Impact of cross-equatorial flow on intra-seasonal variability of Indian summer monsoon rainfall

Authors


Abstract

[1] 850-hPa zonal wind gradient anomaly between the region-1 [Equation-25°N; 30°E–75°E] and region-2 [Equation-10°S; 30°E–75°E] is used to represent the cross-equatorial flow. Long-term daily mean of 1st June to 30th September, for the period 1951–2000, of gradient time series is correlated with Indian summer monsoon daily mean rainfall, for the same period. The correlation coefficient is 0.92, which is statistically significant at 1% level. In order to understand the impact of 850-hPa zonal wind gradient on daily rainfall activity over India during monsoon season, the composite analysis of daily rainfall activity over India during extreme positive and negative 850-hPa zonal wind gradient anomaly is performed. This analysis reveals that when 850-hPa zonal wind gradient anomaly in last 20 days of May is extremely positive then almost all days in June show above-normal rainfall activity and vice versa. The correlation coefficients between 20-day mean of rainfall departure and previous 20-day mean of 850-hPa zonal wind gradient anomaly, for the period 1951–2003, are computed. This correlation analysis suggests that statistically significant (at 5% level) relationship is seen only up to the 30th June rainfall and from this point onward the relationship becomes insignificant. The result also indicates that 850-hPa zonal wind gradient anomaly, in last 20-days of May, is important for daily rainfall activity over India during onset phase of monsoon but, when the monsoon season sets in then it loses its significance.

1. Introduction

[2] The relationship between the seasonal mean Indian summer monsoon rainfall and the intra-seasonal fluctuations has been discussed extensively both from the point of view of the precipitation and the large scale circulation [Charney and Shukla, 1981, Krishnamurthy and Shukla, 2000; Sperber et al., 2000; Goswami and Ajaya-Mohan, 2001]. Webster and Yang [1992] identify indices based on the dominance of the first baroclinic mode in monsoon dynamics. These indices involve the vertical shear of the zonal wind over the monsoon region. Goswami et al. [1999] focus on the local Hadley cell associated with the monsoon to define an index based on the meridional wind shear. All these studies are mainly confining to either residual hypothesis or boundary forced hypothesis. The residual hypothesis suggests that the seasonal mean is the residual of chaotic weather systems and their low-frequency intra-seasonal modulation where as boundary-forced hypothesis suggests that the seasonal mean has a separate physical origin related to the slowly varying boundary conditions such as the sea surface temperature (SST), soil moisture, snow cover etc. The residual hypothesis implies very little predictability for seasonal means and according to the boundary-forced hypothesis, the slowly varying land and ocean states play an important role, so there is some potential predictability in the monsoon system for the seasonal mean. In this paper, daily rainfall variability in monsoon season is discussed for the extreme monsoon years and an attempt has been made to understand the linkage between daily persistency in rainfall activity and extreme monsoon season.

[3] Another purpose of the study is to examine the impact of cross-equatorial flow on intra-seasonal monsoon rainfall activity over Indian subcontinent. The observational studies of Saha [1974], Pisharoty [1976], Cadet and Reverdin [1981] and modeling studies of Washington et al. [1977] and Shukla [1984] have established the importance of the role of cross-equatorial flow over Indian Ocean and moisture flux from both Indian Ocean and Arabian Sea regions in Indian monsoon rainfall, particularly over the west coast. However, there is some difference of opinion regarding the relative dominance of fluxes from Indian Ocean and Arabian Sea. Cross-equatorial flow develops during the onset phase of the monsoon season and its strength, if predictable, can provide an important predictor for Indian summer monsoon rainfall (ISMR). Hastenrath [1987] and Hastenrath and Greischar [1993] suggested the use of SSTs over the Arabian Sea and Parthasarathy and Sontakke [1988] attempted to represent the strength of cross-equatorial flow in the Indian Ocean region by Nouvelle-Agalega SLP difference as well as SSTs. However, none of them could find practical utility in the development of LRF schemes for ISMR. Thus, in spite of the known physical link of cross-equatorial flow with monsoon rainfall, there has been very little progress in identifying useful predictors based on it. In this paper, the strength of cross-equatorial flow is measured as the 850-hPa zonal wind gradient anomaly between region-1 [Equation-25°N; 30°E–75°E] and region-2 [Equation-10°S; 30°E–75°E]. The longitudinal span of both regions is same, but latitudinal span of region-1 is larger and is extended up to 25°N, which includes parts of Pakistan, Afghanistan, Iran, Saudi Arabia etc. Observational and modeling studies by Love [1985], Rodwell and Hoskins [1995], and others suggest that synoptic scale fluctuations over these regions may affect the East African low-level jet, which is one of the most important manifestations of the cross-equatorial flow involving large-scale moisture and momentum transport [Findlater, 1966, 1969, 1977]. The strong south easterly winds on the periphery of Mascarian high which turn to southerlies near equator and become southwesterly winds over Arabian Sea and Peninsular India suggests positive gradient anomaly. Thus positive zonal wind gradient anomaly at 850-hPa is indicative of strong cross-equatorial flow.

2. Data Used

[4] 1. Daily grid point 850-hPa zonal wind data, for the region 10°S–30°N and 25°E–90°E, have been obtained from CDAS-NCEP/NCAR reanalysis data. The web address is nomad3.ncep.noaa.gov.

[5] The gradient between 850-hPa zonal wind, averaged over the region-1 [Equation-25°N; 30°E–75°E] and region-2 [Equation-10°S; 30°E–75°E], is computed for each day from 1st May to 30th September for the period 1951–2006. Long-term daily mean from 1st June to 30th September for the period 1951–2000 of this time series is computed and gradient anomaly series, from this long-term mean is used to represent the cross-equatorial flow.

[6] 2. Daily Indian summer monsoon rainfall data, for the period 1951–2003, have been obtained from Rajeevan et al. [2005] (National Climate Centre, India Meteorological Department, Pune 411005, India).

3. Discussion

[7] The intra-seasonal variability is the manifestation of daily rainfall variability. Therefore, it is logical to expect that this daily rainfall variability may affect the seasonal rainfall variability.

3.1. Daily Rainfall Variability During Extreme Monsoon Years of ISMR

[8] Monsoon season consists of 122 days from 1st June to 30th September. The 122-day mean rainfall departure (in %) for each year from 1951 to 2003 is calculated. Mean (M) and standard deviation (S) for 1951–2000 period of 122-day mean rainfall departure (in %) is computed. We define a particular year as excess (deficient) if 122-day mean rainfall departure (in %) value for that year is greater than M + S (less than M − S). During the period 1951–2003 (53 years), there are 28 years when seasonal averaged daily rainfall anomaly for monsoon season is positive (above-normal) and 25 years when the same is negative (below-normal). Out of 28 above-normal rainfall years, 10 years are excess monsoon years (1954, 1958, 1959, 1961, 1970, 1975, 1983, 1988, 1990 and 1998) and 18 years are positive-normal monsoon years. Similarly, out of 25 below-normal years, 10 years are deficient monsoon years (1951, 1952, 1965, 1968, 1972, 1979, 1982, 1986, 1987 and 2002) and 15 years are negative-normal monsoon years. The composite mean of daily rainfall departure (in %) in monsoon season for excess and deficient monsoon years is depicted in Figure 1. It reveals that, almost all days in excess monsoon year show positive rainfall departure and almost all days in deficient monsoon year show negative rainfall departure. Moreover, in excess monsoon years, all days in September and at least 8–10 days in July and August show rainfall departure (%) greater than 10.3 (Mean + S.D. for the period 1951–2000) and in deficient monsoon years, about 15–20 days in each month from June to September show rainfall departure (%) less than −10.3 (Mean − S.D. for the period 1951–2000). The study indicates that during drought years there is a persistence of extreme negative daily rainfall departure values in June, July and September whereas during flood years there is a persistence of extreme positive rainfall departure values in first few days of July and all days in September. Thus, persistence of extreme negative daily rainfall departure values in June and July may gives rise to the seasonal drought.

Figure 1.

Composite mean of daily rainfall departure from normal (in %) for (a) 10 excess monsoon years and (b) 10 deficient monsoon years for data period 1951–2003.

3.2. Daily Variability of 850-hPa Zonal Wind Gradient Anomaly

[9] The gradient between 850-hPa zonal wind, averaged over the region-1 [Equation-25°N; 30°E–75°E] and region-2 [Equation-10°S; 30°E–75°E], is computed for each day from 1st May to 30th September for the period 1951–2006. This gradient will represent the cross-equatorial flow. Long-term daily mean from 1st June to 30th September for the period 1951–2000 of this time series is computed and its correlation coefficient with Indian summer monsoon daily mean rainfall, for the same period, is calculated (Figure 2). The correlation coefficient is 0.92, which is statistically significant at 0.1% level. Since both the series have persistence, effective degrees of freedom (NE) is computed by the formula NE = N/ [1 + 2*(r*1r'1 + r*2r'2 + r*3r'3 + r*4r'4 + r*5r'5)], where ri and r'i are auto-correlation coefficients of two respective series, and the correlation coefficient is found to be statistically significant at 1% level for this effective degrees of freedom. It gives the climatological relationship between these two atmospheric processes.

Figure 2.

Daily mean of monsoon season (1st June to 30th September) for the period 1951–2000 of Indian rainfall and 850-hPa zonal wind gradient.

[10] A time series is prepared for 1951–2006 by averaging 850-hPa zonal wind gradient anomaly of last 20 days of May (12–31 May) for each year. The mean and standard deviation (S.D.) of this time series is computed for the period 1951–2000. Statistically speaking the values lying outside the range (mean + S.D., mean - S.D.) are considered to be extreme values. In order to find threshold range, we narrowed the range as (mean + half S.D., mean - half S.D.) and defined the extreme value as: the value which is greater than [mean + (S.D.*0.5)] is considered as extreme positive and the value which is less than [mean- (S.D.*0.5)] is considered as extreme negative. During the period 1951–2003, there are 14 years when average of last 20 days in May of 850-hPa zonal wind gradient anomaly is extremely positive and 16 years when it is extremely negative. The extremely positive years are 1955, 1956, 1960, 1961, 1966, 1973, 1978, 1986, 1989, 1990, 1999, 2000, 2001 and 2003 where as extremely negative years are 1951, 1952, 1953, 1954, 1957, 1959, 1963, 1967, 1972, 1980, 1982, 1987, 1991, 1992, 1997, and 1998. In order to understand the impact of average of last 20 days in May of 850-hPa zonal wind gradient on intra-seasonal rainfall activity over India, the composite analysis of daily rainfall activity over India during extreme positive and negative 850-hPa zonal wind gradient anomaly is performed. This analysis reveals that when 850-hPa zonal wind gradient anomaly in last 20 days of May is extremely positive then there is above-normal rainfall activity during almost all days in June only (Figure 3). It suggests that the extreme positive or negative value of 12–31 May average 850-hPa zonal wind gradient anomaly is indicative of the above-normal or below-normal rainfall activity during onset phase of monsoon only and it is not indicating the daily rainfall activity from July to September.

Figure 3.

Composite mean of daily rainfall departures (%) in June during extremely positive (14 years) and negative (16 years) 850-hPa zonal wind gradient anomaly.

[11] 20 day running mean is computed for each year from 1951 to 2003 of rainfall departures (in %) from 1st June to 30th September and 850-hPa zonal wind gradient anomaly from 1st May to 30th September. The correlation coefficients between 20-day mean of rainfall and previous 20-day mean of 850-hPa zonal wind gradient anomaly, for the period 1951–2003, are computed and plotted on graph. The correlation value is represented at the last day (Figure 4). This correlation analysis suggests that statistically significant (at 5% level) relationship is seen only up to the 30th June rainfall and from this point onward, the relationship becomes insignificant. This result also indicates that 850-hPa zonal wind gradient anomaly, in last 20-days of May, is important for daily rainfall activity over India during onset phase of monsoon but, when the monsoon season sets in then it loses its importance.

Figure 4.

Correlation between 20-day mean rainfall with previous 20-day mean of 850-hPa zonal wind gradient anomaly for the period 1951–2003.

4. Conclusion

[12] 1. The excess seasonal mean rainfall for a particular year is possible if there is a persistence of daily excess rainfall activity for at least 8–10 days in July and August each and almost all days in September. However, the deficient seasonal mean rainfall for a particular year is possible if there is a persistence of daily deficient rainfall activity for at least 15–20 days in each month of the monsoon season. Thus, the deficient monsoon year can be inferred by observing persistent daily deficient rainfall activity in June and July.

[13] 2. On climatological scale, daily 850-hPa zonal wind gradient anomaly, in monsoon season, is strongly associated with daily rainfall activity over India.

[14] 3. Extreme positive (negative) 850-hPa zonal wind gradient anomaly in last 20 days of May is linked with above-normal (below-normal) rainfall activity during June.

[15] 4. 850-hPa zonal wind gradient anomaly, in last 20-days of May, is important for daily rainfall activity over India during onset phase of the monsoon but, when the monsoon season sets in then it loses its statistical significance.

Acknowledgments

[16] The authors are grateful to the director, I.I.T.M., for providing necessary facilities for completing this study and to the head, Forecasting Research Division, for encouragement and valuable suggestions.

Ancillary