakened rainfall events, and the MC in the years of high temperature and weakened rainfall events and that in the years of low temperature and enhanced rainfall events are nearly symmetric (Figure S4). The spatial distribution of the regressions of the area mean SAT and rainfall over the YRV onto the zonally averaged vertical velocity of the MC resembles that of the actual vertical velocity anomaly of the MC (comparing figures 7(a)–(c)), implying that the quintuple distribution of the vertical velocity anomaly of the MC is the main contributor to extreme high temperature and weakened rainfall events in August 2022.

Zoom In Zoom Out Reset image size

To explore the possible causes of the abnormal MC in August 2022, an EOF analysis was applied to the vertical velocity anomaly of the MC in August (figure 8). Generally, the variances of the three leading empirical orthogonal function modes (EOFs) are 32%, 21%, and 15%, respectively. The first leading mode (EOF1) is characterized as a 'negative-positive-negative' triple distribution with rising motion over the YRV (figure 8(a)). The corresponding time series of EOF1 (i.e., PC1) is nearly zero in 2022 (figure 8(d)), implying that the anomalous MC is unrelated to EOF1. However, it should be noted that PC1 reaches the third largest value in 2021 since 1979, favoring the sinking motion south of the YRV and the rising motion over the YRV, which further leads to the reoccurrence of Meiyu in August 2021. The second leading mode (EOF2) and third leading mode (EOF3) are both characterized as the 'negative-positive-negative-positive-negative' quintuple distribution (figures 8(b) and (c)). However, the range and intensity of the centers are different in EOF2 and EOF3. PC2 is almost zero, and PC3 reaches its largest value since 1979 (figures 8(e) and (f)), implying that the anomalous MC in August 2022 is closely connected to EOF3.

Zoom In Zoom Out Reset image size

To verify the results proposed above, the regression of PC3 onto the 500-hPa geopotential height, SAT, and rainfall in August during 1979–2022 (figure 9) was carried out. Figure 9 shows that when PC3 is positive, there is an abnormally strong WPSH, resulting in anomalous high temperatures and negative rainfall over the YRV. Namely, the spatial distributions of the regressions and those of the anomalous geopotential height, SAT, and rainfall are rather similar (comparing figures 9(a)–(c) and 1(a), (b), 2(a)). Therefore, the causes of the highest PC3 values in 2022 should be investigated.

Zoom In Zoom Out Reset image size

To investigate the possible causes of the largest value of PC3 in 2022, this study regressed PC3 onto the SST in August during 1979–2022 (figure 10). The regression field of SST resembles the actual SSTA (comparing figures 10(a) and (b)), implying that the SSTA played a crucial role in the abnormal MC. The SSTAs are mainly characterized by the following three features. First, the SSTA is negative (positive) over the tropical west (southeast) Indian Ocean, which is an Indian Ocean dipole (IOD) pattern. Second, the SSTA is characterized as positive (negative) in the tropical west (east) Pacific, which is a La Niña pattern. Third, the North Atlantic SSTA is characterized by a North Atlantic triple (NAT) pattern with a positive (negative) SSTA over the central (north and south) North Atlantic.

Zoom In Zoom Out Reset image size

Three SSTA indices are adopted in this study to explore the impact of SSTA on MC. The area mean SSTA over the tropical central–eastern (5°S–5°N, 160°E–150°W) Pacific is used to define the Niño 4 index (Tang et al 2023). The difference in the area mean SSTA between the tropical west (10°S–10°N, 50°E–70°E) and southeast (10°S–0°, 90°E–110°E) Indian Ocean is used to define the IOD index (Jiang et al 2022, 2023). The NAT index is defined as the difference between double the area mean SSTA over the central North Atlantic (40°N–55°N, 60°W–20°W) and the sum of the area mean SSTA over the northern (60°N–70°N, 40°W–10°W) and southern (25°N–35°N, 35°W–15°W) North Atlantic (Jiang et al 2023).

Figure 11 displays the three SSTA indices and the regressions between the three SSTA indices and the vertical velocity of the MC. Figure 11 shows that the Niño 4 index (multiplied by –1) is the second largest, and the IOD index (multiplied by –1) and NAT SSTA index reach the largest since 1979 (figures 11(a), (d), (g)). After removing the influences of PC1 and PC2 using multivariate linear fitting, the main conclusions remain unchanged (comparing the first and second columns of figure 11). The spatial patterns of the regressions are all characterized as the 'negative-positive-negative-positive-negative' quintuple distribution, which is similar to EOF3, confirming the significant influence of the SSTA.

Zoom In Zoom Out Reset image size

Specifically, when the Niño 4 index (multiplied by –1) and the IOD index (multiplied by –1) are in the positive phase, anomalous clockwise ZC over the Pacific and anticlockwise ZC over the Indian Ocean exist, which can lead to an abnormal rising motion over the Maritime Continent (figures 12(b) and (c)). The rising motion of the ZC further favors the rising motion of the MC by coupling effect. On the other hand, when the NAT SSTA index is positive, the abnormal SSTA in the Atlantic can lead to anomalous Rossby wave train transporting from the Atlantic (figure 13) (Tang et al 2023). The positive action center over East Asia can lead to high-pressure anomalies, which can further result in a sinking motion over the YRV. Therefore, the joint effect of the SSTA over the three oceans leads to the anomalous MC, which ultimately results in extreme high temperature and weakened rainfall events in August 2022.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size