Do Deltas Remain Attractive? Testing the Migration to Coast - - PDF document

Do Deltas Remain Attractive? Testing the Migration to Coast Hypothesis Abu, Mumuni and Codjoe, N.A. Samuel Regional Institute for Population Studies, University of Ghana Corresponding author: Mumuni Abu. P. O. Box LG 96, University of Ghana, Legon. mabu@ug.edu.gh Abstract Earlier studies have found net movement into coastal regions globally - net loss of populations in drylands and mountain areas, and net in-migration to coastal areas. This is because while deltas are at risk from environmental degradation, they tend to have large urban areas which have such economic primacy that they are always protected and act as sources for net in-

• migration. This paper examines whether delta regions continue to be a magnet for populations

and the drift to the coast is continuing. We hypothesise that urban Districts have less or zero net out-migration, and therefore more net in-migration compared to rural districts. We do so by examining a range of deltas, including the most densely populated large deltas in the world (Ganges-Brahmputra-Meghna – Bangladesh and India) along with smaller deltas (Mahanadi and Volta) and estimating net migration from the most recent census interval. Keywords: Coastal regions, net-migration, deltas, vulnerability, climate change Introduction Coastal areas continue to host significant concentrations of people and livelihoods in spite of their high exposure to environmental hazards (McGranahan, Balk, & Anderson, 2007; Neumann et al., 2015). Earlier studies have found net movement into coastal regions globally

• net loss of populations in drylands and mountain areas, and net in-migration to coastal areas

((Nicholls & Cazenave, 2010; Seto, 2011). This is because while coastal areas are at risk from

environmental degradation, they tend to have large urban areas which have such economic primacy that they are always protected and act as sources for net in-migration (Seto, 2011). We hypothesise, in this paper, that coastal regions will not necessarily continue to be a magnet for populations and that the drift to the coast may not continue. We do so by examining a range

• f deltas, including the most densely populated large deltas in the world (Ganges-Brahmaputra-

Meghna – Bangladesh and India) along with smaller deltas (Mahanadi, India and Volta, Ghana). We hold this contrary position due to the following. First, the literature on migration into deltas in Africa and Asia reveal deltas are receiving areas because of the presence of “primate cities” which gives them greater economic advantage over other areas (Nicholls & Cazenave, 2010; Seto, 2011). However, not all deltas contain major or capital cities, as shown in the study by Seto (2011), and so may not necessarily act as points of attraction for migrants. On the contrary, there may be major economic settlements which, due to their proximity to the deltas, may attract migrants from the delta. Apart from the contrary hypothesis to the predominant position of earlier studies, we make a methodological contribution by using population census data at the local level which is a departure from previous studies. These studies used population estimates which were not so accurate because they used low-resolution data which makes it impossible to accurately estimate population density and mobility at the level of local administrative units (McGranahan et al., 2007; Small & Nicholls, 2003). They were however, an improvement upon earlier estimates of coastal population dynamics which were based on figures that were “unsubstantiated” but “widely repeated statements” (Small & Nicholls, 2003, p. 584). In this paper, we use data from the two most recent population and housing censuses of Bangladesh and India (2001 & 2011) and Ghana (2000 & 2010), to estimate net migration for the GBM, Mahanadi and Volta deltas respectively. The paper commences with a synthesis on migration

in deltas and a demographic analysis of the deltas with reference to their respective countries to set the context within which migration occurs. Climate Change and Migration in deltas Global climate change has dire implications for delta populations which are considered highly vulnerable to environmental dynamics as well as socioeconomic challenges, particularly in developing contexts (Nicholls & Cazenave, 2010). In the most extreme case scenario, large- scale displacement of people living in deltas is expected. Deltaic regions are among the most vulnerable types of coastal environment due to the coincidence of vulnerable physical characteristics (i.e. low elevation and high flood probability, significant land erosion and gain, dependence on fluvial inputs of water and sediment, high sensitivity to small changes such as climate) and socio-economic characteristics (i.e. high population density, high prevalence of poverty, gender inequalities, low levels of socio-economic development and lack of connectivity with the main market places). Climate change impacts could reinforce many of the baseline stresses that already pose a serious impediment to development in deltas (Agrawala et al., 2003). These include heavier and more erratic rainfall leading to increased flooding and river bank erosion; warmer average temperatures; and changing intensity of tropical cyclones with higher wind speeds and storm surges (MoEF, 2008). The anticipated sea level rise in the bay of Bengal, for instance, is expected to submerge low lying land, increasing the penetration

• f storm surges and increasing saline intrusion (Karim and Mimura, 2008; Khan et al., 2011).

Also, extensive human activities interfere with the integrity of deltas’ naturally dynamic water and sedimentary systems, thus increasing the risk of relative sea-level rise, inundation and erosion (Church et al., 2013; Tessler et al., 2015). These anthropogenic geophysical modifications interact with socioeconomic characteristics of populations to determine the

• verall risks in deltas (Tessler et al., 2015). Deltas have some of the highest population

densities in the world with about 500 million, often poor residents. The adaptive strategies available to delta residents (e.g., disaster risk reduction by building shelters, or land and water use management) may also exacerbate gender inequalities, and may not be adequate to cope with pervasive, systemic, or sudden changes associated with climate change. Hence, large movements of people are often projected from deltas under climate change. A simple projection

• f existing trends suggests that more than 8 million people could be displaced across deltas

globally by 2050, with the Ganges-Brahmaputra-Meghna (GBM), Nile and Mekong deltas having the largest estimated displacement (Ericson et al., 2006). With additional climate- induced rises in sea level, tens of millions of men and women could be displaced during the 21st Century in the GBM and Nile deltas alone (Milliman et al., 1989; Woodroffe et al., 2006). Ericson et al. (2006) studied 12 deltas in Africa and Asia and estimated that 5.4 million people might be displaced by 2050 based on observed trends of sea-level rise and subsidence out of a global total of 8.7 million displaced people. Globally, coastal deltas are popular destinations for migrants due to the immense economic and social opportunities that are available in these areas. Historically, deltas are known to have very productive ecosystems which have attracted human settlement and agricultural activities. This has overtime transformed the global major deltas from agrarian economies into industrial and service-driven cities which continue to attract migrants (Small and Nicholls, 2003; Okonjo- Iweala and Osafo-Kwaako, 2007). Migration into deltas is driven by spatial inequalities between receiving delta areas and their sending areas. It is expected that with continuous urbanisation and a built up momentum, primate cities in these deltas will continue to attract in- migrants (De Souza et al., 2015; Seto, 2011).

Migration is already an established household adaptation to environmental and economic

• change. This can be both a successful form of adaptation, increasing the resilience of the

migrant household, and unsuccessful, perpetuating vulnerability in a new location with differential impacts on men and women. Population growth, poor access to education and employment opportunities, and environmental change all interact and result in complex patterns of migration to and from deltas. Thus, while the world’s coastal zones generally seem to have witnessed significant inward migration since 1970 compared to inland regions (McGranahan et al., 2007; de Sherbinin et al., 2012), it is necessary to identify that deltas have significantly different ecosystems and sensitivity to climate change impacts. The observed inward movement into coastal areas may not apply, particularly for rural delta areas in developing countries. Migration is a complex process (Foresight, 2011), which has a non-linear relationship with climate change and other drivers. The migration outcome of any environmental change depends on the characteristics of the household and the individual migrant, the nature and speed of the environmental change and the socio-economic and geographic context (Gray & Mueller, 2012; Feng et al., 2010; Black et al., 2012; Brown, 2008). Also, the impacts of environmental change vary across different groups of people and genders (Martin, 2013). Migration due to climate change will interact with existing migratory processes in all the major Asian and African deltas which are all associated with urbanising centres within or near the deltas (Foresight, 2011). Thus, deltas with their peculiar migration problems are going to experience high migration, which can be associated with high vulnerability, gender inequality, unstable regimes, and breakdowns of social resilience (Adger, 2000). However, it is also a strategy to spread risk and increase assets, forming an integral part of a household livelihood strategy (Stark & Bloom, 1985). Migration can therefore represent both an adaptation of choice

and the adaptation of last resort when all other avenues have failed (Black et al., 2006; McLeman & Smit, 2006; Hugo, 1996). With the ongoing debate and enhanced scholarship on climate change and migration not much has been done to highlight the peculiar situation of deltas and the migratory patterns of deltaic populations in developing contexts. Studies that have made attempts to estimate migrations in the deltas are also inconclusive. In this paper, we assess migration in delta areas with the use of census data from three deltas in Asia and one in Africa. The aim is to establish the levels of migration in deltas as a background to investigating the relationship between environmental change processes and migration patterns in deltas. Study areas This paper focuses on migration in four deltas, namely the Ganges-Brahmaputra-Meghna (GBM) Delta which traverses Bangladesh and India (presented in the paper as GBM and India Bengal), the Mahanadi Delta in India and the Volta Delta in Ghana. These deltas constitute the study sites for the Deltas, Vulnerability and Climate Change: Migration and Adaptation (DECCMA) Project whose main focus is to investigate the vulnerability of deltas in Africa and Asia and the use of migration as an adaptation option. Each delta has pressing development needs and potentially will be significantly affected by climate change. All four deltas across the two continents are different geo-physically, economically, and they have very different social, governance and cultural characteristics. The Ganges-Brahmaputra-Meghna (GBM) delta made up of the GBM Bangladesh and India Bengal is one of the most densely populated areas in the world, with a population of about

58 million in 2011 and an average population density of about 1075 persons per km2. There is significant poverty, as well as severe development and urbanisation pressure due to the rapid expansion of the major cities of Chittagong, Dhaka and Khulna in the GBM Bangladesh delta and and Kolkata (Calcutta), Sonarpur and Baraipur in the Indian Bengal delta. The Mahanadi delta (MD) in India is formed by the discharge of three major rivers: Mahanadi, Brahmani and Baitarini, covering a coastline of 200 km, which stretches from south near Chilika, the largest coastal lagoon in Asia, to the north up to Dhamra River. The Delta covers an area of 5,900 km2 covering 3% of state’s geographical area. The delta is considered as the ecological and socio-economic hub of the state of Orissa, supporting a large population, of which most are farmers with incomes on or close to the poverty line. The population in the Mahanadi delta area is estimated at 8 million in 2011, with an average of 613 persons per km2, and is growing rapidly. The Delta comes under three sub-divisions- Kendrapara, Cuttack and

• Puri. Bhubaneswar, Cuttack and Puri are the major urban centres in the delta.

The Volta catchment is the ninth largest basin in Sub-Saharan Africa and has a population of about 14 million people who depend directly or indirectly on the resources of the Volta River. The basin is shared among six countries: Ghana, Benin, Burkina Faso, Cote d'Ivoire, Mali and

• Togo. The Volta delta extends for 82 km along the coast with associated wetlands extending

75 km upstream. It had a population of about 900,000 people, who depend largely on fisheries, agriculture, and salt production. There are several small towns, such as Anloga (35,000 people), while the capital of Ghana, Accra (population 2.3 million) is less than 100 km away from the

• delta. The Volta River was dammed in 1965 for hydroelectric power generation. It now has a

regulated flow of approximately 900 m3/s. The river Volta sediment discharge reduced from about 71 million m3/a to about 7 million m3/a after the dam construction. This has caused

chronic erosion of the delta and for example in Keta, 5,000 houses are reported to have been lost to erosion since 1965. More recently significant soft coastal engineering measures have been implemented to reduce the erosion. [Insert Table 1 about here] Data and Methods Estimating Net Migration using the Residual Method Data from the last two population and housing censuses of Ghana (2000 & 2010), Bangladesh and India (2001 & 2011) is used to estimate net migration. It comprises the number of persons classified by age and sex as enumerated in each area at two successive censuses, and a set of survival ratios which is applied to the population at the first census to derive an estimate of the number of persons expected to survive to the second census. The difference between the enumerated population at the second census and the expected population is the estimate of net

• migration. National census survival rates represent the ratio of the population in a given age

group from one census period to the population in the same age group in the prior census (Shryock and Siegel, 1973). The basic assumptions of this method are that (i) there is no abnormal influence on mortality and (ii) the census information is accurate. Thus, estimated net migration is the difference between actual population in year t and the population at year o that survived to year t. The average estimation method is used in this paper. In doing so, it averages estimates from both the forward and reverse methods. The forward estimation method assumes that all migrants survived to the end of the time interval when they joined the population (none of the migrants died during the period between o and t) and provide estimates of the number of persons expected to survive to year t. M1 = Pta + 10–CSR*Poa. The reverse method assumes that

all migrants come at the beginning of the time interval (all migrants are subjected to the pattern

• f mortality among the age group for the period between o and t). The reverse of the forward

method is used to estimate the expected population in year o given the age distribution of a district population in year t which is specified as follows: M2 = (Pta+ t / CSR) – Poa Following from the forward and reverse methods, the average method assumes that all migrants come at the middle of the time interval (or, all migrants are subjected to the pattern of mortality among the population during half of the period between year o and t). This is specified as: M3 = (M1 + M2) /2 Poa = population in age-group “a” in previous census i.e. year o (2000 for Ghana and 2001 for Bangladesh and India) Pta+t = Population in age-group “a+t” at later census i.e. year t (2010 for Ghana and 2011 for Bangladesh and India) CSR = Census Survival Rate The results are presented for the four deltas by district in each country. A positive value indicates more people migrating into a district than leaving it (In-migration > Out-migration), while a negative value means more people leaving than entering a district (In-migration < Out- migration). Data Adjustments Some districts in Volta Delta in 2000 were split by 2010 because of population growth and for effective local government administration. Thus, we merged the data from the 2010 census for

these districts to mimic the 2000 data for affective analysis. These districts include: Ada East and Ada West, which were formed from Dangme East; Central Tongu and North Tongu, which were originally North Tongu; Ketu North and Ketu South which were originally Ketu. Akatsi South and Ningo-Prampram presented a special case. Ningo-Prampram was originally part of Dangme West District together with Shai Osudoku, which is not part of the delta area. Similarly, Akatsi South formed part of Akatsi with Akatsi North, which is not in the delta area. We merged them into their original districts to estimate net migration and their proportions of the total population were used to estimate their proportion of total net migration. South Tongu and Keta remained unchanged between 2000 and 2010. Demographic characteristics of the deltas Compared to the populations of the respective countries in which the deltas are located, the GBM has the largest proportion (27% of the population of Bangladesh), and the Mahanadi had the least proportion (0.7% of the population of India). The population of the India Bengal and the Volta deltas constitute 1.5% and 3.6%, respectively, of the populations of India and Ghana. Deltas are expected to be densely populated areas and three of the deltas have higher population densities compared to their respective countries. The India Bengal and Mahanadi have 1293 and 613 persons per km2, compared to 382 persons per km2 for India, and the Volta has 151 persons per km2 compared to 103 persons per km2 for Ghana. The only exception is the GBM where the population density, i.e. 857 persons per km2 is less than that of Bangladesh, i.e. 1023 persons per km2. While the Volta has the highest proportions of population less than 15 years (38.0%) and aged 65 years and above (7.1%), the Indian Bengal has the highest proportion (66.5%) of population aged 15-64 years. This is reflected in the highest age dependency of 82 recorded in the Volta,

a situation which may have implications for out-migration. Furthermore, while proportions of population less than 15 years are lower in all the deltas compared to the respective national proportions, proportions of the population aged 15-64 years are higher in all the deltas except the Volta. Regarding the proportions of population aged 65 years and over, they are higher for the India Bengal, Mahanadi and Volta, but lower for the GBM compared to the respective national proportions. Finally, age dependency ratio is higher in GBM and Volta but lower in the India Bengal and Mahanadi. The analysis of the sex ratio shows that there are more females in all the deltas with the highest sex ratio of 97.8 found in the GBM, and the lowest of 87.8 recorded in the Volta. The sex ratio for the Volta is particularly low and it could give an indication of a possible high male migration from the Volta delta. While the Volta recorded the highest Crude Birth Rate (CBR) of 28.1 per 1,000 population, and Total Fertility Rate (TFR) of 3.6, the India Bengal recorded the lowest CBR of 11.1 per 1,000 population, and TFR of 1.5. In addition, the CBR and TFR are lower for all the deltas compared to the respective national figures, the only exception being the Volta, where CBR and TFR values are higher than that of Ghana. Furthermore, while Crude Death Rate (CDR), Infant Mortality Rate (IMR) and Under Five Mortality Rate (UFMR) are highest in the Volta, recording values of 12.1 per 1,000 population, 58 and 88 per 1,000 live births, respectively, CDR and UFMR are lowest in the India Bengal, with figures of 2.5 per 1,000 population and 35 per 1,000 live births, respectively, and IMR is lowest in GBM with a figure of 30 per 1,000 live births. When compared to the respective national figures, CDR is the same in the GBM, lower in the India Bengal and Mahanadi but higher in the Volta. With regards to IMR, it is lower in the GBM, India Bengal and Volta deltas, but higher in the Mahanadi. Finally, UFMR is the same in the GBM, lower in the India Bengal and Volta deltas and higher in the Mahanadi.

An analysis of urbanisation in the deltas show that the India Bengal has the highest proportion

• f urban areas (43.0%), and the GBM has the lowest urbanisation proportion (11.3%). In

addition, when compared to their respective national urban proportions, all the deltas have lower proportions, the only exception being the India Bengal which has a higher proportion. The GBM and the Volta deltas, respectively, have the lowest and highest annual population growth rates of 0.7% and 1.6%. Compared to the annual population growth rates of the respective countries, all the deltas have lower growth rates. [Insert Table 2 about here] Net migration in the Deltas Table 3 presents estimated net migration rate for the GBM Bangladesh Delta. The estimated total number of out-migrants is about 2.6 million people including about 1.6 million males and 1.0 million females. This indicates that the GBM Bangladesh delta is a net sender of migrants to other areas, and there is more out-migration of males compared to females. In addition, all the 19 districts in the GBM Bangladesh delta experienced negative net migration. The district

• f Bhola, an isolated island, recorded the highest (11.2% of total population), and Cox's Bazar

recorded the lowest net out-migration (3.6% of total population). [Insert Table 3 about here] The estimation shows that the Indian Bengal delta is a net receiver of migrants representing 1% of the total population (Table 3). Apart from Kolkota, which is a net sender of migrants (5.1% of total population), the two other districts, namely, North 24 Paraganas and South 24

Paraganas are both net receivers of migrants. Furthermore, as shown in Table 4, the Mahanadi delta is also a net receiver of migrants (2.5% of the total population). Apart from Bhadrak which is a net sender of migrants, all the other districts are net receivers of migrants. [Insert Table 4 about here] Overall, the Volta Delta has a negative net migration of about 41,000 people, representing 4.8%

• f the total population. While Keta, Ketu North, Ketu South, Central Tongu and Ningo-

Prampram districts are net senders of migrants, South Tongu, Ada East, Akatsi South and Ada West districts are net receivers of migrants (Table 5). [Insert Table 5 about here] Discussion and conclusion Net migration estimates differ for all the deltas studied. The GBM Bangladesh and Volta deltas are net senders of migrants and the India Bengal and Mahanadi deltas are net receivers of migrants. Rural-urban migration has been most pronounced in Bangladesh, and is contributing to rapid

• urbanization1. Migration to cities in Bangladesh used to be a predominantly male phenomenon

in the past but more recently, there has been a growing feminisation of migration. Female migration in Bangladesh is linked to non-economic reasons including marriage. However, in Bangladesh, due to declining economic opportunities in rural areas, there are some indications

• f female and family migration to cities. The ready-made garments sector is the single most

important contributing sector in attracting female migrants to cities in huge numbers

1 Currently, urbanization rate is around 4% per year.

particularly, the capital Dhaka 2. This population includes females who migrated as a result of marriages3 and net external migrants during the inter-censal period. North 24 Paraganas and South 24 Paraganas in the Indian Bengal are net receivers of migrants. This is contrary to expectation since the two districts have unfavourable bio-physical characteristics for human habitation and it is expected that there would rather be high out- migration from those districts. Although, the Indian Bengal is a net receiver of migrants, Kolkota is the only district that is a net sender of migrants. This may be due to the fact that Kolkata is densely populated and people are migrating from the inner-city to adjoining peri- urban areas including Barackpur, Barasat, and Rajarhat in North 24 Parganas, and Sonarpur, Baruipur, and Garia in South 24 Parganas. This situation may probably explain the slight reduction in the population density of Kolkota from 24,718 people per km2 in 2001 to 24,252 people per km2 in 2011. The migration trend in the Mahanadi delta can be attributed to intra-state migration. This involves the movement of people from adjoining rural communities to urban areas including Puri, Bhubaneswar, and Cuttack for better employment opportunities, and social well-being. Other migrants move to receiving areas including the states of Maharashtra, Kerela, Karnataka, and Andaman to work in the brick kilns, construction and transportation sectors. The Volta delta is a net-migrant sending area, and biophysical factors have been stated to be mainly responsible for this situation. First is the issue of shoreline recession mainly created by the construction of the Akosombo dam in 1964 (Anthony et al. 2016) and two other minor

2 Currently, only the RMG sector employs more than 4 million female workers. 3 In Bangladesh culture, after marriage most usually females permanently move to husbands’ houses.

• dams. Since the construction of the dams no peaks in flow discharge occur anymore, and the

sediment transport has reduced to only a fraction of the original transport (Bollen et al., 2011). It is estimated that about 90% of the total sediment yield of the Volta River is intercepted by the dam (Boateng, 2012). The reduction in sediment supply has created a shoreline recession (Armah, 1991), estimated to range between 4 and 8 m/yr (Boller et al. 2015) which has resulted in the loss of significant proportions of the total land mass of the Keta, Ketu North and Ketu South districts since the 1960s. Thus, land used for crop and vegetable cultivation has been taken over by the sea. In addition, quite a number of landing sites for fishers in most of the communities along the coast have been destroyed and fresh water fishing in the lagoon has also been affected by salinisation. Furthermore, the destruction of large acres of coconut trees by the Cape Saint Paul Wilt Disease from the mid-1980s through the early 1980s has further resulted in the loss of livelihoods for people engaged in coconut oil production. Second, is the issue of land subsidence, sea-level rise and saltwater intrusion. Contributing to these is the abstraction of underground water for irrigation of crop and vegetable farms mainly through “tube irrigation” which is very common in the delta (DARA 2012). Sea level rise is estimated at 3.1 mm/yr and expected to accelerate significantly (Sagoe-Addy and Appeaning Addo, 2013). Although sea defence walls have been constructed in Keta and Ada to reclaim some of the land and prevent further erosion, it is also influencing regional sediment transport in littoral zone (Appeaning Addo, 2015; Bolle et al., 2015; Danquah et al., 2014). However, socio-economic and political factors account for the reasons why some districts in the Volta delta are net-migrant receiving. For example, the Ada East and Ada West districts are net-migrant receiving districts due to their close proximity to Accra, the capital of Ghana and Tema, a major port in the West African sub-region. The two districts are part of the Accra-

Tema city region which is described as a quality residential sprawl with unicentric tendencies. This city-region came about as a result of a nexus of global forces – Ghana’s structural adjustment programme, trade liberalisation, and foreign currency liberalisation – and local forces that predate the Structural Adjustment Programme (SAP) – local economic conditions, innovations in housing, institutional factors and Ghanaian cultural imperatives (Yeboah, 2003). According to Yeboah et al. 2013, the two districts are receiving a number of migrants because people are simply interested in owning a house in the region, but because Accra and Tema are already built up and congested, they end up building or buying in the peri-urban area, where land is readily available. In the case of the South Tongu District, political factors explain its attraction to migrants. This commenced in 1988 when Ghana introduced the decentralisation programme that established district assemblies through the Local Government Law (PNDCL 207). One of the main goals was rural development so as to reduce migration to the large towns and cities, thus redirecting population movement from high concentration areas to low concentration areas (Ayee, 1995). The South Tongu District which was hitherto unknown gained district capital status and became prominent since administrative status brought with it public infrastructure and an influx

• f population (Yaro et al. 2011).

The net-migrant receiving districts in the Volta Delta, namely, Ada East, Ada West and South Tongu are also home to the tourism and hospitality industry. The main tourist attractions are marine turtle breeding sites located in the estuary at Totope, Lolonya, Akplabanya and Kewuse, bird watching on the Songhor Lagoon (designated wetland and Ramsar site which is a sanctuary for about 80% of migratory birds that transit in Ghana), fetish shrines, sacred groves, Fort Konestem built in the seventeenth century, cemetery for early missionaries, and

Asafotufiam festival (Ghana Districts, 2014). Hotels, guesthouses and resorts located on the Volta River, provide water sporting activities including swimming, sailing and boat cruising, which attract migrants who work as tour guides, translators, caterers, drivers, and security guides (Codjoe et al., 2017). Finally, there are a number of livelihood options in the net-migrant receiving districts that attracts migrants. These include crop and vegetable cultivation (Adjei et al. 2016), fishing,

• yster shell harvesting ((Codjoe et al., 2017), salt, sand, and gravel mining (Wiafe et al., 2013;

Anim et al., 2013; Angnuureng et al., 2013; Jonah et al., 2015) and acquaculture (Adjei- Boateng et al., 2012; Amponsah et al., 2015). Vegetable (onions, cabbages, tomatoes, okra and pepper) cultivation mainly undertaken through sprinkler irrigation and therefore done all year round is used to support fishing. To conclude, there is no clear migration pattern in coastal regions. Our findings show that some coastal regions are net in-migration areas whilst others are net out-migration areas. Over all, the four deltas that were analysed in the study does not support the hypothesis that the migration to coast will be a continuum. The impact of climate change coupled with environmental degradation has made it impossible for coastal regions to continue to serve as nodes of attraction. There has been significant migration from deltaic regions to primate cities that are closer to these regions because of the alternative economic opportunities that these primate cities provide. Although, migration trends show a similar pattern for both males and females, there are few anomalies. For example, in the Mahanadi delta, apart from Bhadrak which is a net sender of both male and female migrants, two districts, namely, Kendrapara and Jagatsinghpur are net senders of female migrants but net receivers of male migrants.

Acknowledgement: This review was carried out under the Deltas, vulnerability and Climate Change: Migration and Adaptation (DECCMA) project (IDRC 107642) under the Collaborative Adaptation Research Initiative in Africa and Asia (CARIAA) programme with financial support from the UK Government's Department for international Development (DFID) and the International Development Research Centre (IDRC), Canada. The views expressed in this work are those of the creators and do not necessarily represent those of DFID and IDRC or its Boards of Governors.

References Adger, W.N. 2000. Social and ecological resilience: are they related? Progress in human geography, 24(3), pp. 347-364. Afify, A.A., Mohamed, A.A., and Yacoub, R.K. 2007. Study on the urbanization encroachment impact on Nile Delta alluvium. J. of Appl. Sci., 22(11), pp. 255-271. Agrawala, S., Ota, T., Ahmed, A.U., Smith, J., and Aalst, M.v. 2003. Development and climate change in bangladesh: Focus on coastal flooding and the sundarbans. Organisation for Economic Co-operation and Development. Ayee, J.R.A. 1995. Financing sub-national Governments in Ghana: The District Assemblies’ Common Fund. In Regional and Federal Studies, 5:305-15. Black, R., Bennett, S.R., Thomas, S.M., and Beddington, J.R. 2011. Climate change: Migration as adaptation. Nature, 478(7370), pp. 447-449. Black, R., Arnell, N.W., Adger, W.N., Thomas, D., and Geddes, A. 2012. Migration, immobility and displacement outcomes following extreme events. Environmental Science & Policy. Brown, O. 2008. Migration and Climate Change. Report No 31. International Organization for Migration, Geneva, Switzerland. Chowdhury, U.K., Biswas, B.K., Dhar, R.K., Samanta, G., Mandal, B.K., Roy Chowdhury, T., Chakraborti, D., Kabir, S., and Roy, S. 1999. Groundwater arsenic contamination and suffering of people in Bangladesh. In: W.R. Chappell, Abemathy, C.O., and Calderon, R.L. (eds.) Arsenic exposure and health effects. UK: Elsevier Science B. V. Church, J.A., P.U. Clark, A. Cazenave, J.M. Gregory, S. Jevrejeva, A. Levermann, M.A. Merrifield, G.A. Milne, R.S. Nerem, P.D. Nunn, A.J. Payne, W.T. Pfeffer, D. S. and A. S. U.

• 2013. Sea Level Change. In T. F. Stocker, D. Qin, G. K. Plattner, M. Tignor, S. K. Allen, J.

Boschung, … P. M. Midgley (Eds.), Climate Change 2013: The Physical Science Basis. Contribution of Working Group 1 to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (pp. 1137–1216). Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. Codjoe, S.N.A. Nyamedor, H.F., Sward J., Dovie D. B. 2017. Environmental hazards and migration intentions in a coastal area in Ghana: a case of sea flooding. Popul Environ. DOI: 10.1007/s11111-017-0284-0. de Sherbinin, A., Levy, M., Adamo, S., MacManus, K., Yetman, G., Mara, V., Razafindrazay, L., Goodrich, B., Srebotnjak, T., Aichele, C., and Pistolesi, L. 2012. Migration and risk: net migration in marginal ecosystems and hazardous areas. Environmental Research Letters, 7(4). El-Raey, M. 1997. Vulnerability assessment of the coastal zone of the Nile delta of Egypt, to the impacts of sea level rise. Ocean & Coastal Management, 37(1), pp. 29-40. Ericson, J.P., Vörösmarty, C.J., Dingman, S.L., Ward, L.G., and Meybeck, M. 2006.

Effective sea-level rise and deltas: Causes of change and human dimension implications. . Global and Planetary Change, 50, pp. 63-82. Feng, S., Krueger, A.B., and Oppenheimer, M. 2010. Linkages among climate change, crop yields and Mexico–US cross-border migration. Proceedings of the National Academy of Sciences, 107(32), pp. 14257-14262. Foresight 2011. Migration and Global Environmental Change. The Government Office for Science, London, UK. Foufoula-Georgiou, E., A vision for a coordinated international effort on delta sustainability, in Proceedings of “Knowledge for the Future” Joint Assembly of IAHS - IAPSO – IASPEI July 2013. 2013: Gothenberg, Sweden. Gray, C.L. and Mueller, V. 2012. Natural disasters and population mobility in Bangladesh. Proceedings of the National Academy of Sciences, 109(16), pp. 6000-6005. Hugo, G., 1996. Environmental concerns and international migration. International migration review, pp.105-131. Karim, M.F. and Mimura, N., 2008. Impacts of climate change and sea-level rise on cyclonic storm surge floods in Bangladesh. Global Environmental Change, 18(3), pp.490-500. Khan, A.E., Ireson, A., Kovats, S., Mojumder, S.K., Khusru, A., Rahman, A., Vineis, P. and Labrese Ej, V.P., 2011. Drinking water salinity and maternal health in coastal Bangladesh: implications of climate change. Environmental health perspectives, 119(9), pp.1328-1332. Lonergan, S. 1998. The Role of Environmental Degradation in Population Displacement. Environmental Change and Security Project Report, (4), pp. 5-15. Martin, M., Billah, M., Siddiqui, T., Black, R., and Kniveton, D. 2013. Climate change, migration and human cognition in an Asian mega delta. Working paper 1. RMMRU of University of Dhaka and SCMR of University of Sussex research on Climate Change related migration in Bangladesh, Supported by Climate and Development Knowledge Network (CDKN). McLeman, R. and Smit, B. 2006. Migration as an adaptation to climate change. Climatic Change, 76(1-2), pp. 31-53. <Go to ISI>://000239060600003 Milliman, J.D., Broadus, J.M., and Gable, F. 1989. Environmental and economic implications

• f rising sea level and subsidence deltas: The Nile and Bengal examples. Ambio, 18, pp. 340-

345. MoEF 2008. Bangladesh climate change strategy and action plan 2008. Ministry of Environment and Forests, Government of the People's Republic of Bangladesh, Dhaka, Bangladesh, pp. xvi + 68 pp. Panda, A. 2010. Climate Refugees: Implications for India. Economic and Political Weekly, xlv(20), pp. 76-79. http://ssrn.com/abstract=1935922.

Seto K.C. 2011. Exploring the dynamics xof migration to mega-delta cities in Asia and Afriaca: Contemporary drivers and future scenarios. Global Environmental Change, S94–S107. DOI:10.1016/j.gloenvcha.2011.08.005 Small C. and Nocholls RJ. 2003. A global analysis of human settlements in coastal zones. Journal of Coastal Research, Vol. 19, (3) pp.584-599 Stanley, D.J. and Warne, A.G. 1993. Nile Delta: Recent Geological Evolution and Human

• Impact. Science, 260, pp. 628-634. DOI: 10.1126/science.260.5108.628

Stark, O. and Bloom, D.E. 1985. The new economics of labor migration. The American Economic Review, 75(2), pp. 173-178. Syvitski, J.P.M. and Saito, Y. 2007. Morphodynamics of deltas under the influence of

• humans. Global and Planetary Changes, 57, pp. 261-282.

Tessler, Z. D., Vorosmarty, C. J., Grossberg, M., Gladkova, I., Aizenman, H., Syvitski, J. P. M., &Foufoula-Georgiou, E. 2015. Profiling risk and sustainability in coastal deltas of the

• world. Science, 349(6248), 638–643.

Woodroffe, C.D., Nicholls, R.J., Saito, Y., Chen, Z.Y., and Goodbred, S.L. 2006. Landscape variability and the response of asian megadeltas to environmental change. In: N. Harvey (ed.) Global change and integrated coastal management : The asia-pacific region. Springer Netherlands,pp. 277-314 Yaro, J.A., Codjoe, S.N.A., Agyei-Mensah, S., Darkwah, A., and Kwankye, S.O. 2011. Migration and Population Dynamics. Changing Community Formations in Ghana. Centre for Migration Studies. Migration Studies Technical Paper Series. University of Ghana. Yeboah, I.E.A, Codjoe, S.N.A. and Maingi, J.K. 2013. Emerging urban system demographic trends: Informing Ghana’s National Urban Policy and Lessons for Sub-Saharan Africa. Yeboah, I.E.A. 2003. Demographic and Housing aspects of Structural Adjustment and Emerging Urban Form in Accra, Ghana. Africa Today 50:107-119.

Table 1: A comparison of the GBM, Mahanadi and Volta Deltas Features GBM Delta, Bangladesh and India Bengal Mahanadi Delta, India Volta Delta, Ghana Rivers/catchment area (103 km2) Ganges, Brahmaputra, Meghna (1,730) Mahanadi, Brahmani & Baitarani (141) Black Volta, White Volta and Red Volta (398) Size of delta (103 km2) 87.3 (66% in Bangladesh; 33% in West Bengal, India) 5.91 2.43 Annual (and peak) discharge (m3/s) 35,500 (138,700 - average annual) 1800 (45,000 -- 1 in 50 year event) 900 (dam at Akosombo) Sediment input (tonnes/yr) 1 x109 29.8 x 106 7 x 106 since dam construction Catchment interventions Significant, but much less affected than other three deltas to date Hirakud Dam in 1957 Akosombo Dam (1961-1965) stopped all upstream influence Current relative sea-level rise (mm/yr) 11.0 3.3 3.0 Key current land use issues and hazards Floods, erosion, low dry season flows, water logging, salinisation, surge Floods, erosion, low dry season availability, water logging, salinisation, surge Erosion (especially at Keta), floods, salinisation Typical crops Rice (main crop), wheat, jute, pulses, oilseeds, sugarcane, potatoes, vegetables, spices Rice (main crop), wheat, jute, pulses,

• ilseeds, sugarcane,

potatoes, vegetables, spices Shallot, maize, cassava, tomatoes, okro, yams, rice. Typical livelihoods Agriculture, fisheries, urban workers/labourers, Sundarban dependent livelihoods Agriculture, fisheries, tourism Fisheries, agriculture, salt production, tourism Key cities and towns in the delta Kolkata, Dhaka, Khulna Bhubneswar, Cuttack, Puri, Keta, Aflao, Sogakope From Ericson et al. (2006) and other sources

Table 2: Demographic characteristics by delta

Demographic characteristic GBM, Bangladesh 2011 Bangladesh 2011 India Bengal 2011 Mahanadi 2011 India 2011 Volta 2010 Ghana 2010 Total population (million) 40.4 150.9 18.2 8.0 1210.9 0.9 24.7 Population density (persons per km2) 857 1023 1293 613 382 151 103 Proportion of population less than 15 years (%) 35.3 45.9 27.6 27.6 33.0 38.0 38.3 Proportion of population aged 15-64 years (%) 59.4 46.1 66.5 65.4 61.5 54.9 57.0 Proportion of population aged 65 years and above 5.3 8.0 6.0 7.0 5.5 7.1 4.7 Age dependency ratio 69 56 51 53 63 82 76 Sex ratio 97.8 100.2 95.5 96.5 94.3 87.8 95.2 Crude birth rate (per 1,000 population) 18.3 19.2 11.1 16.6 21.4 28.1 25.3 Total fertility rate 2.0 2.1 1.5 1.7 2.3 3.6 3.3 Crude death rate (per 1,000 population) 5.5 5.5 2.5 5.4 7.0 12.1 6.8 Infant mortality rate (per 1,000 live births) 30 35 31 51 40 58 59 Under five mortality rate (per 1,000 live births) 44 44 35 66 49 88 90 Proportion urban (%) 11.3 23.3 43.0 21.6 31.2 33.0 50.9 Annual population growth rate (%) 0.7 1.5 1.5 1.4 1.8 1.6 2.5

Source: Civil Registration System (CRS), 2011, Census of India; Sample Registration System (SRS), 2013, Census of India; Annual Health Survey (AHS), 2011-12, Census of India (available only for Odisha); Tables on Number of Women, Children Ever Born and Child Surviving, F-Series, Census of India, 2011 (for Indirect Measures). BBS (2015): population & housing Census 2011, national report, volume – 1, Analytical report, Dhaka. BBS (2015): population & housing Census 2011, national report, volume –4, Socio-economic & Demographic Report, Dhaka. BBS (2013): Sample vital registration system 2011.

Table 3: Net migration by District for GBM Bangladesh Delta

GBM-Bangladesh Population 2001 Population 2011 Migration residuals Net migration rates Male Female Total Male Female Total Male Female Total Total Bagerhat 786260 730560 1516820 740138 735952 1476090

• 90583
• 71292
• 161875
• 10.8

Barguna 435220 409840 845060 437413 455368 892781

• 42754
• 27002
• 69756
• 8.0

Barisal 1196220 1152220 2348440 1137210 1187100 2324310

• 136997
• 102206
• 239203
• 10.2

Bhola 884820 818380 1703200 884069 892726 1776795

• 112014
• 83292
• 195306
• 11.2

Chandpur 1112180 1128840 2241020 1145831 1270187 2416018

• 108940
• 81799
• 190739
• 8.2

Chittagong 3440640 3103220 6543860 3838854 3777498 7616352

• 226160
• 79947
• 306107
• 4.3

Cox's Bazar 915520 844040 1759560 1169604 1120386 2289990

• 46337
• 25552
• 71888
• 3.6

Feni 593240 612740 1205980 694128 743243 1437371

• 30650
• 21942
• 52593
• 4.0

Gopalganj 579460 572340 1151800 577868 594547 1172415

• 66167
• 59702
• 125869
• 10.8

Jessore 1282480 1187200 2469680 1386293 1378254 2764547

• 85077
• 39941
• 125018
• 4.8

Jhalokati 344200 348480 692680 329147 353522 682669

• 38806
• 36396
• 75202
• 10.9

Khulna 1234320 1123620 2357940 1175686 1142841 2318527

• 135977
• 98401
• 234377
• 10.0

Lakshmipur 745220 741320 1486540 827780 901408 1729188

• 63184
• 38701
• 101886
• 6.3

Narail 350700 344200 694900 353527 368141 721668

• 36373
• 30234
• 66608
• 9.4

Noakhali 1267060 1303580 2570640 1485169 1622914 3108083

• 86054
• 60886
• 146940
• 5.2

Patuakhali 742200 722600 1464800 753441 782413 1535854

• 78361
• 63344
• 141705
• 9.4

Pirojpur 553620 546160 1099780 548228 565029 1113257

• 58377
• 51744
• 110121
• 10.0

Satkhira 936200 908920 1845120 982777 1003182 1985959

• 57021
• 45161
• 102182
• 5.3

Shariatpur 543360 537320 1080680 559075 596749 1155824

• 71853
• 50473
• 122325
• 10.9

Total 17942920 17135580 35078500 19026238 19491460 38517698

• 1571685
• 1068015
• 2639700
• 7.2

India Bengal Male Female Total Male Female Total Male Female Total Total North 24 Parganas 4633571 4291526 8925026 5119389 4890392 10009781 130765 162000 290423 3.1 South 24 Parganas 3559186 3336739 6895925 4173778 3988183 8161961 80947 76021 156464 2.1 Total 8192757 7628265 15820951 9293167 8878575 18171742 211712 238021 446887 5.2

Table 4: Net migration by District for Mahanadi Delta

District Population 2001 Population 2011 Migration residuals Net migration Male Female Total Male Female Total Male Female Total Total Bhadrak 602681 589997 1192678 676485 665779 1342264

• 2079
• 7551
• 9355
• 0.7

Kendrapara 608380 619488 1227868 675053 681774 1356827 6817

• 5767

1742 0.1 Jagatsinghapur 480442 472738 953180 523326 511429 1034755 4526

• 4082

605 0.1 Cuttack 877063 844689 1721752 995733 940577 1936310 40201 15464 55697 2.2 Khurda 553663 536204 1089867 651947 621839 1273786 75130 74364 149236 7.2 Puri 657199 641455 1298654 733687 710788 1444475 33716 14212 48258 3.0 Total 3779428

370451 7483999

4256231

4132186 8388417

158311 86640 246183 2.5

Table 5: Net migration by District for Volta Delta

District Population 2000 Population 2010 Migration residuals Net migration Male Female Total Male Female Total Male Female Total Total Keta 62827 70834 133661 68,556 79,062 147,618

• 6962
• 6204
• 13166
• 8.9

South Tongu 29407 35404 64811 40,019 47,931 87,950 503 1,425 1928 2.2 Ketu North 42341 48370 90711 46,551 53,362 99,913

• 6,081
• 8,870
• 12,951
• 13.0

Ketu South 68809 77741 146550 75,648 85,108 160,756

• 9879
• 10929
• 20808

12.9 Central Tongu 26496 29062 55558 27,790 31,621 59411

• 2573
• 2925
• 5498
• 9.3

Ada East 24018 27007 51025 34,012 37,659 71,671 1,473 1,371 2844 4.0 Akatsi South 33715 38096 71811 45,497 53,187 98,684 1760 3609 5369 5.4 Ada West 20181 21906 42087 28,579 30,545 59,124 1238 1112 2350 4.0 Ningo-Prampram 26529 29363 55892 33,514 37,409 70,923

• 625
• 530
• 1,155

1.6 Total 334323 377783 712106 400166 455884 856050

• 21146
• 21941
• 41087
• 4.8