On the epochal variation of intensity of tropical cyclones in the Arabian Sea

Transcription

1 ATMOSPHERIC SCIENCE LETTERS Atmos. Sci. Let. 14: (2013) Published online 26 June 2013 in Wiley Online Library (wileyonlinelibrary.com) DOI: /asl2.447 On the epochal variation of intensity of tropical cyclones in the Arabian Sea M. Rajeevan, 1 J. Srinivasan, 2 K. Niranjan Kumar, 2 C. Gnanaseelan 3 andm.m.ali 4 1 Ministry of Earth Sciences, Earth System Science Organization, New Delhi, , India 2 Divecha Centre for Climate Change, Indian Institute of Science, Bangalore, , India 3 Indian Institute of Tropical Meteorology, Pune, , India 4 National Remote Sensing Centre, Indian Space Research Organisation (ISRO) Balanagar, Hyderabad, , India *Correspondence to: M. Rajeevan, Ministry of Earth Sciences, Earth System Science Organization, Prithvi Bhavan, Lodi Road, New Delhi , India rajeevan61@yahoo.co.in Received: 5 February 2013 Revised: 30 May 2013 Accepted: 30 May 2013 Abstract During recent years, an increase in the intensity of pre-monsoon tropical cyclones (TCs) is observed over the Arabian Sea. This study suggests that this increase is due to epochal variability in the intensity of TCs and is associated with epochal variability in the stormambient vertical wind shear and tropical cyclone heat potential (TCHP). There is a significant increase (0.53 kj cm 2 year 1 ) of TCHP during recent years. The warmer upper ocean helps TCs to sustain or increase their intensity by an uninterrupted supply of sensible and latent heat fluxes from the ocean surface to the atmosphere. Keywords: tropical cyclones; wind shear; ocean heat content; aerosol; global warming 1. Introduction Over the Arabian Sea, tropical cyclone (TC) activity is observed mainly in two seasons, pre-monsoon (April June) and post-monsoon (October December). Compared to the Bay of Bengal, the frequency of TCs over the Arabian Sea is much smaller. On average only one TC forms over the Arabian Sea in a calendar year, which contributes to just 3% of the global total (Gray, 1968; Singh et al., 2000). Strong vertical wind shear associated with the Indian summer monsoon circulation is responsible for the suppressed TC activity during the monsoon season, especially during July and August. In a recent article, Evan et al. (2011) have reported an increase in the intensity of pre-monsoon (May June) Arabian Sea TCs during the period They showed that this trend in intensity is a consequence of a simultaneous upward trend in anthropogenic black carbon and sulfate emissions. They argued that anthropogenic aerosols cause anomalous atmospheric circulation over south Asia, which then reduce the basin-wide vertical wind shear. Reduced wind shear thus creates an environment more favorable for intensification of TCs. However, in a recent article, Wang et al. (2012) argued that the decrease of vertical wind shear in the recent epoch is caused by substantially advanced (by 15 days) TC occurrences on the early onset of the Asian summer monsoon. Evolution of the intensity of TCs depends on many factors like the storm s initial intensity, the thermodynamic state of the atmosphere through which it moves, vertical wind shear and the heat exchange between the atmosphere and the upper layer of the ocean. The model simulations of Emanuel (1999) suggested that the evolution of the intensity of TCs is controlled mostly by their initial intensity along with the thermodynamic profile of the atmosphere and upper ocean through which they move. The vertical wind shear is a major factor that influences the intensity of TCs (DeMaria, 1996; Paterson et al., 2005). Kimball and Evans (2002) in an idealized numerical simulation of a hurricane trough interaction confirmed the negative impacts of vertical wind shear on hurricane intensification. Owing to large wind shear, hurricane attains asymmetry and thus weakens due to subsidence in the eye region. However, the study by Hanley et al. (2001) demonstrated the positive effects of wind shear on intensity of TCs. While analyzing the Atlantic TCs between 1985 and 2006, they confirmed that TCs over warm water are more likely to intensify than weaken after an interaction with an upper-level trough. Recent studies have demonstrated the importance of the upper ocean thermal structure in the intensification of TCs in different ocean basins (Shay et al., 2000; Wada and Usui, 2007; Mainelli et al., 2008, Goni et al., 2009; Lin et al., 2009). The role of a warm upper ocean layer can be explained in the following way. The shear in the oceanic mixed layer due to the passage of a TC initiates internal waves on the thermocline. These waves bring cooler (below thermocline) waters near the surface where mechanical mixing by TCgenerated surface turbulence can complete the mixing to cool the upper ocean. In the case of an anomalously warm mixed layer, the thermocline inversion is stronger if everything else is equal. Therefore, it takes much stronger ocean mixed layer shear (stronger TC surface winds) for the cooler deep ocean water (from below the thermocline) to be transported to the surface. This process is one of the factors that control the intensity and intensification of TCs (Goni and Trinanes, 2003; Emanuel et al., 2004; DeMaria et al., 2005; 2013 Royal Meteorological Society

2 250 M. Rajeevan et al. Ali et al., 2007; Lin et al., 2009). Using a combination of observations and model simulations, in a recent study, Balaguru et al. (2012) demonstrated that the TC intensification is significantly affected by salinityinduced barrier layers. When TCs pass over regions with barrier layers, the increased stratification and stability within the layer reduce storm-induced vertical mixing and sea surface temperature cooling. This in turn leads to an increase in enthalpy flux from the ocean and intensification of TCs. There is a seasonal cycle in the barrier layer over the Arabian Sea. In addition to the in situ rainfall and river discharge, the negative wind stress curl also supports barrier layer formation by deepening the isothermal layer. The seasonal cycle reveals two peaks, one during November March (stronger in the southeastern Arabian Sea) with maximum barrier layer thickness of about 45 m, and the other during July September (mostly confined to the central Arabian sea) with climatological barrier layer thickness of about 60 m. However during the premonsoon period, the seasonal barrier layer is very weak (less than 15 m). Previous studies (Singh et al., 2000) have shown that there are significant decadal variations in the frequency of TCs over the north Indian Ocean which can be attributed to natural variation. In this study, we examine whether the intensity of pre-monsoon TCs over the Arabian Sea exhibits an epochal variation. We further discuss the role of epochal variability of vertical wind shear and upper ocean heat content on the observed epochal variation in the intensity of TCs over the Arabian Sea. Section 2 discusses the details of the data sets used in the study. Results on epochal variability of vertical wind shear and ocean heat content are discussed in Section 3. The conclusions are drawn in Section Data and methodology The tracks of TCs and the associated details like intensity were taken from the archives of the India Meteorological Department (IMD). The details of the tracks of TCs in digital format are now available in the new atlas brought out by the IMD recently. The TC e-atlas for the north Indian Ocean is available at ( The vertical wind shear over the Arabian Sea was obtained from the National Centers for Environmental Prediction/National Centre for Atmospheric Research (NCEP/NCAR) reanalysis data. For examining ocean heat content, we considered two different data sets. The first data set is a new data product developed by the National Oceanic and Atmosphere Administration (NOAA), namely, tropical cyclone heat potential (TCHP). It is defined as a measure of integrated vertical temperature from the sea surface to the depth of the 26 C isotherm. This parameter is computed globally over the tropical oceans from the altimeter-derived vertical temperature profiles estimates in the upper ocean (Shay et al., 2000). Previous studies have shown that variations in the sea surface height anomaly (SSHA) and the thermal structure of the upper ocean are strongly correlated (e.g. Goni et al., 1996; Goni et al., 2009). More details of the data set and the results of real-time monitoring of TCHP (Goni and Trinanes, 2003) are available at The TCHP is calculated as the integrated heat content excess per unit area relative to the 26 C isotherm, integrated from the 26 C isotherm depth to the surface. The TCHP data is derived according to Equation (7) of Shay et al. (2000). TCHP (x, y, t) is calculated as: TCHP (x, y, t) = C ρ p z T (x, y) H 2 (x, y, t) (1) where C p is the heat capacity of the seawater at constant pressure, ρ is the average seawater density of the upper ocean, T z (x, y) is the mean vertical temperature gradient between the surface and the 26 C isotherm. H is the depth of the 26 C isotherm. More details of the calculations are given in Shay et al. (2000). A summary of the estimation of TCHP from altimeter observations and its validation with in situ measurements over the north Indian Ocean during is given by Nagamani et al. (2012). As the TCHP derived from the satellite altimeter data is limited to recent years ( ), we have also used the TCHP derived from the inter-annual three monthly mean temperature profiles (using the Equation (1)) of the World Ocean Atlas (WOA), which is available for the period of at gov/oc5/3m_heat_content/anomaly_data.html. To examine the consistency of TCHP with other data sets, monthly mean TCHP was derived from the SODA analysis (Carton et al., 2000) also for the common period There is a reasonable similarity in the spatial distribution of mean TCHP over the north Indian Ocean among the three data sets during the common period of There is a good correlation (0.85) between TCHP over north Arabian Sea derived from satellite data and WOA data for the April June period of The statistical significance of correlation coefficient was determined by testing a null hypothesis that the correlation in the population is zero and performing a t-test. The correlation of 0.85 is thus statistically significant at 99% significance level. The correlation between TCHP derived from satellite and SODA data for the same period is 0.70, which is also significant at the 99% significance level. The TCHP values calculated may be sensitive to the depth of 26 Cisothermused for the calculations. For this analysis, we have further used TCHP derived from satellite and WOA data. 3. Results and discussions In the following subsections, we address the role of epochal variability of vertical wind shear and the ocean

3 Intensity of tropical cyclones 251 (a) (b) (c) Figure 1. Tracks of intense tropical cyclones formed over the Arabian Sea during the pre-monsoon season (May June) of (a) , (b) and (c) epochs. Low-pressure systems with maximum wind exceeding 48 knots (Severe Cyclonic Storm category) were considered as intense TC.

4 252 M. Rajeevan et al. heat content over the Arabian Sea on the observed increase in intensity of TCs Epochal variability Assuming that low vertical wind shear is crucial for intensification of TCs during their lifetime, it is important to examine whether any natural variability like multidecadal variation exists in the vertical wind shear and the frequency of intense TCs. Previous studies (Singh et al., 2000) have noted significant decadal natural variability of TC activity over the Arabian Sea. The study by Singh et al. (2000) revealed significant cycles of 13 and 29 years over the Arabian Sea during the May June season. Knutson et al. (2010) discussed the role of natural variability in the multidecadal variation of TCs over the north Atlantic Ocean. Evan et al. (2011) considered two epochs ( and ) to examine the changes in the vertical wind shear (difference of vector winds at 850 and 200 hpa) and associated link with the intensity of TCs. However, these two epochs are of not equal periods. For this analysis, three epochs are considered for which a longer time series was considered. The tracks of TCs, upper air data from the Indian region and ocean data with reasonable quality are available from mid-1950s. Therefore, the data for a longer period of 57 years ( ) were considered for this analysis. The wind shear and ocean heat content data of 57-year period were divided into three equal epochs of 19 years, i.e , and The last two epochs considered in this study have close similarity with the two epochs ( and ) considered by Evan et al. (2011). The three epochs considered in this study also bear close correspondence with large-scale indices of global climate variability such as the Pacific Decadal Oscillation (PDO). The first and third epochs have correspondence with the positive phase of the PDO, whereas the second phase has correspondence with the negative phase. Evan et al. (2011) used the maximum wind speed observed in the lifetime of a storm as a measure of the intensity. However, this parameter can be calculated only using accurate wind speed estimations or observations. Since satellite data were not available before the mid-1970s, we have considered two other parameters, frequency of intense TCs (storms with maximum wind speed of 48 knot or more, which will include the categories of severe cyclonic and very severe cyclonic storms) and number of calendar days in a season (April June) with intense TCs. Figure 1 shows the tracks of intense TCs observed over the Arabian Sea during the pre-monsoon season (May June) in the three epochs, , and These tracks are taken from the best tracks prepared by the IMD. The increasing trend of intense TCs during the recent epoch of is clearly evident in Figure 1. During Table I. Frequency of intense TCs and total number of intense TC days observed over the Arabian Sea during the three epochs Frequency of intense TCs (severe cyclonic storm and above) Days of intense TCs the epoch , there were only four intense TCs, whereas during the epochs of and , there were six and seven intense TCs, respectively. Appreciable epochal variability is also noted in intense TC days as shown in Table I. During the first epoch , there were 17 intense TC days, which is comparable with the number of days during the recent epoch, (20). However, during the epoch , the number of days with intense TCs was just 6. However, four intense TCs formed during this epoch were very short-lived. Therefore, there is a statistically significant multidecadal variation of number of intense TC days over the Arabian Sea during the pre-monsoon season. To demonstrate the influence of reduced vertical wind shear on the observed trend of intense TCs, Evan et al. (2011) estimated ambient storm wind shear (vector wind difference between 850 and 200 hpa) during the two epochs. The storm-ambient wind shear is defined as the vertical wind shear at the location of the storm averaged over the period during which the TC s intensity increases from 17 m s 1 to its lifetime maximum intensity (LMI). Evan et al. (2011) averaged vertical shear values for every cyclone position 48 h before the arrival of the storm. This was performed to decrease the contamination of the reanalysis fields by the presence of storms themselves. Their results suggested a pronounced shift toward lower values of the shear from the earlier to the later period. As mentioned earlier, the two epochs considered by Evan et al. (2011) are almost similar to the last two epochs considered in this study. To examine the variability of vertical wind shear during the different epochs considered here, we have calculated the composite wind shear for these periods as performed by Evan et al. (2011). The wind shear is calculated as the vector difference of winds between 850 and 200 hpa levels. Vertical shear values for every cyclone fix 48 h before the arrival of the storm were averaged to decrease contamination of the reanalysis fields by the storms themselves. To examine whether the variability is predominantly natural, the analysis should include consideration of the base climate in which the TCs are evolving. Thus, we have calculated the shear anomaly the shear excess/deficit of the TC environment over the corresponding epoch mean at the same time relative to genesis. This accounts for both changes in the underlying climate and the local TC environment. The results are shown in Figure 2(a). It shows the significant epochal variability of wind

5 Intensity of tropical cyclones 253 (a) (b) Figure 2. (a) Storm-ambient vertical shear anomaly (m s 1 ) calculated during the pre-monsoon season (April June) in the periods , and The box plots show the mean (square) and interquartile range (25th 75th percentile) along with 5th and 95th percentiles. (b) Same as (a) but for the storm-ambient vertical shear (m s 1 ). shear anomaly of the TC environment. During the epoch of and , the composite wind shear anomaly was much less than the composite wind shear anomaly of the epoch, There is ample modeling-based and observation evidence showing that TCs will only intensify if the vertical wind shear in the region of the storm is below a range of 8 12 m s 1 (Paterson et al., 2005). Therefore, the epochal variation of magnitude of vertical wind shear during different epochs is also calculated as done by Evan et al. (2011) and shown in Figure 2(b). The results suggest an epochal variation of wind shear as shown in Figure 2(a) also. Thus, the recent decrease in the reduction of the storm-ambient wind shear could be just a part of the epochal variability of wind shear observed over the Arabian Sea basin during the premonsoon season. This epochal variability in the vertical wind shear can explain partly the observed epochal variation of the intense TCs over the Arabian Sea. In the next section, we analyze the role of the oceanic heat content on the intensification of TCs over the Arabian Sea Oceanic heat content To examine the possible role of oceanic heat content on the observed intensification of TCs over the Arabian Sea, the monthly data of TCHP during the premonsoon season were examined. The spatial variation of linear trends of April June averaged TCHP calculated using the WOA data for the period, is shown in Figure 3 (top panel). The results show an increase in TCHP over the north Indian Ocean during the recent epoch of Over the Arabian Sea, the trends are of the order of 0.5 kj cm 2 year 1. As the epochal variation is our primary concern, it is worthwhile to examine how TCHP changed between the consecutive epochs. The spatial plots of differences in mean TCHP between the two consecutive epochs are shown in Figure 3. The middle panel in Figure 3 shows the difference in mean TCHP between the consecutive epochs and The bottom panel shows the same but for the consecutive epochs and The results suggest an increase in TCHP over the north Arabian Sea during the recent epoch ( ), with respect to the previous epoch ( ). However, there was also a decrease in mean TCHP during the middle epoch ( ) with respect to the first epoch ( ). These differences of TCHP over the north Arabian Sea shown in middle and bottom panels are statistically significant at 90% level. These results suggest an epochal variation of TCHP over the Arabian Sea. The same inference can be derived by examining the time series of area averaged TCHP over the north Arabian Sea (15 25 N, E), which is shown in Figure 4. The 11-year running mean shows epochal variations, with higher TCHP during the first and third epochs compared to the second epoch. These results are therefore consistent with the results shown in Figure 3. Mesoscale ocean features with minimum TCHP values of 50 kj cm 2 contribute to intensification of strong storms (Goni et al., 2009). Time series of percentage area with TCHP values exceeding 50 kj cm 2 over the north Arabian Sea also revealed epochal variations. Out of 19 years in the middle epoch of , in 13 years, the area with TCHP exceeding 50 kj cm 2 was below the long-term average. Compared to the middle epoch, the recent epoch witnessed an increase of about 6% in the area with TCHP exceeding 50 kj cm 2 and 26% in TCHP exceeding 90 kj cm 2. Thus, the observed epochal variations of intense TCs over the Arabian Sea are consistent with the epochal variations of TCHP. Over the period of analysis, there is an increasing trend of TCHP also as shown in Figure 3 (top panel). The increasing trends of TCHP are also evident in the area averaged time series of TCHP derived from satellite and WOA data as shown in Figure 4. Both the time series show a warming trend of TCHP during the recent epoch of The area averaged trend of TCHP derived from the satellite data is 0.53 kj cm 2 year 1, whereas the same trend derived

6 254 M. Rajeevan et al. Figure 3. Linear trends of pre-monsoon TCHP (kj cm 2 year 1 ) derived from WOA data during the period (top panel). Middle panel shows the difference in the mean TCHP between two periods, and The bottom panel shows the same but for the periods and from the WOA data is 0.41 kj cm 2 year 1. These trends are statistically significant at 95% significance level. Therefore, both the TCHP data sets show similar increasing trends in pre-monsoon TCHP over the Arabian Sea during the recent years. As discussed above, warmer upper ocean heat content helps TCs to sustain or increase intensity during the course of its life time. Therefore, we conclude that the increase in TCHP during the recent epoch ( ) could also be responsible (in addition to reduced wind shear) for the observed increase in pre-monsoon intense TCs over the Arabian Sea during the recent years. 4. Conclusions In this study, the role of epochal variability of vertical wind shear and ocean heat content on the epochal variations of intense TCs during the pre-monsoon season over the Arabian Sea is examined. The frequency of intense TCs (severe cyclonic storm and above) and days with intense TCs over the Arabian Sea during the pre-monsoon season shows significant epochal variations. The three epochs considered in this study also bear close correspondence with large-scale indices of global climate variability such as the PDO. The TC activity during the recent epoch, was stronger with a greater number of intense TCs and intense TC days, compared to the previous epoch of However, TC activity was stronger during the first epoch of also. While assuming the important role of vertical wind shear on the intensification of TCs, we have demonstrated that the higher frequency of intense TCs over the Arabian Sea during the epochs and may be attributed to the epochal variability in the storm-ambient vertical wind shear anomaly over the ocean basin. Another important factor which may explain the recent increase in storm intensity is the role of upper ocean heat content. The trend analysis using ocean heat content revealed a significant increase in the heat content over the Arabian Sea in the recent years. The increase in TCHP implies a warmer upper ocean, which helps TCs to sustain or increase their

7 Intensity of tropical cyclones 255 Figure 4. Time series of area averaged (15 25 N, E) pre-monsoon TCHP over the Arabian Sea for the period derived from the WOA data set. The time series of TCHP derived from the Satellite data for the period is shown as red line. The linear trend lines calculated for the period are also shown. The 11-year running mean of WOA-derived TCHP is shown as a blue line. intensity by an uninterrupted supply of sensible and latent heat fluxes from the ocean surface to the atmosphere (Shay et al., 2000; Goni et al., 2009). Acknowledgements We are thankful to Dr G. J. Goni, NOAA for kindly supplying us the satellite derived TCHP data set for our analysis. We are also thankful to two anonymous reviewers for their constructive comments which helped us to improve the quality of the paper. References Ali MM, Jagadeesh PSV, Jain S Effects of eddies on Bay of Bengal cyclone intensity. EOS Transactions American Geophysical Union 88(8), DOI: /2007EO Balaguru K, Chang P, Saravanan R, Leung LR, Xu Z, Li M, Hsieh J Ocean barrier layers effect on tropical cyclone intensification. Proceedings of the National Academy of Sciences 109: Carton JA, Chepurin G, Cao X, Giese B A simple ocean data assimilation analysis of the global upper ocean , Part I: methodology. Journal of Physical Oceanography 30: DeMaria M, Mainelli M, Shay LK, Knaff JA, Kaplan J Further improvements to the statistical Hurricane Intensity Prediction Scheme (SHIPS). Weather and Forecasting 20: DeMaria M The effect of vertical shear on tropical cyclone intensity change. Journal of Atmospheric Sciences 53: Emanuel K Thermodynamic control of hurricane intensity. Nature 401: Emanuel K, DesAutels C, Holloway C, Korty R Environmental control of Tropical cyclone intensity. Journal of Atmospheric Sciences 61: Evan AT, Kossin JP, Chung CE, Ramanathan V Arabian Sea tropical cyclones intensified by emissions of black carbon and other aerosols. Nature 479, DOI: /nature Goni GJ, Kamholz S, Garzoli S, Olson D Dynamics of the Brazil-Malvinas confluence based on inverted echo sounders and altimetry. Journal of Geophysical Research 101(16): Goni GJ, DeMaria M, Knaff J, Sampson C, Ginis I, Bringas F, Mavume A, Lauer C, Lin II, Ali MM, Sandery P, Ramos-Buarque S, Kang K, Mehra A, Chassgnet E, Halliwell G Applications of satellite-derived ocean measurements to tropical cyclone intensity forecasting. Oceanography 22: Goni GJ, Trinanes JA Ocean thermal structure monitoring could aid in the intensity forecast of tropical cyclones. EOS Transactions American Geophysical Union 84(51): Gray WM Global view of the origin of tropical disturbances and storms. Monthly Weather Review 96: Hanley D, Molinari J, Keyser D A composite study of the interactions between tropical cyclones and upper tropospheric troughs. Monthly weather review 129: Kimball SK, Evans JL Idealized numerical simulations of hurricane-trough interaction. Monthly weather review 130: Knutson TR, McBride JL, Chan JCL, Emanuel K, Holland G, Landsea C, Held I, Kossin JP, Srivastava AK, Sugi M Tropical cyclones and climate change. Nature Geoscience 3.3: Lin II, Chen C, Pun I, Liu WT, Wu C Warm ocean anomaly, air sea fluxes and the rapid intensification of tropical cyclone Nargis Geophysical Research Letters 36, DOI: / 2008GL Mainelli M, DeMaria M, Shay LK, Goni G Application of oceanic heat content estimation to operational forecasting of recent Atlantic category 5 Hurricanes. Weather and Forecasting 23: Nagamani PV, Ali MM, Goni GJ, Pedro D, Pezzulo JC, Udaya Bhaskar TVS, Gopalakrishna VV, Kurian N Validation of satellite-derived tropical cyclone heat potential with in situ observations in the north Indian Ocean. Remote Sensing Letters 3: Paterson LA, Hanstrum BN, Davidson NL, Weber HC Influence of environmental vertical wind shear on the intensity of Hurricanestrength tropical cyclones in the Australian region. Monthly Weather Review 133: Shay LK, Goni GJ, Black PG Effect of a warm ocean ring on hurricane Opal. Monthly Weather Review 128: Singh OP, Ali Khan TM, Rahman MS Changes in the frequency of tropical cyclones over the north Indian Ocean. Meteorology and Atmospheric Physics 75: Wada A, Usui N Importance of tropical cyclone heat potential for tropical cyclone intensity and intensification in the western north Pacific J. Oceanography 63: Wang B, Xu S, Wu L Intensified Arabian Sea tropical storms. Nature 489, DOI: /nature11470.