เป็นเรื่องที่น่าสนใจในการจัดการศึกษาเรื่องนี้ ข้าพเจ้าได้นำเสนอในหัวข้อสัมนา ในการสัมนาวิทยาศาสตร์เชิงคำนวณ ประจำภาคการศึกษา 
จึงนำมาเสนอไว้เผื่อท่านใดสนใจ พรีเซนเตชั่น ที่ทำไว้ หรืออื่นๆ สามารถขอได้ครับ ขอบพระคุณเจ้าของลิขสิทธิ์ทุกท่าน
Impact of Climate Variability on Infectious Disease in West Africa   
Madeleine C. Thomson,1 Stephen J. Connor,1 Neil Ward,1 and David Molyneux2
1International Research Institute for Climate Prediction (IRI), The Earth Institute at
Columbia University, Lamont Campus,
Monell Building, 61 Route 9W, Palisades, NY 10964-8000
2Liverpool School of Tropical Medicine, Pembroke Place, Liverpool L3 5QA, UK
Abstract: The importance of infectious disease as a determinant (as well as an outcome)
 of poverty has recently become a prominent argument for international and national 
investment in the control of infectious disease, as can be seen in the recently articulated
 United Nations (UN) Millennium Development Goals (MDGs). Climate variability and 
land use change have an enormous impact on health in West Africa, and may
yet undermine the potential for achieving the MDGs, in certain economic-ecological
zones. However, theirunderlying role in determining the burden of disease in the 
region on a yearly or decadal basis has never been
systematically studied. In order to improve our understanding of the future
impacts of climate change, it may be more effective to start by investigating the impact
 of inter-annual climate variability, and short-term shifts inclimate (e.g., decadal), 
on disease transmission dynamics. This information may inform both current and
future policy decisions with regard to prediction, prevention, and management of
adverse climate-related health outcomes. This article reviews current knowledge of 
changes in the epidemiology of infectious diseases associated with climate variability 
in West Africa over the last 40 years. Selected examples are considered from
bacterial (meningococcal meningitis), protozoan (malaria), and filarial (onchocerciasis
 and lymphatic filariasis) infections where spatial and temporal disease patterns have 
been directly influenced by seasonal, interannual, or decadal changes in climate.
INTRODUCTION
While infectious and parasitic diseases are no longer the leading causes
of mortality in the developed world, theyremain the main cause of mortality in 
middle and lower income countries (Connor,2002a), accounting for morethan 
half of all deaths in Africa (Murray et al., 2001). The importance of infectious 
disease as a determinant (as well as an outcome) of poverty has recently become 
a prominent argument for international and national investment in the control of 
infectious disease (WHO, 2001a) as can be seen in the recently articulated United 
Nations (UN) Millennium Development Goals (MDGs) (Sachs, 2002). Climate variability 
and land use change have an enormous impact on health in West Africa, and may yet
 undermine the potential for achieving the MDGs, in cer economic-ecological zones. 
However their underlying role in determining the burden of disease in the region on 
a yearly or decadal basis has never been systematically studied. Clearly appropriate 
evidence-based economic, environmental, and public health policy instruments are 
needed to reduce the potentially adverse impacts. Developing suchevidence in a 
systematic way is a prerequisite for its effective use in guiding policy development. 
In the countries belonging to the World Health Organization’s sub-region D 
(all West African countries), respiratory infection, malaria, and human 
immunodeficiency virus/acquired immunodeficiencysyndrome (HIV/AIDS) are 
leading causes of mortality and Disability Adjusted Life Years, and their effective 
control is perceived as an essential requirement for sustainable livelihood development
 in the region. This article reviews current knowledge of changes in the epidemiology 
of infectious diseases associated with climate variability in West Africa over the last 
40 years. Selected examples are considered from bacterial, protozoan, and filarial 
infections where spatial and temporal disease patterns have been directly influenced
 by seasonal, inter-annual, or decadal changes in climate. Climate of West Africa
The climate of West Africa ranges from the Saharan desert (bounded by 100 mm
mean annual rainfall isohyet) through the Sahelian zone whose southern border is 
demarcated by the 400 mm isohyet. South of this line is the savanna zone, where 
the Malian horse empire was founded in the 13th century. Further south, the 1000 m 
isohyet marks the start of the woodland savanna zone and the northern limits of the 
tsetse fly (Glossina morsitans) belt—a formidable barrier to all those dependant on 
cattle and horses. Cattle, horses, donkeys, and camels are all susceptible to the often 
fatal parasitic disease trypanosomiasis (sleeping sickness) transmitted by this fly.
The timing of the rainy season in West Africa is highly predictable and is
governed by the movement of the Inter- Tropical Convergence Zone (ITCZ), a 
discontinuous belt of thunderstorms marking the convergence of the northern and 
southern hemisphere’s surface trade winds which moves northwards and southwards 
with seasonal changes in the overhead sun. The ITCZ reaches its maximum northward
 extent around July–August and its southern extent around January–March (during this dry 
period, the harmattan winds blow Saharan/Sahelian dust southwestwards across the
 region). 
The movement of the ITCZ’s northern edge is not regular and short-term oscillations 
of as much as 500–1000 km may occur over a few days. This can cause erratic starts 
to the onset of the rainy season. Rainfall in West Africa is the most significant 
determinant of the landscape in terms of natural vegetation and land use activity. 
The amount of annual rainfall at any geographic location depends heavily on the 
duration of the rainy season, that is, the mean time in total during which the ITCZ is 
positioned well north of the particular location thus allowing rainfall to occur. Drought
 in West Africa West Africa provides one of the most dramatic examples
worldwide of climate variability that has been directly and quantitatively measured
(Hulme, 2001). Averaged over 30- year intervals, annual rainfall across the Sahel
/Guinea-savannah fell by between 20 and 30% between the pre-independence period
 (1930s to 1950s) and the decades since (post-1960s) (Fig. 1). Thus, the famines that 
struck the region in the 1970s and 1980s were embedded in a drought that has lasted
 from the 1960s to the present day (with some amelioration in recent years), which 
not only affected the Sahel but extended southward to the Guinea coast. It has been 
observed that most of the rainfall deficit of
the dry period 1971–1990 in the Sahel is correlated with a general decrease in the
number of rain events rather than a shortening of the length of the rainy season (Le 
Barbe and Lebel, 1997; Le Barbe et al., 2002). This, along with the fact that the 
rainfall deficit was equal to 180–200 mm all over West Africa, indicates that there 
was no crucial difference in nature between the Sahelian drought and the drought 
of the Guinean savanna regions. The explanation of such features resides in either 
an abnormal southward position of the ITCZ and/or a less productive ITCZ during 
the dry spell. It has been proposed that sulphurous emissions from factories in 
industrialized countries may have created the conditions for, or exacerbated the 
extent of, this recent drought by weakening the northwards movement of the ITCZ 
(Rotstayn and Lohmann, 2002). As a result, the intense thunderstorms which produce 
the rains needed for the Sahel croplands fail to materialize. However, the most favored 
current hypothesis for the extended drought continues to be large-scale changes in 
near-global patterns of sea surface temperature (SST) anomalies (Folland et al., 1986)
 with a possible role of landsurface changes in reinforcing the drought. Anomalies in 
SST in the world’s oceans, including those related to the El Nin˜o Southern Oscillation 
(ENSO), contribute to rainfall variability in the Sahel (Lamb, 1978; Folland, et al., 1986;
Ward, 1998). Lamb et al. (1978) observe that droughts in West Africa correlate with 
warm SST in the tropical South Atlantic. Examining oceanographic and meteorological 
data from the period 1901–1985, Folland et al., (1986) found that persistent wet and 
dry periods in the Sahel wererelated to contrasting patterns of SST anomalies on a 
nearglobalscale. When SST in the northern hemisphere’soceans were cold, rainfall in 
the Sahel was low. A recentattempt to quantify the importance of the oceanic 
influence
on the variability of the northern African summer monsoonclimate
(including the Sahel) on interannual to interdecadaltimescales has been made 
(Giannini et al., 2003).

Using an ensemble of simulations of the global atmospheric circulation forced
 with the observed record of SST, but no year-to-year changes in vegetation cover or 
greenhouse gas concentrations, these authors demonstrate that a warming of the 
tropical Pacific and Indian Oceans, of the order of 0.2_C or larger, is associated with 
a negative rainfall anomaly in the Sahel PRINCIPAL component (0.73 mm/ day in 
observations). Furthermore, a negative rainfall anomaly is consistent with local land 
surface temperature anomalies up to 0.6_C warmer than average. From this work, 
the authors conclude that SST variability is instrumental in determining the sign of 
rainfall anomalies in the Sahel, while coherent land-atmosphere interaction acts to 
amplify them. The theory that SST anomalies are the driving force behind the 
drought contrasts dramatically with earlier perceptions of the causes of the Sahelian 
drought. The responsibility for this regional disaster had been put upon the shoulders 
of local peasant farmers, whose overuse of the land was deemed to have resulted in 
reduced vegetation cover, the advance of the desert, and localized impacts on the 
regional climate regime (McCann, 1999). Major river systems endow some, but not 
all, of the countries of West Africa with substantial water resources. One of the first 
consequences of the decrease in precipitation has been a significant reduction in
 discharge within these river systems. The fall in precipitation observed before and 
after the discontinuity ranges between 15 and 30%; for river discharge, the 
corresponding fall varies between 30 and 60% (Gustard and Cole, 2002). This
 translates into a decline in surface water resources across the sub-region in 
recent years, which has serious consequences for the management of these resources.
 A further consequence is the increasingly rapid degradation of vegetation and 
associated desertification, mainly in the Sahelian zones, which
renders the Sahelian ecosystem even more fragile in the face of predominant
socio-economic activities including agriculture and animal husbandry. Whatever 
their cause, droughts in the Sahel are not new. In fact, what characterizes rainfall 
in the Sahel is its variability in space and time from year to year, decade to
decade (Nicholson et al., 1990). Human society has adapted to this variability
in rainfall supply, for example, through pastoralism, diversification of income, 
and mobility (Batterbury and Warren, 2001). However such adaptability has not 
freed the region from vulnerability to periodic famine, the most tragic of which are 
recorded as having occurred in 1680, 1750, 1820, 1920, 1968, 1973, and 1984 
(Egg and Gabas, 1997). In response to the most recent famines, the international 
community has focused efforts on the development of famine early warning systems 
to reduce the region’s vulnerability to such disasters. These warning systems use 
information on current household vulnerability to shocks and climate/environmental 
data in order to assess a region’s vulnerability to famine. The potential value of 
seasonal climate forecasts to agriculture (Hansen, 2002), disaster management 
(Dilley and Heyman, 1995; Thomson et al., 2003), and health (WHO, 2001b) has 
been highlighted and new initiatives are currently underway to assess the value of 
such information to regional and local decision makers. However, because of 
limitations in the predictability of SST, in particular equatorial and South Atlantic 
SST (key determinants of the West African climate), during the spring period, the 
potential to predict the seasonal rainfall over West Africa prior to May or June is poor 
(Goddard and Mason, 2002). Forecasts are therefore limited in this region in the lead 
time they can offer decision makers. While most research to date has focused on the
 predictability of the rainy season, new efforts to predict the length and character 
of the dry season are being prompted by health sector interest in view of the potential
 relationship of low humidity and dust torespiratory infections (Molesworth et al., 
2002a).

Geographical and Socio-economic Characteristics of West Africa There are 15
countries in mainland West Africa (Fig. 2), 3 of which are landlocked. The sub-region 
stretches from the southern fringes of the Sahara Desert to the Atlantic Ocean along
 the Gulf of Guinea, from hot dry desert, through
semi-arid savannah, to moist humid forests. Generally the West African countries can
be currently characterized by: high population growth, young populations, high and 
weakly controlled urbanization, and weak human development. Periodic war and civil 
unrest, often driven by the desire to control minerals (Veeken, 1994; Davies, 2000) 
plague countries which are well endowed with natural resources resulting in large 
refugee/displaced populations in some countries. While substantial gains have been 
made in reducing infant mortality in the last 40 years, deaths rates remain high, life 
expectancy is approximately 50 years, and national Gross Domestic Product (GDP)
 figures indicate that these countries include some of the poorest populations in the 
world. Two important factors currently driving both economic and political change are 
the 10-fold increase of the population over the last century, and the recent opening 
up of the region to the world economy, as globalization advances (Cour, 2001). 
Political and economic activity in West Africa has always been heavily influenced by 
climate, and historians have delimitated key West African historical zones by rainfall 
isohyet boundaries which shift periodically. These boundaries impose specific 
constraints on human activity (McCann, 1999). For instance, while in all West 
African countries there is a predominance of the agricultural sector in the national 
economies, the type of agriculture practiced and its potential for economic growth 
and development is largely governed by the climatically determined ecological 
zones of the region. Recent work for the World Bank by the Food and
Agriculture Organization (FAO, 2001) identified a number of ecological zones,
determined largely by rainfall resources, which characterize particular farming 
systems (Fig. 3). These include: Pastoral Systems with a high incidence of severe 
poverty and low potential for poverty reduction; Agro-Pastoral Millet-Sorghum 
Systems with a high incidence of severe poverty but with a high potential for poverty
 reduction; and Cereal-Root Crop Mixed Systems with a lower incidence of poverty 
and a high potential for agricultural growth. The climate and environment in each 
of the farming system zones described correspond closely to different zones of 
epidemic disease risk (Fig. 4). Therefore, the development policies and practices
 that are followed have important implications not only for agricultural development
 and food security, but also disease transmission risk and vulnerability to severe-
disease outcome. Clearly this has profound implications for sustainable livelihoods 
development in the region over the medium–long term.
Climate and Health Interactions in the Arid and Semi-arid Farming Zones of West
Africa Climate impacts on health through a number of distinct, direct and indirect, 
mechanisms. For example, directly through heat stress, UV induced melanomas, 
etc., or perhaps more significantly, indirectly, through:
a) its role in determining agricultural output and consequently food
security, which directly effects nutritional status;
b) its role in the economy via agricultural exports, hydrological power,
etc., which affects the ability of individuals and communities to maintain nutritional 
status, prevent infection (e.g., through the purchase of mosquito nets) and, when
 necessary, obtain health care. The impact of climate on the economy will also 
affect the ability of governments to provide health services;
c) its role in determining seasonal and inter-annual demographic
 processes (e.g., seasonal labor migration and environmental refugees), 
increasing the likelihood of individuals contracting certain infectious diseases, 
such as HIV;
d) its impact on the spatial and temporal distribution of climate-related
infectious diseases (e.g., malaria, Rift Valley fever, meningococcal meningitis). 
Many of these indirect mechanisms interact, and when climatic anomalies coincide 
with societal vulnerability, disastrous drought and famine can occur (e.g., in West 
Africa in the early 1970s and early 1980s). These situations result in further social 
disruption and population displacement, with the major cause of mortality during
 such disasters being epidemics of infectious diseases (Prothero, 1994). Furthermore
, some of the linkages are bidirectional. For example, while agricultural production
 and food security impact on health, infectious disease episodes impact on 
agricultural production, affecting the availability of labor at critical times in the 
annual agricultural calendar.
Predicting Infectious Disease Transmission
There is currently extensive interest in predicting the possible impacts
of climate change on the major infectious diseases (Colwell et al., 1998; Kovats 
et al., 2001; Patz and Reisen, 2001). However, considerable uncertainty is associated 
with such predictions. While there is widespread acceptance that the climate is 
changing (Houghton et al., 2001), the future impact of such change on the actual 
weather patterns in particular geographic regions remains largely speculative 
(McCarthy et al., 2001). Furthermore, while many infectious disease vectors and 
agents are known to vary in their spatial and temporal distribution, as a function of
 climate processes (Thomson and Connor, 2000), the detailed 
parameterization of key variables, such as vector survivorship, are made difficult by 
the paucity of relevant empirical data which relate to disease transmission under 
field conditions in Africa. This, combined with the many nonlinearities in the
transmission dynamics of infectious diseases and the importance of nonclimatic
 factors in determining the outcome of transmission, make predicting the impact 
of climate change on specific health outcomes a contentious one (Martens et al., 
1997; Rogers and Randolph, 2000). In order to improve our understanding of the 
future impact of climate change, it may be more effective to start by investigating
 the impact of inter-annual climate variability, and short-term shifts in climate 
(e.g., decadal), on disease transmission dynamics, since this information may
 inform both current and future policy decisions with regard to prediction, 
prevention, and management of adverse climate- related health outcomes. 
Consequently, research aimed at the use of seasonal climate forecasts by health
 services in Africa needs to emphasize the importance of understanding the 
probabilistic nature of future climate/ disease predictions (Thomson et al., 2000b)
 [Palmer et al., manuscript in preparation]. Many of the infectious diseases that 
contribute to the very high child mortality rates in West Africa occur within a 
‘‘climate envelope’’ (for example: malaria, Rift Valley fever, yellow fever; relapsing 
fever, lymphatic filariasis, onchocerciasis) and are further restricted in their 
distribution by the microclimate/environment which supports their transmission
 (Thomson and Connor, 2000). However, transmission of such diseases in the 
region may vary in intensity between years as a result of inter-annual variation 
in climate and human vulnerability (Table 1). Many of these diseases are vector-borne
, but climate may also be a major determinant of the variability of nonvector-borne 
diseases such as diarrheal disease, airborne infections such as measles (Brewster 
and Greenwood, 1993), and meningococcal meningitis (Cheesbrough et al., 
1995;Molesworth et al., 2002b). Furthermore, climate ‘‘shocks’’ may change an 
individual’s exposure to infectious disease through migration (Prothero, 2001) 
and changes in behavior (Findley, 1994). A number of examples are presented
 below.
Malaria
The most obvious result of rainfall decline on infectious disease
transmission in West Africa is the associated change in malaria prevalence and 
incidence across the region. A greater than 80% decline in malaria prevalence
 rates have been observed in the semi-arid areas of northern Senegal and Niger 
from the early 1960s to the mid 1990s. This reduction is presumably associated
 with a loss of vector breeding sites, a shortening or reduction in intensity of the 
malaria transmission season as a result of lower vector survivorship and a loss 
of specific vector species. Previous studies have indicated that the different
 cytotypes  of Anopheles gambiae s.s. are highly climate sensitive (Toure et al., 
1994;  Thomson et al., 1997). Observations of malaria prevalence rates before and 
after the droughts which occurred since the 1970s, indicate that in the Sahel 
(Niayes region, Senegal), endemic malaria decreased drastically after the
 disappearance of the principal mosquito vector, Anopheles funestus, due to 
the destruction of its larval sites by cultivation (Mouchet et al., 1996). No 
reestablishment of this species was observed in the Niayes region after the very
 wet year of 1995, however, the presence of this vector has been noted more 
recently in the irrigated areas of northern Senegal (Konate et al., 2001). It has
 recently been proposed that climate and geography are more significant than 
poverty per se in determining levels of malaria burden, and that a reduction in 
malaria is likely to proceed a reduction in poverty (Sachs and Malaney, 2002). 
The major reductions in malaria transmission in the Sahel that have occurred 
have not, however, always resulted in a reduction in poverty, as many countries
 in the region currently rank at the bottom of the development indices. As well
 as climatic variability, other factors such as poor governance, conflict, international 
debt, poor terms of trade, and restricted access to the international markets have
 undermined the region’s development agenda. Changes in malaria endemicity 
as a result of the Sahelian drought have affected large areas of semi-arid West
 Africa as regions have moved from endemic to epidemic. This change presents
 a major obstacle to the development of regional malaria risk maps (Kleinschmidt et al., 2001) as prevalence rates may change dramatically over a few years. The increasing epidemic nature of malaria transmission has, however, been acknowledged in terms of developing control strategies for the region, with new efforts to develop malaria early warning systems based on routinely available meteorological station- or satellite-based information on excess rainfall (WHO, 2001b). Evidence supporting the use of rainfall-driven malaria early warning systems in West Africa is slowly building (Ndiaye et al., 2001). However, rainfall alone may not be a sufficient indicator of malaria transmission in those parts of the Sahel where its impact on the local ecology is governed by hydrological process associated with topography and soil type. In such a situation, an early warning system based on the water levels in rivers or dams may be more valuable—as has been proposed for the Senegal River basin (Faye et al., 1998). Support for the use of hydrological monitoring comes also from Sudan where river flooding is a key trigger for local increases in malaria cases (El Sayed et al., 2000). The flooding of rivers will depend on the rainfall distribution and timing in the entire river catchment rather than on local rainfall. It has been noted that the extent to which flooding is associated with increases in malaria cases is dependent on the timing of the floods in relation to other factors such as local rainfall and humidity (Najera, 1999). Changes in the presence of surface water as a result of the development of irrigation schemes along the major river
systems in West Africa has important implications for malaria transmission. Expectations that the consequent increases in available breeding sites for malaria vectors will result in an inevitable increase in malaria transmission in the vicinity of such schemes has not, however, always been borne out (Ijumba and Lindsay, 2001). The explanation for this finding is still unresolved but, in some cases at least,
can be attributed to displacement of the most endophilic and anthropophilic malaria vector Anopheles funestus by An. arabiensis with lower vectorial capacity.

Lymphatic Filariasis 
The ecological determinants of lymphatic filariasis infection are not well understood, although a generalized ecological distribution was proposed by Brengues in the 1970s (Brengues, 1975); see Figure 5. The parameters used to create Brengues’ filariasis map were elevation, vegetation cover, and annual rainfall. The importance of understanding the geographical determinants of this disease have reemerged with the development of efforts to control lymphatic filariasis in Africa as part of a global program for its elimination as a public health problem (Molyneux and Zagaria, 2002). The results of a recent study of the spatial distribution of prevalence rates of Wucheria Bancrofti-(the causative agent of lymphatic filariasis) has shown that this disease is largely absent from some of the Guinea coastal countries such as Togo and Benin—despite these countries being highly endemic for malaria (Gyapong et al., 2002). The apparent contrast in the frequency of human infection between W. Bancrofti and P. falciparum is all the more striking since both parasites are transmitted by the same mosquito vector (namely members of the Anopheles gambiae species complex). Gyapong and colleagues speculate that differences in the spatial distribution of these two diseases may be related to differences in the cytotypes of the vectors since high prevalence levels of filariasis seem to be coincident with the spatial distribution of Mopti and savannah forms of Anopheles gambiae s.str. Given that malaria infection rates have changed dramatically over the Sahel as a result of reduced rainfall, it would seem unlikely that filariasis transmission by Anopheles gambiae s.l. has not also been affected. There is however no evidence to date which can be used to substantiate such a claim.

Onchocerciasis
Prior to the 1970s, onchocerciasis (known as river blindness) was a neglected disease. Its devastating effects where largely borne by those rural populations of West Africa living near the fast flowing rivers of the Sahel. When the Onchocerciasis Control Programme (OCP) was started in 1974, some of West Africa’s richest river lands were uninhabited. In villages sited in river valleys near to the breeding sites of the blackfly vector of onchocerciasis (Simulium damnosum s.l.), it was not unusual to find 60% of the adults afflicted with the disease and 3–5% blind. Communities were forced to abandon their villages en masse. Today, more than 20 years and US $600 million after the program was first launched, the disease has been controlled through one of the most successful public health campaigns in history (Benton et al., 2002). The governments of 11 African nations and 24 donor agencies combined their resources and energies to spray the rivers where the blackflies breed, and to develop and distribute the anti-helminthic drug ivermectin for treatment or prevention of the disease. Villages once emptied by river blindness have been rapidly resettled—bringing new challenges to the delivery of health care in the region (Richards et al., 2001). Understanding the spatial and temporal distribution of the vectors of onchocerciasis has been key to their successful control (Boakye et al., 1998), given that the S. damnosum species complex comprises many distinct sibling species with varying capacities to transmit the pathogen, Onchocerca volvulus. The distribution of the different members of the Simulium damnosum species complex is generally related to phtyogeographic zones, forest and savannah, but seasonal changes in their distribution occurs on an annual cycle as the monsoon winds and their accompanying rainfall aid dispersal and result in enhanced river flow and the creation of breeding sites. According to Baker et al. (1990), members of the S. damnosum complex move average distances of 15–20 km daily and may migrate over a total distance of 400–500 km. Unusual migrations of savannah species of S. damnosum s.l. (the species most commonly associated with the blinding form of the disease) into the forest zones have been observed (Thomson et al., 1996), and this has led to speculation on the possible role of deforestation and rainfall decline on the distribution of different species of S. damnosum s.l. (Walsh et al., 1993). Rainfall has also had a more direct impact on the control program through its role in determining river flow and, therefore, the amount and type of insecticide required to treat specific areas (Hougard et al., 1993). With the extension of the control of onchocerciasis into countries hitherto not included (through the African Programme for Onchocerciasis Control, APOC), a new obstacle has been placed in the way of effective disease control by the presence of another filarial worm, loa loa, in forested West and Central Africa. This is a consequence of recent reports of severe and fatal encephalopathic reactions to ivermectin in individuals with high loa loa microfilarial counts. Consequently, mapping the spatial distribution of loa loa has become a priority for APOC in order to modify its treatment protocols appropriately (Thomson et al., 2000a).
Meningococcal Meningitis
Neisseria meningitidis (the meningococcus) is responsible for endemic and epidemic meningococcal disease throughout the world. In Africa, meningococcal meningitis often occurs as extensive epidemics—with many thousands of deaths, particularly in the so-called meningitis belt of sub Saharan Africa. This belt is defined as an area between latitudes 4_ and 16_ north where high incidence and recurring epidemics of cerebrospinal meningococcal meningitis coincide with the 300–1100 mm mean annual rainfall isohyets south of the Sahara (Lapeyssonnie, 1963); see Figure 4. Thus, the belt comprises much of semi-arid sub- Saharan Africa and, in particular, the Sahel epidemic waves occur every 5–10 years in the Sahel region of the sub- Sahara and are mostly due to group A meningococci. Risk factors for invasive disease and epidemic outbreaks are not completely understood, but it is likely that a combination of conditions is necessary for an epidemic to occur; immunity, transmission of a virulent strain of meningococcus, poor living conditions, and population movements and crowding are all considered important. Concurrent infections have also been implicated. In addition, environmental factors may play a key role. Asymptomatic carriage of the meningococcus in the nasopharyngeal mucosa is common in the semi-arid zones of Africa, systemic infection causes invasive disease, and most cases result from infection recently acquired from exposure to carriers (Molesworth et al., 2002b). The geographic predominance of epidemics in the belt region and their seasonal occurrence at dry, dusty times of year, ceasing with the onset of the rains (Molesworth et al., 2001) suggest that environmental conditions are important, although the mechanisms by which they may work are poorly understood. It has been postulated that low humidity and dust may damage the mucosal barrier, and/ or inhibit its immune defense mechanism, thus increasing mucosal invasion and the risk of overt disease; indoor overcrowding in adverse environments (intense dust events) may enhance the spread of infection (Greenwood et al., 1987). In a recent review of the spatial and temporal occurrence of meningococcal meningitis epidemics across Africa, it has been clearly shown that the Sahel bears the greatest epidemic burden of meningococcal meningitis with over two-thirds of documented outbreaks and high attack rates (Molesworth et al., 2002b). A recent analysis of epidemic distribution and environmental data indicates that the absolute humidity profiles and land-cover types can be used to distinguish between areas with high and low risk of epidemics, and also that population density and dust may be implicated in some regions (Molesworth et al., 2003). An increase in epidemic occurrence in the southern extent of the meningitis belt may be related to an increase in dust storms in the region (Molesworth et al., 2002a), which may be a consequence of the recent drought periods.
DISCUSSION 
Climate has had an enormous impact on health in West Africa. However, its underlying role in determining the burden of disease in the region on a yearly or decadal basis is not mentioned in key health policy documents, such as those that reflect the framework and indicators promoted for monitoring the achievements of Roll Back Malaria (RBM) (Remme et al., 2001). Its absence is all the more remarkable given that the stated aim of RBM is ‘‘to halve the burden of malaria through interventions adapted to local needs and strengthening of the health sector.’’ If observations indicating that climate variation alone may be responsible for changes in prevalence rates of 50% in some areas, then indicators of the effectiveness of RBM interventions (which do not incorporate climate variability as a confounding factor) may be either overly optimistic or pessimistic as a result. This is considerable evidence that many infectious diseases such as malaria occur within a climate ‘‘envelope’’ and vary between years in intensity as a result of interannual variation in climate and vulnerability. There is a clear need for researching and documenting more closely the extent to which morbidity and mortality from infectious disease, and their broad socio-economic impacts, is determined by climate, and whether or not climate/environmental information (such as seasonal forecasts or rainfall or vegetation monitoring) can, in practice, improve decision making for control purposes in the affected countries. Currently, there is much interest in developing adaptation strategies for poor countries to cope with the likely impacts of climate change, and yet we know little about
how shifts in climate have affected health outcomes in the recent past. Whether or not the recent drought in West Africa is part of the natural cycle of climate variability or indicative of anthropogenic climate change, we can learn much from the region’s experience in terms of impacts and adaptation to extreme shifts in the region’s climate over a relatively short period of time.
ACKNOWLEDGMENTS
The authors thank Alessandra Giannini for providing Figure 1, and Valerie Obsomer for providing the digital data used in Figures 4 and 5, respectively. M.C.T., S.J.C., and N.W. are supported by a grant/cooperative agreement number NA67GP0299 from the National Oceanic and Atmospheric Administration.
REFERENCES
Baker RHA, Guillet P, Seketeli A, Poudiougo P, Boakye D, Wilson MD, et al. (1990) Progress in controlling the reinvasion of windborne vectors into the western area of the Onchocerciasis Control Program in West-Africa. Philosophical Transactions of the Royal Society of London. Series B: Biological Sciences
(London). 328:731–750
Batterbury S, Warren A (2001) The African Sahel 25 years after the great drought: assessing progress and moving towards new agendas and approaches. Global Environmental Change
11:1–8
Benton B, Bump J, Seketeli A, Liese B (2002) Partnership and promise: evolution of the African river-blindness campaigns. Annals of Tropical Medicine and Parasitology 96:5–14 Boakye DA, Back C, Fiasorgbor GK, Sib APP, Coulibaly Y (1998)
Sibling species distributions of the Simulium damnosum complex in the West African Onchocerciasis Control Programme area during the decade 1984–93, following intensive larviciding since 1974. Medical and Veterinary Entomology 12:345–358
Brengues J (1975) La Filariiose de Bancroft en Afrique de L’Ouest, Paris: Orstom, pp 299 Brewster DR, Greenwood BM (1993) Seasonal-variation of pediatric diseases in the Gambia, West-Africa. Annals of Tropical
Paediatrics 13:133–146
Cheesbrough JS, Morse AP, Green SDR (1995) Meningococcal meningitis and carriage in western Zaire—a hypoendemic zone related to climate. Epidemiology and Infection 114:75–92 Colwell R, Epstein P, Gubler D, Hall M, Reiter P, Shukla J, et al. (1998) Global climate change and infectious diseases. Emerging Infectious Diseases 4(3):451–452; available: http://www.cdc. gov.ncidod/EID/Vol4No3/colwell.htm Connor SJ (2002a) Managing health and disease in developing countries. In: Arnold Companion to Development Studies, Desai V, Potter R (editors), London: Arnold, pp 396–400
Connor SJ (2002b) The use of environmental information systems (EIS) for malaria control planning in Africa. Faculty of Medicine, Liverpool, UK: University of Liverpool (unpublished thesis) Cour JM (2001) The Sahel in West Africa: countries in transition to a full market economy. Global Environmental Change 11:31– 47 Davies VAB (2000) Sierra Leone: Ironic tragedy. Journal of African Economies 9:349–369
Dilley M, Heyman BN (1995) ENSO and disaster: droughts, floods and El Nin˜o/southern oscillation warm events. Disasters 19:181–193
Egg J, Gabas JJ (editors), (1997) Preventing Food Crisis in the Sahel: Ten Years of Network Experience in Action: 1985–1995, Paris: OECD El Sayed BB, Arnot DE, Mukhtar MM, Baraka OZ, Dafalla AA, Elnaiem DE, et al. (2000) A study of the urban malaria transmission problem in Khartoum. Acta Tropica 75:163–171
Emerson PM, Bailey RL, Mahdi OS, Walraven GEL, Lindsay SW (2000) Transmission ecology of the fly Musca sorbens, a putative vector of trachoma. Transactions of the Royal Society of Tropical Medicine and Hygiene 94:28–32