Reduce Exposure to Reduce Risk

National Institute for Disaster Prevention, NIDP effective decision making for urban flood risk developed GIS based disaster information analysis management and can estimate damages from system through information sharing and analyzing disaster for efficient emergency recovering shown with GIS based information based on IT in Figure 3. High resolution satellite and aerial technology to assess disaster risk. Some data and images and CCTV images are also useful to information are used for analyzing, calculating and estimate damage boundary and help to make a simulating of natural phenomena. For example, decision because real time images can help to Rainfall data is used for input data of rainfall- understand easily all situations regarding on runoff model and water surface elevation, flow and disasters. Real-time rainfall data are monitored on velocity data is used for calibration and validation concentrated area based on GIS system for warning to represent flow or estimate risk. Some data and and emergency management. The rainfall data with information are combined and integrated to create three hrs leading time forecasted by MAPLE is new knowledge, input data of a system, decision used for flash flood warning especially in the support information. For example, inundation area mountain area which area has severe flood risk represented by numerical model overlapped on GIS because water depth in stream suddenly increase information such as population, land use, vacancy and many people enjoy leisure in the mountain of hospital, school and utilities can support valley. Figure 3: GIS based Decision Support System The integrated or linked real-time CCTV images, at disaster risk area, stage gauges, major bridges especially established in riverside and flood risk and control CCTV for rotation, angle and camera area can help to understand local disaster situation zooming for monitoring area condition. for effective disaster risk management. Detecting technologies to capture disaster case automatically, Global Disaster Information Networks for example automatically detecting water depth over limit line in the river and warning to an The 4th Asian Ministerial Conference on Disaster operator is important because numerous real-time Risk Reduction (AMCDRR) was held in Songdo, images from various regions are presented in same Incheon on Oct. 25 for a three-day run under the time and longtime monitoring through CCTV is not main theme of disaster risk reduction through easy. Effective disaster management, integrated climate change adaptation. Along with ministers network system to link the CCTVs and integrate for disaster management from 53 nations, more analysis is needed. For operation of integrate than 900 participants were attend the international network system, the role of each governmental conference to discuss the mutual cooperation and organization at city, province and national is common resolution in the Asian region for disasters defined as i) national disaster agency monitor all induced by climate change, such as typhoon and images of CCTVs in a country to detect disaster heavy rain. Among those to participating in the situation for supporting decision making, ii) event were three heads of state - the Prime province manage all CCTVs in the province to Ministers from Nepal, Bhutan and Vietnam, monitor and send the video images to the representatives from UN agencies, the Vice organizations and institutes to share information President of the Asian Development Bank (ADB), and iii) city collect video data from CCTV installed Commissioner of the European Commission, and Reduce Exposure to Reduce Risk 90

representatives of NGOs. AMCDRR is the only Raising Awareness and Building Capacity for official international ministerial conference for DRR & CCA disaster management in the Asian region. It has The number of disasters can be considerably been sponsored by the United Nations International reduced if people are well informed and Strategy for Disaster Reduction (UNISDR) government and technical agencies pay high secretariat since 2005 for disaster risk reduction in attention to capacity building at all levels. There Asian region after the tsunami occurred in South are several plans to raise awareness and building Asia in December 2004. disaster risk capacity, which are outlined below. Particularly, globally 38 percent of the disasters Developing and Sharing Information, happen in Asia. However, 90 percent of the global Technology, Sound Practices, and Lessons disaster victims are found in Asia. Therefore, Asia Learned in Climate and Disaster Risk is the most vulnerable area, and because of the Management difference in the disaster management systems and With the increased disaster risks brought by climate technical levels, it is difficult to draw practical change, many countries in the region are taking cooperation. The main purpose of the 4th actions to alleviate their vulnerabilities to disasters. AMCDRR was to derive agreement for common To achieve development and assistance for the response to the disasters induced by climate change countries in the region better, sharing previous in the Asian area for the first time. The Incheon lessons, and cooperating between stakeholders in REMAP and Action Plan is an action plan government, civil society organizations and composed of five-year projects implemented to international or regional organizations is change each nation's disaster management system indispensable. Information is indispensible in into a disaster management system with climate reducing disaster risk. Timely and accurate change adaptation. It reorganizes the disaster information helps policy and decision makers and prevention standard and vulnerability analysis local communities and ownership. It improves through systematic study of disaster status and community and individuals‟ capacity and resilience climate ideas in Asia and the Pacific area, in order in the disasters. Disaster risk reduction knowledge, to strengthen disaster management capacity. tools, and good practices can be shared through information sharing system. It trains climate change adaptation and disaster prevention experts through education and training To share information, technology, sound practices, for disaster risk reduction and climate change and lessons learned in climate and disaster risk adaptation, targeting officials and stakeholders of management, NEMA is developing the Platform of each nation. The plan is also designed to introduce Information and Technology for CCA and DRR a global website and technology related to climate with budget of 2.5 Million USD during 2012~2015. change and disaster prevention by building a The global platform have four main menus such as climate change adaptation technology-information “System and Technology,” “CCA & DRR platform, and improve response capability by Platform,” “Education & Training” and “4th sharing good and bad cases for large disasters. AMCDRR”. The System and Technology provides advanced technologies and system to share and The country can derived preemptive international transfer technologies regarding CCA and DRR. disaster prevention guidelines for climate change The CCA & DRR Platform provide an information disasters through suggesting and agreeing on the of CCA and DRR to analysis linked information to Incheon REMAP and Action Plan for the Asian the global platforms in the world which new region in this conference, and enhanced national integrated and combined knowledge is used for prestige by building a climate change information early warning and decision making. The Education and technology sharing platform. For contributing & Training support information about schedule and to disaster risk reduction in Asia the Pacific region, program education and training operated by Global NEMA provided Korea's excellent up-to-date Education Training Institute (GETI) which GETI disaster prevention systems, such as WEB GIS can develop the training program for technical Based Typhoon Committee Disaster Information transfer and share information of CCA and DRR. System (WGTCDIS), earthquake disaster response The 4th AMCDRR provide information such as system free of charge to each nation (Cheong, history of AMCDRR, results of 4th AMCDRR and 2009; Yi et al., 2008). Through series of also results regarding on Incheon REMAP and participatory consultations, the participants of the action plain decided on 4th AMCDRR. fourth Asian Ministerial Conference on Disaster Risk Reduction have confirmed an action plan Promoting Integration of DRR & CCA into based on Incheon REMAP as follows: Development for green growth Disasters are a big concern for most countries and have great impact on our economy. Disasters can Reduce Exposure to Reduce Risk 91

take significant number of lives and leave long Operation System Setup (II). NIDP-PR-2010- lasting negative impact on our properties. Without 02-02, P. 189. disaster prevention and reduction, we cannot attain Yi, W., Cheong, T.S. and Jin, K. 2008. Typhoon millennium development goal and sustainable Committee Expert Mission Report. WMO/TD- development. Disaster prevention is not only No. 1148, ISBN 978-89-90564-88-7 93560. environmental and social issue, but development issue as well. As there is a saying “Disasters can be an opportunity,” we need to turn disaster into opportunity and create new type of growth. Thus, ensuring disaster reduction policy is a part of broader economic and development policy in environmentally-friendly way. Conclusions NEMA has been developing NDMS as a comprehensive nationwide information system to support disaster management processes in terms of prevention, preparation, response and recovery. NIDP developed GIS based disaster risk management system through information sharing and analyzing with GIS based information based on IT technology to assess disaster risk. The disaster risk management system as a successive prevention oriented disaster information system integrated local and GIS based information to analysis complex and compound disasters. It is provided that comprehensive nationwide information system is useful for disaster risk reduction and supporting decision making on emergence response and recovering. A global disaster information network as a tool for disaster risk reduction and climate change adaptation is suggested in here. The global platform for sharing disaster information and related technology for CCA and DRR is international cooperation system considering emerging disaster risk due to climate change, developed to response Asia-Pacific's initiatives for climate change adaptation and disaster risk reduction considering vulnerabilities in the region. NEMA has developing the global platform by 2015 as an international cooperation system which global platform will be used to reduce disaster risk and adapt climate change in Asia and the Pacific region. References Cheong, T.S., 2009. WEB GIS Based Typhoon Committee Disaster Management Manual, WMO/TD-No. 1508, SBN 978-89-90564-92-4 93560. Ministry of Land, Transport and Maritime Affairs. 2009. Master Plan of the Four Major River Restoration Project. National Institute for Disaster Prevention. 2010. Disaster Information Analysis Center Reduce Exposure to Reduce Risk 92

NATURAL DISASTER MANAGEMENT T.M.N.Peiris, D.G.Fernando Meteorologist, Department of Meteorology, Baudhaloka Mawatha , Colombo 07 [email protected] Abstract: Natural disasters are extreme events within the earth's system that result in death or injury to humans, and damage or loss of valuable goods, such as buildings, communication systems, agricultural land, forest, natural environment etc. The economic losses due to natural disasters have shown an increase over the past four decades, caused by the increased vulnerability of the global society, but also due to an increase in the number of weather-related disasters. Synthetic aperture radar (SAR) plays a very important role in monitoring hazards, especially flood. Because of the presence of clouds and bad weather conditions in almost all flood events, SAR is often the only available tool to monitor such events when they occur and also to contribute in post event damage assessment analysis. Since SAR can acquire images independently day and night and in all weather conditions. SAR data are often thought to be optimal for mapping flood inundation. Flood monitoring and impact analysis is essential for early warning. High resolution SAR images can be utilized for flood monitoring and impact analysis. Satellite-derived data can provide timely geographical data and overlay these maps in GIS database can provide useful information for flood prevention decision making. Identification of vulnerability of flood hazards as an annual feature is beneficial as the country suffers periodical flooding during monsoon season. The article describes the role of SAR technology in evolving a suitable strategy for flood preparedness, monitoring, assessment and mitigation. Introduction aperture radar (SAR) can achieve much higher spatial resolution. Electromagnetic (EM) spectrum is showing atmospheric opacity in the microwave (MW), SAR is a new remote sensing tool and has infrared (IR), visible (VIS) and ultraviolet bands. considerable potential for various applications. Low opacity regions are called the atmospheric Especially radar interferometry and radar windows, allowing the transmission of EM polarimetry have demonstrated some quantitate radiation. Particularly, 10mm to 1m wavelength measurements of geophysical parameters. microwave has little disturbance not only the atmosphere but also cloud and light rain. SAR is a side-looking radar that uses two methods Advantage of the microwave remote sensing is (one for each dimension) to achieve higher observation capability under all-weather condition. resolution. A SAR typically generates a very short In addition, the microwave has different pulse and a series of range gates to resolve range. interactions with the earth surface from the optical The motion of the satellite relative to a point on the wave. This is because that the component size of earth produces a Doppler shift in the returned objects on the Earth is much larger than the signal, permitting azimuthal resolution. The wavelength 0.4µm to 15 µm of the main optical azimuth resolution can be one half the azimuthal region, while it is comparable to the wavelength of dimension of the transmitting antenna and is the MW region. independent on the radar wavelength or the satellite altitude. MW imaging systems are divided into two types, passive system and active system. Passive system SAR uses the synthetic aperture technique to bring is a microwave radiometer and detects energy the high azimuth resolution with a small antenna. It radiating from the earth surface or the atmosphere. uses also a pulse compression technique to achieve On the contrary, active system is a radar system the high range resolution and improvement of that transmits radar pulses toward the earth surface signal to noise ratio. and receives the reflected echoes. Active system uses a stable energy source. Therefore the image Radar imagers represent the reflectivity of the data from MW imaging radar systems can be surface or near-surface in backscatter direction. compared to each other after the proper processing. The image brightness is directly proportional to the Among MW imaging radar systems, synthetic surface backscatter radar cross section. The backscatter intensity is independent primarily on Reduce Exposure to Reduce Risk 93

the radar parameter (i.e. operation frequency 0r larger than HH and HV and VH response become wavelength, incidence angle and so on) and the zero ideally. As for a simple model of vegetation surface physical properties (i.e. average slope and consisting of only thin vertical linear scatters, a small-scale roughness), and surface and near horizontally polarized wave will not interact with surface dielectric constant, which is a function of vegetation, but vertically polarized wave will the surface material, its composition and moisture interact strongly with vegetation. In practice, a content. vegetation canopy is usually consisting of linear scatters with various directions. Therefore, Advanced SAR system – Polarimetric SAR polarization vector is rotated from horizontal to (POLSAR) vertical or from vertical to horizontal at scattering, and then HV and VH response increases. In a Polarization of EM waves is defined by the trace of vegetation canopy or at an extremely rough surface, the electric field vector (E-vector) tip. The most multiple reflections/ scattering also increase HV general polarization state is elliptical polarization; and VH response. in this state, the tip of the vector in the plane orthogonal to the direction of propagation traces an Polarimetric SAR is an advanced imaging radar elliptical curve. Linear polarization and circular system which operates at four polarizations (HH, polarization are special cases, when this curve is a HV, VH and VV). straight line and a circle respectively. For example, ground TV waves are linearly polarized, and Why SAR data for flood monitoring satellite TV waves are circularly polarized. Observation capability of flood conditions in any Combination of linear co-polarization is ordinarily weather conditions, while optical sensors are employed by present airborne and spacebone SAR difficult to observe flooded conditions timely due system, i.e. horizontally polarized transmit and to usually bad weather conditions during the period horizontally polarized receive (HH) for JERS-1 and of flooding. RADARSAT-1 SARs, and vertically polarized transmit and vertically polarized receive (VV) for Easy to use the absolute values of backscattering ESR-1/2 SARs. Horizontally polarized receive or intensity (backscattering coefficient (0)) for vertically polarized receive means that the detecting flooded areas without the effects by horizontally or vertically polarized component of season, time and atmospheric condition, while the scattered wave is received. optical sensor data are much affective by season/time (sun elevation) and atmospheric The polarization response from the earth surface condition. object usually varies with polarization. In case of a slightly rough surface object (e.g. sea surface) relative to the wavelength, VV response becomes Physical basis for flooded area detection. Land cover Non-flooded flooded Roughness Non-water surface Water surface Scattering type Rough Smooth Backscatter Diffused specular High Low Therefore SAR backscattering intensity generally changes to be lower according to the land cover change from non-water surface to water surface by flooding. STUDY AREA We choose flood event occur Gampaha district in 2008.06.03. In that event, Gampaha District was For this case our study area was Gampaha district seriously affected with 10 Divisions or 151 Grama in Sri Lanka. This area located in the Western parts Niladhari areas under water and 55,613 people of of the island, which receives most of the south-west 4,331 families affected. monsoon rainfall making the vulnerable for frequent floods. Reduce Exposure to Reduce Risk 94

Data Used 4. Microsoft Office (excel) In this study ALOS/PALSAR HH and HV Methodology polarization was selected to detect flood extent. Satellite scenes were acquired during the same Using the absolute value of backscattering intensity season and with and without flood. (backscattering coefficient (σ0)) can measure detecting flood areas without the effects of season, PALSAR L1.5R time and atmospheric condition. -2008/06/03, FBD During the flood time, ground surface changes ALPSRP125690130 (slave) rough to smooth. So, SAR backscattering intensity -2008/07/19, FBD generally changes to be lower to the land cover change from non-water to water. ALPSRP125690130 (master) Vector data- Land/Sea, Districts, Administrative Inundated area or flooded area could be identified Divisions, Rivers and Tanks by generating RGB colour composite image by calculating difference, correlation and NDSI Software Used (Normalized Difference Soil Index). For example the areas where there was no change in land cover 1. PALSAR appeared in grey while the areas of land cover 2. ReMap changes will appear as colourful patches in the 3. ArcGIS RGB image. Reduce Exposure to Reduce Risk 95

Make Make backscatter backscatter image (for Slave) image (for Geocording Master) slave image Geocording master image Adjust the position between master and slave imagers and register RGB Calculati Calculati Calculati Calculation Calculation composite on of on of on of of of image differenc correctio NDSI inundation inundation creation e n area area (slave) RESULT HH POLARIZATION HV POLARIZATION HH POLARIZATION HV POLARIZATION Reduce Exposure to Reduce Risk 96

Discussion between masters and slave imagers register those imagers. Though our area of interest was Gampaha District on 3rd June 2008, the SAR image was covered only Following the methodology of generate RGB some part of the District. But it also showed colour composite image, calculation of difference, flooded area clearly. Here HH polarization image correlation, and NDSI we can identify the flooded showed good relation with water bodies. Most area easily. When we have good geo-database of difficult part was to calculate threshold value for land use land cove as well as population or detect the water bodies. It should have to done property of this area help us to estimate degree of carefully with ground measurements and flooding damages, estimate of inundated water meteorological data. Using fine-tuned pre level/volume in flooded areas and estimate of calculated threshold values; we can identify flood agricultural damages. events with the correlation of rainfall data. Automatic weather stations operating by the Using SAR analysis along with meteorological Department of Meteorology covers most parts of data, flood monitoring and impact analysis can be Sri Lanka and gives real time rainfall data to derived for early warning for future flood events. identify the vulnerability of food events for early Therefore, high resolution SAR images can be warnings. utilized for flood monitoring and impact analysis. Satellite-derived data can provide timely Due to the cloud penetrating property of geographical data and overlay these maps in GIS Microwave, SAR is able to acquire \"cloud-free\" database can provide useful information for flood images in all weather phenomena. This is prevention decision making. Identification of especially useful in the tropical regions which are vulnerability of flood hazards as an annual feature frequently under cloud covers throughout the year is beneficial as the country suffers periodical like Sri Lanka. Being an active remote sensing flooding during monsoon season. device, it is also capable of night-time operation too. REFERENCES Therefore, SAR able to observe flooded conditions 1. Scott T. Shipley , Classroom Exercises in timely due to usually bad weather conditions GIS Meteorology,George Mason during the period of flooding. University A radar altimeter sends out pulses of Microwave 2. Nanshan Zheng, Kaoru Takara, Yasuto signals and record the signal scattered back from Tachikawa and Osamu Kozan, the earth surface. The height of the surface can be ‗Vulnerability analysis of regional flood measured from the time delay of the return signals. hazard based modis imagery and demographic data in the Huaihe river The rough degree of surface is relative to radar basin‘, China. incidence frequency. If the frequency changes, surface roughness of ground object will change 3. Saseendran S. A., ‗A GIS application for with it. Surface roughness is usually decided with weather analysis and forecasting, National Rayleigh criterion. Therefore, in SAR image Centre for Medium Range Weather change detection, it demands the same sensor, Forecasting‘, Department of Science and namely, the frequency of incident wave does not Technology, New Delhi -3 change. The roughness of ground objects changes and the backscattering intensity also changes 4. Samarasinghe, S.M.J.S ‗Application of following it, which leads to final SAR image remote sensing sensing and GIS for flood change. risk analysis: Case study at Kalu-ganga river, Sri Lanka‘International Archives of When more than one radar images are available of the Photogrammetry, Remote Sensing and the same area acquired, use the before flood image Spatial Information Science, Volume as ‗master‘ and after flood image as ‗slave‘ image XXXVIII, Part 8, Kyoto Japan 2010 in ‗PALSAR Re Map‘ software and do geocoding to those two imagers. After adjust the position Reduce Exposure to Reduce Risk 97

GIS tools and approaches for Mainstreaming Disaster Risk Reduction into Local Development Plans;ACase Study from Ambalanthota Nandana Mahakumarage1, L.D.C.B. Kekulandala2, Buddika Hapuarachchi2, Vajira Hettige2 1 TMS Consultants, 2 Practical Action Sri Lanka, No 05, Lionel Edirisinghe Mawatha, Colombo 05 Corresponding Author: Nandana Mahakumburage, E-mail: [email protected] ABSTRACT: Sri Lanka is in a rapid urbanization phase in par with global urbanization trends and estimated average urban growth rate is 3.1 percent per year during 2005 – 2015. Disasters are a key determinant in the development process as it can compromise the goals and aspirations of the development. Walawe, is one of the largest rivers of Sri Lanka and falls into the sea through Ambalanthota. 10 Grama Niladhari divisions in the riverine coastal areas of Ambalantota and Hambantota Divisional Secretariat divisions are subjected to frequent flooding. Wanduruppa and Ambalanthota South GN divisions are the severely affected GN divisions. Both these GN divisions are situated along the walawe river and close to river mouth that closes due to the formation of a sand bar that increase the exposure risk to flood hazard. Project used both participatory and technical GIS approaches to map flooding area of the villages. A community led flood mitigation mechanism was introduced by practical action to minimize the flood hazard to local communities and water gauges were installed along the river and villagers were trained to read/interpret water gauges and predict flood. GIS tools and approaches were used to build on this existing knowledge base as villagers have good knowledge and memories about the historical flood events and demarcations. The flood levels shown by the villagers were plotted using GPS. The project also used LIDAR high resolution DEM for elevation data and derived contour intervals at 30cm. These contour lines and contour areas matched with community knowledge and finally identified flood risk areas with different levels. Finally these flooding levels superimposed with physical, human activities and land use. This yielded specific vulnerability scenarios of people and their assets. These vulnerabilities were discussed at a multi stakeholder forum and appropriate measures were drafted. These results were given to UDA to be used for the development of integrated development Plan for Ambalanthota. This will result in guidelines for development and land use. These are being finalized for a gazette notification. KEY WORDS: Key Words: Disasters, local development planning, participatory risk analysis, GIS Reduce Exposure to Reduce Risk 98

Introduction Material and methods Sri Lanka is in a rapid urbanization phase in par Identification and mapping of the land use with global urbanization trends and estimated and buildings average urban growth rate is 3.1 percent per year during 2005 – 2015, whereas annual population Land use data for the study was obtained from growth rate will be less than 1.1 percent1. The rapid the Urban Development Authority. The scale was pace and the magnitude of this urbanization 1:5000 and it has been developed in 2008. These necessitate the development and application of layers were used to characterize the land use appropriate policies, strategies, tools and patterns. Study team verified this data using approaches in terms of safeguarding human well Google satellites image and field observations. This being. Disasters are a key determinant in the layer provided the road network and irrigation development process as it can compromise the systems. goals and aspirations of the development. Furthermore development can induce new disasters GPS was used to mark households and public or enhance the severity of existing disasters. Flood places (government buildings, religious places) that data in most parts of the country have clearly were affected by floods. Finally household level shown an upward trend over the years2. Although data were collated using questionnaire survey and the unprecedented rainfall is a factor for some of linked with each household through GIS. the severe flood events reported in urban areas, it cannot simply be accounted to the rainfall alone. Mapping of flood inundation area The severity of the impact by natural hazards is linked to degradation of ecosystems (storm water Mapping of areas affected by flood was one of retention areas), improper constructions and the major tasks of the project. Community disaster insensitive land use/development planning. consultations were carried out as villagers have a Disaster sensitive localized planning with the very good knowledge about the historical flood participation of all stakeholders can provide easy, events and demarcations. GPS used to plot flood cost effective solutions for disaster prone levels shown by the villagers. In parallel, LIDAR communities. high resolution DEM was used to derive contour intervals at 30cm. Then these were used to Walawe, is one of the largest rivers of Sri characterize different elevation categories (contour Lanka and falls into the sea through Ambalanthota. areas). Finally these contour lines and contour 10 GN (Grama Niladhari) divisions in the riverine areas matched with community knowledge and coastal areas of Ambalantota and Hambantota identified flood risk areas with 3 different levels divisional Secretariat divisions are subjected to that High (1‟ – 5‟) medium (5‟ – 6‟) and low (6‟ – frequent flooding. Wanduruppa and Ambalanthota 7‟). These flooding levels again verified with South GN divisions are the severely affected GN villagers, irrigation officers and other stakeholders. divisions. Both these GN divisions are situated along the walawe river estuary and close to the Mapping of the special areas/reservation Walawe river mouth that closes due to the areas formation of a sand bar that increase the exposure risk to flood hazard. Ambalantota town consists with 4 types of reservation areas and those are Here we report a case study where GIS tools was used to support local communities and local • Irrigation Department development planners to characterize the local • Road Development Authority disaster risk and identify appropriate mitigation • Coast Conservation Department strategies through a participatory process. • Wildlife Conservation Department 1 http://www.makingcitieswork.org/files/pdf/south- Identification of these reservation areas was asia/SriLanka.pdf very useful for urban planning. Buffer analysis 2www.Desinventa.lk were used to demarcate these reservation areas and although it was very complicated. Figure 01 shows reservation areas for different canal types. Reduce Exposure to Reduce Risk 99

Figure 01: Different type of canals and its area revealed that the floods are due to complex reservations reasons. There were agreement that flooding is due to the formation of the sand bar at the Analysis of vulnerability and flood risk Welipatanvila river mouth and water heading up along the river. This creates floods in the low lying Overlaying methods were used to identify areas of the river bank. In addition specific vulnerability scenarios of people and their assets. flooding issues were identified in the highland These vulnerabilities were discussed at a multi areas beyond the low lying river bank zone. stakeholder forum and appropriate response However there was disagreement among the measures were drafted. communities and the government officials about the exact causes for these flooding. There were no Other hand every output maps can use for urban quantifiable data available to assess the flood risk. planning and institutional decision making. As an example Irrigation department can easily identify Maps prepared through participatory exercises and building on their reservations. LIDAR data clearly showed that areas that were Data sources affected by river flood and floods created due to drainage issues in several specific areas. Figure two Project used both primary and secondary data shows the overall flood risks in 10 GN divisions for base mapping and analysis. Those are, along the walawe river. Secondary data Figure 02: Flood risk area of  Land use /Building layer – Urban Ambalantota Development Authority (UDA) The analysis revealed that 208  LIDAR data- Disaster Management buildings are situated in the high risk zone (elevation of 1-5 feet from mean Centre sea level). Table 01 summarizes the  Water bodies/ Irrigation system - UDA results. It also revealed that almost  Google Satellites image 80% of the buildings are houses.  Demographic data – Divisional Figure 03 shows the distribution of Secretariats Office highly vulnerable buildings in the Wanduruppa and Ambalanthota south Primary data GN divisions. Hence the communities  Questionnaire survey and the government officials were able  GPS mapping to identify vulnerable houses and engage in a discussion to minimize the Results and discussion vulnerability. The options of relocation to raising foundation height were People in the most affected GN divisions of evaluated. Wanduruppa and Ambalanthota south approached Practical Action seeking support to address the flood issue, as the areas subjected floods 10 to 12 times a year. The initial consultation with the communities and the government officers in the Reduce Exposure to Reduce Risk 100

The analysis was able to show that the natural drainage paths have been obstructed by constructing houses and other buildings. This was clearly evident from the case of Wanduruppa cemetery. The cemetery ground submerges even after a small rain and creates a standing water body, hence creates a public health hazard. The health risk is multiplied due to the government hospital situated next to the cemetery ground and nearby primary school. This is highlighted in figure 05 Figure 03: Highly vulnerable buildings School Table 01: Number of buildings situated in high Hospital vulnerable zone Cemete GN_NAME No of buildings ** Figure 05: Flooding in the cemetery ground in Wanduruppa Wanduruppa 64 The analysis further revealed that one of the Godawaya 41 contributory factors for the high vulnerability against flood is buildings in the flood reservation Ambalantota South 29 areas. Data provided by government agencies on the limits of reservation areas were mapped using Walawa 17 buffer operations and these were superimposed with building layer. Dehigahalanda 14 The analysis showed that there are more than 200 Ekkassa 13 buildings in the irrigation reservation areas. This has led to deterioration of the irrigation canals Manajjawa 12 creating public health issues, drainage and localized flooding and reduced water flow that Thawaluvila 12 affects the paddy farming. Hence encroachment into river and irrigation reservations was Malpettawa 5 highlighted as major hindrance factor for the development of the area. Welipatanvila 1 Total 208 Flooding issues created by drainage issues were highlighted by the 3D analysis. It showed that are several sinks in the area (figure 04). The local government agency was investing large amount of money to pump out water from these areas as they were a public nuisance and a health issue. Figure 04: Sinks in the area that has led to localized flooding 101 Reduce Exposure to Reduce Risk

Figure 06: Buildings on reservation of the Knowledge in Hong Ha Commune, Thua Thien Irrigation department Hue, Viet Nam. EOSTEM Project Milestone 9 Report. Hatfield Consultants, Integrated development plan for Ambalanthota West Vancouver. The study findings were presented to the local Phong Tran, Rajib Shaw, Guillaume Chantry and government, administration officials, local John Norton, 2009, GIS and local knowledge in community representatives and urban development disaster management: a case study of flood risk authority. The urban development authority was mapping in Viet Nam, Overseas Development provided with the data and maps created by the Institute, study and these were used to strengthen the Published by Blackwell Publishing, 9600 integrated development plan. Garsington Road, Oxford, OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA Conclusions Pradan, A. ,2004, GIS and remote sensing for flood The study clearly demonstrated that GIS disaster identification: a case study of Koshi approaches can be adopted at local level to address River basin in Nepal. In Proceedings of the Global localized flood issues. Symposium for Hazard Risk Reduction: Lessons Learned It is also important to note that traditional from the Applied Research Grants for Disaster Risk knowledge can be effectively harnessed to assess Reduction Program. 26–28 July. Washington, DC. disaster vulnerabilities and GIS techniques can be pp. 139–147. used to capture these knowledge elements. Wisner, B., P. Blaikie, T. Cannon and I. Davis, Community participation and consultation was 2004, At Risk: Natural Hazards, People’s continuously used to complement the study. The Vulnerability, visual elements in GIS can be effectively used to and Disasters (second edition). Routledge, London communicate technical information in a non and New York. technical manner. This was clearly evident from the analysis that was done to address the localized Virany SENGTIANTHR, 2007, Flood Risk Map flooding issues (especially near the cemetery). using RS & GIS: Case Study of Champhone District in Savannakhet Province, Lao PDR, Local governments and other government agencies Environment Research Institute, Science who does not have access to extensive data sources Technology and Environment Agency and technologies can use simple mapping Department of Meteorology and Hydrology, techniques to identify and analyze local risk Ministry of Agriculture and Forestry scenarios. Acknowledgements References The authors like to acknowledge the support and Hatfield ,2006, Using Participatory Methodologies, guidance provided by Practical Action staff Geographic Information Systems and Earth including Ranasinghe Perera, Nialantha Kumara, Observation Data to Map Traditional Ecological Menaka, Chandana Siriwardana and other who provided various support. Special thanks goes to members of the Integrated flood mitigation committee of Ambalanthota including president and other members. Authors would like to appreciate the support and guidance of Mr. Chandrasa, Ms. Anoja and Mr. Shrimal. Disaster Management centre provided LIDAR data and urban development Authority provided land use data, all officials are acknowledged by the authors. Local government officials are also acknowledged for their support in collecting ground level data. Reduce Exposure to Reduce Risk 102

MAINSTREAMING GIS IN NATURAL DISASTER MANAGEMENT – A CASE STUDY OF PAKISTAN 2010 SUPER FLOODS Muhammad Ali Kashmala Ikram, IGIS, NUST, Islamabad, Pakistan, E-mail: [email protected] ABSTRACT: Natural disasters are the hazards caused by the forces of nature that cause unacceptably wide-scale destruction both in terms of life and property. Flood is a form of natural disaster which occurs when water level in a natural or a manmade water body exceeds its holding capacity in such a way that it overflows uncontrollably into the adjacent lands. There are many phenomenon that may cause this event but the most prominent are heavy rainfall or storm surges. In 2010, Pakistan had witnessed one of the most catastrophic floods in its history. About one-fifth of the country’s area was immersed in flood waters for many days and even months. Twenty million people were affected. Although, adequate warning was issued by the Met office, yet the sheer scale of devastation was unprecedented. GIS is relatively a new field in Pakistan. Its applications are now in mainstream use in the disaster management practices in developed countries. In this paper, taking the case scenario of 2010 flood of Pakistan, application of Geo-informatics technology (GIS) in all phases Natural Disaster management i.e., mitigation, preparedness, response and recovery will be discussed. It will be highlighted that how we could mainstream the GIS applications to minimize the extent of damage of natural disaster by having a multi-sectoral, multidisciplinary approach. KEYWORDS: GIS, Natural Disasters, Floods, Disaster Management INTRODUCTION Floods are devastating, causing large-scale destruction and significant loss of life and Economy. The Natural Disaster is a sudden adverse or unfortunate rise in water level can be due to numerous causes, incident that occurs rapidly, instantaneously, although the most common in Pakistan is heavy rainfall indiscriminately causing great damage to human beings in Monsoon. as well as flora and fauna. A natural disaster is the effect of the earth's natural hazards, however, it should have an Floods in Pakistan – An Overview element of human involvement. A physical event such as When Natural Disasters in Pakistan are discussed, flood a volcanic eruption, which does not affect human beings, is described as the most occurring and destructive is a natural phenomenon but not a natural hazard. The disaster for Pakistan. Here is the list of floods that natural disasters can lead to financial, environmental or occurred in Pakistan and the destruction they caused. human losses. Flood is formed when the water level in a creek, river, lake or the sea rises where there is no place Year Lives Cost Villages Cost for the water to go and covers land that is dry in normal 1950 2,190 10,000 days, Sometimes there is a blockage in the run-off due to 1956 160 11,609 certain natural (such as land-slides) and manmade factors 1957 83 4,498 where the water overflows and goes on the land and 1973 474 9,719 causes flood. Floodplains are lands bordering rivers and 1976 425 18,390 streams that are normally dry but become covered with 1978 393 9,199 water during floods. By giving the excess water a place 1988 508 1,000 to go, floodplains help control the extent and severity of 1992 1,008 13,208 floods. Placing buildings and other structures in 1995 591 6,852 floodplains can therefore be dangerous for two reasons: 2001 219 50 1) The structures can be damaged by floods 2003 484 4,376 2) By obstructing the flow of water (thereby increasing 2004 85 47 2005 59 1,931 the width, depth, or velocity of flood waters) such 2007 918 structures increase flooding and flood damage on 2010 1781+ 2 million+ nearby property. 2011 434 20 Million 8.9 Million   Reduce Exposure to Reduce Risk 103

It may be noted that the human impact of natural breaching at Lehri Flood Protection Dam and Flood disasters in Pakistan can be judged by the fact that 6,037 water outburst 43 km of upstream of Sibi inundating people were killed and 8,989,631 affected in the period more than 20 villages along its banks. from 1993 to 2002 (World Disasters Report 2003). 6. The second upsurge of flooding took place by unprecedented rains across the KPK before releasing Main Causes for floods in Pakistan are: to southern Punjab and Sindh. 1. Monsoon patterns and heavy rainfall responsible for Source:OCHA flash floods 7. National Disaster Management Authority has been 2. Apparent depth of the local streams (Nullah) is less quoted as saying that denudation is one of the main that causes an overflow before the time and making reasons for aggravating the floods as only 5.2 per it to carry less amount of water then it could carry in cent of land in Pakistan is covered by forests. actual this is due to the presence of garbage and Waste 3. Melting of glaciers in spring season in Northern Areas. The water makes its way to lower areas moving towards rivers and the sea. 4. Sudden opening of spillways of dams due to various logical reasons. The 2010 Super Floods in Pakistan The Monsoon of the year 2010 brought with its worst flooding in past 80 years in Pakistan. The unprecedented floods began in July 2010 following heavy Monsoon Rains in the KPK, Sindh, (Lower) Punjab and Baluchistan regions. These rains over the large area made the rivers surges and overflow. 1. On 21 June, the Pakistan Meteorological Department cautioned that urban and flash flooding could occur from July to September in the north parts of the country. 2. According to the Meteorological Department, within a short span of time of 3 to 4 days, heavy rain fell in the catchments areas of Indus and its tributary rivers causing heavy floods in River Indus. 3. These floods were driven heavy monsoon rain. 4. Heavy rainfall of more than 200 mm (7.9 inches) Map (16 August, 2010) (Source: OCHA) was recorded during the four day wet spell of July 27 to July 30, 2010 in the provinces of KP and Punjab. 8. During 2010 flood era, flows in western rivers were very high and some of the discharge levels were 5. The 2010 floods started on July 22, 2010 after few comparable to those recorded during floods of 1956, hours of heavy rainfall in Baluchistan which caused 1976, 1988, 1992, 1995 and 1997.   Reduce Exposure to Reduce Risk 104

9. The monsoon rainfall of 2010, over whole country, GIS – A brief Introduction was excess of 87 per cent and was highest since 1994 and ranked second highest during last 50 years The term describes any information system that stores, of period. integrates, edits, analyzes, shares, and displays geographic information for informing decision making. Planning purpose of seasonal monsoon floods in 2010: In other sense GIS is defined as combining map and data Meteorological department suggested that all monsoon and working on both simultaneously i.e. “Map + rains were likely to be normal (+10%). Due to high sea- Attribute”. It provides us with a facility to work on map land temperature contrast, the development of some mid along with attribute table at the same time. Attribute table tropospheric circulations are likely in north Arabian Sea contains data that we want to work on along with map. that may cause HEAVY RAINFALL EVENTS over This map helps us to show the areas of disasters like southern areas of Pakistan(Sindh) during July to August. flood, earthquake, landslide etc. And GIS helps in Due of interactions of westerly-easterly waves, few very evolving a seemly strategy in catastrophe management heavy rainfall events would also occur over north and occupational context for their monitoring, Pakistan that may cause urban/ostentatious flooding assessment and mitigation. It helps us to identify areas of during July to September. This was seasonal conjecture major disasters which have been ruined and those areas with assurance level of 80% and intended for the where there is a risk of disaster in future, and planning purpose only. recommends appropriate strategies of disaster management using different technologies. These Damages and destruction caused by 2010 floods: technologies have also proved to be useful in the Risk Flood across Pakistan has affected millions of people assessment, mitigation and preparedness phases of making them homeless, devastated thousands of square Disaster management. The exact steps taken depend upon kilometers of land. Following are the effects of the the disaster phase and how time critical the need is. Natural disaster flood. MATREIALS & METHODS Pakistan has suffered from the worst flood of its history in monsoon season of 2010. As per DAMAGE NEED Firstly we may have an outlook of the flood forecasting ASSESSMENT (DNA) report of ADB/World Bank, the system prevalent in Pakistan. floods affected FLOOD FORECASTING SYSTEMS IN PAKISTAN  An area of about 160,000km2 (one fifth of the Working on the pre-disaster stage for reducing the country) vulnerability and for efficient preparedness in disaster of flood, Flood Forecasting Division and Pakistan  Suing about 1,985 lives. Metrological Department has wide-ranging system in  Damaging around 1.5 Million houses. function. This system also known as Flood Early  Smearing out cropped area of more than 17 Warning System and consists of the processes of close monitoring of the level of the dams and rivers and also Million acres. observing any abnormal amount of precipitation in  Displacing a population of 20 Million. twenty four hours.  Resulting in economic loss of PKR 10 Billion.  The Flood early warning system is operational for Role of Geo-Informatics Technology in Natural the Indus River below Tarbela and the tributary Disaster particularly Flood rivers, Chenab, Ravi, and Sutlej.  Flood Forecasting Division consist of large network Geo-Informatics plays an important role in the of monitoring stations spread in all over the country assessment and relief of Natural Disaster. The proper whose function is to observe any change in dam, structure of information system for disaster management river water level, precipitation and weather. should be available to tackle the disaster and to  Flood Forecasting Division also has a very detailed amalgamate all the information in all such projects. GIS HR Radio Communication Network So that in case enables us for disasters monitoring, modeling, mitigation, of emergency flood warnings all over the country rescues operation management, and rehabilitation can be summoned. strategies development providing its user with the ability  Communication with India is also established to of having a “Digital Map” of the area that can give you communicate hourly, three hourly, and six hourly to the maximum information. Moreover, GIT also enables keep informed about flood flows in the rivers on the its user to apply query and find its specific result. following river sites.   Reduce Exposure to Reduce Risk 105

1. SUTLEJ (Below Rupar, below Harike and below obtained from the analysis of precipitation, soil Ferozpur) saturation, reoccurrence interval of the past year’s floods in attribute table and observed its Map and 2. RAVI (Below Madhopur) worked on it depending on the requirements. 3. Releases from Bhakra Dam Reservoir on Sutlej and Basically GIS helps us in this stage in a way that we can be informed before the time about the disaster Beas (Pong) Dam Reservoir on Beas River, are also that May come. to be communicated to Pakistan on SUTLEJ telephone daily at 10:30 hours (PST) along with base 2) Post Disaster stage: It is related to the Response, flow data of and Ravi. Recovery, Mitigation, Rehabilitation and Reconstruction. In our data of 2010 flooding we IMPLEMENTATION OF GIS DATA: mapped the flood extent ,the area which it has covered, damage to number of houses, property, We were concerned with the data of 2010 floods in crop, health of people, Population due to flood . Pakistan, so that we can show the flooded areas, This data supports us to know rapidly that how much obliteration, maximum rainfall on Pakistan map and of the area has been destructed, where there is more Develop GIS by smearing certain Queries. need of relief, accommodation to people. GIS data helps us in storing emergency management related  Firstly we developed a shape file (.shp) of Pakistan departments location where we can take the refugees cities and another shape file of waterways of immediately, contact details and capacity. This GIS Pakistan and linked both the shape files. which we have developed act as a tool for linking real time online systems for tracking and  Then we added Data (as per our requirement) communications of hazard and damage speed. And e.g. if we want to work on Destruction made by through this available data we can develop certain Flood 2010 per city We’ll develop GIS by using City models of risk analysis level Map and Death toll, no of villages effected, no of people effected etc and by this we can even apply CONCLUSIONS & DISCUSSIONS: query that select cities with population effected greater then 1000, all those cities which has  GIS based software applications would continue to population effected greater than 1000 will be prefer for operative environment risk assessment and selected. management. Or we developed a thematic Map of Rainfall which has occurred in different cities of Pakistan , In order  These GIS software are also used to represent the to show the area of extreme precipitation-> spatial and non-spatial environment hazards and represented those polygons with different colors to coverage data in a fast and accurate way. consign symbology and may select city on the basis of value of any of the feature provided in data based  Maps of areas at risk from natural disasters are a on its Location or through its attributes. valuable information and communication tools. They can be used for a wide variety of purposes ranging  The same technique can be applied to show the from flood plain delineation, zoning and land use affected areas, population, damaged houses, villages planning. on shape file of Pakistan Provinces.  The thematic Maps derived from satellites and build HOW GIS APPLICATIONS WERE USED IN OUR in GIS software’s are important scientific inputs for DATA? mapping range of different calamities, these data layers can be integrated with case data for GIS applications are extensively used for assembling, envisaging risks of natural and man-made hazards. analyzing data, emergency preparedness and response planning. This disaster management consists of two  Usually the analysis was a difficult chore, and large phases: number of parameters, techniques, going for assemblage of data in field may be very exorbitant, 1) Pre-Disaster stage: It can also be termed as difficult and time consuming. So GIS software is “Disaster prevention stage”. It consists of disaster used for detailed and rapid analysis of natural prevention, mitigation, preparedness of disaster hazards. management. As per according to our data of 2010 floods we collected analyzed the previous data of  The GIS data can also be used to generate floods befallen in past Years, ostentatious and Effective Disaster Information E.g. we gave the input of monitored data of speed, System (DIMS).So if after few years again flood velocity, extent of previous floods, and also the data occur in these areas of Pakistan we can simply add   Reduce Exposure to Reduce Risk 106

the new data in attribute table and show it with help fig 2 of using thematic maps. In the second figure GIS software has helped us in different ways, such that it can help us to illustrate data related to any disaster. It is  Firstly we added the Shape file of Provinces of risk management. We are concerned with our data of Pakistan and linked it with waterways of Pakistan as 2010 flood in Pakistan so we represented the amount of we did in above case. Rainfall that occurred in different areas of Pakistan, which caused flash flooding and caused damaged to  Collected the entire statistical data for areas which houses, villages, lives of people at a large Scale. were destroyed by 2010 flood ,which comprised smashed houses, affected villages ,affected districts, So here our first figure shows that we collected data of total affected population, number of died persons Rainfall in different cities of Pakistan in mm and inches and persons who are injured. and from this data we examined that where maximum rainfall has occurred which cause flood equal to the flood  Our purpose was to collect data and to display it on a in previous years. So briefly I the first figure: Thematic map. fig 1  Applied certain queries -> by going on the selection tab and selecting “selection by attributes” from the We firstly linked the waterways and Pakistan Shape File, drop down list. And by selecting a layer of so that we can know from which routes the rivers pass. PAK_adm2 in the layers list of selection by attribute table ,keeping the create new selection bar selected  Then the we added our collected Data of Rainfall in and applying query on any of the column of attribute attribute table of cities shape file by adding fields of table giving it some unique value. “RF_mm” and “RF_inches” and entering the statistical data by keeping the editor bar active,  Such that we can easily point out those areas where which is telling us that which city of Pakistan has maximum or minimum destruction has ensued or received maximum Rainfall. which can help us to know that in which areas there are more victims of flood. Consequently by merely  Assigned different symbology to each polygon of looking on the map we can have an idea to identify city by going in layers property of “PAK_ adm3”, destructed areas or affected people. which can differentiate in the amount of Rainfall and its respective city.  This collected data helped us to know those areas where frequently Rainfall occurs or where there are more chances of monsoon Rains and can also help us in post-disaster risk management. fig 3   Reduce Exposure to Reduce Risk 107

In the third figure fig 5  Again the same process was done and here two In the 5th figure queries were applied at the same time that the person who died were greater than or equal to 100 and  Simply we have shown the attribute table of person injured were greater than or equal to 1000 damaged property due the flood all these columns were added in the attribute table by keeping the  This query worked very efficiently that we can come editor tool bar DE active and selecting the Add Field to know about two things at the same time button and then keeping the editor toolbar active- >selecting start editing we added the collected data  As a result of which those rows in attribute table in the respective column of each field. were selected where died and injured persons were greater than 100 and 1000 respectively. And at the REFERENCES same time those areas or provinces polygons boundaries of Punjab, Sindh, Baluchistan and 1. Disastrer Management by Dr. S.R Singh. N.WF.P were highlighted with blue color as shown 2. Natural Disaster Management by Arun Kumar in the figure Talwar & Satish Juneja fig 4 3. Natural Hazards & Disaster Management by P.S In the 4th figure Chawla  In this figure again the selection by attribute query 4. http://www.criticalthreats.org/sites/default/files/Paki has been applied stan_Flood_Severity_by_District_-_High_Quality_-  Here in this figure we wanted to know that how _8-26-2010.pdf much of the population is effected along with 5. http://www.springerlink.com/content/t27648/front- damaged houses so for this we used the AND matter.pdf operator and used it in our query to know about the 6. http://mpra.ub.uni- common areas where both population and damaged muenchen.de/11052/1/MPRA_paper_11052.pdf houses are greater than equal to 4 Million and 30000 7. http://www.uninets.net/~dsrowley/All%20Disasters respectively. %20are%20Personal.pdf 8. http://www.ndma.gov.pk/Publications/livingwithdis  Applying this query selected two provinces in asters.pdf attribute table and on map. 9. Federal Flood Commission 10. Wikipedia  Telling us the effected population and damaged 11. http://repec.org/ houses provinces according to the applied query. 12. http://www.ffc.gov.pk/ 13. http://floodmaps.lums.edu.pk/wp- content/uploads/2010/12/OSF-MAP.png 14. http://www.pmd.gov.pk/   108 Reduce Exposure to Reduce Risk

GIS BASED MODEL FOR GEO-HAZARD ASSESSMENTS IN MOUNTAINOUS AREAS OF PAKISTAN; A CASE STUDY OF HUNDUR VILLAGE Shareef Hussain Focus Humanitarian Assistance Pakistan, E-mail: [email protected] ABSTRACT Rapid growth in population of mountainous communities living in the Gilgit – Baltistan and Chitral region of Pakistan, coupled with climate change has alarmingly increased their exposure towards geo-hazards, i.e. debris flow, rock fall, landslide, flood and GLOFs. GIS based geo-hazards assessment model is a useful tool for identification and assessment of prevailing hazards, their risks and vulnerability of mountainous communities. The model comprises of two components; one, remote hazard assessment, i.e. satellite image analysis and base map preparation, and two, on-site hazard assessment, i.e. identification and validation of types and intensities of hazard zones and data collection of critical facilities through Global Positioning System (GPS). The collected data is fed into the GIS system, which comprises of geodatabase - for data input, storage and processing, and Model Builder queries - for spatial analysis of living zones, activity zones, hazard zones and critical infrastructures. The model provides critical information’s related to geo-hazards which can be used by decision makers and planners for community preparedness and sustainable development. The products of the assessment model are hazard and risk maps which have been found useful for land use planning, i.e. identifications of safe areas for future development as well as emergency planning, including identification of evacuation routes and safe havens. Keywords: Geo-hazards; Geographic Information System; Global Positioning System; Hazard and Risk Maps 1. INTRODUCTION recover from disasters2.The impact and intensity of Pakistan suffers from excessive natural and human- these hazardous events depend on the triggering induced hazards that negatively affect the lives and factors and frequency of events. Some of the livelihood means of its inhabitants, specifically in prominent geo-hazards are debris flow, landslide, the mountainous regions, i.e. Gilgit-Baltistan, rock fall, flood, and snow avalanche. According to Chitral and the adjacent areas. These areas are statistics3, 228 casualties have taken place as a result located in the lap of theworld’s highest mountain of 189 geo-hazard incidents in the Gilgit – Baltistan ranges, namely Karakuram, Himalayas and and Chitral region during the last three years. Hindukush, having vulnerable landscapes, exposed Considering these challenges, FOCUS Pakistan to hazards. The northern regions, where some of started working in the region, with the mission of Pakistan’spoorest communities live, suffer a creating resilient communities, through disaster disproportionally high number of disasters with 33% response training, awareness-raising and of the area located in high risk zones 1 . preparedness measures, including establishment of Deforestation, deteriorating and ill-designed stockpiles in some of the remotest areas. infrastructure, weak or non-existent early warning forecasting systems, poor community mobilization for disaster risk reduction and mitigation contribute to the increasing need to prepare for, cope with and 1World Bank, 2011. Disaster Risk Management Programs for 2 Sustainable Development Policy Institute (2011). Policy Priority Countries. Pages 387-398 Recommendations, A year after Pakistan Floods 2010, Islamabad Roundtable Discussion 3 FOCUS Pakistan, 2007, Participatory Evaluation Of Community Based Disaster Risk Management (CBDRM) Programme (Internal Report) Reduce Exposure to Reduce Risk 109

The overall objective of FOCUS Pakistan’s efforts Iceland [1]. Greiving, S., 2006, has designed a is to save lives, reduce suffering and create methodology, integrated risk assessment of multi- resilience in communities prone to man-made or hazards affecting spatial development of European natural disasters. During the last two decades, region [2]. Geological hazards may not necessarily FOCUS Pakistan has progressively established an turn into disasters, whenever they happen; there are effective hazard and risk assessment mechanism other factors such as extent of exposure to hazards, which has been tested during small and large scale social structures of the community and climate disasters, in the urban and rural settings. Recent change, if any, in the region, which play role in advancement in geo-technologies has helped increasing vulnerability of the community [3]. Geo- FOCUS improve the accuracy, efficiency of geo- Technologies play vital role in the development of hazard assessment studies. This paper, with the help efficient hazard assessment and early warning of a case study, explains the process of GIS based system models. GIS based multi-criteria decision geo hazard assessment which has been applied by making for landslide hazard zonation (Malaysia, FOCUS Pakistan in different settings. 2011)[4], GIS-based landslide hazard zonation model and its application (China, 2009)[5] and 2. LITERATURE REVIEW landslide hazard analysis for Hong Kong using Risk anticipation and dissemination of information landslide inventory and GIS (China, 2003)[6] are is one of the emerging fields of social science. some recent example of GIS application in hazard Hazard and risk assessment is a process involving assessment. collection, evaluation, processing and analysis ofgeo-hazards that make a community vulnerable to 3. STUDY AREA disasters. Different researchers and organizations The study area, Hundur Village of Yasen Valley, is use different models for the geo-hazard and risk geographically located at 36◦30'53.56\"N and assessment. There are two prominent multi-hazards 73◦24'56.47\"E. Hundur village is situated in the NW risk assessment methods, deterministic and side of the Gahkuchtown, headquarter of District probabilistic. According to R. Bell, T. Glade, Ghizer, at a ground distance of almost 71km. The separate investigations of single processes might village is home to almost 317houses, having an lead to misjudgment of the general natural risks for approximate population of 3000 individuals. The an area. To avoid this trap, natural risk assessments Yasin River flows through the muddle of Hundur should not focus on a singular process but on village dividing it into tow portion. Hundur village multiple processes. They develop a general has been hit thrice by devastating debris flow, in methodology for Multi-Hazard Analysis in Natural 1978, 2009 and 2010. Risk Assessments of Bíldudalur village, NW Figure 1: Location map of Hundur Village Reduce Exposure to Reduce Risk 110

4. CONCEPTUAL FRAMEWORK AND This comprise of five steps, including selection of METHODOLOGY case study, procurement of images, off-site The conceptual framework and methodology used assessment, on-site assessment and, data for this research is summarized in the following consolidation and analysis. Each step has been chart (Fig:3). explained below: Regional Level Rapid Hazard Off-site Image visual interpretation assessment for Identification of Assessment Demarcation of Living and Priority areas Participatory Risk Activity zones Assessment Acquisition of remotely sensed data Preparation of Base (satellite images) Coordination with Community Organizations Data Consolidation and Analysis Hazards Assessment On-site Assessment Validation of Activity and Living Zones Data Collection using GPS Outputs Hazards Map Risk Map Reports Figure 2: Research Methodology Figure 3: Rapid Hazard Assessment Map Figure 4: Satellite Image of the study area 111 Reduce Exposure to Reduce Risk

4.1 Regional Level Rapid Hazard Assessment 4.3 Off-Site Assessment Rapid hazard assessment is a process carried out in It is one of the important phases of the order to identify about the number and extend of methodology. This is done before going in to field. geo-hazards exists in a valley (up to 20 villages). The acquired satellite image is analyzed by the Hazard indication map is prepared considering the team, for identification of prominent hazards and following indicators: Type of Hazards, Extend of sites for detailed assessment during the field visit. In Hazards, Past Events, Catchment area, vulnerable this phase the team members, under the supervision population. of team leader, gather basic information about the case study area, i.e. distance from the office and 4.2 Acquisition of Remotely Sensed Data town, mode of accessibility, coordination with the After the selection of the case study, second phase community organizations (Local Support of the assessment is procurement of satellite image Organization, Village organization, Community from the authorized vendors, Geo-eye, IKNOS and Emergency Response Team) and the regional SUPARCO, in Pakistan. Mostly 0.6 meter government. Based on the above information, the resolution maps repurchased with false color team collectively demarcates the activity zones4 and composite and four bands. This is used for the living zones [7] within the village. The initial preparation of base map for the field survey, by the observations and findings are verified and validated team. at the field level. Figure 5: Digitization of Activity zone Figure 6: Digitization of living zones Figure 9: Field hazard assessment map Figure 7: Participatory risk assessment in the village Figure 8: Hazard assessment during Figure 10: Vectorization of hazards zones in field visit ArcGIS 4 All areas the community visits on a daily basis. Reduce Exposure to Reduce Risk 112

4.4 On-Site Assessment the designed geodata base (Fig;10). The geodata The team consists of geologists, GIS specialist, data base consists of two feature classes, i.e. Risk enumerator and community representatives. Most of Assessment and Tables. Risk Assessment consists the data for the assessment and analysis is collected of activity zones, living zones and hazards zones during this phase. The phase is divided into two shape files, whereas the Tables are used for the parts, i.e. Participatory Risk Assessment (PRA) and entry of vulnerability data of settlements and living Geo-hazard assessment. During PRA, groups of zones. community members from different backgrounds are gathered for collection of basic information 4.5.1 Vectorization of Data about the village. To ensure inclusiveness, Manual hazard mapincludes information regarding vulnerability data related to susceptibility, coping type, intensity and classes of each hazard, i.e. debris and adaptive capacities and exposure of critical flow, rock fall, flood, snow avalanche and landslide. facilities to hazards, as well as population statistics, It is scanned and converted into a digital format, for are collected during the PRA exercise. The second further processing. In ArcGIS, geo-referencing tool phase is completed by a team of geologists. The is used to convert the .jpg format of scan map to .tiff team mostly comprises of two geologists, equipped or .img format, which is compatible for digitization with Inclinometer, Range Finder, Brunton compass, in ArcGIS software. All the raster data on the binocular, digital cameras and GPS. The manual hazard map is converted into vector data. classification and intensity definition of each hazard Attribute table is populated from the information is based on preset indicators, i.e. catchment area, mentioned in the field hazard map. Activity and types of hazard, slope, geology, geomorphology, living zones are digitized in separate shape files. sediment composition, impact area, historical profile Data about all critical infrastructures is entered into and triggering factors. The team develops manual the infrastructure shape file. In ArcGIS, geo- hazard map on the base map prepared for the field in referencing tool is used to convert the .jpg format of the first phase through visual observations and scan map to .tiff or .img format, which is compatible measurement. Here, they define the type, classes for digitization in ArcGIS software. All the raster and intensities of each hazard. The team enumerator data on the manual hazard map is converted into is responsible for collection of statistics related to vector data. Attribute table is populated from the critical facilities and houses that are exposed to any information mentioned in the field hazard map. hazards. Activity and living zones are digitized in separate shape files. Data about all critical infrastructures is 4.5 Data Consolidation and Analysis entered into the infrastructure shape file. The data collected during the off-site and on-site assessment is entered, processed, and analyzed in Figure 12: Risk Analysis ModelBuilder query Figure 11: Geodatabase structure Reduce Exposure to Reduce Risk 113

4.5.2 Populating Attribute Tables: 4.5.4 Risk Index Calculation Here, data gathered during field survey is entered The method6 of calculation of risk index involves into the tables. There are two formats that are used four indicators, i.e. Exposure (E), Susceptibility (S), during participatory risk assessment, i.e. vulnerable Coping capacities (C) and Adaptive capacities (A). parameters of settlement and living zones. Exposure of activity and living zones are extracted from the overlay analysis in ArcGIS whereas the 4.5.3 Spatial Analysis for Calculation of values of the remaining three indicators are taken Exposure to Hazard from the composite coding of the data collected Exposure is defined as people, property, systems, or during participatory risk assessment. other elements present in hazard zones that are The total vulnerability index V is calculated as: subject to potential losses. There are two functions that need to be performed for calculation of V = (S + (1 - C) + (1 - A)) / 3…… 3] exposure, i.e. intersectional area of activity and living zones over hazards zones, and significance5 The total Risk Index7 R = E x V, of hazard, as calculated by FOCUS Afghanistan Where; HVRA team. Spatial analysis tool in ArGIS is used E = Exposure, V = Vulnerability to calculate the intersectional area. A query in Model Builder is developed in order to automate the 4.5.5 Hazard Maps process of spatial analysis. Hundur village has five The combination of all the shape files, i.e. hazards living zones named as Shigathen, Hundur Bar, zones, living zones, activity zones, infrastructures Burakot, Rahimabad and Mominabad. These graphs and settlements points, along with preset symbols, show exposures of the activity zones. Hundur are used to create Hazard Maps. The geo-hazards village has five living zones named as Shigathen, found in Hundur village are debris flow, rock fall, Hundur Bar, Burakot, Rahimabad and Mominabad. snow avalanche, flooding and landslide. The These graphs show exposures of the activity zones. following map will clearly define the hazards situation in the village. Figure 13: Exposure of activity zones to geo-hazards Figure 14: Exposure of living zones to geo-hazards Risk Indices of Living and Activity zones Zones Total Exposure (%) Vulnerability (%)=(S+(1-c)+(1-C))/3 Risk Index (%) Shigthan 42.71 26.99 11.53 Hundur Bar 28.5 26.99 7.69 Burakot 53.99 26.99 14.57 Rahimabad 13.90 26.99 3.75 Mominabad 39.53 26.99 10.67 Living Zones 48 Activity Zone 43 26.99 12 Table 1: Table showing the calculated values for risk Figure 15: Graph showing risk index of both living and index of the living and activity zones activity zones 5 The significance of a hazard is an indication of its potential for 6 This method is largely inspired by BündnisEntwicklungHilft, destruction and is expressed as a function of the intensity and the return period. (Disaster Risk Management Initiative, AKDN, World Risk Report 2012, available at 2012) http://www.worldriskreport.org 7 Calculation of the global risk index (source: http://www.worldriskreport.com) Reduce Exposure to Reduce Risk 114

Table 3: Table showing vulnerable critical facilities Focus Humanitarian Assistance Pakistan Hundur Village, District Ghizer Sr. Name of Facility Total Risk Hazard Hazard Probability of Mitigation Measures/Recommendations No. No. Status Class Type Damage 1 Army Public School 1 1 Confirm Debris Flow High This school was demaged in 2010 floods, and its construction in the same place is in progres.School mangement should be informed to rebuild it in relativley safer area. 2 Jamat Khana# 3 1 Potential Debris Flow Low Strong outer boundary wall can be constructed to minimize the possible risk. 3 Jamat Khana 4 Jamat Khana# 4 2 Relatively Safe 6 1 Inferred Rock fall >Mdeium to Strong outer boundary or gabion wall can reduce the risk of jamat khana damage in future disasters. High< 5 Jamat Khana# 5 1 Potential Debris Flow >Mdeium to Construction of Gabion walls at the vulberable spot High< 6 Jamat Khana# 6 1 Potential Debris Flow Low Strong outer boundary wall can be constructed to minimize the possible risk. 7 Brokhud Transformer 1 1 Inferred Bank >Mdeium to It may be shifted away from the bank collapse hazard Collapse High< 8 Water Tank 1 1 Confirm Debris Flow High It should be shifted to relatively safer area 9 Houses 56 Confirmed Debris Flow High The damaged houses should be rebuild in a relativley safer area 10 Houses 80 Potential Debris Flow Low Nil 11 Houses 12 Houses 104 Relatively 13 Houses Safe 14 Houses 43 Inferred Debris Flow >Mdeium to Strong boundary wall can be constructed to reduce the risk. 15 Houses 317 High< 16 Brokhud Bridge 5 Potential Rock Fall Low Nil 27 Inferred Rock Fall >Mdeium to Retention wall may be errected at the back of the houses 2 Inferred High< 1 1 Confirm Rock Fall >Mdeium to Retention wall may be errected at the back of the houses and Debris High< Fall Debris Flow High Construction of protective wall at the overtopping area of debris in case of debris flow can reduce the risk of bridge damage The scale of the hazard map is 1:10,000 meters. Several snow avalanches fall into nallah in Hazard map explains the existing situation of the catchment area but it has no impact on settlement prevailing geo-hazards with specific symbols and area directly but it contributes to increase volume of probability of occurrence with different colors water level due to melting. scheme according to the following criteria in the Table 2. Hundur is surrounded by highly rugged 4.5.6 Risk Maps mountains and possesses moderate to high relief. The same scale is used for preparation of risk and Barren, highly weathered, and shattered rock out hazard maps. The range of vulnerability can be crop covers the village. Yasin River has made its identified by overlaying the infrastructure shape file path through these huge mountain ridges from East over the hazard zones. Table 3 shows the to West direction. Two main Nallah’s are flowing vulnerability of critical facilities. The risk map through the village, HundurNallah in the direction guides the community and government departments of East and Thirchilnallah in West. On both sides of about safer construction sites within the village. Hundurnallah unconsolidated material is deposited due to debris flow and flooding, overlying by 5 KEY FINDINGS glacier moraines. Small and large size debris flow The study covers detailed assessment of all the geo- and flooding events have occurred several times, hazards and their potential threats to the village and resulting in damages to houses and productive land. its dwellers. The study reveals the following Highly steep igneous as well as metamorphic rocks findings: cover the area on both sides. The rocks exposed in the watershed are mostly slates along with some  The noticeable geo-hazards in the village are skeletal moraines. In some places rock fall is debris flow, rock fall, flood, snow avalanche common, like on both sides of Nallah there is and landslide. inferred threat of rock falling due to shattering, intersecting joints and also presence of glacier  A large fan has formed as a result of debris moraines. Large catchments area, high gradient and flow events in Terchilnala, which divides into abundant material in Yasin River make the area three main confirm creeks. flood prone. Reduce Exposure to Reduce Risk 115

 Less vegetation and high slope gradient make  Proper structural mitigation (Gabion wall, the nallah8flood prone protective wall) can be carried out to save the land from riverine flood and erosion along the  The village was hit by three devastating debris bank of Yasin River. flow events, in 1978, 2009 and 2010  New construction should follow Building  Experts from FOCUS had successfully Seismic Codes 2007 of Pakistan. anticipated risk of floods and high intensity debris flow in 2009 through rapid hazard  Seasonal calendar of hazards can be prepared assessments9. for the. The village emergency response team should be well trained and aware of the  Debris flow is the most destructive geo-hazard different natural disasters. in the village. In 2010, debris flows hit the village and destroyed several houses, partially  Awareness sessions about natural hazards damaged 2 schools and also destroyed land should be conducted in the village  Exposure analysis concluded that 21% area is  The risk of snow avalanche and rock fall can exposed to rock fall, 18 % to debris flow, 3.3 be reduced through plantation % to flood and 1.03 % to bank erosion, whereas landside has very low impact on the 7. CONCLUSION activity zone and no impact on living zones. The GIS based geo-hazard assessment model is useful tool for assessing multi-hazards collectively,  Exposure analysis of living zones reveals that because it can be used for calculating the risk to a 21% area of Shigathen is vulnerable to debris community. The outcomes of such models are flow, 20% to rock fall, 2% to flood and the rest efficient and have the potential to guide the end is relatively safer area. Hundur Bar is 40% users, i.e. land-use planners, decision makers and prone to rockfall, 14% to debris flow, while 13 stakeholders for developmental activities in hazard % area of Rahimabad is vulnerable for rock prone mountainous areas. fall. Mominabad is 22% vulnerable to rockfall, 12% for floods and 6% area is exposed to ACKNOWLEDGEMENT debris flow. The author acknowledges the support offered by FOCUS HVRA team members and the  There are many dry creeks on both sides of the management. All data for this report has been river and may damage the cultivable land provided by Focus Humanitarian Assistance during heavy rains. Pakistan.  Seismically, Hundur village is located in the REFERENCES zone-3 in the Pakistan Seismic Map, developed by Geological Survey of Pakistan. [1] R. Bell, T. Glade., 2004, Multi-Hazard Analysis in Natural Risk Assessments. Department of 6. RECOMMENDATION Geography, University of Bonn, Germany. Based on the key findings, the following recommendations are being made: [2] Greiving, S. 2006,Integrated risk assessment of multi-hazards: a new methodology. Natural and  The debris flow path may be properly technological hazards and risks affecting the channelized, to reduce the impact on both sides spatial development of European regions. of the village Geological Survey of Finland, Special Paper 42, 75–82.  Multiple gabion walls may be constructed on both side of thenallato reduce future damages [3] Birkmann. J. et al., 2011, World Risk Report. Germany:BündnisEntwicklungHilft (Alliance  Construction of check dams in the nalla is Development Works) essential for minimizing the threat of flooding [4]Othmana .A. N. et al. 2011, GIS based multi-  Vulnerability of the nallahcan be reduced criteria decision making for landslide hazard through plantation zonation. Department of Surveying Science and Geomatics, Faculty of Architecture, Planning  Construction should be avoided in the red zone and Surveying, UniversitiTeknologi MARA, Shah of the alluvial fan and other hazard prone. The Alam, 40100, Selangor, Malaysia. area can be used for cultivation and plantation purposes [5]Jian.W, Xiang-guo .P. 2009, GIS-based landslide hazard zonation model and its application. 8 It is an Urdu word used by community instead of channel or open path for debris or flood flow 9 FOCUS Pakistan, 2009, (Internal Report) Reduce Exposure to Reduce Risk 116

School of Environment Science and Spatial Polytechnic University, Yuk Choi Road, Hung Informatics, China University of Mining & Hom, Kowloon, Hong Kong, China Technology, Xuzhou 221116, China. [7]Henriod.S. and Merchant.F., 2012, AKDN [6]Chau. K.T. et al., 2003, Landslide hazard Disaster Risk Management Initiative – FOCUS analysis for Hong Kong using landslide Standardized HVRA Mapping Strategy, DRMI inventory and GIS. Department of Civil and Coordination office Dushanbe-Tajikistan. Structural Engineering, The Hong Kong Reduce Exposure to Reduce Risk 117

DEVELOPMENT OF AUTONOUMOS MULTI AGENT SYSTEMS FOR QUALIATAIVE RISK ASSESSMENT IN DISASTER MANAGEMENT D.S. Kalana Mendis1, Asoka S. Karunananda2, Udaya Samaratunga3 and Uditha Rathnayake4 1Department of Information Technology, Advanced Technological Institute, Dehiwala, Sri Lanka E-mail: [email protected] 2Faculty of Information Technology, University of Moratuwa, Sri Lanka, E-mail: [email protected] 3Gampaha Wickramarrachi Ayurveda Institute, University of Kelaniya,Sri Lanka E-mail: [email protected] 4Department of Electrical and Computer Engineering, Open University of Sri Lanka, Sri Lanka E-mail: [email protected] ABSTRACT Developing autonomous multi agent systems are to be considered an advancement of multi agent systems can be applied in both physical and logical world. Constructions of multi hazard risk assessment using spatial data for disaster management have a problem of effective communication because of implicit knowledge. Risk management is the identification, assessment, and prioritization of risks followed by coordinated and economical application of resources to minimize, monitor, and control the probability and/or impact of unfortunate events. Constructions of risk assessment using spatial data for disaster management have a problem of effective communication because of implicit knowledge Risk assessment is the determination of quantitative or qualitative value of risk related to a concrete situation and a recognized hazard. Quantitative risk assessment requires commonsense knowledge related with the hazard. This complicates the effective communication of data to the user in real-time machine processing in support of disaster management. In this paper we present an approach to modeling commonsense knowledge in qualitative risk assessment using Autonomous multi agent system. This gives three-phase knowledge modeling approach for modeling commonsense knowledge in, which enables holistic approach for disaster management. At the initial stage commonsense knowledge is converted into a questionnaire. Removing dependencies among the questions are modeled using principal component analysis. Classification of the knowledge is processed through fuzzy logic module, which is constructed on the basis of principal components. Further explanations for classified knowledge are derived by expert system technology. We have implemented the system using FLEX expert system shell, SPSS, XML and VB. This paper describes one such approach using classification of human constituents in Ayurvedic medicine. Evaluation of the system has shown 77% accuracy. KEY WORDS: Risk Assessment, Disaster Management, Commonsense Knowledge Modeling, Autonomous Multi Agent Systems IMPLEMENTATION OF DISASTER MANAGEMENT INSTITUTIONS IN BALUCHISTAN aSyed Ainuddin and bJayant Kumar Routray a Faculty of Earth and Environmental Sciences, Department of Geography, University of Baluchistan Quetta, Pakistan b Professor and Coordinator of Regional and Rural Development Planning and Disaster Preparedness, Mitigation and Management, Asian Institute of Technology, Thailand. ABSTRACT Baluchistan is vulnerable to a number of natural hazards and disasters such as earthquakes, droughts, floods, and tsunamis. After 2005, Earthquake in Kashmir, Pakistani government formulated National Disaster Management Authority (NDMA) in 2007, to handle the disaster crisis and subsequently formulated the Provincial Disaster Management Authorities in all the provinces. The objective of this paper is to analyze the decentralization of the disaster management institutions in Baluchistan, using Key Informant interviews of the involved government and non-governmental organizations in disaster risk reduction and management activities. The results revealed that disasters in Baluchistan are handled at the provincial level and the disaster institutions are not yet Reduce Exposure to Reduce Risk 118

implemented at the district and community levels. This has exacerbated the people’s vulnerability to the impacts of disasters to a considerable level. The paper recommends efficient preparedness and coordination of provincial and national level agencies to enhance community awareness and preparedness. In addition, the paper concludes that disaster management authorities should implement programs/projects and activities to empower communities for disaster risk reduction and management in the province. KEY WORDS: Disasters, Institutions, Decentralization, Risk Reduction, Baluchistan. Reduce Exposure to Reduce Risk 119

Reduce Exposure to Reduce Risk 120

Technical Session-5 [Hall C]: Earthquake & Tsunami Tsunami Damage Assessment Using Artificial Neural Networks 122 Saher Murad, Umair Ahsan Cheema 126 Early Earthquake & Tsunami Warning through Low Cost Vibration Detection 130 Mechanism 136 K.P.D.H De Silva, M. P. A. W. Gamage, G.R.S.S Jayarathne, K.P.K Srilal, G.I.U Geeganage, J.C.R Wimalasuriya Impact after Tsunami 2004 Ghouse Mohamed Shaik, Kumar D, Ramakrishnan.V. V, Radhika .J. G Preliminary Study on” Intraplate” Earthquakes in the Indian Ocean Jayathilaka, R.M.R.M., Jinadasa, S.U.P, Wijayadeva, D. A Reduce Exposure to Reduce Risk 121

TSUNAMI DAMAGE ASSESSMENT USING ARTIFICIAL NEURAL NETWORKS Saher Murad and Umair Ahsan Cheema National University of Computer & Emerging Sciences A.K. Brohi Road, H-11/4, Islamabad, Pakistan [email protected] ABSTRACT This paper focuses on how Remote Sensing and Geographic Information System (GIS) can be used in damage assessment for Tsunamis. Tsunami is a calamity that can cause extensive damage to buildings, infrastructure, transportation networks, agriculture and industry throughout the afflicted area. The project is primarily based on multi-sensor data fusion i.e. combining different bands of data (optical, radar) in order to get more meaningful information and improve the resulting damage assessment map. An Artificial Neural Networks (ANN) based algorithm is used for preparing damage assessment map for North Coast of Banda Aceh-Sumatra. Erstwhile Damage Assessment Maps were produced by using change detection algorithms, expert systems and fuzzy classification. To implement this, tools such as ENVI, ESRI ARCGIS, ERDAS IMAGINE, Java Neural Network Simulator (JNNS) and Arc Macro Language (AML) are used for image processing and manipulation. The different datasets from satellites are processed and subsequently used for Building Density Map and Change Maps. This ancillary data will be fed into ANN as an input along with high resolution satellite imagery to produce Damage Assessment Map. The Neural Network Based Damage Assessment Map has correctly identified 83% of the severely damaged area and hence is a viable option for preparing damage assessment map. Remote Sensing and GIS technologies have become a valuable resource for the disaster management agencies. GIS technology enables emergency planners to easily calculate emergency response times in the event of a natural disaster. Natural disasters are inevitable and can cause massive damage to buildings, infrastructure and environment consequently leading to heavy casualties. It is though impossible to avoid occurrence of these disasters, however with effective development plans, its intensity can be greatly reduced and can assist in recovery and rehabilitation from the disaster. These Damage Assessment Maps are therefore of immense value to disaster management agencies. Keywords: Disaster Management, Remote Sensing, GIS, Change Detection, Artificial Neural Network 1. INTRODUCTION along the coasts of most landmasses bordering the Indian Ocean, killing large numbers of people and inundating Remote sensing and Geographical Information System coastal communities across South and Southeast Asia, (GIS) techniques have become increasingly important including parts of Indonesia, Srilanka, and Thailand. source to provide effective and useful tools for the Its aftermath was even more horrifying as 126000 people disaster assessment. Previously, damage assessment lost their lives, 94470 people were reported missing and maps had been produced by using change detection 412438 were internally displaced [Bohun et al., 2005]. algorithms such as subtraction, ratio, correlativity and These statistics motivated me to prepare assessment map comparability, expert systems and fuzzy classification of North Coast of Banda Aceh, Sumatra; in order to [Cheema, 2006]. contribute with my skills in not only coming up with more innovative ideas, but also assist in minimizing 1.1 Objectives: damage ensuing a disaster and subsequent rehabilitation. Aim of this paper is to produce damage assessment map 1.3 Study Area for natural calamity such as Tsunami etc. by using Table 1: Geographical extent of the study area Remote Sensing and GIS along with Artificial Neural Networks (ANN). The project is primarily based on data Latitude 5°27'40.66\" N 5°26'16.72\" N fusion relying on the combination of multi-sensor images Longitude 95°14'11.70\" E 95°15'42.54\" E to improve accuracy of damage assessment map. 122 Damage assessment maps are essential to fore plan and prioritize relief efforts by determining the extent of expected damage to life, property and environment caused by a disaster. 1.2 Motivation: The Indian Ocean earthquake, known as the Sumatra- Andaman earthquake, was an undersea earthquake which occurred at 00:58:53 UTC (07:58:53 local time) on December 26, 2004 with a magnitude of 9.0 and its epicenter was off the west coast of Sumatra, Indonesia. The earthquake triggered a series of devastating tsunamis Reduce Exposure to Reduce Risk

Source UNOSAT UNOSAT ESA ESA Aster Data Type IKONOS (Optical Imagery) IKONOS (Training Data) ERS-2 (RADAR) ASAR-2 (RADAR) DEM Figure 1: Location map of the study area Figure 2: IKONOS images of (a) Pre-Tsunami & (b) (Source: Google Earth) Post-Tsunami; Radar images of (c) Pre-Tsunami (ERS-2) & (d) Post-Tsunami (ASAR-2) 2. METHODOLOGY: For the most accurate results, images pre-process 2.2 Building Density Maps: included geometric correction (geocoding, ortho- Building density maps were created for both pre and post rectification using DEM) and image enhancement tsunami images. For this, the optical images were first (speckle filtering). classified into three classes - i.e. land, water and vegetation using maximum likelihood classification. The 2.1 Datasets: resulting classified maps were rescaled between 0 and 1, For each layer spatial reference was defined as: where „0‟ shows water and vegetation and „1‟ shows Datum: WGS_1984 areas with more proportion to buildings. Projection: UTM Zone: 46N Table 2: Data used in this study area Data Acquisition January 25, 2005 (Post-Tsunami) December 29, 2004 Data Acquisition November 12, 2002 (Pre-Tsunami) January 10, 2003 Resolution 25 m 30 m 30 m Figure 3: (a) Maximum likelihood supervised 4m classification of pre-tsunami IKONOS image, (b) Building density map (multispectral) 4m 2.3 Change Detection Map: (multispectral) 123 Reduce Exposure to Reduce Risk

Change detection algorithm was used to create change map from supervised pre and post tsunami maps. Figure 4: Change Map Figure 5: Process Diagram for Tsunami Damage Assessment The fig. 4 shows a change map with three change classes - i.e. Negative change identifies that the object does not 2.5 Results and Discussions: exist anymore which previously existed in pre-tsunami, and vice versa in positive change. 2.5.1 Accuracy Analysis: The result of the accuracy analysis for optical imagery The resulting change map was rescaled between 0 and 1, was 76% and for RADAR imagery it was 83% which i.e. damage and no-damage classes respectively. were achieved by comparing testing and damage assessment maps with high degree of confidence level. 2.4 Training Artificial Neural Network: In order to prepare damage assessment map for study area, Java Neural Network Simulator (JNNS) was used. JNNS is a Java based simulator for Artificial Neural Network which is used for preparing Damage Assessment Map. Arc Macro Language (AML) was used to prepare the dependent and independent datasets for JNNS. 2.4.1 JNN Classification: Red band in building density maps for both pre and post tsunami yielded maximum statistics i.e. maximum mean and standard deviation, therefore it was selected to train the JNN for ERS-2 and ASAR respectively. The results were then differenced in order to obtain change map. 2.4.2 Damage Assessment Maps: Firstly, training dataset which composed of two classes i.e. damage & no-damage, was split into training set and was used to train the JNN and testing set so as to verify the generalization ability of JNN, using ratio of 70% and 30% respectively. Different training parameters were experimented, such as learning rate, momentum, etc. in order to attain the best result. Secondly, sequentially optical and radar change maps were fed as independent layer. Finally, JNNS used „Feed- Forward Neural Network‟, back-propagation technique to create damage assessment map. Reduce Exposure to Reduce Risk 124

Figure 6: Damage Assessment Map for North Coast of ACKNOWLEDGEMENTS: Banda Aceh, Sumatra derived by Neural Networks First and foremost, I thank Allah Almighty for all His Fig. 6 shows an overview of damage assessment map for blessings and courage He granted me to complete this North Coast of Banda Aceh, Sumatra. On this map two research project. classes are highlighted, red color indicates the damage areas which includes coastal shipping and industrial I owe special thanks to my parents and husband for their facilities, pristine beaches, lush vegetation, road network support, motivation, encouragement and assistance with and commercial and residential areas. It can be full love and affection all along. ascertained from this map that tsunami had great impact not only on land but also on populated areas alike. I am also grateful to my supervisor Mr. Umair Ahsan Cheema, who gave me an opportunity to explore and The extent of the inundation and total area destroyed can study this topic. This project could not have been be calculated against the magnitude of the tsunami which completed without his valuable contribution and can enormously assist in damage assessment and risk therefore, I thank him for helping me all the way through reduction. with enormous patience. 3. CONCLUSION: The datasets used for this project were graciously Damage Assessment Maps can be prepared by using provided by European Space Agency (ESA). I am also many different algorithms. Certain algorithms require indebted to Dr. Shawn Laffan of the University of New only satellite images while others also necessitate South Wales, Australia for providing the AML tools. inclusion of ancillary data to improve accuracy. These maps can be very useful for disaster risk management REFERENCES: (DRM). Bohun, V. and Bank., A. D., 2005, Report and Neural Networks can prove quite useful for image recommendation of the president to the board of classification and damage assessment as these can directors on proposed grants to the Republic of incorporate multi-sensor datasets. The accuracy obtained Indonesia for the Earthquake and Tsunami for identification of most severely damaged areas along Emergency Support Project and contribution to the the North coast of Banda Aceh, Sumatra is around 76% Multidonor Trust Fund. Asian Development Bank, for optical imagery and 83% for RADAR imagery and (Manila, Philippines). can be improved by using datasets with improved spatial and spectral resolution, qualitative field data and Cheema, U., 2006, IEEE International Conference on temporal images. The key finding of this study is that Advances in Space Technologies “Expert systems Neural Networks is a viable option for conducting natural for earthquake damage assessment”. disasters damage assessment by the fusion of optical and radar imagery. Jensen, J. and Lulla, D., 1987, Introductory digital image processing: a remote sensing perspective. Ur, K. and Shad, R., 2006, IEEE International Conference Advances in Space Technologies “Disaster management and GIS”. Yang, X., Lan, R., and Yang, Q., 1999, Asian Conference on Remote Sensing, Hong Kong, China “Change detection based on remote sensing information model and its application on coastal line of yellow river delta”. Reduce Exposure to Reduce Risk 125

EARLY EARTHQUAKE & TSUNAMI WARNING THROUGH LOW COST VIBRATION DETECTION MECHANISM K.P.D.H De Silva, M. P. A. W. Gamage, G.R.S.S Jayarathne, K.P.K Srilal, G.I.U Geeganage and J.C.R Wimalasuriya Sri Lanka Institute of Information Technology, Sri Lanka [email protected], [email protected], [email protected], [email protected], [email protected], [email protected] ABSTRACT Warning to the general public about an on coming earthquake or a Tsunami in a timely manner could prevent loss of invaluable lives and properties and prepare for a quick recovery with less cost. This research aims to find a solution for early warning of such a disaster through a low cost vibration detection unit. Third world countries, especially islands like Sri Lanka, don’t have their own systems to detect vibrations within or near and therefore will have to rely on information provided by out sources. The signals received by the vibration sensor are calibrated and standard formulas were applied to convert the readings to Richter scale. The detected signals are sent to the system through GSM technology and the generated warnings are sent to general public through SMSs. The software component acquires related information through USGS using RSS feeds to produce necessary information of an earthquake occurred. It also helps to calculate the arrival time of a Tsunami in a certain region. KEYWORDS: Vibration Detection, Seismic vibration sensing, GSM ( Global System for Mobile Communication ) 1. INTRODUCTION Vibration detecting unit of the system is designed to Disaster handling is a very important part of the global work with low power consumption. The device detects security. Most of the social organizations try to develop vibrations within few seconds and communicates the more reliable and efficient disaster handling systems result with main system to send warning messages to the with the evolution of information technology. users immediately. It also retrieves hazard information Earthquake and Tsunami handling is one of the major from an external source. System manipulates the areas in the disaster management. information acquired and presents it to the people in meaningful form. System contains some functions to Earthquake and Tsunami events are very destructive than facilitate the refugee needs in a post disaster situation. some other disaster situations in the world. Therefore handling these types of situations are very critical for the 2. METHODOLOGY global conservation. Reliability, efficiency, availability There are two major components of the system, and security are the main considerable factors to make a successful system in this field. 1. Vibration Detection Unit 2. Alert System Earthquake detection is a complex process ( NOAA, 2009) that needs high technical knowledge, expensive 2.1 Vibration Detection Unit detection equipments, effective and reliable This hardware component is used to detect vibrations communicational modules and accurate measuring of the from the earth and sends it to the server machine. This related scales. As a developing country, Sri Lanka cannot unit contains the following parts; afford this kind of high standard development. However at the moment, disaster management center handles the a. Main controller unit post disaster matters of the country effectively.Main goal b. Sensing unit of this research is to implement the system with least c. Power supply unit cost but high efficiency and accuracy. It contains d. TTL to RS232 converting unit mathematical formulas to calculate accurate results, e. LED indicator reliable and effective communication models, ability to Main controller does the processing of the vibration detect external information and handles post disaster detecting unit. Microcontroller manipulates and stores matters. the input signal that is given by the sensing unit. It also makes the communication link with the GSM modem. Reduce Exposure to Reduce Risk 126

Sensing Power LED Unit Supply Indicator Main TTL to Controller RS232 GSM Figure 3: Structure of Vibration Sensor Modem The power supply unit acts as the distribution unit of Figure 1: Vibration Detecting Unit Design. power for the sub circuit components. It delivers power to the controller unit, sensing components and RS232 Sensing unit was introduced by considering the operation converter unit. This unit is capable of handling 9 volts to principle of actual Seismometer. This unit converts 20 volts AC or DC voltages and reduces it into pure 5 seismic energy inputs (or vibrations) into electrical volts DC for fulfilling the powering requirements of the voltage which can be accurately measured. The sensor previously mentioned components. consists of a mass, which contains an electrical coil that is suspended by a spring between the poles of a magnet. TTL to RS232 converter is to match the interfaces TTL and RS232. The GSM modem supports RS232 port +5V +5V +5V +5V +5V connection but microcontroller has TTL interface. The IC MAX232 used to convert the two interfaces. There R5 C5 1 are two LED indicators in this unit. First LED is to 10K 0.1uf 2 indicate that a vibration is detected by the unit. Second 3 LED is to verify the vibration period is end. VD D 11 13.5V ICD-MCLR 4 VD D 32 ICD-RB7 5 Figure 4: Vibration Sensor 6 JP6 ICD-RB6 ICD2 GSM (Global System for Mobile Communications) technology is used to make the connectivity between the ICD-MCLR 1 MCLR/VPP RC2/CCP1 17 device and server machine. GSM technology ignores the ICD-RB7 40 RB7 RC4/SDI/SDA 23 ionosphere and impact to the connection. Therefore the ICD-RB6 39 RC3/SCK/SCL 18 system contains a reliable connection without any noise. 24 After retrieving device signal, system sends the warning +5V +5V RB6 RC5/SDO SMS automatically to people within few seconds without involving any 3rd party. R5 2 RC0/T1OSI/T1CLK 15 10K 3 16 Vib_Lev el 4 RA0 RC1/T1OSO LED Vibration Detect C13 Vib_Detect 5 19 1uf 6 RA1 20 Vib_Lev el 7 21 Vib_Detect RA2 22 33 27 1 C6 34 RA3 RD0/PSP0 28 2 15pf 35 29 3 36 RA4/T0CLK RD1/PSP1 30 4 Y1 37 5 4Mz 38 RA5/SS RD2/PSP2 6 7 C7 RD3/PSP3 8 15pf RD4/PSP4 JP5 VibrationSensor RB0/INT RD5/PSP5 RB1 RD6/PSP6 RB2 RD7/PSP7 RB3 RB4 RB5 13 OSC1/CLK RE0/RD 8 RE1/WR 9 RE2/CS 10 14 OSC2/CLKOUT RC6/TX 25 Tx RC7/RX 26 Rx 12 GN D 31 GN D U3 PIC16F877A Figure 2: Circuit diagram of main controller When the earth moves, the magnet and spring support move with the earth. The mass tends to remain stationary, so its motion will lag behind that of the magnet. This relative motion produces a voltage that is proportional to the velocity of this motion. Then the generated voltage signal drives through an operational amplifier and low pass filter in order to amplify the signal and cutoff unwanted signals respectively. Then the amplified signal is fed to the microcontroller for further processing. Reduce Exposure to Reduce Risk 127

Figure 5 : Structure of the Proposed Warning System 2.2 Alert System 3. RESEARCH FINDINGS AND RESULTS The purpose of the Warning system is to handle and Many number of experiment readings were used to make manipulate the hazard details and post disaster situations. the device specific amplitude scale. Standard Richter It has the following functions. formula and device base formula are as follows; a. Warning function Richter=log 10 (amplitude) b. Geographical representation Amplitude should be in micro meter c. Automated external information reader d. Concessionary camp function Amplitude = 0.011167917*1.0091547^x e. Hazard indication x=device value Vibration detecting unit is detecting the vibrations and Following table and the graph indicate the part of sends the details to the main server immediately. If that readings that are received by the unit and the relationship vibration level is over or near to the risky level, among them. notification service is automatically activated and passes the notification message through SMS to the registered Table I: Device Readings people in the web site. Circuit Values Amplitude (mm) Geographical representation provides basic details about (x) (y) the vibration generated areas. It provides nearest city and country to the point earthquake occurred, distance to the 568 2 coastal areas from earthquake origin points, nearest sea 613 3 and depth of the earthquake from the surface as the 640 4 geographical features. Google map represent the points 670 5 of the earthquakes. 700 6 790 16 Automated external information reader is very important 809 17 to handle some other functions in the system. The system 816 18 receives those data from RSS feed that is supplied by 821 20 United State Geological Survey ( USGS ). Hazard indication represents Tsunami speed, Tsunami arrival time and wave length using external details. These calculations are based on specific formulas and these values may change because of existing environmental situations. Reduce Exposure to Reduce Risk 128

Figure 6:The graph of device values vs. amplitudes REFERENCES 4. CONCLUSION AND FUTURE WORK [1] T.Chalko.(2008).“Quake Energy Monitor Future improvements and modifications can be done to software.” make the developed system more efficient. Some of them can be listed as follows. http://nujournal.net/quakes.rar [2] NOAA (.2009) “Pacific Tsunami Warning Center” a. Improvements of the sensibility to the higher http://www.prh.noaa.gov/ptwc rangers. [3] Disaster Management Centre of Sri Lanka.( b. Improvements of the smoothness and 2010).About Disasters consistency of the detections. http//www.dmc.gov.lk/Training%20&%20Learning/Ab c. Add another warning function through the outDisasters2.html system like siren alarm. [4]International Strategy for Disaster Reduction.( d. Introduce other registering ways with the 2004).Basic terms of disaster risk Reduction notification service. http://www.unisdr.org/eng/library/lib- The overall scope of this research is measuring vibration terminologyeng% 20home.htm level, warning people, to save human lives and properties [5]Auto Calculator Organization.(2009).Calculate from natural disasters like tsunami and earthquakes. The seismic vibration detection unit is the most important distance part which measures vibration level accurately, http://www.autocalculator.org/Calculators/Driving- efficiently and in a cost effective way. System also capable of sending tsunami warnings and save time, Distance-Calculator.aspx person effort, resources to boost disaster management in [6] Moveble type Sripts.( 2007) .Calculate distance, a more efficient way. When implementing this kind of system, related authority can further improve seismic bearing and more between Latitude/Longitude vibration sensing capacity and accuracy of the system by points locating more sensor nodes in different geographical http://www.movable-type.co.uk/scripts/latlong.html areas in the country. [7]W. A. Morrissey, “Tsunamis: Monitoring, Detection, and Early Warning Systems” (2005) http://www.fas/org/sgp/crs/RL32739.pdf Reduce Exposure to Reduce Risk 129

Impact after Tsunami 2004 Ghouse Mohamed Shaik*, Kumar. D1, Ramakrishnan.V.V2, Radhika.J.G2 Sri Venkateswara College of Engineering and Technology, Thirupachur, Anna University, Chennai E-mail: [email protected], [email protected], [email protected], [email protected], ABSTRACT Tsunami 2004 has created the awareness of managing the disasters and protecting the people in the Indian subcontinent. The Tsunami 2011 of Japan which is violent like 2004 Tsunami had affected Japan’s East coast resulting in loss of infrastructure near the eastern coastal zone. Considering the Tsunami of 2004 the loss of life is less. These pacific rim zone areas in American to Asian Coasts have earth quakes followed by mild tsunamis frequently and the people have the awareness to protect the society and property. Indian Ocean region from Northern Australian Islands, Indonesia, Malaysia, Thailand, Myanmar, India, Sri Lanka, Indian Ocean island nations and Himalayan region of India, Pakistan, Afghanistan experience about 4 to 5 Richter scale at an interval of about 6 months. One or two Heavy Earth quakes occur with Tsunami warning once in two year in Islands bordering Australia in the North, Indonesia, and Andaman Islands region of India. The residents in coastal zones are alerted by Tsunami warning systems to move to safer elevated areas. The buildings and structures are designed and constructed now adopting safer designs. Unfamiliarity of Tsunami in 2004 resulted in huge damage to life and property. The awareness we have and the availability new technology gives us strength to face the challenges and protect the society .This paper stress the need of management of coastal zones with vegetation and protective structures. A collaborative research is required in Indian Ocean region nations to understand the causes of frequent earth quakes after 2004 at Andaman Region of India, islands of Malaysia, Indonesia which are in Active zone. The south and east coasts of India and Sri Lanka meet erosive waves if earthquakes of higher intensity occur in this Indonesia – Andaman India region or in Indian Ocean ridge in the South which are facing these earth quake zone. Even in 2004 the coastal areas with mangroves and protective coastal civil works are least affected. The earthquakes above 4 Richter scale can affect the life in this zone as buildings and infrastructures are constructed with earthquake protection measures only from 2004. Research programmes are to be taken up or strengthened in these zones in association UN and International institutes like AIT by Indian Ocean nations in Asia and Australia using the computer modelling techniques with current data with 10th century Tsunami which swallowed part of Mahabalipuram Port city of Pallava kings and 1st Century Tsunami of submergence of Poompuhar Port city of Chola kings of Ancient in South India and popular tectonic theory of disintegration of Australia and South Indian subcontinent and its movement to North which resulted in the formation of High and active tectonic young Himalayan mountain region. With the computing facility for modelling and handling voluminous data from USGS and Asian region research centre it is observed we can know more about tectonic activity in the core of the earth, movement of tectonic plates and understand the causes for the active Tectonic activity from 2004 to take further protective measures to live in this region. KEY WORDS: Tsunami, Indian Ocean region, USGS, active tectonic zones, Research, linking present and pre historic tectonic events 1. INTRODUCTION 2. TSUNAMIS IN HISTORY A tsunami watch means there is the potential for a The lists notable historic tsunamis, which are sorted tsunami, not that one is imminent. Tsunamis are not by the date and location that the tsunami occurred, massive waves. Tsunamis come and they keep are given in Annex. The seismic and volcanic activity coming for hours. The water swells and swells and at tectonic plate boundaries along the Pacific Ring of swells for hours until the ocean owns the beach and Fire, tsunamis occur most frequently in the Pacific everything else in its path. Any Tsunami can hit the Ocean, but are worldwide it is a natural phenomena. coastal cities in the region with three or four Very small tsunamis, non-destructive and damaging Tsunami waves within an hour to four. undetectable without specialized equipment, occur frequently as a result of minor earthquakes and other Reduce Exposure to Reduce Risk 130

events. As early as 426 BC, the Greek historian needs the investigation of this active Indian and Thucydides inquired in his book History of the Pacific Ocean regions. Peloponnesian War (3.89.1-6) about the causes of tsunamis. He argued rightly that it could only be 4. HIGHEST OR TALLEST MEGA TSUNAMI explained as a consequence of ocean earthquakes, The tallest tsunami ever recorded so far is the 1958 and could see no other possible causes for the Lituya Bay megatsunami, 524 m (1740 ft). phenomenon. Crete and the Argolid and other locations were destroyed by a tsunami caused by the The others are: eruption of Thira, which destroyed Minoan 1980 Spirit Lake megatsunami, 260 m (780 ft) 1963 civilization on Crete and related cultures in the Vajont Dam megatsunami 250 m (750 ft) Cyclades and in areas facing the eruption on the 2011 March 11, off the Pacific coast of Japan, a 9.0 Greek mainland such as the Argolid. During the magnitude earthquake produced a tsunami 33 feet Persian siege of the sea town Potidaea, Greece, in (10 m) high along Japan's northeastern coast. 479 BC,the Greek historian Herodotus reports how The wave caused widespread devastation, with an the Persian attackers who tried to exploit an unusual official count of around 20,000 people confirmed to retreat of the water were suddenly surprised by \"a be killed/missing. The highest tsunami which was great flood-tide, higher, as the people of the place recorded at Ryōri Bay, Ōfunato, reached a total say, than any one of the many that had been before\". height of 97 feet (30 m). Herodotus attributes the cause of the sudden flood to the wrath of Poseidon. Other tsunamis that have 5. DEADLIEST 2004 TSUNAMI occurred include the following: The deadliest tsunami in recorded history was the 2004 December 26 which killed almost 230,000. ca. 500 BC: Poompuhar, Tamil Nadu, India, Tsunami 2004 has created the awareness of managing Maldives the disasters and protecting the people. Tsunami 2011 1000 AD Mahabalipuram. Tamil Nadu and Andaman of Japan which is violent like 2004 Tsunami affecting 1541 AD European settlement in Brazil, São Vicente. Japan’s East coast with loss of infrastructure near the which completely destroyed the Ancient Ports. eastern coastal zone is consider to have less loss of 20000 life though the property loss is huge as the Major Tsunami’s recorded in South Asia from pacific rim zone areas in American to Asian Coasts 1500 AD to 1950 are have mild tsunamis and earth quakes frequently and 1524 AD Near Dabhol, Maharashtra the people have the awareness to protect the society 2 April 1762AD Arakan Coast, Myanmar and property. 16 June 1819 AD Rann of Kachchh, Gujarat, India 31 October 1847 AD Great Nicobar Island, India 6. THE ACTIVE INDIAN OCEAN REGION 31 December 1881AD Car Nicobar Island, India Indian Ocean region from Northern Australian 26 August 1883 AD Krakatoa, Sunda Strait, Islands, Indonesia, Malaysia, Thailand, Myanmar, Indonesia India, Sri Lanka, Indian Ocean island nations and 28 November 1945 AD Mekran coast, Balochistan Himalayan region of India, Pakistan, Afghanistan experience about 4 to 5 Richter scale at an interval of No of Tsunami’s recorded in the past centuries are about 4 to 5 months from 2004 and one or two heavy 1 Before 1001 AD ( 6100 BC to 1000 AD) 9 earth quakes once in two years affecting the Islands 2 1000–1700 700 years 8 Tsunamis bordering Australia in the North, Indonesia, 3 1700s 100 Years 6 Tsunamis Andaman Islands region of India. 4 1800s 100 Years 7 Tsunamis 5 1900–1950 50 Years 8 Tsunamis 7. TSUNAMI WARNING SYSTEM AFTER 2004 6 1950–2000 50 Years 12 Tusamais The residents living in coastal zones are provided 7 2000s 11 Years 9 No of Tsunamis with Tsunami warning systems to move to safer As we have to consider historical records available elevated areas. The buildings and structures are now few number of tsunami’s could have been designed and constructed now adopting safer norms recorded till 1900 AD. However it should be noted in and designs. Unfamiliarity of Tsunami in 2004 1900 -1999 ( 20th Centrury) in 100 years period about resulted in huge damage to life and property. The 20 Tsunami’s have been recorded with modern awareness we have and the availability new devices. The occurrence of about 10 Tsunamis technology gives us strength to face the challenges within decade that too from 2004 Mega tsunami and protect the society. Reduce Exposure to Reduce Risk 131

8. TSUNAMI WARNING AND The ITEWC also issued an alert for coastal Tamil ALERTNESS IN 2011 Nadu, Andhra Pradesh and the Andaman islands An illustration is given here to show the alertness in forecasting the arrival time of the first wave. For the event of Tsunami. In 2011 India issued tsunami Tamil Nadu, the first wave was expected to reach the warning for Andaman and Nicobar Islands, coastal coastal regions at around 4:33 pm. Strong tremors regions of Tamil Nadu and Andhra Pradesh following were felt in Chennai and some other nearby areas. a massive 8.7 magnitude earthquake off Sumatra People in multi storeyed apartments and those coast. The Chennai port has been shut down due to working in high rise buildings rushed out to open tsunami alert. Government has evacuated some areas. Kolkata and its neighborhood were rocked by people from South Andaman as a precautionary tremors which were also felt in North 24 Parganas measure. National Disaster Response Force (NDRF) adjoining Kolkata and in the north Bengal town of teams have already been rushed to Andaman and Siliguri. A large number of people rushed out of Nicobal Islands, coastal regions of Andhra Pradesh, offices in Park street and downtown BBD Bagh area Tamil Nadu and Kerala. here as windows and doors rattled. According to reports, some buildings on Park Street developed Tsunami alert: Time of the likely impact in India cracks. Metro Rail services in the metropolis were is given in detail: suspended from 2:42 pm and passengers were asked - Tsunami likely to hit Little Andaman at 15:46 IST to vacate stations. Later India has scaled down the - Tsunami likely to hit Port Blair at 16:02 IST tsunami threat, saying there is no cause for panic - Tsunami likely to hit North Andaman at 16:22 IST along India's coastline and the government is ready to - Tsunami likely to hit Chennai at 16:57 IST deal with any eventuality. Govt of India has informed - Tsunami likely to hit Kakinada at 17:35 IST there was no cause for panic and the states of Andhra, - Tsunami likely to hit Trivandrum at 17:38 IST Orissa, Tamil Nadu and Andaman and Nicobar. - Tsunami likely to hit Mangalore at 19:06 IST Collectors in coastal districts have been asked to - Tsunami likely to hit Bombay at 21:38 IST broadcast on radio so that fishermen at sea can come - Tsunami likely to hit Gulf of Kutch at 22:04 IST back.\" Meanwhile, Tsunami warning has been scaled down for Indira Point and Katchal in Nicobar Islands Strong tremors were felt in different cities of the country on Wednesday afternoon. People in multi- 9. EARTHQUAKE HAZARD IN SRI LANKA storeyed apartments and those working in high rise Sri Lanka is not located near any of the 12 or 13 main buildings rushed out in panic after they felt the plate boundaries that are prone to earthquakes. tremors. Tremors were felt in Kolkata, Chennai, Indeed, it was squarely on a plate that extended from Bangalore, Patna, Kochi, Thiruvananthapuram, Australia to India. In the last decades, however this Cuttack, Bhubaneshwar and several other cities on plate is beginning to rotate on account of the eastern coast of India. Ministry of Earth Sciences accumulation of runoff from the Himalayas in the of India had issued a tsunami warning for Nicobar Bay of Bengal and other reasons leading to a fissure islands. The projections issued by the Indian Tsunami between the \"Australian\" and \"Indian\" plates. Some Early Warning Centre (ITEWC) showed the tidal scientists believe that this is leading to a new plate waves triggered by the quake hitting parts of Nicobar, boundary across the Southern Indian Ocean. This Komatra and Katchal minutes after it struck the boundary is still approximately a 1000 km from the region at 14:08 IST. south of Sri Lanka. Yet, these shifts have to be more carefully monitored and seismological studies need to be carefully followed. The Geological Survey and Mining Bureau is the government agency entrusted with seismological studies and it hosts a seismic station at Pallekelle in the Kandy District which is part of the global network of seismographs. Further research is needed to elucidate the consequences of compressions set up in the India plate and the impact of the recent earthquake on the regional hazards and also more precisely estimate the probabilities of the risk of earthquakes closer to Sri Lanka. Reduce Exposure to Reduce Risk 132

10. TSUNAMI HAZARD SRI LANKA Tsunamis are likely to have a more modest impact on While Sri Lanka is far away from the plate the coastal zone from Killinochchi to Puttalam as the boundaries, yet it is close enough to the highly active Tsunami wave shall not build into a wall due to the seismic zone near Sumatra and other regions to its presence of neighbouring India. South-East that earthquakes generated in these regions may lead to a Tsunami Hazard in Sri Lanka. 11. ARCHEOLOGICAL EVIDENCE Tsunamis are rarer in the Indian Ocean as the seismic FROM 2004 TSUNAMI activity is much less than in the Pacific. Tsunami's are extremely infrequent - the last major volcanic Tsunami gives a lead to submergence of explosion in the Indonesian island of Krakatau led to Mahabalipuram 1000 and 1500 years ago and for the a Tsunami in Sri Lanka in August of 1883. The wave people in eleven countries across the Indian Ocean heights that resulted however were much smaller than this event provides a unique template for future the 2004 Tsunami. While earthquakes could not be studies. Companion studies are also needed to predicted in advance, once the earthquake is detected identify palaeoseismic evidence in Aceh (Sumatra) it is possible to about an hours notice of a potential and the Andaman– Nicobar Islands to constrain the Tsunami for every 500 km distance from the causative earthquakes. Indian Ocean may have a epicenter. Such a system of warnings is in place geologic history of tsunami events similar in size and across the Pacific Ocean. Once the large amount of source area comparable to that of the 26 December pent-up energy in the compression zones of the plate 2004 event. Archeological Survey of India has boundaries have been released, it takes another observed from two trenches in the Mamallapuram buildup of energy for another event of similar (Mahabalipuram) beach, 55 km south of Chennai magnitude. Thus another Tsunami at the same (Madras) on the east coast of India, an area also location is unlikely in the short-term from the same affected by the 2004 tsunami. The possibility that the epicenter. In the future, Indian Ocean littoral regions sections in question may contain evidence to suggest should generate and pay attention to earthquake and two pre-2004 tsunami events occurring ~ 1000 years tsunami warnings. and ~ 1500 years ago respectively. Their ongoing work in the Andaman–Nicocbar region (near-source) indicates a tsunami occurrence along submergence of vegetation around 1000 years ago, which agrees with the date of the suspected penultimate tsunami event inferred at Mamallapuram. Further, analyses of the pre- and post-seismic GPS data from the Andaman– Nicobar region indicate 700–1000 years of recurrence for the giant tsunamigenic earthquakes at the 2004 source. Expanding the search for anomalous sand layers to other areas along the east coast, particularly closer to ancient cultural settlements should be a major source for future work. Such work would help to constrain ages of suspected tsunami sand sheets and also to characterize them in terms of flow hydraulics, sediment logy, mineralogy and lithology, so as to discriminate them from other types of coastal sedimentation, including storms or flood deposits. Reduce Exposure to Reduce Risk 133

12. WORST IS EXPECTED FROM LESSER dreadful and dangerous damaging event for INTENSITY EARTH QUKES earthquakes above 4 to 5 Richter scale as the The relocation of people, moving them to safer places population intensity varies from 200 persons to 1500 can be done within 30 min to 1 hr as the Tsunami persons in these areas. Even in 2004 the coastal areas waves will affect the Indian Nations coasts within 30 with mangroves and protective coastal civil works are min in worst case to 4 hours in nations above at 1000 least affected. The management of coastal zones with km from active Indian Ocean islands. But it is a vegetation and protective structures is a must. 13. RECOMMENDATIONS A collaborative research is required in Indian Ocean region nations to understand the causes of frequent earth quakes after 2004 Andaman Region of India, islands of Malaysia, Indonesia which are in Active zone. The south and east coasts of India and Sri Lanka meet erosive waves if earthquakes of higher intensity occur in this Indonesia – Andaman India region or in Indian Ocean ridge in the South which are facing these coasts. The computer modeling earth quakes and tsunamis after 2004 for Indian – Australian plate region over laying the USGS data will lead to give more information on movement of tectonic plates and epi centre and depth of earth quakes. Though we do not have clue to stop an earth quake or tsunami we can relocate the people to safer zones. We can adopt buildings and infrastructures which can with stand the shocks of earth quakes as it is being practiced in Japan and Pacific Coast of USA Reduce Exposure to Reduce Risk 134

REFERENCE Terry Machado T. Satyamurthy,P2. Aravazhi 3 and Thucydides: A History of the Peloponnesian War, 3.89.1-5 Manoj Jaiswal3 Jan 1, 2006,Evidence of ancient sea surges Smid, T. C.: \"'Tsunamis' in Greek Literature\", Greece & at the Mamallapuram coast of India and implications for Rome, 2nd Ser., Vol. 17, No. 1 (Apr., 1970), pp. 100-104 previous Indian Ocean tsunami events,, ias.ac.in, ,Current (102f.) Science Herodotus: http://www.perseus.tufts.edu/cgi- Shaik Mohamed Ghouse et al ,Asian Tsunami 2004 bin/ptext?lookup=Hdt.+8.129.1 \"The Histories\", 8.129 GTDM2, AIT Bangkok Bondevik, Stein; Dawson, Sue; Dawson, Alastair; Lohne, Tsunamis in South Asia,Source: Amateur Seismic Centre, Øystein (5 August 2003). India \"Record-breaking Height for 8000-Year-Old Tsunami in www.usgs.com. the North Atlantic\". EOS, Transactions of the American www.google.in, Geophysical Union 84 (31): 289, 293. www,Wikipedia.com, ANNEX Historic Tsunamis 6.2 1958: Lituya Bay, Alaska, USA 1 Before 1001 AD ( 6100 BC to 1000 AD) 9 Tsunamis 6.3 1960: Valdivia, Chile 1.1 ≈6100 BC: Norwegian Sea 6.4 1963: Vajont Dam, Monte Toc, Italy 1.2 ≈1600 BC: Santorini, Greece 6.5 1964: Niigata, Japan 1.3 426 BC: Malian Gulf, Greece 6.6 1964: Alaska, USA 1.4 373 BC: Helike, Greece 6.7 1976: Moro Gulf, Mindanao, Philippines 1.5 79 AD: Gulf of Naples, Italy 6.8 1979: Tumaco, Colombia 1.6 365 AD: Alexandria, Eastern Mediterranean 6.9 1980: Spirit Lake, Washington, USA 1.7 684 AD: Hakuho, Japan (白鳳大地震) 6.10 1983: Sea of Japan 6.11 1993: Okushiri, Hokkaido, Japan 1.8 869 AD: Sendai, Japan 6.12 1998: Papua New Guinea 1.9 887 AD: Ninna Nankai, Japan (仁和南海地震) 7 2000s 9 No of Tsunamis 2 1000–1700 700 years 8 Tsunamis 7.1 2004: Indian Ocean 2.1 1293: Kamakura, Japan (鎌倉大地震) 7.2 2006: South of Java Island 7.3 2006: Kuril Islands 2.2 1303: Eastern Mediterranean 7.4 2007: Solomon Islands 2.3 1361: Shōhei Nankai, Japan (正平南海地震) 7.5 2007: Niigata, Japan (新潟県中越沖地震) 2.4 1498: Meiō Nankai, Japan (明応地震) 2.5 1541: Nueva Cadiz, Venezuela 7.6 2009: Samoa 2.6 1605: Keichō Nankaido, Japan 7.7 2010: Chile 2.7 1607: Bristol Channel, Great Britain 7.8 2011: New Zealand 2.8 1698: Seikaido-Nankaido, Japan 7.9 2011: Pacific coast of Japan 3 1700s 100 Years 6 Tsunamis 3.1 1700: Vancouver Island, Canada 3.2 1707: Hōei, Japan (宝永大地震) 4.2 1854: Nankai, Tokai, and Kyushu Japan (安政南海地震) 3.3 1741: W. Hokkaido, Japan 3.4 1755: Lisbon, Portugal 3.5 1771: Yaeyama Islands, Okinawa, Japan 3.6 1792: Mount Unzen, Nagasaki Prefecture, Kyūshū, Japan (島原大変肥後迷惑) 4 1800s 7 Tsunamis 4.1 1833: Sumatra, Indonesia 4.2 1854: Nankai, Tokai, and Kyushu Japan 4.3 1855: Edo, Japan 4.5 1868: Arica, Chile 4.6 1883: Krakatoa, Sunda Strait, Indonesia 4.7 1896: Meiji Sanriku, Japan 5. 8 Tsunamis 5.1 1906: Tumaco-Esmeraldas, Colombia-Ecuador 5.2 1908: Messina, Italy 5.3 1923: Kanto, Japan 5.4 1929: Newfoundland 5.5 1933: Showa Sanriku, Japan 5.6 1944: Tonankai, Japan 5.7 1946: Nankaidō, Japan 5.8 1946: Aleutian Islands 6 1950–2000 12 Tusamais 6.1 1952: Severo-Kurilsk, Kuril Islands, USSR Reduce Exposure to Reduce Risk 135

PRELIMINARY STUDY ON” INTRAPLATE” EARTHQUAKES IN THE INDIAN OCEAN Jayathilaka, R.M.R.M., Jinadasa, S.U.P and Wijayadeva, D. A National Institute of Oceanography and Marine Sciences (NIOMS), National Aquatic Resources Research and Development Agency (NARA), Crow Island, Colombo 15, Sri Lanka, E-mail: [email protected] ABSTRACT The occurrence of seismicity at the surface of the globe largely along the well defined lines recognizes as mid- oceanic ridges and subduction zones. The equatorial region of the northern Indian Ocean defined Indo- Australian plate has been recognized as containing intense “intraplate” deformation. The identification of anomalous seismicity near the Ninetyeast and Chagos-Laccadive Ridges demonstrated the existence of the deformation. Twelve intraplate earthquakes were recorded with magnitude is greater than six in the Indo- Australian plate during the last 20 years. Ten of the events have depths at least 17 km below the seafloor, well into the upper mantle; two have depths as great as 39 km. There is a significant accumulation of earthquakes in eastern part of the Ninetyeast ridge compared to the western side. Possibility of these seismic events is development of stress due to two contrast plate movements. West of the Ninetyeast Ridge there is a continent- continent collision, and east of the ridge oceanic lithosphere subducts along the Sumatra trench. The Ninetyeast aseismic ridge therefore appears to be a mechanical border separating two distinct deformed areas. In addition there is an increasing trend of earthquake occurrences in this region. This trend is more exaggerated after the major earthquake in 2004. KEY WORDS: Epicenters, aseismic Reduce Exposure to Reduce Risk 136

Technical Session-6 [Hall A]: Landslide Spliting of a Typhoon Passing by Korean Peninsula and an Aftershock of Stormy 138 Wind in the Coastal Sea 144 Hyo Choi, Mi Sook Lee 150 150 A Conceptual Model for Landslide Prediction In Sri Lanka 151 L.D.C. S. Subhashini, H.L. Premaratne 152 152 Landslide Vulnerability Assessment along Four Lane Road Expansion in Between Visakhapatnam to Bhimunipatnam, Andhra Pradesh – A GIS Approach P.Jagadeeswara Rao, P.V.V.Satyanarayana Geomatics Based Integrated Slope and Land Use/Cover Modelling and Landslide Risks of Nilgiri Mountains, South India M. Muthukumar Post Collision Tectonics and the Phenomenon of Landslides in Peninsula SM. Ramasamy Improving Existing Landslide Hazard Zonation Map in KMC Area, Sri Lanka Oshadee Lasitha Potuhera, Vithanage Primali Anuruddhika Weerasinghe Landslide Hazard And Risk Analysis Of Kandy Municipal Council Area using Geoinformatics Techniques Sivapatham Thavavathani, N.D.K. Dayawansa, Ranjith Premalal De Silva Reduce Exposure to Reduce Risk 137

SPLITING OF A TYPHOON PASSING BY KOREAN PENINSULA AND AN AFTERSHOCK OF STORMY WIND NEAR THE EAST SEA OF KOREA Hyo Choi1 and Mi Sook Lee2 1Dept. of Atmospheric Environmental Sciences, Gangneung-Wonju National University, Gangneung 210-702, Korea, [email protected] 2Research Institute of East Sea Life Sciences, Gangneung-Wonju National University, Gangneung 210-702, Korea, [email protected] ABSTRACT: Splitting process of a typhoon Songda from September 5~8, 2004 which produced strong wind storm was investigated, using a 3D-Weather Research & Forecasting (WRF)-3.3 model. The typhoon Songda of 950hPa and maximum wind speed of 40m/s originally generated in the western Pacific Ocean had been gradually developed following a typhoon track. The typhoon track maintained toward northwest in the South China Sea and then it had its turning toward the East Sea (Japan Sea) passing near Korean strait between Busan, Korea and Fukuoka, Japan on September 7, 2004. Before the landfall of the typhoon, it began to split into two parts of air masses due to the shallow sea depth of the South Sea of Korea and surrounding, and it became a tropical depression. On September 7, as the tropical depression passing by the Korea Strait, it split into four parts as the first one (I) in the Yellow Sea, the second one (II) in the southern sea of the Korea, the third one (III) in the East Sea of Korea and the fourth one (IV) near Kyusu Island, Japan. Then, even though parts I and II were vanished, part III was reinforced in the East Sea, due to the deepening of surface atmospheric pressure for 24 hours of -15.42hPa in the East Sea by the terrain-induced channel effect of surrounding lands of Korea, Russia and Japan, showing very stormy wind of 14~22m/s. Simultaneously, negative geopotential height tendency for 24 hours at 500hPa level could induced the shrunken atmospheric layer between the level of 500 hPa to the ground surface, resulting in further the intensification of surface wind and maintaining the aftershock of stormy wind. KEY WORDS: Splitting of typhoon, Stormy wind, GOES-IR satellite, WRF-3.3 model, Geopotential tendency 1. INTRODUCTION toward the sea surface in the open ocean by the Tropical cyclones (in the northeast Asia), Cyclones changes in wind speed or direction in the wake of a (in India), hurricanes (in the North Atlantic and hurricane. Gilbes et al. (2001) and Babin et al. North Pacific) and Willy Willy (in Australia) (2002) presented that Hurricane-induced frequently produce severe disasters through strong phytoplankton blooms supplied by nutrient-laden winds, torrential precipitation with severe flooding waters through upwelling process of bottom colder and storm surges with up and down of sea levels sea waters toward the sea surface. (Anthes and Chang, 1978; Cheung and Chan, 2009). Elsner (203) and Jian and Wu (2008) indicated that As main purpose of this research is to investigate when a tropical cyclone approaches complex terrain the splitting of a typhoon and stormy wind by the during its landfall period, more complicated typhoon, using Weather Research and Forecasting asymmetric structures of winds cause the deflection Model (WRF)-version 3.3 for calculating wind, air of the typhoon track. temperature, relative humidity, atmospheric pressure Monaldo et al. (1997), Cione and Uhlhorn (2003) and gepotential tendency during the passage of and Knauss (2005) explained that cyclonic surface Typhoon Songda in the northeastern Asia-China, winds in a hurricane result in upwelling of bottom Korea, Japan and Russia from September 5 to 8, colder waters to the sea surface, because the surface 2004. Further consideration was given to the wind stress can cause surface divergence of sea surface water and upwelling of bottom colder water Reduce Exposure to Reduce Risk 138

analysis of Geostationary Operational typhoon moved northwestward continuously until Environmental Satellite (GOES)-Infrared image of September 4. cloud and surface weather maps for tracking of a typhoon passage. 2. NUMERICAL MODEL AND INPUT DATA (b) A 3D-WRF-3.3 meteorological model was Figure 1: Track of Typhoon-18, Songda on August adopted for the generation of meteorological 27 ~ September 8, 2004. elements - wind, air temperature, relative humidity, cloud, precipitation and 500 hPa height change for At 0000 UTC (0900LST), September 5, its strength 24 hours (i.e., geopotential tendency (/t; m/day) began to weak and at 0000 UTC, September 7, as in the study area (Hong and Lim, 2006). In the Typhoon Songda, it changed its direction to numerical simulation, one way, triple nesting northeastward across the East China Sea and made process from a coarse-mesh domain to a fine-mesh landfall on the northwestern coast of Kyushu Island, domain was performed using a horizontal grid Japan. Cheju Island, Korea to the left of Kyushu spacing of 27 km covering a 91 x 91 grid square in Island was in the strong impacted by the typhoon. the coarse mesh domain. The second and third At 1800 UTC, September 7, its strength began to domains also consist of the same grid square of 91 x slowly decrease and was downgraded to a tropical 91 with 9 km and 3 km horizontal grid intervals. storm and at 0600 UTC, September 8, when it became an extra-tropical cyclone (i.e., a low National Centers for Environmental pressure) (Fig. 6). Prediction/National Center for Atmospheric Research (NCEP/NCAR) reanalysis-Final Analyses 3.2 Surface wind fields and atmospheric pressure (FNL) 1.00 x 1.00 resolution data were used as The typhoon Songda of 950hPa and maximum wind meteorological input data to the model and were speed of 40m/s for 10 minutes average was vertically interpolated onto 36 levels with originally generated in the western Pacific Ocean sequentially larger intervals increasing with height and then it had been gradually developed following from the surface to the upper boundary level of 100 a typhoon track. The typhoon track maintained hPa (Choi et al., 2009; Choi, 2010). Hourly archived toward northwest in the South China Sea until at data set of wind, air temperature, relative humidity, 0900 LST, September 5 with asymmetric wind cloud and geopotential tendency by Cheju structures and deepening of – 16hPa for a day (Figs. Meteorological Office were used for the verification 2, 3a and 5a). of numerical results of meteorological elements. At 0900LST, September 6, it had its direction 3. RESULTS AND DISCUSSION northward and atmospheric pressure change for a 3.1 Typhoon track day reduced to – 12.86hPa, resulting in surface wind At 1100 UTC, August 26, 2004, Joint Typhoon to be slightly weaker than one day before (Figs. 3b Warning Centre, USA (JTWC) reported a new area and 5b). At 2100LST, September 6, before the of convection which had developed and persisted landfall of the typhoon in the Korean peninsula, the approximately 210 nautical miles northeast of typhoon began to split into two parts of air masses Kwajalein (Fig. 1). JTWC at 1200 UTC, August 27 due to the shallow sea depth of the Yellow Sea and gave the first warning on Tropical Depression-22W the South Sea of Korea and surrounding lands - the with its center at 270 nautical miles east of Eniwetak Atoll in the Pacific Ocean. At 0000 UTC, August 28, the tropical depression with its surface wind speed of 35 kts was assigned the name as Songda as a tropical storm. At 1800 UTC, August 30, it became a typhoon with maximum surface wind speed of 95 kts and situated about 17 nautical miles north-northeast of Agrigan Island in the Northern Mariana Islands. Then the Reduce Exposure to Reduce Risk 139