Association of the Indian summer monsoon rainfall variability with the geophysical parameters over the Arctic region

Authors


Abstract

A study is carried out to understand whether the Indian summer monsoon rainfall variability is associated with the geophysical parameters over the Arctic region. The correlation analyses of the satellite-derived sea ice data for 29 years indicate that out of 9 sectors of the Arctic region, the Kara and Barents Seas sector's Sea Ice Extent (KBS SIE) during October has a strong relationship with the All-India Summer Monsoon Rainfall (AISMR) in the following year. This relationship is more pronounced for the extreme cases, which are identified as the drought or the excess monsoon years. Moreover, the composites of certain geophysical parameters over the Arctic also behave in tandem with the monsoon rainfall. In order to test the relationship of these geophysical parameters with the monsoon rainfall, a case of the recent drought of 2009 is independently evaluated. The results obtained in this study bring out that KBS SIE and some other parameters of the Arctic region can be used as potential predictors in the long-range forecasting of AISMR with a lead period of more than six months. A table indicating the qualitative forecast of the monsoon rainfall is presented on the basis of some parameters of the Arctic region. The mean sea level pressure anomaly tendency over northwest Europe during winter, which is one of the predictors used for forecasting the AISMR, is significantly correlated with the KBS SIE during the preceding October. As such, with the knowledge of October KBS SIE, this parameter can also be foreseen a few months in advance. Copyright © 2011 Royal Meteorological Society

1. Introduction

It is a well known fact that the polar regions act as an important component of the global climate system through the processes of the large-scale melting or the formation of sea ice. The sea ice originates from the freezing of sea water under sub-zero conditions. During the formation of sea ice, the salts are liberated making the surrounding water more saline and dense. This cold dense brine sinks to form the bottom water current towards the equator and affects the ocean circulation. It is known that the albedo of open water lies between 10 and 15%, whereas the average albedo of sea ice is around 80%. Thus, the albedo effect provides a feedback mechanism that can influence the climate variability over very long time periods. The presence of sea ice leads to the reduction of the absorbed solar radiation by the ocean. The flux of oceanic heat to the atmosphere from the open water is nearly two orders of magnitude higher than the adjacent ice-cover (Badgley, 1966; Maykut, 1978). In this way, the process of build-up and decay of the polar sea ice have a link with the large-scale environmental changes occurring over the different parts of the globe. With the advancement of satellite technology in the microwave spectral region and due to the availability of highly reliable sea ice data (Comiso, 2000), the exploration and understanding of polar sea ice variability on different spatio-temporal scales have become substantially viable over the past three decades.

The Indian sub-continent, which is a region in south-central Asia, accounts for about 34% of Asia's population. Agriculture in India is the means of livelihood of almost two thirds of the work force in the country. Hence, the success or failure of the crops, and water scarcity in any year, is always viewed with great concern. A major portion of annual rainfall (75–80%) over the region is received during the summer monsoon season (June–September). There are known vagaries of the monsoon regarding the total amount of rains received over the region. Therefore, a deficit (excess) in All-India Summer Monsoon Rainfall (AISMR) in any particular year, leads to a drought (flood) condition causing a devastating impact on the Indian economy.

The AISMR depends upon the moisture flow and formation of cyclonic circulations like the low pressure areas, monsoon depressions, over the region. It is also associated with a number of factors, namely the strength of the cross-equatorial flow, the orientation and movement of the monsoon trough, the existence of offshore troughs, interaction of the extra tropical weather phenomena, etc. (e.g. Pant and Rupa Kumar, 1997; Webster et al., 1998). The continuation of such an activity for about four months over a large area like the Indian sub-continent is a part of the general circulation of the atmosphere. As such, the performance of monsoon in any particular year is a manifestation of an interaction between the various global and regional parameters. Therefore, there are teleconnections between certain parameters over different parts of the globe and the AISMR. Such teleconnections are useful for forecasting the monsoon rainfall. In the present study, we have explored the possibility of finding an association of AISMR with the sea ice extent and a few other geophysical parameters over the Arctic region.

In view of the importance of the monsoon as stated above, there are a large number of publications on the global and the regional teleconnections of the monsoon and its Long Range Forecasting (LRF) (e.g. Walker, 1924; Thapliyal and Kulshrestha, 1992; Krishna Kumar et al., 1995; Kripalani et al., 1997; Webster et al., 1998; Krishamurthy and Goswami, 2000; Hu et al., 2005). However, the relationships of the parameters with AISMR are not robust and undergo changes over the decades (Rajeevan, 2002). Nevertheless, an accurate prediction of monsoon rainfall is vital for better planning of finance, power and water resources. The research work in this area is, therefore, enduring in an attempt to identify the useful predictors that can further aid in developing a suitable statistical model for LRF of the monsoon rainfall.

The El Niño is characterized by the warming of the tropical eastern Pacific Ocean with the periodicity of 2–7 years and the Southern Oscillation (SO) refers to an oscillation in surface air pressure between the tropical eastern and the western Pacific. These two coupled phenomena together are called as the ENSO event. The inter-annual variation of AISMR is one of the strongest signals of the earth's climate variability, particularly its interaction with the ENSO. It is well recognized that there exists a negative correlation between AISMR and ENSO (e.g. Webster and Palmer, 1997), in which a weak (strong) monsoon is related to a warm (cold) ENSO event through an anomalous Walker cell driven by tropical east Pacific Sea Surface Temperature (SST) anomalies. The teleconnection of the Antarctic sea ice with ENSO and other global parameters have also been studied (e.g. Peterson and White, 1998; Yuan and Martinson, 2000). The links of the Antarctic sea ice with the large-scale precipitation over different parts of the globe, for instance, over Australia (White, 2000), China (Xue et al., 2003; Wu et al., 2009) have been brought out. Similarly, the relationships of the Arctic sea ice extent with certain parameters like the North Atlantic oscillation, Hadley circulation and the Asian summer monsoon rainfall are also highlighted in some recent studies (e.g. Li and Zeng, 2008; Zhou and Wang, 2008; Strong and Magnusdottir, 2010). These studies lead us to infer that AISMR may have an association with the polar sea ice coverage. Working in the same direction, relationship of the Antarctic sea ice extent with AISMR was brought out (Prabhu et al., 2009, 2010). These studies revealed that the sea ice extent of the western Pacific Ocean sector (Long. 90°E–160°E) in the month of March (same year) and the Bellingshausen and Amundsen Seas (BAS) sector's sea ice extent (Long. 90°W–120°W) during the Austral summer (October–December) of the previous year show strong positive and negative relationships respectively with the AISMR. Similarly, the geophysical parameters over the Southern Ocean namely the mean meridional transport of heat in the upper troposphere during Austral summer of the previous year and the SST anomaly over the east Pacific during February (same year) also provide the useful signals in advance about the ensuing monsoon rainfall.

The present study is an endeavor to know what kind of relationships hold true for the other polar region specifically the Arctic with the AISMR, and to what extent it can be used in LRF of monsoon rainfall, if at all it exists.

2. Data

The Arctic Sea Ice Extent (ArSIE) data are available from Nimbus-7's Scanning Multi-Channel Microwave Radiometer (SMMR) and Defense Meteorological Satellite Program's Special Sensor Microwave Imager (SSM/I). The Sea Ice Extent (SIE) refers to the area covered by the sea ice from the continental boundary to the ice edge in the open ocean as per the definition and the procedure followed by Gloersen et al., 1992. We have used the monthly total SIE data for the period of 1979–2007 generated from SMMR and SSM/I sensors and provided by National Snow and Ice Data Center (NSIDC) (Cavalieri et al., 1997) (website: ftp.sidads.colorado.edu). The SIE data have been used in several studies (e.g. Bhandari et al., 2005; Cavalieri et al., 1999; Budikova 2009; Parkinson et al., 1999). The Arctic region is subdivided into 9 sectors (Gloersen et al., 1992) as shown in Figure 1(a). Sector-wise data are considered for the present study.

Figure 1.

(a) Various sea ice sectors of the Arctic region, and (b) Geographical location of the Indian region and the KBS sector. This figure is available in colour online at wileyonlinelibrary.com/journal/joc

The data for other geophysical parameters over the Arctic region: air temperature (T), meridional wind velocity (V) and the Mean Sea Level Pressure (MSLP) are acquired from National Centers for Environmental Prediction (NCEP) as described by Kalnay et al. (1996). The data series during the period of 1979–2007 for the parameter Northwest (NW) Europe MSLP anomaly tendency is also developed using the source mentioned above. The AISMR dataset is the area-weighted average June–September (JJAS) rainfall based on 306 rain gauge stations distributed over the entire Indian sub-continent (Parthasarathy et al., 1995). It is available on the website of Indian Institute of Tropical Meteorology (IITM), Pune, India: ‘Homogenous Indian Monthly Rainfall Data Sets (1871–2008)’, website: ftp://www.tropmet.res.in/pub/data/rain/iitm-regionrf.txt. Although, these data are available from 1871, we have utilized them from 1980 to 2008 as per the well-calibrated Arctic sea ice data as mentioned above.

3. Methodology

The methodology adopted to examine the relationship between the ArSIE and the AISMR is the same as followed in the case of the Antarctic region (Prabhu et al., 2010). The anomalies of the geophysical parameters and the standardized values of the series are computed using the statistical equations (Prabhu et al., 2010). The mean climatology for the MSLP, T and V is taken with the base period of 1971–2000, whereas, for the AISMR, it is taken for the period of 1975–2008. First, the linear Correlation Coefficients (CC) are computed between the standardized series of AISMR and the monthly sea ice extent data over all the sectors of the Arctic region for the months preceding the monsoon season (i.e. from May of the current year to October of the previous year). Then the correlation analyses are carried out and the statistical significance of CC is tested using Student's t-test.

In order to understand the impact of Arctic region on the variability of AISMR, the composite analyses of the anomalies of the parameters over the Arctic region, that is, the MSLP, air temperature and meridional wind velocity, corresponding to the extremes of the monsoon (drought and excess) years are carried out. The years with Standardized Value (SV) of AISMR between − 1 to + 1 are considered as the normal years and the years with SV above + 1 are referred to as ‘excess’ and below − 1 as the ‘drought’ years. During the period under study, three cases of excess monsoon years: 1983, 1988 and 1994, and 6 cases of drought years: 1982, 1985, 1986, 1987, 2002 and 2004 have occurred (Figure 2).

Figure 2.

Time series plot of standardized KBS SIE in the month of October with the standardized values of the subsequent AISMR

4. Results and discussion

The relationship of the ArSIE and other geophysical parameters over the Arctic region namely MSLP, T and V with respect to the extreme monsoon years are considered and the results are presented below. Further, the point about the linkage of the NW Europe MSLP anomaly tendency, ArSIE and AISMR is discussed. In order to evaluate the linkage of the new parameters of the Arctic region brought by this study, a case of the 2009 drought is independently tested as this year was not taken into account for the composite analyses carried out in this paper. The salient features of this study are also briefly discussed in this section.

4.1. Relationship of the parameters over the Arctic region and AISMR

4.1.1. ArSIE and AISMR

The correlation analyses reveal that among different sectors of the Arctic region, Kara and Barents Sea sector's (Long. 30°E to 90°E) Sea Ice Extent (KBS SIE) during the month of October is robustly related to the ensuing AISMR with a positive correlation coefficient (CC = 0.5). It is significant at 99% level of confidence. It may be mentioned that KBS lies almost in the same meridional belt of the Indian region (Figure 1(b)).

In order to confirm this association on year-to-year basis, a plot of October KBS SIE and the AISMR of the following year is generated with their standardized values (Figure 2). It clearly depicts a strong positive relationship between them. In general, the drought (excess) years were associated with a large negative (positive) anomaly of KBS SIE. These observations bring out that the ArSIE is directly related to the AISMR. In general, it is seen that when the KBS SIE during October was below (above) the normal, the ensuing monsoon activity was subdued (enhanced).

4.1.2. Mean sea level pressure

The composite analyses of the MSLP anomaly during the month of October over the region (Long. 30°E to 90°E and Lat. 50°N to 80°N) corresponding to the extremes of monsoon in the succeeding years are carried out. It is noticed that the anomaly of MSLP shows the variation from + 5 to − 5 hPa with the positive and the negative sign signaling the excess monsoon and the drought years respectively (Figure 3). As such, the AISMR and MSLP are directly related to each other.

Figure 3.

Composite mean of MSLP anomaly (hPa) during the month of October preceding (a) drought and (b) excess monsoon years

4.1.3. Air Temperature

The composites of the mean air temperature anomaly [T′] at pressure levels from 1000 to 100 hPa for the month of October corresponding to the extremes of monsoon in the succeeding years are constructed over the region (Long. 30°E to 90°E and Lat. 75°N to 85°N). In the case of drought, a positive temperature anomaly of the order of 4–5 K is seen throughout the vertical column (Figure 4). The anomaly is negative for the excess monsoon years. The contrast is more pronounced throughout the pressure levels from 800 to 100 hPa at around 80°N. In this way, the air temperature over the KBS sector of Arctic region is inversely related to the monsoon rainfall.

Figure 4.

Composite mean anomaly of air temperature [T′] in K averaged over the longitudinal belt 30°E–90°E at different pressure levels in the month of October preceding (a) drought and (b) excess monsoon years

4.1.4. Meridional wind velocity

The anomaly of the mean meridional wind velocity [V′] at pressure levels from 1000 to 100 hPa is depicted for October corresponding to drought (Figure 5(a)) and excess monsoon (Figure 5(b)) years. It brings out the nature of anomaly of the meridional velocity along the KBS sector and thus, aids in the understanding of the overall pattern of the heat transport before the monsoon months. Positive (southerly) wind anomaly of about 2–3 m/s is seen above 700 hPa over the region (Long. 30°E to 90°E and Lat. 75°N to 85°N) for the drought cases and negative (northerly) for the excess monsoon years. Thus, the reversal of anomaly wind pattern is noticed in the upper atmosphere for the extreme monsoon years in the polar cell.

Figure 5.

Composite mean anomaly of meridional wind velocity [V′] in m/s averaged over the longitudinal belt 30°E–90°E at different pressure levels in the month of October preceding (a) drought and (b) excess monsoon years

4.2. Linkage of the NW Europe mean sea level pressure anomaly tendency, ArSIE and AISMR

One of the predictors that is currently used in the statistical models by the India Meteorological Department (IMD) for LRF of AISMR is the ‘North West Europe (region: Lat. 65°N–75°N, Long. 20°E–40°E) MSLP Anomaly Tendency (NWEPAT)’ (Rajeevan et al., 2007). The region is located around the Barents's sector (Figure 1(b)). This parameter is measured as the difference of the MSLP anomalies averaged over the area mentioned above between September to November (denoted as ‘− 1’) and December to February (denoted as ‘0’). The negative relationship (CC = − 0.4 for the base period of 1951–2000) between the NWEPAT and the AISMR is explained through the changes in the mid-latitude wind pattern over Eurasia (Rajeevan et al., 2002).

During the deficient monsoon years, the winter north–south mean sea level pressure gradient over the NW Europe is weaker than normal. The reduced pressure gradient reduces the mid-latitude zonal flow across Eurasia and increases the chances of blocking highs and intrusion of dry westerlies into the Indian region during the spring season (Raman and Maliekal, 1985). Although, our aim is to find out if there is any significant link between the Arctic sectors' sea ice and AISMR, we would like to understand whether there is any connection between the KBS sea ice and this predictor as both of them are associated with the European continent and are in close vicinity to each other. Therefore, the relationship of these parameters is tested as follows.

The CC between NWEPAT and KBS SIE for the period of 1979–2007 is computed. It is found to be − 0.61, which is significant at 99.9% level of confidence. Further, a plot of these parameters shows year-to-year relationship between them (Figure 6). As such, the knowledge of KBS SIE during October can provide a clue about NWEPAT, three to four months in advance, thereby extending the lead period of monsoon forecast.

Figure 6.

Time series plot of standardized KBS SIE in the month of October with the standardized values of NW Europe MSLP anomaly tendency (DJF(0)–SON(−1))

4.3. Evaluation of the parameters for an independent test case of the 2009 drought

India experienced one of the worst droughts in 2009 with AISMR being 22% below normal (http://www.imdpune.gov.in/research/ncc/climatebulletin/bulletin_index.html). Hence, we have evaluated the parameters discussed in this paper to verify their applicability for LRF of monsoon 2009 similar to the procedure followed for the Antarctic region (Prabhu et al., 2010). It is noticed that the anomalies of parameters namely: MSLP, [T′] and [V′] during October of 2008 show negative, positive and positive tendencies respectively (Figure 7) implying the probability of 2009 as a drought year.

Figure 7.

Patterns of geophysical parameters in the month of October 2008 over the northern hemisphere preceding drought of 2009 over India (a) MSLP anomaly (hPa), (b) and (c) shows anomalies of temperature [T′] in K and meridional wind velocity [V′] in m/s respectively that are averaged over the longitudinal belt 30°E–90°E at pressure levels from 1000 to 100 hPa

4.4. Qualitative forecast based on the Arctic region parameters

A table which would be useful in generating a qualitative forecast of AISMR based on the geophysical parameters over the Arctic region is presented (Table I). Accordingly, the negative anomalies of KBS SIE, MSLP and the positive anomalies of the air temperature, and the upper meridional wind velocity during October would indicate likelihood of a drought year. The opposite behaviour of these parameters would indicate an excess or good monsoon year.

Table I. Qualitative forecast of AISMR on the basis of the anomalies of the geophysical parameters over the Arctic region during October in the previous yearThumbnail image of

4.5. Influence of the polar regions on monsoon: Arctic versus Antarctic

It is noticed that over the Antarctic region, the BASS SIE during the preceding Austral summer (October through December) is negatively correlated to the AISMR with the CC = − 0.42 (Prabhu et al., 2010). However, as seen from this study over the Arctic region, the KBS SIE during the same season shows a significant positive relationship (CC = 0.50) with the ensuing monsoon. These observations imply that the sea ice in the Arctic region is directly related to the AISMR, whereas the Antarctic sea ice has an inverse relationship during the Austral summer. Moreover, the Arctic region is more influential than the other.

4.6. Remarks

The present study has brought out that the Arctic region could be an important player which can influence the summer monsoon rainfall over the Indian sub-continent. It is worth noting that out of the geophysical parameters discussed above, one pertains to the cryosphere (KBS SIE) and the others to the atmosphere (MSLP, T and V). The vast ocean is an interface between the cryosphere and atmosphere. However, it is difficult to say at this stage as to what is the physical mechanism for the cryosphere-ocean-atmospheric complex interactions in affecting the monsoon variability at such a long distance with different atmospheric cells embedded in the circulation patterns. Therefore, a suitable dynamical model coupling such complex interactions is needed to address this issue. An intensive research work in the coming years could possibly answer this question.

5. Conclusions

The geophysical parameters over the Arctic region show distinct features with respect to their association with the Indian summer monsoon rainfall. Although, the performance of monsoon over the Indian region depends upon many factors, the positive anomaly of KBS SIE in the month of October indicates an excess monsoon year, and the negative anomaly signals the drought year. Similarly, a negative MSLP anomaly along with positive anomalies in both the air temperature and the upper level meridional wind velocity would indicate the likelihood of the following year to be a drought year over the Indian region. The reversal of the signs of these parameters would indicate an excess monsoon year.

One of the predictors of the summer monsoon rainfall, the NW Europe MSLP anomaly tendency during winter, could also be foreseen 3–4 months ahead of time with the knowledge of October KBS SIE. The results obtained in this study reveal that KBS SIE and some other parameters of the Arctic region can be used as potential predictors in the long-range forecasting of AISMR with a lead period of more than six months.

It is interesting to note that both the Antarctic and the Arctic regions influence the Indian summer monsoon during the Austral summer (Boreal Winter) season, but in opposite ways, with the latter being more dominant.

Ancillary