Synoptic Climatology of Extreme Cyclone Events in Bangladesh using Global Climate Model (GCM)

Transcription

1 Synoptic Climatology of Extreme Cyclone Events in Bangladesh using Global Climate Model (GCM) MS THESIS MD. AL MUSSABBIR HOSSEN Department of Environmental Science Bangladesh Agricultural University Mymensingh-2202 June 2016

2 Synoptic Climatology of Extreme Cyclone Events in Bangladesh Using Global Climate Model (GCM) A Thesis By Md. Al Mussabbir Hossen Examination Roll No. 15 Ag. ENVS JJ 10 M Semester: January-June 2015 Registration No Session: Submitted to the Department of Environmental Science Bangladesh Agricultural University, Mymensingh in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE (MS) IN ENVIRONMENTAL SCIENCE Department of Environmental Science Bangladesh Agricultural University Mymensingh-2202 June 2016 i

3 Synoptic Climatology of Extreme Cyclone Events in Bangladesh Using Global Climate Model (GCM) A Thesis By Md. Al Mussabbir Hossen Examination Roll No. 15 Ag. ENVS JJ 10 M Semester: January-June 2015 Registration No Session: Approved as to the style and content by Dr. M. A. Farukh Supervisor Dr. Md. Shahadat Hossen Co-Supervisor Dr. Md. Shahadat Hossen Chairman, Defense Committee and Head, Department of Environmental Science Bangladesh Agricultural University Mymensingh-2202 June 2016 ii

4 iii

5 ACKNOWLEDGEMENTS All praises are due to the Almighty Allah, the Lord of the Universe, Who has kindly enabled the author to complete this piece of research work and to write this successfully in this field of science and to build up this thesis for the degree of Master of Science (MS) in Environmental Science. The author would like to thank, heartfelt respect, deepest sense of gratitude and profound sincere appreciation to his Supervisor, Dr. M. A. Farukh, Associate Professor, Department of Environmental Science, Bangladesh Agricultural University, Mymensingh, for his scholastic guidance, valuable suggestions, generous help and encouragement throughout the study period and preparation of this dissertation. The author wishes to express her sincere appreciation and indebtedness to her teacher and Co-supervisor, Dr. Md. Shahadat Hossen, Associate Professor and Head, Department of Environmental Science, Bangladesh Agricultural University, Mymensingh for his valuable assistance, constant inspiration and constructive criticism in preparing this thesis work. The author feels it a pleasure to extend her heartiest respect, deepest gratitude and cordial thanks to all course teachers Professor Dr. Md. Abdul Baten, Professor Dr. Muhammad Aslam Ali, Professor Dr. Rehana Khatun, Dr. Md. Badiuzzaman Khan, Associate Professor, Dr. Md. Azharul Islam, Associate Professor, Department of Environmental Science who encouraged and inspired him to conduct the research study and writing the thesis in time. The author is also grateful to the Bangladesh Meteorological Department for providing weather data and Hokkaido University for the opportunity of using their server. The author expresses her deepest sense of love and heartfelt gratitude to her beloved parents, younger brothers, younger sister, and all relatives for their continuous blessing, inspiration, encouragement and helps towards to process of author higher study and research work.the author also extends her thanks to the entire official and laboratory staff of the Department of Environmental Science, for their cordial help in the research work.the author humbly desires to acknowledge her heartfelt appreciation and profound thanks to all of her friends, classmates, roommates and well-wishers for their cooperation, encouragement and help in thesis research work. The author would like to acknowledge those who helped directly and indirectly during the tenure of this research study. The Author June iv

6 Abstract Bangladesh is a disaster prone country where cyclone occurs more frequently in recent decades. In this study, cyclone events from have been. Forty years ( ) temperature, relative humidity (RH) and daily sunshine hours (SH) have been analyzed from the month of July to June to find out the variability behind cyclogenesis. Cyclone data were obtained from Bangladesh Meteorological Department (BMD); Disaster Preparedness Centre, AIT; and Bangladesh Bureau of Statistics (BBS). Temperature, RH and SH data were also collected from BMD. Principal Component Analysis (PCA) and Clustering was used to find out the temperature ( 0 C) variations mostly responsible for the formation of cyclone in last 40 years ( ), and NCEP NCAR reanalysis data were used to find out the distribution of temperature anomaly over Bangladesh and Bay of Bengal (BB). Among the 22 extreme cyclones in last 40 years, classified cluster 1 belongs only 1 cyclone when the temperature was about 42 0 C which indicates too hot condition on that particular day. It is also visible from the synoptic feature that there was a formation of relatively warmer zone (positive anomaly) near to surface level in the southern part of Bangladesh. Again cluster 2 belong 9 cyclones those were occurred when the temperature was about 34 0 C. In case of cluster 3, rests of the 12 cyclones were grouped when temperature was about 35 0 C. Therefore, cluster 2 and 3 indicates that temperature of about 34 0 C to 35 0 C were mostly responsible for the formation of total 21 cyclones in past 40 years. For the clusters 2, and 3, existence of a relatively cooler zone (strong negative anomaly) near to surface level was also evident. At the same time, the upper atmosphere at 850hPa, 700hPa, 500hPa, and 300hPa level was dominated by the development of relatively warmer zone (strong positive anomaly). So, this study reveals that the whole of Bangladesh was dominated by positive anomaly up to 300 hpa level implies formation of a deep warmer zone over this region. The larger warmer air mass at upper atmosphere could create profound influence to develop huge instability throughout the whole atmospheric column and may led to the genesis of extreme weather phenomenon like severe cyclones in Bangladesh s coastal areas. The findings of this research could be very useful for the environmental scientists to forecast severe storm weather as well as to understand the process of cyclogenesis in the coastal areas of Bangladesh. v

7 Table of Contents Title Acknowledgement Abstract Table of Contents List of Tables List of Figures Abbreviations Symbols Chapter I. Introduction Chapter II. Review of Literature 2.1. Synoptic Climatology Synoptic climatology (world context) Synoptic climatology (Bangladesh context) 2.2. Study on Tropical Cyclone 2.3. Extreme Temperature Event Extreme temperature event (world context) Extreme temperature event (Bangladesh context) 2.4. Global Climate Model (GCM) GCM study (world context) GCM study (Bangladesh context) 2.5. Climatological Factors Sun-Shine Hours and Relative Humidity 2.6. Coastal Regions in Bangladesh Chapter III. Materials and Methods 3.1. Study Area 3.2. Selection of the Study Year 3.3. Data Sources Cyclone data Weather Data 3.4. Data Processing and Analyses Principal component analysis (PCA) Clustering Page No. iv v vi viii ix xii xiv vi

8 Table of Contents (Continued) Title Page No Composite mapping and General Circulation Model (GCM) Chapter IV. Results and Discussion 4.1. Temperature, Relative Humidity and Sunshine Hours Satkhira Khulna Mongla Khepupara Barisal Bhola Feni Sandwip Sitakundo Chittagong Kutubdia Cox s Bazar Teknaf Ensemble Means Clustering Synoptic Climatology using GCM Chapter V. Conclusions References vii

9 Table No. List of Tables Title of the Table Page No Stations in Southern Coastal Region of Bangladesh with their Coordinates and Elevation from Sea Level. 20 viii

10 List of Figures Figures No. Title of the Figures Page No GIS plotting of 16 stations in southern coastal region of Bangladesh 4.1. (a) Location of Satkhira, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) relative humidity (%) and (d) sunshine hours from (a) Location of Khulna, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) relative humidity (%) and (d) sunshine hours from (a) Location of Mongla, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) relative humidity (%) and (d) sunshine hours from (a) Location of Khepupara, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) relative humidity (%) and (d) sunshine hours from (a) Location of Barisal, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) relative humidity (%) and (d) sunshine hours from (a) Location of Bhola, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) relative humidity (%) and (d) sunshine hours from (a) Location of Patuakhali, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) relative humidity (%) and (d) sunshine hours from (a) Location of Hatiya, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) relative humidity (%) and (d) sunshine hours from (a) Location of Chandpur, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) relative humidity (%) and (d) sunshine hours from (a) Location of Feni, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) relative humidity (%) and ix

11 (d) sunshine hours from (a) Location of Sandwip, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) relative humidity (%) and (d) sunshine hours from (a) Location of Sitakundo, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) relative humidity (%) and (d) sunshine hours from (a) Location of Chittagong, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) relative humidity (%) and (d) sunshine hours from (a) Location of Kutubdia, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) relative humidity (%) and (d) sunshine hours from (a) Location of Cox s Bazar, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) relative humidity (%) and (d) sunshine hours from (a) Location of Teknaf, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) relative humidity (%) and (d) sunshine hours from Ensemble means of 16 stations air temperature ( 0 C) from Ensemble means of 16 stations relative humidity (%) from Ensemble means of 16 stations sunshine hours (hr.) from Clustering of temperature ( 0 C) with the sea level pressures (hpa) of 22 cyclones occurring days from The patterns of composite air temperature ( K) anomaly for cluster 1 compared with climatology for cyclone occurring at (a) surface level, (b) 850 hpa level, (c) 700 hpa level, (d) 500 hpa level, and (e) 300 hpa level The patterns of composite air temperature ( K) anomaly for cluster 2 compared with climatology for cyclone occurring at (a) surface level, (b) 850 hpa level, x

12 (c) 700 hpa level, (d) 500 hpa level, and (e) 300 hpa level The patterns of composite air temperature ( K) anomaly for cluster 3 compared with climatology for cyclone occurring at (a) surface level, (b) 850 hpa level, (c) 700 hpa level, (d) 500 hpa level, and (e) 300 hpa level 64 xi

13 Abbreviations AGCM AIT BB BMD ESRL GIS IMD JMA JRA LSMPS MRI MTCLIM NASA NCAR NCEP NIO NOAA NSLCSD PCA RCM Atmospheric General Circulation Model Asian Institute of Technology Bay of Bengal Bangladesh Meteorological Department Earth System Research Laboratory Geographic Information System Indian Meteorological Department Japan Meteorological Agency Japanese Reanalysis Project Large Scale Meteorological Patterns Meteorological Research Institute Mountain Microclimate Simulation Model National Aeronautics and Space Administration National Center for Atmospheric Research National Center for Environmental Protection North Indian Ocean National Oceanic and Atmospheric Administration Non Severe Local Convective Storm Days Principal Component Analysis Regional Climate Model xii

14 RH SAARC SH SLCSD SLP SLR SOM SSE SST WRF Relative Humidity South Asian Association for Regional Cooperation Sunshine Hour Severe Local Convective Storm Days Sea Level Pressure Sea Level Rise Self-Organizing Map Surface Meteorology and Solar Energy Sea Surface Temperature Weather Research and Forecasting xiii

15 Symbols % Percent ~ Approximately 0 C Degree Celsius 0 K Degree Kelvin hr. 300hPa 500hPa 700hPa 850hPa hpa Tmean Tmax Tmin Hour Geo-potential height at 300hPa level Geo-potential height at 500hPa level Geo-potential height at 700hPa level Geo-potential height at 850hPa level Hectopascal Mean Temperature Maximum Temperature Minimum Temperature RH mean Mean Relative Humidity RH max Maximum Relative Humidity RH min Minimum Relative Humidity SH mean Mean Sunshine Hour SH max Maximum Sunshine Hour SH min Minimum Sunshine Hour xiv

16 Chapter I Introduction

17 Chapter I Introduction Climate change is an important issue, with a variety of influences on agriculture, water, health and economy. It is now recognized that climate variability and extreme events affect society more than changes in the mean climate (IPCC, 2001), and suggests a visible human influence on global climate (AR4, 2007). The exact magnitude of the changes in the global climate is still uncertain and subject to worldwide scientific studies. It is broadly recognized that Bangladesh is more vulnerable to these changes. Indeed, it has internationally been argued that Bangladesh, as a country, may suffer the most severe impacts of climate change (Akter and Ishikawa, 2014). Bangladesh is highly vulnerable because it is a low-lying country located in the deltaic plain of the Ganges, the Brahmaputra and the Meghna and densely populated. Its national economy strongly depends on agriculture and natural resources that are sensitive to climate change and Sea Level Rise (SLR). The impact of higher temperature and more extreme weather events such as floods, cyclone, severe drought and SLR are already being felt in south Asia and will continue to intensify (Huq et al., 1999; Ali, 1999). In this connection proper planning and analyses of extreme weather events are essential for this area. Southern regions of Bangladesh have already faced some super cyclones in recent past where, the northern areas are facing drought in every year. The southern region mainly known as a coastal region of Bangladesh experiences cyclone almost every year and suffers some havoc destruction which are not manageable during the incidence. As found from the statistics that, changes in temperature played a great role to cause such disastrous cyclones in the southern coastal region of Bangladesh. Studies showed that one of the main reasons behind cyclone occurrence is the extreme temperature. As a result, there is an increasing research interest in different parts of the world on extreme temperatures and their variations. Mearns et al. (1984) and Hansen et al. (1988) stated that relatively small changes in the mean temperature could produce substantial changes in the frequency of temperature extremes. Alexander et al. (2006) showed that annual trends in the lowest and highest daily minimum and maximum temperatures in the latter half of the twentieth century increased at many locations throughout the world. Further global warming ranging 1

18 between 1.4 o C and 5.8 o C is expected by the end of the twenty-first century (IPCC, 2007), which could also lead to an increase in temperature extremes. Few researches in Bangladesh studied with the same issue specially by Mahtab (1989), Pramanik (1983), BCCCSP (1997), Farukh and Baten (2015) etc. and all have the same view that Bangladesh is one of the foremost countries extremely susceptible to the unpleasant effects of extreme temperature events. The average annual temperature of Bangladesh is expected to increase by 1.4 ± 0.6 o C by 2050 (IPCC, 2007; MoEF, 2008). The BUP-CEARS-CRU (1994) study reported 0.5 o C to 2.0 o C rise in temperature by the year It was also reported that that, for 2010 the temperature would rise by 0.3 o C and for 2070, the rise would be 1.5 o C (ADB, 1994). Again temperature has decent relationship with the sunshine hour (SH) and relative humidity (RH). The greater the SH is available, the greater the temperature will be available (Matuszko and Weglerczyk, 2014) although the increase in temperature depends on some other factors also. SH is not showing same trends in all over the world. For example, Jeglic (2006) showed different areas of Slovenia where it was reported that July had the highest SH among the twelve months. Lorenzo et al. (2009) showed same result in the Iberian Peninsula where Goodale et al. (1998) showed highest amount of sun-shine hours in the month of May and June in Ireland. But in Bangladesh, it is lowest in the months of May to August (CDMP II, 2014). It has been said that the average total SH is decreasing day by day which is also known as dimming around the world reported in many articles (Singh et al., 2012; Sayeda and Nasser, 2012). So it is surely essential to find out the current situation of SH in Bangladesh also, in relation to extreme temperature induced cyclone events RH on the other hand, has positive relationship with the temperature. It has proved several times that increase in temperature also increases the RH in the adjacent area (Lawrence, 2004; Skilling, 2009). The trends showing of average RH is increasing over the years in all over the world as well as in Bangladesh (Islam, 2014; CDMP II, 2014) The southern part of Bangladesh showed the more increasing trend of RH than the other parts (Saniruzzaman et al., 2015). This increasing pattern of RH has some severe effect on our environment as well as, it has adjacent relationship with the occurrence of cyclone also (Wu et al., 2012). So, the trends of this parameter in 2

19 Bangladesh need to be investigated to cope with the origination of severe cyclone events. Among all the climatic variables temperature, SH and RH have been used to in this study. For investigating the most effective variable, Principal Component Analysis (PCA) and clustering have been used all over the world. PCA mainly known as the variable-reduction procedure similar to factor analysis. This approach of PCA was quite successful in predicting flooding due to a major tsunami in Chile (Hebenstreit et al., 1985). Hibberd and Peregrine (1979) studied the run-up and back-wash by considering the long wave equations together with the wave front condition represented by a bore. Farukh and Yamada (2014) applied PCA to find out the synoptic climatology related to extreme snowfall events in Japan. So, PCA provides us the guidelines about how to determine the number of components to retain, interpret the rotated solution, create factor scores, and summarize the results (Fernandez, 1995). Cluster analysis applied to meteorological variables is a suitable approach for redefining the climate divisions, and its use is becoming increasingly more common in atmospheric research (Kalkstein et al., 1987; Fovell and Fovell, 1993). In climate classification, the variability of long-term temperature data is the most readily available variables (Fovell and Fovell, 1993). Unal et al. (2003) intended to define spatially homogeneous climate regions of Turkey by using a mathematical methodology called cluster analysis. Mizuta et al. (2014) made a cluster analysis on tropical SST changes at the end of the 21st century projected by CMIP5 models under the RCP8.5 scenario. Clustering has mainly used in this study so that groups with common climatic conditions of cyclone events can be identified easily. On the other hand, a General Circulation Model (GCM) is a type of climate model of the general circulation of a planetary atmosphere based on the Navier Stokes equations on a rotating sphere. GCMs are the main tools available for developing projections of an extreme weather event (Houghton et al., 2001). A little work is done on climate change scenarios for Bangladesh using regional climate model (Islam et al., 2009). Despite continuous model development, Atmospheric General Circulation models (AGCMs) still have systematic preferences in simulating the East Asian summer monsoon (Kang et al., 2002). Therefore, nowadays, the use of GCMs for 3

20 climatic variables is indispensable to assess extreme weather event like cyclone and also their effects on surrounding environment. Keeping these things in mind, the present study has done with the following objectives: to find out the synoptic climatology of extreme temperature induced cyclone events in Bangladesh, and to find out the variability of temperature, RH and SH behind cyclogenesis. 4

21 Chapter II Review of Literature

22 Chapter II Review of Literature World s climate is changing day by day due to various environmental issues. This changing climate has direct effect on the world s temperature (Akter and Ishikawa, 2014). So rising up of temperature is a vital issue now-a-days around the world which has been studied by various scientists where synoptic, extreme events, climate model etc. are some common words mainly used for their research purpose. Bangladesh is not also far behind from that though over here the scope of this type of research is very limited. Still there are some evident of these types of research of which some are closely relevant to this study. Therefore, in this chapter, some important and relevant information of articles from both around the world and also from Bangladesh perspective will be reviewed Synoptic Climatology Long time ago in 1957, Court described synoptic climatology as the totality of weather conditions over a small region. According to Barry and Perry (1973), synoptic climatology mainly studied with the atmospheric circulation classifications and the assessment of the relationship between atmospheric circulation categories and the weather elements at a given locations or region. Yarnal in 1993 stated that synoptic climatology examined climate variability of the surface environment and focused on the spatial unit of a region. More recently, Barry and Carelton (2001) defined the field as the study of the relationship between local and regional climate conditions to the atmospheric circulation. So, one of the main research directions of synoptic climatology in recent years has been its application to the output of general circulation models Synoptic climatology (world context) Cassano et al. (2005) studied about large scale circulation patterns associated with temperature and high wind extremes at Barrow Alaska. The Self-Organizing Map (SOM) algorithm was used to study circulation patterns associated with temperature and high wind extremes at there. The analysis was done by using the SOM algorithm to produce an automated 55 year synoptic climatology of daily mean temperature. Results were indicating that high temperature extreme anomalies were associated with 5

23 patterns that produce strong, southerly flow at Barrow, while low temperature extreme anomalies were associated with patterns producing strong, northerly flow at Barrow, Alaska. Crimmins (2006) used a synoptic climatological approach (weather typing) to examine the seasonal climatology of extreme fire-weather conditions across the southwest United States (Arizona and New Mexico) during the period of Three key circulation patterns representing broad southwesterly flow and large geopotential height gradients which were associated with over 80% of the extreme fire-weather days identified in the study. Nolan et al. (2013) reported that they developed a synoptic climatology for Lake El gygytgyn, Chukotka Russia, and explored modern climate trends affecting air temperatures there to aid in pale climate reconstructions of a 3.6 million-year-old sediment core taken from the lake. Their SOM approach identified 35 synoptic weather patterns that span the range of synoptic patterns influencing the study domain over the National Center for Environmental Predictions (NCEP)/ National Center for Atmospheric Research (NCAR) analysis periods Synoptic climatology (Bangladesh context) Yamane et al. (2013) examined synoptic situations of Severe Local Convective Storms (SLCS) (mesoscale severe weather associated with deep convections such as tornado and hail) during the pre-monsoon season (from March to May) in Bangladesh. They compared composite meteorological fields on Severe Local Convective Storm Days (SLCSD) with those on Non-Severe Local Convective Storm Days (NSLCSD). The temperature was higher at 800hPa over the inland area of the Indian subcontinent including Bangladesh on SLCSD than NSLCSD. At 550hPa, a trough over Bangladesh developed on SLCSD compared with NSLCSD. That led to the development of a thermal trough over the inland area of the Indian subcontinent That synoptic situation produced great potential instability of the atmosphere in Bangladesh on SLCSD during the pre-monsoon season. Akhter and Ishikawa (2014) studied the synoptic features and environmental conditions of Brahmanbaria tornado event that caused 36 fatalities, 388 injuries and huge damages of properties on 22 March, Various factors for initiation of that 6

24 terrific event were investigated through analysis of Japanese 55-year Reanalysis (JRA-55) project (50 km horizontal resolution) data and Multi-functional Transport Satellite images by Japan Meteorological Agency. In addition, radar images, radiosonde data and three hourly synoptic data of Bangladesh Meteorological Department were used to verify the reanalysis data. Ahasan et al. (2013) studied the synoptic analysis of the heavy rainfall event of 7 September 2011 that was carried out using the Weather Research and Forecasting (WRF) Model. This extraordinary rainfall event was localized over the southeast region of Bangladesh The model performance was evaluated by examining the different predicted parameters like upper and lower level circulations, wind shear, relative vorticity, convergence, moisture and rainfall. The results indicate that the WRF model was able to simulate the heavy rainfall event, and associated synoptic features reasonably well which suggested that highly localized heavy rain over southeast Bangladesh was the result of an interaction of the large scale monsoon system with cyclonic disturbances Study on Tropical Cyclone Cyclone, also known as hurricane or typhoon, is intense low pressure areas where the pressure increases from center towards the outer part (Abbott, 2004). The amount of the pressure drops in the center and the rate at which it increases outwards gives the cyclones, the intensity and the strength of winds. Most of the countries located along the periphery of the North Indian Ocean are threatened by storm surges associated with severe tropical cyclones. Recent studies indicate that the frequency and intensity of tropical cyclones originating in the Pacific Ocean have increased over the last few decades (Fan and Li, 2005). In contrast, cyclones originating from the BB and Arabian Sea have been noted to decrease since 1970 but the intensity has increased (Lal, 2001). An increase of 10 to 20% in tropical cyclone intensities for a rise in sea-surface temperature of 2 to 4 C relative to the current threshold temperature is likewise projected in East Asia, South- East Asia and South Asia (Knutson and Tuleya, 2004). Amplification in storm-surge heights could result from the occurrence of stronger winds, with increase in seasurface temperatures and low-pressures associated with tropical storms resulting in an 7

25 enhanced risk of coastal disasters along the coastal regions of East, South and South- East Asian countries (Cruz et al., 2007). The impacts of an increase in cyclone intensities in any location will be determined by any shift in the cyclone tracks (Kelly and Adger, 2000). Ali (1999) reported that Bangladesh experiences about 0.93% of the world s total tropical cyclones. Depending upon the pressure distribution in the atmospheric winds and its mode of circulation in the BB, South Asian countries experience all types of cyclones viz. cyclonic storm, severe cyclonic storm, very severe cyclonic storm and super cyclonic storm (SDMC, 2009). It is to be noted that major cyclone disasters are still continuing in Bangladesh and India (Ali, 1999). Examples are the recent Bangladesh cyclones of 1985, 1988, 1991, 1994, 1995, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2007 and 2009 (BBS, 2011). Cyclones in the South Asian region originate primarily in the North Indian Ocean (NIO) which covers the BB and Arabian Sea basins. As per the assessment of the Indian Meteorological Department (IMD), there are on an average fifteen cyclones in a year in the NIO. The frequency in the BB is roughly five to six times more than that in the Arabian Sea. This basin s season has a double peak: one in April and May, before the onset of the monsoon, and the second in October and November, just after the monsoon season (SDMC, 2008). Most of the countries located along the periphery of the North Indian Ocean are threatened by storm surges associated with severe tropical cyclones (SDMC, 2009). About 80 tropical storms (tropical cyclones with wind speeds greater than or equal to 17 ms -1 ) form in the world s waters every year (McBride, 1995). Of these, about 6.5% form in the North Indian Ocean (BB and Arabian Sea) (Neumann, 1993). Since the frequency of cyclones in the BB is about 5 to 6 times than the frequency in the Arabian Sea and the BB share comes out to be about 5.5% (IMD, 1979). Tropical cyclones generally originate on the eastern side of the North Indian Ocean and move in a west north-westerly direction. More cyclones occur in the BB than in the Arabian Sea and the ratio is 4:1 (SDMC, 2008). Further, tropical cyclones in the BB striking the east coast of India and Bangladesh usually produce a higher storm surge compared to elsewhere in the world because of the special nature of the 8

26 coastline, shallow coastal ocean bottom topography and characteristics of tide (SDMC, 2008). The Indian Ocean is inherently vulnerable to tropical cyclones. Some of the favorable conditions identified through observational facts and scientific studies in SAARC Disaster Management Centre (SDMC, 2008) for the formation of tropical cyclone especially with respect to the BB and Arabian Sea. These are (i) A warm sea (temperature in excess of 26 to 27 0 C) to a depth of 60 meters with abundant water vapor in the overlying air (by evaporation), (ii) high relative humidity of the atmosphere to a height of above 5000 m facilitates condensation of water into water droplets and clouds, (iii) releases heat energy, thereby inducing a drop in pressure, (iv) atmospheric instability that encourages formation of massive vertical cumulus clouds due to convection with condensation of rising air above the ocean, (v) Low vertical wind shear between lower and higher levels of the atmosphere that does not allow generated heat released by the cloud to get transported from the area (vertical wind shear is the rate of change of wind between higher and lower levels of the atmosphere), (vi) presence of cyclonic vorticity (vorticity is the rate of rotation of air) that initiates and favors rotation of the air, and (vii) a location over the ocean, at least 4-50 latitude away from the equator. Cyclones normally originate from deep depression which is a consequence of high amount of temperature in the coastal region which is quite common in Bangladesh. They generally occur in early summer of the month from April-May or late rainy season from November-December (Choudhury, 1992; Wisner et al., 2004; Ali 1999; Paul 2009; Farukh and Baten, 2015) Extreme Temperature Event In In this part, articles related to extreme temperature event have been reviewed Extreme temperature event (world context) Domonkos (2003) showed an extreme temperature event as various threshold values of daily temperature or daily temperature anomaly. Herrera et al. (2005) examined about the effect of extreme summer temperatures on daily mortality in two large cities: Lisbon (Portugal) and Madrid (Spain). Results revealed that in both cases there was a triggering effect on mortality when maximum daily temperature exceeded a given threshold (34 0 C in Lisbon and 36 0 C in Madrid). Lisbon showed a higher impact, 31%, as compared with Madrid at 21%. The analysis of 500hPa geopotential 9

27 height and temperature fields revealed that, despite being relatively close to each other, Lisbon and Madrid had relatively different synoptic circulation anomalies associated with their respective extreme summer temperature days. The extreme temperature trends are analyzed for a meteorological data collection station in Jeddah, Saudi Arabia over approximately last four decades stretching between years 1970 and The ratio between the seasonal mean temperatures (T mean ) of the daily mean of hottest in July and coldest in January months was Similarly the ratios between the seasonal mean temperature of daily maximum (T max ) of hottest and coldest months was while for seasonal mean temperature of daily minimum (T min ) was Significant increase was observed in hot days per year and relatively smaller decrease in hot nights. The monthly and annual mean maximum temperatures had increased more than the mean and mean minimum temperatures (Rehman and Hadhrami, 2012). Kenawy et al. (2011) investigated the spatial and temporal characteristics of extreme temperature events in northeastern Spain. The results indicated a significant increase in the frequency and intensity of most of the hot temperature extremes. An increase in warm nights, warm days, tropical nights and the annual high maximum temperature was detected in the 47-yr period. Spatially, the coastal areas along the Mediterranean Sea and the Cantabrian Sea experienced stronger warming compared with mainland areas. Given that only few earlier studies analyzed observed changes in temperature extremes at fine spatial resolution across the Iberian Peninsula. The findings of this paper help us to understand the climatology of extremes temperature. (Kenawy et al., 2011). Grotjahn et al. (2015), focused upon extreme temperature events that affect parts of North America. Extreme temperature events were associated with Large Scale Meteorological Patterns (LSMPs). Methods to define extreme events statistics and to identify and connect LSMPs to extreme temperature events were presented. Various LSMPs, ranging from synoptic to planetary scale structures, were associated with extreme temperature events. Systematic studies of: the physics of LSMP life cycles, comprehensive model assessment of LSMP-extreme temperature event linkages, and LSMP properties were needed. Modeling studies had identified the impact of large- 10

28 scale circulation anomalies and land atmosphere interactions on changes in extreme temperatures Extreme temperature event (Bangladesh context) Ali (1999) discussed the possible impacts of climate change in Bangladesh through tropical cyclones, storm surges, coastal erosion and back water effect where he focused mainly on the effect of sea surface temperature. Both qualitative and quantitative discussions were made on cyclone intensity increase for a sea surface temperature rise of 2 and 4 C. Different scenarios of storm surges under different climate change conditions are developed by using a numerical model of storm surges for the BB. Possible loss of land through beach erosion because of sea level rise due to increase temperature on the eastern coast of Bangladesh was examined. MoEF (2005) reported about the climate change which was associated with hotter summers and colder winters. The report said that the temperatures in Bangladesh had increased about 1 0 C in May and C in November between 1985 and 1998, and further temperature increases were expected. However, although the overall climate was warming, temperature extremes were increasing, and winter temperatures as low as 5 0 C had been recorded in January 2007, reportedly the lowest in 38 years. Another article where Meteorological Research Institute (MRI) global 20-km mesh Atmospheric General Circulation Model (AGCM), called MRI-AGCM has been used to simulate mean surface air temperature. Mean surface air temperature projection for Bangladesh was experimentally obtained for the period of This work found that the MRI-AGCM temperature were not directly useful in application purpose. However, after validation and calibration, acceptable performance was obtained in estimating mean surface air temperature in Bangladesh. Change of mean surface air temperature was projected about C during the period of (Rahman and Ferdousi, 2011). Razib et al. (2012) studied about the temperature variability appears to be a major aspect as a consequence of extreme climatic changes. In their paper, they tried to demonstrate the temporal as well as spatial variation patterns of extreme temperature conditions in the North-western region of Bangladesh, based on some temperaturerelated climate change indices involving Average Monthly Maximum and Minimum 11

29 Temperature (T max and T min ). The average maximum and minimum temperature records from 5 meteorological stations situated in the North-western region were applied in calculating the indices, within a 30 year study period of 1971 to Then the same indices were calculated using a Regional Climate Model (RCM) temperature projections in the future time scale of Comparison of these indices showed significant enhancement in temperature extremes in future times. Rakib (2013) demonstrated the temporal variation patterns of extreme temperature conditions in a climate-vulnerable country Bangladesh, based on some simulated temperature-related climate change indices. Averaged over 18 stations, the indices of temperature extremes indicated warming of both daily minimum and maximum temperature in the monsoon and pre-monsoon season, which influenced the most in overall temperature rise in Bangladesh in the past reference period That increase had been found more in eastern and south-western region than the other parts. Minimum temperature of the coastal and central areas of Bangladesh had shown a significant increase. The level of discomfort had also shown an upward shift. Nasher and Uddin (2013) studied monthly, seasonal and yearly highest maximum and lowest minimum temperatures of two cities. Chittagong, as the coastal city and Rajshahi, as Barind track were selected as a study area due to its respective geographical location. Monthly highest maximum and lowest minimum temperature data from for Chittagong and for Rajshahi were used for analysis. In Rajshahi, significant rising trends were found in highest maximum postmonsoon temperature, lowest minimum monsoon temperature and highest maximum temperature from July to October, June and August for lowest minimum temperature. For Chittagong, significant increasing trends found in post-monsoon highest maximum temperature, June to December highest maximum temperature except July and December lowest minimum temperature. No significant decreasing trend was found in Chittagong. Shahid et al. (2014) studied about the quantify of the human-induced changes in temperature. In his paper, a number of studies had been carried out on rainfall and temperature trends in Bangladesh in recent years, but none of the studies considered the natural variability of climate that was present in time series in the form of autocorrelation that inflated the variance of the test statistics and changes the chance of 12

30 significance. The study showed that temperature was found to increase over the entire country, a similar trend to that suggested by other authors. Farukh and Baten (2015) represented the temperature anomaly events in the southern coastal regions of Bangladesh. In this paper they used the NCEP NCAR reanalysis data to find out the distribution of temperature anomaly over Bangladesh and BB. They also used Skew-T Log-P analysis to assess vertical instability of upper atmosphere. Hasanuzzaman (2012) reported that the temperature was greatly varied in different months. From his analyses it was observed that highest and lowest maximum temperature was recorded in the month of May ( C) and January ( C), respectively. The highest and lowest minimum temperature was recorded in the month of June ( C) and December ( C), respectively. Among the districts of his paper, the highest maximum temperature was recorded in Jessore ( C), while the lowest maximum temperature was observed in Chittagong and Mymensingh ( C). In contrary, the highest minimum temperature was observed in Cox s Bazar ( C), while the lowest was observed in Sylhet ( C). One necessary condition for tropical cyclone formation is that the sea surface temperature (SST) should have a minimum temperature of about 26 to 27 C (Ali, 1999). This leads to the speculation that any rise in sea surface temperature (SST) due to climate change is likely to be accompanied by an increase in cyclone frequency (Ali, 1999). In Bangladesh, average value of temperature ranges from 18 to 28 0 C where the average maximum has found in the month of May and the minimum has found in the month of January (Rahman and Ferdousi, 2011). For the formation of cyclone the effective temperature is about 27 to 29 0 C which is responsible for category 1 to category 5 cyclone in the context of Bangladesh (Kumar et al., 2011; Khan, 2012). Again Storms are found to initiate at prevailing temperature between C and C, with an average of C and in the months of November to December, the temperature had found about C for the formation of cyclone (Chowdhury et al. 2012; Saha and Khan, 2014). Shamsad et al. (2012) reported that increase of C temperature from August to September may gear up an average of 40% cyclone activity. 13

31 2.4. Global Climate Model (GCM) Climate models use quantitative methods to simulate the interactions of the atmosphere, oceans, land surface, and ice. They are used for a variety of purposes from study of the dynamics of the weather and climate system for the projections of future climate (IPCC, 2007). NOAA (2010) defined it as a complex mathematical representation of the major climate system components (atmosphere, land surface, ocean, and sea ice), and their interactions. Researchers have utmost interest in climate modeling all over the world including Bangladesh. Some important article related to global climate modeling both in the world and also in Bangladesh has been reviewed in later part GCM study (world context) Hungerford et al. (1989) showed the Mountain Microclimate Simulation Model (MTCLIM) model predicted daily solar radiation, air temperature, relative humidity, and precipitation for mountainous sites by extrapolating data measured at National Weather Service stations. Evaluation of model outputs (solar radiation, air temperature, relative humidity, and precipitation) were presented by comparing MTCLIM-predicted data with observed data. Adjustments to coefficients were discussed relative to these evaluations and how the values might be changed to reflect different conditions from the Northern Rocky Mountains where MTCLIM was developed. Salathe et al. (2006) explained that the Global climate models did not have sufficient spatial resolution to represent the atmospheric and land surface processes. Regional climate models explicitly simulated the interactions between the large-scale weather patterns simulated by a global model and the local terrain. They had performed two 100-year climate simulations using the Weather and Research Forecasting (WRF) model developed at the National Center for Atmospheric Research (NCAR). The mesoscale simulations produced regional changes in snow cover, cloudiness, and circulation patterns associated with interactions between the large-scale climate change and the regional topography and land-water contrasts. These changes substantially altered the temperature and precipitation trends over the region relative to the global model result or statistical downscaling. 14

32 Thomas et al. (2014) represented the performance of non-hydrostatic Regional Atmospheric Modeling System (RAMS) in simulating the tropical cyclones of different intensities, formed during pre and post monsoon. Study was carried out for Orissa, Sidr, Mala, h04b, 01A, and Agni cyclones. The simulated cyclone track, winds at 850hPa and 200hPa and other thermo dynamical features associated with the development of cyclones such as vertical wind shear, mid tropospheric humidity and sea surface temperature were compared with observations. The model represented the low level circulation and upper air divergence of wind during all the cyclone cases GCM study (Bangladesh context) Rajib and Rahman (2012) featured the development of future surface temperature projections for Bangladesh on monthly resolution for each year from 2011 to 2100 applying Providing Regional Climates for Impacts Studies (PRECIS). It was found to perform reasonably well in simulating future surface temperature of Bangladesh. The applied change field in average annual temperature showed only C 1 0 C deviation from the observed values over the period from 2005 to Eventually, from the projected average temperature change during the years , it was apparent that warming in Bangladesh prevails invariably every month, which might eventually resulted in an average annual increase of 4 0 C by the year The paper also illustrated their spatial distribution with a view to identify the most vulnerable zones under consequent warming through future times. Debsarma (2013) reported about the numerical simulation of storm surges in the BB. Long continental shelf, shallow bathymetry, complex coastal geometry with lots of kinks and islands having over all funneling shape of the BB were well-known features for the highest storm surge and of the longest duration. IIT Kharagpur Model (2002) for storm surge was used for numerical simulation of storm surges near Orissa, West Bengal and Bangladesh coasts. High resolution IIT Model had been used for the simulation. Dastagir (2015) synthesized about extreme climatic events in Bangladesh in the context of the climate modeling data. The modeling results of extreme events showed significant trends in Bangladesh due to climate change. The results of these climate models were significant to show the importance of climate modeling in Bangladesh 15

33 and it would help to promote research on climate modeling in least developed countries like Bangladesh Climatological Factors Sun-Shine Hours and Relative Humidity Sun-shine hours and relative humidity another two important parameters under climatology. Lots of work has been done throughout the world about sun-shine hours and relative humidity. Bangladesh is also not far behind from that. In this part, articles mostly related to sun-shine hours and relative humidity will be reviewed. Khan et al. (2012) examined the average annual sunlight hours in Bangladesh and was compared with other developed countries like Germany and Spain, which were notable for their development in renewable energy sector. Suitable locations for solar power plants were proposed based on the maximum efficiency factors like sunlight hours, cloud coverage limits, amount of solar radiation received, type of plane etc. Some social, economic and environmental constraints regarding the implementation of solar technology were highlighted and some possible solutions were offered. Podder and Islam (2015) studied the global solar radiation over the southern coastal region of Bangladesh which was estimated from the duration of relative sunshine hours. Five models were considered to estimate the solar irradiance. A quadratic logarithmic model, relating the relative solar radiation and the relative sunshine hours was proposed for southern coastal region of Bangladesh. NASA Surface Meteorology and Solar Energy (SSE) had record of solar radiation data all over the world, measured from satellite. The estimated solar irradiance from the proposed regression model was compared with the data recorded by NASA SSE. Sunshine hours have positive relationship with the temperature (Matuszko and Weglerczyk, 2014) and temperature has direct effect on the formation of cyclone as it is stated earlier in the temperature part. In Khulna Division, the average sunshine hour is about 6.7 hours from , whereas in Barisal and Chittagong division the average is about 6.4 and 6.8 hours respectively (Sayeda and Nasser, 2012). Khan et al. (2012) reported that in these areas maximum values were ranging from 8.49 to 9.40 hours available in the month of April and May. 16

34 Islam (2014) reported the annual average humidity of 30 meteorological stations of Bangladesh which had been studied over the period ( ). Trends, periodicities and frequency distribution of the annual average humidity were found by the standard statistical techniques. It was seen that the frequency distribution of most of the stations of Bangladesh follow normal distribution. Positive trends were shown for the data of Dinajpur, Rajshahi, Mymensingh, Ishurdi, Jessore, Madaripur, Satkhira, Hatiya, Sitakunda, Teknaf and Patuakhali, while Dhaka the capital of Bangladesh had negative trend. The periodigram analyses of the annual average humidity of most of the stations showed a significant cycle of range 8 to 12 years. Khan et al. (2015) showed the effects of relative humidity on mango saplings. He reported that the effects of temperature and relative humidity on the incidence and severity of noted diseases were observed in the aforesaid locations of Bangladesh. Hasanuzzaman (2012) in his research paper showed that the maximum relative humidity was recorded in the month of September which was 82.88%, while the minimum relative humidity was recorded in the month of March which was 62.67%. He also observed that the maximum relative humidity was recorded in the district of Barisal (80.88%), while the minimum relative humidity was recorded in the district of Dhaka (66.32%). The maximum total mean of the coastal region of Bangladesh has found in the month of July which is around 88% which is much higher than the value in July of Bangladesh, recorded 81.78% (BBS, 2009). Though, according to BBS (2009), the highest value has found in the month of September pointing at 82.88%. Temperature and relative humidity relates to each other (Lawrence, 2004 ; Skilling, 2009) and relative humidity has a good influence on the formation of cyclone (Kaplan and DeMaria, 2003; Emanuel et al., 2004; Hendricks et al., 2010; Kaplan et al., 2010). Wu et al. (2012) showed that about 82% relative humidity is responsible for the occurrence of North Atlantic tropical cyclone Coastal Regions in Bangladesh Sarwar (2005) showed the impacts of sea level rise on Bangladesh which were assessed using secondary sources. The study revealed that a one meter sea level rise will affect the vast coastal area and flood plain zone of Bangladesh. It will also affect 17

35 national and food security of the country. The Sundarbans, the most important ecosystem of the country will be totally lost with one meter rise in sea level. Dasgupta et al. (2010) integrated information on climate change, hydrodynamic models, and geographic overlays to assess the vulnerability of coastal areas in Bangladesh to larger storm surges and sea-level rise by A 27-centimeter sealevel rise and 10% intensification of wind speed from global warming suggested the vulnerable zone increased in size by 69 percent given a +3-meter inundation depth and by 14% given a +1-meter inundation depth. Nasreen et al. (2013) showed their concern with climate change related risks and hazards that affected the inhabitants of coastal Bangladesh. The study findings demonstrated that the climate change had affected the livelihood of coastal people in many folds. They also reported that climate change had also created a state of unemployment among the people of coastal communities. The study identified the alternative adaptation strategies adapted by the affected coastal women and men in coastal Bangladesh. The paper exhibited that the coastal community people try to solve their problems through adopting and exploring alternative employments. Minar et al. (2013) represented the vulnerability of Bangladesh to climate change in the world. The country has already been facing several climate change effects such as increasing cyclones, flood frequency probabilities, erosion, inundation, rising water tables, salt water intrusion and biological effects. Coastal environments particularly at risk include mangroves, tidal deltas and low-lying coastal plains, sandy beaches, coastal wetlands, estuaries and coral reefs. These bio-geophysical possessions would have consequent effects on ecosystems and eventually affect socio-economic systems in the coastal zone. 18

36 Chapter III Materials and Methods

37 Chapter III Materials and Methods 3.1. Study Area Bangladesh is located between 20 o to 26 o North and 88 o to 92 o East. Traditionally the North Indian Ocean (NIO), especially the coastal region of Bay of Bengal (BB), is prone to tropical cyclone. About 6.5% of all tropical storms in the world s water form in the NIO every year (Neumann, 1993). However, geographically Bangladesh is a disaster-prone country. Almost every year the country experiences one or more tropical cyclones. From , approximately 178 severe cyclones with wind speeds of more than 87 kmh -1 formed in the BB, causing extensive loss of life and destruction of property (Alam, 2003). Most of these cyclones severely affect the southern part of the country. Therefore, the coastal area of Bangladesh surrounding the BB was selected as the study area. Satkhira, Khulna, Mongla, Khepupara, Barishal, Bhola, Patuakhali, Hatiya, Chandpur, Feni, Swandip, Sitakunda, Chittagong, Kutubdia, Cox s Bazar and Teknaf was our study area which have shown in the Fig and listed with their coordinates in Table 3.1 provided from Bangladesh Meteorological department (BMD). Fig GIS plotting of 16 stations in southern coastal regions of Bangladesh 19

38 Table Stations in Southern Coastal Regions of Bangladesh with their Coordinates and Elevation from Sea Level Sl No. Name of the Observatories Latitude (N) Longitude (E) Elevation (Meter) Degree Minutes Degree Minutes 01 Satkhira Khulna Mongla Khepupara Barisal Bhola Patuakhali Hatiya Chandpur Feni Sandwip Sitakunda Chittagong Kutubdia Cox's Bazar Teknaf Selection of the Study Year From 1975 to 2014 lots of tropical cyclone hits on Bangladesh. For example, some of the disastrous cyclone were reported in 1985, 1988, 1991, 1994, 1995, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2007, 2009 and 2013 (BBS, 2014). Among these, cyclone SIDR in 2007 and cyclone AILA in 2009 were most destructive in terms of damage of properties and fatalities. Most importantly, all the cyclones stroke in the southern part as the area is a coastal area of Bangladesh. That is why this area is selected for the synoptic study of cyclone events in the past 40 years Data Sources Cyclone data Severe cyclone events in the southern coastal regions of Bangladesh will be investigated. Here, the severe cyclone events denote the deadliest cyclone occurring days from 1975 to The in-situ data of climatic variables was collected from Bangladesh Meteorological department (BMD) at 16 measuring points (Fig. 3.1.) from 1975 to Historical data of these tropical cyclone that attacks on Bangladesh from were collected from Disaster Preparedness Center, Asian Institute of Technology (AIT), Bangkok, and Bangladesh Bureau of Statistics (BBS). The data then used to find out the cyclone occurring days from

39 Weather Data In this study, daily weather data of temperature, sunshine hours and relative humidity were analyzed which were obtained from Bangladesh Meteorological Department, Agargaon, Dhaka, Bangladesh. The collected data were three hourly daily temperature ( 0 C), average air temperature ( 0 C), maximum air temperature ( 0 C), minimum air temperature ( 0 C), three hourly relative humidity (%) and sunshine hours (hr.). To compare the parameters comprehensively, weather data from Satkhira, Khulna, Mongla, Khepupara, Barisal, Bhola, Patuakhali, Hatiya, Chandpur, Feni, Sandwip, Sitakundo, Chittagong, Kutubdia, Cox s Bazar and Teknaf stations were carefully analyzed (shown in Fig. 3.1.) Data Processing and Analyses All the collected data were then compiled, tabulated, and analyzed according to the objectives of the study. Daily average air temperature ( 0 C), maximum air temperature ( 0 C), minimum air temperature ( 0 C), relative humidity (%) and sunshine hours (hr.) were calculated using frequently used statistical software Minitab and MS Excel 2007 to analyze the variation and trend lines. The MS Office Picture Manager and Paint, NCEP (National Centers for Environmental Prediction) Global Circulation Modeling were applied to develop map and to analyze 850hPa, 700hPa, 500hPa and 300hPa air temperature and surface air temperatures as per objectives of the study Principal component analysis (PCA) Principal component analysis (PCA), also known as Empirical Orthogonal Function (EOF) analysis, is a powerful technique for the objective characterization of low dimensional linear structure in multivariate datasets. In PCA, the original variables are transformed into a smaller set of linear combinations, and attempts to identify a small set of factors that represents the underlying relationships among a group of related variables. However, PCA based on S mode (grid points as variables and days as observations) data matrix was carried out (Yarnal, 1993) after standardizing the temperature data from a spatial point of view to keep variable field s intensities in the analysis, and maintains the daily temporal scale of the original data (Esteban et al., 2005). The most efficient representation of variance in the data set is provided by correlation matrix (Barry and Carleton, 2001), and the patterns of spatial correlation is detected without possible domination by the 21

40 grid points with the largest variances (Jolliffe, 1986). Finally, the retained components were rotated using a varimax procedure to facilitate the spatial interpretation of the principal components (Yarnal, 1993) Clustering The component scores (standardized values) generated by the PCA represent the relationship between observations and retained orthogonal components. The nonhierarchical K-means method is used to cluster the observations (Hair et al., 1998). To decide the number of groups and centroids, we considered the spatial variation patterns established by PCA, i.e. components in positive and negative phases as potential groups for circulation patterns were used. Birkeland et al. (2001) and Tait and Fitzharris (1998) method were followed to create the centroids of the groups. Finally, the centroids were obtained from the average of the cases included into every group. The K-means also produces the final classification of all observations (days) with similar distribution of temperature. As the centroids were well established by the PCA grouping (reflecting a circulation pattern), no repetition of the K-means clustering was used Composite mapping and general circulation model (GCM) The composite mapping for GCM has been done through the National Centers for Environmental Prediction (NCEP) National Center for Atmosphere Research (NCAR) reanalysis project (Kalnay et al., 1996), which ensures a good resolution of atmospheric data with a grid of 2.5 resolution. Global Climate Model also known as General Circulation Model (GCM) is the most complex of climate model, since they attempt to represent the main components of the climate system in three dimensions. GCMs are the tools used to perform climate change experiments from which climate change scenarios (possible representations of how the climate will evolve) can be constructed. The design and structure of an individual GCM determines the climate change experiments that can be performed. These characteristics are limited by the scientific understanding of the climate system and by the available computing resources (Viner, 2000). Temperature anomaly for the severe 22 cyclone occurring days was obtained from Japanese 25-year reanalysis project (JRA-25) by the Japan Meteorological Agency (JMA) (Onogi et al., 2007), encompassing the region 12 N 32 N by 78 E 102 E with a spatial resolution. 22

41 Chapter IV Results and Discussion

42 Chapter IV Results and Discussions This chapter represents the results of analyses and findings of temperature, relative humidity (RH) and Sunshine hours (SH) of coastal region as it is stated in the materials and methods part that the cyclones are mostly hit in the coastal areas. So, 16 stations in the coastal regions have been represented here with their maximum, minimum, mean and mean±sd value of the parameters. In later part, clustering of the 22 cyclones has been represented and composite mapping of temperature anomaly for severe cyclones have been studied Temperature, Relative Humidity and Sunshine Hours Temperature is an important climatic parameter for creating any extreme events in the world as well as in Bangladesh. As the temperature increases from its average value then the weather tends to become unstable and the possibility of causing weather related extreme events increases. Cyclones originate from deep depression which is a consequence of high amount of temperature in the coastal region which is quite common in Bangladesh. They generally occur in early summer of the month from April-May or late rainy season from November-December (Choudhury, 1992; Wisner et al., 2004; Ali, 1999; Paul, 2009; Farukh and Baten, 2015). Two cyclonic seasons have also considered that frequently occur in Bangladesh. In Bangladesh, average value of temperature ranges from 18 to 28 0 C where the average maximum has found in the month of May and the minimum has found in the month of January (Rahman and Ferdousi, 2011). For the formation of cyclone the effective temperature is about 27 to 29 0 C which is responsible for category 1 to category 5 cyclone in the context of Bangladesh (Kumar et al., 2011; Khan, 2012). So, it can be said that any temperature with in or above of range would trigger the formation of a cyclone. RH is another important climatic factor and is also responsible for the formation of any kind of unsteady condition. Study on this parameter is also important because temperature and RH relates to each other (Lawrence, 2004; Skilling, 2009) and RH has a good influence on the formation of cyclone (Kaplan and DeMaria, 2003; Emanuel et al., 2004; Hendricks et al., 2010; Kaplan et al., 2010). Wu et al. (2012) showed that about 82% RH is responsible for the occurrence of North Atlantic 23

43 Tropical cyclone. So, RH above this level is very much responsible for the formation of cyclone in the coastal region of Bangladesh. SH (SH) is one of the most important climatic parameters though it has no direct effect on the formation of cyclone. In fact there are some indirect consequences of the SH, because SH have positive relationship with the temperature (Matuszko and Weglerczyk, 2014), and temperature has direct effect on the formation of cyclone as it is stated earlier in the temperature part. In Khulna Division, the average SH is about 6.7 hours from , whereas in Barisal and Chittagong division the average is about 6.4 and 6.8 hours, respectively (Sayeda and Nasser, 2012). Khan et al. (2012) reported that in these areas maximum values were ranging from 8.49 to 9.40 hours available in the month of April and May. The statistics of the 16 stations and the ensemble means of those 16 stations has represented and discussed in the following part of this chapter. In the first part, temperature ( 0 C), RH (%) and SH (hr.) of the 16 stations have represented where the figures are drawn by using the raw data from the year of 1975 to The box plotting shows in the integer expose the long term mean, mean+sd and mean-sd of temperature, RH and SH in these regions. The line through the box plot indicating the mean value of temperature, RH and SH. Upper portion of the box shows the mean+sd and the lower portion of the box shows the mean-sd value. The red, blue and red markers in the three types of graph represent the maximum temperature, RH and SH while the purple, green and blue markers represent the minimum temperature, RH and SH in the definite area. In the second part, study has carried out to find out the possibilities of creating unstable condition by the parameters for the definite area. Statistics of temperature, RH and SH have compared to the findings of the graph where it is seen if the findings are responsible for the formation of cyclone or not. District wise analyses with their ensemble means has been studied in the following part starting from south-western to the south-eastern region. 24

44 Satkhira Satkhira is in the south-western part of Bangladesh as it is shown in the Fig (a) and it s an ecologically important area for the country, because the huge part of the largest mangrove forest Sunderban is situated here. The area is badly affected by many cyclones occurred in the pasts and recent years in Bangladesh. So, study on temperature behavior, percentage of RH and the total SH for this station is so important to find out the influence of these parameters on cyclogenesis. Mean air temperature (T mean ), maximum air temperature (T max ), minimum air temperature (T min ) and mean±sd of Satkhira from 1975 to 2014 shown in Fig (b). It shows clearly that T mean, T max, T min and mean±sd is in decreasing trend from June to January and then it shows an increasing trend from January to May. The extreme T max of Satkhira is found in the month of May ( C) where the extreme T min ( C) was in the month of January. From the figure, it is clearly seen that the highest amount of temperature is C where the highest mean value is C both of which is in the cyclone occurring month May. It is also clear that the weather started to become unstable in the premonsoon (March to May) and post-monsoon season (August to October) mostly in this area. The maximum temperature throughout the year ranges from 29 0 C to 38 0 C which is high enough to create unstable condition in the surrounding area. This unstable condition may induce devastating cyclone. Mean RH (RH mean ), maximum RH (RH max ), minimum RH (RH min ) and mean±sd have shown in the Fig (c). The trend is increasing from July to September and then it dropped from September to March, and started to rise up again from March to June. The RH max in this area is 99.03% (January) whereas RH min is 20.23% (March). The RH max in Satkhira area was seen in August and September (99%) where their mean is around 86%. Again in April and May, the highest RH was 98% though the mean value was around 70-75%. The highest RH throughout the whole year ranged from 98-99% which indicated hot and humid situation on that particular month. The higher RH also indicates higher temperature. Both of the parameters together may create any unsteady situation. Again, the highest RH range in this area indicates powerful signal of cyclogenesis. 25

45 (a) (b) (c) (d) Fig (a) Location of Satkhira, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) RH (%) and (d) sunshine hours from SH have shown the reverse trend compared to temperature and RH. The trend is shown in the Fig (d) which increasing from the month of July to November and then in the next three months (November to January) it is decreasing. It was in increasing trend again from the month of January to April and then decreased from April and continued up to June. The highest period of SH (SH max ) in this region was in the month of April (11.30 hr.) and the lowest (SH min ) was in the month of August (0.04 hr.). From the figure, it is clearly seen that the maximum mean of SH was in April which is more than 8.5 hours. The longest period of SH in April may have a strong influence in uprising of temperature which in consequence may form any cyclonic event in Satkhira region. 26

46 Khulna Khulna is the divisional town and situated just following of Satkhira in the southwestern part shown in Fig (a) It is also a part of the largest mangrove forests Sunderban and many industries have situated here by the side of the river Poshur. As like as Satkhira, cyclone has also affected here in 1901, 1909, 1961, 1966, 1971, 1974, 1977, 1991, 1998, 2007 and 2009 in the cyclone occurring days and caused havoc destruction. For Khulna region, temperature ( 0 C) has shown in the Fig (b). From the figure, it has shown that temperature in this region in same trend like it is in Satkhira. T mean, T max, T min and mean±sd are in decreasing trend from June to January and then it shows an increasing trend from January to May. Here, T max is C that has found in the month of April and the T min has found in the month of January ( C). From the figure, it is clearly seen that the highest mean value is about C which is in the month of May. These two months (April and May) indicated the T max and highest mean value respectively are known as the cyclone occurring month. The T max ranges throughout the year is about 29 0 C to C which is very high amount and favorable for creating unstable disorder in the surrounding area which may form deep depression and cause shattering cyclone in later. RH max, RH min, RH mean and mean±sd of RH in Khulna area has shown in the Fig (c) where the results come from the analysis of due to the data 1975 was unavailable. Mean RH in Khulna area is uppermost in the month of September pointing about 89%. Again in April and May, the mean value is around 78-81%. The highest value throughout the year ranging from 98-99% and any area with in this range indicated how hot and humid it was on that certain day. These high amounts of RH also indicate great amount of temperature on such day both of which may responsible for creating any unsteady situation which in later may cause some extreme events. The highest range in this area is a strong indicator for the formation of cyclone. 27

47 (a) (b) (c) (d) Fig (a) Location of Khulna, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) RH (%) and (d) sunshine hours from The trend of SH in this region has shown in the Fig (d) where it s increasing from the month of July to November and then in the next three months (November to January) the trend is decreasing. It was in increasing trend again from the month of January to April and fallen down from the month of April and continued up to June. The SH max in this region is in the month of April (12.08 hr.) and the SH min is in the month of July (0.02 hr.). The maximum mean has also found in the month of April pointing at 8.5 hr. This highest amount of SH also indicating higher amount of temperature in April at Khulna. This higher amount of temperature may create deep depression and form cyclonic activity in this area. 28

48 Mongla Stations Mongla is a part of Khulna standing in the lower southern part of Khulna district showing in Fig (a). At this present, it is the second largest port of Bangladesh and an Export Processing Zone (EPZ) is located here depending on the ports. So the area has an equal importance in the context of economic activity due to its several times damaged by the devastating cyclone. Temperature in ( 0 C) has shown in the Fig (b) for this region. T mean, T max, T min and mean±sd is in decreasing trend from July to January and then it s increasing from January to May but drops little bit in the next month. Extreme T max has shown in the month of April which is about C and the extreme T min temperature has shown in the month of January which is about C. The highest of T mean has found in the month of May where the temperature was about C. From August to October, the T mean is ranging from 27.5 to C where the T max has found about C. Again the range of T max throughout the year is about C. So, these maximum values indicating strong possibilities for unstable weather conditions in the studied area. Difference of maximum temperature from the mean is about C which sudden rise may cause unstable condition which later may turn into cyclonic storm. RH max, RH min, RH mean and mean±sd has shown in the Fig. 4.3 (c). Data from 1975 to 1988 has not analyzed due to its unavailability. The RH max in this region has shown in the month of September, November, January and March (99.92%) whereas the minimum RH has shown only in the month of March (23.19%). So, trend of RH of Mongla region is little bit different from the stations of Satkhira and Khulna. The highest RH mean is about 89% in the month of September which indicated hot and humid circumstances in that specific month in Mongla. September is also known as the cyclonic month and this high amount of RH also may cause shaky condition in the specific area which may turn into deep depression later and ultimately forms cyclone. 29

49 (a) (b) (c) (d) (a) Fig (a) Location of Mongla, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) RH (%) and (d) sunshine hours from The trend of SH in this region has shown in the Fig (d). Long term data of SH from was not available for the analyses. The trend in the figure is increasing from the month of July to November and drops then in the next two months (November and December). It again rises up from the month of December to April, then dropped again from the month of April and continued up to June. The SH max in this region is in the month of May which is about hr. and the SH min is in the months of June, July, August and September where the days were covered by cloud and as a result no sunshine was available. From the figure, it has found that the maximum of SH mean is in the month of April (8.75 hr.). This maximum of SH mean in 30

50 April and the SH max in May in a day may have a strong influence in uprising of temperature which in consequence may form any cyclonic event in Mongla region Khepupara Khepupara is the local name which is officially known as Kalapara upazila of Patuakhali district under the division of Barisal, Bangladesh shown in the Fig (a). This area is located in the lower southern part of Bangladesh and is environmentally important, because the Daughter of Sea Kuakata is located here and it is one of the most attractive tourist places in Bangladesh. Cyclone has affected here also due to its position in the lower southern part that experiences great destruction. The station also represents the Barguna district which is just beside of the Barguna. So, study on temperature variation, behavior of RH and deviation in SH in this area is also important. For Khepupara, temperature in ( 0 C) has given in the Fig (b). It has seen from the figure that T mean, T max, T min and mean±sd of temperature is in decreasing trend from the month of July to January. Then it again rises up to May and drops little bit in the following month (June). In this area, T max has shown in the month of April ( C) and T min has shown in the month of January ( C). The range of T max in this area throughout the year is about 29 to 35 0 C where the highest T mean is about 30 0 C which is in the month of May though in post monsoon cyclonic period September and October, the mean is about C which is very much comfortable and is not suitable enough for creating any unstable condition. The high amount of temperature in the premonsoon period (April and May) indicating unstable condition at the surrounding area. Again the difference between the highest temperature and the higher mean value is about C which also indicating the strong positive signals for the formation of cyclone. The trend of RH in Khepupara area has shown in the Fig (c) where RH max, RH min, RH mean and mean±sd have represented. July has represented the highest RH mean (88.56%) and the lowest RH mean is about 73% pointing in the month of February here in Khepupara. In fact, from June to October the RH mean is more than 85%. These highest RH mean in the area indicated hot and humid circumstances in that specific months in Khepupara. September and October is also known as the cyclonic month and the high amount of RH showed in the figure also may cause insecure 31

51 (a) condition in the specific area which in later may form deep depression and ultimately forms cyclone. (b) (c) (d) Fig (a) Location of Khepupara, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) RH (%) and (d) sunshine hours from The trend of SH (hr.) in this location has shown in the illustration 4.4. (d). The data from 1975 to 1987 is missing for this area. The trend is rising from the month of July to November and then decreases from November to January. It has again risen up from the month of January to April, then it again drops from the month of April and continuing up to June. The SH max in this location is in the period of April (11.68 hr.) and the SH min is in the months of June, July, August and September where the phase was 0 period. The highest SH mean is about 8 hours which has also found in the month of April. The longest period of SH in this area indicating also higher amount of 32

52 temperature in April at Khepupara. This higher amount of temperature may form also cyclonic activity in that area Barisal The location of Barisal has shown in the Fig (a) which is a divisional area and is mostly renowned for the riverine transportation by the giant launches which is its main way of transportation. So the area is mostly covered by river. Here, the maximum, minimum, mean and mean±sd of temperature have shown in the Fig (b). The trend is similar as it is shown in the Satkhira, Khulna, Mongla and Khepupara area. The trend is increasing first from July to August, then it s started to decrease and continued up to January. Then it again rises up to May and after this month decrease slightly from May to June. The T max has found in the month of April ( C) and the T min has shown in the month of January where the temperature is C. In the pre-monsoon season (March to May), the highest T mean is about 29 0 C which is in the cyclone occurring month May. In post-monsoon season (August to October), the maximum mean value of temperature is about 28 0 C pointing at the month of September. Again the T max in this region throughout the year ranges from 28 to 36 0 C, are strongly responsible for creating unstable condition. Again the difference of the maximum temperature from the highest mean value is about C which have indicated powerful signal for creating deep depression in the southern part which may later turn into shattering cyclone. For Barisal area the RH has shown in the Fig. 4.5 (c). The highest RH mean is about 90% in the month of July and the lowest RH mean is about 76% that is in the month of March here in Barisal. This highest mean value in the area indicated hot and humid circumstances in that specific month in Barisal. This high amount of RH also may cause unstable condition in the specific area which ultimately may forms cyclone though the highest RH occurring month is not known as cyclone occurring month. SH max, SH min, SH mean and mean±sd of SH for Barisal area has shown in the figure 4.5. (d). From the figure, it has shown that the trend is rising from July to November and then drop slightly to December, but again has lifted up to February. There is little bit ups and down in the month of March and April, but after April it has started to fall and goes up to June. 33

53 (a) (b) (c) (d) Fig (a) Location of Barisal, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) RH (%) and (d) sunshine hours from The highest period of SH is in the month of April (11.55 hr.) and the lowest period of SH is in the month of June (0.05 hr.). From the figure, it is clearly seen that the maximum SH mean is in the month of April which is more than 8 hours. The maximum has found in the month of April also as it stated earlier. The maximum mean of SH and the highest period of SH both in April may have a strong effect in uprising of temperature which in concern may form any cyclonic event in Barisal region Bhola Bhola is another important area in the southern part of Bangladesh. It is covered by the rivers and that s why it is known as the largest island in Bangladesh. Maximum, 34

54 minimum, mean and mean±sd of all the parameters have shown in the Fig In 4.6 (b), the figure of temperature has shown where it is seen that like Barisal, the trend of temperature is increasing from July to August and then the trend has fallen down to January. Then it s again rises up from January to May and then drops little bit to June. The T max has shown in the month of May ( C) and the T min has shown in the month of January ( C). (a) (b) (c) (d) Fig (a) Location of Bhola, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) RH (%) and (d) sunshine hours from The T mean in this area throughout the year is ranging from C where the maximum T mean is about C is in the month of May. The T max in this area is ranging from C throughout the year also. The T max in this area has strong 35

55 influence in creating unstable condition. Again T max also shows powerful indication for the formation of cyclone by creating low atmospheric pressure in the adjacent area. The abrupt C difference of the temperature between the maximum range and the mean value are strong indication of changing temperature in the area in such a day where the condition of the environment may become unstable which in later may turn into cyclone. RH of Bhola has given in the Fig (c). July has represented the RH max (99.97%) where the RH mean is about 91%. This highest mean value in the area indicated hot and humid circumstances in that specific month in Bhola. In both cyclone occurring months, the RH mean is comparatively low and is not in at that level of forming cyclone though the maximum range is high enough in almost every month. The high amount of RH also may cause unstable condition in the specific area which in later may form deep depression and ultimately forms cyclone. SH of Bhola have given in the Fig (d). Data from was not available for the analysis of SH in Bhola. From the figure, it has seen that the trend is increasing from its starting point July and continued up to November. Then it s fall slightly from November to December but again started to rise. After April, the trend has again started to fall and continued to June. The SH max is hr. has found in the month of April and the SH min is 0.05 hr. has found in the month of July. Again the highest SH mean is about 7 hours also found in the month of April. This highest level of SH also indicating higher amount of temperature in April at Bhola which may induce atmospheric low pressure in the area and may form cyclonic activity Patuakhali Location of Patuakhali has shown in Fig (a) which is another important area of the coastal regions, as it is mostly vulnerable to cyclones. The most of the cyclones in Bangladesh had stroked here and caused massive destruction. So, it s significant to find out the temperature movement, variation of RH and the duration of SH in this area. T mean, T max, T min and mean±sd of temperature have shown in the Fig (b). The trend is increasing first from July to August and then drops from August to January. From January, it s again started to rise up to May and then drops from May 36

56 (a) to June. The T max in this area is in the month of April ( C) and the lowest one is in the month of January ( C). (b) (c) (d) Fig (a) Location of Patuakhali, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) RH (%) and (d) sunshine hours from The T mean is ranging throughout the year from C where the maximum T mean has found in the month of May which is a cyclone occurring month. Again, the T max in this region is ranging from C throughout the year which is too high for creating any unstable condition on that day in the area. Again temperature above 28 0 C is vulnerable for occurring cyclone in the coastal region of Bangladesh in the cyclone occurring months. So, here the maximum temperature showing in the graph mainly in the two period of cyclone occurring months are strong enough to create any unstable 37

57 condition which in later may generate deep depression in the surrounding area and as subsequent, cyclonic storm may form. RH from of Patuakhali area has shown in the Fig (c). Like the other one, RH max, RH min, RH mean and mean±sd of RH also has shown here. July has represented the highest RH mean pointing at 91% and the lowest RH mean is about 76% pointing in the month of February here in Patuakhali. This highest mean value in the area indicated hot and humid circumstances in that specific month in Patuakhali. The high amount of RH also may cause insecure condition in the specific area which in later may form low atmospheric arc and ultimately forms cyclone. The figure 4.7. (d) has shown the SH for the area of Patuakhali. 10 years data of SH from was not available and as a result has not shown in the graph. SH max, SH min, SH mean and mean±sd have shown in the figure. From the figure, it has seen that the line graph is increasing from its turn taper July and continuing up to November. Then it s losing slightly from November to December and then again started to rise. After April, the discernment has again started to change and dropped to June. The SH max has shown in the month of April (11.13 hr.) and the SH min has shown in the month of August and September (0.02 hr.). From the figure, it is also clearly seen that the maximum SH mean is in the month of March which is more than 7 hours. The maximum of SH mean and the SH max both in the months of March and April may have a strong effect in uprising of temperature which in concern may form any cyclonic event in Patuakhali region Hatiya Hatiya is an island mainly situated in the last part of Bangladesh showing in the Fig. 4.8 (a). It s in southern central part of Bangladesh. The area is ecologically important for Bangladesh and also badly affected by many cyclones. Analysis of temperature has also done here which is shown in Fig (b). The T max in this area is in the month of May pointing at C where the T min is C in the month of January. The uppermost T mean in this area is C which is in the month of May. The T max in this area is ranging from 28 to 35 0 C throughout the year which is not comfortable at all and blamable for unsteady situation. Again the temperature above 28 0 C is responsible for the formation of cyclone. So the maximum values in this area in the 38

58 cyclonic period are too high and powerful components for the formation of cyclonic activity in the area. (a) (b) (c) (d) Fig (a) Location of Hatiya, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) RH (%) and (d) sunshine hours from RH max, RH min, RH mean and mean±sd of RH of Hatiya have shown in the Fig. 4.8 (c). Like some of other stations, Hatiya has also shown the Highest RH mean value in the month of July which is about 89% and the lowest RH mean is about 73% pointing in the month of February here in Hatiya. This highest mean value in the area indicated hot and humid circumstances in that specific month in Hatiya. This high amount of RH also may cause insecure condition in the specific area which in later may form deep depression and ultimately forms cyclone. 39

59 The trend of SH in this part has shown in the illustration 4.8. (d). The SH max in this area is in the month of May (11.8) and the RH min is in the months of July and August (0.02) hours. The highest of SH mean is more than 7 hours in the month of November. The SH max which has found in the month of May indicating higher amount of temperature in May on such day at Hatiya. This higher amount of temperature may create low atmospheric pressure and form cyclonic activity in that area Chandpur Chandpur is another riverine district in the coastal regions of Bangladesh which has shown in the Fig (a) and almost every year, it is affected by several cyclone attacked in Bangladesh. So the area is also important to study the variations of temperature and find out its possibility for the formation of cyclone. Here the T mean, T max, T min and mean±sd have shown in the Fig. 4.9 (b). The T max in Chandpur area is in the month of May pointing at C. The T min is in the month of January which is 11 0 C. So, from the figure, it is clearly seen that the highest amount of temperature in this area is in the month of May which is a cyclone occurring month. Again the maximum of T mean has also found in the month May which is around C. Both the temperature are high enough to create unstable situation in that area. This unstable situation in later may form from low atmospheric pressure in the surrounding area which may turn into cyclonic storm. The trend line from of RH has shown in the Fig. 4.8 (c). RH max, RH min, RH mean and mean±sd of RH of last forty years has expressed here. July has represented the highest RH mean where the it is about 86% and the lowest RH mean is about 73% pointing in the month of March here in Chandpur. This highest RH mean in the area indicated hot and humid circumstances in that specific month in Chandpur. The high amount of RH also may cause insecure condition in the specific area which in later may form deep depression and ultimately forms cyclone. In the Fig. 4.9 (d), SH max, SH min, SH mean and mean±sd of weather hours for Hatiya area has shown. The SH max in this area is hr. which is in the period of April and the SH min has found in the month of June and July which is 0.02 hr. From the figure, it is clearly seen that the maximum SH mean is in the month of April which is more than 7 hours and the maximum has also found in the month of April, as it stated earlier. The maximum of 40

60 (a) SH mean and the SH max are both in April, indicating strong effect in uprising of temperature which in concern may form any cyclonic event in Chandpur region. (b) (c) (d) Fig (a) Location of Chandpur, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) RH (%) and (d) sunshine hours from Feni Feni is faced towards south and is the entrance to the south eastern part of Bangladesh. Location of Feni in Bangladesh map has shown in Fig (a). Destruction of cyclone also occurred here. The T mean, T max, T min and mean±sd of temperature have shown in the Fig The T max in this area in the month of May which is C and the T min is in the month of January pointing at C. 41

61 (a) (b) (c) (d) Fig (a) Location of Feni, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) RH (%) and (d) sunshine hours from The T max is ranging here from C which is very high and any temperature above the average value that is about 28 0 C is responsible for creating any types of unsteady situation. Both in the two cyclone periods, the maximum temperature is about 35 0 C where the T mean is about 27 0 C. So, the difference here is about 8 0 C which is very high from its mean value and this type of range may play a great role in the formation of cyclone. Analysis of RH of Feni from has shown in the Fig (c) where it has shown the RH max, RH min, RH mean and mean±sd of RH of that area. The RH max in this area is in the month of August and February and the lowest point is in the month of March. The RH max of this area is 99.85% (August and February) and the RH min of this 42

62 area is 24.65%. RH mean in Feni area is extreme in the month of July pointing about 88%. Again in April and May, the RH mean is around 85 and 84%. Any area with in this value indicated how hot and humid it was on that certain day. These high amounts of RH also responsible for creating any unsteady situation which in later may cause some extreme events. The highest range in this area is a strong indicator for the formation of cyclone. The wideness of sun hours in this endeavor has shown in the illustration (d). SH max, SH min, SH mean and mean±sd of SH has shown here. Data of 1975 and 1976 of SH in this area is missing. The SH max in this area is in the month of May (11.88 hr.) and the SH min is in the month of July (0.02 hr.). The highest of SH mean is about 8 hours which is in the month of February. The SH max in May indicating higher amount of temperature on such day at Feni. This higher amount of temperature may create deep depression and form cyclonic activity in that area Sandwip Sandwip is an island facing towards southern eastern part of Bangladesh and is vulnerable to the effects of cyclone. T mean, T max, T min, and mean±sd of temperature has shown in the Fig (b). Data of 1975 of temperature in Sandwip area was not available for analysis. The T max in this area is in the month of April ( C) which is also known as cyclonic period. The T min is in the month of January and the temperature is C. In another time period, from September to October, the maximum temperature is also about 34 0 C which indicate that both in the cyclone period, the temperature is very high and is enough to create any types of unstable condition which later may introduce some extreme events like cyclone by the formation of low atmospheric pressure in the surrounding area. RH max, RH min, RH mean and mean±sd of RH for Sandwip area has shown in the Fig (b). Data of RH of 1975 was unavailable. July has represented the highest RH mean where it was about 91% and the lowest RH mean is about 76% pointing in the month of February. This highest RH mean in the area indicated hot and humid circumstances in that specific month in Sandwip. The high amount of RH also may cause insecure condition in the specific area which in later may form deep depression and ultimately induce cyclone. In Illustration (d), SH mean, SH max, SH min and mean±sd of light hours for Sandwip area has shown where the data from 1975 to 1986 of SH was not available. 43

63 (a) The SH max in this area is hours which is in the month of April and October and the SH min is nil which is in the period of July. From the figure, it is clearly seen that the maximum SH mean is in the month of April which is about 8 hr. and the SH max has also found in the month of October and April, as it stated earlier. The maximum of SH mean and the sudden longest period of sunshine hr. in a day of both in April and October may have a strong effect in uprising of temperature which in concern may form any cyclonic event in Sandwip region. (b) (c) (d) Fig (a) Location of Sandwip, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) RH (%) and (d) sunshine hours from

64 Sitakundo (a) Sitakundo is located in the south-eastern part of Bangladesh shown in the Fig (a) and is a part of Chittagong district. Part of Sitakundo is also covered by the hilly area. The warmer falls also situated here. So this area is important for its geographical location and as well as for its ecological resources. Temperature trend line of T mean, T max, T min and mean±sd has shown in the Fig (b) where the data of 1975 and 1976 was missing. The T max in this area is in the month of April which is C and the lowest temperature is in the month of January and the value is C. The highest T mean is about C has shown in the cyclone occurring month May. (b) (c) (d) Fig (a) Location of Sitakundo, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) RH (%) and (d) sunshine hours from

65 Again in September and October, the highest value of temperature is also about 34 0 C. So, it is very clear from the figure that the weather started to become unstable mainly in the pre-monsoon and post-monsoon season mostly in this area. Again, the T max almost throughout the year ranges from 29 0 C to 34 0 C which is very high amount and favorable for creating deep depression in the surrounding area. This deep depression in later may turn into the devastating cyclone. The trend analysis of RH from of RH max, RH min, RH mean and mean±sd has shown in the Fig (c). Previous two years data of 1977 was not available for the analyses. RH mean in Sitakundo area is highest in the month of July pointing about 87%. Again in April and May, the RH mean is around 85 and 84%. Any area with in this value indicated how hot and humid it was on that certain day. These high amounts of RH also responsible for creating any unsteady situation which in later may cause some extreme events. The highest range in this area is a strong indicator for the formation of cyclone. In Fig (d), SH max, SH min, SH mean and mean±sd of light hours for Sitakundo region from has shown where data of 1975 and 1976 was not accessible for the analysis purpose. From the illustration, the highest SH mean is more than 7 hr. in the month of March. Again the SH max has found in the month of April which is greater than 12 hours. These highest values of SH also indicating higher amount of temperature in March and April on such day at Sitakundo. This higher amount of temperature may create deep depression and induce cyclonic activity in that area Chittagong Chittagong is located in the south-eastern part and also economically important area for Bangladesh. It is normally known as the port city and the largest port of Bangladesh is located here. The city is basically covered with the hilly areas along side with the sea. Maximum, minimum, mean and mean±sd of temperature ( 0 C) has been studied here from which has shown in the Fig (b). The T max in this area is C (in May) and the T min is C (in January). The highest of T mean has found in the month of May where the point is about C. In the post-monsoon season (August to October) the T mean is in the range of 27.5 to C where the T max has found about C. Again the range of T max throughout 46

66 the year is about C. So, these maximum values indicating strong possibilities for unstable weather conditions in the studied area. Again, sudden rise of temperature like this maximum values may cause deep depression and then may turn into cyclonic storm which may affect the whole coastal areas of Bangladesh. (a) (b) (c) (d) Fig (a) Location of Chittagong, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) RH (%) and (d) sunshine hours from RH for the Chittagong area has shown in the Fig (c). RH max, RH min, RH mean and mean±sd of RH have been analyzed here from the data of July has represented the highest RH mean pointing about 86% here in Chittagong. This highest mean value in the area indicated hot and humid circumstances in that specific month in Chittagong. September is also known as the cyclonic month and the high amount of 47

67 RH showed here also may cause insecure condition in the specific area which in later may form atmospheric low pressure and ultimately forms cyclone. SH for port City Chittagong has shown in the Illustration (d). SH max, SH min, SH mean and mean±sd of SH have shown here which has up risen from the analyses of It is clearly seen that the maximum of SH mean is in the month of March which is more than 8 hr. and the S max has found in the month of April (11.60 hr.). The maximum mean of SH and the sudden longest period of SH in a day of both March and April may have a strong effect in uprising of temperature which in concern may form any cyclonic event in Chittagong region Kutubdia Kutubdia is situated in the lower southern-eastern part of Bangladesh. It is mostly faced to the Bay of Bengal (BB) and the BB is located to its western part. So, the cyclones also had badly affected here in several times. The line graph of temperature has shown in the Fig (b). The analyses have done from the data of , because data from was not exist. The T max in this region is C and the T mean is about 29 0 C both of which is in the month of May. The highest value of temperature showed in the figure has strong influence in creating unstable condition. Again the T max also shows powerful indication for the formation of cyclone by creating deep depression in the adjacent area. The abrupt C difference of the temperature between the maximum range and the mean value are strong indication of changing temperature in the area in such a day where the condition of the environment may become unsteady and in later, it may turn into cyclone. RH max, RH min, RH mean and mean±sd of RH of Kutubdia have shown in the Fig (c). The data was missing also here of 1975 and1976 and is not accessible for the analysis. The highest RH mean presented in the month of July where it is about 89%. The highest RH mean in the area indicated hot and humid circumstances in that specific month in Kutubdia. High amount of RH also may cause unstable condition in the specific area which in later may form deep depression and ultimately forms cyclone. SH max, SH min, SH mean and mean±sd for Kutubdia realm from has shown in figure 4.14 (d)where the information from has been skipped due to its unavailability. The SH max in this area is in the period of March (11.77 hr.) and the 48

68 (a) highest SH mean is more than 8.5 hr. in the month of February. This highest value of SH also indicating higher amount of temperature in March on such day at Kutubdia. This higher amount of temperature may create deep depression and form cyclonic activity in that area. (b) (c) (d) Fig (a) Location of Kutubdia, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) RH (%) and (d) sunshine hours from Cox s Bazar Cox s Bazar is located in the lower southern-eastern part of Bangladesh and it is the last part of Bangladesh. As shown in Fig (a). It is mostly bounded by the Bay of Bengal in the southern and western part. Fewer part of the eastern sight bordered by the Naf River and the rest of the part is surrounded by the hilly area and plain land. It is also known as the fishing port and healthy place of Bangladesh. The largest sea 49

69 beach of the world also situated here and that s why it is also the famous tourist spot of Bangladesh. The temperature graph has shown in the Fig (b). (a) (b) (c) (d) Fig (a) Location of Cox s Bazar, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) RH (%) and (d) sunshine hours from The graph has shown little bit variations in this area than the other area. The T max in this area is in the month of May which is C and the T min is in the month of June which is C. The range of T max in this area throughout the year is about 29 to C where the highest T mean is about 29 0 C which is in the month of May though in post monsoon cyclonic period September and October, the T mean is about C which is very much comfortable and is not suitable enough for creating any unstable condition.the high amount of temperature in the pre-monsoon period (April and 50

70 May) indicating unstable condition at the surrounding area. Again the difference between the highest temperature and the higher mean value is about 5 0 C which also indicating the strong positive signals for the formation of cyclone. The trend of RH of Cox s bazar has shown in the Fig (b) where it has shown that the RH max in this area is in the month of July and August and the point is 99.95%. July has also presented the highest RH mean where the RH mean is about 89%. This highest mean value in the area indicated hot and humid circumstances in that specific month in Cox s Bazar. This high amount of RH also may cause insecure condition in the specific area which in later may form deep depression and ultimately forms cyclone. SH for Cox's Bazar area has shown in the figure (d). SH max, SH min, SH mean and mean±sd for Cox's Bazar expanse from has shown there. The SH max in this area is in the month of April which is hours and the maximum of SH mean is in the month of January which is about 9 hours. This sudden longest period of SH in a day of April may have a strong effect in uprising of temperature on such day which in concern may form any cyclonic event in Cox s Bazar region Teknaf Teknaf is the last part of Bangladesh and it s an upazila under the Cox s Bazar District. The area is covered by the Bay of Bengal and the Hilly areas. Plain lands are also found here but it is in small amount. The Naf River is flowing through its eastern site and bordered the area as well as Bangladesh from the Myanmar. The area is geographically as well as ecologically important for our country. Temperature has been studied here also from excluding previous two years which was missing has shown in the Fig (b). The T max in this area is in the month of April which is about C. In the pre-monsoon season (March to May), the highest T mean is about 28 0 C which is in the cyclone occurring month May. In postmonsoon season (August to October), the maximum T mean is about 27 0 C pointing in the month of October. Again the T max in this region ranges from 29 to 34 0 C, are strongly responsible for any kind of extreme situation in any area. Again the difference of the T max from the highest T mean is about 6 0 C which sudden rise may warm the adjacent air of the surrounding area and form deep depression in that part. This deep depression as subsequent may turn into shattering cyclone in the area. 51

71 (a) (b) (c) (d) Fig (a) Location of Teknaf, (b) maximum, minimum, mean and mean±sd of temperature ( 0 C), (c) RH (%) and (d) sunshine hours from RH of Teknaf area has shown in the Fig (c), the data of 1975 and 1976 was not available for the analysis. Here, RH max, RH min, RH mean and mean±sd of RH have shown. The highest RH mean is about 90% which is in the month of July here in Teknaf. The highest mean value in the area indicated hot and humid circumstances in that specific month in Teknaf. The high amount of RH also may cause insecure condition in the specific area which in later may form deep depression and ultimately forms cyclone. SH max, SH min, SH mean and mean±sd of SH for Teknaf area from has shown in the figure 4.16 (c) where the data of 1975 and 1976 was missing. The highest of SH mean is more than 9 hours in the month of January. Again the SH max has found in the month of May which is about hr. This highest value of SH also 52

72 indicating higher amount of temperature in May on such particular day at Teknaf. This sudden higher amount of temperature may create deep depression and form cyclonic activity in that area Ensemble Means The Fig represents the comparison of mean values of temperature ( 0 C) of 16 coastal areas which have been studied here. The figure also represents the average value of temperature of 16 coastal regions of Bangladesh. From the illustration, it is understandably seen that except Satkhira, Khulna, Mongla, Cox's Bazar and Teknaf, the others 11 stations showed almost homogenous starting point from the month of July where the temperature was around C to C. Among the rests, Satkhira, Khulna and Mongla which have positioned at the last south-western part of Bangladesh have showed the similar starting point where the temperature is ranging from C to C. In fact, these three areas showed the uppermost T mean in the coastal regions of Bangladesh. Cox's Bazar and Teknaf shows little bit lower temperature of the starting point than that of others areas. The temperature of July in these two areas ranged from C to C which is the minimum mean temperature of July in the coastal regions of Bangladesh and strongly indicated the relatively cooler area than that of others areas. The total mean of July is about C also indicates moderately warm days in such monsoon season in the coastal regions. From July to August, all the stations with their total mean started to rise up indicates little bit increment of temperature in the coastal areas, except Satkhira, Mongla and Khepupara where the temperature is in decreasing trend. But from August, the mean is started to fall and drops up to January though for Cox's Bazar and Teknaf, the perceptiveness is dissimilar. In Cox s Bazar, the trend has dropped with a good range in the months from September to October and then rise up again in late October, but from mid-october, the trend has started to fall. For Teknaf, the trend is uprising from its starting point to mid-october and then fall down to January. 53

73 Fig Ensemble means of 16 stations air temperature ( 0 C) from Farukh and Baten (2015) reported that the months of September and October is mostly vulnerable for cyclone occurence. In fact more than 50% of cyclone from has occurred in these months (Ali, 1999; Farukh and Baten, 2015). So, variations of temperature in this area indicate unstable condition of that surrounding area. From the figure, it has seen that Satkhira, Khulna and Mongla shows relatively higher mean temperature in the month of September and October. This situation strongly indicates the possibility of unstable condition in that area. Shamsad et al. (2012) reported that increase of C temperature from August to September may gear up an average of 40% cyclone activity. Now, if we compare the mean values of Satkhira, Khulna, and Mongla with the total mean value for the period from August to October, then we can see that the difference of mean of the areas with their total mean is about 1 0 C which strongly indicating the possibility of creating unstable condition and formation of cyclone. Ali (1999), Farukh and Baten (2015) again in their report stated that the months from March to June is second most vulnerable time for the formation of cyclone, because on an around 45% of cyclone from has occurred in this time. The cyclone Mohashen in 2013 and Ruyano in 2016 also attacked in the coast-line during this time. So study on the variations of temperature is very important for this time. 54

74 In Fig , the mean value of RH for 16 stations have represented from where the comparison of the stations is clearly seen along with their total mean value. The result is important as because temperature and RH relates to each other (Lawrence, 2004; Skilling, 2009) where the RH has a good influence on the formation of cyclone (Kaplan and DeMaria, 2003; Emanuel et al., 2004; Hendricks et al., 2010; Kaplan et al., 2010). From the figure it is clearly seen that all the stations show almost similar relationship at its starting point except Chittagong and Satkhira. Fig Ensemble means of 16 stations RH (%) from These two stations show much lower RH than that of others in the month of July. From July to August, the entire trend is decreasing except Satkhira and Mongla and in the following months the entire trend follow similar characteristics except Khulna that showed different characteristics where RH was increasing like Satkhira and Mongla. Then for the preceding months, the mean value has started to fall and continued up to February-March. The maximum of total mean has found in the month of July (88%) which was much higher than the average value in July for whole Bangladesh, recorded 81.78% (BBS, 2009). In cyclone occurring months specially from August to October, the mean value of the coastal region is on around 84~88% and in May to June, the value is about 82-55

75 86% both of which indicate positive sign for the formation of cyclone as Wu et al. (2012) showed. In fact, Khulna, Mongla, Khepupara and Barisal area are mostly responsible for creating deep depression if the RH plays any role to the formation of cyclone, as the highest amount of RH has found in this area, though Satkhira showed the lowest amount of RH in these months. In fact, almost every station was not in homogenous trend and showed different characteristics which may have reduced the importance of RH to the formation of cyclone for the coastal areas of Bangladesh. Fig Ensemble means of 16 stations SH (hr.) from Mean of SH for the 16 stations has been represented in the Fig with their total mean which indicating the average mean of the coastal region. The starting point is almost same for all the 16 stations from the month of July ranging from 3.12 to 4.15 hours. Now from this month to November, the time period is increasing for all the 16 stations, except the Sitakundo. From November to April, the lines are not in homogenous trend and there are lots of ups and down in these months in the 16 stations. But from April, the lines have followed similar movement and decreased from this month up to June. SH have not direct effect on the formation of cyclone though there are some indirect consequences of the SH. Because SH have positive relationship with the temperature (Matuszko and Weglerczyk, 2014) and temperature have direct effect on the formation of cyclone as it is stated earlier in the temperature part. From the figure, it is clearly seen that in the months of April, the average mean is 56

76 maximum which indicating the higher amount of temperature in that period. This higher amount of temperature in the following month may cause unstable condition in the area which later may form the unwanted cyclonic events in that area Clustering Fig represents the clustering of 22 cyclones from that we have considered in our study. From the figure, it is clearly seen that the ultimate factors (SLP and temp.) responsible for the formation of 22 cyclones have grouped into 3 clusters (centroid). The cyclonic factors which have similar characteristics belongs to the same centroid. Cluster 1 Cluster 2 Cluster 3 Fig Clustering of temperatures ( 0 C) with the sea level pressures (hpa) of 22 cyclone occurring days from As we can see from the figure that circle with red color is considered as centroid 1 where the temperature was C and sea level pressure was hpa. So, centroid 1 indicates that C temperature is responsible for 1 cyclone out of 22 cyclones occurred from In centroid 2, 9 circles with blue color have presented which have considered as in same characteristics for occurring 9 cyclones in these 40 years. The temperature triggered behind the formation of 9 cyclones was from 30 0 C to 57

77 C where the sea level pressure was from hpa to hpa. The centroid 2 is pointing at C which is the average of these 9 blue circles. So, these results indicate that temperature above 30 0 C is very much favorable for happening atmospheric low pressure which in consequence turn into cyclonic activity. The rest of the 12 yellow circles have considered under the third centroid where the circles are in same characteristics for the formation of rest 12 cyclones from Synoptic Climatology using GCM In this part, the composite geographical distribution of synoptic temperature anomalies compared with the climatology from for the cyclone occurring days. Composite geographical distribution of temperature anomalies are seen for the 3 clusters and compared among the surface air temperature, 850 hpa air temperature, 700 hpa air temperature, 500 hpa air temperature and 300 hpa air temperature. Fig to Fig illustrates air temperature ( 0 K) from 12 0 N to 32 0 N latitude and 78 0 E to E longitude derived from NCEP, NOAA/ESRL Physical Science Division. factor or condition which has played a big role for the formation of that cyclone has fallen under cluster 1. Comparison of composite geographical distribution of synoptic temperature from the specified level has shown in Fig From the results we can say that surface level air temperature in Bangladesh alongside with south-western India was comparatively hotter than some part of Myanmar, Thailand and Laos where the air was much cooler on the cyclone occurring day. In clustering part, we have seen that only one cyclone which has passed in the stated 40 years have followed different characteristics among 22 cyclones. These single Fig (a) shows the surface air temperature anomaly where positive surface air temperature anomaly zone over the western and south-western part of Bangladesh. This positive anomaly was prominent over the whole part of Bangladesh as well as prominent over West Bengal of Indian territory. This positive anomaly zone was also available in the eastern border of Bangladesh also covering India and some part of Myanmar strong negative anomaly zone is seen in the south-eastern part of Myanmar and northern-western part of Thailand and Laos. 58

78 Fig The patterns of composite air temperature ( K) anomaly for cluster 1 compared with climatology for cyclone occurrence at (a) surface level, (b) 850 hpa level, (c) 700 hpa level, (d) 500 hpa level, and (e) 300 hpa level 59

79 From the results we can say that surface level air temperature in Bangladesh alongside with south-western India was comparatively hotter than some part of Myanmar, Thailand and Laos where the air was much cooler much cooler on the cyclone occurring day. Fig (b), (c), (d), and (e) shows the composite air temperature anomalies at 850, 700, 500, and 300 hpa level of atmosphere on the days of cyclone occurrences, respectively. Fig (b) depicts a strong positive air temperature anomaly zone over the southe-western part of Bangladesh at 850 hpa level. In fact, all the part of Bangladesh from N to the south of its area is covered by the strong positive anomaly zone. The transition zone between positive and negative anomalies existed over the central part of Myanmar keeping one strong positive zones over southeastern India and central-western part of Myanmar. But at around 3,000 m above (700 hpa) at Fig (c) from surface there was complete disappearance of positive anomaly zones. Appearance of some part of negative zone is found in the southe-eastern part of Bangladesh and in from central to southern part of Myanmar. Thus, this cooler zone at relatively upper atmosphere would not have effect on creating any unstable atmosphere. The temperature distribution at Fig (d) shows the anomaly of temperature at the most influential weather zone of upper atmosphere i.e., at 500 hpa level. A strong negative zone encircled over the south-western part of Bangladesh and over Bay of Bengal (BB) indicates development of a cooler zone at around 6,000 m (500 hpa) above from surface. The most dramatic phenomena is seen at 300 hpa level (~10,000 m above) where, a huge area is covered by positive anomaly extending from Tibet to North Indian Ocean (NIO) via eastern Indian territory (Fig (e)). Whole of the Bangladesh was dominated by positive anomaly values implies formation of a deep warmer zone over these regions. This larger warmer air mass at upper atmosphere could influence a lot to develop huge instability throughout the whole atmospheric column. There were nine cyclones those had passed in the listed 40 years following almost same characteristics among 22 cyclones. These conditions which have played a big role for the formation of those cyclones have fallen under cluster 2. 60

80 Fig The patterns of composite air temperature ( K) anomaly for cluster 2 compared with climatology for cyclone occurrence at (a) surface level, (b) 850 hpa level, (c) 700 hpa level, (d) 500 hpa level, and (e) 300 hpa level 61

81 Comparison of composite geographical distribution of synoptic temperature from the specified level is shown in Fig Fig (a) shows the surface level air temperature anomaly where the positive surface air temperature anomaly zone is prominent over the eastern and south-eastern part of Bangladesh. On the day of cyclone occurrences, the strong negative anomaly zone from Indian territory expanded toward south-western territory of Bangladesh. At the same time another strong negative anomaly zone is also evident over the southern part of Myanmar. This coexistence of negative zone over the Indian Territory and the southern part of Myanmar keeping the transition zone over the southern part of Bangladesh to the BB. The result suggests dominance of relatively cooler air temperatures in the southern zone near to surface level especially on the day of cyclone prevalence. The area is mostly cooler on the south-western part than the south-central, and south-eastern part of Bangladesh on the cyclone occurring days at surface level. Fig (b), (c), (d), and (e) shows the composite air temperature anomalies at 850, 700, 500, and 300 hpa level of atmosphere on the days of cyclone occurrences, respectively. Fig (b) depicts a strong positive air temperature anomaly zone over the north-eastern part of Bangladesh at 850 hpa level. The transition zone between positive and negative anomalies was existed over the BB keeping two negative zones over south-eastern India and southern Myanmar. But at around 3,000 m above (700 hpa) in Fig (c) from surface there was complete disappearance of negative anomaly zones. The positive zone expanded from north-east toward south-east and over BB. Thus, this warmer zone at relatively upper atmosphere could make unstable atmosphere through the thermal instability interacting with the surface layer. Whole of the Bangladesh was dominated by positive anomaly values implies formation of a deep warmer zone over these regions. This larger warmer air mass at upper atmosphere could influence a lot to develop huge instability throughout the whole atmospheric column. 62

82 The rest of 12 cyclones those had also passed in the recorded 40 years following same characteristics among them belong to cluster 3. Comparison of composite geographical distribution of synoptic temperatures from the specified level is shown in Fig Fig (a) has shown the surface level air temperature anomaly where the positive surface air temperature anomaly zone was almost disappeared from the map except some small area of NIO, though they were not strong enough. On the day of cyclone occurrences, the strong negative anomaly zone from Indian territory expanded toward all over of Bangladesh as well as expanded to the whole south-east Asia. These coexistence of positive zone and negative zone over the Indian Territory and the whole part of Myanmar keeping the transition zone over the BB to some part of NIO Territory. The result suggests dominance of relatively cooler air temperatures in the southern zone near to surface level especially on the day of cyclone prevalence. The area is mostly cooler on the southern-western and central western part then the south-central and south-eastern part of Bangladesh on the cyclone occurring days at surface level. Fig (b), (c), (d), and (e) shows the composite air temperature anomalies at 850, 700, 500, and 300 hpa level of atmosphere on the days of cyclone occurrences, respectively. Fig (b) depicts a strong negative air temperature anomaly zone over the southwestern part of Bangladesh at 850 hpa level. The transition zone was existed over the BB keeping two negative zones over south-western India and southern Myanmar. But at around 3,000 m above (700 hpa) in Fig (c) from surface there was complete disappearance of negative anomaly zones over Bangladesh. The positive zone expanded from north-east toward southern part of Bangladesh and extended up to BB. Thus, this warmer zone at relatively upper atmosphere also could make unstable atmosphere through the thermal instability interacting with the surface layer. 63

83 Fig The patterns of composite air temperature ( K) anomaly for cluster 3 compared with climatology for cyclone occurrence at (a) surface level, (b) 850 hpa level, (c) 700 hpa level, (d) 500 hpa level, and (e) 300 hpa level 64