The location and shift in the position of the maximum lightning activity with respect to the position the of monsoon trough has been investigated in the region of 20°N–32°N and 75°E–95°E during the Indian summer monsoon seasons of 2002 and 2003 from the lightning imaging sensor (LIS) data from the Tropical Rainfall Measuring Mission (TRMM) satellite. The polynomial curve fitting the maximum lightning activity followed the position of the monsoon trough in both the normal monsoon and drought years. Observations of the maximal lightning activity in the TRMM-LIS data collected during a period of 13 years, from 1998 to 2010, confirm these findings in the monsoon season. When the seasonal position of the monsoon trough moved from the plains to the submontane region of the Himalayas during a monsoon break period, the polynomial fitting of the maximum lightning activity also moved from the central Indian plains to that region. An empirical orthogonal function (EOF) analysis is applied to two different data sets of precipitation and observed high lightning flash rates over the monsoon trough region for the period of 1995–2005. The first and second components, EOF1 and EOF2, of precipitation values are out of phase with those of the flash rates. Our results imply that deep electrified convective systems during monsoon periods are responsible for the lightning activity in the monsoon trough region and that the lightning activity in this region can be used as a proxy for the location and shift of the seasonal monsoon trough.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 Measurements of global lightning from the ground and from space in the last few decades have greatly increased the application and integration of lightning in climate studies. Several efforts have been made to find out the correlations between lightning activity and related meteorological properties. These studies suggested strong connections between lightning and thunderstorm updraft growth [Goodman et al., 1988], rainfall rate [Petersen and Rutledge, 1998; Tapia et al., 1998], cloud-top height [Williams, 1985; Price and Rind, 1992], and mesocyclone occurrence [MacGorman et al., 1989; Williams et al., 1999]. In an interesting study in the subtropical southern hemisphere, Ortega and Guignes  found a correlation between lightning activity and the position of the South Pacific Convergence Zone (SPCZ). Their study showed that location of the maximum lightning activity in this zone roughly coincides with the monthly mean position of the SPCZ. Following this study, in this paper, we examine whether similar relationship exists in the monsoon trough region over India, which is the semipermanent feature of the Indian southwest monsoon season and is also characterized by the low pressure and moist convection over the surface.
 During the Indian summer monsoon season, the trough of low pressure in the lower and middle tropospheric westerlies is one of the most significant features over central India. This trough is fully developed in the months of July–August and, on the mean, its axis at sea level runs from west-northwest (with its western end over west Rajasthan and Pakistan) to east-southeast (with its eastern end over northern Bay of Bengal), almost parallel to and ~300 km south of the Himalayas [Normand, 1943]. The position of the monsoon trough varies on different temporal and spatial scales, which has definite bearing on the Indian summer monsoon activity [Rao, 1976; Sikka and Narsimha, 1995]. Rainfall weakens over most of north India when the trough moves north of its normal position and lies close to the Himalayas. On the other hand, there are heavy rains in the north Indian plains when the trough moves back to its normal position. The location of the axis of the monsoon trough is responsible for variations in the intensity of monsoon rainfall over northern India [Raghavan, 1973]. Paul and Sikka  showed that the mean position of the trough moved between 28.9°N at 73.25°E and 21.6°N at 93.75°E during the July–August period of 1946–1965. Large differences exist in the dynamical and convective characteristics of the eastern and western parts of the monsoon trough. Sikka and Narsimha  have reviewed and summarized some characteristics as those shown below. The western end of the monsoon trough is characterized by the heat low with very shallow ascent with limited convergence in the lowest half kilometer or so and clear sky above whereas the eastern end is characterized by the dynamic trough with convergence up to midtroposphere and moist ascent throughout the troposphere with intense convection and clouding. Deep moist convection and unsaturated moist processes occur alternatively in the middle region of the trough. The trough, a region of instability associated with vortices on the scales of 500–2000 km, forms and moves within the trough with quasi–regularity.
 The objective of this study is to analyze the relation between lightning activity, precipitation, and the position of the monsoon trough during the Indian summer monsoon season from June to September for two years, 2002 (a drought year) and 2003 (a normal monsoon year) from the data obtained from the lightning imaging sensor (LIS) mounted on the Tropical Rainfall Measuring Mission (TRMM) satellite. The monthly distributions of the maximum lightning activity are compared with the monthly mean position of the monsoon trough. The monsoon trough data during 23–28 July 2002, when the monsoon trough shifts to foothills of the Himalayas, is also studied. An empirical orthogonal function (EOF) analysis is performed on the lightning flash rates and precipitation data in the study region. The first two dominant modes of lightning flash rates along with those of the precipitation are analyzed in the monsoon trough region for the 1995–2005 period.
2 Data Sets
 In this study, we use the lightning imaging sensor (LIS) data (version 2) from the Tropical Rainfall Measuring Mission (TRMM) satellite. The lightning flash locations are available on a monthly basis. From the monthly distribution of global lightning data, the lightning activity distribution in the study region (20°N–32°N latitude and 70°E–95°E longitude) can be drawn. The LIS data allows us to locate the spots of maximum lightning activity in the area covered by the monsoon trough. The lightning flash rate is obtained from the NASA website for the period 1995–2005. The lightning flash count data for the 1998–2010 period is obtained from the same website. The Indian Daily Weather Report (IDWR) provides the daily position of the sea level monsoon trough for the year 2002 and 2003. Precipitation data is obtained from the National Centers for Environmental Prediction reanalysis data sets (http://nomads.ncep.noaa.gov) for the study region.
 The region of our study ranges from 20°N–31°N and 70°E–95°E and is divided into meshes of constant size 0.5°. The LIS can provide flashes in the 0.5° grid size. Broadly, we follow the method of analysis described in detail by Ortega and Guignes  and summarized below. The maximum lightning (ML) grid at a fixed longitude (in a column) is located. Thus, average position of the maximum lightning activity is obtained by taking the latitudinal average of the maximum lightning activity along a fixed longitude. Thus, a spatial distribution of ML with longitude is obtained in the study region. The best polynomial function of p degree fitting the ML distribution is drawn and compared to the monthly position of the monsoon trough, which is obtained from the 30 day average and standard deviation for each month of the period June–September in 2002 and 2003. To confirm our findings with larger statistics, we use the 13 year (1998–2010) flash count data obtained from the TRMM-LIS. Monthly distribution of ML for the summer monsoon months is obtained separately for each year during the 1998–2010 period. Mean values of the ML positions are calculated as the average of annual values over the 1998–2010 period. The data when the monsoon trough shifts close to the Himalayas, e.g., during 23–28 July 2002, are also obtained from the IDWR.
 The band of the maximum lightning activity is represented by the polynomial function Pa(x), where the residues of the fitting remain relatively small. The correlation of the monsoon trough position and the polynomial function is optimized by increasing the mesh size from 0.5° to 1°, i.e., from 250 km2 to 12,100 km2. To choose the polynomial curve of the best fitting, the degree of the polynomial function is varied from 2 to 5. A minimum value of ML is imposed in order to avoid large residual values. This threshold value of ML cannot be a constant and is expressed as
where X and Y are the number of meshes along the latitude and longitude, respectively, and Tnf is the total number of monthly flashes detected in the study region.
4 Results and Discussion
 Our analysis provides an interesting description of lightning activity in the region of study. Figure 1 shows the ML distributions during the summer monsoon months (June–September) of 2002 and 2003. The maximal lightning activity is located in the rectangle formed by 22°N–31°N and 70°E–95°E. The monthly mean position of the monsoon trough, the normal position of the trough, and the polynomial fitting of the ML distribution are plotted in each panel for comparison with one another. All the three curves are very close to each other in both the normal monsoon and drought years. Position of Pa(x) and the monsoon trough position are observed to follow each other during the whole monsoon period. The distribution of ML is represented by a Pa(x) (degree 2) in Figure 1. It can be seen that the position of the polynomial fitting of the ML distribution moves to the north in the month of July and August and to the south in the month of September along with the seasonal monsoon trough. This northward shift of the polynomial fitting of the ML distribution and the seasonal monsoon trough is more prominent in the extreme eastern and western parts of the area of study.
 Particularly noteworthy is the northward shift of the position of Pa(X) and monsoon trough in July 2002 as compared to that in July 2003 (Figure 1). It is noteworthy in this respect that drought in 2002 occurred mainly due to a long monsoon break during the month of July. July is normally the wettest month in a normal monsoon year but turned out to be the driest month in the 2002 monsoon season. In Figure 2, we plot the position of Pa(X), which also shifted when the monsoon trough shifted to the foothills of the Himalayas during the 23–28 July 2002 period. It clearly indicates that the position of Pa(x) varies with the position of the seasonal monsoon trough and is not much sensitive to the normal position of the monsoon trough. Therefore, similar to the result of Ortega and Guignes  in the region of the SPCZ in the southern hemisphere, our results in the region of the monsoon trough during the Indian summer monsoon season in the northern hemisphere also show that the position of polynomial function fitting the lightning maximum follows the seasonal mean position of trough.
 To check the validity of our results with larger statistical data, we studied the above relationship from 13 year (1998–2010) observations of the TRMM-LIS data. Figure 3 shows the mean ML distributions during the summer monsoon months averaged for the 1998–2010 period. A polynomial function (degree 2) fitting the ML distribution and the normal position of the trough are also plotted in Figure 3. Vertical bars on each ML point show the standard deviation of each position in the annual mean ML values over the 1998–2010 period. The agreement between the positions of normal trough and mean maximal lighting is very good in the middle portion of the trough. However, the deviation between the two somewhat increases in opposite directions on the extreme western and eastern ends of the normal trough. It is worth noting in this context that there are major differences in the convective and dynamical characteristics of the eastern and western parts of the trough. The western end corresponds to a heat low with very shallow ascent and convergence limited to the lowest half kilometer or so and clear skies above [Sawyer, 1947]. On the other hand, the eastern end is a dynamic trough. In this part, convergence occurs up to the midtroposphere and moist ascent occurs throughout the troposphere with intense convection and clouding. In the middle portion of the trough (70°E–80°E), the deep moist convection and unsaturated moist processes keep alternating.
 We have examined the relationship between flash rates and precipitation in terms of their seasonal variations by applying the empirical orthogonal function (EOF) analysis to their data sets for the period of 1995–2005. Figures 4 and 5 show the spatial distribution and time series, respectively, of the first two EOF modes, EOF1 and EOF2, of the flash rate. The first EOF mode (EOF1) is characterized by high flash rates over the western part of the monsoon trough and extends over the whole region up to the eastern end of the monsoon trough in the premonsoon months of April–June, and it accounts for 80% of the variance. The second EOF mode (EOF2), accounting for 13% of the total variance, is characterized by the variation in flash rate in the western part of the monsoon trough in the month of July–August, indicating the high flash rate values in this region during the monsoon season.
 Figures 6 and 7 show the first two modes of EOF and the corresponding time series, respectively, as applied to precipitation. Contrary to the spatial and temporal distributions of flash rate, the first mode (EOF1) of precipitation appears in the eastern part of the monsoon trough and extends to the western end of the monsoon trough. It peaks in the June–August months during the monsoon season and accounts for 78% of the variance. The second EOF mode (EOF2) of precipitation, accounting for 8% of the total variance in the eastern part of the monsoon trough and the central India region, peaks in the month of August–September.
 Thus, the first modes in flash rate and precipitation are out of phase with each other and have opposite tendencies in their spatial distributions in the area of study. Thus, the variance in precipitation alone cannot account for the variance in the lightning activity in this region. This result in this tropical area of our study is in contrast with the studies of Goodman and Buechler  and Chèze and Sauvageot  in midlatitudes asserting that flash rate and precipitation are in phase with each other.
 The position of a polynomial function fitting the maximal lightning activity in the region of the monsoon trough during the Indian summer monsoon season of 2002 (a drought year) and 2003 (a normal monsoon year) reasonably matches the position of seasonal monsoon trough throughout the monsoon season in both the normal monsoon and drought years. The TRMM-LIS data collected over a period of 13 years (1998–2010) confirm such matching of normal position of trough with the position of maximum lightning activity during the monsoon season. The correlation can be optimized by the choice of the size of the mesh grid and the definition of the limits of the rectangle. The curve of the polynomial function representing the maximal lightning activity shifts to the monsoon trough position in the foothills of the Himalayas during a monsoon break. This polynomial curve, representing the maximal lightning activity, can therefore be used as a proxy to locate the mean position of the monsoon trough in this area. The EOF analysis of flash rates and precipitation clearly shows that the spatial distribution of flash rate is out of phase with the precipitation in the first mode of the EOF. Thus, the variance in the flash rate in this region cannot be associated with the variability in precipitation.
 PRK is thankful to the Director of the Indian Institute of Tropical Meteorology, Pune, for a Research Fellowship and the necessary facilities to do this work. AKK acknowledges the support under the INSA Senior Scientist program. We acknowledge NASA for making the global lightning data available on their website (http://ghrc.msfc.nasa.gov) and NOAA for making the precipitation data available through theirs (http://nomads.ncep.noaa.gov).