Reversal nature in rainfall pattern over the Indian heavy and low rainfall zones in the recent era

This study presents the reversal nature in rainfall over heavy rainfall zone (HRZ; more than 80% of the long period average (LPA) of the Indian summer monsoon rainfall (ISMR)) and low rainfall zone (LRZ; less than 40% of ISMR-LPA) in India. The India Meteorological Department (IMD) high-resolution (0.25° × 0.25°) dataset is used from 1901 to 2016. The single and multiple change-point detection techniques are used to find the change in rainfall patterns over both regions. Further, the study period is divided into two halves P1 (1901–1958) and P2 (1959–2016) to examine the change in rainfall patterns in the recent and past periods. In P2, the rainfall pattern gets reversed, and ISMR has shown a significant increasing (decreasing) trend over the LRZ (HRZ). The increasing/decreasing number of moderate- and high-intensity rainfall events is one of the causes for this reversal pattern. Additionally, the number of dry days is increased over the HRZ and decreased over the LRZ. This study further confirms that the “dry becomes drier and wet becomes wetter” paradigm is not solely acceptable for India. The present study provides information about changes in dry days and ISMR variability in the context of climate change, which will be useful to agricultural risk management, water resources, drought monitoring, model developers, and policy planner on the adaptation strategies for climate change.


Introduction
The Indian summer monsoon (ISM) is an important meteorological phenomenon, occurred each year without failing and has a year-to-year variation (Gadgil et al. 2019). This ISM phenomenon is crucial for the country's economy because more than half of the employment of India directly or indirectly depends on agricultural and allied sectors (Economic Survey 2017-20182018. The western and eastern branches of the ISM intersect over east India due to which the eastern zone receives an ample amount of rainfall compared to western India (Rajeevan et al. 2010). The monsoon slowly sets over India in June and withdraws in September (Ding and Sikka 2006). There are two prime hypotheses that explain monsoon circulation over India; in the first one, it is believed that the Indian monsoon is a gigantic land-sea breeze driven by land-sea contrast (Gadgil 2007); alternatively, it is a movement of the Inter-Tropical Convergence Zone (ITCZ) (Gadgil 2018). In both hypotheses, distribution of rainfall is abiding, i.e., eastern India experiences heavy rainfall and the northwestern part receives less rainfall. To date, this spatial distribution of the long period average of seasonal rainfall over all-India remains almost the same; however, year-to-year fluctuation in rainfall amount is observed for ISM rainfall (ISMR).
Previously, many attempts have been made to understand the year-to-year variability of ISMR (Krishnamurti et al., 1998;Goswami and Mohan, 2001;Mohanty et al. 2005;Mishra et al. 2012;Sinha et al., 2013;Walker et al. 2015;Krishnan et al. 2020). The major concern of the ISMR variation is its spatiotemporal distribution which is non-uniform, and its variability is independent of latitude or longitude. For example, in Gujarat, out of 16 stations, 11 stations have experienced an increased in summer monsoon rainfall (SMR) by 5% per decade since 1960 (Dave et al. 2017), while over Jharkhand, which is located at the same latitude as Gujarat, it has shown the 15.65% reduction in ISMR for the same period (Sharma and Singh 2017). On the other hand, the southern parts of Western Ghats have shown significantly decreased in ISMR, whereas the northern parts 1 3 of Western Ghats have encountered a significant increase in ISMR which lies almost in the same longitude (Rajendran et al. 2012;Kothawale and Rajeevan 2017;Varikoden et al. 2019). Moreover, the other parts of the country also show random increased or decreased in ISMR, e.g., Sindh river basin, located in Madhya Pradesh, has shown an increase in the trend of ISMR since 1952 (Gajbhiye et al. 2016), Punjab state has also shown increased in the trend of ISMR for a 1901-2002 period (Krishan et al. 2015), while Wainganga basin in central India has shown a significant decreasing in ISMR since 1949 (Taxak et al. 2014).
The monsoon core zone has also shown a declining in ISMR rainfall in the recent decades (Kulkarni 2012;Guhathakurta et al. 2015). Northeast India, which is one of the homogeneous regions and has famous for its enormous rainfall, has experienced a significant reduction in rainfall after 1950 (Guhathakurta et al. 2015). A study on rainfall over different climatic zones of India based on the distribution of the Koppen climate reveals that after 1975, the region of tropical moist evergreen rain forest is becoming tropical dry land and semiarid dry climate becoming desert land (Rao et al. 2016). Recently, Ramarao et al. (2019) have illustrated that recent decades have experienced 10% expansion of the semiarid region over the Indian main landmass compared to the previous decade. A significant decreasing trend in ISMR is observed for northeast India; however, the significant increasing trend is observed for parts of west India (Krishan et al. 2015;Sharma and Singh 2017;Preethi et al. 2017;Barde et al. 2020).
Meanwhile, the well-known fact is that the long-term mean of ISMR remains unchanged (Mishra et al. 2012), which make us rethink whether it is possible that a region which has experienced low rainfall previously is getting more rainfall recently compare to its long period average (LPA) and other regions may experience a reduction in rainfall than corresponding regions LPA, and it would be interesting whether the famous "dry gets drier and wet gets wetter" paradigm is applicable to the Indian region or not.
So, it is of interest to examine the spatiotemporal rainfall variability for the regions which receive ample and scant rainfall over India. Keeping this in mind, we focused our study over the Indian landmass into two regions where rainfall occurrence is less than 40% (referred as a low rainfall zone; LRZ) and above 80% (referred as a heavy rainfall zone; HRZ) of LPA of all-India ISMR, to perceive the changes in rainfall pattern under the umbrella of climate change. The areas that receive different amounts of rainfall with respect to all-India ISMR including HRZ and LRZ are shown in Fig. 1a. In this study, we tried to figure out the change in ISMR variability of HRZ and LRZ by using the India Meteorological Department (IMD) daily rainfall datasets. We have carried out several statistical tests to confirm those results.

Data and methodology
IMD has developed the high-resolution (0.25° × 0.25°) daily rainfall analysis dataset (Pai et al. 2014) over India using varied station observations with time. In the preparation of the analysis dataset, the observations are not equally spread over India, about 1,450 station records have been used for 1901 and with the progress of time, and around 6,955 station records have been used in the recent period. Several studies have compared these datasets against the other available analysis datasets and validated with the station observations and confirmed that the quality of the high-resolution rainfall analysis of IMD is better than other available datasets (Pai et al. 2014;Nageswararao et al. 2016Nageswararao et al. , 2019aSharma and Mujumdar 2017). This dataset has also shown its potentiality in analyzing longterm changes in extreme rainfall events at met-subdivision, district/local scales (Swain et al. 2019(Swain et al. , 2020Nageswararao et al. 2019b).
In this study, we have used the same IMD daily rainfall dataset as discussed above for 116 years  during the ISM season, because ISM season receives an annual rainfall of about 80% (Bollasina 2014). Some standard homogeneity tests such as the Buishand range test (BR), Pettitt homogeneity tests (Wijngaard et al. 2003;Kang and Yusof 2012) have been applied to examine homogeneity in the rainfall distribution. Previous studies have reported that these methods were found to be robust in detecting the homogeneity (Deni et al. 2008;Kang and Yusof 2012;Goyal 2014;Nageswararao et al. 2019b). The null hypothesis has been set for the test as H 0 , data time series is consistent, and the alternate hypothesis is H 1 , data time series is not homogeneous. Further, the occurrence of consecutive dry days and various rainfall events based on IMD classification (Nageswararao et al. 2019a) has been examined for both the HRZ and LRZ regions. For a particular day, the occurrence of rainfall in between 0.1 and 2.5 mm is considered a very low rainfall (VLR), followed by 2.5-to 7.5-mm rainfall which is considered low rainfall (LR); 7.5-to 35.5-mm rainfall which is considered moderate rainfall (MR); 35.6-to 64.4-mm rainfall which is termed as rather heavy rainfall (RHR); 64.4-to 124.4-mm rainfall which is considered heavy rainfall (HR); 124.5 to 244.4 mm which is considered very heavy rainfall (VHR); and more than 244.4 mm which is considered extremely heavy rainfall (EHR). The day having rainfall less than 0.1 mm is considered a dry day.
Mann-Kendall trend test has been employed to explore the statistical significance of the trend for various rainfall events. Further, the dataset is divided into two halves (1901-1958 and 1959-2016) to examine the change in ISMR pattern over HRZ and LRZ. It is reported that the seasonal mean rainfall at all-India scale does not change significantly; however, notable changes in the extreme rainfall events have been noticed (Goswami et al. 2006); thus, it is indeed noteworthy to investigate the changes of the contribution of various rainfall events to the seasonal total. The contribution of the various rainfall events to the seasonal rainfall is calculated as follows: We have evaluated the recent changes in the contribution of rainfall by different events to the monthly and seasonal total during June-September (JJAS).
(1) events contribution = Total rainfall amount by a particular event in the season total seasonal rainfall × 100

Characteristics of heavy and low rainfall zone of India
The percentage distributions of various rainfall amounts have been shown in Fig. 1a; the rainfall more than 80%, i.e., HRZ, is shown in Fig. 1b; and less than 40%, i.e., LRZ, is shown in Fig. 1c. Out of 34 met-subdivisions over the Indian mainland, an average of 7-8 come under the HRZ category (b) the zones in blue color (shaded) represent the area of India where climatological rainfall is 80% or more (heavy rainfall zone: HRZ) with respect to all-India seasonal rainfall climatology; and (c) the zones in orange color (shaded) represent the area of India where climatological rainfall is 40% or less (Low rainfall zone: LRZ) with respect to all-India seasonal rainfall climatology; (d) contribution of various rainfall events to the seasonal rainfall for the HRZ and e contribution of various rainfall events to the seasonal rainfall for the LRZ. The computation has been carried out using India Meteorological Department (IMD) rainfall analysis at 0.25° × 0.25° horizontal resolution for the period of 116 years  and 13-14 are the LRZ category ( Fig. 1b and Fig. 1c). It may be noted that some met-subdivisions are experiencing both the categories, i.e., high and low rainfall; however, the locations are different. Therefore, we carry rainfall analysis based only on HRZ and LRZ regions instead of state or metsubdivision level. The contribution of various rainfall events to the seasonal rainfall is shown in Fig. 1d and Fig. 1e. It is seen that MR is the major contributing event for both HRZ (43%) and LRZ (48%). Apart from MR events, the heavy rainfall category rainfall events such as RHR, HR, VHR, and EHR have a substantial contribution in seasonal rainfall for the HRZ (Fig. 1d); however, LR and MR have more contribution to the seasonal rainfall for the LRZ (Fig. 1e).
To get a clear picture of the changes in ISMR distribution, it is very important to identify the area where the heavy or more intense rainfall events (hereafter referred to as HRm) and the total number of dry days (DD) are increased and/ or decreased. Here, we have examined four combinations with increasing/decreasing of HRm and DD. Figure 2 shows the spatial pattern of the different combinations for the increased/decreased in DD and HRm events over India and employed the Mann-Kendall test at 90% confidence level for the significant test. It may be noted that very limited study is available for combinations of DD and HRm events; that is why we choose 90% confidence level. It is found that the north Western Ghats which experienced the increase in seasonal rainfall also have experienced decreasing in HRm and increasing in DD; hence, the intense rainfall events are concentrated over a smaller region. Apart from that region, central Maharashtra, north-central parts of India, and some parts of northwest India also have experienced decreasing in HRm and increasing in DD (Fig. 2a). These findings are commensurable with the results presented in the previous study by Prathipati et al. (2019). Over central India (from east of Gujarat to Odisha), some parts of Uttarakhand, northeast India, and the south of Western Ghats, both HRm and DD have increased (Fig. 2b). Furthermore, along east coast, parts of the Indian peninsular regions, parts of northeast India, and western Rajasthan, HRm events are increased and DD events are decreased (Fig. 2c), while over south Andhra Pradesh coast, parts of Tamil Nadu, Arunachal Pradesh, and parts of Rajasthan, both DD and HRm events are decreased (Fig. 2d). It is also observed that major parts of the HRZ come under the increase in DD and decrease in HRm events (more than half of the regions) and very small parts are experiencing an increased HRm events (Fig. 2a,   Fig. 2 Spatial distribution of Mann-Kendall trends for heavy and more intense rainfall (HRm) events and total number of dry days at 90% of confidence level. a The regions where dry days increased and HRm decreased shown in red color, b the regions where both dry days and HRm events increased shown in yellow color, c the regions where dry days decreased and HRm events increased shown in green color, and d the regions where both dry days and HRm events decreased shown in blue color during 116 years period  Fig. 2b), while the major parts of the LRZ are experiencing an increase in HRm events (Fig. 2b, Fig. 2c). This implies that the HRZ and LRZ both are losing their identity based on changes in heavy or more rainfall events. Thus, it is interesting to analyze the change in overall seasonal rainfall for these regions.
For the single change-point detection, the Standard Normal Homogeneity Test (SNHT) is used to calculate the change-point detection at the starting and the endpoint of data series; however, the Buishand range (BR) test and Pettitt test are used to detect change point in the middle of the data series (Martinez et al. 2010;Nageswararao et al. 2019b). But SNHT shows more than one point for the inhomogeneity or change point because of various climatic or non-climatic factors and it gives a poor performance in detection of change point at the beginning of the series (Toreti et al. 2011). So, the BR test is mostly used for any distribution data (Taxak et al. 2014). Figure 3 shows the BR and Pettitt tests which are used for the single change-point detection and multiple change-point detections over the HRZ and LRZ. The HRZ shows less variability in data (Fig. 3a) and more fluctuations are observed for the LRZ (Fig. 3b) as data is not equally spread over the space, and the number of stations is also get changed with time. The year 1961 is a change-point year for HRZ and 1930 is for LRZ by using the BR test. The single change-point year by using Pettitt test in terms of change in mean rainfall is depicted in Fig. 3c and Fig. 3d. For the HRZ, 1961 is the change in mean rainfall year and the rainfall over this region get reduced after 1961 (Fig. 3c); similarly, for the LRZ, 1930 is a change in mean rainfall year, and rainfall is increased after this year (Fig. 3d).
The multiple change-point detection technique is used to find changes in rainfall with a small time interval (Fig. 3e and Fig. 3f). For the multiple point detection technique, two algorithms are used, offline and online. In the offline algorithm, the entire data is checked once, to find the change points within it, and the online algorithm (which is also known as real-time change-point detector) runs simultaneously with a process to check the change in the current point with the previous one (Aminikhanghahi and Cook 2017). In this study, the offline method is used to detect change points in the time series. The R software's change-point package is We found that 1920, 1930, 1950, 1952, and 1999 are change points for the mean rainfall over the HRZ (Fig. 3e), and 1905and , 1915and , 1917and , 1930and , and 1964) are the change points for the LRZ; interestingly, the year 1930 is a changing point similar with the single point detection as noticed in the previous section. It is also noted that after 1960, both regions, i.e., HRZ and LRZ, are experiencing the change in mean rainfall. To understand the change in the standardized rainfall anomaly for both regions, we have divided the study period into two equal halves, i.e., Period l (P1: 1901Period l (P1: -1958 and Period 2 (P2: 1959-2016). The midpoint of this division is near to the change point of the HRZ (1961 by single and multiple change-point detections) and LRZ (1964 by multiple change-point detection). Over the HRZ, the difference between changing points is larger compared to the LRZ.

Observed climate change over HRZ and LRZ
The special report on "Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation" of the Intergovernmental Panel on Climate Change (IPCC) explains possible factors that can be accounted for climate change. According to that report, climate change can be identified by using various statistical tests such as changes in the mean and changes in variability, and that change must be continued for decades or more (Field et al. 2012). The report also defines the climate extremes as "the occurrence of a value of a weather or climate variable above (or below) a threshold value which is near the upper (or lower) ends of the range of observed values of the variable" (Field et al. 2012). The probability density function (PDF) is a mathematical function and widely used to identify climate change (Field et al. 2012;Paeth et al. 2013;Matthews et al. 2016).
The PDF of seasonally averaged rainfall for each LRZ and HRZ has been computed for both periods to understand whether there has been any shifting in the mean rainfall (in millimeter) in the recent period compared to the earlier period. The Gaussian curves of PDF for ISMR during P1 and P2 have shown in Fig. 4. The mean rainfall of HRZ shows a shift to the left side in P2 and the distribution extends beyond 25th percentile of P1 which implies that the seasonal mean is decreased (Fig. 4a). At the same time, the flattered PDF curve in P2 indicates that the variance of ISMR over the HRZ is increased in P2 (Fig. 4a). On the other hand, the mean rainfall of LRZ has shifted towards the right side, suggesting that the mean seasonal rainfall is increased and variance is also slightly changed (Fig. 4b). In the recent period P2, the mean rainfall is decreased by 50 mm over HRZ and increased over LRZ by 10 mm. It is also noticed that in P2, climate extremes (excess and deficit) are increased over both zones; over HRZ, deficit rainfall condition is increased and over LRZ, excess rainfall condition is increased. It is also found that the number of excess (deficit) is increased (decreased) in P2 over the LRZ, while a reverse picture is depicted for the HRZ. The standardized rainfall anomaly for HRZ and LRZ is also computed (Supplementary Fig. 1). It is seen that 11 excess, 7 deficit, and 40 normal rainfall years have occurred during P1 and 8 excess, 13 deficits, and 37 normal monsoons took place during P2 over HRZ (Supplementary Fig. 1a). When analyzing the monsoon conditions Fig. 4 The probability density function of seasonal rainfall for (a) HRZ and (b) LRZ during P1 (1901-1958) and P2 (1959-2016). The blue color line shows the earlier period P1 and red line shows the recent period P2 for the LRZ during two periods, it is observed that 4 excess, 10 deficit, and 43 normal rainfall years have been ensued during P1, while 13 excess, 8 deficit, and 37 normal rainfall years are observed during P2 (Supplementary Fig. 1b). It is fascinating to notice that the total number of normal rainfall years is changed and a total number of climate extremes (excess + deficit) are increased over both regions during the recent period. Further, our findings are supported by Duncan et al. (2013), where they have reported that interannual ISMR variability showed a decreasing trend over the eastern (which is within the HRZ) and an increasing trend over the west and south India (which is within the LRZ) during 1951-2007 period.
The spatial distribution of the various rainfall events is presented in Fig. 5. The left and right columns represent the spatial distribution of various rainfall events for the HRZ (Fig. 5a, c, e, g, i, k) and the LRZ (Fig. 5b, d, f, h, j, l), respectively. The detailed analysis suggests that VLR and LR events have increased over northeast India (Fig. 5a, c), and LR to MR events are decreased over the HRZ (Fig. 5c,   Fig. 5 Spatial distribution of the difference of various rainfall events between P1 and P2. The hatched region shows value at 90% of confidence level, " + " used for an increase dry spell and " − " used for a decrease dry spell. The left column (a, c, e, g, i, and k) represents various rainfall events for the HRZ and right column (b, d, f, h, j, and l) represents various rainfall events for the LRZ e). On the other hand, over the LRZ, VLR and LR events are decreased in P2 (Fig. 5b, d), while MR to VHR events are increased over the LRZ (Fig. 5f, h, j). Recently, Prathipati et al. (2019) pointed out an inconsistency in the frequency of rainfall events over India during the ISM season. They have shown that the number of HR events is increased over most parts of central Maharashtra, Jammu and Kashmir, and west-central India, which are the parts of LRZ. The trends for various rainfall events are examined and employed the Mann-Kendall trend test and the results are presented in Table 1. It is found that the MR and RHR events are significantly decreased over the HRZ and significantly increased over the LRZ.
It is indeed interesting to understand the difference of dry spells having different time lengths between P1 and P2 for the HRZ and LRZ (Fig. 6). The difference between the total numbers of dry days between the two periods reveals an increase in dry days over HRZ (Fig. 6b) and a decrease   (1901-1958) and P2 (1959-2016) Fig. 6 Spatial distribution of the difference between total number of dry days and various dry spells between P1 and P2. The hatched region shows value at 90% of confidence level, " + " used for an increase dry spell and " − " used for a decrease dry spell. a Total number of dry days over HRZ, c one to two consecutive dry days (CDD) for HRZ, e short dry spell (3 and 4 CDD) for HRZ, g normal dry spell (5 to 7 CDD) for HRZ, i long dry spell (8 to 14 CDD) for HRZ, k very long dry spell (> 14 CDD) for the HRZ. b, d, f, h, j, and l are same as a, c, e, g, i, and k, respectively, but for the LRZ over northwest LRZ (Fig. 6b) in P2. Also, the trends of different consecutive dry days (CDD) are presented in Supplementary Fig. 2. The increase in the total number of dry days is a consequence of the decrease in seasonal rainfall over the HRZ; on the other hand, seasonal rainfall is increased and the total number of dry days has decreased over the LRZ. Several studies have shown that monsoonal rainfall has been decreased over India especially monsoon core zone due to land use land cover, reduction in moisture supply from the Bay of Bengal, and weakening of tropical easterly jet which are the consequences of global warming (Rao et al. 2004;Naidu et al. 2011;Kulkarni 2012;Panda and Kumar 2014;Paul et al. 2016). Here, we have found that the rainfall is reduced over the HRZ, and rainfall is increased over the LRZ. In P2, the major contributing events MR and HRm (HRm accounts combined effect of heavy rainfall events HR, VHR, and EHR) have decreased over the HRZ except for September (Fig. 7a,  b), and in reverse, all these events are increased over the LRZ (Fig. 7c, d). The decreased in percentage (%) of the contribution of MR and HRm events (Fig. 7) over the HRZ during JJAS is due to a decrease in MR and HR events almost every month in P2 (Supplementary Fig. 3). In June and September, the number of low-intensity rain events (VLR and LR) has decreased (more than 200 events) along with MR events (Supplementary Fig. 3); that is why the percentage (%) of the contribution of HRm events get increased (Fig. 2b). Over the LRZ, the MR and VHR events have increased for each month in P2 (Supplementary Fig. 3) which results in an increase in percentage (%) of the contribution of the MR and HR events at seasonal and monthly scales (Fig. 7c, d). We have also analyzed the contribution of various rainfall events to the seasonal rainfall in different decades and shown in Supplementary  Fig. 4. The percentage distribution of the seasonal rainfall increases (decreases) from the decade D4 (1931)(1932)(1933)(1934)(1935)(1936)(1937)(1938)(1939)(1940) onwards in LRZ (HRZ). Decade-wise analysis depicts the low-intensity events have increased over HRZ and the medium-intensity events MR and RHR have increased (decreased) over LRZ (HRZ). The difference in mean rainfall between HRZ and LRZ get reduced since 1901 ( Supplementary Fig. 5). Over the HRZ, low-and medium-intensity rainfall events have shown a decreasing trend and VHR and EHR have shown an increasing trend at 95% confidence level. Over the LRZ, the LR, MR, RHR, and VHR have shown an increasing trend at 95% confidence level. Overall, the HRZ is experiencing a decrease in seasonal rainfall and slightly increase in heavy rainfall events where LRZ is experiencing an increase in both seasonal and heavy rainfall events in P2.

Discussion
The part of HRZ, mostly an extended area of the Indian summer monsoon core zone, has experienced a reduction in ISMR in the warming environment (Kulkarni 2012;Roxy et al. 2017). Several studies have shown that the Arabian Sea Surface Temperature (SST) is increased during the last few decades (Roxy et al. 2014), and the increased Arabian SST have a good association with below normal rainfall of the Gangetic Plains (Mishra et al. 2012) which is also parts of HRZ. The occurrence of monsoon depressions over the Bay of Bengal is reduced and moves southward in the recent period (Vishnu et al. 2016) probably seed to the reduction in atmospheric moisture for rainfall over HRZ. At the same time, monsoon rainfall is reduced and shows a poleward shift in a warming climate (Sandeep et al. 2018). As a result of both, the HRZ containing the monsoon core region experienced a reduction in seasonal rainfall. The Western Ghats has also come under the HRZ region. The south of Western Ghats has shown a significant decrease in seasonal rainfall because of the weakening in southwesterly winds, slackened vertical velocity, weakening of mean meridional circulation, and northward movement of low-level jet (Rajendran et al. 2012;Varikoden et al. 2019). Moreover, Rajendran et al. (2012) Fig. 7 The percentage (%) of contribution of major contributing rainfall event: moderate rainfall (MR) and heavy rainfall events (HR) to seasonal as well as monthly basis, a MR events for the HRZ, b HR events for the HRZ, c MR events for the LRZ, d HR events for the LRZ have shown that the lapse rate is getting reduced because of increased temperature in the upper atmosphere compared to the lower atmosphere. Further, the presence of a large number of aerosols over the foothills of the Himalayas causes a dimming effect of the sun; the reducing surface temperature over north-central India causes the less convection rate and therefore reduction in seasonal rainfall over the HRZ. At the same time, Solmon et al. (2015) have shown that the increased in aerosol depth over the Arabian Sea and increase in rainfall over the southern peninsula are highly correlated. So, aerosols play an important role to change the pattern of rainfall over HRZ and LRZ.
We found that the long-term seasonal rainfall has shown a decreasing trend for the HRZ and an increasing trend for the LRZ (Fig. 4). After 1970, a significant increasing trend in seasonal rainfall is observed over Gujarat state and western districts of Maharashtra, which are the parts of LRZ (Guhathakurta and Saji 2013;Dave et al. 2017). A study of rainfall analysis of 30 met-subdivisions of India has shown that major parts of Punjab (a part of the LRZ), Haryana (a part of the LRZ), and coastal Katakana has significantly increased in rainfall whereas Chhattisgarh (a part of the HRZ) has significantly decreased in seasonal rainfall (Kumar et al. 2010).
There is a clear indication in the reduction of seasonal rainfall over the HRZ with an increase in rainfall variability in the recent era, while an increase in seasonal rainfall and heavy to extreme rainfall events is noticed for the LRZ in the recent era. The MR and RHR events play a key role in both increased seasonal rainfall over the LRZ and decreased seasonal rainfall over the HRZ. The VHR events are increased over both regions. Mishra et al. (2019) have shown that the temperature and heavy rainfall events are increased over India since 1900 and the light precipitation events are reduced in the warming temperature years. Our finding is supported by these reports as various rainfall events are increased over LRZ (Supplementary Fig. 3 and Fig. 4). When the air temperature is increased by 1 °C, the water holding capacity of the atmosphere is increased approximately by 7% (Trenberth 2011;Nageswararao et al. 2016), and the surface air temperature over northwestern and southern India (LRZ) is increased, and over central north India, it is decreased (HRZ) (Ross et al. 2018). So, more (lesser) atmospheric water is available for rainfall over LRZ (HRZ) during the recent period compared to the earlier period. Recently, Saha et al. (2017) have shown that east-west coastal asymmetry is observed for the summer time near surface wind.
The correlation between El Niño and below normal rainfall is one of the signs of seasonal rainfall over India (Yadav et al. 2013;Yadav and Roxy 2019); recently, Equatorial Indian Ocean Oscillation (EQUINOO), an atmospheric circulation, sets a new relation with Indian seasonal rainfall.
The positive period of EQUINOO enhances the convection over the western Indian Ocean and suppresses convection over the eastern Indian Ocean and it has a good correlation with ISMR. During the positive period of EQUINOO, India has experienced normal to excess rainfall and the situation becomes opposite in the negative phase (Surendran et al. 2015;Gadgil et al. 2019).
The Indian Ocean Dipole (IOD) is an oceanic phenomenon, calculated by the difference of western Indian Ocean (WIO) SST and eastern Indian Ocean SST; the positive phase of IOD (pIOD) causes the convection over WIO and good rainfall over the Indian subcontinent (Ashok et al. 2001;Saji and Yamagata 2003). Cai et al. (2014) have shown that during extreme pIOD events, westerlies and eastward flowing upper ocean currents reverse that cause anonymous dry condition over central and eastern Indian and strong convection over the westward side which triggers the flood-like conditions in eastern African countries. We found that the excess rainfall events are also increased over LRZ, which is the west side of India. They have also reported that in the warming climate, extreme pIOD events will likely occur more frequently. The frequent occurrence of deficit rainfall events over the HRZ and heavy rainfall events over the LRZ confirms the result presented by Cai et al. (2014). In the 2019 ISM season, LRZ has experienced a greater number of flood events and a large part of the HRZ experienced below normal rainfall conditions till mid-September. This shows the rainfall distribution slowly changes and getting reverse over India; LRZ is experiencing a positive deviation from LPA and HRZ is experiencing a negative deviation from LPA.
The understanding of the reverse in variability is important to water resources, flood and drought management and policy-making, and model developers of the country. The results are highly helpful to the policy planner for making adaptation strategies in the context of changing climate.

Conclusions
The analysis of seasonal rainfall for heavy rain zone (HRZ) and less rainfall zone (LRZ) is carried out by using 116 years (1901-2016) IMD high-resolution (0.25° × 0.25°) gridded rainfall analysis dataset for the summer monsoon season. The Indian main landmass is divided into two zones based on the amount of seasonal rainfall, the HRZ, and the LRZ. The HRZ area which receives rainfall higher than 80% of all-India LPA and LRZ is the area that receives rainfall lesser than 40% of all-India LPA. By using different single change-point detection techniques, the pattern of rainfall gets changes during 1930 and 1961 onwards over the LRZ and HRZ, respectively. The mean rainfall is decreased over the HRZ and from 1964 over the LRZ since 1960. The HRZ has experienced an increase in both dry days and heavy rainfall events; while the LRZ has experienced an increase in heavy rainfall events and decrease in dry days.
Further, the dataset is divided into two equal halves Pl (1901-1958) and P2 (1959-2016) (which is nearest to the change point for HRZ and LRZ) to investigate the changes in the spatiotemporal distribution of rainfall. It is found that the HRZ has experienced more excess rainfall years in Pl and more deficit rainfall events in P2. Conversely, the LRZ has experienced a greater number of excess rainfall years in P2. The seasonal rainfall analysis for two zones (HRZ and LRZ) of India for the last century has shown that the "dry gets drier and wet gets wetter" paradigm is not absolute over India, particularly for the summer monsoon season, as HRZ (LRZ) has experienced a reduction (an increase) in seasonal rainfall. The number of moderate and heavy rainfall events (LR, MR, RHR, HR, and VHR) reduces over the HRZ as their contribution is largely important to the seasonal total rainfall. The seasonal rainfall has increased over the LRZ because of an increase in number of low, moderate, and heavy rainfall events. The extreme rainfall events are increased over both regions.
The understanding of the effect of this change is important as the HRZ contains a large number of tributary areas, a large number of rivers including India's one of the biggest dams-Hirakud-and large agricultural area. Similarly, the LRZ is also not fully prepared for frequent extreme rainfall events and necessary action plan may be needed for the LRZ region for adaptation under the scenario of the climate change. In addition, almost one third of the districts which is more than half of the agricultural land is predominantly rainfed regions and the agriculture solely depends on the summer monsoon rainfall. Thus, there are needs for proper planning of the agricultural practice and suitable adaptation strategies for both the HRZ and LRZ regions.