Linear regression of precipitation against Niño3 SST of the prediction (left panels) and the observation (right panels) in November, December and January during the ENSO peak phase for the period 1996–2009. The difference between January and December regressed catchment area with respect to Niño3 SST for the prediction (left panels) and the observation (right panels). The monthly mean precipitation climatology for the prediction (left panels) and the observation (right panels) in November, December, and January from 1996–2009.
Changes in winter atmospheric circulation over East Asia and the North Pacific associated with ENSO in a seasonally varying period. It is suggested that the intrawinter changes are governed by the relative role of the central Pacific (CP) and western North Pacific (WNP) equatorial precipitation anomalies on the extratropical teleconnection over the North Pacific Ocean. Further analyzes of the ensemble distribution of the forecast data support the relative role of CP and WNP precipitation anomalies in influencing the extratropical circulation over the North Pacific.
Introduction
Although the Philippine Sea anticyclone plays a rather important role in the link between the ENSO climate system and East Asia, Son et al. 2014) pointed out that it is insufficient to explain the association of East Asia with the Philippine Sea anticyclone alone. For example, the Philippine Sea anticyclone is confined to the subtropics, making it difficult to influence the midlatitude region north of 30°N. 2014) suggested that in addition to the Philippine Sea anticyclone, the Kuroshio anticyclone, located in the North Pacific, significantly affects climate variability in East Asia, covering East China, Korea, and Japan. Anomalous diabatic cooling due to suppressed convective heating over the WNP is known to be responsible for the development of the Philippine Sea anticyclone in the lower troposphere and cyclonic flow in the upper troposphere.
The upper-level cyclonic flow leads to the extratropical anticyclone through the propagation of Rossby wave energy, which is a key process in the development of the Kuroshio anticyclone. Son et al (2014) argued that the Kuroshio anticyclone exists in early winter due to relatively strong WNP negative precipitation anomalies. In particular, the compensating role between WNP and equatorial CP precipitation anomalies in the development and sudden disappearance of the Kuroshio anticyclone will be intensively examined.
Data and model
Reanalysis and observation data
The atmospheric circulation observational data are taken from the monthly mean National Centers for Environmental Prediction/The National Center for Atmospheric Research (NCEP/NCAR; Kalnay et al., 1996) reanalysis data to compare with seasonal prediction data. The grid spacing in the reanalysis data is a 2.5˚ latitude–longitude resolution, and the data set used is from 1996 to 2009. For the monthly mean CPC Merged Analysis of Precipitation (CMAP; Xie and Arkin, 1997), data of the same resolution and period are used.
Also, the NOAA Extended SST reconstructed version 3 (ERSST.v3d; Smith et al., 2008) is analyzed for tropical Pacific SST variation.
Verification of the prediction skill
It is interesting that there is a significant skill over East Asia in November and January. 2014) showed that the precipitation variability in the Korean Peninsula is highly correlated with the El Niño index in November and December, but their relationship is weakened in January. This significant skill in East Asia may be the result of its close relationship with ENSO, which will be discussed below. So far, we have shown that the model has good skill where the El Niño-related signal is strong in the observation.
This suggests that the model has the ability to simulate the observed El Niño patterns and may provide a good test bed to evaluate the observational hypothesis of the intra-winter changes in the El Niño teleconnection to East Asia.
Teleconnection over East Asia and North Pacific
Alternatively, the cyclonic flow in the North Pacific is stronger than in the November observational data. Almost simultaneously, the negative precipitation anomalies occur with relatively suppressed convection in the equatorial western Pacific. The anomalous diabatic cooling due to the weak convection over the WNP induces an upper-level cyclonic flow and the lower-level Philippine Sea anticyclone (Gill, 1980).
The upper-level cyclonic flow over the WNP leads to an anticyclone in the midlatitudes through the propagation of Rossby wave energy (Hoskins and Karoly, 1981). On the other hand, the intensification of convective activity in the equatorial CP leads to anomalous cyclonic flow in the North Pacific through a PNA-like pattern and ultimately plays a role in suppressing the Kuroshio anticyclone. On the other hand, in January, the negative precipitation anomalies in the WNP weaken, and the positive precipitation anomalies in the equatorial CP strengthen.
Therefore, the sudden disappearance of anticyclone Kuroshio in January is due to the dominant effect of positive precipitation anomalies in the equatorial CP ( Son et al. 2014 ). The observational patterns exhibit typical rainfall features during the peak phase of El Niño, such as positive rainfall anomalies in the equatorial CP and negative rainfall anomalies in the WNP. Compared to December anomalies, negative precipitation in WNP decreases significantly in January, while positive precipitation in equatorial KP tends to increase with southward displacement.
As expected from the precipitation correlation skill shown in Figure 3 , GloSea5 captures the overall patterns of negative precipitation in the WNP and positive precipitation in the equatorial CP ( Figure 6a,c,e ). On the other hand, the positive anomalous precipitation in the equatorial CP becomes weaker in January (Figure 6e). Overestimation of equatorial CP precipitation also leads to a stronger cyclonic flow in the North Pacific.
This plays a role in keeping the Kuroshio anticyclone in the model forecast different from the observed development.
Teleconnection difference from ensemble spread
One is for the strong negative precipitation in the WNP and the weak positive precipitation in the equatorial CP compared to the ensemble mean precipitation anomaly in each region. The other had the opposite condition, the weaker negative precipitation in the WNP and the stronger positive precipitation in the equatorial CP than the ensemble mean precipitation anomaly. As expected, the strong WNP case is characterized by the strong negative precipitation in the WNP and the weak positive precipitation in the equatorial CP.
On the other hand, the strong CP case shows the weak negative WNP precipitation and strong positive equatorial CP precipitation. Because the precipitation in the equatorial CP is a direct response to the SST warming, the spread between the models is relatively small. Based on the two groups, we can isolate the impact of the different tropical diabatic heatings on the teleconnection patterns by comparison in the same month.
In the strong CP case, it is clear that the cyclonic flow overwhelms the North Pacific, while the Kuroshio anticyclone is almost absent (Figure 12b). In the strong CP case, on the other hand, the cyclonic current still overwhelms the North Pacific, and the anticyclonic current is weakly found in East Asia. In January, the weak anticyclonic flow is found in the western North Pacific and the cyclonic flow is dominant over the North Pacific, where the strong WNP cases are clustered (Figure 12e).
In the strong CP case, the explosive cyclonic flow settles over the entire North Pacific, including East Asia (Figure 12f). It is very interesting that the strength of the cyclonic flow becomes stronger in January than in December, although the positive precipitation is somewhat weakened in the equatorial CP. Moreover, the weaker negative WNP precipitation in January plays a crucial role in the strong cyclonic flow development.
In summary, the strong WNP case consistently shows a relatively strong anticyclonic flow near East Asia and a weak cyclonic flow over the North Pacific during all months.
Climatological field in tropics and extratropics
Summary and discussion
That is, the ensemble distribution clearly supports the relative role of the WNP and the CP precipitation anomalies on the extratropical teleconnection over the North Pacific. First, the hypothesis suggested in the observational data is confirmed by connecting the tropical precipitation estimation problem to the incorrect teleconnection. Second, the performance of the KMA operational seasonal forecast model in predicting the teleconnection is evaluated.
Based on the relationship between tropical rainfall and teleconnection pattern simulated in the forecast model, forecast skill for extratropical regions can be improved using statistical post-processing (e.g. Kug et al., 2008a,b). Long-term hindcasts are needed to assess systematic biases in tropical precipitation and the teleconnection model. In particular, since most CMIP5 climate models have difficulty simulating the Kuroshio anticyclone, multi-model analyzes can provide insight into what the most recent climate models often lack in simulating the El Niño teleconnection. in the extratropics including East Asia. Analyzes of the CMIP5 multimodel ensemble are not included in this paper, it also failed to simulate teleconnection models.
CMIP5 captures well the typical patterns of the negative precipitation in the WNP and the positive precipitation in the equatorial CP. In January, the anomalous positive precipitation in the equatorial CP becomes stronger, similar to the observational data. Yamagata, 2001: The impact of the Indian Ocean Dipole on the relationship between the Indian monsoon rainfall and ENSO.
Kim, 2013: Impact of the dominant large-scale teleconnections on winter temperature variability in East Asia. Kumar, 1997: Teleconnective response of the Pacific-North American region atmosphere to large central equatorial Pacific SST anomalies. Nitta T., 1987: Convective activities in the tropical western Pacific and their impact on the Northern Hemisphere summer circulation.
Jin, 2002: The role of Indian Ocean warming in the development of the Philippine Sea anticyclone during ENSO, Geophys. Zhang, 2012: Different evolutions of the Philippine Sea anticyclone during the eastern and central Pacific El Niño: Possible SST effects in the Indian Ocean. The black dots indicate that the correlation abilities are higher than the inertia and the correlation is 0.3.
