Recovery, Rehabilitation and Development Project for Tsunami Affected Area of Southern Region in the Democratic Socialist Republic of Sri Lanka

Transcription

1 Questionnaire Survey (1) Necessity of Questionnaire Survey As described above, available hazard maps lack sufficient accuracy to estimate the actual inundation area. Furthermore, existing marks that indicate the height of the flood waters have disappeared as time passes by. As such, a questionnaire survey was conducted to: (i) collect responses from a sufficient pool of respondents in the affected areas and make record height measurements left behind by the waters from the 2004 tsunami and 2003 flood; and (ii) investigate the actual results of inundation, damage, and evacuation during the noted tsunami and flood. This survey was subcontracted to the ENGINEERING & LABORATORY SERVICES (PVT) LTD from the middle of October 2005 to the middle of December (2) Methodology 1) Survey Area The survey area covered the Matara Division, which was affected by the 2004 tsunami and 2003 flood as shown in Figure The number of samples for the questionnaire was to be at least 500 samples from each disaster. 2) Questionnaire Sheet Questionnaire sheets were prepared for the 2004 tsunami and for 2003 flood, respectively. The items in the questionnaire survey are shown in Table 10.8 and the survey sheet is attached in Annex Table 10.8 Items of Questionnaire Survey Item Questionnaire a) General information Name Sex Disaster experience, etc. b) Result of Inundation Location of high water marks Distance form coastline/river Inundation depth and inundation period Current direction c) Result of Number of affected people, casualties, and fatalities Tsunami/Flood Damage Damage to houses, movable assets, crops, livestock, and d) Warning and Evacuation system infrastructure Tsunami/flood forecasting and warning system Publication of hazard areas and evacuation areas Results of evacuation (place, route, period, method, etc.) e) Hazard Map Awareness of hazard maps Utilization of hazard maps Source: Project Team 10-47

2 3) Methodology To improve the reliability of the survey, the questionnaire survey was conducted by adopting the following procedures: i) Survey respondents were selected from residents actually affected by the disasters; ii) The sheet was completed via face-to-face interviews; iii) Inundation depth was measured from the ground and floor levels; iv) High water mark location was confirmed with portable GPS; v) Survey responses were compared with responses from neighbors to determine if considerable disparities were present in the responses; vi) The survey sheet, which showed an outline of the survey results, was pasted on the mark indicating the highest water level; vii) Photographs were taken of all height measurements (corresponding to their survey sheet) and compiled in photo books; and viii) Some height measurements of floodwaters were surveyed by leveling, so as to convert the inundation depth to the elevation data. (3) Results of Questionnaire Survey for 2004 Tsunami 1) Characteristics of Respondents The total number of respondents for the tsunami disaster was 586, with 44% being male and 56% being female. Approximately 14% of 4,330 houses damaged by tsunami in Matara Division were queried. Out of all respondents, 79% experienced the tsunami disaster in % of respondents lived within 100 m of the coast, 67.3% lived between 100 m to 500 m away, and 10.7% lived between 500 m to 1,000 away, respectively

3 Distance from coastline (m) <1200 <1000 <0800 <0600 <0400 <0200 < Source: Project Team No. of houses Figure Accumulated Number of Houses vs. Distance from the Coast 2) Results of Inundation by the 2004 Tsunami Inundation conditions reported by respondents are summarized in Figure The maximum inundation depth was 410 cm from the ground at 40 m from the coastline. In regards to the inundation depth from the ground, 54% of houses were more than 2.0 m, 31% were between 1.0 m and 2.0 m, and 32% were between 0.5 m and 1.0 m, respectively. <0.5m, 53, 9% 2.00m <, 316, 54% 1.0m-2.0 m, 30, 5% m, 187, 32% <0.5m m 1.0m-2.0 m 2.00m < Source: Project Team Figure Inundation Depth above the Ground during the 2004 Tsunami 10-49

4 Regarding the inundation period, the average time experienced by the respondents was estimated at 5.9 hours. Some 43.2% of houses were inundated from longer than four hours. Regarding the inundation speed, 75% of respondents felt it was very fast (within a few seconds), while 17% felt it was fast (within a minute). Regarding the direction of the tsunami wave, it was basically from south to north, and came through the coastal line. It was very rare that the direction changed due to the effect of drain canals and small ponds. Also, respondents who lived close to the Nilwala River experienced the tsunami from the river. 3) Results of Damage from 2004 Tsunami Since the respondents for the survey were selected from those actually affected by disasters and living in their original locations, the following data may be less accurate: 4% of houses were damaged completely and 19% were partially damaged; 80% of respondents lost their movable property, 39% completely lost their crops, while 6% completely lost their livestock; From the survey, 60 people from the homes of respondents died due to tsunami; and 4% of respondents noted road damage. 4) Evacuation and Warning System 89% of respondents were unaware of tsunami-prone areas and evacuation locations. Before the 2004 tsunami, people had never experienced a tsunami, so there was no tsunami forecasting and warning system. Regarding conditions for dwelling, 150 respondents thought they did not live in a dangerous place, 286 respondents thought they lived in a dangerous place, but the danger was not imminent, and 150 respondents thought they lived in a very dangerous place, respectively. Regarding the first evacuation, the average distance from the house was 3,132 m; 47.1% of respondents did not evacuate from their house during the 2004 tsunami. Evacuation locations were schools, temples, two-story buildings, and elevated areas. Regarding the final evacuation, 19% of respondents moved to a safer location and 81% returned to their houses after the tsunami water subsided. Evacuation locations included relatives homes, schools, and temples

5 Recovery, Rehabilitation and Development Project Source: Project Team Figure Results of First Evacuation during the 2004 Tsunami 5) Tsunami Hazard Map No one knew about the hazard map. Expected information to be displayed on a hazard map included evacuation centers, evacuation routes, hazard areas on evacuation routes, medical care facilities, and special signs along evacuation routes, etc. Many respondents expected a hazard map to be distributed to each house. Some expected publishing to be done by public servants, with maps displayed at public places. Also, there was an idea that media publicity would facilitate understanding of the findings of the survey. (4) Results of Questionnaire Survey for the Flood in ) Characteristics of Respondents The total number of respondents for the flood disaster survey was 503, with 44% being male and 56% being female. Out of the respondents, 94% experienced the flood disaster in Those who experienced the flood and responded to the questionnaire had lived within 6,500 m of the Nilwala River. Those who experienced floods from the branch rivers (Tributaries) had lived within 2,500 m of the tributary

6 <7000 Distance to Nilwala river (m) <6000 <5000 <4000 <3000 <2000 <1000 < Source: Project Team No. of respondents Figure Number of Houses vs. Distance from the Nilwala River 2) Result of Inundation by the Flood in 2003 Inundation conditions related to the number of houses and distance from the Nilwala River are summarized in Figure The maximum inundation depth was 386 cm from the ground, at a distance of 200 m from the Nilwala River. Regarding the inundation depth from the ground, 18% of houses were inundated less than 0.5 m, 24% were between 0.5 m and 1.0 m, 51% were between 1.0 m and 2.0 m, and 7% were more than 2.0 m, respectively. 457 houses were inundated above the floor level

7 >2.0m, 6.96, 7% m, 17.69, 18% m 0.5m-1.0m 1.0m-2.0m >2.0m 1.0m-2.0m, 51.29, 51% Source: Project Team 0.5m-1.0m, 24.06, 24% Figure Results of Inundation Depth by 2003 Flood Regarding the inundation period, 60% of houses were inundated for more than six days, while 22% were inundated at least 10 days. Regarding the inundation speed, 26% of respondents felt it was very fast (within a few minutes), 28% felt it was fast (within a hour), and 36% felt it was slow (within a few hours), respectively. Regarding the direction of the flood, it was basically from the Nilwala River or tributary. At certain locations, some respondents had experienced the flood from both the Nilwala River and tributary. 3) Result of Damage by the Flood in % of houses were completely damaged, 48% were partially damaged, and 43% were not damaged, respectively

8 Inundation levels varied from 40 cm below the floor level to 380 m above it. 72% of respondents lost their movable items and 54% lost their crops. Only 4%, however, lost their livestock. According to the survey, only one person died and 503 were injured from the homes of respondents, respectively. 2% of respondents mentioned damage to the roads. 4) Evacuation and Warning System Only 71 respondents out of 503 (14%) were aware of flood disaster warning. Flood warnings were generally disseminated by the police. 12% of respondents were aware of flood-prone areas and official evacuation locations. Most evacuation locations during the flood were schools and temples. Regarding dwelling conditions, 42% of respondents thought they lived in a dangerous place, 51% thought they lived in a dangerous place, but the danger was not imminent, and 7% thought they lived in a very dangerous place, respectively. Regarding the first evacuation, 390 of the respondents (78%) evacuated to schools and temples, while 113 (22%) did not evacuate and stayed in their homes. Respondents stayed an average of 10 days at their first evacuation locations. Regarding the final evacuation, only 5% of respondents moved to a safer location, while 95% returned to their house after floodwaters subsided. Evacuation locations were mainly the homes of relatives. Source: Project Team Figure Results of First Evacuation during the 2003 Flood 5) Flood Hazard Map Only three out of 503 respondents knew of the hazard map. Many respondents expected that hazard map would be distributed to each house. Some expected publishing by civil servants, with maps being displayed at public places

9 Tsunami/Flood Disaster Map (1) Basic Specifications for the Disaster Map Considering the improvements to the hazard map and ease of utilization, basic specifications for the disaster map were defined as below: Type: Colored Map; Size: A2 size; Area: Matara Division; Scale: 1:20,000; Topographic Information: Referring to the map data of the UDA GIS Division Survey Department; and Disaster Information: Results of the questionnaire survey for the 2004 tsunami and 2003 flood. (2) Basic Information Basic information was plotted on the disaster maps to utilize: (i) topographic information that was collected from the Survey Department and UDA GIS Division; and (ii) disaster information based on the results of questionnaire surveys as listed below. 1) Topographic Information i) Topographic data (based on 1:10,000 scale map) ii) Contour line (from 0 m to 5 m MSL, with intervals of 1 m) iii) Niliwara River, tributaries, and drainage canals iv) Main roads v) Divisional boundaries vi) Location of houses vii) High elevation areas (higher than 5 m MSL) 2) Disaster Information i) Boundaries of inundation area ii) Locations of inundation area (grid of 50 m x 50 m) or contour data iii) Inundation depth (0.0 m to 0.5 m - green, 0.5 m to 1.0 m - yellow, 1.0 m to 2.0 m - cyan, more than 2.0 m - blue) iv) Numerical data of inundation depth v) Flow direction vi) Evacuation direction vii) Location of present evacuation shelters 10-55

10 (3) Main Findings Based on the above works, the 2004 tsunami disaster map and 2003 flood disaster map were prepared as shown in Figures and 30. Main findings are noted below. 1) Tsunami Disaster Map Detailed and reliable flood-affected areas were defined. Precise tsunami movement was indicated with a scale of 50 m as noted below: i) The 2004 tsunami penetrated up to 4 km upstream from the river mouth, along the Nilwala River; ii) Around the U-shape bay in the Pholhena area, the inundation height was relatively high; and iii) Tsunami movement was affected by small canals. The area that was not indicated in the available hazard map was newly identified as a disaster area. The tsunami inundation area was spread approximately 800 m from the coastline. Actual flow direction indicated on the map shows a high correlation between the flow direction and the direction of inundation. Actual evacuation direction indicated on the map shows that people evacuated to high elevation or inland areas. 2) Flood Disaster Map Detailed and reliable flood-affected area were defined. Precise flood movement was indicated with a scale of 50 m as noted below: i) The 2003 flood spread from the northwestern area along the main body of the Nilwala River; ii) The inundation depth on the right side of the river was especially high; iii) The area along some tributaries was seriously affected by inland water; and iv) The lower river basin was not flooded. Actual flow direction indicated on the map shows a highly correlation between the flow direction and the direction of inundation. Actual evacuation direction indicated on the map shows that people evacuated to high elevation or low inundated areas

11 Source: Project Team Inundation Depth 0 0.5m m m > 2.0m Figure Tsunami Disaster Map (Grid Data) 10-57

12 Source: Project Team Inundation Depth 0 0.5m m m > 2.0m Evacuation Direction Tsunami Direction Existing Refugee Shelter Elevated Area Figure Tsunami Disaster Map (Contour Data) 10-58

13 Source: Project Team Figure Flood Disaster Map (Grid Data) Inundation Depth 0 0.5m m m > 2.0m 10-59

14 Source: Project Team Figure Flood Disaster Map (Contour Data) Evacuation Direction Inundation Depth Flood Direction Evacuation Location Elevated Area 0 0.5m m m > 2.0m 10-60

15 Tsunami Simulation (1) Outline Tsunami simulation was carried out to estimate accurate data with respect to the area of inundation, the inundation depth, and the arrival time to Matara City -this in order to make a preliminary tsunami hazard map, in case of a tsunami on a similar scale to the 2004 one. The tsunami simulation process is detailed as shown in Figure As shown in Figure 10.34, the simulation area includes the seismic center, the wave-propagating sea, and Matara city. The grid size for computation varies and lessens toward Matara city (with a 50 m grid in Matara city). Setup of Analysis Conditions Development of Geographical Configuration 1) Grid data in sea area 2) Grid data in land area Setup of Tsunami Wave Source Model 1) Collection of tsunami wave source model 2) Tsunami simulation trial by each model 3) Calibration by actual tsunami height 4) Determination of the model Calculation of Tsunami Approach 1) Tsunami simulation in sea areas 2) Tsunami run-up distance in land areas Source: Project Team Figure Process of Tsunami Simulation 10-61

16 Source: Project Team Figure Area for Tsunami Simulation (2) Calculation Condition 1) Calculation Method The behavior of tsunamis is known to be dominated by the law of conservation of mass and the law of conservation of momentum, and is described by the equation of continuity and equation of motion, respectively. In this model, the displacement of sea levels caused by the movement of a tectonic plate is given as the initial condition. Solving the equations with a time parameter, tsunami behavior can be expressed numerically. The numerical analysis method to estimate the inundation situation in Mantra City is summarized in Table Table 10.9 Numerical Analysis Method for Tsunami Simulation Area Applied Method From seismic center to offshore Liner long wave equations From offshore to coastal line Non-liner long wave equations (shallow water theory) On-shore Iwasaki and Shinno Theory (1979) Source: Project Team 10-62

17 2) Data of Sea and Land Geographic Configurations The grid data for the geographical configuration are based on the data collected and surveyed by the JICA Project Team. As stated before, the minimum size of the grid is chosen to be 50 m. In order to carry out numerical analysis, continuous geographical configuration data is translated to an individual elevation value in each grid. When determining the size of the grid, guidelines advise that it should be at least 20 m within a wave length. As a tsunami wave approaches shore, the wave length becomes shorter, the wave height becomes higher, and the flow speed becomes faster. Thus, the size of a grid in the open sea including the seismic center is taken to be larger, while that nearer the shore to be smaller, making computation speed faster without sacrificing reliability of the analysis. At the smallest grid area (50 m spacing), the grid data includes not only elevation values, but also land utilization information (roughness coefficient, and with/without building), so that the simulation of the tsunami run-up on the land area can be more accurate. Table10.10 shows the data used for numerical analysis, while Figure shows a 50 m grid and its elevation values on the land area. Table Data Used for Numerical Analysis Item Name Issue Water Depth ETOPO2 2 Mesh NOAA/NGDC CHART WELIGAMA-MATARA S=1/25,000 National Aquatic Resources and Development Elevation on Land Topographic Information (DWG) Survey Department Source: Project Team 10-63

18 Source: Project Team Figure m Grid and Elevation on Land Area 10-64

19 3) Consideration of Obstacles Such As Structures It is well known that the existence of seawalls along the coastline, relatively large reinforcement concrete (RC) structures, greatly impacts the extent of a tsunami s reach inland. Thus, these obstacles are treated as boundary conditions in the numerical analysis. The locations, figures, and heights of such obstacles are surveyed by the JICA Project Team, and the information is simplified and assigned to each grid (as shown in Figure 10.36). 4) Initial Conditions The displacement of the sea level caused by the movement of tectonic plates is calculated by the earthquake fault model, and its value is used as the initial condition of the tsunami inundation analysis. The earthquake fault model explains that a sudden slip of the adjacent strata is caused by accumulated energies due to deformation at the crusts. If such a slip occurs at the seabed strata, the seabed will rise up or subduct, resulting in a tsunami (refer to Figure 10.37)

20 Source: Project Team Figure Location of Boundary Conditions of On-shore Structures for Numerical Analysis

21 Source: Nature Vol433 p352 Figure Basic Concept of Seismic Occurrences and Seismic Foul Model 5) Tsunami Source Model a Selection Method of Fault Model Fault Model for the Tsunami simulation was chosen from issued models for Sumatran Earthquake Tsunami. Flowchart of Fault Modeling Approach is shown in Figure 10.38, and characteristic of the issued models is shown in Table

22 Issued Fault Models Checking with 4 tidal level records and choice the model Adjustment of Parameter Fault Model Selection Figure Choice Flowchart of Fault Model Table Characteristic of issued Models Organization Characteristics of fault Modeling approach 1.Tohoku Univ. Site investigation result at Kirinda in Sri Lanka 2.Kyoto Univ. Seismographic date by Tokyo Univ. Seismic research laboratory 3.Koshimura Tohoku Univ. Wave height of sea leveling by artificial satellite over the Indian Ocean 4.JAMSTEC ditto 5.AIST ditto 6.Akita Univ. Compression with site investigation result at the South West area in Thailand According to the Coastal Development Institute of Technology (April 2004) the fault models are categorized in Table Because there were no issued models for checking with the site investigation result in the Matara City, the model for this project was basically chosen by the data of site investigation result in Sri Lanka and water levels recorded by satellite remote sensing. Table Category of Fault Mode Category of Fault Model Approach 1 Modeling by Seismographic and Geodesic data 2 Optimum fault model by re-create tsunami simulation: 3 Modified model by site investigation based on optimum fault model by re-create tsunami simulation accuracy Low High b Selection of Fault Model Re-create simulations by six issued fault models were carried out and were checked with tidal level records at 4 places (Colombo in Sri Lanka and Male, Hanimaadhoo and Gan in Maldives). The parameters of the models and the location maps of fault are shown in Annex10.3. From the result of comparison of re-create simulation with tidal level record as shown in Annex10.3, the fault model by Tohoku University was most similar to the first and the second waves. The validity of the model was identified with site investigation result in Sri Lanka. So the fault model by Tohoku University was selected for this project

23 The selected fault model will be used for not only the Matara City but also the other cities in Sri Lanka and Maldives. c Parameter Modification of Fault Model The parameters of Fault Model by Tohoku University were modified to be consistent with inundation conditions at Matara city. Since the wave height of the Tsunami simulation result by Tohoku University tended to be a little smaller than the tidal level record at Colombo, the parameter of slipping volumes was modified. Figure show inundation area at Matara city through 5 cases (11m to 18m) of slipping volume. The simulation result of 14 m as a slipping volume matched with site investigation result. Table shows the modification result of parameter fault model. (a) 11m (b) 13m (b) 14m (c) 15m (d) 18m Figure Tsunami inundation area by fault slipping volumes 10-69

24 Fault plate Upper edge depth Table Modification Result of Parameter Fault Model slipping volume direction angle of inclination slipping inclination Length width Base point Crashing time HH(km) D(m) TH( ) DL( ) RD( ) L(km) W(km) Y0 X0 Fault Fault Fault d Comparison with Tidemark The maximum tsunami water levels at various points within two hours from the time of the earthquake were calculated by an applied numerical model, and summarized as a univariate data distribution. Such calculated values are then compared with the tidemarks (highest inundation levels) caused by tsunami inundation. To evaluate the fit of the model, the calibration indices, Κ and κ, defined by Aida s series of study are used. Geometrical Average 1 Κ logκ = n i= Ri log 1 H i n Logarithm Standard Deviation R i H n i Observation Inundation Level Estimated Inundation Level Number of Observations n 1 R i κ logκ = log n i= 1 H i 2 2 (logκ) Κ Index Indicating Actual Tsunami Inundation Height Is Κ -fold of Estimation κ Logarithm Standard Deviation of R / H i i 1 2 Κ is the index meaning that K times the calculated height is, on average, equal to the observed height. Thus, if K is closer to 1.0, the numerical model can be considered to fit better to the actual data. κ is the index indicating the variance of the ratio ( R / H ), which is comparable to the standard deviation in the normal distribution. Also, it is construed that the majority of the values of R / H at the observed points are less than Κ x κ. According to Souda i i (1986), apart from the value of Κ, the fit of the numerical model can be evaluated to be fairly accurate if κ 1. 2, and very accurate if κ i i 10-70

25 Table Comparison Simulation Result with Site Investigation Result No (i) Investigation Simulation log(r/h) (log(r/h)) 2 Result (R) Result (H) Σ logk K logκ κ

26 Source: JICA Project Team, Red: Investigation Result, Blue: Simulation Result Figure Comparison investigation result with simulation result 10-72

27 e Comparison with Recorded Tide Levels Since there is not any available recorded tide data of the Tsunami wave in Matara City is available, the calculated water levels within 2 hours from the time of earthquake occurrence are compared with the actual water levels based on the questionnaire survey data. Under this section, the calculated water levels varied with time are analyzed and evaluated by the abovementioned method. Figure shows that Arrival time at Matara city is approximately 2 hours from the time of earthquake occurrence, the first wave of the tsunami was highest and the following waves came at intervals of 20 or 30 minutes. After 20 minutes, the first wave came from the coastal line and arrived at the farthest inundation area. Table shows the tsunami arrival time at Matara city and additional cities in the Southern area of Sri Lanka. Since the arrival time was estimated by questionnaire and/or site survey, it seems that it was inundation time. Since the simulation results are the same as the results from the above data, the simulation was determined to be valid. Figure Tsunami Time Series Table Tsunami Approach Time to the South Area in Sri Lanka Earthquake time (Local Time) Tsunami Arrival time (Local Time) Arrival hour Yala 6: 58 9:10 1) 2h 12m Hambantota 9:22 2) 2h 24m Matara 9:20 1) 2h 22m Galle 9:20 1) 2h 22m Source 1) Kiobuchi, JSCE Coastal Engineering Paper (No. 52) ) NIshihata, JSCE Coastal Engineering Paper (No. 52)

28 (3) Result of Tsunami Simulation The results were summarized by inundation area and depth at each area. Table Inundation Area and Depth GN Inundation Area (%) (Rate for Total Area) Inundation Depth (m) Madiha West Madiha East Walgama Central Walgama South Welegoda West Welegoda East Polhena Pamburana Noope Thotamuna Kadaweediya West Kadaweediya South Fort Uyanwatta North Uyanwatta Kotuwegoda North Kotuwegoda South Weragampita Navimana South Weraduwa Eliyakanda North Eliyakanda South Meddawatta South (4) Conclusion Simulated and surveyed results of inundation depth generally matched (refer to P Comparison with Tidemark). The arrival time and inundation area of Tsunami wave generally matched, too. Accordingly, it was confirmed that the simulation results were suitable. Therefore, these results are satisfactory for planning the disaster prevention and mitigation of Matara city

29 Figure Results of Tsunami Simulation Preliminary Tsunami Hazard Map (1) Overview A preliminary tsunami hazard map for the tsunami of 2004 was prepared for basic information on planning of disaster prevention and mitigation, as well as enhancement of disaster awareness for both residents and local authorities. This preliminary tsunami hazard map does not include some information that needs to be authorized by local agencies as shown in Table As described in Section 10.3, local authorities are expected to use the map for various measures as soon as possible

30 Table Comparison of Tsunami Hazard Map and Preliminary Tsunami Hazard Map Item HM Remarks Classification Item Residents Author-i PHM ties External Force Information of Inundation Information about External Force Unused Information about External Force Expected Inundation Area Area for Refuge Maximum Inundation Depth Forecasted Arrival Time Starting Point of Inundation Current Speed Shelter In case of a tsunami equivalent to the 2004 tsunami Have to be careful to fix disaster image for inhabitants To indicate a draft buffer zone that is decided by elevation Information of Disaster Prevention Evacuation Route Disaster Prevention Shelter Base Inundation Record On Disaster Map Others Area Condition Shelter Condition Ground Elevation Necessary Rescue Institutions Evacuation Guidelines, etc. * HM Tsunami Hazard Map * PHM Preliminary Tsunami Hazard Map Source: Project Team. (2) Methodology 1) Necessity of Input for Preparation of Preliminary Tsunami Hazard Map The necessities of input for the preparation of a preliminary tsunami hazard map are extracted from the results of the tsunami simulation, and are indicated on a base map made from satellite photographs of Matara city

31 Source: GIS Division, UDA Figure Satellite Photograph of Matara City 2) Presumption of Damage Displaying the number of victims and damage to the houses are effective for gathering basic information on planning disaster prevention and mitigation, as well as enhancement of disaster awareness for both residents and local authorities. The number of victims and damage to the houses are commonly presumed from structure (Brick, Wooden or RC Structure) and people in houses based on the disaster caused by the past Tsunami in the area. Therefore, the number of victims calculated from the damaged houses was presumed from the disaster record of the 2004-Tsunami. The Presumption methodology is shown in Figure Relation with Tsunami current and damage 1) Damaged area is presumed from site survey Tsunami current are calculated by Tsunami simulation by equivalent 2004-Tsunami Current by equivalent 2004-Tsunami 1) Current are presumed from Tsunami simulation Presumption of Damage Damage of houses are presumed roughly Source: Project Team Figure Procedure of Presumption of Damage by Tsunami 10-77

32 Recovery, Rehabilitation and Development Project (3) Preliminary Tsunami Hazard Map Preliminary Tsunami Hazard Map is shown in Figure Basic information was mentioned for planning of disaster prevention and mitigation in Matara city (refer to Table 10.15). The information is the following: - Information about External Force; - Expect Inundation Area; - Maximum Inundation Depth; - Forecast Arrival Time; and - Current Speed. Animation of simulated Wave Spreading from the seismic center and Inundation Condition are made in order to increase the understanding of disaster awareness as shown in Figure10.45 and Figure Figure Animation of simulated Wave Spread from the seismic center Figure Animation of simulated Inundation Condition 10-78

33 Source: Project Team Figure Preliminary Tsunami Hazard Map 10-79

34 10.6 Implementation of Awareness Activities JICA-Net Seminar (1) Background Tsunami-affected people and officers in the local public administration in Matara District are well aware that Japan is a very experienced nation in terms of disasters such as earthquakes, typhoons, tsunamis, and flooding. Local people are extremely interested in mitigating damage from the next possible disaster. The JICA Project Team has continually been asked questions such as, How are Japanese people preparing for disaster? In order to satisfy this interest in Japanese disaster management, the JICA Project Team conducted seminars and transferred lessons learned from Japanese disasters to residents and officers of the Matara District. (2) JICA Net Seminars JICA-Net is an interactive Internet-television-computer system that connects studios in Colombo with JICA Headquarters in Tokyo, Japan. Seminar participants in the studios, who are thousands of kilometers apart, can attend a seminar together in real-time. In the seminars, resource persons in Japan come to the Tokyo studio and make presentations on various topics. Participants in Sri Lanka assemble in a studio in the JICA Colombo Office and are free to ask questions and discuss topics from their own viewpoints so that they can obtain a good understanding of the topic in the context of Sri Lanka. The JICA Project Team carefully prepared and held two seminars on disaster management and rehabilitation in order to facilitate the transfer of knowledge. This included: Community Initiatives in Disaster Management and Rehabilitation, and Role of the Local Authorities in Disaster Management. Detailed program information is discussed below. (3) Seminar 1 - Community Initiative in Disaster Management and Rehabilitation This seminar was held on the 13 th December at the JICA Sri Lanka Office as one of the activities of the JICA Project. There were 30 participants from the pilot project organizations, such as refugee camp societies, fishery corporative societies, and small 10-80

35 industry associations. For seminar participants, it required a full-day round trip from Matara to Colombo and back, spending eight hours in the bus. Despite this, the seminar was greatly appreciated. The program was as follows. Table JICA-Net Program for Community Initiative on Disaster Management Module Facilitator or lecturer 1 Introduction to JICA-Net Seminar Katsura MIZUNO, PADECO, Tokyo 2 1st lecture Okushiri Tsunami and disaster management of the town council Mr. Kakemi NAGASAKI, Officer in charge of tourism and planning, Okushiri Town Council 3 2 nd lecture What we learned from the Hanshin-Awaji Earthquake and the importance of community initiative in disaster management Mr. Masakiyo MURAI, Director, CODE 4 Question & Answer Mr. Dahanayake 5 Group discussion on what we can do in Matara, Coordinators of the JICA Project Team and presentation from three groups 6 Mechanisms of a tsunami Mr. Ryo MATSUMARU, JICA expert for disaster management Source: Project Team. Based on the evaluation sheets collected after the session, it appeared that the participants recognized the importance of this kind of disaster management seminar for communities, and were satisfied with information and lessons offered by the seminar. (4) Seminar 2 - Disaster Management for Local Public Administrations This seminar was held on the 14 th December at JICA Sri Lanka Office as one of the activities of the JICA Project. In total, 24 participants came (mainly from the Matara area, but also some from Colombo). Participating organizations included the police, army, divisional secretariat, Urban Development Authority, Fishery Department, Corporative Department, and Matara Municipal Council, etc. Since co-assistance among community people produces significant impacts in terms of disaster management and rehabilitation works, it is very important for officers of local public administrations to be aware of this and to know how to draw such effects out of communities. This seminar explained Japanese disaster experiences and systems for disaster management (particularly how to enhance community participation). The program was as follows

36 Table JICA-Net Program for Public Administration Module Facilitator or lecturer 1 Introduction to JICA-Net Seminar Katsura MIZUNO, PADECO, Tokyo 2 1 st lecture Okushiri Tsunami and disaster management of the town council Mr. Kakemi NAGASAKI, Officer in charge of tourism and planning, Okushiri Town Council 3 2 nd lecture Hanshin-Awaji Earthquake and community involvement in disaster management Mr. Isamu OKADA, Director, Suma Fire Station, Kobe Municipal Council 4 Question & Answer Mr. Dahanayake 5 Group discussion on what we can do in Coordinators of the JICA Project Team Matara, and presentation from three groups 6 Mechanisms of a tsunami Mr. Ryo MATSUMARU, JICA expert for disaster management Source: Project Team. The public administration participants understood the necessity to involve local communities in counter-disaster administration. (5) Feedback from Seminar Participants Since the participants are all interested in disaster management, the vast majority of the participants appreciated the seminars. Requests and suggestions mentioned in the evaluation sheets included: - A half day was insufficient for this kind of seminar; - Many more people should attend the seminar; - This seminar should be held in every region of the country; - This seminar should be aired as a television program; and - Besides disaster preparedness, subjects such as rehabilitation should also be discussed. Participants mentioned in the evaluation sheets that they have: - Shared Japanese disaster experience; - Learned how to protect themselves from tsunamis; - Received useful knowledge and information; and - Realized the importance of co-assistance. As a part of the seminar program, participants held group discussions on what they can do for disaster management in Matara. Three groups were formulated for discussion, with each group proposing dozens of ideas for effective preparation against disaster. They are motivated to have the same kind of discussions in their communities and offices. Some ideas derived from the group discussions are the following: 10-82

37 - Community and government should conduct annual evacuation drills like Japan; - Every family should check evacuation routes in the case of an emergency; - We (participants) will organize disaster committee in every community under Grama Niladari (community chief); - School children should learn disaster management to protect themselves; - We will discuss the warning system of our village; - Our community should prepare a storehouses to keep tools, equipment, and food to survive disaster; - We should edit a manual of lessons learned from our disaster experience and keep it for generations; - Police officers need practical training for disaster management; - Community should cooperate on protecting children, the elders, and disabled in a disaster; and - Every public agency should prepare for next tsunami in order not to be affectedand still maintain its ability to save citizens Seminars on Disaster Management The JICA Project Team conducted seminars on disaster management to raise awareness of disaster management for both government staff and community members. The details of the seminars are found below. Date and Place 6 th of March 2006 in Matara and 8 th of March 2006 in Colombo Agenda Opening Speech Presentation of the Project and Outcomes Presentation of Disaster Management Coffee Break Presentation by a Sri Lankan Expert Open Discussion Handover Ceremony of the Hazard Map Best Practice Award Giving Ceremony (only in Matara) Closing Speech 10-83

38 Attendees Matara - Government officials: GA office, Police, DMCC - Pilot Project related agencies. - PP representatives - Donors and NGOs - Academics Colombo - Government officials: Central ministries, DMC, - McRAP Committee - Donors and NGOs - Academics Presentation materials used for the seminars are found in Annex

39

40 Chapter 11 Conclusions 11.1 Conclusions Summary of Conclusion The JICA Project Team has done various work for in rehabilitation process in the southern region of Sri Lanka. To examine the achievements, the results of the Project were summarized below against the objectives of the Project, as stated in Chapter 2. To formulate a plan for the recovery, rehabilitation, and development of tsunamiaffected areas in the southern region of Sri Lanka. The JICA Project Team introduced two approaches for developing the rehabilitation program, which were based on the Japanese experience in the Great Hanshin-Awaji Earthquake, that is the bottleneck identification approach and co-assistance approach. The bottleneck identification approach is based on lessons learned in the aftermath of the Great Hanshin-Awaji Earthquake and was used for identifying rehabilitation bottlenecks and identifying necessary activities quickly. Co-assistance, defined as mutual assistance among victims, is employed for expediting livelihood rehabilitation. Utilizing the bottleneck identification approach, strategies for rehabilitation of the southern region were developed. The strategies included assessment and identification of necessary rehabilitation activities. For verification of the co-assistance approach, three Pilot Projects for community assistance were formulated and implemented. A fisheries sector rehabilitation strategy was also developed. The JICA Project Team implemented three Pilot Projects in the following aspects: refugee camp support, fishery cooperative support, and small industry association support. These Pilot Projects consisted of various community activities with full consideration of the co-assistance approach, such as the establishment of associations and steering committees, publishing of newsletters, and lending of equipment. Through implementation of the Pilot Projects, the JICA Project Team assessed these two approaches. It was found that the two approaches generally worked well for the rehabilitation process in Sri Lanka. Based on these results, lessons learned and follow-up work of the Pilot Projects were prepared. 11-1

41 The JICA Project Team also developed suggestions for the disaster management plan in the Matara area. These suggestions included disaster prevention measures and emergency preparedness and response measures. The community activities for disaster preparedness were also included, based on the co-assistance approach. As a part of disaster management, the JICA Project Team prepared a preliminary tsunami hazard map and tsunami/flood disaster maps. Disaster awareness workshops were also held. To assist and monitor technically the implementation of recovery and rehabilitation projects to be funded under Japanese Non-project Grant Aid and ODA Loans. The JICA Project Team technically assisted and monitored two rehabilitation projects funded under Japanese Non-project Grant Aid and ODA Loans: (i) the reconstruction of the Matara Aqueduct Bridge; and (ii) the rehabilitation of fishery harbors at Galle and Tangalle. The JICA Program Team has been in charge of technical assistance of design, tender, and monitoring of construction. Construction work is being implemented by the Non-Project Grant Aid scheme. To share Japanese experiences in disaster management through implementation of the Project. The two approaches: the identification of bottleneck and the co-assistance were adapted from Japanese experience. These approaches formed the backbone of the Project. In particular, Pilot Projects applying the co-assistance approach were well-accepted and appreciated by the subject persons. This success, however, was derived partly from the traditional propensity for group work in Sri Lankan rural societies. Nevertheless, differences between Sri Lanka and Japan in coping with tsunamis on the one hand and earthquake on the other, each with different social settings, were made apparent in details. It is very much desirable to formulate approaches fully reflecting Sri Lankan realities. Also, some concrete measures concerning disaster management suggested for the Matara area were adapted from measures developed in Japan, including tsunami hazard maps and tsunami signboards, etc. 11-2

42 Lesson Learned from the Project (1) Identification of Bottlenecks The list of possible bottlenecks adapted from ones prepared in Japan proved to be quite useful. However, it is strongly recommended that the list be modified to reflect the actual circumstances in Sri Lanka by planners of the regional rehabilitation plan especially in the following aspects: - The dual structure of local administration in Sri Lanka; - Involvement of a large number of donors and NGOs; - The co-assistance approach; - Availability of land for evacuees; - Roles of the Urban Development Authority, Provincial Government, and Municipality; - Adequately reflecting the needs and opinions of victims themselves; - Reviving small-scale industries; and - Expanding to include public health and education anticipated in earlier stages. (2) Co-Assistance The co-assistance approach proved to be quite effective in post-disaster management. There are a number of points that have emerged in the course of implementing the Pilot Projects, which, if properly done, would make the approach even more effective. These points are as follows: - Members of a newly formed society should be of a similar background (as much as possible); - Societies should be given an opportunity to negotiate with external organizations to the extent possible; - Meetings should be held as frequently as possible; - Micro-finance can serve as a glue to keep society cohesive, while its interest payments can be used for the society s own activities; - Training can enhance the standing of society to individuals, a mutually reinforcing cycle; and - Formation of internally-cohesive and outwardly-oriented societies enables them to link up with existing organizations to enlarge their scope. 11-3

43 (3) Priority Projects for Infrastructure Smooth Project Identification The Steering Committee, which included both Sri Lankan and Japanese representatives concerned with Non-Project Grant Aid, has selected projects quickly. In the course of selecting candidate project components, the Steering Committee has widely gathered information on components with urgent restoration needs from local and central government agencies. This action was performed in a short time period, resulting in the selection of project components that obviously required urgent restoration as claimed by the local residents affected by the tsunami. Quick Preparation of Facility Design Draft tender documents were prepared within two months after the commencement of the Project due to the urgency of facility rehabilitation and to resume their functions as early as possible. The other reason why the JICA Project Team prepared the design and tender documents in such a short time was that damaged facilities simply required restoration to their original forms, without time-consuming studies Recommendations This section summarizes recommendations on rehabilitation activities in Sri Lanka, based on achievements and evaluations of the Project Preparation of a List of Lessons and Checkpoints on the Disaster Rehabilitation In the Project, the JICA Project Team made the hypothesis that the methodology of using a list of lessons and checkpoints for disaster rehabilitation would be useful to identify bottlenecks and necessary administrative activities quickly. The bottlenecks and necessary activities were clarified according to the list developed after the Great Hanshin-Awaji Earthquake in Japan. Through the evaluation, it became clear that the methodology was quite useful for tsunami rehabilitation in Sri Lanka, although it was necessary to adapt contents to the situation in the Sri Lanka case. It is recommended that one government agency collect issues on tsunami rehabilitation from various government agencies, donors, and NGOs, and summarize them into one list. This list should be open to the public and disseminated to local authorities for rehabilitation of additional disasters. 11-4

44 Since the situation varies considerably according to the phase of rehabilitation, lists for each phase need to be prepared. In the Project, the JICA Project Team used only the list for initial recovery phase. At least, it is necessary to prepare lists for the emergency phase, evacuation phase, initial rehabilitation phase, and full-scale rehabilitation phase. Besides the list for tsunamis, it is also considered beneficial to have ones for floods and landslides, as these will be quite useful for effective rehabilitation in Sri Lanka Enhancement of Co-assistance Activities for Community Rehabilitation Based on Japanese experience, the JICA Project Team introduced the concept of co-assistance to develop the strategy of community rehabilitation. Three Pilot Projects for community assistance were formulated and implemented from the point of view of enhancement of co-assistance activities such as the establishment of committees, recovery of micro-finance schemes, and provision of training opportunities. Through the evaluation of these activities, it became clear that co-assistance activities improved internal relationships, mutual trust, and external networks. It was also observed that these improvements improved synergies in terms of assistance for physical, financial and human assets, and promoted rehabilitation of community. Considering the above findings, it is recommended that assistance of co-assistance activities be expanded to other tsunami-affected communities nationwide. As the first step, it is very important to disseminate a concept of co-assistance among government authorities, communities, and NGOs. Local authorities need to take the initiative of enhancing co-assistance activities in all affected communities, closely cooperated with NGOs and donors. The details are shown in Section 11.1 in the report Promotion of Regional Disaster Management As well as rehabilitation, preparation for future disasters is quite an important issue in this stage. The JICA Project Team focused on disaster management in the Matara urban area. Through a review of existing measures, it was clear that disaster management in Matara was insufficient. It is recommended that comprehensive disaster management measures be implemented in Matara area as soon as possible. Disaster prevention measures are expected to prevent or reduce physical damage from disasters directly, such as the construction of seawalls and strong buildings in dangerous 11-5

45 area. Emergency preparedness and response are measures for planning and preparing for emergency activities, including early-warning systems and emergency evacuation areas/routes, as well as disaster awareness measures. In addition, establishment of the Disaster Management Committee in Matara is recommended for promotion of the implementation these measures. The details of each measure are shown in Section 10.3 in the report. As a priority measure for disaster management, the JICA Project Team prepared a preliminary tsunami hazard map and tsunami/flood disaster maps. It is recommended that local authorities effectively use these maps in the following ways: (i) to authorize dangerous zones in reference to them; and (ii) to distribute them to government staff and community members for disaster awareness. 1) Phasing of Disaster Rehabilitation As is explained in Chapter 2 of this report, there are several phases in disaster rehabilitation, including the: (i) emergency rescue phase; (ii) evacuation phase; (iii) initial full-scale rehabilitation phase; and (iv) full-scale rehabilitation phase. Co-assistance can be applied in any of these phases. 2) Motivation to Prepare for Another Disaster Many communities and governmental agencies are still very much motivated for disaster prevention. It is best to disseminate the lessons learned while people are motivated. 3) Integration of Disaster Preparation and Rehabilitation If the co-assistance approach is to be disseminated, the three steps mentioned in the previous section (i.e., survey, policy, and training) should include aspects of disaster preparedness Dissemination of the Project Results As described above, the Project produced valuable recommendations on regional rehabilitation processes. It is very important to widely disseminate these results. There are two targets for dissemination, communities and government officers. All three recommendations (including a list of lessons and checkpoints), as well as co-assistance, and disaster management aspects, needs to be disseminated to central government agencies and local authorities through workshops and seminars. For communities, it is useful to raise the awareness of co-assistance and disaster management through 11-6

46 workshops, seminars, and public relations. The details of the awareness activities are shown in Sections 10.4 and (1) Dissemination of Co-assistance Approach Through implementation of the Pilot Projects, the JICA Project Team realized that the co-assistance approach works. The question is how this approach can be spread to other parts of Sri Lanka. Since the Pilot Projects targeted a very limited part of the tsunami-affected population (even in Matara) and only a few government officers were actually involved in their implementation, the immediate influence of the co-assistance approach has been quite limited. People in Sri Lanka are relatively well-educated and have a tradition of mutual help. Community development projects of the National Housing Development Authority, as well as of Sarvodaya, are widely-practiced in Sri Lanka in line with the co-assistance approach. Communities in this country have substantial foundations to build co-assistance structures upon. (2) Target Community Groups One year after the disaster, all the tsunami-affected population has received external assistance to some extent, recovering their livelihoods that existed prior to the disaster. As such, it may not be the best time to organize new societies and associations in order to start co-assistance work for rehabilitation. However, strengthening existing grassroots organizations has enough justification to accelerate the well-being of the people. Co-assistance work can be applied to many kinds of organizations such as: Various cooperatives (agriculture, fishery, consumers, etc.); Community-Based Organizations like committees under Grama Niradari; Age-based groups like community youth associations; Parent and teacher associations at each school or student groups; and Religious groups at each Buddhist temple, Hindu kovil, mosque, and church. These parties or Community-Based Organizations (CBOs) can be the targets for dissemination work, however, there should also be facilitators of co-assistance extension work. 11-7

47 (3) Three Steps of Co-assistance Dissemination The JICA Project Team has experienced various constraints concerning how to buildup co-assistance during the course of Pilot Projects. This experience should be shared with a broad set of disaster rehabilitation stakeholders in Sri Lanka. Three steps to disseminate co-assistance are proposed as follows. 1) Survey Of Existing Community-based Organizations and Their Achievements This survey is to identify effective CBOs for rehabilitation. Good practices were collected from all the potential target areas. These results may be used to develop CBO database and used to repeat good practices. Survey results can be used for subsequent steps. 2) Policy Development at the Government Level A central government agency, together with the District Secretaries of each area, should take initiatives to establish policies and rules of co-assistance. An example of this policy is the organization of residents associations in each refugee camp during the next disaster. Typical dissemination tools include seminars and training sessions for government officials at the central or local level as following. Training on public administration for co-assistance facilitation This training is for high-ranking government officials and includes investigation of disasterprone communities and discussions about disaster rehabilitation, disaster awareness, and disaster preparedness. Training on enhancement of co-assistance activities for tsunami-affected communities This training is designated for government officials on the rural or grassroots level. Venues include best practices of community rehabilitation in Tsunami-related countries and training of facilitators for enhancement of community activities. Participants have received the training in a practical manner, so that officers can obtain distinct ideas on how co-assistance works on the community level. Policy coordination with NGOs NGOs activities contribute a great deal to community rehabilitation. Government agencies, especially at the local level, need to coordinate with NGOs for the dissemination of co-assistance. Local agencies should have seminars or meetings with NGOs on how to introduce and implement co-assistance activities for sustainable rehabilitation. 11-8

48 3) Dissemination on the Community Level The following measures are proposed for dissemination of co-assistance on the community level. Group training for community leader Until each community understands the importance of co-assistance, co-assistance activities will not be disseminated widely. Government agencies closely cooperating with NGOs should provide group trainings for community leaders and committee members in order to introduce the co-assistance activities into the community as a first step to disseminate co-assistance to each community. Assistance of establishment of community organization The establishment of community organization is essential. During the Pilot Projects preceding the actual Project, the JICA Project Team promoted target communities to establish steering committees and the committees took an important role for various community activities. Local agencies, or NGOs, should assist to establish such committees in target communities, and encourage them to take leadership in their communities. Monitoring of co-assistance activities Even though co-assistance activities can promote sustainable rehabilitation with the initiative of communities, it is expected that there are many implementation issues on various co-assistance activities, especially in the introduction period. Local agencies and NGOs should frequently monitor community activities and advise them or provide assistance if it is necessary. Financial or physical assistance through committees is highly recommended because such assistance can enhance internal coordination and bonding of communities. On the other hands, assistance directly to each community member can easily discourage the initiative of coordination and cooperation among community members Subsequent Follow-up for the Project 1) Necessary Activities for Project Follow-up In regards to the implementation of recommendations, the following activities are recommended for short-term implementation by the Sri Lankan side and JICA. Sri Lankan side: - Dissemination by distributing copies of the report to concerned agencies; - Handover of the Pilot Project from the JICA Project Team to government agencies; 11-9

49 - Continuation of the Pilot Projects by government agencies, with close cooperation from Berendina; - Promotion of co-assistance dissemination through the following three steps: (i) survey of existing community-based organizations and their achievements; (ii) policy development at the government level; and (iii) trainings and seminars; - Preparation of lists of lessons and checkpoints according to the phase of rehabilitation and type of disaster (by a committee specifically formed for the purpose); and - Promotion of short-term measures on disaster management, such as the implementation of awareness programs with hazard maps and the establishment of a disaster management committee. JICA: - Support of the continuation of Pilot Projects, which are to be implemented by GA and Berendina (NGOs). In regards to the follow-up of the Pilot Project, the following institutional settings should be implemented. 2) Proposed Institutional Settings for Pilot Projects The JICA Project Team has built the following structures to execute the Pilot Projects: Each project has one or two government agencies as the counterparts; For each Pilot Project, beneficiaries formed a society or an association for the purpose of achieving project goals; Berendina, a Sri Lankan NGO, has managed the operation of the Pilot Projects; and The steering committee of each Pilot Project includes representatives from beneficiaries, government agencies, local NGOs, and the JICA Project Team. Co-assistance is a mechanism that tsunami-affected people use to organize their own society or association to effectively control and facilitate rehabilitation. This does not mean that co-assistance mechanism should be free of external assistance, however. On the contrary, obtaining assistance from government, donors, and NGOs is a very important part of the co-assistance mechanism. The diagram below illustrates the operational structure of the Pilot Projects

50 JICA Project Team Government Agencies Steering Committee (JICA Team as Secretariat) Local NGO (Contract with JICA) Association of Affected People Association of Affected People Beneficiary Beneficiary Beneficiary Beneficiary Beneficiary Beneficiary Source: Project Team Figure 11.1 Organizational Structure for Pilot Project Implementation Figure 11.2 shows a suggested structure for Pilot Project implementation after the JICA Team departs in March Local NGO Government Agencies Additional Support from Donors (Including Japan) Association of Affected People Association of Affected People Beneficiary Beneficiary Beneficiary Beneficiary Beneficiary Beneficiary Source: Project Team Figure 11.2 Pilot Project Structure after March 2006 Finally, it was concluded that governmental agency involvement is essential. The JICA Project Team exchanged Memorandum of Understanding (MOU) with the governmental agencies mentioned in Table 11.1 below. MOUs for the three Pilot Projects are attached in 11-11

51 Annexes 4-8, 5-8, and 6-7. These government agencies are also in a position to extend co-assistance approaches to other parts of the nation. Table 11.1 Government Agencies that Supported Pilot Projects Pilot Project Government Agencies 1 Refugee Camps District Secretary, Divisional Secretary 2 Fishery Cooperative Societies 3 Small Industries Industrial Development Board Source: Project Team Fishery Department (Matara branch office of Central Government Department) Cooperative Office (Matara branch office of Southern Provincial Council) A local NGO, the Berendina Foundation, has been undertaking the manager of Pilot Project implementation, while the JICA Team has acted as the supervisor. Considerable know-how in managing co-assistance projects has been accumulated in this NGO. The JICA Team also expects Berendina to stay in Matara and continue its support of beneficiary societies and associations. The associations have reached a point where they do not require daily support for their activities, thus only monitoring and occasional consultation activities would be sufficient for sustainable co-assistance activities

52