Forecast-Based Financing (Fbf) Flood Early Action Protocol (EAP) Bangladesh

Forecast-Based Financing (FbF) Flood Early Action Protocol (EAP) Bangladesh

Bangladesh Red Crescent Society

Last update: 29/05/2019 Submitted to IFRC: 31/03/2019

Early Action Protocol Summary

Hazard Floods Potential high risk prone areas Districts along the Jamuna River Basin where the FbF mechanism could be activated Prioritized risks to be tackled by 1. Human causality (deaths due to drowning and others) early actions 2. Loss of livelihood linked to livestock 3. Loss of movable assets and food grains Proposed Early Actions 1. Unconditional cash grant 2. Evacuation transportation by boat Potential No. of household to be 3,300 households reached Expected activation budget CHF 249,862 Forecast sources of information National: Flood Forecast and Warning Centre (FFWC) Global: Global Flood Awareness System (GloFAS) Expected lead time for activation Pre-activation- 10 days; Activation- 5 days Responsible focal point for this Deputy Secretary General and Head of DRM Division, BDRCS EAP Governmental coordinating Department of Disaster Management, Bangladesh Water agency Development Board and Bangladesh Meteorological Department

1. Executive summary The Forecast-based Financing for the Bangladesh Red Crescent Society (BDRCS) supported by the German Red Cross (GRC) is a tool to guide timely and effective implementation of early actions when a certain weather or climate forecast shows a high likelihood of impacting people. If the event materializes and there are no preparedness actions in place, there is a high potential of humanitarian crisis. These crises could be avoided or minimized by the joint and timely action of the Government, communities at risk, stakeholders and the national society.

This Early Action Protocol (EAP) has been designed and will be implemented by several organizations, such as the Bangladesh Post Office (BPO) and/or private sector entities (e.g. commercial banks and telecom service provider) for Electronic Transfer System (ETS), the Swiss Red Cross (SRC), the American Red Cross, the German Red Cross, the International Federation of Red Cross Red Crescent Societies (IFRC), and the Bangladesh Meteorological Department (BMD). Everyone involved in this plan has a key role during the preparation and activation of the Forecast-based Financing mechanism. Depending on the context, this document could be an essential component of contingency plans at national and local level for the Government and the Bangladesh Red Crescent Society.

The Early Action Protocol (EAP) presented here covers flood related early actions which are triggered by a forecast provided by the Flood Forecast and Warning Centre (FFWC) along with global forecast models. This EAP provides step-by-step instructions for the selected actions to be implemented in a particular fashion and order, at the time it is activated. The EAP defines clearly: what are the actions to be taken by whom, when and where with what funds.

The EAP does not work in isolation. It is connected to already existing Disaster Risk Reduction and Management Plans and Contingency Plans of BDRCS. It ensures that early actions planned to be implemented in the window of time between the given forecast and the potential disaster event are carried out in a timely manner and successfully.

List of Acronyms

AmCross : American Red Cross BBS : Bangladesh Bureau of Statistics BDRCS : Bangladesh Red Crescent Society BMD : Bangladesh Meteorological Department BPO : Bangladesh Post Office BWDB : Bangladesh Water Development Board CDMC : Community Disaster Management Committee DDM : Department of Disaster Management DDMC : District Disaster Management Committee DRM : Disaster Risk Management Department, BDRCS DREF : Disaster Relief Emergency Fund EAP : Early Action Protocol FbF : Forecast based Financing FFWC : Flood Forecast and Warning Centre FSC : Food Security Cluster FSP : Financial Service Providers GoB : Government of Bangladesh GRC : German Red Cross HH : Household IFRC : International Federation of Red Cross and Red Crescent Societies MoFA : Ministry of Foreign Affairs, Germany MoU : Memorandum of Understanding NDRT : National Disaster Response Team, BDRCS RCCC : Red Cross/Red Crescent Climate Centre RCY : Red Crescent Youth volunteers, BDRCS SRC : Swiss Red Cross TO : Technical Officer GRC UDMC : Union Disaster Management Committee ULO : Unit Level Officer, BDRCS UzDMC : Upazilla Disaster Management Committee WB : World Bank WFP : World Food Programme

2. Key actors This section describes the key actors involved in developing this EAP along with their role in activation of this EAP.

The Bangladesh Red Crescent Society (BDRCS) The Bangladesh Red Crescent Society is present in all 68 districts of the country with around 874 full time staff and more than 81,804 life members, 55,515 CPP volunteers, 720,800 Red Crescent Youth (RCY) Members, 5000 RCY volunteers. 1500 Sr. volunteers along with significant numbers of community volunteers, Given the recurring natural disasters, BDRCS has extensive experience in responding to cyclones, floods, and other disasters and is coordinating the humanitarian response with the government and Red Cross Red Crescent counterparts. The Forecast-based Financing approach sits in BDRCS’ Disaster Risk Management (DRM) department and is actively supported with technical expertise by the German Red Cross and the Red Cross Red Crescent Climate Center. However, implementation of this EAP requires active role of other departments within BDRCS (for. e.g. the Disaster Response department, department of youth and volunteers, department of unit affairs etc.). In coordination with relevant departments, the DRM department has been responsible for orienting and training BDRCS Unit Level Officers (ULO) and Red Crescent Youth volunteers (RCY) on the FbF project, ensuring that FbF early actions are in line with the organization’s Strategic Plan for 2017-2020, and enabling linkages with national stakeholders such as the Flood Forecast and Warning Centre (FFWC). Strong linkages with BDRCS’ Disaster Response Department are established to support the EAP implementation and facilitate the link with a potential response operation. In the ‘Agreement of Cooperation’ between WFP, GRC, and SRC, BDRCS signed as a witness. The agreement highlights and acknowledges BDRCS’ crucial role and expertise in implementing the early actions of this EAP. The agreement will further strengthen the effort to collaborate more closely on FbF and flood related early actions at the national and local level.

German Red Cross (GRC) With more than 30 years of presence in Bangladesh and its global FbF experience, the German Red Cross is a strong technical partner for BDRCS and supports the development and implementation of the EAP. GRC provides substantial HR and technical support to ensure that BDRCS is in a position to leverage the strategic advantages of FbF. In close collaboration with BDRCS’ DRM department, GRC facilitates partnerships and alliances within the Red Cross Red Crescent movement but also with other humanitarian and governmental partners such as the World Food Programme or the Ministry of Disaster management and Relief (MoDMR), Department of Disaster Management. Given that GRC is involved in FbF projects in other countries in the region, GRC Bangladesh ensures that regional linkages are established and maintained for horizontal learning and broader policy alignment. For this EAP, GRC is providing overall coordination support through its FbF project staff at Dhaka and field level. Once a trigger condition is reached, GRC will provide technical support to BDRCS to activate Flood EAP and implement the early actions on affected districts. The German Red Cross signed an ‘Agreement of Cooperation’ with WFP and SRC to collaborate more closely on FbF and improve the timely implementation of early actions.

Red Cross Red Crescent Climate Centre (RCCC) The RCCC supports BDRCS, GRC, BMD, FFWC and other Movement and non-Movement partners to anticipate changes in climate-related risks and provides technical support to the FbF project by analysing climate data for improved early action triggers. Technical guidance is available both at HQ level for policy and strategy matters as well as at country level via a technical advisor. In addition, the RCCC fosters the exchange of climate scientists at the regional and global level also through its strong linkage with academic institutions. The RCCC was instrumental in developing the trigger for floods and will be consulted for its assessment when trigger conditions are being reached. The RCCC Technical Adviser based in Dhaka together with FFWC and BDRCS will monitor the forecasts and trigger conditions.

International Federation of the Red Cross Red Crescent (IFRC) As the umbrella organization, IFRC is involved in the broader strategic discussions of the FbF implementation and facilitates close cooperation within the Red Cross Red Crescent movement as well as with other humanitarian and governmental partners. Linkages with other ongoing humanitarian programmes and initiatives are established to ensure horizontal learning and avoid duplication of efforts. Ahead of an EAP activation, IFRC will be involved in preparatory steps such as the coordination and orientation of BDRCS Unit Offices and the Framework Agreement with suppliers. In addition, IFRC will coordinate the FbA by the DREF to ensure that funds are available for a potential activation. During the implementation of the EAP, IFRC will provide support and guidance for the procurement, financial and PMER processes at the field level. In the ‘Agreement of Cooperation’ between WFP, GRC, and SRC, IFRC signed as a witness. American Red Cross (AmCross) As a strategic partner of BDRCS, the American Red Cross has actively supported the FbF approach, especially in terms of research for quick beneficiary selection and cash distribution modalities. Through its ongoing projects, AmCross has continuously strengthened BDRCS’ implementation capacity and can quickly mobilise BDRCS volunteers and resources if necessary. In the Agreement of Cooperation between WFP, GRC, and SRC, the American Red Cross signed as a witness, highlighting its technical support to the FbF work in Bangladesh.

Swiss Red Cross (SRC) The Swiss Red Cross has actively contributed to BDRCS’ efforts to implement FbF for floods. SRC is implementing an FbF component as part of its DRM programming in Gaibandha and will support BDRCS during an EAP activation in Gaibandha. SRC continues to strengthen relationships with local level government counterparts and community-based disaster management systems in order to implement FbF quicker and more efficiently. SRC-BDRCS signed agreements with Union Parishads and Upazila Parishad of Fulchari in Gaibandha to strengthen the capacity of UDMCs and UzDMC on disaster risk management. The Swiss Red Cross is a party in the Agreement of Cooperation between WFP, GRC, and SRC that highlights the will to collaborate more closely on FbF and the flood EAP.

Government of Bangladesh’s Ministry of Disaster Management and Relief and Department of Disaster Management (DDM) Through its extensive humanitarian programming, BDRCS has already an excellent collaboration mechanism set up with the Ministry of Disaster Management and Relief (MoDMR) and the Department of Disaster Management which is strengthened through the advocacy for the FbF approach. In turn, MoDMR and DDM supports BDRCS’ EAP by granting access to government data on vulnerability and by discussing best approaches to implement early actions in a logistically feasible way. The FbF approach provides an opportunity for the MoDMR and DDM to adopt a more forward- looking humanitarian way of working and include a higher number of early action activities as part of

its response plan. During an activation, MoDMR and DDM will be informed of the early action implementation.

Bangladesh Post Office (BPO) Under the government’s Ministry of Post and Telecommunication, the Bangladesh Post Office provides Electronic Money Transfer Services (EMTS) via its wide network of post offices. It provides nationwide coverage and extensive experience in money transfers which make the BPO a reliable partner for the EAP’s early action of cash distribution. The BPO was actively involved in the flood EAP simulation during which the process of cash distribution was successfully tested. BDRCS/GRC has signed an agreement with BPO for EMTS under FbF. In addition, IFRC has signed a new agreement with BPO starting in September 2018 with a particular focus on Cox’s Bazar.

Flood Forecasting and Warning Centre (FFWC) Under the Bangladesh Water Development Board (BWDB) the Flood Forecasting and Warning Centre is situated to monitor and forecast the water level of 54 stations in Bangladesh. Along the Jamuna river basin, there are six FFWC stations. For this EAP, the 5-day deterministic forecast and the flood depth map provided by FFWC will be used.

World Food Programme (WFP) The World Food Programme is implementing the FbF approach for floods and has been part of a technical working group together with BDRCS and GRC to jointly advance the research on beneficiary selection and impact analysis as well as to align the FbF strategy with government counterparts. Evolving from the technical working group, an ‘Agreement of Cooperation’ was signed between WFP, GRC, and SRC, with BDRCS, AmCross, and IFRC as witnesses. The agreement outlines the parties’ will to collaborate more closely on FbF and the early action implementation of the EAP. WFP also remains a strategic partner especially for advocacy purposes at the national and regional level, and will support the EAP implementation during an activation in its pre-identified area in the Kurigram district. WFP is using the same trigger methodology as BDRCS/GRC which allows for a coordinated implementation of early actions.

3. Risk Analysis This section briefly describes the flood risk in Jamuna flood plains taking into account of the historical flood impacts, exposure, and vulnerability.

3.1 Hazard Selection The geographic location, land characteristics, multiplicity of rivers, and the monsoon climate make Bangladesh highly vulnerable to natural hazards. Flood is an annually recurring hazard with a larger impact- affecting lives, livelihoods and assets of poor and vulnerable populations.

Jamuna, Padma and Meghna are the major river systems in Bangladesh and this EAP covers flooding in Jamuna. People living along the Jamuna river are vulnerable to flooding; especially during the monsoon (June to September). Floods have the greatest impact on people who are living in low-lying areas, live below the poverty line, live in fragile houses and have a number of dependent family members to care for.

Using the global and national forecasts- Global Flood Awareness System (GLOFAS) and Flood Forecasting and Warning Centre (FFWC), floods in Jamuna river can be forecasted in advance and based on these forecasts early actions can be undertaken.

3.2 Past Impact In 1998, Bangladesh suffered one of the worst floods in living memory; affecting more than 68% area of the country. About one thousand people were killed, 30 million people rendered homeless (for about 2 months), substantial damage was accrued to crops and livestock, and outbreak of waterborne disease was high. The economic cost was also significant – public infrastructure damage was high, and factories were shut down. Initial estimates suggested that 1998-99 GDP growth dropped to 3.3 percent, from the 5.6 percent attained in 1997-1998.

In 2016 floods, five districts experienced the greatest impact. The National Disaster Response Coordination Committee (NDRCC), as of 6 August, estimated that almost 3.7 million people across 19 districts of Bangladesh were affected by the monsoon-induced flood.

Table 1: Historical analysis of flood impacts at different discharge and duration at Bahaduarabad River Jamuna. Description Flood Impact parameters

Flood year 1988 1998 2004 2007 2016 2017

Observed Discharge 68,700.00 102,535.00 96,105.52 71,860.11 76,554.92 78,262.42 (cubic meter per second- m3/s)

Duration of water 11.00 32.00 16.00 10.00 14.00 12.00 level above danger level (days)

Number of homes 7.20 0.98 4.00 1.10 0.25 0.62 fully/partly damaged (million)

People affected 45.00 31.00 36.00 14.00 3.70 6.7 (million)

In this historical impact analysis, the discharge data has been considered for Bahadurabad station. It is to be noted that in 1988 the flood was also induced from the water discharge of Ganges basin and hence damage was higher even though discharge was relatively lower than that of 1998. Moreover, the investment in increasing the structural, mitigation, and adaptation measures also contributed to lesser damage in subsequent years. The historical analysis suggests that damage starts at a discharge of 70,000 m3/s and is significant when the water level exceeds danger level for a minimum of 10 days.

Disaster impacts per sector Flood is a recurring hazard in Bangladesh, especially during the monsoon (June to September) that threatens human life, health and well- being. The direct impacts include personal security (deaths- by drowning, water-borne diseases, snakebites etc.), building and infrastructure damage/destruction (especially slum, mud huts, tin walled huts, public properties- schools, health facilities etc.), and loss of grains, crops and livestock. The economic cost of flood is huge; for instance, the Government of Bangladesh in 1999 estimated a total economic loss of 1998 floods at 2 billion USD- equivalent to 6% of the FY 1997/98 GDP.

Table 2: Flood impacts in different years by sectors Year- No. of No. of houses Crop damage No. of dead Road damage Embankment flood dead fully damaged (Acre) livestock fully (Km) damages events people

1986 57 196,803 990,573 42,374 3,094 13

1987 1,470 71,572 2,983,362 370,129 12,624 1,272

1988 1,621 1,151,189 1,119,998 398,018 47,042 1,718

1990 41 14,101 37,987 8,716 171 125

1991 818 340,043 1,220,225 46,306 1,692 649

1993 162 234,393 778,513 29,512 4,367 1,013

1994 10 19,177 55,325 8,666 60 18

1995 246 898,708 2,823,751 58,100 8,881 2,876

1996 76 218,275 404,456 47,946 1,635 448

1997 125 13,252 167,586 4,726 3,490 586

1998 918 980,571 1,423,320 26,564 15,927 4,528

2000 37 437,050 14,262 1,643 409 118

2002 26 115,511 321,355 25,237 3,720 4,734

2003 104 109,147 373,376 7,197 2,019 1,535

2004 747 894,954 1,605,958 15,143 14,271 3,158

2007 970 81,817 890,898 1,459 3,705 88

2016 42 NA 250,000 NA NA 58

2017 121 71,628 548,175 NA 403 56

Source: Department of Disaster Management (DDM).

3.3 Exposure People living in the char islands in the middle of the river, outside the embankment protection, and high erosion risk areas are most exposed to flood risks. People- especially children, elderly and person with disability, who cannot swim, houses- fragile houses- mud houses, tin walled houses, and livestock living in these areas are the exposed elements. For further information on these regions, please see Appendix A on risk analysis. The populated unions in these three regions are shown in map 1.

Map 1: Exposed elements in the Jamuna river basins

3.4 Vulnerability Three elements of vulnerability- socio-economic vulnerability, topography and physical infrastructure- will be considered in this EAP. The vulnerability elements, sources of data, and updating frequency is presented in table 3.

Table 3: Vulnerability indicators, data source, data scale and updating frequency Category Sub-category Data Source Data Scale Year Frequency

Socio-Economic Poverty Line BBS / WB Upazila 2010 5 years Vulnerability Dependent BBS Census Union 2011 10 years population (> 60 and <5 yrs.)

Topography Land type water depth from 2017 Yearly FFWC model

Physical House Type BBS Census Union 2010 10 years Infrastructure

First, socioeconomic indicators in these regions are used as indicators of vulnerability to flooding. Poverty is a significant contributor to people’s vulnerability to flooding and frequent flood impacts lead to an increase in poverty and hence vulnerability. It has been observed that poor people mostly live closer to the rivers, lack adequate means to take protective measures, and have very limited capacity to cope with the loss of property and income. Map 2 shows percentage of population living below poverty line along Jamuna. It is also assumed that areas with a high percentage of elderly and children (dependent population) require more assistance to reduce the risk of flooding and prepare for flooding. Map 3 shows the percentage of dependent population along Jamuna.

Map 2: Percentage of population below poverty Map 3: Percentage of dependent population in line in Jamuna flood plain (Source: BBS/WB, Jamuna flood plain (BBS, 2011) 2010)

Second, we assume that people living in the low-lying areas will experience higher damages for the same river level. Therefore, houses in low-lying areas are more vulnerable to the same river water height. Map 4 shows the topography of Jamuna flood plain.

Third, we assume that people living in less sturdy houses are more vulnerable to flooding. We calculate a relative measure of vulnerability by assigning a value of 100% to the district with the maximum number of fragile houses (Jhupri and Kutcha) and then normalized this for all other districts relative to that district. Map 5 shows the composite score of house types.

Map 4: Topography of Jamuna flood plain Map 5: Composite score for house type structure Combining all three of these indicators, a composite vulnerability map can be seen in map 6. For more information, please see Annex A.

Map 6: Composite Vulnerability map of different unions in Jamuna flood plain area

3.5 Prioritized impact Vulnerable households in the Jamuna basin might lose their lives, houses, assets, food grains, and income from livestock because of the lack resources and means for timely and efficient evacuation. They are also forced to skip meals, reduce portion size and eat lower quality food. Hence, this EAP, has prioritized impact on peoples’ lives, household assets, and livelihoods linked to livestock. See 5.1 for detail.

4. Trigger Model (Impact Forecasting model) The Trigger model will address two things: 1. When to activate 2. Where to activate

To identify, ‘When to activate’, at first, the available Global and National level forecasting systems have been evaluated taking into account the type of forecasting system, content and source of the system, lead time of the forecast and the skill and error of forecasting. Based on this evaluation particular forecasting systems have been selected for trigger activation.

Based on the historical data, thresholds have been identified for the parameters- discharge, water level, duration, and flood depth- so as to decide when the hazard can be declared as an extreme event. In the meantime, impact level has been identified for different discharge and flood depths. As the historical data has been developed at a national level, Key Informant Interviews (KII) and expert consultations were carried out to understand the impact for the Jamuna flood plain area.

Then from the forecasted water level, the impact level has been assessed. When this impact crosses the limit the trigger will be activated. The method has been summarized and presented in figure 1.

[Q = Flow in volume (m3) per unit time (s) at river cross section, WL = Water level]

Figure 1: Trigger model (when to activate)

The following sections will describe the process for prioritizing areas for EAP activation (i.e. where to activate).

4.1 Inventory of Forecasts and selection of forecasts We analyzed several global and national forecasts for use in this region. Please refer table 4 and Annex B. This EAP uses a global forecast model- GLOFAS- for pre-activation and FFWC’s deterministic model for activation of early actions.

GLOFAS is a global hydrological forecast and monitoring system that couples weather forecast with a hydrological model and it is calibrated for the Brahmaputra/Jamuna river. It issues probabilistic flood forecasts more than 15 days in advance. However, this model does not ingest local data in making its forecasts. For pre-activation trigger, GloFAS 10 days lead time forecast will be used. In the recent 2019 floods, in 10-day lead time flood forecasting system the probability did not exceed 70%, but was above 50%. So, for pre-activation trigger the probability of 50% will be considered- to get ready in the field.

The FFWC deterministic flood model uses rainfall data upstream. FFWC receives the near-real time rainfall data and forecasts from BMD and BMD processed their own data and also from different sources like the ECMWF, WRF and IMD. The FFWC 5 days deterministic forecast has been proposed for activation of EAP.

There are also other national forecasts available for 6 locations along the Jamuna river from the Flood Forecasting and Warning Centre (FFWC). However, the 5-day deterministic forecast is considered the most reliable. It was evaluated for the government established Danger Level at the Sariakandi station

and the result showed that the forecast successfully predicted 67% of floods and had only 29% of false alarms at the 5-day lead time. Recent flood forecasting accuracy assessment shows about 15 cm error at the 5-day lead time.

Table 4: Forecast type, lead time and skill of different model Forecast Sources/availability Forecast type Lead time Skill

GLOFAS – http://www.globalfloods.eu/ Probabilistic 1-15 days 10 days Global glofas-forecasting/ Discharge and forecast R2: return period on 0.760; 15 the river days forecast Brahmaputra(Ja R2: 0.610 muna) at (limited data Bahadurabad – 1 year)

FFWC http://www.ffwc.gov.bd/ind Deterministic 1-5 days 5 days National ex.php/hydrograph/forecast Water level forecast R2: short term forecast at 52 0.72 deterministic locations FAR (5days) forecast 0.29

FFWC http://www.ffwc.gov.bd/ind Probabilistic 1-10 days 10 days National ex.php/hydrograph/medium Water level R2 = 0.61 short term -range-1-10-days-forecast forecast – mean (FFWC probabilistic +sd, mean and annual forecast mean – SD report 2016)

R2 = Regression coefficient of observe vs forecast curve. Higher value refers to good acceptability of forecast FAR = False Alarm Ratio, higher value refers to less acceptability of forecast

4.3 Exposure map The people living in the char island and outside embankment protection are exposed to flood, especially the children, elderly and persons with disabilities. Fragile houses (Kuchha and Jhupri) are disproportionately exposed. See section 3.3

4.4 Vulnerability Index Socio-economic vulnerability (poverty and dependent population), topography (land type), house types are key vulnerability indicators used to arrive at the vulnerability index. Although this EAP aims at protecting livestock and assets of the most vulnerable families, it was not possible to find a direct vulnerability indicators with enough quality in relation to livestock and assets, therefore the selected vulnerability indicators for the trigger are proxy indicators of vulnerability, that will help to identify the key communities and households where early action will take place. See section 3.4

4.4 Definition of the Impact Level The government Danger Level is set at the level of agricultural fields in the region. This corresponds to a return period of approximately 1 in 5 years. Livestock sheds and people’s houses are usually built at higher levels than the agricultural fields. Houses are usually about 1 meter higher than the agricultural fields and livestock sheds are also temporarily raised during the flooding season. Flooding to people’s houses happens at a return period of approximately 1 in 10 years. To define the extreme event, the FFWC forecasting error which is 15 cm has been considered. Hence, the flood exceed the

water level DL + 0.85 meter (at Bahadurabad) or discharge 84, 000 m3/s is considered as the extreme flood event for this FbF trigger.

4.5 Justification of the Impact Level Historical analysis of national data on flood impact shows that floods have affected fewer people over time, due to increasing resilience and protective measures.

Figure 2 presents the impact of different flood heights on household assets in each of the exposure areas- char island, outside embankment protection and erosion risk areas; based on the key informant interview.

Figure 2: Impact on household assets at different flood depths (Source: Survey and KII)

This curve maybe used to compute the impact on household assets based on forecast water depth in FbF trigger and the damage limits 25%.

Besides, another curve has been generated with affected population against flood depth as shown in Figure 3. The figure shows that, when the flood depth crosses the limit, about 40% people living in the area will be affected and this may be used as an impact limit in the trigger.

Figure 3: Affected population by flood depth

As shown in the figure 4, people reported greater losses as soon as the plinth of their house was flooded, and losses were greater for longer durations of floods. So, considering this fact, the extreme event of DL+ 0.85 meter at least for 3 days duration is justified- as soon as the water level crosses 0.85 meter above the DL (plinth level), damage increases rapidly.

Figure 4: Household asset damage at different durations (Source: Survey and KII)

To make the best use of available forecasts and increase time to prepare for activation, we propose a two-step trigger using both the global and the local models, and using the house structures (plinth at DL + 0.85 meter inundation) as the level of impact warranting intervention. See Annex B for further details.

Stage I: Pre-activation trigger – Get ready with the resource and identify the beneficiaries likely to be impacted based on the forecasted exposed area. Pre-activation Trigger will be reached when flow at Bahadurabad (forecasted by GloFAS with lead time 10 days) likely (50% or above) to cross the 1 in 10- year return period and flowing over at least for three days. This corresponds to an impact forecast of 3.7 million population will be affected and about 1million household will be affected.

Stage II: Activation trigger – Activate EAP in the vulnerable households within forecasted exposed area. Trigger will be reached when forecast water level by FFWC at Bahadurabad (5 days lead-time) shows to cross the DL+0.85m and remain above for 3 days. This corresponds to an impact forecast of loss of household assets over 25% and affected population over 40%.

4.6 Intervention Map To identify where to activate, as shown in figure 5, at the first stage, the unions will be identified where the impact crosses the limit. Following steps are undertaken to identify the unions where impact crosses the limit: 1) Flood Forecast – flood depth map – union wise average water depth 2) Apply affected population based on flood depth (map 7) 3) Impacted HH asset based on flood depth (map 7) Thereafter, vulnerability index layer will be overlaid to identify and rank the unions with potential largest impacts. The process methodology is illustrated in figure 5 and a dummy result presented in map 7.

Figure 5: Methodology to identify ‘Where to activate’

Average forecast flood depth (in Affected population Impacted HH asset meters) limit 40% limit 25% Map 7: Forecasted water depth and Impacts on population and HH asset

Following the above motioned process, the unions crossing the impact limit (i.e. over 25% HH asset damage and/or more than 40% affected population) ranked by the vulnerability index will allow prioritization of areas and allocation decisions based on this model. Map 8 shows priority of FbF investment unions in the Jamuna flood plain.

Map 8: Prioritization of areas EAP implementation 5. Selection of Actions This section briefly describes the flood impacts this EAP will address. In doing so, a brief description has been provided on the early actions, their selection, theory of change, usefulness of the early actions in case of non-occurrence of the hazard, beneficiary selection criteria and feasibility of the early actions.

5.1 Prioritized impacts As discussed in the above sections, people living in Char islands and in areas outside the embankment protection are disproportionately exposed to flood risks and flood impact in those areas is further exacerbated by poverty, dependency and housing types. The major flood impacts experienced by the vulnerable communities are: human causality (deaths due to drowning, diseases e.g. diarrhea, dysentery, cholera and snakebites ref), damage to house structures and household assets, loss of food grains and livestock. Lack of resources for evacuation was the key issue raised by communities, which not only posed a threat to life but also led to damage of household assets and loss to livestock that could have been prevented should there have been adequate evacuation measures. Hence, this EAP has prioritized three impacts:

● Human casualty (deaths due to drowning and others) ● Loss of household assets and food grains, ● Loss of livelihood linked to livestock 5.2 Summary of selection process By looking at the prioritized impacts, the impact survey in the flood districts, and an analysis of the expected impact/risk reduction, as well as through joint discussions with key actors including, BDRCS, GRC, WFP, SRC, IFRC, RCCC and CWG, the following early actions were selected:

● Distribution of unconditional cash grants at household level ● Provision of boat evacuation

BDRCS/GRC conducted an impact survey in the flood prone areas of Tangail, Gaibandha and Kurigram districts (307 households covering three different scenarios- char island, mainland outside embankment and erosion prone area) to determine the linkages between flood levels, humanitarian impact, and potential early actions. As shown in figure 4 the findings indicated that the majority of the population want to evacuate during the flood but they either do not have enough money to afford for evacuation transportation or there is no transportation available at all. Out of the 307 sample households, 48% of respondents cited lack of money preventing them from evacuating, while 34% reported that no transportation for evacuation was available during the flood event. The proposed early actions- unconditional cash and boat evacuation- will address both the issues, lack of sufficient money for evacuation as well as the lack of evacuation transportation during the flood event.

Figure 6: Problems faced by households during evacuation (Source: Survey and KII)

When people do not evacuate, they are exposed to casualty and lose their household assets, food grains, and livestock-one of their major sources of livelihood. During a flood event, the household assets and food grains will get washed away or damaged while livestock will suffer from sickness or shortage of fodder or get drowned or washed away. The unconditional cash grant will allow beneficiaries to take appropriate measures to secure and/or move their household assets. At the same time, with the unconditional cash they will be able to evacuate themselves, their movable assets, food grains, and their livestock. When asked what potential early actions would be helpful ahead of a flood event, as shown in table 5, 38% of respondents mentioned that a cash grant would allow them to take early actions while 24% responded that the provision of transportation for evacuation would reduce the impact.

Table 5: Suggested early actions to reduce the problems faced during flooding Early action Percentage of respondents No of respondents

Cash grant 38% 128

Transportation 24% 81

Early warning 22% 75

Shelter 16% 54 Source: Survey and KII

During the Asia Pacific Regional FbF M&E workshop by the Climate Centre in Hanoi in June 2018, BDRCS/GRC developed a theory of change for two priority impacts: loss of livelihood linked to livestock

and loss of household assets. The two priority impacts were previously identified in discussions with BDRCS field level officers as well as beneficiaries.

Please find here a more detailed ToC for the loss of livelihood linked to livestock: https://drive.google.com/open?id=1s301wkTYalLVaC0l2uYoYW-bMjw5FLGV

Theory of change

Reduce the impact of floods on most vulnerable HH in Jamuna river basin

Improved Reduce Reduce health, loss of loss of HH safety and livestock assets food security of families

Availability of local Availability Availability of transport local transport of local and place Market is and place to Market is to Markets is transport evacuate functional evacuate functional and place functional to Protect HH Evacuation Buy food and Evacuation Buy fodder Dismantle of affected medicine for of livestock and and assets HHs family medicines for transport (furniture, members livestock HH assets grains and others

ACTIVATION Evacuation MPCG 5 days lead support time

Beneficiary PRE ACTIVATION selection 10 days lead time

Identification Agreement Orientation to Training of ferry with financial BDRCS (NDRT, UDRT, PREPAREDNESS service service units/volunteers others) providers providers

LEGEND

Goal

Outcome Activities

Assumptions

Outputs

Table 6: Theory of Change Long term Available capacities to Risk Action Outputs Short term outcomes outcomes implement If the trigger ● Agreement with Compared to the is reached electronic money Up to comparison group, over a ten- Poor transfer services for Provide 3,300 beneficiaries report that year period, households in quick and secured unconditional households the loss of their lives, compared to the Jamuna distribution of cash to cash grants to in the household structure, control basin might the beneficiaries. the most Jamuna assets, food grains, and group lose their ● BDRCS Unit Office at vulnerable basin area income from livestock beneficiaries lives, houses, district level have households in will receive have been minimized by report: more assets, food trained volunteers the exposed BDT 4,500 timely and efficient assets, less grains, and ● Trained BDRCS/ areas in immediate evacuation. Moreover, debts, better income from IFRC/GRC staffs at Jamuna basin before the they avoid negative food livestock Dhaka and Unit Office before the flood for coping mechanisms and security; and because of the ● FbF and EAP have been flood. their early are able to eat better may have lack resources introduced in actions. quality and quantity of more and means for NDRT/NDWRT Training food. productive timely and capacities. efficient About evacuation. 1,500 HH They are also Provide As compared to the with their forced to skip support for comparison group (CG), livestock meals, reduce evacuation of the intervention group and portion size people, (IG) lost less household moveable and eat lower livestock, and structure household will be quality food. moveable assets and food grains evacuated assets by efficient evacuation. to the safe place.

Unconditional cash grant: Justification for the amount of the cash grant

Beneficiaries will receive BDT 4500 as an unconditional cash grant. In consultation with BDRCS, the Cash Working Group, SRC and WFP the amount of BDT 4500 was decided based on several criteria:

Taking the amount of FbF’s first phase as a basis for discussion (BDT 5000, implemented by BDRCS/GRC from 2015-2017), the amount was reviewed and discussed with the Cash Working Group. According to the Multi-Purpose Cash Grant (MPCG) guideline for Bangladesh, the cash grant package for a single household is BDT 4500 for a month. The package considers a household’s average daily wage and food basket requirements over a month. BDRCS, GRC, SRC, and WFP agreed to use the same amount, being aware that households have a much shorter time to use the cash grant. Yet, households also face higher investment costs for the early actions, such as additional labour costs for moving shelters and livestock, or higher transportation, and food expenditures for the evacuation period of 3-5 weeks. The cash grant is designed to allow beneficiaries to strengthen their resilience while maintaining their daily household needs.

In addition, since the amount follows existing guidelines it facilitates coordination and acceptance among CTP implementing partners. More importantly, it will allow BDRCS/IFRC/GRC/SRC and WFP to jointly implement an EAP activation and complement each other’s intervention without creating misunderstandings among beneficiaries.

Looking at the previous cash distribution with BDT 5000 as part of an FbF activation in July 2017 (see also 5.4 Evidence Base), BDRCS/GRC achieved good results in reducing the flood impact with the chosen amount. By lowering the amount slightly by BDT 500, more beneficiaries can be reached during an EAP activation while maintaining the same level of impact reduction.

Boat evacuation Boat evacuation will start once the cash distribution is completed and will continue until flood reaches its peak (3-0 days before the peak flood). BDRCS and GRC undertook a preliminary assessment on feasibility of the boat evacuation and it was learned that big boats (ferries) are already available in the areas which can be mobilized for evacuation. If needed, the existing ferry service providers can hire additional boats. For the provision of evacuation transportation via boat the location and practical considerations will be discussed with the respective BDRCS Unit Offices as well as local government units such as UDMCs.

5.3. Usefulness of actions in case of non-occurring event The two-stage trigger as well as the stop mechanism will reduce the risk of acting in vain. The pre- activation trigger (based on GLOFAS) does not entail physical movement of beneficiaries, their assets and does not warrant cash distribution unless affirmed by the activation trigger. In case of the early actions being implemented but the extreme flood event not occurring, the early actions will still help reduce the impact because a flood event is still occurring, simply at less extreme levels. The provided cash and evacuation transportation by boat will help beneficiaries to reduce the flood impact. In case of the cash grant, beneficiaries will be informed about suggestive usages to help them make the best use of it.

5.4. Evidence Base RCCC together with BDRCS/GRC conducted a post-distribution analysis of the cash grant provided as part of the July 2017 activation. The assessment shows that the cash grants contributed to greater household food security, less psychosocial stress during and after the flood period, and a reduction in high-interest debt accrual of vulnerable households, compared to a control group of similarly vulnerable and flood-affected communities that did not receive the forecast-based cash assistance. The intervention may also have prevented households from being forced to make destitution sales of valuable assets, as indicated by qualitative data collected in July. Please find the analysis here.

During the Asian Ministerial Conference on DRR (AMCDRR) in Mongolia in July 2018, the findings were presented as a case study at a thematic event. Please find the case study here.

5.5 Beneficiary selection process and criteria The beneficiary data collection starts with the pre-activation trigger which is 10 days before the flood event. At this stage the unions with the potentially highest impact are identified yet cover a larger area of potentially affected people than this EAP can address. This means that a higher number of beneficiary data is being collected which will be narrowed down when the second trigger is reached, which is 5 days before the flood event. To narrow down beneficiary data, a set of criteria was developed to identify the most vulnerable population (see below).

Logistically, BDRCS has sufficient capacity to carry out large-scale beneficiary identification activities. BDRCS has 68 Unit Offices with around 50 to a few hundred active volunteers in each unit including the units in the Brahmaputra/Jamuna river basin area. Considering that the EAP will be activated in 2- 3 districts most vulnerable and affected districts and each volunteer can collect data of around 15 beneficiaries per day. The volunteers will be using ODK (Open Data Kit) software on smartphones to collect beneficiary information. The collected data can be analyzed centrally and digitally. Taking BDRCS’ capacity into account, data of around 4,500 households can be collected in 02 (two) days which then will be narrowed down to 3,650 for the cash distribution. Depending on the location of the forecasted impact and on the BDRCS Unit capacity, the total number of beneficiaries can be less. Please see also 5.6 for the beneficiary selection criteria as tested during a simulation.

After the flood simulation, the beneficiary selection criteria were revised based on the learnings and in collaboration with the RCYs. Final set of beneficiary selection criteria can be found here.

Beneficiary evacuation with Boat: ● Evacuation support of the people with their livestock and movable assets will be provided as per requirement. ● These beneficiaries may be the same who received cash but not exclusively. ● Beneficiary selection will not be undertaken for the evacuation support. People living in the char island whoever need to evacuate can avail this support. However, people who didn’t receive cash will be prioritized as the cash grant beneficiaries have the means of hiring boat. The members of the UDMC (Union Disaster Management Committee) will be involved in the process to avoid any potential conflict. ● Big boat already available in the areas which are being used as ferry.

5.6 Feasibility With BDRCS’ long-standing experience in disaster relief operation, they have demonstrated the staff and logistical capacity to implement large-scale humanitarian programmes in a short period of time. Similarly, for the current ‘Population Movement Operation’ in Cox’s Bazar, BDRCS shows its capacity to deliver and coordinate humanitarian aid within short notice. Section 9 also highlights BDRCS’ capacity as indicated in its involvement in strategic logistical processes of this EAP.

To test the feasibility of the quick beneficiary identification process and the cash distribution through the Bangladesh Post Office, BDRCS/GRC together with the financial support of the American and Swiss Red Cross conducted a simulation. It was carried out in a condensed form within 5 days and in 3 different districts. Within 2 days a total of 2580 beneficiary data was collected and then narrowed down to 1700 beneficiaries out of which 1617 received a cash grant. The simulation successfully proved that BDRCS has the capacity and process in place to identify a large number of beneficiaries in an efficient and quick way. The simulation report with more details can be found here.

6. Monitoring, Evaluation, Accountability and Learning (MEAL) BDRCS has a Planning, Monitoring, Evaluation and Reporting (PMER) unit with 4 PMER staffs under the Planning and Development Department and the FbF MEAL system links with that PMER unit. This unit has the capacity of project planning, monitoring and evaluation. They are also responsible for reporting and capacity building of different departments of BDRCS. The unit also develops the mid- term and annual reports of BDRCS, contributing to project development, designing project-based M&E plans, and conducting capacity assessment of branch offices. One GRC supported PMER Officer in BDRCS will provide full time support in all MEAL activities at field level from GRC and BDRCS end along with the support from other PMER officers in BDRCS. In addition, IFRC’s PMER unit will be involved in the MEAL aspect of this EAP. The respective IFRC staff will be involved in any preparatory activities and deployed during an activation to lead the MEAL team and collect the relevant data. The BDRCS and IFRC PMER department will be engaged in the whole MEAL process. MEAL data will be collected in the intervention community as well as another community having similar geographical location and vulnerability (comparison community) to assess the result given in the ToC. The MEAL process will again be conducted before the next flood season (around 4-6 months) to assess the impact.

Implementation of the early actions will follow BDRCS’s Community Engagement and Accountability (CEA) guidelines. BDRCS CEA team will also be involved in the overall process to ensure beneficiary participation and accountability. In order to ensure transparency, the beneficiary list will be made available for the community for validation. In addition, a hotline number will be made available for the beneficiaries to put their complaints and comments. The beneficiary selection process will be

objective and will make use of vulnerability criteria like single women, lactating mothers, pregnant women, women headed families, families with disabled members etc. For the ease of community, the distribution point will organized as close to the communities as possible.

CEA and Protection, Gender and Inclusion (PGI) budget will be used to conduct beneficiary satisfaction survey and exit survey with the beneficiaries. PGI will ensure separate sitting arrangement, breastfeeding room, and proper washroom facilities for the beneficiaries in the distribution points. BDRCS CEA official will be involved in the whole process to support in monitoring CEA activities. CEA officials will communicate with program people after receiving any feedback or complaint from any beneficiary or community and will reach back the beneficiary with proper answer and justification of their query or complain.

6.1 Monitoring the EAP implementation FbF team will coordinate with the BDRCS DRM department to monitor the field level activity after the trigger being reached with RCY volunteers under close coordination with the ULO of the branch office. This will entail activity monitoring and Post Distribution Monitoring (PDM). BDRCS’ and IFRC’s PMER units will coordinate the overall monitoring process of the EAP activation with support from BDRCS DRM.

Table 7: Activity monitoring and responsibilities Trigger reached

Activity Responsibility

1 BDRCS M&E volunteer to be deployed to each response area for BDRCS M&E volunteer cash distribution BDRCS/IFRC PMER team 1 BDRFS/IFRC PMER team to be deployed to one cash distribution site

1 BDRCS staff to track EAP implementation practice at Dhaka level Project Coordinator, FbF, BDRCS-GRC

Early action implementation starts

● Monitoring and analysis of implementation activities BDRCS M&E volunteers according to pre-agreed indicators, BDRCS/IFRC PMER team ● verifying the number of people served by cash and document the process for monitoring purpose ● Monitoring distribution site setup and accountability ● Internal process evaluation ● M&E data is collected via smartphones

6.2 Trigger Evaluation The trigger model will be evaluated at the technical and implementation level: ● An EAP Activation committee consisting of members from BDRCS, IFRC, GRC, SRC, AmRC, RCCC will jointly review the trigger thresholds and activation in terms of lead time, accuracy and feasibility. The methodology committee that is led by FFWC can be a good forum for this review. ● At the implementation level, BDRCS, GRC, IFRC, SRC, AmRC will jointly evaluate the trigger’s thresholds and accuracy from an implementation point of view.

6.3 Evaluation of impact of early actions after activation BDRCS M&E volunteers and IFRC PMER officer will be involved in impact evaluations after the flood event that include sample surveys, and focus group discussions. GRC will support BDRCS and IFRC on developing questionnaire and checklists to reflect the changes in the community due to FbF intervention. The survey will be designed jointly by BDRCS, IFRC and GRC with technical support provided by GRC, and will make use of smartphones to collect data on the Open Data Kit (ODK) platform. The BDRCS M&E volunteers will collect data on the ground while IFRC PMER will take the lead in reporting the findings. Results will be shared after initial analysis in a joint meeting with all relevant partners.

For the evaluation of impact after activation, BDRCS volunteers in support from the ULO and unit office will collect data from FbF intervention area and also outside intervention area. A comparative analysis between the two areas on people receiving cash grant, mortality of livestock, casualty and others will be done using impact data collected by BDRCS volunteers. BDRCS PMER unit and FbF team will work together to develop a comprehensive format of data collection for evaluation reporting. The analysis report of FbF beneficiary community and other community of same geographical location will give a comparison to evaluate the impact of FbF intervention after activation.

6.4 Identification and Documentation of Lessons Learned Once all data is compiled and collated, a lessons learned workshop will be held to analyze the results and decide on next steps. The workshop is not internal and participants will include BDRCS, AmCross, SRC, GRC, RCCC, BMD, DDM, and IFRC. It will be facilitated by BDRCS and GRC. As an output, the lessons learnt workshop should deliver an outcome report as well as concrete next steps in terms of adapting the EAP and trigger. IFRC will lead the reporting activities in order to foster transparency among partners and as a means of serving internal monitoring mechanisms.

7. HR and General Capacity This section briefly describes the National Society’s capacity to implement the EAP and undertake monitoring and evaluation as described above.

7.1 National Society Early Action Capacity The Bangladesh Red Crescent Society has extensive experience and a strong operational capacity in implementing relief operations related to natural hazards. As such BDRCS is in a strong position to implement FbF early actions, drawing on its existing institutional structures and administrative processes. BDRCS has 55 RDRT members, 11 PECT training holders, 180 NDRT, 90 NDWRT members that can be deployed within a short time. In addition, BDRCS has an extensive volunteer network of more than 720,800 Red Crescent Youth (RCY) members, 5000 RCY volunteers. 1500 Sr. volunteers, 81,804 life members along with significant numbers of community volunteers, which play a vital role in distribution and other humanitarian operations. For the EAP activation, specially trained BDRCS volunteers and NDRT members will be selected to implement the early actions. They have received training on the FbF project as well as specialized trainings such as First Aid, and Search & Rescue.

7.2. Thematic capacity Through its past and ongoing humanitarian responses, BDRCS has sufficient thematic experience and capacity for this EAP. In particular, for cash transfer and evacuation support:

Evacuation: For Cyclone Mora, BDRCS activated around 55,515 Cyclone Preparedness Program (CPP) volunteers. Besides, the District Branches of BDRCS in Cox’s Bazar, Chittagong and other coastal districts mobilized RCY volunteers to disseminate early warning messages, evacuating vulnerable people to shelter house along with CPP volunteers and local administration.

Cash transfer: BDRCS has a vast experience of cash based interventions in emergency as well as long term/development programs. Table 8 presents BDRCS’s experience in cash transfer few in emergency operation with different modalities. As shown in the table, since 2007 BDRCS has reached 104,180 households with CHF 6,550,002.00.

BDRCS Strengths in Cash distribution: ● Approved Cash SOP (Standard Operating Procedure) ● Memorandum of Agreement with Bangladesh Post Office for Money Transfer using Mobile Phone ● In country RCRC Cash Working Group is functioning, also members of national CWG ● 12+ years of experience in cash transfer? ● 11 PECT (Practical Emergency Cash Transfer Programming) trained Personnel ● 85 CBI (Cash based Intervention) Level-2 trained staffs and volunteers ● 150 CBI Level-1 trained staffs and volunteers (3 days training in Country)

Table 8: BDRCS’s experience in cash transfer (Emergency operation) Name of the Year # of HHs Amount in CHF Modalities Mechanism Operation Cyclone Sidr 2007 6,250 685,976 CCG Cash in Bank operation Cyclone Aila 2009 1,643 159,158 CCG Cash in Bank operation Monsoon Floods Cash in Envelope & Cash 2011 1,850 231,707 CCG & UCG operation through Bank Floods and Landslides Cash in Envelope & Cash 2012 22,500 1,055,585 CCG & UCG operation through Bank Tropical Cyclone Cash in Envelope & Cash 2013 6,100 234,146 CCG & UCG Mahasen operation through Bank Cash in Envelope & Cash 2014 Flood operation-2014 9,000 445,000 CCG & UCG through Bank Cash in Envelope, Cash Cyclone Komen 2015 4,000 341,463 UCG through Bank & Mobile Operation-2015 Money Transfer Cyclone Roanu Cash in Envelope & Cash 2016 10,994 423,293 CCG & UCG Operation through Bank Flood operation – Cash in Envelope & Cash 2016 1,470 283,171 UCG, CCG 2016 through Bank Electronic Cash Transfer Flood-2016 (Forecast 2016 1,729 105,210 UCG through Post Office using based Financing) Mobile Phone Cyclone Mora Cash in Envelope & Cash 2017 3,000 1,060,976 CCG & UCG Operation through Bank Floods-2017 2017 25,000 657,098 UCG Cash in Envelope operation Population 2017 6,750 196,646 UCG Cash in Envelope movement operation Electronic Cash Transfer 2017 Cyclone Mora-2017 2,255 139,397 UCG through Post Office using Mobile Phone Electronic Cash Transfer Floods-2017 (Forecast 2017 1,039 60,937 UCG through Post Office using based Financing) Mobile Phone

Cash through bank 2018 Flood 2017 Recovery 600 321,429 CCG account transfer Cash through bank 2018 Flood 2017 Recovery 500 148,810 CCG account transfer Grand Total 1,04,180 6,550,002 * CCG: Conditional Cash Gant, UCG: Unconditional Cash Grant

7.3 Strategies and plans The Bangladesh Red Crescent Society has a Strategic Plan (SP) for 2017-2020. Under Goal 1 (Strengthened preparedness, response, and recovery services in reducing impacts of disasters, emergencies, and other humanitarian consequences), it is clearly mentioned that BDRCS will be responsible for ensuring forecast based early actions to enhance preparedness and response. BDRCS also has a Flood Contingency Plan for quick implementation of humanitarian actions during times of emergencies. The contingency plan provides clear guidelines on how to respond to flood disasters. Within 24 hours BDRCS will establish a control room and start preparations for deploying emergency teams for rescue, evacuation and first aid activities.

7.4. Financial capacity to advance funds BDRCS NHQ as well as unit offices have some capacity to advance funds for the short duration.

8. Finance and Logistics The total estimated budget for this EAP is CHF 249,862. Details can be found here.

8.1. Budget BDRCS/GRC has the experience of two activation for floods. The budget has been developed based on the experience of two activations and BDRCS and GRC’s other experiences on the same area.

The amount of cash for each household has been decided with consultation with the Bangladesh Cash Working Group as well as the other actors of the FbF e.g. WFP.

8.2 Readiness cost The readiness cost, which stands at CHF 15,059, is planned for some workshops with the RCYs to orient them on the EAP and associated travels. This cost will be managed by relevant Department in BDRCS.

8.3 Activation cost The EAP activations cost is estimated at CHF 234,803 and it includes costs for beneficiary selection, cash transfer, evacuation and associated travels. The EAP will be activated by BDRCS. However, the fund flow mechanism is yet to be sorted out.

8.4 Prepositioning Since cash transfer and evacuation support are the key early actions envisaged, this EAP doesn’t envisage any pre-positioning to be done.

9. EAP Activation Process 9.1. Preparatory measures

Please find all relevant information here.

9.2 Communication protocol (this could be an annex) Please find the communication protocol here.

9.3 Stop Mechanism The stop mechanism occurs 5 days ahead of the flood event when more accurate forecasts are available. If trigger conditions are not reached at that point, the EAP activation will be stopped. Already incurred costs are related to the beneficiary data collection. No items or money will have been distributed before the stop mechanism.

9.4 Trigger activation system Since an automatic system to monitor the forecast does not exist, the RCCC technical advisor together with BDRCS designated staff (project coordinator) will regularly monitor the forecast during the monsoon (June to September) and keep the BDRCS DRM Director updated. In an event of pre- activation, the BDRCS DRM Director will call EAP activation committee meeting, which is chaired by BDRCS DSG. Based on the information provided by the RCCC technical advisor, the activation committee will activate the EAP (pre-activation). Thereafter, EAP implementation committee will meet and preparation of field activities will begin. Similar activities be carried out for the Activation trigger. Proof of trigger condition is crucial to activate the pre-activation and activation process. Please find the activation process here.

9.5 Implementation process Please find all relevant information here.

10. EAP endorsement BDRCS DRM department has worked closely on the EAP development and will push for the internal endorsement, final endorsement by the SG is likely to take some time. The endorsement with involved actors is in process. An EAP validation and endorsement meeting was organized in 2019.

11. Conclusion This EAP and the identified early actions will effectively reduce the prioritized impacts of the flood hazard, targeting up to 3,300 households through unconditional cash grants. With the two-stage trigger and a lead time of 10 days, the pre-agreed early actions will allow BDRCS and implementing partners to reduce the suffering of people and maintain their livelihoods by ensuring that their movable assets and their livestock related livelihood remain safe. Through the impact-based forecast the people highest impacted will be targeted, allowing BDRCS and implementing partners to bring humanitarian early actions to those who need it most. Thanks to BDRCS’ extensive experience in responding to the yearly flood events, the capacity to act immediately exists on the ground. The M&E system that will accompany the EAP ensures that learnings can be captured which will feed into the updating of the flood EAP.

Sustainability of this intervention? Streamlining through the IFRC DREF? As mentioned above, some 3,300 households will be reached through FBA by DREF. WFP, using the same trigger and thresholds, will cover additional households in Kurigram district. The Red Cross Red Crescent Movement together with WFP and other actors (like Care Bangladesh and Start Fund) are now advocating with the Government to operationalize FbF into the Government’s DRM sector as it is already included in the Government’s Standing Order on Disasters (SODs).

Annex A: Detailed Risk analysis The risk profile has been prepared evaluating the exposure and the vulnerability. This step will identify who is vulnerable in the Jamuna flood plain area of Bangladesh. The FbF team has gathered data for six criteria so as to best capture the range of elements that inform vulnerability. Team has designed a composite index through a combination of exposure and vulnerability. Unlike the exposure, the vulnerability has been grouped into three categories: topographical vulnerable, socio-economic vulnerable and physical infrastructure vulnerable. At the first stage, the exposure has been considered with two parameters: i) the island ii) the area out of embankment protection and not an island. And then vulnerability index has been identified using land type, poverty line, dependent population and types of housing material.

Table 9: Flood risk indicators Flood Risk Indicators

Vulnerability Index

Topography Socio-economic Infrastructure Exposure 30% 40% 30%

· The island Mean water depth: Poverty Dependents: House Structure: · The area out of 30% Line: 30% 10% 30% embankment protection · The area prone to river bank erosion

A brief description of the said parameters has been given below.

Topography: The land topography is the key factor for land use in the flood plain. For example, land use for crop cultivation in the monsoon where in average monsoon water depth not high enough to submerge the standing crop in the field. The assets/commodity in the low elevation topography (low lying area) inf the floodplain areas are highly vulnerable to the flood as flood water comes early and stay longer period with higher depth. In this analysis use flood depth as a proxy for the topography.

Poverty Line: population has been considered living under the lower poverty line who has not able to fulfill the basic needs (for this particular indicator we refer to food security).

Percentage of population in upazila of Jamuna flood plain area living below poverty line

Dependents: Dependent population has been considered under two age groups: ● Elderly: % of population over the age of 60 years plus ● Children: % of population under the age of 5 years

House Construction: During flood, house damage is occurred due to water velocity, water depth and duration of inundation. As water velocity data is difficult to have access, water depth has been considered as the reason for house damage. There are four house types Jhupri, Kacha, semi-pucca and pucca as recorded by BBS. The Jhupri for its construction type is most vulnerable to flood and Kutcha houses are little bit less vulnerable than Jhupri. So, almost same index has been given to these types of houses. And as Semi-pucca and Pucca is more resistant to flood, lower value has been given to these

types of houses. For instance, for a certain flood, if the Jhupri is 100% damaged, semi-pucca building will be damaged 50% of it.

● Very High: 100 value x percentage of house type Jhupri in the area/district ● High: 90 value x percentage of house type Kacha in the area/district ● Moderate: 50 value x percentage of house type semi-pucca in the area/district ● Low: 25 value x percentage of house type pucca in the area/district

3.1 Exposure At the first stage, the exposure has been considered with three parameters: i) The island – Chars ii) The area outside of embankment protection iii) The area prone to riverbank erosion

A brief description has been given on the exposure categories in the following part.

3.1.1. The island (char land): Island or Char refers to a tract of land surrounded by the waters of an ocean, sea, lake, or stream; it usually means, any accretion in a river course (map 9). In the dynamics of erosion and accretion in the rivers of Bangladesh, the sand bars emerging as islands within the river channel (island chars) or as attached land to the riverbanks (attached chars). And these chars are vulnerable to flood. In this research, the chars throughout the Jamuna River have been considered as highly exposed to floods (map 11). There are around 2 million people (around 450 hundred thousand households), who are considered as vulnerable.

Map 9: Char land in the Jamuna floodplain 3.1.2. The area outside of embankment protection Government of Bangladesh over the period constructed 10,000 km earthen embankment to protect the agriculture land from flooding. The Brahmaputra right embankment (BRE) has been constructed along the right bank of the river Jamuna and along the left bank there is no embankment. The area along the riverbank and outside the embankment is more vulnerable than the protected area (inside the embankment). The unions which are located outside of the embankment protection along with the bank of the river has also been considered as another exposure categories (Figure 3). The population living in these unions is about 2.2 million. In these unions, there are around five hundred thousand households which are exposed to flood.

3.1.3. The area prone to riverbank erosion The area prone to riverbank erosion has been considered as another exposure (map 10). Riverbank erosion, one of the major natural disasters in Bangladesh, causes untold miseries every year to thousands of people living along the banks of the rivers of Bangladesh.

Erosion prediction has been done for both banks of the Jamuna River for 2018. Erosion is predicted for 15 locations (7 locations along the right bank and 8 along the left bank), the lateral extension of which would be more than 100 m. Among these locations, three are in Kurigram, four in Gaibandha, one in both Kurigram and Gaibandha District, two in Bogra, two in Sirajganj, two in Tangail and one both in Sirajganj & Tangail District. About 783 ha of land and 123 ha of settlement are vulnerable to erosion. Map 10: Erosion prone area with the riverbank line

There is more than 50% probability that active flood embankments may breach at two locations along the right bank. About 1140 m active embankment is vulnerable to riverbank erosion. In this study, the unions having the probability of river erosion has been considered as another exposure whose total population is 5, 57,709. And there are more than one hundred thousand households located in these unions among which 90.91% (on an average) households are kutcha.

These three categories of exposure is shown in map 11.

Map 11: Exposure of Jamuna flood plain

3.2 Vulnerability Index For identifying the Vulnerability Index three types of parameters has been considered which has been elaborated in the following section.

3.2.1 Socio Economic To identify the socioeconomic vulnerability, two indicators have been evaluated for the specified area: Poverty line and Dependent people.

I) Poverty Line in relation to Food security The poverty estimates are based on poverty line developed by the Bangladesh Bureau of Statistics (BBS) jointly with World Bank staff, employing a Cost of Basic Needs (CBN) approach (2010). Intuitively, CBN poverty lines represent the level of per capita expenditure at which a household can be expected to meet their basic needs (food and non-food).

If the frequency of the upazilas is identified against the percentage of population living under the lower poverty line the following map is depicted (map 12).

II) Dependent Population Population data has been collected at union level. Dependent population has been considered under two age groups: i) elderly (over the age of 60 years) and ii) children (under the age of 5 years). In Bachamara union, Daulatpur thana, Manikganj district, 24.4% population is categorized as dependent population (map 13).

Map12: Percentage of people living under lower poverty line of Jamuna flood plain Map 13: Percentage of dependent population

3.2.2 Topography Low lying areas are more vulnerable to flood and proxy to represent the topography by flood depth which depth on DEM (Digital Elevation model). The higher water depth represents low land area and vice versa. In map 14, the red refers to low land area and the green refers to the high land area. From the average water depth for each union (calculated using recent historic data on water depth derived from FFWC model) the topography has been defined (shown in map 14).

3.2.3 Infrastructure To assess the vulnerability of the infrastructure, the housing condition of the area has been assessed.

I) House Construction techniques and materials Type of housing structures data has been collected from “Housing Condition in Bangladesh”, 2015. The types of the houses have been categorized in four types: Pucca house; Semi pucca house; Kutcha house and Jhupri. In Jamuna flood plain area, most of the houses (on an average 77.06%) are kutcha houses.

In this study, all types of houses have been considered as vulnerable to flood. But, as all types of houses are not vulnerable to same extent, different value has been given for different type of houses and a composite value for vulnerable structures has been generated for all districts (map 15). The value is calculated to be 100% for the district with the maximum of houses that are Jhupri and Kutcha, and then normalized for all other districts.

This composite score is highest in Ward No-01 union, Sreebardi thana, Sherpur district (92.65).

Map 14: Topography of Jamuna flood plain Map 15: Composite score for house type structure

3.2.4 Vulnerability Index Map Finally, using the above-mentioned parameters the Vulnerability Index map has been prepared. To identify the vulnerability index, the normalized value has been identified for the said parameter and then, specific weightage has been given to the parameters. Finally, by adding all the parameters, a

vulnerability index map (map 16) has been prepared for the unions where the higher index expresses the higher vulnerability. In Narayanpur union, Nageshwari upazila, Kurigram district the index is highest (91.99).

The Vulnerability Index is mostly higher in Kurigram district, then in Gaibandha and Jamalpur district and partially higher in Tangail district.

Map 16: Vulnerability score of different unions of Jamuna flood plain area

3.5 Prioritized Impact Priority impacts: ● Human causality (deaths due to drowning and others) ● Loss of household assets ● Loss of livelihood linked to livestock

In joint discussions with BDRCS, RCCC, BMD, DDM, and SRC, this Early Action Protocol will address the loss of household assets and the loss of livelihood linked to livestock. These priority impacts were chosen taking into account the FbF capacity to address these impacts in a meaningful way. Based on past experiences with flood response and keeping in mind the two-stage trigger of 10 and 5 days, it was agreed that these priority impacts could be reduced by the FbF system.

Annex B: Detailed trigger analysis 4.1.1 Menu of Forecast The global and national level forecast systems have been considered to select the forecasting system for trigger.

Global forecast The GloFAS system is composed of an integrated hydro-meteorological forecasting chain and of a monitoring system that analyzes daily results and shows forecast flood events on a dedicated web platform. To run the model, Meteorological data and Hydrological modeling is needed.

In Bangladesh, Bahadurabad station can be a good point to exercise GloFAS model. An example has been given from the experience in 2017 (shown in figure 7). On 1st of July the forecast shows flood water will cross 1 in 10-year return period on 11th July and it was reached.

Figure 7: GloFAS 1-15days probabilistic forecast on 1 July 2017 at Bahadurabad

For GloFAS forecast, skill has been assessed for the period: 1 June to 31 August 2018 at Bahadurabad station, Brahmaputra River. The r2 are 0.76 and 0.61 for 10 days and 15 days lead time respectively. The 10 days GloFAS forecast works very well in July and August in 2017 flood in the Jamuna floodplain.

National forecast (FFWC): The Flood Forecasting and Warning Centre (FFWC) is under the Bangladesh Water Development Board (BWDB) and monitors and forecasts water level at 54 locations in Bangladesh (map 17). There are six FFWC stations along the Jamuna river.

Map 17: FFWC stations in Bangladesh

Currently FFWC produces three different forecasts, two deterministic forecasts and one probabilistic. One deterministic forecast is utilizing mainly on ground-based data and produces forecasts for 5 days, while the other deterministic forecasts takes satellite based altimetric data and produces an 8-day prediction. The probabilistic forecast uses also ground data and predicts water levels up to ten days in advance.

FFWC produces flood depth map for 1-5 days lead-time. The map is produced with all the model forecast water level nodes. Here is an example of 5 days lead time FFWV forecast on flood depth and extension of August 2017 shown in map 18.

Map 18: Flood depth map from FFWC (as an Example)

Among the forecasts, five-day deterministic forecast is the most reliable one and it will be used to trigger this EAP. Figure 6 illustrates FFWC’s deterministic forecast with an example of forecast made on 11 August 2017. As shown in the figure, on 11th August 2017 FFWC 1-5days forecast shows water level to cross the DL on 13th (Figure 8) and cross DL + 0.5 meter on 16th August. The forecast performance in terms of Coefficient of Determination (r2) is 0.77 at the Bahadurabad station.

Figure 8: Flood forecasted 1-5 days by FFWC at Bahadurabd on 11th August 2017

The Forecast Performance has been assessed at Sariakandi station. For different lead-time at Sariakandi station, the coefficient of determination in 72 hour and 120 hours lead time are 0.92 and 0.76 respectively. The Hit Rate (HR) and False Alarm Ratio (FAR) are 0.67 and 0.289 respectively for 5 days lead-time using the government danger level.

Summary of the available forecast from different sources with skill is shown in Table 10.

Table 10: Forecast type, lead time and skill of different model Lead time and Forecast Source Forecast type error in time Skill and location

GLOFAS http://www.globalfloods.eu/glo Probabilistic 1-15 days 10 days fas-forecasting/ Discharge and forecast R2: return period on 0.760 the river 15 days Brahmaputra at forecast R2: Bahadurabad 0.610 (limited data – 1 year)

1-5 day forecast http://www.ffwc.gov.bd/index. Deterministic 1-5 days 5days forecast php/hydrograph/forecast Water level R2: 0.72 forecast at 52 FAR (5days) locations 0.29

1-10 day http://www.ffwc.gov.bd/index. Probabilistic 10 days probabilistic php/hydrograph/medium- Water level R2 = 0.61 forest range-1-10-days-forecast forecast – mean (FFWC annual +sd, mean and report 2016) mean - SD

4.1.2 Extreme event Extreme hydrological events refer to such events which have the potential to cause considerable harm to human populations and their related activities. In Bangladesh, extreme event is considered 1 in 5 year considering mainly the agricultural impact. But, for the damage of houses and people the extreme event is considered 1 in 10 year.

Figure 9 presents the historical data on maximum water level at Bahadurbad station. As shown in the figure, the danger level is at 19.5 m. In last 31 years (1987-2017), the water level has crossed the danger level 24 times and DL + 1 meter 4 times.

Figure 9: Yearly distribution of Maximum Water level at Bahadurabad station

Figure 10 shows yearly maximum discharge (m3/sec) at Bahadurbad station for last 34 years (1984- 2017). As shown in the figure, in last 34 years, the discharge crossed 75,000 m3/s 11 times, 80,000 m3/s seven times (i.e. 7 in 34) and 85,000 m3/s four times.

Figure 10: Yearly distribution of Maximum Discharge at Bahadurabad station Then the return period of maximum water level and discharge has been identified at Bahadurabad station. The return period at Bahadurabad station is 1 in 10 years. For 1 in 10-year return period, the discharge is 82,000 m3/s and the water level is 20.4m pwd (DL +0.9m). In general, the Danger level is set at 0.5m above of agricultural field, at which point agricultural damage starts. And normally the houses are constructed at 1m above the agriculture field. So, when the water level crosses DL+1m, the houses are submerged at plinth level. Hence, the flood exceeds the water level DL + 1.0 meter (at Bahadurabad) or discharge 84 000 m3/s is considered as an extreme flood event for FbF trigger.

4.1.3 Impact Limit The increase in discharge will increase the water level- meaning more flooding. Here, a relationship has been established between the discharge and damage. Flood damage is also related with the duration of water at a certain level. But, the data on duration of flood was not available at the national level and hence relationship has been drawn with the discharge. And from field level data, detail analysis has been done on damage against the duration and water level. For national level analysis, discharge data of Bahadurabad station has been considered as sample. A relationship has been drawn between Peak Discharge data at Bahadurabad station and affected population (Figure 11).

Figure 11: Relation between peak discharge data at Bahadurabad and affected population due to flood

From the above relation it can be observed that up to a certain discharge value (60,000 m3/s), there is no impact. However, after that point increase in discharge increases the number of affected populations exponentially. The major flood impacts start at a discharge of around 80 000 m3/s at Bahadurabad. The big events like 1998 represent 1 in 100 year and the corresponding discharge records shows the flow around 100 000 m3/s and causing 70% of the population affected.

An analysis was carried out to find a relationship between Peak Discharge data at Bahadurabad and Affected housing due to flood (Figure 12).

Figure 12: Relation between Peak Discharge data at Bahadurabad and Affected house due to flood

Figure 12 shows that, to a certain level of discharge (like 60,000 m3/s), there is no impact. Further to that will increasing the damage on houses. A discrepancy can be observed in the above chart, which is the damage of 1988 floods that was also due to floods from Ganges basin.

The historical major event of flood impact varies with flow and duration as shown in Table 11. In all those major flood events, IFRC activated the DREF. In all those floods the flood duration above the DL at Bahadurabad was 10 days and the discharge was over 68 500 m3/s. Hence it may be the lower limits for triggering DREF.

Description Flood Impact parameters

Flood year 1988 1998 2004 2007 2016 2017

Discharge 68,700.00 102,535.00 96,105.52 71,860.11 76,554.92 78,262.42

Duration (water level 11.00 32.00 16.00 10.00 14.00 12.00 above danger level)

Number of homes 7.20 0.98 4.00 1.10 0.25 0.62 fully/partly damaged (million)

People affected 45.00 31.00 36.00 14.00 3.70 6.7 (million) Table 11: Historic impacts with respect to flood discharge and duration at Bahaduarabad River Jamuna.

Impact Survey

Key Informant Interview (KII) A field assessment was carried out in Bogra and Kurigram districts to analyze impacts of different flood heights on household assets in three categories of exposure; char land, potential erosion risk area and outside embankment expose area. It was found that, the impact was not significant until the submergence of plinth level. In Raniganj union (Char) in Kurigram district flood height at plinth (DL+1m) for 3 days and 10 days led to household asset damage respectively at 10% and 15%. However, in the mainland outside the embankment at Jatrapur the damage was observed at 25% and 35% respectively.

Figure 13: Percent of household asset damage for different flood heights and duration, Raniganj union, Kurigram district.

Household Survey GRC and the alliance partners (Swedish Red Cross and WFP) Was BDRCS involved in this study? conducted a household survey with 307 HHs in 6 unions in Tangail, Gaibandha and Kurigram districts. The HH survey showed that about 48% house were inundated half of the wall (DL + 1.5 meter) and

68% of house flooded above plinth level (DL + 1.0m). The survey showed that about 40% of HH had agriculture field that got inundated in 2017 flood. Of the total surveyed households, 87%, 63% and 45% households experienced inundated yard, plinth and half of the wall of house respectively in 2017 flood.

The flood impacts on assets at different stage of flooding is shown in figure 12. The survey data showed that household assets damage in terms of monetary value in the Char land is almost half of that in the mainland. This is because household asset in the Char is in the range of BDT 115,000 – 175,000 which is about BDT 50,000 to 75,000 less than house in the main land.

Figure 14: Total asset damage in three explore areas (source: HH survey)

From figure 15 it is evident that structural damage and livestock loss have the largest monetary implications to the households. The cow shed, furniture, toilet, tube-well and other damage are bellow BDT 5000.

Figure 15: HH asset damage for different flooding (source: GRC 307 HH survey data)

The survey data shows about average BDT 32000, 18000 and 11000 per HH loss of livestock if the flood reach up to half of the wall, plinth and yard respectively. Hence impact start when agriculture/grazing field goes under water.

During and after flood, health impact is a major concern in the vulnerable community. From the HH expenses data shows about BDT 2000 to BDT 3000 per household for health during the flood.

To compare the impacts across three exposure groups normalization with carried out with respect fo the mainland outside embankment protection. We used the monthly income and HH asset for this task and computed normalized factors at 1, 1.2, and 1.6 for HH outside embankment in the mainland, erosion prone, and char islands respectively; as shown in table 12.

Table 12: Computation of normalized factor HH Asset Exposure Monthly Income (BDT) HH Asset* Monthly Normalized (BDT) type 2016 HH baseline survey Income factor From KII

Char 150,000 5,760 864,000,000 1.6

Erosion 185,000 6,548 1,211,380,000 1.2

Mainland OE 215,000 6,548 1,407,820,000 1

After using the normalize factor on the impacts, a graph has been generated, which is presented in figure 16. This curve maybe used to compute the impact on household assets based on forecast water depth in FbF trigger and the damage limits 25%.

Figure 16: Flood Impact HH asset in Char, Mainland outside embankment and erosion area for different flood depth.

In addition to the HH asset damage an attempt was made to estimate affected population for different flood level using FbF pilot survey data. The topography survey data (four pilot area in Bogra district, 2016), household demography data and water level measurement are used to generate affected population against flood depth as shown in Figure 17. As shown in figure 17 when the flood depth is over one meter, about 40% people living in the area will be affected and this may be used as an impact limit in the trigger.

Figure 17: Affected population by flood depth

Hence, when forecasted flood depth impacts 25% of household asset and 40% population the impact limit is in place to trigger this EAP.

To activate trigger, two-stage trigger methodology has been proposed for FbF: (i) pre-activation; (ii) forecast based activation. The trigger criteria for different lead-time such as 10 days for pre-activation and 5 day for EAP activation are summarized in table 13. There are three trigger elements that are the forecast, extreme event and thresholds, and impact level that are described in table 13.

Table 13: Trigger criteria at different stage of flood Trigger Elements Stage I: Readiness Stage II: Activativation

Forecast

Source Global model GloFAS National model: http://www.globalfloods.eu/glofas- http://www.ffwc.gov.bd/index.php/hydrograph forecasting/ /forecast

Lead time 10 days 5 days

Skill R2 =0.61, chance 70 percent R2: 0.72 and FAR 0.29

Content Discharge and return period at Bahadurabad Water level and wrt DL at Bahadurabad

Flood Depth and JBA produced flood map / use FFWC 2016 FFWC 120 hr Flood depth map Extent map for 1 in 10 year flood and 1 in 50 year http://www.ffwc.gov.bd/index.php/map/inund flood use ation-map/bangladesh-120hr-forecast

Extreme event and threshold

Flow 84 000 m3/s or 1 in 10 year

Water height DL + 0.85 meter /level 20.5 m pwd at Bahadurabad

Duration 3days 3days

Impact Limits

Affected 3.7 million (Table 2 and Figure 16) 40% (figure 22) population

Damage houses 0.26 million (Table 2 and Figure 17)

Household assets HH Asset damage (Figure 21) cross 25%

Follow are the statements for two-stage trigger. Stage I – Pre-activation- Get ready with the resources and identify beneficiaries in the forecasted exposed area. Pre-activation trigger will be reached when flow at Bahadurabad (forecasted by GloFAS with lead time 10 days) likely (50% or more chance) to cross the 84,000 m3/s and flowing over at least for three days and impact forecasting 3.7 million population will be affected. Stage II – Activation- Activate EAP in the vulnerable households within forecasted exposed area. Trigger will be reached when forecast water level by FFWC at Bahadurabad (5 days lead-time) has over 70% chance to cross the DL+0.85m and remain above for 3 days and impact forecasting household assets over 25% and affected population over 40%.