Solar cycle effects on Indian summer monsoon dynamics

Journal of Atmospheric and Solar-Terrestrial Physics ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Contents lists available at ScienceDirect

Journal of Atmospheric and Solar-Terrestrial Physics journal homepage: www.elsevier.com/locate/jastp

Solar cycle effects on Indian summer monsoon dynamics M. Venkat Ratnam a,n, Y. Durga Santhi b,1, P. Kishore c, S. Vijaya Bhaskara Rao b a b c

Department of Physics, S. V. University, Tirupati 517502, India National Atmospheric Research Laboratory, Department of Space, Gadanki, India Department of Earth System Science, University of California, Irvine, CA, USA

art ic l e i nf o

a b s t r a c t

Article history: Received 2 December 2013 Received in revised form 10 June 2014 Accepted 23 June 2014

Solar activity associated with sunspot number influences the atmospheric circulation on various time scales. As Indian summer monsoon (ISM) is the manifestation between warmer Asian continent and the cooler Indian Ocean, changes in the solar cycle are expected to influence the ISM characteristics. Among several elements of ISM, Tropical Easterly Jet (TEJ), Low Level Jet (LLJ), and rainfall are important features. As a part of CAWSES India Phase II theme 1 (solar influence on climate (0–100 km)) programme, we made an attempt to investigate the role of solar cycle variability on these ISM features using long-term data available from NECP/NCAR (1948–2010) and ERA-Interim (1979–2010) re-analysis products. To check the suitability of these data sets, ground based observations available over the Indian region are also considered. ISM characteristics are studied separately for the maximum and minimum as well as increasing and decreasing solar cycle conditions. Amplitudes corresponding to the solar cycle observed in TEJ, LLJ and rainfall are extracted using advanced statistical tool known as intrinsic mode function. Longterm trends in TEJ reveal decreasing trend at the rate of 0.13 m/s/yr (between 1948 and 2000) and no perceptible trend in LLJ. There exists inverse relation between TEJ strength and Central India rainfall. Large difference of 2 m/s (5 m/s) in the zonal winds of TEJ between solar maximum and minimum (increasing and decreasing trend) is noticed. There exists a difference of
2 m/s in LLJ winds between solar maximum and minimum and increasing and decreasing trend of the solar cycle. However, no consistent relation between the ISM rainfall and solar cycle is noticed over Indian region unlike reported earlier but there exists a delayed effect around 13 years. We attribute the observed features as linear and non-linear relation between dynamics of ISM, rainfall and solar cycle, respectively. & 2014 Elsevier Ltd. All rights reserved.

Keywords: Solar cycle Monsoon Tropical Easterly Jet Low Level Jet

1. Introduction The Indian summer monsoon (ISM), which is a part of a large scale circulation pattern of Asian Summer monsoon (ASM), develops due to the differential heating between the warm Asian continent and cooler Indian Ocean. The strong southwesterly winds in the lower troposphere brings significant amount of moisture to the Indian subcontinent, which is released as precipitation. This rainfall plays a very important role in agricultural productions and economical conditions, thus its variability and long-term trends is of great interest. ISM is characterized by a few important features in the troposphere such as seasonal wind reversal Low Level Jet (LLJ) stream in the lower level and Tropical Easterly Jet stream (TEJ) in the upper level. TEJ is an important parameter of the ISM circulation which can be seen in the upper tropospheric levels (100–150 hPa) during the ISM months of June to September n

Corresponding author. Tel.: þ 91 8585 272123; fax: þ 91 8585 272018. E-mail address: [email protected] (M.V. Ratnam). 1 Expired on 30 September 2013.

(Krishnamurti and Bhalme, 1976). This jet has wind speed of roughly 40–50 m/s with the strongest winds being found in northern hemispheric summer over the Arabian Sea (Reiter, 1961). The study of these features is important as they reveal the strength of the monsoon and its year-to-year variability. To the present knowledge, many specific changes during the monsoon have been identified on long-term scales and based on these, forecasts are being made. However, many issues remain unexplained and monsoon is playing on its own way. Though the monsoon system is well understood at present, but it is evident that the effect of solar cycle on monsoon climate is least explored. Since many years the influence of solar activity on climate is on debate. Recent advances in reconstruction of the past climate with fine temporal resolution clarified the relationship between the solar cycles and the monsoon rainfall in South Oman with multiple time scales from decadal to millennial and the direct cause of higher rainfall in South Oman was explained by stronger northward surface winds (Neff et al., 2001; Burns et al., 2002; Fleitmann et al., 2003). The multidecade to century scale variations in the monsoon winds were much larger in the early Holocene

http://dx.doi.org/10.1016/j.jastp.2014.06.012 1364-6826/& 2014 Elsevier Ltd. All rights reserved.

Please cite this article as: Ratnam, M.V., et al., Solar cycle effects on Indian summer monsoon dynamics. Journal of Atmospheric and Solar-Terrestrial Physics (2014), http://dx.doi.org/10.1016/j.jastp.2014.06.012i

2

M.V. Ratnam et al. / Journal of Atmospheric and Solar-Terrestrial Physics ∎ (∎∎∎∎) ∎∎∎–∎∎∎

coincident with increased sunspot numbers (Gupta et al., 2005). A possible association between Indian monsoon rainfall and solar activity was reported by Bhattacharyya and Narasimha (2005, 2007). Study by Agnihotri et al. (2011) indicates that anomalous dry periods of the Indian monsoon are coincident with negative total solar irradiance (TSI). Note that understanding the influence of solar variability on the Earth's climate requires knowledge of solar variability, solar– terrestrial interactions, and the mechanisms determining the response of the Earth's climate system which can be associated with variations in ozone, temperatures, winds, clouds, and precipitation (Gray et al., 2010). Numerous studies have been established to identify the influence of solar cycle on various atmospheric parameters like sea surface temperature (SST), sea level pressure (SLP), rainfall and tropospheric temperatures (Reid, 1987, 1991; Haigh, 1996; Hiremath and Mandi, 2004; Gray et al., 2010; Roy and Haigh, 2010). Long term variations in Earth's temperature are closely associated with variations in the solar length, and it closely matches with the long term variations in the land temperatures in the Northern Hemisphere (Christensen et al., 1991). As the major source for the ISM to occur is the thermal gradient between the Asian continent and the Indian Ocean, changes in solar cycle are expected to influence the monsoon winds correspondingly. Solar variability has to be carefully monitored in decadal to multi-decadal timescale, in conjunction with other greenhouse parameters, in order to achieve predictive capability of monsoon rainfall (Agnihotri and Dutta, 2003). Reddy et al. (1989) found that there is a correlation between rainfall and solar cycle with phase delay of 0.16 year (solar cycle leads). The spring and southwest monsoon rainfall variabilities are positively correlated with the sunspot activity (Hiremath and Mandi, 2004). It was also proposed that solar influence on monsoon activity is not due to a change in radiative heating in the troposphere but rather originates from the stratosphere through modulation of the upwelling in the equatorial troposphere, which produces a north–south see-saw of convective activity over the Indian Ocean sector during summer. Increase in solar activity resulting increased in precipitation over Arabia and India (Kodera, 2004). Larger precipitation is found along the western Pacific Ocean and reduced over equatorial Indian Ocean due to the high solar activity (Claud et al., 2008). Solar cycle influences the mean meridional circulations (Hadley and Brewer–Dobson circulations) in northern summer especially during the east phase of the Quasi-Biennial Oscillation (Labitzke, 2003). Although the observational data suggests that solar activity has influenced temperature on decadal, centennial and millennial timescales; however there could be some other sources co-exists which need to be differentiated from other factors such as volcanic eruptions and the El Niño Southern Oscillation (ENSO) (Haigh, 2003). Using a multiple linear regression analysis Roy and Haigh (2010) identified solar cycle signals in 155 years of global SLP and SST in the Northern Pacific and found a statistically significant weakening of the Aleutian Low and a northward shift of the Hawaiian High in response to higher solar activity in SLP and a weak El Niño like pattern in the tropics in SST. The solar response shows largest warming in the stratosphere and bands of warming of more than 0.4 K throughout the troposphere in the midlatitudes. In this present study we made an attempt to investigate the role of solar cycle variability on ISM features like TEJ, LLJ and rainfall. As mentioned earlier, the common feature that can be observed during ISM which occurs at 16 km (
100 hPa) is the TEJ and it affects the formation of storms during this season. By considering about 50 years data up to 1998, Sathiyamoorthy (2005) has shown that the strength of TEJ has decreased in the recent years. But recent results by Raman et al. (2009) and Venkat

Ratnam et al. (2013) showed that there exists an increase in the strength of TEJ from 2000 onwards at about 1 m/s/year. We further extend the study made by them to explore the variability of TEJ characteristics and monsoon rainfall over India due to the solar cycle using advanced statistical tool like Intrinsic Mode Function (IMF). Similar analysis has been extended for the LLJ that occurs at
850 hPa during ISM months and on the rainfall.

2. Database 2.1. NCEP/NCAR and ERA-Interim reanalysis datasets The 64 years (1948–2010) monthly mean zonal wind data from National Centers for Environmental Prediction (NCEP)/National Center for Atmospheric Research (NCAR) reanalysis data available with 2.5° latitude  2.5° longitude (Kalnay et al., 1996) grid on 17 pressure levels from 1000 hPa to 10 hPa has been used to study the spatial distribution and long-term trends in TEJ and LLJ strength. In addition, European Centre for Medium-Range Weather Forecasts (ECMWF) Interim (ERA-Interim) reanalysis data available with 1.5° latitude  1.5° longitude with 37 vertical pressure levels from 1000 hPa to 1 hPa from 1979 onwards is also used to verify the results obtained by relatively less vertical and horizontal resolution NCEP/NCAR data. It is worth to mention here that there is a concern about the use of reanalysis data for calculating climate trends as they can overestimate/underestimate the true trends (Bengtsson et al., 2004). Note that in case of ERA-Interim, warming of the lower stratosphere by approximately 0.2 K in December 2006 with the introduction of GPS radio occultation data from the COSMIC constellation is being reported and slight excess warming of upper-tropospheric temperatures due to the assimilation of growing numbers of warm-biased temperature measurements from aircraft, beginning in 1999 (Dee and Uppala, 2009). After December 2006, this drift is somewhat reduced with the introduction of GPS radio occultation data from the COSMIC constellation. This might have some effect on the observed winds only in the latter half of the recent decade as gradient in temperature result in the winds. In the case of NCEP, the final product from this model again put through the quality check and assimilated with a data assimilation system kept unchanged over the reanalysis period. Thus, we are not expecting a big change except in the last decade as in the case of ERA-Interim. However, in the present study we restrict to variation in the winds from these reanalysis data which are more than two-sigma variations. Further we also validate the use of reanalysis from one of the location using high resolution ground based observations. 2.2. MST radar and GPS radiosonde observations Data available from high vertical resolution MST Radar located at Gadanki (13.5°N, 79.2°E) since 1996 has also been considered to verify the accuracy of reanalysis data sets. The Indian MST radar is operated almost daily for about half-an-hour around 1730 LT (LT¼UTþ 0530 h) with vertical resolution of 150 m and accuracy of 0.1 m/s (horizontal winds). MST radar located quite over the core of the TEJ grid provides an opportunity to observe the monsoon features like TEJ with high-vertical resolution (Narayana Rao et al., 2000; Ghosh et al., 2001; Vasantha et al., 2002; Raman et al., 2009; Venkat Ratnam et al., 2013). High-resolution radiosonde observations available daily from the same location during the period of April 2006 to December 2011 is also used to verify the LLJ variability observed in reanalysis data sets. Most of these radiosondes were launched around 1730 LT. In addition to these data sets, we also make use of the gridded (1°  1°) rainfall data obtained from India Meteorological

Please cite this article as: Ratnam, M.V., et al., Solar cycle effects on Indian summer monsoon dynamics. Journal of Atmospheric and Solar-Terrestrial Physics (2014), http://dx.doi.org/10.1016/j.jastp.2014.06.012i

M.V. Ratnam et al. / Journal of Atmospheric and Solar-Terrestrial Physics ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Department (IMD) (Rajeevan and Bhate, 2009) available since 1901 and the monthly mean sunspot numbers (SSNs) information available from NOAA website during 1901–2010. Since rainfall data is available from 1901, we also obtained zonal wind data from 1901 onwards from 20th century re-analysis data sets (Compo et al., 2011).

3. Results and discussion 3.1. Characteristics of TEJ and LLJ Before going into the details on the solar cycle effects on ISM features, the background mean zonal wind and atmospheric conditions over the study region are presented which will be useful while interpreting the ISM characteristics. As mentioned earlier ISM is characterized by few important features such TEJ, LLJ and rainfall. TEJ can be effectively seen in the upper troposphere in the northern hemispheric summer (Koteswaram, 1958), generally between the pressure levels of 100 hPa and 150 hPa (Raghavan, 1973) with core wind speed of the order of 30–40 m/s. LLJ core lies at 850 hPa, which corresponds at a height of 1.5 km MSL, with speeds of the order of 20–30 m/s though it was reported that it can extend up to 710 hPa over some regions (Raman et al., 2011). Note that the speed of TEJ and LLJ varies considerably within the monsoon season. The zonal wind speed is higher in the peak monsoon months of July and August and lower in June and September. Thus, we have examined the prevailing wind condition over Gadanki region during peak monsoon months (July and August) in the lower and upper troposphere using the combined observations of MST radar and radiosonde observations. The climatological mean zonal wind averaged for the months of July and August during 2000–2010 using the combined observations of MST radar and radiosonde over Gadanki is shown in Fig. 1a. The upper wind structure over Gadanki region shows a deep belt of westerlies extending up to the mid-troposphere. Further aloft, there is an easterly flow extending up to the tropopause (
100 hPa). During ISM months the low level winds

3

are westerlies, gradually increases and attain maximum velocity of 7–10 m/s at 710 hPa. It is found that, on average, the LLJ exists at 710 hPa over southeastern peninsular India (Raman et al., 2011) rather than at 850 hPa. The zonal wind velocity decreases gradually and attains maximum speed of 30–35 m/s at 100 hPa. The 32-year (1979–2010) climatology of the zonal wind obtained from ERA-Interim reanalysis averaged during the peak summer monsoon months of July and August at 100 hPa, 850 hPa and 700 hPa representing the spatial distribution of TEJ, LLJ in general, and LLJ in particular to the southern India are shown in Fig. 1b, c, and d, respectively. Several studies have been made earlier to explain the characteristics of the TEJ (Koteswaram, 1958, 1960; Reiter, 1961). According to the WMO definition, the jet stream is a narrow fast moving air current with wind speed exceeds 30 m/s. Note that zonal wind has crossed 30 m/s as highlighted in Fig. 1b and Gadanki location is in the core of TEJ. The peak jet speed is located around 10–20°N and 60–90°E during the period mentioned above. Although reanalysis data sets represent well the spatial variability, they underestimate the peak zonal winds by about 5 m/s when compared to Indian MST radar and radiosonde observations (Raman et al., 2009). From Fig. 1c, it can be observed that the LLJ at 850 hPa is located around 10–15°N and 50–70°E with maximum wind speed of 20 m/s. Kalapureddy et al. (2007) observed that LLJ varies in the height range of 1.8 þ0.6 km with mean jet intensity of 20 m/s using Lower Atmospheric Wind profiler (LAWP) over Gadanki. Recently Raman et al. (2011) have given the detailed characteristics of LLJ over Gadanki with mean core height lying around 3 km, corresponding pressure level of 710 hPa using high resolution radiosonde observations. To see the detailed characteristics of TEJ and LLJ and their year-to-year variability, we have selected the peak zonal wind over the grid 10–20°N and 60–90°E at 100 hPa and 10–15°N; 50–70°E at 850 hPa and 700 hPa. The time series of zonal winds from NCEP/ NCAR (ERA-Interim) reanalysis in peak monsoon season (July– August) averaged over above mentioned grids for the period 1948– 2010 (1979–2010) and the combined observations of radiosonde and MST radar over Gadanki at 100 hPa, 700 hpa and 850 hPa are shown in Fig. 2a, b and c, respectively. In general, reanalysis data

Fig. 1. (a) Climatological mean profile of zonal wind (averaged during July and August from 2000 to 2010) over Gadanki observed using combined observations of MST radar (2000–2005) and GPS radiosonde (April 2006–December 2010). Spatial distribution of 32-year (1979–2010) mean zonal wind at (b) 100 hPa, (c) 850 hPa and (d) 700 hPa obtained from ERA-Interim reanalysis. The location of Gadanki is shown with filled circle.

Please cite this article as: Ratnam, M.V., et al., Solar cycle effects on Indian summer monsoon dynamics. Journal of Atmospheric and Solar-Terrestrial Physics (2014), http://dx.doi.org/10.1016/j.jastp.2014.06.012i

4

M.V. Ratnam et al. / Journal of Atmospheric and Solar-Terrestrial Physics ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Fig. 2. The time series of zonal winds observed (a) at 100 hPa using NCEP/NCAR, ERA-Interim reanalysis and combined observations of MST radar and Radiosonde during peak monsoon months (July and August) averaged over 10–20°N; 60–90°E for the period 1948–2010. (b) and (c) same as (a) but for 850 hPa and 700 hPa, respectively, observed over 10–15°N; 50–70oE. Magenta color line shows linear fit (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.).

sets of NCEP/NCAR and ERA-Interim shows the decreasing trend in TEJ at 100 hPa over the grid, though a slight increasing trend in TEJ is noticed from the period of 2000–2010 over Gadanki. The decreasing trend in the strength of TEJ is also noticed by Rao et al. (2004) using the NCEP/NCAR reanalysis and the radiosonde data over India. The rate of decreasing velocity in TEJ over Gadanki is about 0.13 m/s/yr. Using same data Sathiyamoorthy (2005) reported that the zonal extent of the TEJ has reduced between 1960s and 1990s particularly over the Atlantic and African regions. Note that the decrease in wind speed in NCEP/NCAR reanalysis is more than the ECMWF reanalysis. In addition to this, an increasing trend in TEJ (1 m/s/yr) in peak jet speed is observed particularly from the year 2000 over Gadanki region from the combined observations of MST radar and radiosonde (Raman et al., 2009) which was discussed in more detail by Venkat Ratnam et al. (2013). Similar decreasing trend of about 0.04 m/s/yr in the LLJ at 850 hPa over the Gadanki grid can be noticed from Fig. 2b. No significant trend is seen in LLJ at 850 hPa in the peak zonal wind grid. However, it is interesting to see an increasing trend in the LLJ at 700 hPa over the peak zonal wind grid from 1979 onwards which is about 0.1 m/s/yr (Fig. 2c). Interestingly, the ground measurements match well with that reported using reanalysis data sets in the overlapping periods. The details on rate of change in zonal wind speed at different pressure levels are given in Table 1.

3.2. Relation between TEJ and rainfall Gridded rainfall data available from IMD is obtained from 1901 to 2010, though zonal winds from NCEP is available from 1948 only. In order to consider full time series and also to check how TEJ behaves in the first half of the century, we have considered 20th century re-analysis (20CR) products. The time series of zonal winds observed at 100 hPa using 20CR products during peak monsoon months (July and August) averaged over 10–20°N; 60– 90°E for the period 1901–2010 is shown in Fig. 3. 11-year smoothing is applied and is shown with thick line. This new data set also show similar decreasing trend in TEJ during second half of the century (1950 onwards) similar to that observed by NCEP and ERA-Interim, however, this trend stopped around 1980. It is surprising to see an increasing trend in the TEJ intensity during the first half of century (1901–1950) at a rate of 0.13 m/s/yr similar to the decreasing tend observed in latter half. There exists nearly 60 years oscillation and this demands separate investigation. Central India (16.5–26.5N, 74.5–86.5°E) rainfall obtained by IMD during the same period superimposed in this figure shows an inverse relation between the two i.e., higher the TEJ intensity, lower the rainfall. The correlation between TEJ intensity and Central India rainfall shows  0.43 with standard deviation of 0.85. Similar analysis is made while considering All India rainfall during peak monsoon months of July–August and correlation is found to be  0.12 with standard deviation of 0.88. When we

Please cite this article as: Ratnam, M.V., et al., Solar cycle effects on Indian summer monsoon dynamics. Journal of Atmospheric and Solar-Terrestrial Physics (2014), http://dx.doi.org/10.1016/j.jastp.2014.06.012i

M.V. Ratnam et al. / Journal of Atmospheric and Solar-Terrestrial Physics ∎ (∎∎∎∎) ∎∎∎–∎∎∎ Table 1 Long-term trends observed in the zonal winds at 100, 150, 200, 850 and 700 hPa over the different girds mentioned in the first row during different Indian summer monsoon months. Month

July August

10–20°N; 60–90°E (rate of change in velocity (m/s/yr)

10–15°N; 50–70°E (rate of change in velocity (m/s/yr)

100 hPa

150 hpa

200 hPa

850 hPa

700 hPa

 0.138 7 0.015  0.138 7 0.017  0.13 7 0.014

 0.0154 7 0.012  0.027 0.012  0.017 7 0.01

0.02 7 0.009 0.014 7 0.0096 0.017 7 0.008

0.006 7 0.008  0.00386 7 0.009 0.00107 7 0.007

0.016 7 0.0099 0.0127 7 0.0086 0.0144 7 0.0069

 0.11 7 0.01  0.13 7 0.017  0.12 7 0.013

 0.01 7 0.013  0.03 7 0.013  0.02 7 0.011

0.02 7 0.0107 0.02 þ 0.0101 0.01 7 0.0086

 0.0425 7 0.013  0.0367 7 0.012  0.0396 7 0.0089

 0.0042 7 0.013  0.0054 7 0.012  0.0049 7 0.008

July and August mean Over Gadanki (rate of change in velocity( m/s) per year) July August July and August mean

5

separate the study between first half (1901–1950) and second half (1951–2000) century, the correlation between TEJ and Central India (All India) rainfall is found to be 0.48 (  0.23) with standard deviation of 0.83 (0.90) and  0.40 ( 0.03) with standard deviation of 0.88 (0.89), respectively. Thus, TEJ has inverse relation with Central India rainfall. Since the main interest in the present communication is see the solar cycle effects on the LLJ, TEJ and rainfall, we will concentrate in extracting the amplitudes associated with this in the following sub-sections. 3.3. Sunspot variations To represent the solar activity, several types of indicators have been used, which includes sunspot number, solar diameter, solar radio flux at 10.7 cm and geomagnetic activity index (aa). The sunspot number is one of the best known indicators of the solar activity, which has been recorded since the early 17th century. The monthly mean variation of solar activity associated with sunspot numbers during the period of 1958–2010 is shown in Fig. 4. It is well known that the average period of solar cycle is 11 years though it varies in length and amplitude from one cycle to other. It was already shown that the 11 years solar cycle significantly

Fig. 3. The time series of zonal winds observed at 100 hPa using 20th century re-analysis products during peak monsoon months (July and August) averaged over 10–20oN; 60–90oE for the period 1901–2010. Central India rainfall obtained by IMD during same period is also superimposed. Axis for rainfall is shown in the right. 11 years running means are shown with thick lines.

Fig. 4. Monthly mean time series of sunspot numbers observed during the period of 1948–2010. Maximum and minimum sunspot number years are shown above and below the line of 100 and 40, respectively. The period between sunspot minimum and maximum (black arrows) and between maximum and minimum (red arrows) sunspot years are considered as increasing and decreasing sunspot years, respectively (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.).

Please cite this article as: Ratnam, M.V., et al., Solar cycle effects on Indian summer monsoon dynamics. Journal of Atmospheric and Solar-Terrestrial Physics (2014), http://dx.doi.org/10.1016/j.jastp.2014.06.012i

M.V. Ratnam et al. / Journal of Atmospheric and Solar-Terrestrial Physics ∎ (∎∎∎∎) ∎∎∎–∎∎∎

6

influences the atmospheric parameters such as land temperatures, SST, atmospheric temperatures etc., (Reid, 1987, 2000; Haigh, 1996). In order to differentiate the effect of the solar activity on ISM features, the magnitude of sunspot number is taken into consideration. After careful analysis, we considered the sunspot number exceeding 100 as maximum solar cycle years (solar maxima) and less than 40 as minimum solar cycle years (solar minima) as shown in Fig. 4. Note that in this figure we considered from the year 1949 onwards where we have observations on TEJ and LLJ. The increasing and decreasing trend of years are considered when the sunspot numbers lies in 40 and 100. This analysis is quite different with that done by Roy and Haigh (2010), where they removed climatological mean from the maximum solar cycle condition by keeping cut off in SSN as 80. Since the cut-offs have some time both sides, the effects of lag/lead will also come into it and thus, the cut-offs used in the present study are more robust to identify the solar maximum and minimum periods. The corresponding years of sunspot number maximum, minimum, increasing trend and decreasing trend of solar cycle Table 2 Years that fall under the category of solar cycle increasing trend, decreasing trend, maxima and minima during 1948–2010. Total number of years that fall under each category is also mentioned. Increasing trend years Decreasing trend years

Maximum sunspot years

Minimum sunspot years

1967,1978, 1988, 1998, 1999 Total: 5 1952, 1961, 1962, 1971, 1972, 1973, 1983, 1984, 1993, 1994, 2003, 2004 Total: 12 1948, 1949, 1956, 1957, 1958, 1959, 1960, 1967, 1968, 1969, 1970, 1980, 1981, 1982, 1989, 1990, 1991, 1992, 2000, 2001, 2002 Total: 21 1953, 1954, 1955, 1962, 1963, 1964, 1965, 1966, 1973, 1974, 1975, 1976, 1977, 1985, 1986, 1987, 1995, 1996, 1997, 2005, 2006, 2007, 2008, 2009, 2010 Total: 23

years are given in Table 2. For present analysis purpose we considered peak months of July and August. 3.4. Solar cycle effects on ISM features To investigate the solar cycle effects on ISM features, the zonal wind at 100 hPa (TEJ) is averaged separately for the years (for July and August months) during the solar maximum, solar minimum, increasing trend and decreasing trend. The spatial distribution of mean zonal wind at 100 hPa obtained using ERA-Interim averaged during the increasing trend and the maximum solar cycle is shown in Fig. 5(a) and (b), respectively. In general, Fig. 5(a) reveals that the magnitude of wind speed is high reaching 40 m/s between 10– 20°N and 60–90°E (core region) in the increasing trend of solar cycle. Difference in the zonal wind between increasing and decreasing trend years shown in Fig. 5(c) reveals that the wind speed decreases during decreasing trend of solar cycle. During decreasing trend TEJ core has reduced considerably in its size (figure not shown). The difference in zonal wind speed in TEJ is noticed about 3 m/s in the jet core region during increasing and decreasing trend of solar cycle. The difference in the zonal wind between the maximum and minimum solar cycle shown in Fig. 5 (d) reveals no major difference in wind speed (only about 2 m/s is noticed). The difference in the zonal wind at 100 hPa between increasing trend and decreasing trend of solar cycle, solar maxima and minima suggest that maximum decrease has taken place in the core region over the peninsular India. Christiansen et al. (2007) showed the signals in response to the solar cycle in zonal mean temperatures and winds throughout the troposphere and stratosphere and in all seasons using NCEP/NCAR reanalysis data sets. The annual mean signal in zonal wind shows negative anomaly in low latitudes, representing a weakening of the tropospheric jets. For verification, similar analysis is performed using NCEP/NCAR reanalysis (figure not shown). Interestingly, no big difference in the spatial extent of TEJ is seen between increasing and maximum solar cycles. Further the difference in the magnitude between the increasing and decreasing trend in the solar cycle in ERA-Interim

Fig. 5. Spatial distribution of zonal wind at 100 hPa (averaged over July and august) observed using ERA-Interim reanalysis during (a) increase in sunspot years, (b) maximum sunspot years for the period of 1979–2010. The difference between (c) increasing and decreasing trend of sunspot number years and (d) maximum and minimum sunspot years. (e) and (f) same as (a) and (b) but after removing ENSO years. (g) and (h) same as (c) and (d) but after removing ENSO years.

Please cite this article as: Ratnam, M.V., et al., Solar cycle effects on Indian summer monsoon dynamics. Journal of Atmospheric and Solar-Terrestrial Physics (2014), http://dx.doi.org/10.1016/j.jastp.2014.06.012i

M.V. Ratnam et al. / Journal of Atmospheric and Solar-Terrestrial Physics ∎ (∎∎∎∎) ∎∎∎–∎∎∎

and NCEP/NCAR is similar. However, big difference exists in the magnitude as well as spatial extent in maximum and minimum solar cycle conditions between ERA-Interim and NCEP/NCAR reanalysis data sets (figure not shown but will discuss again later). The difference in the magnitude between increasing and decreasing trend in the solar cycle is restricted to Indian region only (
60–90°E) but the difference extends to complete longitude range (60–100°E). Thus, it may be concluded that TEJ is fundamentally affected by the solar cycle though amplitudes differ among the various reanalysis data sets. Further, it is well known that ENSO will have strong influence on these dynamical features and hence its contribution has to be removed from the analysis. For this ENSO index provided by NOAA is utilized and the months falling in the ENSO years have been removed and is depicted in Fig. 5(e) and (f) for increasing trend years and maximum sunspot years, respectively. A difference of more than 5 m/s can be noticed more clearly from Fig. 5(g) where difference between increasing and decreasing trend in the sunspot number years after removing ENSO is plotted. It is interesting to see increase in the wind speed of about 2 m/s between maximum and minimum sunspot years after removing ENSO years (Fig. 5h). Similar analysis has also been extended for LLJ at 850 hPa and is shown in Fig. 6. Maximum eastward winds of about 12–18 m/s in ERA-Interim reanalysis (Fig. 6a) can be noticed in the core region (10–20°N and 50–70°E) during increasing trend of solar cycle at 850 hPa. Although the maximum difference between the increasing and decreasing trend and between maximum and minimum solar cycle is within 2 m/s, it can be concluded that the LLJ is also affected due to solar cycle. Interestingly the features are similar between ERA-Interim and NCEP/NCAR reanalysis (figure not shown) both during increasing and decreasing and maximum and minimum sunspot conditions unlike that were observed in TEJ. Enhancement in the wind speed between increasing and decreasing sunspot years and also maximum and minimum sunspot years can be noticed after removing ENSO years. Note that in this subsection, we have used composite analysis to see the differences in the solar maximum and minimum and also during increasing and decreasing solar conditions in the monsoon features. However, note that number of years is not uniform in the increasing (5 years) and decreasing (12 years) trend years though more or less

7

similar numbers in maximum (21 years) and minimum (23 years) sunspot years exists. Thus, the difference between them in the former case might be biased. In order to overcome this limitation we have used advanced statistical tool namely Intrinsic Mode Functions (IMF) to delineate the effects of various oscillations present in the monsoon features. 3.5. Intrinsic mode functions (IMF) It is well known that in any long-term data set, the contribution of semi-annual (SAO), annual (AO), quasi-biennial (QBO), ENSO, solar cycle and long-term trend will be embed. The contribution of each oscillation needs to be treated separately. In order to extract the solar components in the ISM characteristics, an advanced statistical tool, IMF has been applied, which was introduced by Huang et al. (1998) for analyzing the nonlinear and non-stationary data using the Empirical Mode Decomposition (EMD) method. This method is empirical because the local characteristic time scales of the data itself are used to decompose the time series. It has distinct advantage in identifying the dominant periods and their amplitudes. The structure of each mode is determined by the natural amplitude variations in the time series, and higher frequencies are captured in the first mode and subsequent modes have lower average frequencies (Huang et al., 2012). In brief, IMF is a data-derived function such that, in the whole data set, the number of extreme and the number of zero crossings must be equal or may differ by one. At any point, the mean value of the envelope defined by the local maxima and the envelope defined by the local minima is zero. Recently, Kishore et al. (2012) explained about the IMF procedure and extracted the planetary wave signatures using MLS temperature data sets. This relatively new and promising EMD technique has been performed on the ISM characteristics of TEJ and LLJ zonal winds obtained from ERAInterim and NCEP/NCAR reanalysis and also on the rainfall obtained from India Meteorological Department. Fig. 7 shows the IMFs extracted using Era-Interim reanalysis zonal wind at 100 hPa. In EMD decomposing procedure, at every time step we construct upper and lower envelops by connecting all maxima or all minima with cubic spline. The mean of the upper and lower envelops is subtracted from the original time series and

Fig. 6. Same as Fig. 5, but at 850 hPa.

Please cite this article as: Ratnam, M.V., et al., Solar cycle effects on Indian summer monsoon dynamics. Journal of Atmospheric and Solar-Terrestrial Physics (2014), http://dx.doi.org/10.1016/j.jastp.2014.06.012i

M.V. Ratnam et al. / Journal of Atmospheric and Solar-Terrestrial Physics ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Data

8

-27 -28 -29 -30 -31 -32

ECMWF wind Date:1979-2011, July 150hPa

Lomb Scargle periodogram

a

f

3 2 1

b

1 IMF1

4

0

IMF2

0.6 0.4 0.2 0.0 -0.2 -0.4 -0.6 0.6 0.4 0.2 0.0 -0.2 -0.4 -0.6

d

12 10 8 6 4 2

i

IMF3

g

0.4 0.2 0.0 -0.2 -0.4

e

14 12 10 8 6 4 2

j

IMF4

-1

6 5 4 3 2 1

1980

c

1985

1990

1995

2000

2005

2010

Year

12 10 8 6 4 2

0

h

5

10 15 20 Peirods (yrs)

25

30

Fig. 7. Time series of (a) yearly mean zonal wind observed using NCEP/NCAR reanalysis data at 100 hPa in peak TEJ grid (10–20°N; 60–90°E). Extraction of IMF1, IMF2, IMF3 and IMF4 are shown in (b)–(e), respectively. Corresponding L-S periodograms (period vs. power in arbitrary units) are shown in (f)–(j), respectively. Dashed horizontal line in (h)–(j) indicates 90% confidence level.

then the envelops of the residual is again found using spline interpolation. The mean of that envelops is then subtracted from the residual and the process is repeated. This procedure continues till the filter data shows only one peak or one valley occurs which will be the first IMF denoted as IMF1. To extract the second IMF, the first IMF is subtracted from the original data and the process is repeated. In the similar way, IMF3, IMF4, IMF5 are hierarchically extracted by repeating the shifting process until the filter data shows no oscillations. Thus, long-term climate trends, centerline drifts, and long period non- stationary features come out as the last IMF. The approach of IMF decomposition to investigate the influence of QBO, ENSO, solar cycle on Indian summer monsoon characteristics is new. Here we examine the QBO, ENSO and solar cycle components in the background wind at 100 hPa, 150 hPa, 850 hpa and 700 hPa. For the time series of wind data in peak TEJ grid (10– 20°N and 60–90°E) shown in Fig. 7; five IMFs can be extracted, but only four IMFs are shown here to characterize the most important components. In order to investigate the gross characteristics of the oscillations with dominant periods, we applied Lomb Scargle (L-S) periodogram analysis at each IMF and resultant amplitude spectral plots are shown in the right panels of Fig. 7. It is observed that the higher frequency oscillations are captured in the first IMF and subsequent modes shows clearly gross characteristics of the oscillations with dominant periods. The first mode IMF (IMF1) contains the equatorial QBO and the maximum amplitude at around 28 months. The second mode IMF (IMF2) corresponds to ENSO mode and an average period of about 58–60 months. Similarly, the third mode IMF (IMF3) is associated with an average period of 10year, and these oscillations are observed with 90% confidence level. In the third IMF, a clear peak between 9 and 11 years, and the maximum peak at around 10-years exists and it corresponds to the

solar cycle. The standard method of calculating the L-S periodogram significance levels employed in the present study is only appropriate for the most significant peak in a spectrum (Horne and Baliunas, 1986). All these oscillations (QBO, ENSO and solar cycle) are also observed in the original time series data sets. Note that peak-to-peak values in the IMFs are decreasing as the amplitude decreases from QBO to ENSO and to the solar cycle. However, panels shown in the right (L-S periodogram analysis) reveal an increase in the power as a particular frequency/period is very dominant. In IMF3, though it seems to be consistent from cycle-to-cycle, note that periodicity is not the same which will come as broad peak shown in the right panels. In IMF4, dominant oscillation of 15–30 years is clearly observed. Agnihotri et al. (2011) also have reported this inter-decadal (16–30 years) variability between TSI and All India rainfall. It is worth mentioning here that the decomposed oscillations are only a small part of the whole wave spectrum. Similar analysis has been performed by Iyengar and Raghu Kanth (2005) using Indian monsoon rainfall data and found that the data exhibit six modes of temporal variation corresponding to QBO, ENSO, sunspot cycle, tidal forcing and climatic average. We further extend our study to see the influence of solar cycle on LLJ at 850 hPa in grid of 10–15°N; 50–70°E, 700 hPa over Gadanki grid and Central India rainfall shown in Fig. 8. Six dominant wave periods are observed in NCEP/NCAR reanalysis in LLJ at 850 hPa, 700 hPa and rainfall data (figure not shown). However the oscillations around 11 years are extracted and presented in Fig. 8. Though clear solar cycle signal can be seen at 700 hPa over Gadanki grid, a broad peak around 10 years is seen at 850 hPa. One does not see a solar cycle component in Central India rainfall (Fig. 8c) but rather a statistically significant (at 90%

Please cite this article as: Ratnam, M.V., et al., Solar cycle effects on Indian summer monsoon dynamics. Journal of Atmospheric and Solar-Terrestrial Physics (2014), http://dx.doi.org/10.1016/j.jastp.2014.06.012i

M.V. Ratnam et al. / Journal of Atmospheric and Solar-Terrestrial Physics ∎ (∎∎∎∎) ∎∎∎–∎∎∎

confidence level) variation with a ∼13-year periodicity. This will be discussed again in later sections. The amplitudes and phases of all these oscillations obtained from both NCEP/NCAR and ERA-Interim reanalysis data sets for different pressure levels are shown in Tables 3 and 4, respectively, and for rainfall in Table 5. Note that additional level available in ERA-Interim data sets at 125 hPa is also considered as it is close to the TEJ level. QBO (period of 32 months) exhibits larger amplitudes (about 4.17 and 4.142) in July and August months at 100 hPa and 125 hPa respectively in ERA-Interim reanalysis. Large amplitudes (8.797, 9.740, and 7.982) are appearing in solar cycle in July at 100 hPa, 125 hPa and 150 hPa in ERA-Interim reanalysis during 1979–2010. In NCEP/NCAR reanalysis, high amplitudes appear in July month at 100 hPa in the solar cycle component during 1948– 2010. Average amplitude of 4–6 m/s is observed from the solar

Fig. 8. L-S periodograms (period vs. power in arbitrary units) obtained for the solar component through IMF analysis for (a) daily mean zonal wind observed using NCEP/NCAR reanalysis data at 850 hPa in peak TEJ grid (10–20°N; 60–90°E), (b) 700 hPa in Gadanki grid (12.5–15°N; 77.5–80°E), and (c) for Central India rainfall. Dashed horizontal line indicates 90% confidence level.

9

cycle component at 100 hPa wind measurements. QBO, ENSO, and solar exhibit higher amplitudes in both 700 hPa and 850 hPa in NCEP/NCAR reanalysis over Gadanki grid. Note that there are changes in the amplitudes and phases between ERA-Interim and NCEP mainly due to considering the data from 1948 in NCEP/ NCAR. 3.6. Influence of solar cycle on ISM rainfall In order to explore further long-term association (1901–2010) between the solar cycle and ISM rainfall (averaged over Central India) variabilities, the anomalies of these two variables during peak monsoon months (July and August) for the period of 110 years (1901–2010) is shown in Fig. 9. For all the different combinations of sunspot years, from the year 1901 to 2010, the rainfall activity is deviated either negatively or positively from overall mean. In some of the cases the anomaly in the rainfall occurrence variability is positive, when the sunspot activity anomaly is negative and vise versa which is found to be inconsistent. We further compute correlation coefficient between the two and are found to be 0.18 with 90% confidence level. Results of correlation coefficients for the peak monsoon months are presented in Table 6. Note that we have made analysis while considering Central India and also All India separately. There is reasonable correlation (0.46) between solar cycle and Central Indian rainfall during maximum solar cycle but not in other cases in the last 60 years. This correlation degrades (0.29) when we consider complete data set (1901–2010). Selvaraj and Aditya (2012) found that Tamil Nadu rainfall is correlated with the sunspot activity with a moderate to high significance and the occurrence of rainfall is low during high sunspot activity. Note that the correlation between the rainfall and sunspot activity may be positive, negative or nonexistent depending upon the geographical location or time interval. Clayton (1923) found that increased precipitation exists in the equatorial regions and decreased precipitation in the mid-latitudes during solar maximum from different stations at different geographic latitudes during 1860–1917. Sun-climate mechanism is important for precipitation variability (which is indirectly affecting) at different areas and is significant with short periods like 2–6 years than 11, 22 or higher order periods (Bal and Bose, 2010). However, note that the large uncertainties in the correlation coefficient suggest that measure-

Table 3 IMF analysis for TEJ over peak grid (10–20°N; 60–90°E) using NCEP/NCAR and ERA-Interim reanalysis at 100 hPa, 150 hPa and 200 hPa. Model

Month

hpa

Amp QBO(28) (m/s)

Phase QBO (28) (deg)

Amp QBO(32) (m/s)

Phase QBO(32) (deg)

Amp ENSO (m/s)

Phase ENSO (deg)

Amp Solar (m/s)

Phase Solar (deg)

1979–2011 ERA ERA ERA ERA

July July July July

100 125 150 200

1.510 0.850 0.517 0.461

281.61 242.79 215.57 217.72

4.170 3.564 2.613 1.593

270.18 263.52 262.23 267.71

3.466 3.429 2.881 2.762

128.98 126.70 144.23 163.68

8.797 9.740 7.982 5.753

44.20 39.44 31.60 28.07

ERA ERA ERA ERA

August August August August

100 125 150 200

2.916 2.319 2.008 1.849

257.49 255.15 245.39 237.85

4.142 3.497 2.797 2.393

259.08 263.00 276.02 281.80

3.641 3.088 2.590 3.376

78.03 119.59 144.41 153.48

3.679 4.724 2.495 1.439

44.60 53.18 73.89 126.08

1948–2010 NCEP July NCEP July NCEP July

100 150 200

1.298 0.416 0.419

2.16 233.25 231.31

0.757 1.415 0.919

129.16 109.98 106.27

1.466 5.188 3.092

138.00 216.02 253.72

12.629 8.958 7.167

334.22 26.79 26.69

NCEP NCEP NCEP

100 150 200

3.175 1.475 1.090

14.96 12.18 2.54

1.306 1.311 1.039

114.41 115.97 119.40

3.203 5.913 5.054

191.89 212.76 240.15

5.279 8.815 7.400

352.25 54.85 51.79

August August August

Please cite this article as: Ratnam, M.V., et al., Solar cycle effects on Indian summer monsoon dynamics. Journal of Atmospheric and Solar-Terrestrial Physics (2014), http://dx.doi.org/10.1016/j.jastp.2014.06.012i

M.V. Ratnam et al. / Journal of Atmospheric and Solar-Terrestrial Physics ∎ (∎∎∎∎) ∎∎∎–∎∎∎

10

Table 4 IMF analysis for LLJ over peak grid and Gadanki grid using NCEP/NCAR analysis at 700 hPa and 850 hPa. Model

Month

hpa

Amp QBO (28) (m/s)

Ph QBO(28) (deg)

Amp QBO (32) (m/s)

Ph QBO(32) (deg)

Amp ENSO (m/s)

Ph ENSO (deg)

Amp Solar (m/s)

Ph Solar (deg)

Gadanki NCEP NCEP

July July

700 850

0.784 0.778

196.96 167.14

0.485 0.873

18.22 327.46

1.482 0.918

96.05 204.40

6.033 7.912

263.41 259.70

NCEP NCEP

August August

700 850

0.781 0.848

285.50 284.36

1.589 1.630

54.10 53.15

2.981 2.410

250.81 274.07

6.869 6.603

293.46 285.86

NCEP Peak LLJ NCEP July NCEP July

700 850

0.489 0.264

144.50 162.66

0.833 0.376

22.39 17.57

1.917 1.164

263.75 288.04

4.248 5.025

261.25 278.79

NCEP NCEP

700 850

0.425 0.989

295.91 268.51

1.178 0.898

50.16 56.03

1.891 0.825

219.05 213.85

1.744 4.772

298.96 292.79

August August

Fig. 9. Anomalies observed in the Central Indian region rainfall (mm/day) averaged over 1901–2010 during peak monsoon season (July–August). Anomalies observed in the sunspot number is also superimposed.

Table 5 IMF analysis for the rainfall (mm/day) for All India and Central India. Model

Month Amp QBO(28) (mm/day)

Ph QBO(28) (deg)

Amp QBO(32) (mm/day)

Ph QBO(32) (deg)

Amp ENSO (mm/day)

Ph ENSO (deg)

Amp Solar (mm/day)

Ph Solar (deg)

All India All India Central India Central India

JJA JA JJA JA

81.35 89.3 94.3 105.03

0.454 0.288 0.476 0.591

260.87 256.32 268.74 317.4

1.022 1.038 0.616 0.57

344.88 307.61 88.65 146.03

2.077 1.666 2.865 2.579

308.32 306.5 284.58 292.43

0.607 0.36 1.008 0.845

ment errors need to be taken into consideration to see the relation between the sunspot activity and the rainfall variability (Hiremath and Mandi, 2004).

3.

4. Summary and conclusions 4. In the present study we made an attempt to study the influence of solar cycle on TEJ, LLJ intensities and rainfall using long-term data sets available from NCEP/NCAR, ERA-Interim and 20th century reanalysis and IMD rainfall datasets. Main findings obtained from this study are summarized below: 1. Consistency between ground based measurements available for 10 years and reanalysis products is checked and found that these reanalysis data sets represents well the observed features for investigating the trends in the winds. 2. Long-term trends in TEJ reveals decreasing trend at the rate of 0.13 m/s/yr from 1948 to 2000 and then starts increasing in the

5.

6.

recent decade, however, no perceptible trend is noticed in LLJ at 850 hPa but show increasing trend at 700 hPa. Interestingly, an increasing trend in the TEJ intensity during the first half of century (1901–1950) at a similar rate (0.13 m/s/yr) to the decreasing tends observed in latter half is noticed in 20CR data sets. There exists nearly 60 year oscillation and this demands separate investigation. There exists an inverse relation between Central India rainfall and TEJ intensity i.e., higher the TEJ intensity, lower the rainfall. Large difference of 2 m/s (5 m/s) in zonal wind is noticed between solar maximum and minimum (increasing and decreasing trend) in TEJ in both NCEP/NCAR and ERA-Interim reanalysis. Easterlies are stronger in solar maximum and increasing trends years. These differences are larger after removing the effect of ENSO. In the LLJ at
850 hPa and 700 hPa, the difference of about 2 m/s in both the reanalysis data sets between solar maximum and minimum (increasing and decreasing trend) is noticed with stronger westerlies. Again, these differences are larger after removing the effect of ENSO.

Please cite this article as: Ratnam, M.V., et al., Solar cycle effects on Indian summer monsoon dynamics. Journal of Atmospheric and Solar-Terrestrial Physics (2014), http://dx.doi.org/10.1016/j.jastp.2014.06.012i

M.V. Ratnam et al. / Journal of Atmospheric and Solar-Terrestrial Physics ∎ (∎∎∎∎) ∎∎∎–∎∎∎

7. Using advanced statistical tool Intrinsic Mode Function (IMF), the solar cycle in TEJ and LLJ are extracted. It is demonstrated that the time series of zonal wind in peak months (July and August) are decomposed into five empirical modes with dominant periods of 2 years in IMF1, 4–7 years in IMF2, and 11 years period in IMF3. 8. The correlation coefficient between the solar activity and rainfall is found to be insignificant indicating no clear link between the two. The results indicate that IMF is able to pick up the solar component clearly, suggesting that TEJ and LLJ characteristics are strongly influenced by solar cycle. However, solar cycle signal is not clear in the rainfall unlike reported by Bhattacharyya and Narasimha (2007) but consistent with that reported by Jagannathan and Bhalme (1973). Note that Jagannathan and Bhalme (1973) reported presence of the solar cycle in only 5 stations among 48 stations considered. This is mainly due to slightly different oscillations observed in rainfall (Fig. 8) when compared to other monsoon circulation features (TEJ and LLJ). Note that Bhattacharyya and Narasimha (2007) considered the band of 8–16 years oscillation for the solar cycle. It is worth to recall recent results by Agnihotri et al. (2011), where they have observe interdecadal (16–30 years) variability between TSI and All India rainfall rather than 11 years cycle. This exactly matched with the IMF4 variability (15–30 years) in TEJ shown in Fig. 7. In the present study, though there is detectable change in the monsoon circulation during different solar conditions, this change is not reflected in the rainfall exactly but with slightly higher periods (13 years). This suggests that the relation between the solar cycle and the monsoon circulation features is linear but with rainfall it has nonlinear relation. Note that lead/lag correlations are already included in the analysis procedure we followed as we dealt solar conditions differently. In a detailed study we could notice (figure not shown) significant difference in the SST between solar maximum and minimum conditions and it varies from cycle-to-cycle. In a nut shell, increase in the SST in the Indian ocean during solar maximum might have caused increased intensities in LLJ and TEJ. These winds in turn bring lot of moisture from sea into the land and releases as precipitation. Note that moisture/cloud cover itself has direct effects of sun which will in turn change according to the solar cycle. Cloud cover/moisture thus, have indirect effect when we deal according to the solar cycle effects on the monsoon dynamics (LLJ, MLLJ). Since several other parameters like monsoon trough, moist and dry static stability etc., will also plays role in the monsoon rainfall, delayed effects might have seen. This needs further investigations while considering other solar parameters.

Table 6 Correlation coefficients obtained between the rainfall data (for Central India) and sunspot number for the whole period (1901–2000) and for 1948–2010 during peak monsoon months (July and August).

Increasing trend years Decreasing trend years Maximum sunspot years Minimum sunspot years Total years

Correlation coefficient 1948–2010

Correlation coefficient 1901–2010

 0.01  0.30

0.24 0.02

0.46

0.29

0.10

0.08

0.22

0.18

11

Acknowledgments This work is part of CAWSES-India Phase-II theme 1 fully supported by Indian Space Research Organization, Department of Space. We thank NCEP/NCAR, ERA-Interim, 20th century reanalysis centers for providing data used in the present study through their ftp sites. One of the authors (YDS) wishes to thank Department of Science and Technology for providing fellowship. We thank Ms. Vedavathi for helping in the analysis after YDS expired.

References Agnihotri, Rajesh, Dutta, Koushik, 2003. Centennial scale variations in monsoonal rainfall (Indian, east equatorial and Chinese monsoons): manifestations of solar variability. Curr. Sci. 85, 4. Agnihotri, Rajesh, Dutta, Koushik, Soon, Willie, 2011. Temporal derivative of Total Solar Irradiance and anomalous Indian summer monsoon: an empirical evidence for a Sun-climate connection. J. Atmos. Sol. Terr. Phys. 73 (13), 1980–1987. Bal, S., Bose, M., 2010. A climatological study of the relations among solar activity, galactic cosmic ray and precipitation on various regions over the globe. J. Earth Syst. Sci. 119 (2), 201–209. Bengtsson, L., Hagemann, S., Hodges, K.I., 2004. Can climate trends be calculated from reanalysis data? J. Geophys. Res. 109, D11111, http://dx.doi.org/10.1029/ 2004JD004536. Bhattacharyya, S., Narasimha, R., 2005. Possible association between Indian monsoon rainfall and solar activity. Geophys. Res. Lett. 32, L05813, http://dx. doi.org/10.1029/2004GL021044. Bhattacharyya, S., Narasimha, R., 2007. Regional differentiation in multidecadal connections between Indian monsoon rainfall and solar activity. J. Geophys. Res. 112, D24103, http://dx.doi.org/10.1029/2006JD008353. Burns, S.J., Fleitmann, D., Mudelsee, M., Neff, U., Matter, A., Mangini, A., 2002. A 780-year annually resolved record of Indian Ocean monsoon precipitation from a speleothem from South Oman. J. Geophys. Res. 107 (D20), 4434, http://dx.doi. org/10.1029/2001JD001281. Christensen, E.F., Lassen, K., 1991. Length of the solar-cycle—an indicator of solaractivity closely associated with climate. Science 254 (5032), 698–700. Christiansen, F., Haigh, J.D., Lundstedt, H., 2007. Influence of solar activity cycles on Earth's climate. DNSC Sci. Rep. (1). Claud, C., Duchiron, B., Terray, P., 2008. On associations between the 11-year solar cycle and the Indian Summer Monsoon system. J. Geophys. Res. 113 (D9), 1–12, http://dx.doi.org/10.1029/2007JD008996. Clayton, H.H., 1923. World Weather, Including a Discussion of the Influence of Solar Radiation on the Weather. Macmillan, New York. Compo, G.P., et al., 2011. The twentieth century reanalysis project. Q. J. R. Meteorol. Soc 137, 1–28, http://dx.doi.org/10.1002/qj.776. Dee, D.P., Uppala, S., 2009. Variational bias correction of satellite radiance data in the ERA-Interim reanalysis. Q. J. R. Meteorol. Soc 135, 1830–1841, http://dx.doi. org/10.1002/qj.493. Fleitmann, D., Burns, S.J., Mudelsee, M., Neff, U., Kramers, J., Mangini, A., Matter, A., 2003. Holocene forcing of the Indian monsoon recorded in a stalagmite from southern Oman. Science 300, 1737–1739. Ghosh, A.K., Siva Kumar, V., Kishore Kumar, K., Jain, A.R., 2001. VHF radar observation of atmospheric winds, associated shears and Cn2 at a tropical location: interdependence and seasonal pattern. Ann. Geophys. 19, 965–973. Gray, L.J., et al., 2010. Solar influences on climate. Rev. Geophys. 48, RG4001, http: //dx.doi.org/10.1029/2009RG000282. Gupta, A.K., Das, M., Anderson, D.M., 2005. Solar influence on the Indian summer monsoon during the Holocene. Geophys. Res. Lett. 32 (17), 2–5, http://dx.doi. org/10.1029/2005GL022685. Haigh, J.D., 1996. The impact of solar variability on climate. Science 272, 981–984. Haigh, J.D., 2003. The effects of solar variability on the Earth's climate. Philos. Trans. R. Soc. A 361 (1802), 95–111. Hiremath, K., Mandi, P.I., 2004. Influence of the solar activity on the Indian Monsoon rainfall. New Astron. 9 (8), 651–662, http://dx.doi.org/10.1016/j. newast.2004.04.001. Horne, J.H., Baliunas, S.L., 1986. A prescription for period analysis of unevenly sampled time series. J. Astrophys 302, 757–75764. Huang, N., Shen, Z., Long, S., Wu, M., Shih, E., Zheng, Q., Tung, C., Liu, H., 1998. The empirical mode decomposition method and the Hilbert spectrum for nonstationary time series analysis. Proc. R. Soc. London A 454, 903–995. Huang, B., Hu, Z.Z., Kinter, J.L., Wu, Z., Kumar, A., 2012. Connection of stratospheric QBO with global atmospheric general circulation and tropical SST. Part 1: Methodology and composite life cycle. Clim. Dyn. 38, 1–23. Iyengar, R.N., Raghu Kanth, S.T.G., 2005. Intrinsic mode functions and a strategy for forecasting Indian monsoon rainfall. J. Meteorol. Atmos. Phys. 90, 17–36, http: //dx.doi.org/10.1007/s00703-004-0089-4. Jagannathan, P., Bhalme, H.N., 1973. Changes in the pattern of distribution of southwest monsoon rainfall over India associated with sunspots. Mon. Weather Rev. 101, 691–700.

Please cite this article as: Ratnam, M.V., et al., Solar cycle effects on Indian summer monsoon dynamics. Journal of Atmospheric and Solar-Terrestrial Physics (2014), http://dx.doi.org/10.1016/j.jastp.2014.06.012i

12

M.V. Ratnam et al. / Journal of Atmospheric and Solar-Terrestrial Physics ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Kalapureddy, M.C.R., Rao, D.N., Jain, A.R., Ohno, Y., 2007. Wind profiler observations of a monsoon low-level jet over a tropical Indian station. Ann. Geophys. 25, 2125–2137. Kalnay, E., et al., 1996. The NCEP/NCAR 40-year reanalysis project. Bull. Am. Meteorol. Soc. 77, 437–471. Kishore, P., Velicogna, I., Venkat Ratnam, M., Jiang, J.H., Madhavi, G.N., 2012. Planetary waves in the upper stratosphere and lower mesosphere during 2009 Arctic major stratospheric warming. Ann. Geophys. 30 (1529–1538), 2012. Kodera, K., 2004. Solar influence on the Indian Ocean Monsoon through dynamical processes. Geophys. Res. Lett. 31 (24), 1996–1999, http://dx.doi.org/10.1029/ 2004GL020928. Koteswaram, P., 1958. The easterly jet stream in the tropics. Tellus 10, 43–57. Koteswaram, P., 1960. The Asian Summer Monsoon and the General Circulation Over the Tropics, Monsoons of the world. India Meteorological Department, New Delhi, pp. 105–110. Krishnamurti, T.N., Bhalme, H.N., 1976. Oscillations of a monsoon system. Part I. Observational aspects. J. Atmos. Sci. 33, 1937–1954, http://dx.doi.org/10.1175/ 1520-0469(1976)033o 1937:OOAMSP 42.0.CO;2. Labitzke, K., 2003. The global signal of the 11-year solar cycle in the atmosphere: when do we need the QBO? Meteorol. Z. 12, 209–216. Narayana Rao, D., Thulasiraman, S., Rao, S. Vijaya Bhaskara, Narayana Rao, T., Kishore, P., Venkat Ratnam, M., Krishna Reddy, K., 2000. VHF Radar observations of Tropical Easterly Jet stream over Gadanki. Adv. Space Res. 26, 943–946. Neff, U., Burns, S.J., Mangini, A., Mudelsee, M., Fleitmann, D., Matter, A., 2001. Strong coherence between solar variability and the monsoon in Oman between 9 and 6 kyr ago. Nature 411 (6835), 290–293, http://dx.doi.org/10.1038/35077048. Rajeevan, M., Bhate, J., 2009. A high resolution daily gridded rainfall data set (1971– 2005) for mesoscale meteorological studies. Curr. Sci. 96, 558–562. Raghavan, K., 1973. Tibetan anticyclone and Tropical Easterly Jet. Pure Appl. Geophys. 110 (1973/IX), 1230–1242. Raman, M.R., Jagannadha Rao, V.V.M., Venkat Ratnam, M., Rajeevan, M., Rao, S.V.B., Narayana Rao, D., Prabhakara Rao, N., 2009. Characteristics of the Tropical Easterly Jet: long-term trends and their features during active and break

monsoon phases. J. Geophys. Res. (114), D19105, http://dx.doi.org/10.1029/ 2009JD012065. Raman, M.R., Venkat Ratnam, M., Rajeeven, M., Jagannadha Rao, V.V.M., Rao, S.V.B., 2011. Intriguing aspects of monsoon low level jet over peninsular India revealed by high-resolution GPS radiosonde observations. J. Atmos. Sc. (68), 1413–1423, http://dx.doi.org/10.1175/2011JAS3611.1. Rao, B.R.S., Rao, D.V.B., Rao, V.B., 2004. Decreasing trend in the strength of Tropical Easterly Jet during the Asian summer monsoon season and the number of tropical cyclonic systems over Bay of Bengal. Geophys. Res. Lett. 31, L14103, http://dx.doi.org/10.1029/2004GL019817. Reddy, R.S., Neralla, V.R., Godson, W.L., 1989. The solar cycle and Indian rainfall. Theor. Appl. Climatol. 39, 194–198. Reid, G.C., 1987. Influence of solar variability on global sea-surface temperatures. Nature 329 (6135), 142–143. Reid, G.C., 1991. Solar total irradiance variations and the global sea-surface temperature record. J. Geophys. Res. 96 (D2), 2835–2844. Reid, G.C., 2000. Solar variability and the Earth's climate: introduction and overview. Space Sci. Rev. 94 (1–2), 1–11. Reiter, E.R., 1961. Jet Stream Meteorology. University of Chicago Press, Chicago, Ill p. 515. Roy, I., Haigh, J.D., 2010. Solar cycle signals in sea level pressure and sea surface temperature. Atmos. Chem. Phys. 10, 3147–3153. Sathiyamoorthy, V., 2005. Large scale reduction in the size of the Tropical Easterly Jet. Geophys. Res. Lett. 32, 1–4. Selvaraj, R.S., Aditya, R., 2012. The solar influence on the monsoon rainfall over Tamil Nadu. J. Indian Geophys. Union 16 (3), 107–111. Vasantha, B., Venkat Ratnam, M., Mohan, K., Kamala, S., Narayana Rao, D., 2002. Characteristics of tropical easterly jet over Gadanki: comparison with radiosonde and rawinsonde. Indian J. Radio Space Phys. 31, 130–139. Venkat Ratnam, M., Krishna Murthy, B.V., Jayaraman, A., 2013. Is the trend in TEJ reversing over the Indian subcontinent? Geophys. Res. Lett. 40, 3446–3449, http://dx.doi.org/10.1002/grl.50519.

Please cite this article as: Ratnam, M.V., et al., Solar cycle effects on Indian summer monsoon dynamics. Journal of Atmospheric and Solar-Terrestrial Physics (2014), http://dx.doi.org/10.1016/j.jastp.2014.06.012i