Prediction and mechanistic analysis of May precipitation in North China based on April Indian Ocean SST and the Northwest Pacific Dipole

Tiejun Xie ,Tihen Feng ,Rong Zhi ,Ji Wng ,Qing Zhng

a Beijing Municipal Climate Center, Beijing, China

b School of Atmospheric Sciences, Sun Yat-sen University, Zhuhai, China

c National Climate Center, China Meteorological Administration, Beijing, China

Keywords:North China May precipitation Indian Ocean sea surface temperature Northwest Pacific Dipole Synergistic effect Annual variability

ABsTRACT North China May precipitation (NCMP) accounts for a relatively small percentage of annual total precipitation in North China,but its climate variability is large and it has an important impact on the regional climate and agricultural production in North China.Based on observed and reanalysis data from 1979 to 2021,a significant relationship between NCMP and both the April Indian Ocean sea surface temperature (IOSST) and Northwest Pacific Dipole (NWPD) was found,indicating that there may be a link between them.This link,and the possible physical mechanisms by which the IOSST and NWPD in April affect NCMP anomalies,are discussed.Results show that positive (negative) IOSST and NWPD anomalies in April can enhance (weaken) the water vapor transport from the Indian Ocean and Northwest Pacific to North China by influencing the related atmospheric circulation,and thus enhance (weaken) the May precipitation in North China.Accordingly,an NCMP prediction model based on April IOSST and NWPD is established.The model can predict the annual NCMP anomalies effectively,indicating it has the potential to be applied in operational climate prediction.

1.Introduction

In recent decades,numerous studies have been carried out on precipitation in North China (34°—44°N,110°—120°E),achieving many important results.However,most have focused on the summer season,when precipitation is more abundant,with relatively fewer studies having been conducted on precipitation in May in North China (e.g.,Wang et al.,2006 ;Ding et al.,2008).May is the transition period from spring to summer in the Northern Hemisphere,and the North China May precipitation (NCMP) plays an important role in springtime crop planting and winter wheat growth in North China (Wang et al.,2006 ;Zhao et al.,2018).As shown in Fig.1(a,b),the annual variability of NCMP anomalies is very significant.For instance,in both 1985 and 1998,the positive NCMP anomalies exceeded 30 mm,while the multiyear average NCMP during 1979—2021 was only～38 mm,and had good spatial consistency across North China.Therefore,exploring the possible precursor signals of NCMP anomalies and thus predicting them effectively is of great scientific and social value.

Water vapor supply is a necessary condition for precipitation in North China (Wei et al.,2005 ;Ma and Gao,2006 ;Jiang et al.,2017).Based on the climatological trajectories method,Jiang et al.(2017) found that the water vapor over the whole year in North China derives mainly from the Indian Ocean and Northwest Pacific.Chen et al.(2020) indicated that the western tropical Indian Ocean sea surface temperature (SST) can influence the early summer precipitation in North China through the latent heat changes caused by first influencing precipitation in central-northern India.Moreover,some studies have also pointed out that the Indo-Burmese trough,the South Asian monsoon,the South Asian high,and other factors related to the Indian Ocean climate (e.g.,Liu and Ding,2008 ;Huang et al.,2012 ;Wei et al.,2014),as well as the East Asian summer monsoon,the western Pacific subtropical high,and other factors related to the Northwest Pacific climate (e.g.,Zhang et al.,2003 ;Ding et al.,2008 ;Sun et al.,2009 ;Wu and Jiao,2017),have a significant influence on precipitation in North China.Therefore,the precursor signals of NCMP anomalies may also be associated with the Indian Ocean climate and Northwest Pacific climate,and focusing on May to further explore these influences on NCMP is necessary.

Fig.1.(a) Time series of NCMP during 1979—2021.(b) Correlation map of NCMP with May precipitation over North China,in which the dotted area shows where the correlation coefficients exceed the 95% confidence level.

In this study,it was found that the NCMP anomalies are closely related to the April SST in the Indian Ocean and the Northwest Pacific,defined here as the Indian Ocean SST (IOSST;Xie et al.,2021) and Northwest Pacific Dipole (NWPD),respectively,which can synergistically influence NCMP anomalies,and therefore serve as predictors of the annual NCMP variability.

2.Data and methods

The monthly precipitation data are from the Global Precipitation Climatology Project (GPCP) Version 2.3 Combined Precipitation Dataset,with a 2.5° × 2.5° horizontal resolution,covering the period from 1979 to 2021 (Adler et al.,2003).For comparison,the daily meteorological dataset of basic meteorological elements of China National Surface Weather Stations (V3.0) was also used.As the results based on the two datasets were found to be consistent,only the results based on the GPCP data are shown in this paper.The monthly wind,geopotential height,and relative humidity data are from the NCEP—DOE AMIP-II reanalysis dataset,with a 2.5° × 2.5° horizontal resolution and covering the period 1979—2021 (Kanamitsu et al.,2002).The SST data used in this study are from the HadISST dataset,with a 1° × 1° horizontal resolution and covering the period 1870—2021(Rayner et al.,2003).

The NCMP time series is defined as the areal-weighted average of May precipitation over the studied North China region (34°—44°N,110°—120°E).April IOSST is calculated as the areal-weighted average SST in April over the Indian Ocean region (20°S—10°N,60°—100°E).The April NWPD index is defined as the difference between the normalized areal-weighted average SST in April over the areas (26°—36°N,150°—170°E) and (18°—24°N,145°—155°E) of the Northwest Pacific.The vertically integrated whole-layer water vapor flux is calculated following the Gill (1982).

3.Results

3.1.Relationship of NCMP with April IOSST and NWPD

It is found that NCMP has a significant correlation with the April SST in the Indian Ocean and Northwest Pacific (Fig.2 (a)).Fig.2 (b)shows the correlations between the April IOSST and May precipitation anomalies on the annul time scale over North China and the surrounding region.It can be seen that there are significant correlations between April IOSST and May precipitation anomalies over North China and the surrounding region,except in the southeastern part of North China.This suggests that there may be other climate factors besides April IOSST that can also be used to predict NCMP anomalies.Further analysis reveals that the April SST in the Northwest Pacific here defined as NWPD is also significantly associated with May precipitation in North China (Fig.2 (c)).The correlation coefficient of April IOSST and NWPD during the period 1979—2021 is 0.268,which is not significant at the 95% confidence level,indicating that April IOSST and NWPD are independent.These results indicate that NCMP anomalies may have a potential link with the IOSST and NWPD in April.

As shown in the time series of NWPD anomalies,April IOSST anomalies,and April NWPD index values (Fig.2 (d)),the three have significant consistency across annual changes,suggesting that NCMP anomalies may be synergistically influenced by the preceding IOSST and NWPD.In detail,the correlation coefficients of NCMP anomalies with April IOSST anomalies and NWPD index during the period 1979—2021 are 0.47 (exceeding the 95% confidence level) and 0.49 (exceeding the 95% confidence level),respectively.It can also be seen that when NCMP is anomalously positive in 1983,1991,2010,2005,and 2020,both April IOSST and NWPD are also anomalously positive;and likewise,when NCMP is anomalously negative in 1981,1986,1989,2000,2001,2017,and 2021,both April IOSST and NWPD are also anomalously negative.In addition,when NCMP is anomalously positive in 1988 and 1998,April IOSST is also anomalously positive,even though there is an insignificant positive or even negative anomaly of April NWPD.Similarly,in 1985,1990,and 2018,April NWPD is anomalously positive,even though there is an insignificant positive anomaly or even negative anomaly of April IOSST.Particularly,in 1985 and 1998,extremely strong positive NCMP anomalies correspond to strong positive NWPD and IOSST anomalies,respectively.This is similar to when there is a negative NCMP anomaly,such as in 1994,1995,and 1996,i.e.,there is also a negative April NWPD anomaly,even though there is an insignificant negative or even positive April IOSST anomaly.

Fig.2.(a) Correlation map of NCMP anomalies with April SST.(b,c) Correlation maps of April (b) IOSST and (c) NWPD with May precipitation anomalies over North China and the surrounding region during 1979—2021.The dotted areas indicate where the correlation coefficients exceed the 95% confidence level.(d) Time series of NCMP anomalies,April IOSST anomalies,and NWPD indices during 1979—2021.

Fig.3.Composite geopotential height (shading;units: gpm) and wind (vectors;units: m s -1) anomalies at 500 hPa in May corresponding to (a) positive and (b)negative NCMP anomalies.(c,d) Regressions of geopotential height (shading;units: gpm) and wind (vectors) anomalies at 500 hPa in May onto (c) IOSST and (d)NWPD in April at normalized values.The dotted areas indicate where the regression coefficients for geopotential height exceed the 95% confidence level.(e,f) As in(a) but for when the April anomalies of both IOSST and NWPD are positive or negative,respectively.(g,h) Regressions of geopotential height (shading;units: gpm)and zonal—vertical wind (vectors) anomalies averaged over 35°—44°N in May onto (g) IOSST and (h) NWPD in April at normalized values.

Fig.4.Composite of vertically integrated whole layer water vapor flux (vectors) and their divergence (shading) anomalies at 500 hPa in May corresponding to NCMP (a) positive and (b) negative anomalies.Regressions of vertically integrated whole layer water vapor flux (vectors) and their divergence (shading) anomalies at 500 hPa in May onto (c) IOSST and (d) NWPD in April at normalized values.(e) and (f) As in (a),but for the composite during the period when both IOSST and NWPD in April are positive anomalies or negative anomalies,respectively.

Fig.5.(a) Observed (red line) and modeled (blue line) NCMP anomalies for the period 1979—2021.Shaded areas indicate the 2-sigma uncertainty range for the modeled NCMP.(b) As in (a) but with the hindcast NCMP for the period 2017—2021 added.

These results indicate that it may be possible to improve the predictability of NCMP by combining the April IOSST and NWPD,and that NCMP may be synergistically influenced by the April IOSST and NWPD.

3.2.Potential mechanisms by which IOSST and NWPD affect NCMP

Further analyses were conducted to explore the mechanisms by which IOSST and NWPD affect NCMP.Composite geopotential height and wind anomalies at 500 hPa in May corresponding to positive and negative NCMP anomalies are shown in Fig.3 (a,b),respectively.It can be seen that when there is a positive NCMP anomaly,there are also positive geopotential height anomalies at 500 hPa in the Indian Ocean—South China Sea—Northwest Pacific region,and negative geopotential height anomalies in northwestern North China.From the wind fields,there is an anomalous cyclone in the Sri Lankan region of the Indian Ocean and another in the western region of Lake Baikal,along with an anomalous anticyclone in the southeastern seas of North China,which together are conducive to the transport of water vapor from the Indian Ocean and Northwest Pacific to North China.When there is a negative NCMP anomaly,there is a “negative—positive—negative ”distribution of geopotential height anomalies across the Indian Ocean through to the Tibetan Plateau and North China,along with negative geopotential height anomalies in the Sri Lankan region of the Indian Ocean and northeastern North China,and positive geopotential height anomalies in the Tibetan Plateau region,which together are not conducive to the transport of water vapor from the Indian Ocean and Northwest Pacific to North China (Fig.3 (b)).These findings also correspond to the results reported by Zhou and Huang (2006).

A spatial regression analysis of the geopotential height and wind anomalies at 500 hPa in May onto the April IOSST and NWPD is also shown,in Fig.3 (c) and Fig.3 (d) respectively,to analyze the mechanisms by which IOSST and NWPD affect NCMP.It can be seen that the patterns are generally consistent with the composite results corresponding to NCMP positive anomalies.In Fig.3 (c),the anomalous anticyclone in the southeastern seas of North China is weaker,while the cyclone near Sri Lanka in the Indian Ocean better reproduced.Similarly,in Fig.3 (d),the anomalous cyclone near Sri Lanka in the Indian Ocean is weaker,while the anomalous anticyclone in the southeastern seas of North China better reproduced.These results further indicate that NCMP anomalies are synergistically influenced by IOSST and NWPD.Composite geopotential height and wind anomalies at 500 hPa in May corresponding to positive and negative anomalies of both April IOSST and NWPD are also shown,in Fig.3 (e) and Fig.3 (f) respectively,showing results that are consistent with Fig.3 (a—d),but this is not discussed in detail here.As shown in Fig.3 (g,h),the role of vertical motion in NCMP is also related to April IOSST,especially NWPD.

As shown in Fig.4,the patterns for the vertically integrated whole-layer water vapor flux anomalies and their divergence also correspond to the results in Fig.3.The water vapor needed for May precipitation in North China comes mainly from the Indian Ocean and Northwest Pacific,and positive anomalies of IOSST and NWPD are conducive to the transport of water vapor from Indian Ocean and Northwest Pacific to North China,respectively,and vice versa.

3.3.NCMP prediction model based on April IOSST and NWPD

To further demonstrate the importance of the IOSST and NWPD for annual NCMP variability and to predict NCMP anomalies in advance,we construct three linear models: one relating NCMP to April IOSST,one relating NCMP to April NWPD,and the other to both April IOSST and NWPD.The results show that the NWPD model based on both IOSST and NWPD is better than the model based on IOSST or NWPD alone.The prediction model for annual NCMP based on IOSST and NWPD in April is established as follows:

As shown in Fig.5 (a),this prediction model for NCMP can simulate changes that are consistent with the observed NCMP anomalies,with the correlation coefficient between the simulated and observed NCMP reaching 0.61 (exceeding the 95% confidence level).Hindcast experiments are conducted to check the prediction skill of this NCMP prediction model.As shown in Fig.5 (b),this NCMP model has good predictability for NCAP anomalies,indicating that this model may be applied in operational climate prediction.

4.Summary and discussion

Based on observed and reanalysis data,the relationship of NCMP with April IOSST and NWPD was analyzed,the possible mechanisms involved were discussed,and an NCMP prediction model based on April IOSST and NWPD was established.The main conclusions can be summarized as follows:

(1) The correlation between NCMP anomalies and both the April IOSST and NWPD is significant,both spatially and temporally.This indicates that there may be some links between NCMP and April IOSST and NWPD,and that NCMP can be predicted in advance by the April IOSST and NWPD.

(2) Mechanistic analysis showed that positive (negative) IOSST and NWPD anomalies in April can enhance (weaken) the water vapor transport from the Indian Ocean and Northwest Pacific to North China by influencing the related atmospheric circulation,and thus enhance (weaken) the May precipitation in North China.

(3) The prediction model for annual NCMP anomalies based on IOSST and NWPD in April can effectively predict NCMP anomalies and has the potential to be applied in operational climate prediction.

(4) Although the relationship of NCMP with April IOSST and NWPD and the related mechanisms have been examined in this paper,it is undeniable that there are other climatic factors that can influence NCMP variability.Wu et al.(2003) suggested that spring rainfall in North China is significantly correlated with preceding-winter Ni?o3.4 SST.Zhang and Sun (2019) also pointed out that the connection between spring precipitation over North China and tropical eastern-central Pacific SST has strengthened.Therefore,on the basis of April IOSST and NWPD,further exploration of these other climate factors that can influence NCMP variability would be helpful for improving NCMP prediction.In addition,extreme rainfall events in North China are underestimated by the prediction model based on April IOSST and NWPD,suggesting that further consideration of nonlinear effects may enhance the prediction of extreme rainfall events in this region.

Funding

This work was supported by the National Natural Science Foundation of China [grant number 41975088].