Arabian Sea cyclone: Structure analysis using satellite data

Available online at www.sciencedirect.com ScienceDirect Advances in Space Research 56 (2015) 2235–2247 www.elsevier.com/locate/asr Arabian Sea cyclone: Structure analysis using satellite data Lubna Rafiq a,⇑, Thomas Blaschke b, Sapna Tajbar a b a Physics Department, Shaheed Benazir Bhutto Women University, Peshawar, Pakistan Centre for Geoinformatics, Geology & Geography Department, University of Salzburg, Salzburg, Austria Received 15 December 2014; received in revised form 27 July 2015; accepted 29 July 2015 Available online 1 August 2015 Abstract Advances in earth observation technology over the last two decades have resulted in improved forecasting of various hydrometeorological-related disasters. In this study the severe tropical cyclone Gonu (2–7 June, 2007) was investigated using multi-sensor satellite data sets (i.e. AIRS, METEOSAT, MODIS and QSCAT data) to monitor its overall structure, position, intensity, and motion. A high sea surface temperature and warm core anomalies (at 200 hPa and above) with respect to the pressure minima in the central core were found to have influenced the pattern of development of the tropical cyclone. High relative humidity in the middle troposphere was aligned with temperature minima at 850 hPa and 700 hPa; high winds (above 120 knots) and closed pressure contours were observed during the intensification stage. A contour analysis of outgoing longwave radiation (OLR) provided an explanation for the direction of movement of the cyclone. The translational movement and velocities (ground speed) of the tropical cyclone were calculated using the surface pressure of the cyclone’s central core. Statistical analyses revealed a strong correlation between the maximum wind speeds within the cyclone and various atmospheric parameters. We conclude with a discussion of the significance of these findings with regard to cyclone forecasting within the framework of early warning and disaster management. Ó 2015 COSPAR. Published by Elsevier Ltd. All rights reserved. Keywords: Arabian Sea tropical cyclones; Warm core anomaly; Relative humidity; Outgoing longwave radiation; Early warning; Disaster risk management 1. Introduction Although cyclones in the Arabian Sea are relatively infrequent and are of much lower intensity than those in the Bay of Bengal, they can influence the climatic conditions over a large area of Pakistan. They are most likely to occur in two periods: May–June and November– December. Two severe cyclones threatened the Pakistani coast in 2007: we carried out a detailed analysis of one of these – tropical cyclone Gonu – which occurred in the Arabian Sea between the 2nd and the 8th of June 2007. ⇑ Corresponding author. E-mail addresses: rafiqlu@stud.sbg.ac.at, (L. Rafiq). URL: http://www.sbbwu.edu.pk (L. Rafiq). drlubna@stud.sbg.ac.at http://dx.doi.org/10.1016/j.asr.2015.07.039 0273-1177/Ó 2015 COSPAR. Published by Elsevier Ltd. All rights reserved. This cyclone (Saffir-Simpson Category 2) brought much destruction to the north-western coastal region of the Arabian Sea, affecting Iran, Oman, and Pakistan (Khalid et al., 2009). Since tropical cyclones spend most of their time over data-void oceanic areas, observations from cyclonic fields are generally rare. Satellite-based observations (from either visible or microwave bands) offer an opportunity to monitor the cyclone’s structure by virtue of its atmospheric temperature, moisture, and cloud imaging and sounding capabilities. Consequently, they also have the potential to improve cyclone forecasting. The combination of Geostationary Operations Environmental Satellites (GOES) and Polar Orbiting Environmental Satellites (POES) is crucial for monitoring meteorological processes over scales that range from a global scale to a synoptic scale, a mesoscale, and finally, to a storm scale 2236 L. Rafiq et al. / Advances in Space Research 56 (2015) 2235–2247 Fig. 1. Large cloud clusters (tropical depression) on May 31, 2007 as seen by METEOSAT-7 IR (right) and MODIS IR (left). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Fig. 2. Track and intensity of Cyclone Gonu, (Source: IBTrACS). (Scofield et al., 2002). At the global level, the World Meteorological Organization’s Tropical Cyclone Programme (WMO-TCP) issues tropical cyclone and hurricane forecasts, warnings, and advisories. It seeks to promote and coordinate efforts to mitigate risks associated with tropical cyclones. Regional bodies worldwide have adopted standardized WMO-TCP operational plans and manuals, promoting internationally accepted procedures in terms of units, terminology, data and information exchange, operational procedures, and telecommunication of cyclone information (UNEP, 2009). Due to the low frequency of tropical cyclones over the Arabian Sea (Lander and Guard, 1998), there are few publications available that deal specifically with cyclone formation in this area. In particular, there are very few publications related to the application of satellite data to L. Rafiq et al. / Advances in Space Research 56 (2015) 2235–2247 2237 Fig. 3. MODIS IR images shows different stages of Cyclone Gonu (June 02 to June 05, 2007); white arrow shows the possible cyclone’s central locations. the study of Arabian Sea cyclones. In this paper we exploit various satellite data sets ranging from visible to near infrared regions of the electromagnetic spectrum, to monitor the surface and vertical structures of cyclone Gonu. The data sets used and some pre-processing steps are described in Section 2, below, which also includes the necessary characteristics of the sensors used. Section 3 presents a synoptic analysis (from birth to landfall) for cyclone Gonu while Section 4 presents analyses of sea surface temperature, calculation of surface velocity via GPH (pressure fields) and outgoing longwave radiation (OLR) for the cyclone Gonu and in Section 5 an analysis of its vertical structure. Section 6 depicts a statistical analysis (correlation) of the wind speed (cyclone system) with various satellite-derived parameters, and finally, Section 7 provides a summary of the main results together with some conclusions related to early warning and disaster management. 2. Data sources and descriptions 2.1. Data sources Data utilized in this study are detailed below:  AQUA AIRS: mean sea level pressure (MSLP), sea surface temperatures; outgoing longwave radiation; pressure levels at different heights (GPH); temperatures at 200 hPa, 700 hPa, and 850 hPa; relative humidity at 700 hPa and 850 hPa.  METEOSAT: visible images 2238 L. Rafiq et al. / Advances in Space Research 56 (2015) 2235–2247 Fig. 4. High sea surface temperatures in the region prior to formation of Cyclone Gonu (i.e. averages from May 23 to 30, 2007). Fig. 5. High positive SST anomalies in the region prior to the formation of Cyclone Gonu. L. Rafiq et al. / Advances in Space Research 56 (2015) 2235–2247 2239 Fig. 6. High SST over the Arabian Sea on June 02, 2007.  MODIS: water vapour IR images  QSCAT: wind speed 2.2. Data descriptions  The Atmospheric Infrared Sounder (AIRS) is a hyperspectral infrared instrument on board the AQUA satellite that is designed to measure the Earth’s atmospheric water vapour and temperature profiles on a global scale. AIRS has 2378 infrared channels in the spectral range between 3.7 and 15.4 microns (Chahine et al., 2005), with a coarse spatial resolution of 13.5 km. We used AIRS Level-3 Daily Global 1  1° data for various variables, which were acquired from GES-DISC Interactive Online Analysis Infrastructure (Giovanni) as part of the NASA Goddard Earth Sciences Information Services Center (DISC).  Meteosat is a series of geostationary meteorological satellites operated by EUMETSAT under the Meteosat Transition Programme (MTP), providing IR satellite imagery every 30 min with a 3 km pixel resolution (Zinner et al., 2008).  The MODIS instruments currently onboard the NASA Earth Observing System (EOS) Terra and Aqua Spacecrafts are NASA facility instruments (Salomonson, 1989; King et al., 1992; Asrar and Greenstone, 1995) designed for global monitoring of land, ocean, and atmosphere. Three near-Infra Red (IR) channels located within the 0.94-mm water vapour band absorption region were implemented on MODIS for water vapour remote sensing. The daily ‘‘pixel-base d” near-IR water vapour products are standard MODIS Level 2 data products with a 1-km spatial resolution, obtained from http://lance.nasa.gov/imagery/ rapid-response/. In this study, we only used METEOSAT and MODIS imageries for synoptic analyses of the tropical cyclone.  Quick Scatterometer (QSCAT) wind data are produced by Remote Sensing Systems and provided by the NASA Ocean Vector Winds Science Team. Wind retrievals are calculated at a 25 km  25 km spatial resolution. The orbital data were mapped to a 0.25° grid and divided into 2 maps based on ascending and descending passes twice a day (Callahan, 2006). Data were obtained from www.remss.com. 3. Synoptic situation 3.1. Formative stages On the mainland of the subcontinent (Pakistan, Nepal, India, and Bangladesh) the onset of the monsoon usually occurs between mid-May and June over Kerala (the southern tip of India). A survey of METEOSAT visible and MODIS Aqua IR images covering this period reveals unusual atmospheric activity over the southern Arabian Sea (5–15° N) during 2007. On May 31st, 2007, three cloud clusters developed a significant formation of depressions 2240 L. Rafiq et al. / Advances in Space Research 56 (2015) 2235–2247 Fig. 7. GPH at 850 hPa; white arrow shows the possible cyclone’s central locations. in the region (Fig. 1). Dense cloud clusters are visible on the METEOSAT visible image (right of Fig. 1), while the same cloud structures are prominent in the MODIS IR image in water vapour channel (left of Fig. 1) i.e., the 6.5 mm channel reveals low temperatures in dark brown regions against the warmer sea background in light blue. Low temperatures (dark regions) represent a water vapour rich cyclonic environment and deep convection. Such abundant moisture is capable of feeding a developing cyclone. It is therefore not surprising that subsequent images in Fig. 3 show continued intensification of the system i.e. on 2nd, 3rd, 4th & 5th June, respectively. Of the three cloud clusters the southernmost large, dense cloud cluster is the more unusual feature, which subsequently developed into a cyclone due to the convergence of these large tropical depressions. This large low pressure area was able to develop into a cyclone because of (a) the very warm temperature of the tropical Arabian Sea (temperatures were 28–31 °C and (b) because of the latitude being P15°N, where the Coriolis force resulting from the Earth’s rotation is strong enough to encourage cyclone development. The cyclone eventually evolved into an organized pattern of dense clouds, high winds, and rain on June 2nd, 2007. 3.2. Birth to landfall stages Fig. 2 shows the overall track (June 02 to June 07 2007) and intensity (based on Saffir-Simpson Intensity scale) of cyclone Gone whereas in Fig. 3 MODIS depicts water vapour IR images indicating the different stages of cyclone Gonu (from June 02 to June 05 2007). MODIS IR images provide important information about the low-level moisture environment that typically surrounds tropical cyclones. It first appeared on June 2nd, 2007, as a discrete circulation moving slowly towards the north. On June 3rd it rapidly intensified to attain wind peaks of 80 knots, moving in a north-westerly direction, then on June 4th it became fully evolved over this region and strong circulatory winds >150 knots appear to have occurred close to the centre of the cyclone, with an eye becoming visible in the centre. On June 5th the cyclone weakened after encountering dry air and cooler waters, and early on June 6th it made a landfall on the eastern coast of Oman as the L. Rafiq et al. / Advances in Space Research 56 (2015) 2235–2247 2241 Fig. 8. OLR data for different dates; white arrow shows the possible cyclone’s central locations. strongest recorded tropical cyclone to hit the Arabian Peninsula. It then turned northward into the Gulf of Oman and dissipated on June 8th after again making a landfall on the south-eastern coast of Iran. As the remaining depression moved northward it also brought heavy rain to south-western coastal areas of Pakistan. 4. Analysis of the sea surface anomaly, GPH & outgoing longwave radiation 4.1. SST anomaly The requirement for a sea surface temperature (SST) greater than 26.5 °C necessary for a tropical low depression to develop into a cyclonic storm and continue intensifying has been referred to in many studies (Frank, 1977; Gray, 1992; Lander and Guard, 1998). Through AIRS data it has been observed that a week before the formation of cyclone Gonu the average SST over the Arabian Sea was very high, ranging between about 28 and 31 °C (Fig. 4), compared to the long-term average of 28 °C for this time of year (Shukla, 1987). A high sea-surface temperature means that a rich supply of warm water vapour is available to fuel the tropical cyclone and to maintain its destructive potential as it tracks towards the north-east. During its formation a tropical cyclone gathers momentum by feeding on warm, moist air over a warm sea. Consequently, the SST is a crucial parameter for determining the location and ultimate intensity of a tropical cyclone. The SST has a positive thermodynamic correlation with the intensity of tropical cyclones. Emanuel (1987) predicted a small increase in tropical cyclone intensity as a result of an increase in SST, in the order of 10% for an increase of 0.5 °C in the sea surface temperature. Similar estimates have also been obtained by Knutson and Tuleya (2004) and Knutson et al. (2001). These estimates, however, pertain to large-scale temporal increases in SST and not to localized spatial variations. Jullien et al. (2012) estimated that the primary energy supply for tropical cyclones is the upward latent heat flux that is directly related to the Sea Surface Temperature. Fig. 5 shows a SST anomaly over the study area with a 2242 L. Rafiq et al. / Advances in Space Research 56 (2015) 2235–2247 Fig. 9. Cyclone Gonu’s uppermost region (200 hPa), showing its “warm core” on June 02, 03, 04, and 05, 2007; black arrow shows the possible cyclone’s central locations. significant deviation of 0.4–0.8 from long term averages. The presence of a warm SST anomaly provide a major promotion of upward latent heat fluxes that supply energy to the tropical cyclone, and hence provides a positive feedback on its intensity (Vincent et al., 2012). The presence of warm SSTs on June 2nd in the eastern/east-north-eastern part of the cyclone system (Fig. 6) may have contributed to the system moving towards this warmer region data at 850 hPa of different dates to calculate the surface velocity of cyclone Gonu. AIRS based GPH data (Fig. 7) for the cyclone from June 4th and 5th reveals that the cyclone was moving north-westward towards Oman at that time. From Fig. 7 it can be concluded that the pressure minimum was centred at 18°N/64.2°E on June 4th and at 21.2°N/62.2°E on June 5th. The transitional movement of the cyclone eastwards and northwards can be estimated using the relationship: northward movement ¼ 0:01745  Rearth  ðchange in latitudeÞ 4.2. Translational movement/surface velocity via GPH The motion of a tropical cyclone is the result of complex interactions between a number of internal and external influences. The dynamic characteristics of the movement and intensification process of tropical cyclones have previously been investigated by Mohanty and Gupta (1997). As stated previously, tropical cyclones spend most of their time over data-void oceanic areas, where surface observations from cyclonic fields are generally rare. We therefore attempted to exploit AIRS GPH (geo potential height) ¼ 0:01745  6371 km  3:2 eastward movement ¼ 0:01745  Rearth  ðchange in longitudeÞ ¼ 0:01745  6371  2 Total movement ¼ northward movement þ eastward movement The total movement is therefore calculated to have been 419 km over a period of approximately 24 h, giving an average ground speed of 18 km/h. L. Rafiq et al. / Advances in Space Research 56 (2015) 2235–2247 2243 Fig. 10. High relative humidity values present at 850 hPa, on June 04 (a) and 05 (b), 2007. Pink lines show; black arrow shows possible cyclone central locations. Relative humidity (%) & blue lines show temperature (°C); black arrow shows the possible cyclone’s central locations. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Fig. 11. High relative humidity values present at 700 hPa on June 04 (a) and 05 (b), 2007. Pink lines show relative humidity (%) while blue lines show temperature (°C); black arrow shows the possible cyclone’s central locations. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) 4.3. Outgoing longwave radiation (OLR) OLR is the total longwave radiative flux (in units of W/m2), emitted to space by the earth-atmosphere system, and integrated from radiances emitted at all angles and all frequencies. OLR is not directly measured but is calculated from the retrieved state. AIRS measures OLR values in IR channels between the 650 cm1 and 2668 cm1 interval and are integrated for the radiances observed at the top of the atmosphere (Edward et al., 2000). Deep convective clouds are usually identified by their cold cloud tops which emit low values of outgoing longwave radiation (OLR). Minima in OLR thus serve as a proxy for deep convective clouds and typically correspond to regions of maximum 2244 L. Rafiq et al. / Advances in Space Research 56 (2015) 2235–2247 Fig. 12. QSCAT wind vector of GONU Cyclone on 3rd June 2007; black arrow shows the possible cyclone’s central location. Table 1 Resulting correlation coefficients between maximum speed and various atmospheric parameters. Atmospheric parameters Correlation coefficients Outgoing longwave radiation (w/sq m) Mean Sea level pressure (hPa) Temperature (°C) at 850 hPa Temperature (°C) at 700 hPa Temperature (°C) at 200 hPa Relative humidity (%) at 700 hPa Relative humidity (%) at 850 hPa Geopotential height at 850 hPa 0.81 0.96 0.93 0.92 0.58 0.77 0.91 0.37 cumulonimbus convection and inner-core intensification (Gray, 1995). A comprehensive analysis of OLR isopleths revealed that these contours became steeper and closer together as the system intensified. The OLR contour patterns within the cyclonic field tended to become aligned in the direction of movement of the cyclone (particularly visible on June 3rd and 4th). The gradient of the isopleths then decreased as the system weakened on June 7th, with landfall taking place on June 8th. 5. Vertical structure convection (Richie et al., 1993). The monitoring of convection centres (OLR minima) can therefore be useful for predicting the movement of a cyclonic storm. However, in this study AIRS-OLR data were used to indicate the direction of the system’s movement, as well as to monitor the structure of the system. Fig. 8 shows the OLR field (in watts/sq m) on June 2nd, 3rd, 4th and 5th. The lowest OLR values were observed to the east/south-east of the centre (shown as a white arrow) of the tropical cyclone for these four days, indicating an increase in deep With 2378 spectral channels, AIRS provides very detailed information on the vertical moist thermodynamic structure of the atmosphere. Therefore, in order to investigate the vertical structure of cyclone Gonu, the air temperature and relative humidity at various pressure levels were analysed. The AIRS-based standard temperature and relative humidity products are the result of the combined IR/Microwave retrievals (Edward et al., 2000). AIRS temperature products are available on a fixed pressure grid at altitudes from the surface to the mesosphere, L. Rafiq et al. / Advances in Space Research 56 (2015) 2235–2247 2245 Fig. 13. Relationship between maximum speed of the system, relative humidity, and air temperature at 700 hPa and 850 hPa. and the nominal resolution is approximately 1 km vertically in the troposphere. AIRS-based profiles of relative humidity are calculated relative to liquid water and relative to the stable phase of water (i.e.: taking into account the phase change from liquid to ice in a freezing layer). The following significant features were observed regarding the vertical structure of the cyclone:  As shown in Fig. 9, a warm anomaly at the 200 hPa level was centred over the pressure minima. No warm anomaly could be observed for lower pressure levels, probably due to the low resolution of the AIRS data.  A relative humidity high was always coincident with the temperature minima at different pressure levels i.e.; 850 hPa and 700 hPa (Figs. 10 and 11). AIRS tropospheric data for the system from different dates reveal the role of minimum vertical shear and warm anomaly with respect to pressure minima in influencing the development pattern of this severe tropical cyclone. During the cyclone’s development a contraction occurred, together with the development of a small warm core nested within a larger cool anomaly. This was particularly evident on June 4th and 5th, as shown in Fig. 9. A temperature difference of about 2–3 °C was observed for any point within a distance of about 200 km from the centre of the cyclone. The observed relationship between the temperature minima (at 200 hPa) and pressure minima (at 850 hPa) is reasonable, since the warmth of the core provides an indication of the energy of the tropical cyclone (Waliser and Graham, 1993).  High relative humidity in the lower troposphere is favourable for the development of intense cyclones since large quantities of latent heat are released and the formation of a tropical cyclone depends on the release of the latent heat contained in cumulus clouds. Figs. 10 and 11 show that a relative humidity high was centred over a corresponding tropospheric temperature at 850 hPa and 700 hPa. 6. Correlation analysis Statistical analyses were performed to determine the relationship between the maximum speed of the tropical cyclone with various environmental parameters (e.g., the mean sea level pressure (MSLP), the pressure at 850 hPa, the outgoing longwave radiation (OLR), the relative humidity (RH) at 850 hPa and 700 hPa, and the air temperature (AT) at 850 hPa, 700 hPa, and 200 hPa levels) for 6 days, i.e. from June 2nd to June 7th 2007, over the Arabian Sea. Scatterometer derived wind vectors can accurately estimate the centre of cyclone (Rao et al., 1995). The basic assumption of this technique is that there is a circular 2246 L. Rafiq et al. / Advances in Space Research 56 (2015) 2235–2247 Fig. 14. As shown in Fig. 13 but showing a relationship between maximum speed of the system and pressure (at surface and at 850 hPa), OLR, and air temperature at 200 hPa. symmetry in the inner core of the cyclone during its intense stage. Closed surface circulation is evident in Fig. 12. The maximum speed (MS) acquired from QSCAT (available on www.remss.com) was taken as a representative measure of cyclone intensity. A simple correlation analysis (Table 1) indicates that a high positive correlation exists between the MS and the RH at 850 hPa (0.93) and at 700 hPa (0.77). Fig. 13 shows this correlation for different dates, indicating that high RH values were present at 700 hPa from June 2nd to June 6th, but high MS values (>100 knots) were only observed on June 4th and 5th. Likewise, the RH values at 850 hPa correlate positively with the corresponding MS: the higher the MS, the higher the RH. The correlation analysis also indicated a good correlation between the MS and pressure values at different levels (i.e. at the surface and at 850 hPa), but a strong negative correlation was found between the MS and the MSLP, and a weak negative correlation between the MS and the pressure at 850 hPa. Fig. 14 shows these correlations for different dates and also reveals high MS values to have occurred at low MSLP values (<950 hPa) and low MS values to have occurred when the MSLP was high. This relationship makes sense since the system is more stable during low pressure conditions, with much higher moisture values, and consequently supports a greater amount of deep convective activity, which ultimately increases the severity of the cyclone, resulting in high MS values. A strong negative correlation was found between MS and AT at 850 hPa and 700 hPa, whereas a weak positive correlation was found between MS and AT at 200 hPa. 7. Summary and conclusion Detailed analyses of various satellite data sets (particularly AIRS data) have enabled us to document the following features of the severe tropical cyclone Gonu that formed over the Arabian Sea in June 2007:  High sea-surface temperatures were recorded at the northern part of the Arabian Sea prior to, and during, the formation of the cyclone. This provided sufficient fuel for the tropical cyclone Gonu to maintain its movement in a north-westerly direction.  A warm anomaly in the troposphere reaching up to the tropopause could be documented, and was found to be spatially centred over the pressure minima for all the relevant days studied i.e.; during the lifetime of the cyclone, while high relative humidity values in the lower-troposphere were centred over the temperature minima. L. Rafiq et al. / Advances in Space Research 56 (2015) 2235–2247  The AIRS data on outgoing longwave radiation (OLR) was found to be very useful for monitoring both structural changes within the cyclone and the direction of movement of the tropical disturbances.  Statistical analyses verified the existence of physical correlations between the maximum speed values for a tropical cyclone and various AIRS-derived atmospheric parameters. An analysis of various satellite data based meteorological parameters as well the corresponding results highlight that unravelling the causes of changes in cyclone activity requires not only an understanding of which factors influence their origin and development (i.e. high sea surface temperatures prior to formation), but also an understanding of which factors (parameters) influence the direction in which they will track (i.e. high relative humidity, outgoing longwave radiation, air temperature etc.) The high temporal frequency of remote sensing data provides good potential for real time monitoring of cloud movements, sea surface temperatures, and tropospheric conditions through various meteorological parameters, producing data with multiple applications for early warnings. With these data, for example, the formation and track of a tropical cyclone can be monitored and specific warnings relating to its location, timing, expected intensity, and the expected rainfall, can be issued several hours ahead of the actual impact. This study provides important insights into the formation and intensification of Arabian Sea tropical cyclones that will be useful for operational analysis and forecasting as well as for designing disaster mitigation measures, and may also play a major role in the development of cyclone warning strategies. References Asrar, G., Greenstone, R. (Eds.)., 2005. Mission to planet Earth/Earth Observing System. Reference Handbook, NASA Goddard Space Flight Center, Greenbelt, Md., 1995. Callahan, P.S., 2006. QuikSCAT Science Data Product User’s Manual, Overview and Geophysical Data Products. V3.0, D-18053-RevA, JPL. Chahine, M., Barnet, C., Olsen, E.T., Chen, L., Maddy, E., 2005. On The determination of atmospheric minor gases by the method of Vanishing partial derivatives with application to CO2. Geophys. Res. Lett. 325. Edward, T. Olsen, Fishbein, E., Hearty, T., Lee, S-Y., Irion, F.W., Kahn, B., 2000. AIRS Standard Product Quick Start. Available at: <http:// airs.nasa.gov>. Emanuel, K., 1987. The dependence of hurricane intensity on climate. Nature 326, 483–485. Frank, W.M., 1977. The structure and energetics of the tropical cyclone II: Dynamics and energetic. Mon. Weather Rev. 105, 1136–1150. 2247 Gray, W.M., 1992. Tropical cyclone formation and intensity change. In: Lighthill, James, Holland, Greg, Zhemin, Zheng, Ommanuel, Keny (Eds.), Tropical Cyclone Disasters. Peking University Press. Gray, W.M., 1995. Tropical Cyclones, Summary of the 8th IMO Lectures, WMO Bulletin April 1995. Jullien, S., Menkes, C.E., Marchesiello, P., Jourdain, N.C., Lengaigne, M., Koch-Larrouy, A., Lefèvre, J., Vincent, E.M., Faure, V., 2012. Impact of tropical cyclones on the heat budget of the South Pacific Ocean. J. Phys. Oceanogr. 42, 1882–1906. Khalid, Ahmad, Al Najar, Salvekar, P.S., 2009. Understanding the Tropical Cyclone GONU. In: 1st WMO International Conference on Tropical Cyclones and Climate Change Muscat, Sultanate of Oman, 8–11 March 2009. King, M.D., Kaufman, Y.J., Menzel, W.P., Tanre, D., 1992. Remote sensing of cloud, aerosol, and water vapor properties from the Moderate Resolution Imaging Spectrometer (MODIS). IEEE Trans. Geosci. Remote Sens. 30, 1–27. Knutson, T., Tuleya, R., 2004. Impact of CO2-induced warming on simulated hurricane intensity and precipitation: Sensitivity to the choice of climate model and convective parameterization. J. Clim. 17, 3477–3495. Knutson, T., Tuleya, R., Shen, W., Ginis, I., 2001. Impact of CO2-induced warming on hurricane intensities as simulated in a hurricane model with ocean coupling. J. Clim. 14, 2458–2468. Lander, M.A., Guard, C.P., 1998. A look at global tropical cyclone activity during 1995: contrasting high Atlantic activity with low activity in other basins. Mon. Weather Rev. 126, 1163–1173. Mohanty, U.C., Gupta, Akhilesh, 1997. Deterministic methods for prediction of tropical cyclone tracks. Mausam 48, 257–272. Rao, B.M., Kishtawal, C.M., Pal, P.K., Narayan, M.S., 1995. ERS-1 surface wind observations over a cyclone system in Bay of Bengal during November 1992. Int. J. Remote Sens. 16, 351–357. Richie, E.A., Holland, G.J., Lander, M., 1993. Contribution of mesoscale convective systems to movement and formation of tropical cyclones. In: Lighthill, James, Holland, Greg, Zhemin, Kerry, Emanues (Eds.), Tropical Cyclone Disasters. Peking University Press, pp. 286–289. Salomonson, V., 1989. MODIS: advanced facility instrument for studies of the Earth as a system. IEEE Trans. Geosci. Remote Sens. 27, 145. Scofield, R.A., Kuligowski, R.J., Qiu, S., Davenport, C., 2002. AWIPS for Satellite-derived Hydrometeorological Applications. In: 18th International Conference on IIPS Interactive Symposium on AWIPS, 2002. Shukla, J., 1987. Interannual Variability of Monsoons. In: Fein, J.S., Stephens, P.L. (Eds.). J. Willey & Sons, Momsoons New York, pp. 452–453. United Nations Environment Programme (UNEP), 2009. Early Warning Systems: State-of-Art Analysis and Future Directions. Available at: <http://na.unep.net/geas/docs/Early_Warning_System_Report.pdf>. Vincent, E.M., Madec, G., Lengaigne, M., Vialard, J., Koch-Larrouy, A., 2012. Influence of tropical cyclones on sea surface temperature seasonal cycle and ocean heat transport. Clim. Dyn., 1–20 Waliser, D.E., Graham, N.E., 1993. Convective cloud systems and warm – pool sea surface temperatures: coupled interactions and self-regulation. J. Geophys. Res. 98, 881–893. Zinner, T., Mannstein, H., Tafferner, A., 2008. Tracking and monitoring severe convection from onset over rapid development to mature phase using multi-channel Meteosat-8 SEVIRI data. Meteorol. Atmos. Phys. http://dx.doi.org/10.1007/s00703-008-0290-y.