WPS7593 Policy Research Working Paper 7593 Impact of Climate Change and Aquatic Salinization on Fish Habitats and Poor Communities in Southwest Coastal Bangladesh and Bangladesh Sundarbans Susmita Dasgupta Mainul Huq Md. Golam Mustafa Md. Istiak Sobhan David Wheeler Development Research Group Environment and Energy Team March 2016 Policy Research Working Paper 7593 Abstract Fisheries constitute an important source of livelihoods for households in the region. Using the salinity tolerance range tens of thousands of poor people in the southwest coastal for each species, 27 alternative scenarios of climate change region of Bangladesh living near the UNESCO Heritage in 2050 were investigated to assess the possible impacts of Sundarbans mangrove forest, and they supply a significant climate change and sea level rise on aquatic salinity, fish spe- portion of protein for millions. Among the various threats cies habitats, and the poor communities that consume the fisheries in the southwest coastal region and Sundarbans affected fish species. The results provide striking evidence that mangrove forest will face because of climate change, adverse projected aquatic salinization may have an especially nega- impacts from increased aquatic salinity caused by sea level tive impact on poor households in the region. The estimates rise have been identified as one of the greatest challenges. indicate that areas with poor populations that lose species This paper focuses on 83 fish species consumed by poor are about six times more prevalent than areas gaining species. This paper is a product of the Environment and Energy Team, Development Research Group. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The authors may be contacted at sdasgupta@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Impact of Climate Change and Aquatic Salinization on Fish Habitats and Poor Communities in Southwest Coastal Bangladesh and Bangladesh Sundarbans Susmita Dasgupta* Mainul Huq Md. Golam Mustafa Md. Istiak Sobhan David Wheeler JEL Classification: Q22; Q25; Q54; Q57 Keywords: climate change; aquatic salinization; fish habitats; poverty; Bangladesh. This research was conducted under the South Asia Water Initiative - Sundarbans Landscape. *Authors’ names are in alphabetical order. The authors are respectively Lead Environmental Economist, Development Research Group, World Bank; United States; World Bank Consultant, Bangladesh; Research Associate, WorldFish, South Asia; World Bank Consultant, Bangladesh; and Senior Fellow, World Resources Institute. We would like to extend our special thanks to Lia Sieghart, Pawan Patil and Michael Toman for their review comments on this paper. 1. Introduction Around 43.2 million people or 30 percent of the population of Bangladesh live in poverty. This figure includes 24.4 million extremely poor people who are not even able to afford their basic needs of food expenditure. In densely populated and land scarce Bangladesh, poor households are disadvantaged with regard to land access, and many end up settling in low-lying regions close to the coast. The poverty map developed by the Bangladesh Bureau of Statistics, World Food Program and World Bank identifies a high incidence of poverty near the coast, where 11.8 million poor people are located in 19 districts (World Bank, 2014). The incidence of poverty is particularly severe in the southwest coastal region, where the area is prone to tidal surges and cyclones, soil and water are saline at certain times of the year, and living conditions are harsh. The vulnerability of coastal regions to flooding, storm surges and salinity will further increase in this century, according to the climate projections of the Intergovernmental Panel on Climate Change and the World Meteorological Organization. Climate change thus poses a serious threat to the livelihoods of the poor in the southwest coastal region, especially because they are burdened by limited mobility due to their economic circumstances, disadvantages with land access, and near-total dependence on local ecosystems for their livelihood. Fisheries make an important contribution to the regional economy, especially in areas close to the Sundarbans mangrove forest (Shah et al. 2010). Marine fisheries, inland open water or capture fisheries and closed water fisheries provide an important source of livelihood for tens of thousands of poor people and supply a significant portion of their protein intake (World Bank 2000; Alam and Thomson 2001; Thilsted 2010; Thisted 2012; Fernandes et al., 2015). Over the years, southwest coastal region inland open water fisheries have faced increasing threats from over-exploitation of resources; indiscriminate fishing with inappropriate fishing gear; 2 destructive fishing practices, such as the use of poisons in closed creeks or canals; increased water pollution; reduction in the freshwater flow of the river system; and intrusion of salinity. Significant threats from human actions are likely to continue in the future, and the stress on fisheries in the region may be further aggravated by climate change. Among climate-related threats fisheries in this region will face,1 one of the greatest challenges will be increased aquatic salinity from sea level rise and climate-induced changes in temperature, rainfall and riverine flows from the Himalayas (Dasgupta et al., 2014; Gain, Uddin and Sana, 2008). These changes will adversely affect many fish species, with significant impacts on their reproductive cycles, reproductive capacities, suitable spawning areas, feeding, breeding, and longitudinal migration. Fishing communities are among the poorest of the poor in Bangladesh, so understanding these impacts is critical for ensuring the future sustainability of fishing-dependent households. Within the southwest coastal region, Sundarbans ecosystem supports a wealth of fish diversity,2 provides a refuge for fish from predators, and serves as a nursery for the larvae and juveniles of 90 percent of commercial fish and 35 percent of all fish in the Bay of Bengal (USAID 2010). In 2008, Gain, Uddin and Sana studied the impact of river salinity on fish diversity in the southwest coastal region near the Sundarbans. Their research area included highly saline conditions in Paikgacha upazila, Khulna district, and moderately saline conditions in Rampal upazila, Bagerhat district. The researchers analyzed river salinity data monitored by the Bangladesh Water Development Board (BWDB) for the Sibsa river in Paikgacha and the Passur river in Rampal, and found a significant increase in salinity from 1975 to 2004. After surveying 1 Other threats include increased water temperature, changes in cyclonic storm patterns, change in surge heights. 2 According to IUCN, water bodies in the Sundarbans (ruvers, streams and canals) covering 1,874 sq. km and marine zones covering 1,603 sq. km upport 27 families and 53 species of pelagic fish, 49 families and 124 species of dermal fish, 5 families and 24 species of shrimps, 3 families and 7 species of crabs, 2 species of gastropods, 6 species of pelecypods, and 8 species of locust lobster. See Shah et al., 2010 for details. 3 local fishermen, the researchers concluded that freshwater fish species declined by 59 percent in Paikgacha and 21 percent in Rampal, with little compensating increase in saline-tolerant fish. The study inferred that reduction in fish diversity is a serious threat to the local ecosystem and food supply. In light of such evidence, the potential impacts of increasing salinity have become a major concern for the Government of Bangladesh and affiliated research institutions. Recently, the Bangladesh Climate Change Resilience Fund (BCCRF) Management Committee has highlighted salinity intrusion in coastal Bangladesh as a critical part of adaptation to climate change. Prior research on salinization has employed a variety of methods (See for example Nobi and Das Gupta 1997; Aerts et al., 2000; IWM 2003; CEGIS 2006 and Bhuiyan and Dutta 2011). Many of these studies have simulated salinity change in rivers and estuaries using hydraulic engineering models and then compared the results with actual measures. In the most comprehensive study to date, Dasgupta et al. (2015) have used 27twenty seven alternative climate change scenarios to project salinity trends in coastal rivers to 2050, with a model that links the spread and intensity of salinity to changes in the sea level, temperature, rainfall, and altered riverine flows from the Himalayas. The study provides new estimates of location-specific river salinity through 2050. Resources will remain scarce, and mobilizing a cost-effective response will require an integrated spatial analysis of threats from salinity diffusion, their socioeconomic and ecological impacts, and the costs of adaptation. The temporal and geographic pattern of appropriate adaptation investments will depend critically on the ecological impacts of salinity diffusion in different locations. Understanding household choices will also be critical, since households may respond to localized threats of salinization by relocating some or all members to areas where expected earnings and survival probabilities are higher (Dasgupta et al., 2014). 4 This paper attempts to contribute by assessing the impact of aquatic salinization on the spatial distribution of fish species that are significant for the livelihoods of poor fishing communities in southwest coastal districts and the Sundarbans region. 3 In absence of comprehensive data on spatial distribution of fish abundance by species, the focus of our analysis is on expected impact of changing aquatic salinity on the extent of fish habitats. Although the importance of Sundarbans mangroves as fish habitats and nursery grounds is recognized in the literature, this paper does not consider the indirect impact that climate-induced changes in the location and composition of mangroves will have on fish species. The remainder of the paper is organized as follows. Drawing on prior work by Dasgupta, et al. (2015), Section 2 develops high-resolution digital maps of aquatic salinity in the Sundarbans region for 2012, and salinity in 2050 projected for 27 combinations of IPCC climate change scenarios, global circulation models, and assumptions about rates of land subsidence in the Ganges Delta. In Section 3, we develop a database for 83 fish species that are important for the livelihoods of poor households in the Sundarbans region. Section 4 combines our salinity projections with information on fish species salinity tolerances to produce maps of projected changes in habitats and local species populations by 2050. In Section 5, we combine projected changes in species populations with upazila-level information on poverty to assess the potential impacts of aquatic salinization on poor households in the Sundarbans region. Section 6 summarizes and concludes the paper. 3 Examples of prior research on climate change and fisheries in coastal Bangladesh can be found in Ali (1999); World Bank (2000); Sarwar (2005); Hassan and Shah (2006); UK DEFR (2007); Chowdhury et al. (2010); World Bank (2011) and Nicholls et al. (2013). However, the bulk of this research makes inferences from descriptive statistics. 5 2. Current and Future Aquatic Salinity in the Sundarbans Region This paper draws extensively on the findings of Dasgupta, et al. (2015), who quantify the prospective relationship between climate-induced changes in sea level, temperature, rainfall, and riverine flows from the Himalayas, and the spread and intensity of aquatic salinization in the coastal zone. Their research takes account of projected land subsidence in the Ganges Delta, as well as alternative levels of upstream freshwater withdrawal. The research develops 27 aquatic salinity scenarios in 2050 that incorporate three global emissions scenarios (B1, A1B, A2)4 from the IPCC’s Fourth Assessment Report (AR4); two estimates of sea level rise by 2050 (27 cm for scenario B1, 32 cm for A1B and A2); three global circulation models (IPSL-CM4, MIROC3.2, ECHO-G);5 and three annual subsidence rates for land in the lower Ganges Delta (2, 5 and 9 mm/year). Each of the 27 scenarios is used to produce high-resolution maps of projected maximum aquatic salinity during December 2049 and six months in 2050: January-June.6 4 Basic elements of the three scenarios are as follows: B1: Rapid economic growth with convergence among regions; global population that peaks in mid-century and declines thereafter; rapid change in economic structures toward a service and information economy, with reductions in material intensity and the introduction of clean and resource-efficient technologies. A1B: Rapid economic growth with convergence among regions; global population that peaks in mid-century and declines thereafter; rapid introduction of new and more efficient technologies; energy from mixed fossil and renewable sources. A2: Non-convergent economic development; continuously increasing population; heterogeneous technologies and energy sources. 5 Model implementing institutions are as follows: IPSL-CM4: Institut Pierre Simon Laplace, France; MIROC3.2: Center for Climate System Research, University of Tokyo, National Institute for Environmental Studies, Japan, Frontier Research Center for Global Change, Japan; ECHO-G: Meteorological Institute of the University of Bonn, Germany, Model and Data Group, Max Planck Institute for Meteorology, Hamburg, Germany, Korea Meteorological Administration. 6 Average salinity concentrations of the rivers in the coastal area are higher in the dry season than in the monsoon because of lack of freshwater flow from upstream. Salinity generally increases almost linearly from October (post- monsoon) to late May (pre-monsoon) with the gradual reduction in freshwater flow. At the end of May, salinity level drops sharply because of rainfall and upstream flow of freshwater through the river system and remains low until early October. 6 Figure 1 displays the estimated spatial distribution and intensity of maximum aquatic salinity in 2012 and two projections for 2050:7 least change (Scenario B1, GCM MIROC-3.2, SLR 27 cm, land subsidence 2 mm/year); and most change (Scenario A2, GCM IPSL-CM4, SLR 32 cm, land subsidence 9 mm/year). The figure is color-coded to highlight changes in relatively low-salinity areas. In 2012 (Figure 1(a)), most of the core Sundarbans region (outlined in black) and its immediate neighborhood display north-south bands of maximum salinity that are highest (25+ ppt) in the west and decline eastward toward 10-15 ppt. Both 2050 scenarios exhibit expansion of these color bands, with somewhat greater change in the A2 case (Figure 1(c)). The eastern part of Figure 1 presents a strong contrast in 2012, with most of the area dominated by very low maximum salinity (0-2 ppt). The 2050 B1 scenario (Figure 1(b) - least change) exhibits notable area reduction for 0-2 ppt, accompanied by expansion in the ranges 3-5 and 6-10 ppt. The shift is more pronounced for A2 (Figure 1(c) - most change), with area dominance shifting to the range 3-5 ppt and further expansion of 6-10 ppt. 7 The data are mean values for seven months (January-June, December). 7 Fig 1: Sundarbans region: estimated maximum aquatic salinity in 2012 and 2050 (a) 2012 (b) 2050 (Least Change) (c) 2050 (Most Change) 8 3. Fish Species in the Sundarbans Region The area changes in Figure 1 have potential significance for the spatial distribution of fish species, since the stable habitat of each species is limited to areas whose salinity ranges fall within its salinity tolerance range.8 In this paper, we focus on 83 fish species that are consumed by households in the southwest coastal region as well as in the Sundarbans region. Appendix A1 identifies these species. We compiled salinity tolerance range of these fish species drawing on secondary literature. (For example, see Hussain et al. 2013; Rahman and Asaduzzaman 2010; MoEF 2010: Robin et al. 2010; Gain et al. 2008; Mustafa 2003: Mustafa and Dey 1994; Kasim 1979.) Table 1 enumerates the species by salinity tolerance range. Table 1: Southwest coastal and Sundarbans regions: fish species consumed by households Salinity Tolerance Number (ppt)* of Group Min Max Species 1 0 2 2 2 0 5 25 3 0 10 14 4 0 15 2 5 0 20 3 6 5 10 3 7 5 20 21 8 5 25 1 9 5 30 7 10 10 30 1 11 10 35 3 12 15 35 1 Total 83 *Salinity tolerance intervals were selected based upon consultation with local experts. 8 We define stable habitat as the area within which a species can survive year-round in any body of water that it inhabits. To illustrate, a species with a salinity tolerance range of 0-2 ppt has a stable habitat in an area whose annual salinity range is 0-1 ppt. In an area with salinity range 0-5 ppt, the species’ habitat is limited to months with salinity in the range 0-2 ppt. 9 Figures 1(a) and 1(c) strikingly illustrate the potential impact of climate change and sea level rise on species in salinity tolerance groups 1 and 2. For the two species in group 1 (tolerance range 0-2 ppt), stable habitat occupies a large swath of the eastern region (approximately 15,363 sq. km) in 2012 but practically disappears from that area by 2050 in the A2 (most change) scenario. The potential stakes are also high for the 25 species in group 2, which comprise 30% of all fish species consumed in the Sundarbans region. In 2012, almost the entire eastern part of the Sundarbans region is stable habitat for group 2. In the A2 scenario for 2050, however, maximum salinity has moved beyond the tolerance range of group 2 in broad north-south swaths at the eastern and western margins of the eastern part. By implication, poor communities in these swaths might face significant drops in fish supply by 2050. 4. The Impact of Salinization on Fish Habitats We generalize the previous illustration using the digital salinity maps provided by Dasgupta, et al. (2015). For each of the 12 species salinity tolerance groups, we assign 1 to a pixel in the map for 2012 that satisfies the stable habitat criterion (pixel salinity range falls within the species tolerance range) and 0 otherwise. We add across 101,600 pixels to determine total stable habitat by salinity tolerance group in 2012.9 We perform the same operations for all 27 salinity scenarios in 2050; calculate percent changes from 2012 to 2050 for each scenario and salinity tolerance group; and tabulate the results in Tables 2 and 3. Table 2 displays all the results, ordered by local subsidence level, IPCC AR4 scenario and global circulation model. We include summary information in Tables 3 and 4 to aid interpretation. 9 We use pixels for numerical convenience, although pixel numbers are readily translated to areas. In our mapping system, one pixel has an area of 0.327 sq. km. This is equivalent to a square cell with side length of 571.54 meters. 10 Table 3 reports median change rates across IPCC scenarios and GCMs for different salinity tolerance groups and rates of local subsidence. 11 Table 2: Stable habitat area change (%) by species salinity tolerance range, 2012-2050 Salinity Tolerance Range (ppt) Local SLR by IPCC Global 1 2 3 4 5 6 7 8 9 10 11 12 Subsidence 2050 AR4 Circulation 0-2 0-5 0-10 0-15 0-20 5-10 5-20 5-25 5-30 10-30 10-35 15-35 (mm/year) (cm) Scenario Model 2 27 B1 ECHO-G -13.8 -8.0 -1.3 -0.3 -0.3 21.5 13.2 7.8 7.4 11.2 11.2 6.4 2 27 B1 IPSL-CM4 -14.4 -8.8 -2.1 -1.4 -0.5 21.7 10.6 7.5 7.4 11.2 11.2 6.4 2 27 B1 MIROC3.2 -13.7 -7.8 -1.2 -0.1 -0.2 21.3 13.7 7.8 7.3 11.2 11.2 6.4 2 32 A1B ECHO-G -20.7 -8.7 -1.3 -0.5 -0.4 21.9 11.4 7.6 7.4 11.9 11.9 7.2 2 32 A1B IPSL-CM4 -21.3 -9.6 -2.1 -1.8 -0.8 22.6 8.6 7.3 7.4 12.0 12.0 7.2 2 32 A1B MIROC3.2 -20.5 -8.4 -1.1 -0.2 -0.4 22.3 12.1 7.7 7.4 11.9 11.9 7.2 2 32 A2 ECHO-G -20.8 -8.7 -1.3 -0.5 -0.5 21.9 11.2 7.6 7.4 11.9 11.9 7.2 2 32 A2 IPSL-CM4 -21.2 -9.3 -1.8 -1.4 -0.7 22.3 9.4 7.4 7.4 12.0 12.0 7.2 2 32 A2 MIROC3.2 -20.2 -8.1 -1.0 0.0 -0.4 22.1 12.1 7.7 7.3 11.9 11.9 7.2 5 27 B1 ECHO-G -45.5 -12.8 -2.0 -1.1 -0.7 28.1 12.5 9.4 9.0 13.3 13.3 8.3 5 27 B1 IPSL-CM4 -46.0 -13.6 -2.9 -2.2 -1.0 27.4 10.2 9.2 9.0 13.4 13.4 8.3 5 27 B1 MIROC3.2 -45.3 -12.6 -1.9 -0.8 -0.8 28.1 12.0 9.5 9.0 13.3 13.3 8.3 5 32 A1B ECHO-G -47.6 -5.4 -1.1 -0.5 -0.4 35.5 14.7 10.2 8.9 14.0 14.0 9.7 5 32 A1B IPSL-CM4 -48.5 -14.7 -2.7 -2.5 -1.2 27.0 8.2 8.8 8.9 14.0 14.0 9.6 5 32 A1B MIROC3.2 -47.7 -13.9 -1.7 -0.9 -0.7 26.8 12.5 9.2 8.9 13.9 13.9 9.7 5 32 A2 ECHO-G -47.9 -14.1 -1.9 -1.2 -0.8 26.8 11.4 9.0 8.9 14.0 14.0 9.7 5 32 A2 IPSL-CM4 -48.4 -14.7 -2.4 -2.1 -1.0 26.6 9.6 8.8 8.9 14.0 14.0 9.7 5 32 A2 MIROC3.2 -47.4 -13.5 -1.6 -0.6 -0.7 27.2 12.4 9.2 8.9 14.0 14.0 9.7 9 27 B1 ECHO-G -52.4 -19.3 -3.0 -2.1 -1.0 32.3 14.3 11.3 11.0 16.1 16.1 11.5 9 27 B1 IPSL-CM4 -53.0 -20.0 -3.9 -3.2 -1.4 31.7 11.2 11.1 11.0 16.1 16.1 11.5 9 27 B1 MIROC3.2 -52.3 -19.2 -2.9 -1.8 -0.9 32.3 15.1 11.4 11.0 16.1 16.1 11.5 9 32 A1B ECHO-G -53.7 -22.0 -2.8 -2.1 -1.0 27.9 13.2 10.6 10.6 16.4 16.4 12.7 9 32 A1B IPSL-CM4 -54.4 -22.6 -3.7 -3.5 -1.8 26.6 6.5 10.2 10.5 16.4 16.4 12.6 9 32 A1B MIROC3.2 -53.4 -21.7 -2.7 -1.8 -1.0 28.1 13.9 10.8 10.6 16.4 16.4 12.7 9 32 A2 ECHO-G -53.7 -22.0 -2.9 -2.1 -1.1 28.1 12.8 10.6 10.6 16.4 16.4 12.7 9 32 A2 IPSL-CM4 -54.2 -22.5 -3.4 -3.0 -1.4 27.4 10.1 10.3 10.6 16.4 16.4 12.7 9 32 A2 MIROC3.2 -53.2 -21.4 -2.6 -1.5 -1.0 28.1 13.9 10.7 10.6 16.4 16.4 12.7 Table 3: Median habitat change (%) by salinity tolerance and subsidence level Salinity Tolerance Local Subsidence (ppt) (mm/year) Number Habitat Size Habitat size in 2012 Group of in 2012 (sq. km) Species Min Max (Pixels) 2 5 9 1 2 0 2 46,982  15,363.11 -20.5 -47.6 -53.4 2 25 0 5 63,692  20,827.28 -8.7 -13.6 -21.7 3 14 0 10 70,964  23,205.23 -1.3 -1.9 -2.9 4 2 0 15 77,826  25,449.1 -0.5 -1.1 -2.1 5 3 0 20 88,906  29,072.26 -0.4 -0.8 -1.0 6 3 5 10 470  153.69 21.9 27.2 28.1 7 21 5 20 9,855  3,222.585 11.4 12.0 13.2 8 1 5 25 16,690  5,457.63 7.6 9.2 10.7 9 7 5 30 22,254  7,277.058 7.4 8.9 10.6 10 1 10 30 12,534  4,098.618 11.9 14.0 16.4 11 3 10 35 12,534  4,098.618 11.9 14.0 16.4 12 1 15 35 5,507  1,800.789 7.2 9.7 12.7 Table 3 highlights three major features of the results in Table 2. The first is a clear division between fresh water tolerant species (minimum ppt 0) and species that require brackish water. The freshwater species (groups 1-5) all exhibit habitat loss with increased salinization, while the brackish water species all exhibit habitat gain. Habitat loss is particularly striking for groups 1 and 2 at subsidence rates of 5 and 9 mm/year. Among brackish water tolerant species, the greatest habitat gain (27-28%) occurs for group 6 (tolerance range 5-10 ppt). Groups 10 and 11 also have relatively large habitat growth. The second feature highlighted by Table 3 is an important asymmetry in habitat scale. The greatest habitat loss rates are for groups 1 and 2, which have large habitats in 2012 (46,982 and 63,692 pixels or 15,363 and 20,827 sq. km respectively). Conversely, the greatest habitat increase rates are for groups 6, 10 and 11, which have much smaller habitats in 2012 (470, 12,534 and 12,534 pixels or 154, 4,099 and 4,099 sq. km respectively). By implication, the scale of habitat losses for freshwater species is far greater than the scale of habitat gains for brackish water species. This difference is particularly striking for freshwater group 2, which comprises 25 species in a habitat of 63,392 pixels (20,827 sq. km) in 2012, and brackish water group 7, which comprises 21 species with a habitat of 9,855 pixels (3,223 sq. km). The third striking feature of Table 3 is the effect of the land subsidence rate on habitat loss in freshwater groups 1 and 2. For group 1, subsidence rates of 2, 5 and 9 mm/year are associated with habitat loss rates of 20.5%, 47.6% and 53.4%. In group 2, which has much greater species representation (25 vs. 2 in group 1), the equivalent loss rates are 8.7%, 13.6% and 21.7%. It is more difficult to determine whether variations in IPCC climate scenarios and GCMs have significant impacts on the results in Table 2. To test these effects, we perform a regression analysis for the 324 change rates in Table 2 (27 scenarios, 12 salinity tolerance groups). We convert change rates to ranks in order to avoid scaling problems.10 We regress the rank of the habitat change rate on dummy variables for salinity tolerance groups, local subsidence rates, IPCC scenarios and GCMs. We exclude one dummy variable from each category to make the regression feasible.11 Table 4 reports results for climate scenarios and GCMs, after controlling for salinity groups and subsidence rates. We find no significance for the IPCC scenarios, but high significance for the GCMs. 10 Change rates are ranked from the greatest decrease (-54.4%, assigned rank 1) to the greatest increase (+35.5%, rank 324). 11 Inclusion of all dummy variables produces total collinearity of regression variables and failure of the regression algorithm. 14 Table 4: Selected rank regression results for habitat change rates Dependent variable: Rank of habitat change rate (Smallest = 1) Full regression dummy variable set: Salinity tolerance group, subsidence rate, IPCC scenario, GCM IPCC Scenario A1B -0.204 (0.07) A2 -0.046 (0.01) GCM ECHO 10.926 (3.54)** MIROC 12.56 (4.07)** Observations 324 R-squared 0.94 Absolute value of t statistics in parentheses ** significant at 1% 5. The Potential Impact of Salinization on Poor Households Figure 1 and Tables 2-4 reveal a spatially-uneven pattern of salinization and fish habitat change with continued climate change, sea level rise and land subsidence in the Sundarbans region. Data from rural areas in Bangladesh suggest that small low-value wild freshwater species are the most common fish consumed and the most important source of dietary protein for the poor (Belton et al. 2011; Thilsted 2010, 2012).12 The potential impact on poor households will depend on their 12 The nutritional contribution of small fish species is generally high. As many small fish species are consumed whole, they provide a significant percentage of recommended intakes of calcium, vitamin A, iron, and some minerals (Thilsted 2010, 2012). 15 vulnerability to changes in fish species in areas where salinization will significantly alter habitats.13 Vulnerability will in turn depend on the relative abundance, average size, commercial value and dietary status of local fresh- and brackish-water fish species. If the aquatic intensity (yield per unit volume) of fish biomass, commercial value and dietary status were always identical for fresh- and brackish-water species groups, then salinization would have no impact on the welfare of poor households. Tropical field research on habitat salinity and fish biomass has revealed diverse patterns in different regions and ecosystems, but no clear, robust relationship between biomass yields in fresh and brackish water bodies (see for example Welcomme, et al. 2010; Nixon 1988; Marten and Polovina 1982). In addition, we have only spotty information about the relative abundance, commercial value and dietary status of the 83 fish species consumed by the poor in the Sundarbans region. Given the lack of robust research results and species-specific data, we cannot project the ultimate impact of salinization on fish consumption by poor households with any confidence. However, it does seem reasonable to assert that transitional risks for poor households will be higher in areas where the greatest changes in fish species will occur. And collective risks will be greater in areas where the settlement density of poor households is also high.14 13 An example is provided by Bombay Duck (Harpadon nehereus), a low price fish that is still caught in abundance and preferred by poor and middle class consumers all along the Bangladesh coast. On average, Bombay Duck accounts for 14 percent of daily fish sales. Using the IPCC A1B emissions scenario, Farnandes et al. (2015) have predicted a 35 percent reduction in production of this species in Banglades’s exclusive economic zone. 14 This is an issue of major concern as fishery experts in Bangladesh indicated that significant gain of  brackish fish species in the study region is unlikely to occur in a changing climate by 2050. Salinity is only  one  of  the  multiple  determinants  of  brackish  fish  behavior  and  habitats.  Wild  marine  and  brackish  fish  species  prefer  coastal  ecosystems  to  river  systems  because  of  their  feeding  habits  and  biology;  and  are  expected to move slowly over time to inland river systems, if at all. On the contrary, many freshwater fish  species have low swimming speed, prefer local habitats and will cease to survive with increase in salinity  (Robin  et  al.  2010).  Gain  et  al.  2008  also  reported  significant  decline  in  fish  diversity  with  increase  in  salinity  in  Sibsa  River  near  Paikgacha.  In  1975,  fresh  water  fish  species  near  Paikgacha  were  abundant,  but in 2005 the field sampling could not locate 17 fresh water species, including Labeo rohita, Catla catla,  Anabas  testudineus  and  Clarius  batrachus.  Experts  also  indicated  that  with  change  in  aquatic  salinity,  a  16 5.1 Fish Species Change Scenarios for Upazilas in the Sundarbans Region For each of the 83 species identified in Appendix Table A1, we build a digital map for 2012 that assigns 1 to pixels that satisfy the species’ stable habitat criterion (pixel salinity range falls within the species’ tolerance range) and 0 otherwise. We add across the 83 maps to determine total species with stable habitat in each of 101,600 pixels. Then we perform the same operations for all 27 salinity scenarios in 2050 and calculate percent changes (2012-2050) in total species for each pixel. Overlaying an administrative map shapefile provided by the Government of Bangladesh, we compute mean percent changes in the 27 scenarios for 110 upazilas in the Sundarbans region. 5.2 Poverty Incidence in the Sundarbans Region We assess collective risk using estimated total poverty populations for upazilas in 2011. These are the product of 2010 poverty incidence estimates provided by the World Bank (2014b) and 2011 population estimates from the Bangladesh Bureau of Statistics. Following World Bank (2014a), we use two standards to determine poverty incidence: the upper poverty line, for households whose food expenditures are at or below the food poverty line established by the Bangladesh Bureau of Statistics;15 and the lower poverty line, for extremely poor households whose total expenditures are at or below the food poverty line. 5.3 Risk Assessment for Upazilas in the Sundarbans Region To illustrate the range of results produced by this exercise, we employ the two bounding scenarios for 2050 that are mapped at the pixel level in Figure 1: least change (Scenario B1, GCM MIROC-3.2, SLR 27 cm, land subsidence 2 mm/year); and most change (Scenario A2, GCM few  coastal  fish  species  may  emerge  gradually  in  inland  water  but  their  harvesting  technology  is  costly  and not affordable to the poor.  15 See Report of the Household Income and Expenditure Survey/HIES 2010. Bangladesh Bureau of Statistics, Government of Bangladesh. 17 IPSL-CM4, SLR 32 cm, land subsidence 9 mm/year). We map the results for 110 upazilas in Figures 2 and 3. The maps illustrate two critical dimensions for priority-setting: percent change in species counts, and poverty populations identified using lower and upper poverty lines. Figure 2 overlays color-coded changes in fish species with black circles scaled by lower poverty line populations. Figure 2(a) displays the scenario with least change (B1), while 2(b) displays the scenario with most change (A2). Although the maps present a wealth of information, three patterns are immediately clear. First, the two scenarios exhibit a very similar pattern of species increase (colored blue) in the southwest, with growth somewhat more pronounced in 2(b). Second, the two scenarios exhibit widespread species decrease in both scenarios, and strikingly higher decrease rates in 2(b). Third, the distribution of the lower level poverty population is strikingly non-uniform across upazilas, with the largest concentrations in the center of the eastern region. Risk assessment should incorporate both species change and poverty population size, focusing particularly on upazilas which have high species loss rates and large poverty populations. Visual inspection reveals two obvious priority candidates in Figure 2(b): Lakshmipur in Chittagong Division, and Bhola in Barisal. Both have large extreme poverty populations (defined by the lower poverty line) and species loss rates greater than 50%. Elsewhere, the diversity of change rates and poverty populations makes it more difficult to identify clear patterns. This is also true of Figure 3, because poverty populations are less skew-distributed when we employ the upper poverty line. 18 Fig 2: Upazila change scenarios, lower poverty line populations (a) Least Change (b) Most Change 19 Fig 3: Upazila change scenarios, upper poverty line populations (a) Least Change (b) Most Change 20 To provide a clearer basis for identifying priority cases, we construct more general risk indicators for all 27 scenarios and 2 poverty definitions. First, we multiply the species change rate in each upazila by its share of the region’s poverty population to create a poverty-weighted species change index. To check for robustness, we generate index values for 110 upazilas in all 54 cases (27 scenarios, 2 poverty definitions) and calculate rank correlation coefficients within and across the two poverty groups.16 Table 5 presents summary statistics for the three correlation exercises. Table 5: Summary statistics: rank correlation coefficients for species change indices Upazilas: 110 Scenarios: 27 Poverty Line Min P10 P25 Median Mean P75 P90 Max Lower 0.744 0.816 0.835 0.886 0.892 0.948 0.985 0.999 Upper 0.757 0.822 0.838 0.892 0.895 0.947 0.986 0.999 Lower vs. Upper 0.744 0.820 0.839 0.892 0.896 0.954 0.991 0.999 These results suggest that our methodology is robust to changes in scenarios and poverty definitions. In all three exercises, the median and mean correlation coefficients are around .89; the first- and third-quartile correlations are .84 and .95, respectively; and the minimum correlation never falls below 0.74. Given these results, we believe that a summary index can provide useful information for identifying priority cases. Accordingly, we compute mean ranks for the 110 upazilas across all 54 cases and use the results to rank upazilas in three classes: species losses, species gains and no change. We provide complete tabulations of our results in Tables A2 (76 upazilas that lose species), A3 (11 upazilas that gain species), and A4 (23 upazilas with no change). Figures 4 and 5 display upazilas with the top-ten index values for species losses and gains. Among the upazilas with top-ten species loss indicators, nine are in Khulna (Bagerhat, Dighalia, Khalishpur, Kotwali, 16 We use rank correlations to eliminate potential outlier effects, and because rankings are the core identifier for priority assessment in any case. 21 Mollahat, Morrelganj, Rampal,17 Satkhira, Terokhada) and one is in Barisal (Char Fasson). All ten upazilas with top-ten species gain indicators are in Khulna (Tala, Assasuni, Batiagahata, Dacope, Dumuria, Kaliganj, Mongla, Paikgachha, Sharsha, Shyamnagar). Fig 4: Upazilas with species losses: top ten index values Fig 5 Upazilas with species gains: top ten index values 17 Our findings for Rampal are in line with the reduction in fish diversity noted by Gain et al. (2008). 22 Since 76 upazilas have projected species losses and only 11 have projected gains, it seems likely that the majority of poor households are in areas with projected losses. Table 6 confirms this difference, which turns out to be very large. Poverty populations in upazilas with losses are 4.0 and 6.6 million for lower and upper poverty lines, respectively. The comparative populations for upazilas with species gains are 0.7 million and 1.2 million, respectively. For both poverty lines, the ratio of populations with losses to those with gains is about 6:1. Table 6: Poverty populations by species change Poverty Population Species Change Lower Line Upper Line Loss 3,993,190 6,578,473 Gain 692,757 1,167,131 None 1,165,526 2,130,843 To provide more concrete illustrations, the figures 6 and 7 below show the minimum and maximum variants from our twenty seven salinity change scenarios to portray projected range changes for a variety of species that are important for fish consumption by poor households. 23 Fig 6: Range Changes for Illustrative Fish Species (0-5ppt) typically consumed by the poor Fish Species:  Puntius sophore Puntius ticto Channa punctatus Salmostoma baccila Xenentodon cancila Glossogobius giurus Nemacheilus botia Channa orientalis Chela taubuca High estimate: Scenario: A2, GCM: IPSL-CM4, SLR: 32 cm, land subsidence: 9mm/year Low estimate: Scenario: B1, GCM: MIROC 3.2, SLR: 27 cm, land subsidence: 2mm/year Fig 7: Range Changes for Illustrative Fish Species (5-20ppt) typically consumed by the poor Fish Species:  Coilia ramkoranti  Cynoglossus  cynoglossus  Harpadon nehereus  Liza parsia  Mugil cephalus  Setipinna taty  Setippina phasa  Sillago domina  High estimate: Scenario: A2, GCM: IPSL-CM4, SLR: 32 cm, land subsidence: 9mm/year Low estimate: Scenario: B1, GCM: MIROC 3.2, SLR: 27 cm, land subsidence: 2mm/year 24 6. Summary and Conclusions Data on water quality indicates river salinity increased significantly in southwest coastal region of Bangladesh over time (IWM 2003; Dasgupta et al. 2015). Scientists and hydrologists unanimously agree that river salinity in Sundarbans will increase due to sea level rise in a changing climate. In the absence of agreement among scientists about the time and spatial profile of climate change, in this paper we have used a detailed scenario analysis for the Sundarbans region to assess possible impacts of climate change and aquatic salinity on fish species habitats, and the poor communities that consume the affected fish species. Drawing on Dasgupta et al. (2015), we use a digital map of aquatic salinity for 2012 and 27 digital maps for 2050, projected from combinations of three IPCC climate change scenarios (B1, A1B, A2), three global circulation models (IPSL- CM4, MIROC3.2, ECHO-G) and three assumptions about the rate of subsidence in the Ganges Delta (2, 5 and 9 mm/year). Our exercise uses 101,600 pixels, at a resolution of 0.327 sq. km per pixel. We focus on 83 fish species that are consumed by households in the region. Using the salinity tolerance range for each species, we construct digital maps of its stable (12-month) habitats for 2012 and 27 scenarios in 2050. We add across maps to generate species counts for each pixel and compute percent changes for 2012-2050. Our results indicate two broad patterns of change, with brackish water expanding moderately into fresh water habitat in the western part of the region and more broadly in the eastern part. Increase in salinity is expected to have adverse impacts on reproductive cycle, reproductive capacity, extent of suitable spawning area, and feeding/ breeding/ longitudinal migration of fish species. 25 To assess the consequences for poor households, we overlay our results with an administrative map of Bangladesh and compute mean percent changes in fish species for 110 upazilas that lie within the region. We construct an impact indicator that weights these results by upazila poverty populations identified using two bounds: an upper poverty line, for households whose food expenditures are at or below the food poverty line established by the Bangladesh Bureau of Statistics; and the lower poverty line, for extremely poor households whose total expenditures are at or below the food poverty line. Our calculations encompass 54 cases (27 scenarios, 2 poverty definitions). We find that potential impact rankings are highly correlated, so we use the mean rank across 54 cases as a robust general impact indicator. This enables us to produce rank-orderings for 76 upazilas that lose fish species and 11 upazilas that gain species (23 upazilas exhibit no change). Among the 20 upazilas with top-ten loss and gain indices, 19 are in Khulna and one (a species loss case) is in Barisal. Our summary results provide striking evidence that projected aquatic salinization may have a strongly regressive impact on poor households in the Sundarbans region. For both poverty definitions, we find that poverty populations in upazilas that lose and gain species have a ratio of approximately 6:1. Given that fish is the main source of protein in the diet of 43.2 million poor people, and the chronic as well as acute malnutrition levels, as indicated by statistics on wasting and stunting of children in Bangladesh, are higher than the WHO’s thresholds for public health emergencies,18 our finding is serious and emphasizes the importance of mainstreaming climate change in relevant policies, action plans and programs in the country. As we note in the paper, we must attach one strong caveat to our results. Our measure of potential risk is simply the change in species count because we do not have good evidence on other 18 Government of Bangladesh: Strategic Plan for Health Population and Nutrition Sector Development Program (HPNSDP) 2011-2016, http://www.bma.org.bd/pdf/strategic_Plan_HPNSDP_2011-16.pdf 26 important factors: species-specific fishing yields, commercial values and dietary status of the poor. It is possible that these factors would reinforce our results, but it is also possible that they could be countervailing, perhaps strongly so. Inclusion of these factors should be a high priority for future research on aquatic salinization, fish habitat changes, and poverty impacts in the Sundarbans region. Our research also highlights the importance of systematic data collection for monitoring impacts of climate change on fish and other aquatic species. Nevertheless, to the best of our knowledge this paper presents the first thorough analysis of expected impacts of climate change and river salinity on habitats of 83 fish species. It is expected that this analysis will serve as a foundation for further analyses of climate change and fisheries in Bangladesh. The paper will contribute to multiple ongoing and future action plans and programs under the Environment Policy 1992,19 National Fisheries Policy 1998,20 the Coastal Zone Policy 2005,21 the Climate Change Action Plan 200922 and Strategic Plan for Health Population and Nutrition Sector Development Program 2011-2016 of Government of Bangladesh.23 It should also be noted that the Government of Bangladesh has already adopted the Ocean/Blue Economy initiative to promote sustainable and inclusive growth and employment opportunities in the maritime economic activities, and highlighted its important role in poverty 19 Bangladesh Environmental Polcy 1992: Conservation of habitats for fish (Stated Objective 3.8.1). 20 Bangladesh National Fisheries Policy 1998, Page 2: Stated objectives ae enhancement of the fisheries production, poverty alleviation through creating self-employment and improvement of socioeconomic conditions of the fisheries, fulfillment of the demand for animal protein, achievement of economic frowth through earning foreign currency by exporting fish and fish and fisheries, maintenance of ecological products’ balance and conservation of biodiversity. http://faolex.fao.org/docs/pdf/bgd149571.pdf 21 Bangladesh Coastal Zone Policy 2005: Provision of basic needs and opportunities for livelihoods (Framework 4.2a), Sustainable management of natural resources (framework 4.4c), http://lib.pmo.gov.bd/legalms/pdf/Costal-Zone- Policy-2005.pdf 22 Bangladesh Climate Change Strategy and Action Plan 2009: Research and knowledge management of impacts of climate change on ecosystems and biodiversity (Pillar 4.3), linkages between climate change, poverty and vulnerability (Pillar 4.5a), linkages between climate change, poverty and health to identify interventions to increase the resilience of the poor and vulnerable households to climate change (Pillar 4.5b), . http://www.climatechangecell.org.bd/Documents/climate_change_strategy2009.pdf 23 Bangladesh HPNSDSP 2011-2016: Action plans for mainstreaming nutrition services of the Directorate General of Health Policy (DGHS). 27 alleviation, ensuring food and nutrition security and sharing prosperity in the short, medium and long time frames.24 In this context, priorities have been assigned to increasing sustainable fishing capacity, promoting sustainable management of small-scale fisheries, supporting artisanal communities’ access to information, technology, finance, regulation and governance processes to ensure their year-round employment, and increasing the share of capture fisheries in fish production through protection and restoration of critical habitats (see Alam 2015 for details). It is well recognized that addressing climate change impacts on fisheries is critical for protection and restoration of critical habitats, increasing sustainable fishing capacity as well as promoting sustainable management of fisheries. Our paper with the baseline of fish habitats in 2012 and the detailed scenario analysis of possible impacts of climate change and aquatic salinity on fish species habitats for the Sundarbans region will provide a science-based approach essential for mainstreaming climate change in adaptive management and decision-making- essential for developing the Blue Economy. In light of our findings, introduction of coastal and/or sea fish breeding programs and sea ranching to enhance diversity of key species, establishment of conservation measures to protect fish breeding areas and nurseries, establishment of protected areas and marine reserves are expected to result in rewarding outcomes. 24 Alam 2014. 28 References Aerts LJ, A. Hassan, H. Savenije, and M. Khan. 2000. Using GIS tools and rapid assessment techniques for determining salt intrusion: STREAM - a river basin management instrument. Physics Chem Earth B Hydrol Oceans Atmos, 25: 265–273. Alam, M.F., Thomson, K.J., 2001. Current constraints and future possibilities for Bangladesh fisheries, Food Policy 26, pp.297-313. Alam, M. K 2015. Ocean/Blue Economy for Bangladesh. http://mofl.portal.gov.bd/sites/default/files/files/mofl.portal.gov.bd/page/d1b6c714_aee6_499f_a 473_c0081e81d7dc/Blue%20Economy.pdf , Accessed Januray 2016. Ali, Youssouf. 1999. Fish Resources Vulnerability and Adaptation to Climate Change in Bangladesh. In Karim, Zahurul, M. Asaduzzaman, and F. Mahtab, eds. Vulnerability and adaptation to climate change for Bangladesh. The Netherlands: Kluwer Academic Publishers, 113-124. Belton, B., M. Karim, S. Thilsted, M. Kondker, W. Collis, and M. Phillips. 2011. “Review of Aquaculture and Fish Consumption in Bangladesh.” Studies and Reviews 2011–53, The World Fish Center. Bhuiyan, Md. J.A.N and Dushmanta Dutta. 2011. Assessing impacts of sea level rise on river salinity in the Gorai river network, Bangladesh. Estuarine, Coastal and Shelf Science, 96(1): 219-227. CEGIS. 2006. Impact of Sea Level Rise on Landuse Suitability and Adaptation Options. Ministry of Environment and Forests, Government of Bangladesh. Report on Coastal Land Use Zoning in Bangladesh. Chowdhury, M. T. H., Z. P. Sukhan, and M. A. Hannan. "Climate change and its impact on fisheries resource in Bangladesh." Procedding of International Conference on Environmental Aspects of Bangladesh (ICEAB’10). 2010. Dasgupta, Susmita, Farhana Akhter Kamal, Zahirul Huque Khan, Sharifuzzaman Choudhury and Ainun Nishat. 2015. River Salinity and Climate Change: Evidence from Coastal Bangladesh. ), in Asia and the World Economy - Actions on Climate Change by Asian Countries, 2015 (eds. John Whalley and Jiahua Pan), World Scientific Press, 205-242. Dasgupta, Susmita, Mainul Huq and David Wheeler. 2014. Facing the Hungry Tide: Climate Change, Livelihood Threats and Household Responses in Coastal Bangladesh. Policy Research Working Paper (forthcoming), Development Research Group, World Bank. Farnandez, Jose A., Susan Kay, Mostafa A.R. Hossain, Munir Ahmed, William W.L. Cheung, Attila N. Lazar and Manuel Barange. 2015. Projecting marine Fish Production and Catch Potential in Bangladesh in the 21st Century Under Long-term Environmental Change and 29 management Scenario. ICES Journal of Marine Science. http://icesjms.oxfordjournals.org/content/early/2015/11/25/icesjms.fsv217.abstract, Accessed December 2015. Gain Animesh K., M.N. Uddin and P. Sana. 2008. Impact of River Salinity on Fish Diversity in the South-West Coastal Region of Bangladesh. International Journal of Ecology and Environmental Sciences. 34 (1):49-54. Hassan, A. and Md. A.R. Shah. 2006. Impact of Sea Level Rise on Suitability of Agriculture and Fisheries. Center for Environmental and Geographic Information Services, Dhaka, Bangladesh. Mimeo. Hussain, Asad. M., Saiful, I.A.K.M., Hassan A.M. and Balakrishnan, B. 2013. Changes of the seasonal salinity distribution at the Sundarbans coast due to impact of climate change. 4th International Conference on Water & Flood Management (ICWFM-2013), 637. IWM. 2003. Sundarban Biodiversity Conservation Project: Surface Water Modeling, Final Report. Institute of Water Modeling, Ministry of Environment and Forests, Government of Bangladesh. Kasim, H. M. 1979. Salinity tolerance of certain freshwater fishes. Madurai Kamaraj University, Madurai – 625021. India. Mimeo. Marten, Gerald and Jeffrey Polovina. 1982. A comparative study of fish yields from various tropical ecosystems, p. 255-289. In Pauly, D. and G.!. Murphy (eds.) Theory and management of tropical fisheries. ICLARM Conference Proceedings 9, 360 p. International Center for Living Aquatic Resources Management, Manila, Philippines. Ministry of Environment and Forest (MoEF), Government of Bangladesh. 2010. Integrated Resources Management Plans for the Sundarbans (2010-2020). Volume I. http://pdf.usaid.gov/pdf_docs/pnaec417.pdf Mustafa, M.G. 2003. Trophic Model of the Coastal Ecosystem in the waters of Bangladesh, Bay of Bengal, p. 263-280. In G. Silvestre, L. Garces, I. Stobutzki, M. Ahmed, R.A.V. Santos, C. Luna, L. Lachica-Alino, V. Christensen, D. Pauly and P. Munro (eds) Assessment, Management and Future Directions of Coastal Fisheries in Asian Countries. WorldFish Center Conference Proceedings 67. (http://www.worldfishcenter.org/trawl/index.asp). Mustafa, M.G. and M. Prava Dey. 1994. The effect of environmental factors in the ecosystem of the Bay of Bengal, Bangladesh. Paper presented in the 9th National Zool. Conf. of the Zoological Society of Bangladesh`` 26-28 Jan. 1994. (Souvenir Abstract No. 61.). Nicholls, R. J., C. W. Hutton, A.N. Lazar, M.M. Rahman, M. Salehin and T. Ghosh. 2013. "Understanding climate change livelihoods in coastal Bangladesh." Hydrolink .2 (2013): 40-42. Nixon, Scott. 1988. Physical energy inputs and the comparative ecology of lake and marine ecosystems. Limnology and Oceanography, 33(4, part Z), 1988, 1005-1025. 30 Nobi, N. and Gupta, A. D. (1997), Simulation of Regional Flow and Salinity Intrusion in an Integrated Stream-Aquifer System in Coastal Region: Southwest Region of Bangladesh. Ground Water, 35: 786–796. Petersen, L. and S. Shireen. 2001. Soil and water salinity in the coastal area of Bangladesh. Bangladesh Soil Resource Development Institute. Rahman, M. R. and Asaduzzaman, M. 2010. Ecology of Sundarban, Bangladesh. J. Sci. Foundation, 8(1&2): 35-47, June-December 2010. Robin L. W., Ian G. C., David, C., Christophe, B., Simon, F. S., Ashley, H. and Lorenzen, K. 2010. Inland Capture fisheries. Review, Phil. Trans. R. Soc. B (2010) 365, 2881–2896. Sarwar, Md. Golam Mahabub, 2005. Impacts of sea Level rise on Coastal Zone of Bangladesh. Masters Thesis. http://fs.wa4.lucklaboratories.com/placemarks/files/225/golam_sarwar.pdf . Accessed December 2015. Shah, M. S., K. A. Huq, and B. S. M. Rahaman. 2010. "Study on the conservation and management of fisheries resources of the Sundarbans." Integrated Protected Area Co- Management (IPAC), Bangladesh. http://www.google.com.sg/url?sa=t&rct=j&q=&esrc=s&source=web&cd=4&ved=0ahUKEwi1n 8z- tPvJAhXOBo4KHU37AwIQFggwMAM&url=http%3A%2F%2Fmarineworldheritage.unesco.or g%2Fwp-content%2Fuploads%2F2012%2F01%2FUSAID_Integrated-Protected-Area-Co- Management-IPAC-2010-english1.pdf&usg=AFQjCNFnABEpfHZVERyLiNAsUHtV1HIckw Thilsted, S.H. 2010. “Improved Management, Increased Culture and Consumption of Small Fish Species Can Improve Diets of the Rural Poor.” In Sustainable Diets and Biodiversity: Directions and Solutions for Policy, Research and Action Farming the Waters for People and Food. Edited by B. Burlingame and S. Dernini. Proceedings of the International Scientific Symposium on Biodiversity and Sustainable Diets United against Hunger, FAO Headquarters, Rome. November 3–5. Pages 176–81. Thilsted, S.H. 2012. “The Potential of Nutrient-Rich Small Fish Species in Aquaculture to Improve Human Nutrition and Health.” In Farming the Waters for People and Food. Edited by R.P. Subasinghe, J.R. Arthur, D.M. Bartley, S.S. De Silva, M. Halwart, N. Hishamunda, C.V. Mohan, and P. Sorgeloos. Proceedings of the Global Conference on Aquaculture 2010, Phuket, Thailand. September 22–25. Pages 57–73. UK DEFR (U.K. Department of Environment, Food and Rural Affairs). 2007. Investigating the Impact of Relative Sea-Level Rise on Coastal Communities and Their Livelihoods in Bangladesh. Dhaka: Institute of Water Modelling in Bangladesh and Center for Environment and Geographic Information Services in Bangladesh. 31 Welcomme, Robin L. Ian G. Cowx, David Coates, Christophe Béné, Simon Funge-Smith, Ashley Halls, and Kai Lorenzen. 2010. Inland Capture Fisheries Philosophical Transactions of the Royal Society B: Biological Sciences. 2010 Sep 27; 365(1554): 2881–2896. World Bank. 2000. Bangladesh: Climate Change and Sustainable Development. Report No. 21104-BD. Rural Development Unit, South Asia Region. December. World Bank. 2011. Survey questionnaire: Livelihoods, Biodiversity Conservation, Adaptation to Climate Variability and Socio- economic Development in Sundarbans –Bangladesh. World Bank Human Development Department and South Asia Sustainable Development Department. World Bank. 2014a. Bangladesh Poverty Map 2010-Technical Report. Dhaka: World Bank, Bangladesh Bureau of Statistics, World Food Programme http://www- wds.worldbank.org/external/default/WDSContentServer/WDSP/IB/2014/11/11/000442464_2014 1111221335/Rendered/PDF/904870v20Bangl0LIC000Sept004020140.pdf , Accessed December 2015. World Bank. 2014b. Bangladesh Poverty Map 2010. Dhaka: World Bank, Bangladesh Bureau of Statistics, World Food Programme http://www- wds.worldbank.org/external/default/WDSContentServer/WDSP/IB/2014/09/10/000442464_2014 0910105709/Rendered/PDF/904870WP0WB0Po00Box385319B00PUBLIC0.pdf, Accessed December 2015. 32 Table A1: Salinity tolerance ranges: Fish species found and consumed in southwest coastal region and Sundarbans Fresh Water Tolerant Species Salinity Tolerance (ppt) Scientific Name Bangladesh Name English Name Min Max Clarias batrachus Magur Walking catfish 0 2 Heteropneustes fossilis Shing Stinging catfish 0 2 Anabas testudineus Koi Climbing perch 0 5 Catla catla Catla Carp 0 5 Channa orientalis Gachua Snakehead 0 5 Channa punctatus Taki Spotted snakehead 0 5 Channa striatus Shol Snakehead murrel 0 5 Chela laubuca Kash khaira Indian grass barb 0 5 Cirrhinus reba Bata Reba carp 0 5 Clupisoma garua Ghaura River catfish 0 5 Dermogenys pussilus Ekthota Wrestling halfbeak 0 5 Eutropiichthys vacha Bacha River catfish 0 5 Gagata cenia Kauwa River catfish 0 5 Labeo calbasu Kalibaus Carp 0 5 Labeo gonius Goinna Carp 0 5 Mystus tengara Bajari tengra Long bled catfish 0 5 Mystus vittatus Tengra Catfish 0 5 Nandus nandus Meni Perch 0 5 Nemacheilus botia Loach Zipper loach 0 5 Notopterus notopterus Foli Bronze featherback 0 5 Ompak bimaculatus Kani pabda Butter catfish 0 5 Ompok pabda Pabda Butter catfish 0 5 Puntius sophore Jatputi Pool barb 0 5 Puntius ticto Tit puti Ticto barb 0 5 Salmostoma bacaila Katari Minnow 0 5 Wallago attu Boal Freshwater shark 0 5 Xenentodon cancila Kakila Garfish 0 5 Aorichthys aor Ayre Long barb catfish 0 10 Gagata gagata Gang tengra Catfish 0 10 Glossogobius giurus Baila Tankqoby 0 10 Macrobrachium birmanicus Nazari icha, shul icha Freshwater prawn 0 10 Macrobrachium dolichodactylus Icha Freshwater prawn 0 10 Macrobrachium lamarrei Thenga icha Freshwater prawn 0 10 Macrobrachium malcolmsonii Boro icha Indian freshwater prawn 0 10 Macrobrachium villosimanus Dimua icha Dimua river prawn 0 10 Macrobrachiurn rudis Kucha chingri Hairy river prawn 0 10 Monopterus cuchia Kuicha baim Mud eel 0 10 Mystus bleekeri Golsha tengra Long bled catfish 0 10 Mystus tengara Tengra Catfish 0 10 Nematopalaernon tenuipes Gura icha Spider prawn 0 10 Pseudambassis ranga Lal chanda Indian glassy perchlet 0 10 Himantura fluviatilis Saplapata Gangetic stingray 0 15 Pellona ditchela Choikka Indian pillona 0 15 Palaemon styliferus Gura icha Freshwater prawn 0 20 Scylla serrata Kakra Mud crab 0 20 Thryssa dussumieri Phasa Dussumiers thryssa 0 20 Saline Water Tolerant Species 33 Salinity Tolerance (ppt) Scientific Name Bangladesh Name English Name Min Max Apocryptes bato Chiring Goby 5 10 Odontamblyopus rubicandas Lal chewa Irubicundus ee!goby 5 10 Parapocryptes batoides Chewa, chirin Goby 5 10 Plotosus Canius Kaim Magur Canine ell tail fish 5 20 Arius caelatus Mad, kata Engraved cat fish 5 20 Arius gagora Mad , kata Gagor cat fish 5 20 Arius thalassinus Mad , kata Giant sea cat fish 5 20 Coilia ramkoranti Olua Tepertail anchovy 5 20 Cynoalossus lingua kukurjib Long tonguesole 5 20 Cynoglossus cynoglossus Kukurjib Gangetic tonguesole 5 20 Eleuthronema tetradactylum Thailla Fourfingor throadfin 5 20 Harpadon nehereus Loytta Bombay duck 5 20 Lates calcarifer Bhetki, koral Seabass, barramundi 5 20 Liza parsia Pashia,bata Gold spot mullet 5 20 Liza spp Bata Mullet 5 20 Mugil cephalus khorul bata Flathoad grey mullet 5 20 Mystus gulio Guilla,nuna tengra Long-whiskered catfish 5 20 Pangasius pangasius Pangas Fatty cat fish 5 20 Polynemus indicus Lakhua Indian threadfin 5 20 Rhinomugil corsula Kholla,bata Yellow tail mullet 5 20 Scatophagus argus Bishtara Spotted scat 5 20 Setipinna taty Tailla phasa Scally hair fin anchovy 5 20 Setippina phasa Phasa Gangetic hairfin anchovy 5 20 Sillago domina Hundra, tolar dandi Ladyfish 5 20 Macrobrachiurn rosenbergii Golda chingri Giant freshwater prawn 5 25 Lepturacanthus savala Chhuri Ribbonfish 5 30 Panna microdon Poa Panna croker 5 30 Pomadasys maculatus Guti datina Blotched grunt 5 30 Tenualosa ilisha Ilish Hilsa shad 5 30 Therapon jarbua Barguni Therapon porch 5 30 Trichiurus leopturus Buri Ribbon fish 5 30 Johnius sp Poa mach Jew fish 5 30 Penaeus Indicus Chaga chingri Indian white shrimp' 10 30 Metapenaeus lysianassa Hanny Brown shri mp 10 35 Metapenaeus monoceros Horina chingri Brown shrimp 10 35 Penaeus monodon Bagda chingri Tiger shrimp- 10 35 Parapenaeopsis uncta Kddi chingri Uncta shrimp 15 35 Sources: Hussain et al. 2013; Rahman and Asaduzzaman 2010; MoEF 2010: Robin et al. 2010; Gain et al. 2008; Mustafa 2003: Mustafa and Dey 1994; Kasim 1979. 34 Table A2: Impact indicator ranks for upazilas with fish species losses Percent Species Loss (27 Scenarios) Poverty Population Lower Upper Rank Division District Upazila Geocode Mean Median Max Min Line Line 1 Khulna Satkhira Satkhira 48782 -47.99 -47.66 -53.06 -43.50 119,832 198,644 2 Khulna Khulna KCC (Kotwali) 44751 -72.82 -73.68 -73.68 -63.64 42,195 79,678 3 Khulna Khulna Terokhada 44794 -51.13 -51.05 -52.21 -51.05 35,013 57,888 4 Khulna Bagerhat Bagerhat 40108 -31.08 -31.26 -42.26 -16.82 49,548 95,634 5 Barisal Bhola Char Fasson 10925 -41.17 -51.00 -55.63 0.64 68,009 128,715 6 Khulna Khulna KCC (Khalishpur) 44745 -43.03 -42.11 -73.68 -15.79 38,845 67,938 7 Khulna Bagerhat Rampal 40173 -37.58 -38.91 -49.41 -29.04 34,867 63,691 8 Khulna Bagerhat Morrelganj 40160 -16.75 -16.45 -25.03 -9.52 79,536 136,978 9 Khulna Bagerhat Mollahat 40156 -31.39 -31.64 -37.29 -23.83 34,944 60,335 10 Khulna Khulna Dighalia 44740 -40.30 -40.36 -42.58 -38.11 25,313 45,425 11 Khulna Bagerhat Fakirhat 40134 -37.63 -36.87 -45.18 -30.81 26,455 50,155 12 Barisal Bhola Lalmohan 10954 -32.16 -38.15 -56.32 0.00 43,151 78,921 13 Khulna Khulna KCC (Sonadanga) 44785 -73.68 -73.68 -73.68 -73.68 12,245 32,374 14 Barisal Bhola Manpura 10965 -52.01 -52.33 -56.82 -44.86 14,857 25,119 15 Khulna Jessore Abhaynagar 44104 -17.56 -17.04 -28.02 -9.86 41,727 94,476 16 Khulna Khulna Rupsa 44775 -18.55 -16.91 -33.76 -8.58 36,263 66,243 17 Khulna Khulna Phultala 44769 -41.36 -41.82 -45.12 -32.17 14,260 28,268 18 Barisal Bhola Tazumuddin 10991 -42.89 -56.03 -56.82 -3.56 14,344 28,308 19 Barisal Pirojpur Mothbaria 17958 -7.75 -8.08 -10.81 -4.03 67,287 99,880 20 Khulna Khulna Koyra 44753 -9.51 -8.58 -17.56 -3.72 56,434 95,220 21 Barisal Barisal Barisal 10651 -3.39 -4.35 -4.35 -0.23 163,375 262,982 22 Barisal Barisal Mehendiganj 10662 -3.35 -2.82 -6.05 -1.37 150,523 193,874 23 Barisal Patuakhali Galachipa 17857 -11.58 -9.12 -20.59 2.16 52,059 93,995 24 Khulna Satkhira Debhata 48725 -13.06 -13.03 -13.13 -12.93 34,473 54,029 25 Barisal Barisal Bakerganj 10607 -3.32 -4.33 -4.33 -0.23 132,443 173,870 26 Barisal Patuakhali Kalapara 17866 -17.65 -19.15 -22.98 -8.39 23,070 48,280 Percent Species Loss (27 Scenarios) Poverty Population Lower Upper Rank Division District Upazila Geocode Mean Median Max Min Line Line 27 Khulna Khulna Khan Jahan Ali 44748 -28.27 -27.38 -35.82 -24.21 13,661 25,939 28 Khulna Narail Kalia 46528 -16.48 -16.44 -24.69 -7.67 21,360 51,307 29 Barisal Bhola Burhanuddin 10921 -18.21 -12.53 -49.48 -0.24 38,119 66,182 30 Khulna Jessore Keshabpur 44138 -6.59 -6.61 -11.87 -2.16 51,671 106,382 31 Barisal Barisal Hizla 10636 -4.07 -4.35 -4.35 -1.90 72,308 91,006 32 Khulna Bagerhat Kuchua 40138 -11.05 -8.99 -26.11 -0.46 22,895 41,230 33 Chittagong Chandpur Haimchar 21347 -4.35 -4.35 -4.35 -4.35 44,926 67,169 34 Barisal Bhola Bhola 10918 -10.05 -0.10 -43.01 -0.10 153,696 211,816 35 Barisal Barisal Muladi 10669 -3.06 -4.02 -4.18 0.00 77,076 101,719 36 Chittagong Lakshmipur Roypur 25158 -10.39 -4.35 -31.52 -4.35 23,939 45,952 37 Khulna Jessore Manirampur 44161 -2.15 -1.89 -5.05 -0.50 80,980 167,803 38 Barisal Patuakhali Bauphal 17838 -3.90 -4.19 -4.39 -2.76 42,295 73,028 39 Dhaka Shariatpur Gosairhat 38636 -2.17 -2.45 -3.44 -0.17 64,170 91,919 40 Barisal Bhola Daulatkhan 10929 -20.05 -5.14 -56.82 0.00 30,342 51,076 41 Barisal Pirojpur Pirojpur 17980 -2.26 -1.69 -5.35 -0.63 46,916 69,802 42 Barisal Patuakhali Patuakhali 17895 -2.90 -4.35 -4.35 0.00 73,736 116,774 43 Barisal Patuakhali Dashmina 17852 -6.28 -3.54 -13.59 -3.29 13,943 26,899 44 Barisal Barguna Barguan 10428 -3.26 -4.37 -5.10 -0.45 25,873 50,178 45 Barisal Jhalokati Nalchity 14273 -2.90 -4.35 -4.35 0.00 62,906 90,004 46 Barisal Barguna Amtali 10409 -2.09 -2.37 -4.04 -0.03 32,496 61,743 47 Barisal Pirojpur Bandaria 17914 -2.27 -2.85 -4.35 0.00 44,300 62,227 48 Barisal Barisal Babuganj 10603 -2.90 -4.35 -4.35 0.00 51,653 68,356 49 Barisal Barisal Wazirpur 10694 -1.60 -2.01 -3.09 0.00 88,815 122,414 50 Barisal Jhalokati Jhalokati 14240 -2.89 -4.35 -4.35 0.00 48,246 81,563 51 Khulna Bagerhat Chitalmari 40114 -1.22 -1.01 -2.53 -0.57 41,643 69,405 52 Chittagong Chandpur Chandpur 21322 -1.61 -2.90 -2.90 0.00 120,673 211,993 53 Barisal Barguna Patharghata 10485 -3.86 -3.61 -4.76 -2.81 10,000 21,147 36 Percent Species Loss (27 Scenarios) Poverty Population Lower Upper Rank Division District Upazila Geocode Mean Median Max Min Line Line 54 Barisal Pirojpur Nazirpur 17976 -0.51 -0.45 -0.84 -0.26 66,029 92,910 55 Barisal Jhalokati Rajapur 14284 -2.66 -3.89 -4.35 0.00 44,251 62,367 56 Barisal Pirojpur Nesarabad 17987 -1.69 -1.73 -3.80 0.00 63,521 91,377 57 Barisal Barisal Banaripara 10610 -1.83 -2.16 -3.69 0.00 56,460 77,354 58 Barisal Barisal Gaurnadi 10632 -1.31 -1.79 -2.35 0.00 75,246 104,665 59 Barisal Barguna Bamna 10419 -2.95 -4.35 -4.35 -0.05 7,081 13,605 60 Barisal Jhalokati Kanthalia 14243 -2.88 -4.35 -4.35 0.00 26,594 42,501 61 Khulna Narail Narail 46576 -0.87 -0.82 -1.29 -0.64 16,645 47,207 62 Chittagong Lakshmipur Lakshmipur 25143 -12.63 0.00 -56.82 0.00 197,114 312,098 63 Khulna Narail Lohagara 46552 -0.17 -0.12 -0.39 -0.04 17,373 45,490 64 Barisal Barguna Betagi 10447 -2.90 -4.35 -4.35 0.00 12,066 22,960 65 Barisal Patuakhali Mirzapur 17876 -2.90 -4.35 -4.35 0.00 11,685 21,665 66 Barisal Pirojpur Kawkhali 17947 -1.67 -0.97 -4.35 0.00 27,771 36,608 67 Barisal Patuakhali Dumki 17855 -2.90 -4.35 -4.35 0.00 9,256 15,544 68 Dhaka Shariatpur Bhedarganj 38614 -0.19 -0.06 -0.62 0.00 96,989 142,571 69 Khulna Jessore Jessore 44147 -0.02 -0.01 -0.15 0.00 121,835 262,243 70 Chittagong Noakhali Hatiya 27536 -9.18 0.00 -56.82 4.55 26,695 72,394 71 Dhaka Gopalganj Gopalganj 33532 0.00 0.00 -0.02 0.00 86,690 141,387 72 Barisal Barisal Agailjhara 10602 -0.06 0.00 -0.22 0.00 57,092 76,372 73 Dhaka Shariatpur Damudya 38625 -0.09 0.00 -0.38 0.00 32,047 52,212 74 Khulna Khulna KCC (Daulatpur) 44721 -0.66 0.00 -8.93 0.00 19,565 38,792 75 Dhaka Madaripur Kalkini 35440 -0.05 0.00 -0.23 0.00 43,175 90,722 76 Khulna Satkhira Kalaroa 48743 -0.10 -0.09 -0.97 0.58 68,304 109,476 37 Table A3: Impact indicator ranks for upazilas with fish species gains Percent Species Gain (27 Scenarios) Poverty Population Lower Upper Rank Division District Upazila Geocode Mean Median Min Max Line Line 1 Khulna Satkhira Tala 48790 75.73 83.01 38.79 100.77 86,648 135,519 2 Khulna Khulna Dacope 44717 29.82 29.96 25.48 32.77 37,927 67,781 3 Khulna Khulna Batiagahata 44712 27.35 27.91 15.85 36.67 38,974 69,535 4 Khulna Khulna Paikgachha 44764 18.37 18.97 15.70 20.66 57,780 105,145 5 Khulna Khulna Dumuria 44730 16.45 17.84 2.23 30.50 59,912 113,711 6 Khulna Satkhira Assasuni 48704 10.48 12.62 2.21 16.95 86,001 130,077 7 Khulna Satkhira Kaliganj 48747 5.09 5.44 3.13 6.56 87,140 131,947 8 Khulna Bagerhat Mongla 40158 11.03 10.65 8.94 14.52 31,005 57,230 9 Khulna Satkhira Shyamnagar 48786 1.88 1.96 1.33 2.36 107,570 159,764 10 Khulna Jessore Sharsha 44190 0.39 0.40 0.34 0.43 66,218 139,262 11 Khulna Bagerhat Sarankhola 40177 1.27 1.79 -1.39 3.17 33,582 57,160 38 Table A4: Upazilas with no change in fish species Percent Species Change (27 Scenarios) Poverty Population Lower Upper Division District Upazila Geocode Mean Median Min Max Line Line Dhaka Faridpur Nagarkandi 32962 0 0 0 0 37,996 71,045 Dhaka Munshiganj Lahajang 35944 0 0 0 0 32,008 53,505 Dhaka Faridpur Bhanga 32910 0 0 0 0 44,035 86,776 Chittagong Chandpur Uttar Matlab 21379 0 0 0 0 83,528 145,736 Khulna Jessore Chaugachha 44111 0 0 0 0 47,894 99,026 Khulna Jhenaidah Kaliganj 44433 0 0 0 0 27,107 67,768 Dhaka Gopalganj Muksudpur 33558 0 0 0 0 86,532 134,574 Khulna Jessore Jhikargachha 44123 0 0 0 0 53,505 116,275 Dhaka Munshiganj Munshiganj 35956 0 0 0 0 57,873 118,045 Dhaka Gopalganj Kashiani 33543 0 0 0 0 51,073 81,177 Dhaka Gopalganj Tungipara 33591 0 0 0 0 26,333 42,980 Dhaka Madaripur Rajoir 35480 0 0 0 0 35,221 71,815 Khulna Jhenaidah Maheshpur 44471 0 0 0 0 31,256 78,473 Dhaka Shariatpur Naria 38665 0 0 0 0 70,651 111,421 Khulna Jessore Bagher para 44109 0 0 0 0 45,331 92,181 Dhaka Madaripur Shib Char 35487 0 0 0 0 64,280 123,469 Dhaka Faridpur Boalmari 32918 0 0 0 0 55,181 100,867 Dhaka Faridpur Alfadanga 32903 0 0 0 0 16,787 32,382 Dhaka Madaripur Madaripur 35454 0 0 0 0 60,509 121,017 Dhaka Gopalganj Kotalipara 33551 0 0 0 0 63,847 100,495 Khulna Magura Shalikha 45585 0 0 0 0 40,424 72,337 Dhaka Shariatpur Zanjira 38694 0 0 0 0 67,713 104,770 Dhaka Shariatpur Palong 38669 0 0 0 0 66,442 104,709 39