The participatory methods of focus group discussions (FGDs) [11] and Key Informant Interviews (KIIs) [12] have been followed in data collection for community-based climate change risk assessment. Location of KII and FGDs are shown in Figure 2. The preliminary vulnerable unions have been identified

through preliminary field visit and conducting KIIs with relevant local level government officials from agriculture department, fisheries department and LGED. A total of 56 FGDs have been conducted in the most vulnerable parts of the study area. Local level stakeholders from community specifically farmers, fishermen, women, local influential people, labourers, marginal people etc. participated in the FGDs.

The sessions were pre-arranged following standard participatory norms of FGDs. The checklists for the KII and FGDs include the following points:

· Major Risks/Hazard (e.g., Drought/Hail Strom/Flash Flood etc.) in the Upazilla

· Most vulnerable area of this Upazilla in terms of risk and hazard

· Local perceptions on risks (i.e. drought/hail storm/flash flood)

· Causes, Frequency, intensity and specific month/season/time for each risks

· Identify possible hazards due of climate change in the area

· Scoring/Ranking of the hazards for the area

· Identify vulnerable groups (e.g., woman, children, elderly people, landless farmers, marginal/small/medium farmers, fishermen, day labourer etc.)

· Possibility of certain climate change induced hazard (likelihood) in past and future

· Hazard impacts and their possible consequences

Vulnerability and risk assessment encompass various approaches and techniques ranging from indicator-based global or national assessments to qualitative participatory approaches of local level vulnerability and risk assessment [13] [14]. In this study, qualitative risk assessment has been performed focusing particularly in agriculture and fisheries sectors, using both participatory and scientific tools. The overall risk assessment process has been illustrated in Figure 3.

Risk score has been calculated following the IPCC (2012) defined formula as

shown in Equation (1) [14].

Risk = P ∗ E (1)

where, P = Probability of occurrences of hazards or Likelihoods, E = Extent of consequences due to climate change induced hazards. Scores for likelihood (occurrence) and consequence have been computed based on scoring. In this way, risk score has been found for each of the 28 Upazillas against each of the identified hazards. Normalization of score has been performed to convert the scores to a standard range of 0 to 1 (Equation (2)) [15]. The normalized scores have been classified to set the level of risk following Table 3.

Normalization formula:

Normalized   Score = 1 − [ ( Maximum − Respective   Scor e ) / ( Maximum − Manimum ) ] (2)

Based on the FGD findings, temperature rise, erratic rainfall, drought, early flash floods, thunderstorm, Afal or wave action, hailstorm, erosion and siltation and cold snaps are major climate change induced hazards which community people in Haor areas face very frequently. Figure 4 presents the overall scores for each hazard in the Upazillas of the study area.

Among the disasters mentioned above, flash flood, thunderstorm and hailstorm are the main concerns for the Haor region. The FGD participants specified that, every year at least one of these disasters occurs and destroys their crops fully or partially. Figure 4 also shows that, most of the participants of northern Upazillas have given more emphasis on early flash flood as the most frequent event while in the southern Upazillas early flash flood got less emphasis.

Potential consequences of climate change induced hazards have been perceived

through the FGDs. According to the assessment, most hazardous consequences would be the loss of crops as frequency of early flash floods is increasing and due to damage of crops before harvesting. Although, they did not anticipate massive crop land loss due to those hazards. Loss of households is the second most considered negative consequence which would occur in the study area. There are chances of loss of fish production too but in limited manner according to community perceptions. Most interestingly, community people did not express major concerns over potential loss of income, which indicates that they are used to with the alternative livelihoods, if their main occupations were hampered. Apart from these, anticipations have been made on deterioration of drinking water quality, health issues and hampered education almost uniformly among all the Upazilla, but not severely as crop production loss. Following Figure 5 illustrates consequence scoring by the community people in the study area.

The FGD participants assessed the likelihood or probability of occurrences of

different climate change induced hazards (Figure 6). Most of the respondents expressed that temperature rise is very frequent phenomena now and would be as usual in future, whereas excessive rainfall will occur once in 2 to 3 years. The frequency of the occurrence of early flash floods would be as excessive rainfalli.e. once in 2 to 3 years, as early flash floods is mostly triggered by excessive rainfall. Respondents from Nabinagar of Brahmanbaria District only opined that there would be very less probability to occur early flash floods in that area.

Thunderstorm and hailstorm are occurring damage very frequently i.e. almost every year according to the perceptions of community people, which will prevail in future too. Mixed feedback has been found regarding likelihood assessment of drought. Either once in each year or within 2 to 3 years, drought may occur in the area. Erosion and siltation problem would be severe once in 2 - 3 years according to their perceptions. People of Sunamganj and Netrokona face severe wave action problem almost in every year than other parts of the study area. In other parts, frequency of occurrence has been perceived as once in 2 to 3 years. Likelihood of occurrence of cold snaps is found also as of highly frequent, not very high.

The risk due to individual hazards for each of the Upazillas has been assessed following the methodology described. Figure 7 illustrates the spatial variation of climate change risk for individual hazards. The results show that, Kalmakanda, Sunamganj Sadar, Dakshin Sunamganj, Dhamrapasha, Khaliajuri, Dowarabazar and Tahirpur Upazillas are in high climate change risk due to early flash floods, whereas, Bishwambharpur, Jamalganj, Derai Upazillas are under medium risk level despite being located in the deeply flooded area. But the normalized scores for these Upazillas are quite close to lower threshold of the high risk category i.e. 0.7. Therefore, the risk map shows the variation of risk in a relative scale, which may vary on the basis of the overall understanding of the local context. Sometimes, the risk map due to individual hazard shows sudden or irregular changes in the risk level, which may be due to the variation of community perceptions and lack of homogeneity in the knowledge level of groups of community people.

The multi-hazard risk map has been prepared taking weighted sum of the normalized score of each hazard for each Upazilla. The weight for multi-hazards has been derived from the FGDs outcomes (Figure 8). According to the community preferences, early flash floods and drought both got highest 12% weightage, 11% weightage for temperature rise, hail storm and thunder storm while river bank erosion and wave action was given less emphasis (8% weightage).

Among all the identified climate change induced hazards. Total weightage has been counted in a scale of 100. Based on this information, the multi-hazard weighted sum of normalized scores has been calculated for each of the 28 Upazillas considering same combination of weights for each Upazilla.

Finally, multi-hazard climate change risk map has been produced from the calculated total normalized score. Figure 9 shows the multi-hazard climate change

risk map with priority rank of different Upazillas. The developed multi- hazard risk map has been found very consistent with other study results (e.g., [1] [16] ).

According to the multi-hazard risk map, 12 Upazillas have been found to be under high risk, 11 Upazillas under medium risk and 5 Upazillas under low risk. Mostly, north-eastern Upazillas are in high risk category. A priority rank of the Upazillas has been prepared using the calculated normalized score. Among all Upazillas, those of Sunamganj District and one Upazilla from Netrokona District are considered as high priority in terms of multi-hazard climate change risk in the study area.

For the geo-spatial analysis, identifying the elements at risk is a key component. The study team tried to identify the most important elements which may face risk due to climate change induced hazards using the community perceptions i.e. socio-economic context analysis from this study as well as from relevant literatures. As, socio-economic resources and infrastructures will be worst affected due to the climate change induced hazards, which are also directly linked with life and livelihood of the community people, most of the elements are identified from these categories. Apart from that, perennial beels and fish sanctuaries have been considered as elements at risk due to the particular focus of this study on agriculture and fisheries sector. The submersible embankments of BWDB or LGED both are very crucial for the study area in connection with crop production as well as for maintaining communication or connectivity with other parts of the country, therefore, roads, embankment and rail networks are also considered as elements at risk.

Density of these elements of risk has been analysed to identify which part of the study area has most densely populated elements at risk. These analyses have facilitated further to identify hotspot risk areas in the study area. Figure 10 illustrates the spatially varied density of elements at risk in the study area along with

the weighted overlay process.

After the weighted overlay, a climate change risk hotspot map has been prepared (Figure 11), which specifically depicts the site-specific areas which

should be considered for immediate interventions in terms of potential risk due to climate change induced hazards. Around 100 unions (Table 4) have been identified as climate change risk hotspot area. Some unions in the southern part of the study area are found as risk hotspot, despite being located in low climate change risk level as the elements at risk are very densely located in those areas.