Influence of PDO and ENSO with Indian summer monsoon rainfall and its changing relationship before and after 1976 climate shift

In this study, we investigated the possible relation between the Indian summer monsoon and the combination of the different phases of Pacific Decadal Oscillation (PDO) and El Niño Southern Oscillation (ENSO) before and after the climate shift in 1976. This study is carried out using IMD’s rainfall dataset, HadISST v1.1 dataset and twentieth century reanalysis dataset by comparing anomalies of the respective parameters from 1901 to 2020. It is found that when positive (negative) phases of PDO and El Niño (La Niña) co-occur, deficit (surplus) rainfall are likely to occur over entire India. The SST signatures of both the PDO and ENSO are showing their respective spatial structures. However, when negative (positive) PDO and El Niño (La Niña) co-occur, the signal is mixed and it is unlikely that either surplus or deficit rainfall conditions will occur over entire India, and the SST signatures are not capturing their proper spatial pattern. In other words, when ENSO and PDO are in (out of) phase they enhance (counteract) the conventional monsoon-ENSO relationship. After confirming the climate shift in 1976, study periods were further divided into pre and post climate shift periods based on Niño 3.4 index and PDO index and analysed their impact on the Indian summer monsoon rainfall. In the pre-shift example, in-phase conditions exhibit similar qualities to those described above. Rainfall patterns are more indicative of ENSO than PDO. In the post-shift situation, the positive anomaly of SST in the PDO and Niño region is significantly stronger than in the pre-shift phase. When compared to the pre-shift example, positive rainfall anomalies are amplified during positive PDO and El Niño, while negative PDO and La Niña show a weakening of positive rainfall anomalies. The out-of-phase condition has a balancing effect due to the counteracting impact, but with an increased positive anomaly of SST. In that combination, rainfall patterns with PDO characteristics rather than ENSO characteristics emerge. Circulation features at 850 hPa during the pre-shift and post shift periods show considerable changes as an indication of the climate shift. During the pre-shift of positive PDO and La Niña, convergence at low level enhances over Indian subcontinent and resulting enhanced rainfall; however, in the post shift period the strength of convergence reduces, and it leads to reduced rainfall. The patterns of stream function and velocity potential are also consistent with rainfall during the pre and post shift periods.


Introduction
Indian summer monsoon rainfall (ISMR) from June to September has various socio-economic impacts across south Asia particularly over the Indian subcontinent (Webster et al. 1998;Gadgil and Gadgil 2006;Rajeevan and Nanjundiah 2009).This monsoon rainfall accounts for 70-90% annual rainfall over India (Shukla and Huang 2016).The ISMR has large spatio-temporal variability, which can impact the economic growth of the country (Mooley et al. 1984;Kripalani et al. 2003) that plays a vital role in agriculture, reservoir management, hydrological power generations and many more, thus it controls the economy of India (Gadgil and Rupa Kumar 2006;Mertz et al. 2009;Saha and Mooley 1978).ISMR exhibits different temporal variabilities ranging from diurnal to multidecadal time scales in addition to its spatial heterogeneity (Krishnamurthy and Shukla 2000;Rajeevan et al. 2010;Varikoden et al. 2010Varikoden et al. , 2012;;Seetha et al. 2020;Reshma et al. 2021).Earlier studies also mentioned that the variabilities in ISMR can be due to the local forcing (Chakraborty et al. 2002;Nair et al. 2018;Shukla and Wallace 1983) as well as by teleconnections like El Niño Southern Oscillations, Pacific decadal oscillations, Atlantic Multidecadal Oscillations (Varikoden and Preethi 2013;Varikoden et al. 2015;Luo and Lau 2020;Malik et al. 2017;Nair et al. 2018).The combinations of such teleconnections can also be a reason for these types of variabilities (Krishnamurthy and Krishnamurthy 2014Krishnamurthy , 2017;;Levine et al. 2017).There are studies mentioned about the empirical and dynamical relationships between ISMR and ENSO (Shukla and Paolino 1983;Webster and Yang 1992;Webster et al. 1998;Kumar et al. 1999;Varikoden et al. 2015;Seetha et al. 2020;Hrudya et al. 2020Hrudya et al. , 2021) ) and ISMR and PDO (Malik et al. 2017;Joseph 2014;Malik and Brönnimann 2018;Krishnan and Sugi 2003).Their combined effect always has greater importance, since both are happening in the Pacific Ocean (Mantua et al. 1997;Mantua and Hare 2002;Rasmusson and Wallace 1983).
The PDO and ENSO have a negative correlation with ISMR (Sikka and Gadgil 1980;Rasmusson and Wallace 1983;Krishnamurthy andKrishnamurthy 2014, 2017;Krishnan and Sugi 2003;Kumar et al. 1999;Sikka 1980;Varikoden et al. 2015).The ENSO influence the ISMR in the presence of PDO on the basis of the phases they coexist (Krishnamurthy and Krishnamurthy 2014).It also determines rainfall pattern, strength and intensity (Krishnamurthy and Krishnamurthy 2014Krishnamurthy , 2017)).Sampling variability regarding ENSO-monsoon correlation and its decrease in strength after 1970s is also mentioned in earlier studies (Cash et al. 2017).
Climate change refers to the change in the environmental conditions of the earth due to many internal and external factors (Crowley 2000;Stern and Kaufmann 2014) and, it has become a global concern over the last few decades (Fischer et al. 2021;Baines and Folland 2007).Alterations to the earth's atmosphere that occur over much longer periodsdecades to millennia-characterise as climate change.
The stepwise changes of the climate system between different states in temporal domain are called climate shifts (Miller et al. 1994) and such shifts were noticed in the North Pacific between 1976 and 1977, and that is manifested in the time series of the PDO (Graham 1994;Mantua and Hare 2002), which ultimately changed the basic mechanisms of a phenomenon that happened in the Pacific Ocean (Meehl et al. 2009).Climate change and climate shift are having a different perspective.Their normal characteristics such as their ordinary SST range, precipitation range, etc., are also changed (Sabeerali et al. 2012).The climate shift has increased the Pacific Ocean SST by 0.75 °C (Nitta and Yamada 1989;Gaffen et al. 1991).Studies show that there exist dynamically consistent changes in the rainfall, SST, surface winds, and organised convection during the mid-1970s (Graham 1994Suhas and Goswami 2008;Sabeerali et al. 2012;Sahana et al. 2015;Kuttippurath et al. 2021). Preethi et al. (2017) also found a similar climate shift in rainfall history over south Asia and East Asian monsoon regimes around 1970s.These changes can also be reflected on the relationship of ENSO and PDO on ISMR.
There are studies on the variabilities of ISMR and its association with ENSO and PDO (Krishnamurthy andKrishnamurthy 2014, 2017), however, the combined influence of ENSO and PDO on the ISMR variability and its changes in the recent decades are yet to be studied systematically.The changes occurred after the climate shift of 1976/1977 has also not been investigated.Therefore, the present study explores the influence of ENSO and PDO on ISMR before and after the climate shift.The changes in the recent decades and their linked dynamics are carried out on the atmospheric, and oceanic parameters to bring out the variability of ISMR by the influence of ENSO and PDO in different combinations.In Sect.2, we briefly explained the datasets and methods used in this study.Variabilities that happened to SST, rainfall pattern and circulation features due to the combination of ENSO and PDO are explained in Sect.3. Section 4 depicts the conclusions of the study.

Data
We have used long period daily gridded rainfall data set over India from India Meteorological Department (IMD) with high spatial resolution (0.25 × 0.25 latitude-longitude) from 1901 to 2020 (Rajeevan et al. 2008;Rajeevan and Bhate 2009;Pai et al. 2014).This data set was prepared by running basic quality control over 6955 rain gauge stations in India and missing data is replaced from neighbouring rain gauge measurements.The climatological and variability aspects of rainfall over India produced from this data set were comparable to the existing gridded daily rainfall data sets.Furthermore, due to its higher spatial resolution and the higher density of rainfall stations used in its development, the spatial rainfall distribution, such as heavy rainfall areas in the orographic regions of the west coast and over northeast, low rainfall in the leeward side of the Western Ghats was more realistic and better presented in IMD gridded dataset.
Sea Surface Temperature (SST) analyses are carried out using monthly SST data from HadISST version 1.1 (https:// www.metof ce.gov.uk/ hadobs/ hadis st), which has a spatial resolution of 1° longitude × 1° latitude (Rayner et al. 2003).It is reconstructed using a two stage reduced space optimal interpolation procedure, followed by superposition of quality improved gridded observations onto the reconstructions to restore local detail (Rayner et al. 2003).Additionally, the zonal and meridional circulations are also used in this study, which are obtained from twentieth century reanalysis version 3 (20CRv3) provided by National Oceanic and Atmospheric Administration-Cooperative Institute for Research in Environmental Sciences (https:// psl.noaa.gov/ data/ gridd ed/ data.20thC_ ReanV3.press ure.html).It has 1° latitude × 1° longitude horizontal spatial resolution with 28 vertical pressure levels (1000-1 hPa) on a monthly time scale (Slivinski et al. 2019).The 20CR ensemble is the first to span over 100 years of sub-daily global atmospheric conditions.This gives the best estimate of the weather at any given location and time, as well as an estimate of its confidence and uncertainty.The version 3 system (NOAA-CIRES-DOE 20CRv3) employs improved data assimilation methods, such as an adaptive inflation algorithm, a newer, higher-resolution forecast model that specifies dry air mass, and assimilates a larger set of pressure observations.
For selecting years of different phases of PDO and ENSO, the PDO index and Niño 3.4 index were taken.PDO index which is known as the leading principal component of monthly sea surface temperature variability of North Pacific Ocean (Deser et al. 2016) and Niño 3.4 index which is defined as the average SSTs of the region in between 5° N-5° S, 120° W-170° W in the Pacific Ocean, (Trenberth 2020) was obtained from the website https:// www.cpc.ncep.noaa.gov/ data/ and https:// www.ncdc.noaa.gov/ telec onnec tions of National Oceanic and Atmospheric Administration-Cooperative Institute for Research in Environmental Sciences (NOAA-CIRES), respectively.

Methods used in the analysis
The study period  was classified into four quadrants based on the Niño 3.4 and PDO indices.The quadrants were selected as per the orientation of the cartesian coordinate system.The quadrant 1 (3) is the combination of the warm (cold) phase of PDO and El Niño (La Nina).The quadrant 2 (4) is the combination of the cold (warm) phase of PDO and El Niño (La Niña).In this category, the PDO and ENSO are in phase in the quadrants 1 and 3, whereas they are out-of-phase in 2 and 4 quadrants.In this study, index values are considered that exceed plus/minus 60% of standard deviations.The warm phase is defined when index values are > 0.36 for ENSO and > 0.55 for PDO, while the cold phase is defined as when the value is < − 0.36 for ENSO and < − 0.55 for PDO.
However, to study the changes in the influence of climate shift on ISMR, the study period was subdivided into preshift (1930-1975) and post-shift (1976-2020) periods.The pre and post periods were categorised into four quadrants as done earlier.However, we modified the above criteria of 60% of the standard deviation to 25% of the standard deviation of the respective indices for incorporating more cases in the sample space for all the four quadrants before and after the 1976/1977 periods to make it more statistically acceptable.
Interannual variations of PDO and Niño 3.4 index are used to identify variation of PDO and ENSO within the period of 1930-2020.The effect of PDO and ENSO on the precipitation patterns of Indian monsoon summer rainfall is studied using correlations between these indices and ISMR.A clear signature of climate shift was noticed in 1976/1977 period from the interannual variability of both the indices.Spatial anomaly is used to compare differences in the pattern of parameters like precipitation, SST, zonal and meridional wind, stream function, velocity potential, that happened before and after climate shift.Here, the anomalies are calculated by considering the period from 1930 to 2020 as the climatological base period.Two tailed Student t test is used to evaluate statistical significance of the analysis.

Climatology of ISMR and SST
The climatological features of rainfall in the Indian subcontinent and SST in the Indo-Pacific domain during the summer monsoon season (June to September) for the period of 1901 to 2020 are given in Fig. 1. Figure 1a depicts high rainfall over the windward side of the Western Ghats and the northeastern regions of the country, where the monsoon rainfall exceeds 16 mm day −1 .Over central India, the rainfall is about 10 mm day −1 and it is less than 5 mm day −1 over southeast, north and northwest Indian regions.The Western Ghats in the west coast region receives high rainfall, with an average annual rainfall of about 3800 mm year −1 in the windward side (Rao 1976).North eastern region also gets high amounts of rainfall compared to the north and central part of India.Rain shadow region of south India receives less rainfall during the monsoon season.
In Fig. 1b, the equatorial Pacific Ocean experiences high SST greater than 27 °C, which is concentrated over the western side between 120° E and 160° E. Eastern side experiences relatively low SST between 21 °C and 25 °C.SST decreases towards higher latitudes beyond 20° N. The equatorial Indian Ocean shows SST more than 27 °C.The North Pacific or PDO region is comparatively cooler than the Niño region.The spatial distribution of rainfall and SST varies considerably during the occurrence of ENSO and PDO, moreover the co-occurrence of both the events can amplify the changes in their spatial pattern.Hence a detailed understanding of the variability in different space time domains and relationship of PDO and ENSO with ISMR is needed.

Interannual variation and relationship of PDO and ENSO
Interannual variations between PDO and ENSO along with linear trend for the period before and after 1976 (1930-1975 and 1976-2020) for a period from 1930 to 2020 during southwest monsoon are given in Fig. 2a.PDO and ENSO interannual variations have had major alterations before and after the 1970s.The standard deviation of Niño 3.4 index is 0.60 and PDO index is 0.92.In the decadal variations, before the 1970s (pre-shift), the phase transitions of PDO took 15-20 years but after the 1970s (post-shift), only 12-15 years were taken for it.In the years after the 1970s, the warm (cold) phase of PDO tends to occur with the warm (cold) phase of ENSO.Pre-shift years showed a less intense ENSO and PDO where magnitude of indices were in between − 1.5 and + 1.5.But the years after 1970, the PDO index have phases with high intensity which are varying in between − 2.5 to + 2.5.During 1930 to 1975, interannual variations of PDO index and Niño 3.4 index are inhomogeneous.Occurrence of El Niño (La Niña) is not coinciding with positive PDO (negative PDO).Intensity of ENSO is smaller (in between − 1 to + 1) than that of PDO (in between − 2 to + 2).During 1976 to 2020, the occurrence of El Niño (La Niña) coincided with the occurrence of positive PDO (negative PDO).Both indices have similar intensities (ranging from − 3 to + 3) and interannual fluctuations.These observable changes indicate that a climate shift occurred in 1976, which is consistent with previous studies (Graham 1994;Chaluvadi et al. 2021;Miller et al. 1994).
It should be noticed that the ENSO trend has changed from negative to positive (trends are not significant), however the PDO trend has shown significant changes in their trend values (from − 0.29 to − 0.01).Global warming and abrupt changes in SST conditions can be a reason for the changes occurring in the prevailing climatic conditions (Philander and Philander 2008;Schuldt et al. 2011).

Combined effect of PDO and ENSO on SST and ISMR anomalies
To  The previously described SST features influence southwest monsoon rainfall patterns.Spatial distribution of rainfall anomalies of southwest monsoon during different combinations of PDO and ENSO in Fig. 4 is used to analyse this concept.The negative PDO and La Niña composite of rainfall shows positive anomalies of rainfall all over India except the North eastern region (Fig. 4a).Higher anomalies (> 3 mm day −1 ) are observed in the northern portion of the Konkan coast.This enhanced excess rainfall condition over almost the entire India indicates that La Niña and nPDO, both of which produce excess rainfall conditions over India, are complementing each other.pPDO and El Niño composite (Fig. 4c) shows below normal rainfall patterns over central India and western coast of the country (Krishnan and Sugi 2003;Kanamitsu and Krishnamurti 1978;Krishnamurthy and Goswami 2000).However the Northeastern region is showing positive anomalies of rainfall.This enhanced signature of negative rainfall anomalies over India seems to be the result of pPDO and El Niño, complementing each other.Figure 4b shows nPDO and El Niño composite of rainfall in which India is experiencing a combination of deficit rainfall and neutral conditions, with a domination of deficit rainfall conditions.In a pPDO and La Niña composite (Fig. 4d), excess rainfall occur in the north and central regions of India, whereas negative rainfall anomalies occur in the southern India and western regions of Gujarat.Hence out-of-phase cases having a mixed pattern of rainfall with ENSO domination.Hatches in the figure represents 5% significance level.

Comparison of SST and rainfall using pure conditions
To further substantiate the relation between monsoon and ENSO and to understand such relation relative to PDO, composites of summer monsoon seasonal anomaly of rainfall and SST of pure (all cases of a particular phenomenon without conditions on the other index) ENSO and PDO cases are shown in Figs. 5 and 6, respectively, 5% significance level is hatched in the figures.In this case, we considered 14, 11, 11 and 16 number of years for positive PDO, negative PDO, The Indian Ocean is likewise witnessing a negative SST anomaly, particularly in the Arabian Sea.Rainfall composite during La Niña period (Fig. 6a) exhibits an above-normal precipitation pattern across India, with the exception of the central and eastern parts, which show below-normal precipitation.El Niño is causing a horseshoe pattern of positive SST anomalies in the central and eastern equatorial Pacific and negative anomalies in the western equatorial Pacific (Fig. 5b).The positive SST anomaly can also be detected in the western Indian Ocean.Rainfall composite during El Niño (Fig. 6b) depicts the deficit rainfall situation in India.
A positive SST anomaly can be evident in the northeastern Pacific during a warm or pPDO (Fig. 5c), but a significant negative SST anomaly can be detected in the central and western Pacific.A positive SST anomaly can also be found across the Indian Ocean.Rainfall composite of pPDO (Fig. 6c) indicating a minor precipitation deficit.A slight negative SST anomaly exists in the northeastern Pacific and significant warming central to northwestern Pacific during nPDO (Fig. 5d).The Indian Ocean is experiencing positive SST anomalies and can observe an above normal rainfall pattern in the rainfall composite of nPDO (Fig. 6d).Deficit rainfall conditions are experienced when the pPDO and El Niño is present.Deficit rainfall severity is greater in El Niño situations than in pPDO.Excess rainfall is the result of both nPDO and La Niña.As noted previously, the La Niña situation causes more positive rainfall anomalies than negative PDO situations.It can be concluded that the ISMR has a negative correlation with PDO and ENSO.Spatial correlation of ISMR with PDO and Niño 3.4 indices can further substantiate this result.Figure 7 is the spatial correlation coefcient of Indian summer monsoon rainfall with PDO and Niño 3.4 indices during southwest monsoon.Contours represent regions with 95% significance level.
The two indices show regionally varying relationships with ISMR.The correlation is significant over most parts of India, except in the northeast, where the correlation is insignificantly out-of-phase as noted in earlier studies (e.g., In some parts of India correlation of ISMR with PDO index is becoming positive.The ENSO (Fig. 7b) is having a strong negative correlation with ISMR compared to PDO.Most of the regions are having a significance of more than 95% confidence level.This can be identified using contours drawn in the figure.When comparing years with a combination of phases and pure ones, it is noted that if they are in-phase, strong and well defined SST features of both events occurring in the considered region.When they are out-of-phase, their intensity drops and various places experience neutral conditions.The horseshoe pattern of SST, which is an important feature of the ENSO SST pattern, is well characterised in in-phase conditions but not in out-ofphase conditions.Rainfall intensifies (weakens) when both are in warm (cold) phase, ie, during in-phase conditions.In the out-of-phase scenario, ENSO patterns are more prominent than PDO patterns.

Climate shift and change in the behaviour of teleconnections
Climate shifts in the Pacific Ocean have a significant impact on the phenomena that occur there (Meehl et al. 2009;Mantua and Hare 2002).During the shift basic characteristics and properties of all the phenomena in the Pacific Ocean changed (Graham 1994).Considering the changes in the Pacific ocean, It is assumed that phenomena like ENSO and PDO, which are produced by the SST and SLP variations in the Pacific Ocean can also be changed along with this.
To recognize changes in the SST anomalies of the Pacific Ocean under distinct ENSO and PDO phase combinations, spatial distribution of SST (°C) anomalies during PDO and ENSO (1930-2020 period were selected as the base period for this analysis) were drawn for the pre-shift and the postshift (Fig. 8).Third panel illustrates the difference of the SST anomalies between the pre-shift and the post-shift.Statistical significance of the distribution with 95% confidence is hatched.Negative SST anomalies can be detected in the northwestern and central Pacific during pPDO and El Niño situations (Fig. 8a), with positive anomalies near the western coast of north America.Positive anomalies can also be seen in the central and eastern equatorial Pacific, while negative anomalies can be seen in the western equatorial Pacific.Positive anomalies detected around the western coast of North America, as well as in the eastern and central Pacific.However, the same relationship exists in the post-shift period with higher magnitudes (Fig. 8b).Negative anomalies in the western Pacific region are diminishing.The Indian Ocean has negative anomalies and neutral temperatures before the shift, but positive anomalies after the shift.The anomaly values are statistically significant at 5% level which is indicated by hatches.
Negative SST anomalies can be detected in the northwestern and central Pacific during positive PDO and La Niña situations (Fig. 8d), with positive anomalies near the Fig. 7 The spatial correlation coefcient of Indian summer monsoon rainfall (mm day −1 ) with a PDO index, b Niño 3.4 during southwest monsoon for 1901-2020 southwest monsoon period.Contours represent regions with a 5% significance level western coast of North America.Positive anomalies can also be seen in the central and eastern equatorial Pacific, while negative anomalies can be seen in the western equatorial Pacific.Positive anomalies are found around the western coast of North America, as well as the eastern and central Pacific, become stronger after the shift (Fig. 8e).Negative anomalies in the western Pacific region are diminishing.The Indian Ocean has negative anomalies and neutral temperatures in the pre-shift condition, however, it turned to positive anomalies during the post-shift period.
In nPDO and El Niño conditions (Fig. 8g), positive SST anomalies in the north Pacific with a weak negative anomaly around it show nPDO characteristics.The horseshoe pattern of positive SST anomalies across central and eastern Pacific with weak negative anomalies at the western Pacific shows El Niño signature.However, the anomalies are weak compared to the in-phase composites.Thus, the nPDO phase and El Niño counteract each other to minimise their proper signals.El Niño characters are visible compared to PDO.The Indian Ocean is having a negative SST anomaly.In the post-shift phase (Fig. 8h), a stronger positive anomaly of SST is observed in the central and western Pacific, with a weaker negative anomaly along north America's western coast.Positive anomaly is present in the equatorial Pacific, reducing El Niño signatures which depicts clear dominance of PDO characteristics.Figure 8j, which shows nPDO and La Niña, has negative SST anomalies over the western coast of North America and positive SST anomalies along the western and central Pacific, both are having strong PDO signatures.Negative SST anomalies over the equatorial Pacific is a strong indicator of La Niña.SST traces of both events are observed.Positive anomalies are evident everywhere over the Oceanic region after the shift (Fig. 8k), with the exception of the eastern side of the Pacific, which has weak negative SST anomalies.Both phenomenon's signature traits are missing, despite the increased positive anomaly of SST.The Indian Ocean is having negative SST anomalies in the pre-shift however positive anomalies in the post-shift.
The difference between the pre-shift and the post-shift (Fig. 8c, f, i, l) shows strong positive anomalies all over the study region and very weak negative anomalies over equatorial Pacific Ocean, which indicates significant increase in the SST after the climate shift.In-phase conditions (Fig. 8c,  i) have slightly decrease in temperature in the North Pacific region where as out-of-phase conditions (Fig. 8f, l) are having comparatively stronger negative SST anomalies in the equatorial and north eastern Pacific region.But the overall distribution has positive anomalies clearly denoting an enhancement in SST after climate shift period.
To investigate variations of ISMR and to identify differences in ISMR pattern before and after climate shift, spatial distribution of rainfall (mm day −1 ) anomalies during PDO and ENSO periods before climate shift and after climate shift is shown in Fig. 9 with significance level of 5% is hatched.In the positive PDO and El Niño case, the preshift condition (Fig. 9a) is showing a negative anomaly of rainfall with some positive rainfall anomaly western coast and northern parts of India.whereas the after shift (Fig. 9b) is showing a more intensive negative anomaly of rainfall.The difference between shifts (Fig. 9c) showing negative anomalies of rainfall over central India and north eastern region indicating deficit precipitation after the shift.The pPDO and La Niña case is having negative rainfall anomalies in the pre-shift period (Fig. 9g), whereas mixed rainfall anomalies are identified in the post-shift (Fig. 9h). Figure 9i shows a mixed signatures of rainfall pattern.During nPDO and El Niño, negative anomalies of rainfall (El Niño favourable situation) can be seen in the pre-shift case (Fig. 9d), whereas positive anomalies (nPDO favourable situation) can be seen in the after shift phase (Fig. 9e).Strong decrease in the rainfall is observed all over India in Fig. 9f.nPDO and La Niña show an increased amount of rainfall anomaly in the pre-shift (Fig. 9j) but during the post-shift (Fig. 9k), positive rainfall anomaly weakens and can see some negative rainfall anomalies in north eastern and central portions of India.The difference in anomalies are indicating decrease in rainfall (Fig. 9l).Thus the analysis shows that during in-phase conditions, SST anomalies for ENSO and PDO combinations are showing their individual SST signatures.Warm (cold) phase has characteristics of El Niño (La Niña) and pPDO (nPDO) in the pre-shift.In the post-shift, negative (positive) anomalies become weaker (stronger).Rainfall decreases (increases) when warm (cold) phases coexist during the preshift.After the climate shift, negative (positive) anomalies of rainfall become stronger (weaker).Out-of-phase conditions counteract each other and minimise their characteristic SST signals.But ENSO features are prominent during the pre-shift and PDO features are prominent during after shift.Rainfall patterns are coinciding with this result.

Enhancement of changes using dynamical parameters
Understanding the low level circulation anomalies helps us to identify the regions where the circulation pattern is favourable for rainfall.Figure 10 shows wind anomalies at 850 hPa (in ms −1 ) for the pre-shift phase and the post-shift JJAS seasons and the difference of the anomalies between the pre-shift and the post-shift.The anomaly values are statistically significant at 5% level and are indicated by contours In pPDO and El Niño conditions, the pre-shift case (Fig. 10a) shows the westerly component of trade winds strengthens over the central Pacific region.This strengthening modifies the ENSO-monsoon relationship.Anomalous westerlies are also observed at 20° N in the Pacific Ocean, which indicates the weakening of trade winds due to the high SST/ low SLP produced by pPDO and El Niño conditions.This combined effect decreases rainfall in India.The divergence pattern of the wind vector (Fig. 10a) which leads to descending motion can also be attributed to decreased rainfall.Figure 10b shows the post-shift condition of pPDO and El Niño, decreased intensity of anomalous westerlies is seen in both Niño and PDO regions.The divergence pattern of wind vectors is prominent over India, which further decreases rainfall than that of the pre-shift condition.In this phase the difference between the pre-shift and the post-shift (Fig. 9c) is overall indicating the strengthening of westerly component of trade winds.The pPDO and La Niña during the pre-shift situation (Fig. 10d), show a convergence of the wind, thereby enhanced rainfall over India.Easterlies are prominent in the Niño region and the Indian Ocean region.Weakened easterlies originated from the western side of the Niño region due to the lower SST and higher SLP due to the climatic conditions caused by the La Niña.In the post-shift case (Fig. 10e), there is divergence all over central India, and strong easterlies are evident, decreasing rainfall.The rainfall is further weakened by westerlies across the Indian Ocean.
The difference (Fig. 10f) is having prominent divergence over India and positive wind anomalies in equatorial central Pacific region and Indian Ocean which shows consistent shift of wind patterns after the shift.The pre-shift (Fig. 10g) has weak anomalous westerlies over the Indian region, which strengthen the prevalent trade winds over the Indian subcontinent that leads to deficit rainfall during nPDO and El Niño conditions.During the postshift period (Fig. 10h), strong westerlies led to convergence over the Indian regions, enhancing convection and thereby enhancing rainfall over the Indian subcontinent.Corresponding changes are seen in the difference as well (Fig. 10i).In nPDO and La Niña conditions (Fig. 10j), a convergence is observed over India in association with very weak easterlies in the western Pacific.This convergence weakens during the post-shift situation (Fig. 10k), which weakens the rainfall.Strong easterlies are observed in the western central Pacific region along with strong westerlies in the central eastern and north Pacific regions.The difference between the preshift and the post-shift (Fig. 10l) is having a weak positive anomalies of wind concentrated in equatorial Pacific with a significant convergence over Indian subcontinent results in the decrease in the rainfall.Analysis of the wind vectors are consistent with the results obtained from the analysis of SST and rainfall.
Further analysis of velocity potential and stream function is done to understand large scale features of horizontal circulation at 850 hPa levels (Fig. 11).Difference of the anomalies between the pre-shift and the post-shift is also determined with statistical significance test.During the preshift case (Fig. 11a), a tendency of divergence across India is noticed that reduces rainfall over there.The equatorial Pacific is showing higher convergence and thereby ascending motion of air.In the post-shift period (Fig. 11b), a positive anomaly of velocity potential intensified, increasing the descending motion through an increase of divergence over India and decreasing rainfall.In the pre-shift phase of pPDO and La Niña (Fig. 11d), intensified convergence over Indian Ocean increases the amount of rainfall.In the post-shift (Fig. 11e) phase, there is a divergence since convergence center shift to east which reduces rainfall.Intensified divergence patterns are also observed in between 0 and 20° N in the post-shift period.The difference in both cases (Fig. 11c,  f) are having a convergence center in equatorial Pacific region and significant divergence over India.They also indicated change of circulation patterns before and after shift.
During the nPDO and El Niño, the divergence pattern caused by the relocation of the convergence centre towards the eastern Indian Ocean resulted in deficit rainfall over India in the pre-shift case (Fig. 11g).The post-shift case (Fig. 11h) shows a small convergence over north India, causing an above normal rainfall.Increased convergence near the Indian subcontinent enhances positive rainfall anomalies in the pre-shift situation of nPDO and La Niña (Fig. 11d).But the post-shift situation (Fig. 11k) shows a slightly divergent pattern over India, which weakens the positive rainfall anomaly.Difference of the anomalies (Fig. 11i-l) are having significant shift with 95% confidence level, which is consistent with the analysis results.Thus in-phase conditions show an increased positive (negative) rainfall anomalies when warm (cold) phases of PDO and ENSO coexist.In the case of out-of-phase conditions, the rainfall pattern shows similar to the rainfall pattern of ENSO during the pre-shift case; however, the rainfall pattern shows as that of the PDO pattern during the post-shift period.
In order to compare large scale circulation features at lower level (850 hPa) between pre-and post-shift periods, anomalies of stream function are given in Fig. 12. Difference between the anomalies of the pre-shift and the post-shift is also included where significance level of 5% is hatched.During pPDO and El Niño during the pre-shift (Fig. 12a), Northern Pacific is witnessing a positive anomaly in stream function, resulting in anticyclonic circulation and descending motion where as equatorial Pacific is having negative anomalies of stream function.In the post-shift (Fig. 12b) case, positive anomalies enhanced and region affecting it is enlarged simultaneously with strong negative anomalies of stream function in the equatorial Pacific.The difference (Fig. 12c) indicating this circulation change correctly.In the pre-shift case (Fig. 12d), pPDO and La Niña shows a feeble negative anomaly of stream function over India, which increases rainfall as per property of La Niña but the post-shift (Fig. 12e) shows just the opposite, which decreases the rainfall pattern as in pPDO.In the nPDO and El Niño pre-shift case (Fig. 12g), a positive anomaly of stream function is observed in north and equatorial Pacific.But after shift (Fig. 12h) shows a positive anomaly of stream function over India with decreased rainfall by the clockwise circulation pattern and descending motion of air is observed.This change in circulation pattern is evident in the Fig. 12i.The nPDO and La Niña (Fig. 12j) case is showing negative anomalies with enhanced rainfall in the pre-shift but post-shift shows a weak negative anomaly (Fig. 12k), and it can be attributed to deficit rainfall.The difference of the stream function anomalies between the pre-shift and the post-shift in Fig. 12c, f, i, l is showing a significant change of distribution of anomalies in the post-shift with respect to the pre-shift.In general, the results of the previous analysis are also consistent with SST and rainfall observations, as well as dynamical characteristics like horizontal wind velocity and velocity potential.

Conclusions
This study has investigated the characteristics of SST during the simultaneous occurrences of pPDO and El Niño, nPDO and El Niño, pPDO and La Niña, and nPDO and La Niña.The impact of these various combinations on rainfall patterns before and after climate shift is also investigated.Following confirmation of effects of the shift, an inquiry into alterations in the circulation pattern was conducted.An association between the decadal and interannual variability of the North and equatorial Pacific SST and the Indian summer monsoon rainfall has been established through correlation and composite analyses.The PDO index exhibits significant negative correlation with the ISMR, similar to the relation between ISMR and Niño 3.4 index.However, the ISMR-PDO correlation is slightly less than that between ISMR and Niño 3.4 index.In order to confirm the year of climate shift, time series of PDO index and Niño 3.4 index were drawn.Significant changes in the intensity and pattern of occurrences of the phases of PDO and ENSO were observed before and after 1976 hence 1976 is selected as the year of climate shift.
When positive (negative) phases of PDO and El Niño (La Niña) co-occur, more intense deficit rainfall (excess rainfall) are likely to occur over entire India.However, when negative (positive) PDO and El Niño (La Niña) co-occur, the signal is mixed and it is unlikely that either severe deficit or severe rainfall conditions occur over entire India.In other words, when ENSO and PDO are in (out of) phase they enhance (counteract) the conventional monsoon-ENSO relationship.When years were divided into the pre-shift and the post-shift periods, the attributes acquired during comprehensive analysis were completely altered.In the pre-shift example, in-phase conditions exhibit similar qualities to those described above; if both are in the warm (cold) phase, rainfall decreases (increases) more than the ambient condition.SST signatures during the in-phase situations exhibit the qualities of both the phenomena.Rainfall patterns are more indicative of ENSO than PDO.The rainfall pattern is determined by the El Niño or La Niña conditions existing in that combination.
In the post-shift situation, the positive anomaly of SST in the PDO and Niño region is significantly stronger than in the pre-shift phase.When compared to the pre-shift example, the in-phase condition shows an amplification of negative rainfall anomalies during positive PDO and El Niño and a weakening of positive rainfall anomalies during negative PDO and La Niña.Because of the counteracting impact, the out-of-phase condition has a neutralising effect, but with an increased positive anomaly of SST.Rainfall patterns exhibiting PDO traits rather than ENSO features occur in that combination.
The difference between the pre-shift and the post-shift anomalies results into larger values, implying that the change in parameters in the post-shift versus the pre-shift is relatively high.The difference is statistically significant in the majority of the study areas, indicating a significant shift is indeed happened in 1976.In pPDO and El Niño conditions, wind anomalies at 850 hPa during the pre-shift show strengthened westerly trade winds in the central Pacific, resulting divergence and in decreased Indian rainfall.Anomalies during post shift period indicate decreased intensity of westerlies and prominent divergence over India, further decreasing rainfall.During pPDO and La Niña pre-shift, wind convergence enhances Indian rainfall, while weakened easterlies in the post-shift weaken the rainfall.The stream function and velocity potential studies are all mostly consistent with the same mode of variation.The divergence and convergence of the circulation pattern have considerable changes before and after the climate shift, which can lead to the plausible changes in the ISMR distribution.All the above results are statistically significant by a 95% confidence level.
Figure 2b is showing the regression analysis of the Niño 3.4 index and PDO index during the southwest monsoon from 1930 to 2020.In the Fig. 2b, the first quadrant represents (warm) positive PDO (pPDO) and El Niño phases (positive values for PDO index and Niño 3.4 index), second quadrant represents (cold) negative PDO (nPDO) and El Niño phases (negative values for PDO index and positive values for Niño 3.4 index), third quadrant represents nPDO and La Niña (negative values for PDO index and Niño 3.4 index) and fourth quadrant represents pPDO and La Niña (positive values for PDO index and negative values for Niño 3.4 index).There exist years having warm (cold) phases of PDO and ENSO.Mixed cases are also seen but the number of occurrences is less.Regression line is showing a positive relationship.Points are staggered and the linear correlation coefcient is 0.5 and regression coefcient of linear relation is 0.3 which denotes an in-phase relationship.Even though

Fig. 1
Fig. 1 Spatial distribution of a climatological rainfall (mm day −1 ) and b sea surface temperature (°C) during the summer monsoon.The base period for calculating climatology is considered from 1901 to 2020 substantiate the relation between ISMR and combination of PDO and ENSO, composites of summer monsoon seasonal anomalies of rainfall and SST are shown in Figs.3 and 4, respectively.The number of pPDO and El Niño, pPDO and La Niña, nPDO and El Niño and nPDO and La Niña years included in the composite analysis are 32, 27, 21 and 39, respectively.The nPDO and La Niña composite of SST is shown in Fig.3aindicating negative SST anomalies in the eastern part along the west coast of North America and stronger positive SST anomalies in the central and western parts of the north Pacific.This gives strong nPDO signatures.A pattern of severe negative anomalies exists in the eastern and central equatorial Pacific, with a neutral situation

Fig. 2 a
Fig. 2 a Interannual variation of PDO and Niño 3.4 indices along with the linear trend for the period before 1976 and after 1976 (1930-1976-pre-shift and 1977-2020-post-shift) during the southwest monsoon.Blue and red lines represent the PDO and Niño 3.4 indices,

Fig. 5
Fig. 5 Spatial distribution of SST (°C) anomalies during pure PDO and pure ENSO periods.a La Niña, b El Niño, c Positive PDO, d Negative PDO.Hatches in the figure represent 5% significance level

Fig. 6
Fig. 6 Spatial distribution of rainfall (mm day −1 ) anomalies during pure PDO and pure ENSO periods.a La Niña, b El Niño, c Positive PDO, d Negative PDO.Hatches in the figure represent 5% significance level

Fig. 8
Fig. 8 Spatial distribution of SST (°C) anomalies during PDO and ENSO periods in the pre-shift and the post-shift JJAS season and the difference of SST anomalies between the pre-shift and post-shift.a Positive PDO and El Niño, b Positive PDO and El Niño, c b − a, d

Fig. 9
Fig. 9 Spatial distribution of rainfall (mm day −1 ) anomalies during PDO and ENSO periods in the pre-shift and the post-shift JJAS seasons and the difference of rainfall anomalies between the pre-shift and the post-shift.a Positive PDO and El Niño, b Positive PDO and

Fig. 10
Fig. 10 Wind anomalies at 850 hPa (in ms −1 ) for the pre-shift phase and the post-shift JJAS seasons and the difference of the anomalies between the pre-shift and the post-shift.a Positive PDO and El Niño, b.Positive PDO and El Niño, c b − a, d Positive PDO and La Niña,

Fig. 11
Fig. 11 The anomalies of velocity potential at 850 hPa (10 -5 m 2 s −1 ) at 850 hPa for the pre-shift and the post-shift phase and the difference of velocity potential anomalies between the pre-shift and the post-shift.a Positive PDO and El Niño, b Positive PDO and El Niño,

Fig. 12
Fig. 12 Spatial anomalies of stream function (10 5 m 2 s −1 ) at 850 hPa for the pre-shift and the post-shift phase.and the difference of stream function anomalies between the pre-shift and the post-shift.a Positive PDO and El Niño, b Positive PDO and El Niño, c b − a, d Positive