---

Groundwater Arsenic in Bangladesh

List of Figures:

Introduction: Fig 40, Fig 41

POPULATION DENSITY AND SURFACE WATER RESOURCES IN BANGLADESH

• Bangladesh has a large population (year 2000 estimate = 128 million) living very close to sea level, mostly on the flood plain of the Brahmaputra and Ganges Rivers. The population density, relative to the total area of the country, is about 900 people per km², almost triple that of India and seven times that of China (Fig 1).
• In the same units, the mean for the global continental area is about 44 and for the USA about 30. Two states in the USA similar in total land area to Bangladesh are Wisconsin and New York, which have population densities of 30 and 150, respectively. The population density of Egypt expressed in these units is only about 60 per km², due to the huge area of desert away from the Nile. Thus population data based on the total area of a country can sometimes be quite misleading.

• Using total area of croplands (cultivated land plus land in "permanent" crops such as orchards), the ratio of population to area of cropland in Bangladesh is about 1300 per km², similar to that for China, and more than double that of India (550 per km²) (Fig 2). Egypt has about 2400 people per km² of total cropland, almost double that for Bangladesh. Comparable ratios of population to total cropland area for the USA and the global land area are 150 and 400, respectively. Clearly the amount of land in Bangladesh available for producing food crops is quite limited, with respect to the large resident population.

• Precipitation in Bangladesh is dominated by the South Asian monsoon pattern of heavy rain during the months of June through September (Fig 3),with very little rain during the winter. The monthly pattern of river discharge is similar, with highest flows and most frequent flooding during July through October.

• If total potentially available renewable water resources are considered, however, Bangladesh appears to have very ample supplies (2400 km³/yr), similar to that estimated for the USA (Fig 4). Current withdrawals of freshwater in Bangladesh (23 km³/yr) are only about 1% of the estimates of renewable water resources. In contrast, Eqypt and Israel currently withdraw fresh water at rates very similar to the estimated total renewable supplies. In the case of Egypt, the renewable supply value quoted in The World's Water (Gleick) is based on the "natural" annual discharge of the Nile, assuming no diversions in upstream countries such as Sudan or Ethiopia, and no losses by evaporation in reservoir storage (Lake Nasser). This value clearly is considerably higher than the amounts actually available for irrigation and other uses in Egypt. Thus there are assumptions and other approximations used to derive water resource estimates that can sometimes be quite misleading. It is important to understand any conventions used to derive such estimates.
• Another example of such issues is the very high percentage of "safe" water reported for Bangladesh (97%). This value is based on the population deriving water from either treated surface waters or groundwaters thought to be free of human pathogens. Comparable estimates for other developing countries, such as Pakistan, Egypt and Nigeria, are much lower, suggesting that pathogen contamination from drinking water in Bangladesh should now be much less of a problem than a few decades earlier. Since the early 1970s, there has been an enormous effort in Bangladesh by UNICEF and other international donor organizations to install tube wells throughout much of Bangladesh, to reduce exposure of the population to infectious disease transmission via contaminanted surface waters. There are now about 10 million shallow wells, accessed by hand pumps, located in many regions of the country. In some areas, which have saline water in the upper layers of groundwater, there are deeper wells that required much more extensive drilling operations to install.

• Per capita withdrawals of freshwater for domestic demands are very low in Bangladesh (6 m³/person/yr), reflecting hand pumped supplies for most of the rural population (Fig 5). Domestic water withdrawals in Egypt average about an order of magnitude greater (60 m³/person/yr), in large part because of the much higher fraction of urban population, including Cairo, which receive treated surface waters. Domestic water withdrawals in the USA average about 250 m³/person/yr, reflecting extensive use for washing of clothes, dishes, showers, flushing of toilets plus watering of lawns and other vegetation. Irrigation withdrawals in Bangladesh (200 m³/person/yr) are appreciably lower than in Egypt (900 m³/person/yr), reflecting reliance on rain as the primary source of water for most crops in Bangladesh. However, use of groundwater for irrigation has expanded rapidly in Bangladesh over the past two decades.

• The population of Bangladesh (Fig 6) increased from about 77 million in 1975 to about 120 million in 1995, representing an annual increase of about 2.3%. The population is currently predominant rural (about 80%), with urban areas growing much more rapidly than the population of the country as a whole.

• The most unique aspect of the natural geography of Bangladesh is that it lies on the flood plain of two of the largest discharge rivers in the world, the Brahmaputra and Ganges, each of which is comparable to the Mississippi River in terms of annual water discharge. The combined annual discharge of the Ganges, Brahmaputra and Meghna Rivers is comparable to that of the Zaire, the second largest river in the world, and suspended particle discharge is the highest in the world (Fig 7). The Brahmaputra and Ganges each carry suspended particle loads in excess of 500 million tons per year, more than twice the current mean annual particle discharge from the Mississippi (200 million tons per year). This enormous quantity of suspended sediments in deposited in the BG delta both above and below sea level and helps maintain soil fertility and maintain the level of the soil surface in opposition to the continued sinking of the delta due to the great weight of sediments accumulating there.

• The drainage basin of the Brahmaputra (population=110 million) includes an appreciable area in China, Bhutan and India, as well as Bangladesh, with a high fraction of the landscape intensively modified for food cultivation (Fig 8, Fig 37). The headwaters region includes large areas of high mountain plateau in the Himalaya.
• Most of the drainage basin of the Ganges (population=450 million) lies in India, but also includes extensive high mountain plateau in Nepal (Fig 9). Except for Nepal, nearly all of the drainage basin of the Ganges has been extensively modified for food production. More than 70% of the entire basin area is crop land, probably the highest for any major river in the world. The portion of crop land in the Mississippi basin is about 35%. Total crop land fraction in the Yangtze basin is about 56%.

• Total dissolved solids (TDS) in the Brahmaputra and Ganges Rivers (Fig 10) are about 100 ppm and 200 ppm, respectively, compared to about 270 ppm in the Mississippi. Thus both of these rivers entering Bangladesh have dissolved major element compositions quite similar to other major rivers, and have calcium and bicarbonate as their highest abundance ions. From dissolved ion data in the major rivers reaching Bangladesh, there would be no reason to expect any particular problems with groundwater chemical compositions.

HEALTH AND ECONOMIC INDICATORS IN BANGLADESH.

• Countries with relatively high economic resources per capita tend to have much lower rates of water-borne and other infectious diseases. Many of the deaths of young children in developing countries are the result of lack of access to adequate quantities of clean water. This general pattern is evident in a scatter plot of child mortality (deaths per 1000 children < 5 years of age) vs GNP per capita (Fig 11). In 1995, annual deaths per 1000 of children < 5 years of age in Bangladesh was about 115, comparable to India, and about an order of magnitude greater than in the USA (10). China had appreciably lower childhood mortality (47), comparable to Egypt (51) in 1995. From the general trend of the scatter plot, much of the gains in improved health for children appear to be feasible at relatively modest increases in GNP per capita.

• Countries with populations > 20 million in 1995 were then grouped in rank order of increasing GNP per capita (Fig 12), with 4 countries discussed having 42% of world population (China, India, Pakistan, Bangladesh).

• Mortality data for these 45 largest population countries illustrate very clearly the much lower rates of young child deaths in countries with higher GNP per capita (Fig 13). Bangladesh, India and Pakistan had comparable young child mortality rates (110 to 140 per 1000/yr), while China had considerably lower young child death rates (about 50).

• Tabulations of access to "Safe" drinking water were greater than 70% for all four of these countries, with Bangladesh being the highest (97%) of the ground of countries with relatively low GNP per capita (Fig 14), reflecting the very high fraction of that population with access to groundwater through shallow tube wells installed since the early 1970s. Note that this indicator may be fairly difficulty to relate in a simple way to something as complex as young child mortality rates.

• Mortality rates for children under five years of age remain quite high (about 100 per 1000 live births), but have been reduced by more than a factor of two since the 1970s (Fig 15). For comparison, most developed countries have a child mortality rate (under five years of age) of less than 10 per 1000 live births. The primary reasons for the substantial improvement of child mortality risks over the past three decades in Bangladesh are not well defined, but probably include at least the following: introduction of shallow groundwater supplies, oral rehydration therapy for rapid treatment of diarrheal diseases, and expanded child innoculations. The relative importance of shifting of water supplies from surface waters contaminated with microbial diseases to groundwaters is currently a contentious issue in Bangladesh.

DISSOLVED ARSENIC [As] IN GROUNDWATERS OF BANGLADESH.

• One of the great tragedies of the latter half of this century is that the well-intentioned efforts of the international community and the Bangladesh government over a number of decades to improve health conditions through installation of shallow tube wells to access groundwater has now resulted in a massive episode of poisoning through dissolved arsenic [As] in drinking water (Fig 39). From evidence available up to now, the source of the dissolved arsenic appears to be almost entirely natural, resulting in mobilization of As from mineral phases in the sediments of the Ganges-Brahmaputra delta. Similar environmental conditions exist in the state of West Bengal of India and As poisoning is now widespread in that area as well (Fig 42).

• International standards on maximum permissible levels (MPL) of dissolved As are currently under extensive review. Standards between various countries now range over a factor of seven (Fig 16), and the proposed new standard for the USA would be a factor of five to ten lower than that currently in force. Thus there is substantial scrutiny as to the level of dissolved arsenic in drinking water which does not represent significant risk to human populations.

• The World Health Organization (WHO) maximum level for As has been 10 parts per billion (ppb) since 1993, while the USEPA and Bangladesh MPL is 50 ppb, a standard that has been in place in the USA for more than half a century (Fig 17). Large number of shallow tube wells in Bangladesh have As concentrations that are in the range of 2 to 20 times the current Bangladesh and EPA standard, clearly far greater than what has been judged to be safe for continuous human consumption.
• Over the next few years, it appears quite likely that the "approved" level for drinking water [As] in the USA will be decreased substantially, probably converging towards the WHO standard or below, based on the level of 10 ppb recently proposed by EPA. Such a change would only make the groundwater As situation in Bangladesh of even greater concern.

• Arsenic occupies a position in the periodic table immediately below phosphorus, and has a valence state of +5 in the presence of dissolved oxygen concentrations typical of most surface waters, as does phosphorus. Arsenic is extremely toxic, while phosphorus is one of the essential elements required by a large range of molecules in living systems. In acute doses, arsenic is lethal on the time scale of a few hours. Chronic low doses cause a range of serious health problems, including skin lesions, fatal skin cancer, gangrene, and a range of fatal organ cancers including those initiated in liver, kidney and lungs (Fig 23). Signs of As related disease were detected in the Bengal basin only in the the mid 1980s (Fig 24).

• Data (through Jan 1999) for dissolved As in 30,000 wells from many areas of Bangladesh indicate that the problem is widespread, but not uniform in geographical distribution. A band of several hundred kilometers width near the middle of the country (Fig 19) has more than 20% of the wells with As levels greater than 50 ppb. Note that the field kits used to obtain this data do not have enough sensitivity to detect dissolved As below about 150 ppb. Thus it is likely that the extent of the problem of dissolved As is appreciably greater than indicated from this distribution of monitoring data.

• Monitoring data for dissolved As using other analytical methods, compiled by the British Geological Survey, also support the observation that elevated As is widespread in Bangladesh, and that deeper wells (> 200 meters below the surface) tend to have considerably lower As concentrations (Fig 38). The degree of local variation in dissolved As is extreme, even on a very local scale (Fig 45). Thus two wells within 10 meters of each other can have very different As concentrations, for reasons that are really not understood at the present time. Monitoring data from a group of scientists in India, working with Bangladesh colleagues, provided some of the earliest data linking As in drinking water to a huge increase in a number of health problems caused by As poisoning. These data also indicate that the problem is probably most severe in a band running east-west across the middle of the country, and also show a large range of As concentrations, ranging from < 10 ppb to greater than 700 ppb in a number of districts.

• The detailed causes of elevated As in Bangladesh and West Bengal groundwaters remain obscure. The most plausible explanation currently under consideration is that arsenic from the sediments has become mobilized by strongly reducing conditions of the groundwaters, which are largely anoxic (ie no dissolved oxygen). Such conditions would lead oxidized iron [Fe(III)], which is very insoluble, to be reduced to a much more soluble form of iron [Fe(II)]. Arsenic in a higher oxidation state [As(V)] has a strong tendency to be sorbed onto solid iron mineral phases. Once iron mineral phases go into solution in reducing groundwaters, As would also tend to go into solution. Thus the iron phases serve to control the amount of As in solution. The detailed processes involved in Bangladesh groundwaters are clearly more complicated than the simple outline provided here, but they do appear to involve primarily "natural" processes, and are not the result of human pollution. Arsenic can be present in a number of chemical forms in natural waters, with considerable variation in species type and oxidation state depending upon the pH and Eh (oxidation potential) conditions in the water. The lower oxidation state As (III) is generally more soluble than higher oxidation state As(V), but people drinking As in any inorganic chemical form are exposed to significant health risk when the amounts of soluble As appreciably exceed 10 ppb. The details of the processes causing mobilization of As are not critical to define here, but are mentioned to provide some indication that the factors controlling water quality can sometimes be quite complicated and poorly understood (Fig 43).

• The land surface of Bangladesh & adjacent state of India (West Bengal) is composed of alluvium deposited over the last several million years. A major fraction has accumulated since the end of the last glacial period as sea level rose from about 130 meters below current sea level. Note there are some areas of Bangladesh where the surface sediments were deposited at least several hundred thousand years ago. These deposits do not appear to have elevated arsenic concentrations in groundwaters. The original source of As was probably from pyrite minerals derived from the uplands in the Himalaya mountains, but major changes in the chemical environment during erosion, deposition and burial have considerably altered the mobility of arsenic as a function of geological environment.

• The alluvial deposits in Bangladesh were derived from three major river systems (Ganges, Brahmaputra and Meghna Rivers), all three of which appear have delivered sediments which can readily lead to high dissolved arsenic in groundwaters. The pattern of accumulation of sediments in the Bengal Basin has varied greatly over the past 20,000 years, as the channels of the major rivers have changed during that period. In general, very deep groundwaters tend to have relatively low arsenic concentrations, but the most common depths of the shallow wells have a very large range of concentrations.

The Columbia University's Arsenic program (An overview)

• There are a number of areas, including the following countries, which have experienced high dissolved As in drinking water which has impacted the health of human populations: Taiwan, West Bengal, Bangladesh, Mongolia, Argentina, Chile and Mexico. Within one region of Bangladesh, there is now a multi-year research program  involving a number of researchers at Columbia University and other institutions, focussed on documenting the effects of arsenic poisoning on the population of that region, as well as the geology and hydrology factors which tend to lead to high dissolved As and could potentially permit reduction in human exposure (Fig 20). The project is located in an area about 20 km east of the capital city of Dhaka  in Arahaizar Thana (Fig 26). This effort involves faculty and staff from the School of Public Health, Lamont-Doherty Earth Observatory, Engineering School, School of International and Public Affairs, plus other units of the university (Fig 21) and partners in Bangaldesh (Fig 22).

• The following pictures provide some visualization of the study area and the construction of the wells (Fig 27, Fig 28). Villages are typically built on slightly elevated artificial "islands" that protect  residents from the big floods in the summer. Wells are being installed by using a wet drilling technique (Fig 29). Almost all existing wells (~5000) in the area of investigation were sampled in 2001/02 (Fig 30, Fig 36). The village wells in this area, as in much of the country are pumped by hand. Concentrations of As (Projection #24.25) are quite variable as a function of location, which makes it especially difficult to predict the likely level of contamination prior to drilling a new well. As an example for the horizontal and also vertical distribution of As, one village in Araihazar is shown (Fig 31). As concentrations show a very large spatial variability, but all of the wells beyond a depth of ~60ft seem to be low in As. There also appears to be a certain depth interval (50-100 ft) where almost no wells exist. A sediment core was eventually collected at the site that showed a massive low permeability clay layer separating the shallow from the deep aquifer (Fig 32). It looks like that at this site the aquifers are well separated and that if not too large quantities of water are pumped, the deper aquifer might provide an alternative source for low-As water. However, not everywhere [As] drops off that dramatically with depth. Other options (Fig 33) might have to be studied there.

• In areas with high spatial varible As concentrations, well switching might be a remediation option. Most residents would not have to walk very far to obtain access to a "safe" well (Fig 34), but major cultural boundaries complicate this scenario.

• A small number of single-well treatment systems to remove arsenic have also been deployed (Fig 35). This particular system uses several plastic buckets, sand filters, and the addition of iron in tablet form to remove dissolved [As]. However it appears difficult to create enough incentives for people to consistently use and maintain these systems.

• Certainly, the As problem poses major challenges to the scientific community to better constrain the processes mobilizing As and the dose/response relationships of As, to engineers to develop feasible solutions, and the social scientists  to help implement them (Fig 44).

• 1. From the excel worksheet, a data set has been organised. The data is from Site A (Satyabandi village) from Araihazar, Bangladesh (Columbia's research site). It shows the concentration of As in ug/L (ppb) in groundwaters collected from different depths. Can you plot a depth profile of this variation in Arsenic concentration for this site. To plot, this, please use a scatter plot with y-axis representing the depth (depth increasing down the axis) and the top horizontal x-axis as the concentration of As. What inference can you draw from the distribution pattern of Arsenic in this site?

• 2. From the excel worksheet, another data set has been organised. This is also from Site A (Satyabandi village)from Araihazar, Bangladesh. Here, several samples of groundwaters have been collected from different locations in site A. Each of these locations have their GPS positions (as Latitude and Longitude). After these groundwaters were collected, they were analysed for their As concentration, in an instrument called High Resolution ICP-MS. In the excel worksheet you can see, each of these samples have their respective GPS positions and As concentration (in ug/L or ppb) charted. Can you plot a bubble plot, to show the variation of As concentration with Latitude and Longitude. Also, try to plot in a way, to show the smaller bubbles (lesser concentrations of As) are on top of the bigger bubbles (higher concentrations of As).