Asian Disaster Preparedness Center

Size: px
Start display at page:

Download "Asian Disaster Preparedness Center"

Transcription

1 Asian Disaster Preparedness Center Sustainable End-to-End Climate/Flood Forecast Application Through Pilot Projects Showing Measurable Improvements Draft Final Report on Post-Flood Assessment of CFAN Flood Forecast 2007 January 2008 A public trust under the Ministry of Water Resources

2 Table of Contents 1. Introduction Method of Post Flood Assessment Performance of CFAN Flood Forecast Performance Evaluation of CFAN Forecast Dissemination at Pilot sites Conclusion and Recommendations...15 Appendix A: Questionnaires Used for Post-Flood Assessment of CFAN Flood Forecast System Appendix B: Matching of Water Level Change between Forecasted and Real Situation in Each Pilot Union List of Tables Table 1: Performance of CFAN Flood Forecast...5 Table 2: Flag Operation Sites in Pilot unions...6 Table 3: Community responses about the flood forecast system...9 Table 4: Crop loss due to flood...10 Table 5: Crop cultivation after flood...11 List of Figures Figure 1: Selected pilot unions in context of Bangladesh...2 Figure 2: Flood extent map of Bangladesh...4 Figure 3: SMS understanding and Flag Conversion by Flag Operators...7 Figure 4: Community Consultation at Flag site, Char Mohanmia, Gazirtek...8 Figure 5: Paddy field destroyed by flood at Rajpur, Gaibandha...10 Figure 6: Farmers could save Jute partially before flood...10 Figure 7: Household Survey in the pilot unions...12 Figure 8: Source of flood early warning...13 Figure 9: Activities of people after getting CFAN flood forecast...13 Figure 10: Activities of people before and during flood

3 1. Introduction With the view of long term flood forecast in Bangladesh, Climate Forecast Application Network (CFAN) produced flood forecast with 10 days lead time in The major objective of the CFAN flood forecast is to increase lead-time. The Flood Forecasting and Warning Centre (FFWC), under BWDB, provided the forecast results to CEGIS along with the CFAN partners. CEGIS was assigned to disseminate the flood forecast to the community during monsoon season. To accomplish this task CEGIS has prepared flood forecast message understandable to the community, installed the dissemination network and instruments. From this year CEGIS has disseminated the flood forecast for four days lead-time in five pilot areas Rajpur union of Lalmonirhat district, Uria union of Gaibandha district, Kaijuri union of Sirajganj district, Bekra Atgram union of Tangail district and Gazirtek union of Faridpur district (Fig. 1). Post flood assessment has been carried for the flood forecast dissemination in The main objective was to assess the effectiveness and drawbacks of the flood forecast and dissemination system. The assessment has been done at community level and individual level in pilot unions. The post flood assessment results are presented in the report. Figure 1: Selected pilot unions in context of Bangladesh 2

4 2. Method of Post Flood Assessment Assessment of flood forecast dissemination system has been carried out at various levels as follows: (a) Assessment of CFAN Flood Forecast: The CFAN flood forecast has been assessed through comparing with actual water level variation in the river and the forecasted rise and fall of water level. (b) Assessment of Flag Operators: The flag operators were assessed about their experience in flag operation, maintenance, community response and difficulties. (c) Assessment of the Community at Flag Sites: The local people at each flag sites were asked about their experiences with the new flood forecasting system, reliability of information, understanding of flag messages, usefulness and suggestion for further improvement. (d) Assessment of Farmers: As the farmer is main target group of this flood forecast system; farmer s community was especially assessed about the flood forecast system. As the starting flood forecast dissemination was delayed this year, real benefit of the farmers due to the flood forecast could not be estimated but potential benefit of the flood forecast has been estimated. (e) Assessment of Households: Household survey was done to determine the benefit of the households who received early warning from the flood forecast system. (f) Assessment of SMS Receiver: The SMS receivers of the flood forecast message were asked about their understanding of message, benefit and difficulties about the system. The questionnaires used for post flood assessments are given in Appendix-A. 3. Performance of CFAN Flood Forecast The flood started on last week of July 2007 in Bangladesh through the Brahmaputra basin and Ganges basin. The districts along the Brahmaputra Jamuna river and Ganges river such as Rangpur, Kurigram, Lalmonirhat, Gaibandha, Sirajganj, Jamalpur, Bogra, Faridpur, Rajbari, Madaripur, and Shariatpur are affected by flood. According to CEGIS (2007), about 44,000 sq. km area of Bangladesh was inundated by flood on 3rd August 2007 (analyzed from RADARSAT image) (Fig. 2). The water level raised above danger level for two times, which caused two times inundation of adjacent areas along the major rivers. As lying in the Brahmaputra Ganges basin, the pilot unions under this study has inundated during the flood. Dissemination of CFAN flood forecast in the pilot union was started from 21 st August 2007 through the established dissemination network. Although it was delay for the first flood, the CFAN flood forecast was regularly disseminated for second flood from August 21, 2007 to October 11, Although 10 days flood forecast was generated by CFAN/ FFWC, the 4 days forecast was disseminated because of higher probability and testing the forecast for the first time in the field. The mean value of flood forecast was used. The estimation method of flood forecast message and dissemination network was provided in the report Piloting Flood Forecast Dissemination 2007 submitted in September CFAN flood forecast information was used in 4 pilot unions - Uria, Kaijuri, Gazirtek and Bekra Atgram, whereas FFWC regular 72 hours forecast for Kaunia station (in Teesta river) were used at Rajpur union because there is no CFAN forecast for Kaunia station, which is nearest to Rajpur union. The daily records of 72 hours forecast for Kaunia station was used to estimate 4 day forecast for Rajpur union. 3

5 District: Lalmonirhat Thana: Lalmonirhat Sadar Union: Rajpur District: Gai ban District: Tangail Thana: Nagarpur Union: Bekra Atgram District: Sirajganj Thana: Shahajadpur Union: Kaijuri District: Faridpur Thana: Char Bhadrasan Union: Gazirtek Figure 2: Flood extent map of Bangladesh (Source: CEGIS 2007, produced from RADARSAT image taken On 3rd August visited on 10 th December 2007) 4

6 Performances of the flood forecast has been assessed in terms of two criteria: (i) matching actual trend of rising or falling water level with the forecast and (ii) accuracy of magnitude of forecast. Flood forecast and actual water level record of gauges from August 21 September 26, 2007 has been taken for analysis of performance for Uria, Rajpur, Kaijuri and Gazirtek unions. For Bekra Atgram union, the forecast and water level records from September 03 October 02, 2007 has been taken for analysis of performance. These timeframes has been taken because reliable actual water level records have been found from the plot sites. The trend of actual rising or falling of water level was analyzed from the record of water level gauges, which were installed at each site. The actual rise and fall of water level has been compared with the forecasted message. It has been observed that about 70% forecast of rising and falling case (out of 37 days in Uria, Kaijuri and Gazirtek; and out of 30 days in Bekra Atgram) were matched with the actual situation at all pilot sties except Rajpur, Lalmonirhat (Table 1). At Rajpur union, only about 40% forecast of rising and falling case (out of 37 days) matched with actual situation. There is possibility of getting good result if CFAN forecast is used for Kaunia station like other sites. The matching of water level rising or falling trend at pilot unions are shown in Figures B1 B5 of Appendix B. Table 1: Performance of CFAN Flood Forecast Pilot Union Matching actual trend of change of water level with forecast Over all Accuracy of Magnitude of Water Level Accuracy for Rising Accuracy for Falling Accuracy for No Change Case Uria, Gaibandha % % % % % Kaijuri, Sirajganj % % % % % Gazirtek, Faridpur % % % % % Bekra Atgram, Tangail % % % % % Rajpur, Lalmonirhat % % % % % Beside the analysis of rising or falling trend, the accuracy of magnitude of flood water level has been analyzed. The magnitude of water level has been converted into number of flag for both actual water level and forecasted water level. Then comparison was made for accuracy of magnitude. Although the trend of rising or falling water level matched in most of the cases, the magnitude of water level in actual situation differed with the forecast in most of the cases. Comparatively good result has been found in Gazirtek and Gaibandha, where about % forecasted water level matched with real situation (Table 1). In case of rising water level at Bekra Atgram union, falling water level at Rajpur union and no water level change case at Kaijuri union, no forecasted magnitude of water level has been matched with real condition. While producing flood forecast for the community, there were some problems/ limitations such as, - CFAN flood forecast was not directly produced for the pilot sites. So conversion / estimation of the flood forecast for the pilot sites might have errors; - There is no CFAN flood forecast for Kaunia station of Teesta river. So estimation of flood forecast for Rajpur union was difficult; and - As there is no water level gauge stations at the floodplains of pilot sites beside the water level gauge station in river, there is low probability of accuracy for magnitude of flood water level with the forecast. 5

7 The results of the analysis of CFAN flood forecast in terms of trend and magnitude has also been reflected by the response of community. 4. Performance Evaluation of CFAN Forecast Dissemination at Pilot sites 5.1 Performance of Flag Operators CFAN flood forecast dissemination was started from August 21, 2007 and ended on October 11, The flag operators were trained to read mobile SMS messages, hoisting flag accordingly and disseminate the information to the community as much as possible. Flag were installed in 18 flag sites in five pilot unions, where 3-4 flag sites in each union (Table 2). Most of the flag operators operated flag regularly except in union parishad (UP) of Gazirtek union and Char Mohanmia, Gazirtek. The responsible UP member of Gazirtek union was not able do his job due to emergency flood situation. And the responsible person at Char Mohanmia, Gazirtek had to shift to other place due to high flood. During assessment, the flag operators were asked about understanding of mobile SMS and conversion of SMS to flag. Among the 19 flag operator surveyed, 53 % flag operator understand SMS message satisfactorily, 42 % understand fairly (Fig. 3). Most of them (about 80%) can hoist right flag converting from SMS. 16% of them fairly can hoist the right flag according to SMS (Fig. 3). Table 2: Flag Operation Sites in Pilot unions Pilot Union Kaijuri Uria Rajpur Gazirtek Bekra Atgram Flag Sites Flag operated 1 Bhatpara Bazar Yes 2 JJB Madrasa Yes 3 Kaijuri Hat Yes 4 Kaijuri Union Parishad Bhaban Yes 1 Union Parishad Bhaban Yes 2 South Ratanpur Mosque Yes 3 Gunvari High School Yes 1 Rajpur Union Parisad Yes 2 Rajpur Hudur Bazaar Yes 3 Tajpur Embankment Yes Remarks 4 Modhuram Govt. Primary School Yes 1 Bepari Dangi, Gazirtek Yes 2 Char Mohonmia, Gazirtek No People shifted to other place due to high flood 3 Somir Bapari Dangi, Char Harirampur Yes 4 Union Parishad, Hazigang Bazar No Union member was responsible. He could not operate due to emergency situation of flood. 1 Bekra Union Parishad Bhaban Yes 2 Barapusa Bridge Yes 3 Mushuria South para Mosque Yes 6

8 % of Flag Operators Satisfied Fair Not Acceptable 0 SMS understanding Flag conversion Figure 3: SMS understanding and Flag Conversion by Flag Operators Most of the flag operators (90%) use their own mobile set for receiving SMS and flag operation. Rest of them uses relative s mobile set for receiving SMS. Almost all day they received SMS and operated flag but they did not write in the flag operator s logbook regularly. In 2007, the flood forecast SMS message has been sent mostly in the afternoon. But most of the flag operators suggest that they need SMS message in the morning preferably between 10: 00 11:00 AM so that they can hoist flag early on the day. During flag operation, people showed interest to know the next flood situation. About 10 to 50 people asked to each flag operators about the flood forecast everyday in the flag sites. While operating flags, flag operators faced different problems. The major problems observed by them are given as follows: The time of SMS receiving was not fixed; Due to flag operation, daily routine work of the operator has been hampered to some flag operators, especially the person who work at Union Parishad; In some places, people annoyed the flag operator about flag message as local people don't know much about flag; Difficulties of understanding message written in English; In some places, flag operators could not hoist flag timely as the operator had to leave for shopping in some places; Facing problem of receiving SMS timely who use relative's mobile; The flags, bamboo & rope were stolen in one site. 7

9 5.2 Result of Assessment of the Community at Flag Sites Community level consultation was conducted at the flag sites in pilot unions (Fig. 4). They were asked about the flag message, visibility and design of flag, coverage area of flag, matching of flood forecast with real condition, etc. They also suggested important ideas to improve the system. The list of questions and the summary of community responses is given in the Table 3. According to the people at flag sites, almost 70% people understand the flag message. They realize that the rise and fall of water level matches with the flag message in all sites although the magnitude of rise and fall did not match in all cases. The 4-day lead-time is helpful for some cases but the lead-time should be increased so that farmers can be benefited more. Community Consultation at Flag site, Char Community Consultation at Flag site, Mohanmia, Gazirtek Gunvari, Uria Figure 4: Community Consultation at Flag site, Char Mohanmia, Gazirtek Community people have no objection about color of flag. The color of flag is visible up to m distance. People living within 250 m 500 m radial distance from the flag regularly informed about the flood forecast through this system. In case of some flag sites hoisted beside road, people within 1 km along the road were informed about the flood forecast. People suggested important ideas to improve the system. They suggested to increase the number of flag sites, use larger flag, taller flag stand, provide forecast early in the morning, and increase lead time of forecast. 8

10 Table 3: Community responses about the flood forecast system Sl. No. Questions Response 1 Do you understand the flag message? 70% of the respondents 2 Does the flag message match with the real condition of flood? 3 Does flag operator operate flag regularly? Yes 4 Does lead time of forecasting helpful? Yes 5 Does flood pillar reflects the real flood situation in the nearby area? Yes (Rise / Fall) Yes 6 Does color of flag visible from far enough? m 7 Catchments of flag (radial distance, km) 250 m m 8 Suggestion for flag site, SMS and lead-time? Flag size should be larger Flag stand should be longer Preferred time for message- before 9 A.M More campaign is required Flag should be installed at comparatively open space Flood forecast should be given before 2 weeks More flag site and pillar is needed 5.3 Result of Assessment of Farmers Farmer s livelihood group was specially assessed for evaluation of the CFAN flood forecast. Farmers were asked mainly two questions about the crops grown before and after flood and what they did after getting flood early warning. Before the flood, Jute, B. Aman, T. Aman and Aus were major crops in all pilot sites. Jute and Aus crops were in ripening stage, which were saved partially (Figs. 5 and 6). According to the farmers, Jute could be saved if they got CFAN flood forecast few days ago. Aus rice was not fully harvestable in all areas. So it could be partially saved in spite of flood forecast. Since T. Aman is major crop growing in monsoon season, the seedlings and seed beds were damaged a lot. The cost of seedling and seed bed could be saved if they got flood forecast early. The farmer could plant seedlings after flood. The affected crops due to flood in the pilot areas are given in the Table 4. 9

11 Figure 5: Paddy field destroyed by flood at Uria, Gaibandha Figure 6: Farmers could save Jute partially before flood Table 4: Crop loss due to flood Union Crop Name Crop Stage Rajpur T. Aman Seedling / Seed bed What they did Nothing Aus Ripening Cultivated at High Land Jute Ripening Partially harvested What can be done Seedling can be delayed or cost can be saved Partially can be saved 100% can be harvested Amount of loss (area of crop) 70% of the total cultivable area (full damaged) Partially damaged 25-40% of the total cultivable area (partially crop damaged) Uria Kaijuri Bekra Atgram Jute Ripening Partially harvested T. Aman Seedling / Seed bed Nothing 100% can be harvested Seedling can be delayed or cost can be saved Partially can be saved 25-60% of the total cultivable area (partially crop damaged) 60% of the total cultivable area (full damaged) Aus Ripening Cultivated at High Land Partially damaged B. Aman Vegetative Nothing Nothing 100% of the total cultivated crop was damage. T. Aman Seedling / Nothing Seedling can be 10% of the total cultivable Seed bed delayed or cost can area (full damaged) be saved Jute Ripening Partially 100% can be Partially damaged harvested harvested B. Aman Vegetative Nothing Nothing 60% of the total cultivable area (fully damaged) 10

12 Union Crop Name Crop Stage What they did Jute Ripening 20% was harvested Gazirtek T. Aman Seedling / Seed bed Vegetables, Chili, oil seed (Teel) Nothing What can be done Jute could be saved Seedling can be delayed or cost can be saved Partially can be saved Amount of loss (area of crop) 20-30% of the total cultivable area (fully damaged) 40-60% of the total cultivable area (fully damaged) 5% of the total cultivable area (full damage). B. Aus Ripening Nothing Partially can be saved 60% of the total cultivated area (fully damaged) T. Aman Seedling / Seed bed Nothing Seedling can be delayed or cost can 100% of the total cultivated crop was damage. be saved Jute Ripening Partially 100% can be saved 25% of the total cultivable saved area (partially damaged) Sugarcane Development Nothing Partially can be saved 10% of the total cultivable area (full damaged) Nut Development Nothing Partially can be saved Full damaged Source: Field Survey, 2007 In most of the pilot unions, Mustard, Pulses (Khesari, Maskolai, Mushuri, etc.), vegetables (Potato, Brinjal, Chili, etc.) were grown immediately after flood. Farmers had chosen these crops from their own experience due to low cost and short term of maturity. Different crops grown in the pilot unions after flood are given in Table 5. Farmers urged that the CFAN flood forecast would be helpful if they got it early of the season. Also longterm forecast will be more helpful for taking decision of crop cultivation. Table 5: Crop cultivation after flood Union Crop Name Why chosen Time of cultivation Rajpur T. Aman (BR-11, High Production August 28) September Mustard Short term crop, October High benefit November Vegetables Short term crops, October High benefit November Uria Mustard Short term crop, High benefit October November Vegetables (Potato, They cultivated it regularly, October Brinjal, Chili) Short term crops November Aman (Kajir Shail, High Production August BR-11, 28) September Kaijuri Mustard Short term crop, October High benefit November Who gave suggestion DAE/BS, Own Experience 11

13 Union Crop Name Why chosen Time of cultivation Pulses (Maskali, October Khesari) Suitable for the area November Wheat, Maize, Short term, October Vegetables Low cost and suitable in the November area Bekra Mustard Short term (2.5 month) Middle of Atgram October Pulses (Maskali, Short term. Middle of Khesari) Fertilizer requirement is October Gazirtek Potato, Corn, Wheat, Vegetables minimum. Short term, Low cost and suitable in the area Middle of October Wheat Short-term and high benefit October November Mustard Short-term and high benefit October November Pulses (Maskali, Short-term and high benefit October Mushuri) November Vegetables Short-term and high benefit October November Who gave suggestion 5.4 Result of Assessment of Households Total 62 households were surveyed in the five pilot unions. The households were selected randomly. Each household head were asked about the flood forecasting system and their activities after getting flood forecast information (Fig. 7). About 70% households know about the flood forecast system using flags and mobile SMS. Most of them clearly understand the flag messages. Household Survey in Berka Household Survey in Rajpur Figure 7: Household Survey in the pilot unions 12

14 Out of 62 households, 56 household got flood early warning through various sources such as flags, neighbors, TV/ Radio/ newspaper, miking by administration, volunteers or community associations, etc. (Fig. 8). It is observed that most of the households (66%) got early warning through observing flags (Fig. 8). Community Associations 7 Source of flood early warning Volunteers Family members Observed Flood Markers TV/ Radio/ Newspaper Miking by administration Neighbors Observed flags % of Household responded Figure 8: Source of flood early warning The households, who observed the flags regularly (i.e. got CFAN flood forecast), were asked about their activities after getting warning message. Most of them were involved with saving livestock; shifting to safe place; storing food, water and fuel (Fig. 9). Since the flood forecasting was delayed in 2007 in the pilot unions, people could not use this information for saving crops. However they saved farming loss (i.e. re-plantation of paddy seedlings) during second time flood using CFAN flood forecast in some areas. Some photographs of the activities are shown in Figure 10. Activities after getting flood forecast Making raft Health Care Saving Crops/ Fisheries Storage of drinking water Children /women/elderly care Making Macha Fuel storage Shift to safe place Food storage Saving livestock % of HHs responded 100 Figure 9: Activities of people after getting CFAN flood forecast 13

15 (a) Shelter for human and livestock on road (b) Collection of drinking water (c) Raising net around pond to protect fish (d) People living on macha (bamboo made structure) during flood (e) Health care of a diarrhea patient during flood (the patient had been shifted to shelter from his house) Figure 10: Activities of people before and during flood 14

16 5.5 Result of Assessment of SMS Receivers Since the mobile SMS receivers were the member of flag operation group at each flag sites, they were asked similar questions as flag operators. Hence, similar responses have been found from the SMS receivers. They suggested that the mobile SMS should be in Bengali so that they can understand easily. In future, the number of SMS receiver will be increased and only the SMS receivers, who are not engaged in flag operation, will be assessed. 5. Conclusion and Recommendations The community based CFAN flood forecast system has generated greater interest of people living in the pilot unions. The flood forecast should be more specific to the pilot unions so that the forecast will match with real condition more accurately. People have realized the benefit of the forecast and start to believe in the system. Although flood forecast disseminati on was delayed, the dissemination network has been established and part of the community has been trained. More campaign and training on the flood forecast sy stem is required in the community. More training and awareness of the Union Disaster Management Committee (UDMC) is required to institutionalize the system. Regular monitoring of flag operators, SMS receivers, flood pillars, water level gauge stations, etc. is required to make sustainable operation of the flood forecast system. 15

17 Appendix A Questionnaires Used for Post-Flood Assessment of CFAN Flood Forecast System

18 Sustainable End-to-End Climate/Flood Forecast Application Through Pilot Projects Showing Measurable Improvements (Post Flood Assessment Survey, 2007) Assessment of Flag Operators Pilot Union: Date: Surveyor: Sex Age Edu. Received SMS Used Own Mobile Is time of receiving SMS suitable to operate flag? Did people ask you about flag message? Understandi ng Forecast Sl. No. Name of Flag Operators Prob lems of flag ope ration No. of days % Y if no, whose no range of receiving SMS Preference No. of people/ day Not SMS understanding (S/F/NA) Flag conversion (S/F/NA) Dissemination (S/F/NA) Note: S = Satisfactory; F = Fair; NA = Not Acceptable

19 Sustainable End-to-End Climate/Flood Forecast Application Through Pilot Projects Showing Measurable Improvements (Post Flood Assessment Survey, 2007) Assessment of Flag Message User through community consult ation Participants people in each flag site Pilot Un ion: Date: Name of Flag Site: Name of Flag Operator: Surveyor: SL. No. Question 1 Do you understand the flag me ssage? 2 Does the flag message match w ith the real condition of flood? 3 Does flag operator operate flag regularly? 4 Does lead time of forecasting helpful? 5 Does flood pillar reflects the real flood situa tion in the nearby area? 6 Does color of flag visible from fa r enough? 7 Catchments of flag (radial distance, km) 8 Suggestion for flag site, SMS and lead time? % of responses Yes No

20 Sustainable End-to-End Climate/Flood Forecast Application Through Pilot Projects Showing Measurable Improvements (Post Flood Assessment Survey, 2007) Damage Assessment of Farmers in Flood 2007 through community consultation Participants farmer in each site Pilot Union: Date: Name of Site: Surveyor: SL. No. Question Crop Name What they What can be Amount of loss did done (area of crop) 1 What crop was in field before the flood came? 2 What crop are you cultivating after flood? Crop Name Why chosen Time of cultivation Who gave suggestion

21 Sustainable End-to-End Climate/Flood Forecast Application Through Pilot Projects Showing Measurable Improvements (Post Flood Assessment Survey, 2007) Household Level Survey SL.# Date: Reporting Period: Pilot Union: Name of the respondent Gender Father/ Husbands Name Age(Year) Occupation Mon thly Income ( Taka) Education Male Female Illiterate Can sign Primary Below SSC SSC

22 SL. # If yes, please indicate whether Do you know the correct or about flood incorrect of flag was hoisted warning flag and (Corr= Correct, Inc= Incorrect) its Use mark interpretation? (use mark) Blue White Have you got early warning? (use mark) If yes, what was the source of information? (use mark) If yes, whether the message received about increase/decrease of water was correct or incorrect If yes, Whether the message received about level of water (quantity) was correct or incorrect) 1 Yes No Corr Inc Corr Inc Yes No Obser ved flags Neighbo rs Other Family membe rs CAs Volunte ers Observ ed Flood Marker Others corr Inc Corr Inc 2 3 4

23 SL. # 1 Was the information received useful to you for reducing losses? (use mark) Useful Partially useful Useless Evacuation to safe place Farming If useful or partially useful, what were the areas of mitigation? (use mark) Livestock management Food storage Fuel storage Children /women /elderly care Health Care Ensuring safe water Making raft Making Macha Others Name of the Enumerator: Signature: Date:

24 Sustainable End-to-End Climate/F lood Forecast Application Through Pilot Projects Showing Measurable Improvements (Post Flood Assessment Survey, 2007) Assessment of SMS Receivers Pilot Union: Date: Surveyor: Sex Age Edu. Received SMS everyday Were SMS msg and flags same? Is the SMS easily understanda ble? Did people ask you about msg? Is time of receiving SMS suitable? Sl. No. Name of SMS Receivers List saved assets due to f lo od EW S MS No. of days % Y (%) N (%) Y N Understandable but Need help from others No No. of people/ day SMS understanding (S/F/NA) Flag conversion (S/F/NA) Dissemination (S/F/NA) What did you do? 1.Name: Amount: 2. Name: Amount: 3. Name: Amount: Note: S = Satisfactory; F = Fair; NA = Not Acceptable

25 Appendix B Matching of Water Level Change between Forecasted and Real Situation in Each Pilot Union n Water Level Difference i Number of Flags District: Gaibandha Th ana : Fulchhari Union: Uria Matching of For ecast with real condition = % Accuracy of Magnitude = % Aug 23-Aug 25-Aug 27-Aug 29-Aug 31-Aug 2-Sep ep 4-S Actual Water Level Difference in No. of Flags Figure B1: Matching of flood forecast with real situation in Uria union under Gaibandha district (Note: One Flag = 22 cm or 9 inch) 6-Sep ep 8-S Dates 10-Sep Sep 12- p 14-Se p 16-Se p 18-Se ep 20-S 22-Sep 24-Sep 26-Sep Forecasted Water Level Difference in No. of Flags Water Level Difference i n Number of Flags District: Sirajganj Thana: Shajadhp ur Union: Kaijuri Ma tching of forecast with r eal condition = % Accuracy of magnit ude = % Aug 23-Aug 25-Aug 27-Aug 29-Aug 31-Aug 2-Sep 4-Sep 6-Sep 8-Sep 1 0-Sep 12-Sep 14-Sep 16-Sep 18-Sep 20 -Sep 22-Sep 24-Sep 26-Sep Dates Actual Water Level Difference in No. of Flags Forecasted Water Level Difference in No. of Flags Figure B2: Matching of flood forecast with real situation in Kaijuri union under Sirajganj district

26 Water Level Difference in Number of Flags Aug 23-Aug District: Faridpur Thana: Char Bhadrasan Union: Gazirtek Matching of forecast with real condition = % Accuracy of magnitude = % 25-Aug 27-Aug 29-Aug 31-Aug 2-Sep 4-Sep 6-Sep 8-Sep Dates 10-Sep 12-Sep 14-Sep 16-Sep 18-Sep 20-Sep 22-Sep 24-Sep 26-Sep Actual Water Level Difference in No. of Flags Forecasted Water Level Difference in No. of Flags Figure B3: Matching of flood forecast with real situation in Gazirtek union under Faridpur district Water Level Difference in Number of Flags District: Tangail Thana: Nagarpur Union: Bekra-Atgram Matching of forecast with real condition = % Accuracy of Magnitude = % Sep 5-Sep 7-Sep 9-Sep 11-Sep 13-Sep Actual Water Level Difference in No. of Flags 15-Sep 17-Sep Dates 19-Sep 21-Sep 23-Sep 25-Sep 27-Sep 29-Sep 1-Oct Forecasted Water Level Difference in No. of Flags Figure B4: Matching of flood forecast with real situation in Bekra Atgram union under Tangail district.

27 4 3 Water Level Difference in Number of Flags Aug 23-Aug District: Lalmonirhat Thana: Lalmonirhat Sadar Union: Rajpur Matching of forecast with real condition = % Accuracy of Magnitude = % 25-Aug 27-Aug 29-Aug 31-Aug 2-Sep 4-Sep Actual Water Level Difference in No. of Flags 6-Sep 8-Sep 10-Sep Dates 12-Sep 14-Sep 16-Sep 18-Sep 20-Sep 22-Sep 24-Sep 26-Sep Forecasted Water Level Difference in No. of Flags Figure B5: Matching of flood forecast with real situation in Rajpur union under Lalmonirhat district.