Figures
Abstract
Anopheles gambiae s.s mosquitoes are important vectors of lymphatic filariasis (LF) and malaria in Ghana. To better understand their ecological aspects and influence on disease transmission, we examined the spatial distribution of the An. gambiae (M and S) molecular forms and associated environmental factors, and determined their relationship with disease prevalence. Published and current data available on the An. gambiae species in Ghana were collected in a database for analysis, and the study sites were georeferenced and mapped. Using the An. gambiae s.s sites, environmental data were derived from climate, vegetation and remote-sensed satellite sources, and disease prevalence data from existing LF and malaria maps in the literature. The data showed that An. gambiae M and S forms were sympatric in most locations. However, the S form predominated in the central region, while the M form predominated in the northern and coastal savanna regions. Bivariate and multiple regression analyses identified temperature as a key factor distinguishing their distributions. An. gambiae M was significantly correlated with LF, and 2.5 to 3 times more prevalent in the high LF zone than low to medium zones. There were no significant associations between high prevalence An. gambiae s.s locations and malaria. The distribution of the An. gambiae M and S forms and the diseases they transmit in Ghana appear to be distinct, driven by different environmental factors. This study provides useful baseline information for disease control, and future work on the An. gambiae s.s in Ghana.
Citation: de Souza D, Kelly-Hope L, Lawson B, Wilson M, Boakye D (2010) Environmental Factors Associated with the Distribution of Anopheles gambiae s.s in Ghana; an Important Vector of Lymphatic Filariasis and Malaria. PLoS ONE 5(3): e9927. https://doi.org/10.1371/journal.pone.0009927
Editor: Photini Sinnis, New York University, United States of America
Received: November 2, 2009; Accepted: March 2, 2010; Published: March 29, 2010
Copyright: © 2010 de Souza et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This study was supported by the JRS Biodiversity Foundation (http://www.jrsbdf.org/v2/home.asp) grant to Prof Boakye. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
The mosquito Anopheles gambiae sensu lato (s.l.) contains seven species, of which An. gambiae sensu stricto (s.s), An. arabiensis and An. melas are three of the major vectors of lymphatic filariasis (LF) and malaria caused by Wuchereria bancrofti, and Plasmodium falciparum respectively in West Africa [1], [2]. In Ghana, previous studies have found the An. gambiae s.l and the An. funestus to be the major vectors of LF in the southern coastal zone and in the northern region of the country [3]–[7].
Vector control is considered an important tool for diseases transmitted by mosquitoes and other insect vectors [8]. The current Ghana policy on vector control against Anopheles vectors prioritizes the use of insecticide treated materials and indoor residual spraying. However, the efficiency of these interventions will depend on information on the distribution and abundance of the main vectors, the specific molecular forms of An. gambiae s.s and the levels of insecticide resistance within them [9], [10].
Studies on the distribution of the An. gambiae complex in Africa [9], [11], highlighted gaps in our knowledge, and Coetzee and colleagues [9] further recommended an urgent need for baseline surveys on the distribution of these malaria vectors in areas where no reliable information exists. Ghana is one of these countries. The largest and most recent study published on the abundance, distribution and levels of insecticide resistance in An. gambiae s.s covered 11 sites across different ecological zones of the country [12]. This study found that the An. gambiae S form predominated across the country, except in the arid north where only the M form was found. While this was the first major attempt at presenting the geographical distribution and ecological variations in An. gambiae s.s in Ghana, it is not known how these data compare with other An. gambiae s.s data, whether there are specific environmental factors driving the distributions of the M and S molecular forms, and if there are any associations with LF and malaria distributions [13], [14].
In order to better understand the ecological aspects of this important vector, their influence on the epidemiology of LF and malaria, this study collated data on An. gambiae s.s in Ghana, and aimed to i) examine the spatial distribution of the An. gambiae M and S forms across the country, ii) identify key environmental factors associated with their distribution, and iii) determine their relationship with LF and malaria prevalence distributions. It is envisaged that this information will help develop a comprehensive profile on the ecology of An. gambiae s.s in Ghana, which will assist researchers and the diseases control programs in the country.
Methods
Study site
The Republic of Ghana is a developing West African country bordered to the north by Burkina Faso, to the east by Togo, to the West by Cote d'Ivoire and to the south by the Gulf of Guinea [15]. It has a total surface area of 239, 460 sq km with a coastline of 539 km, and approximately a population of 23 million inhabitants. The climate is warm and comparatively dry along the southeast coast, hot and humid in the southwest, and hot and dry in north. The ecology can be divided into six zones; the mangrove zones at the coastline, the coastal savanna, the evergreen forest in the south-west, the moist semi-deciduous forest in the central area, the guinea savanna in the north and the sudan savanna in the north-east. The terrain is mostly low plain, with dissected plateau in the south-central area. The elevation ranges from 0 to 880 m above sea level. There are two main seasons; a rainy season from April to October and dry season from November to March.
Entomological data and mapping
Entomological studies on An. gambiae s.l in Ghana were identified from various sources including published articles in peer-reviewed journals, unpublished works from MPhil and PhD theses held at the Noguchi Memorial Institute for Medical Research (NMIMR), Accra-Ghana, as well as on-going studies at NMIMR. The collated data spanned from 2001 to 2008. Information on the location, study period, sample size, collection method, mosquito species and molecular forms, from each study were collated into a database. Various collection methods including; human landing catches, pyrethroid spray catches, larval collections and aspirators were used, depending on the location of the study sites and the objectives of the various studies. However, irrespective of the collection method, data collected from the same location in different years were considered separately. With the exception of few sites, most locations (i.e collection site) in the database were geo-referenced using the latitude and longitude coordinates obtained by cross-checking the names with data from the GEOnet Names Server [16], and Directory of Cities and Towns in the World [17] databases.
All data were imported into the geographical information systems software ArcGIS 9.2 (ESRI, Redlands, CA) for mapping and spatial analyses. First, the overall distribution of the An. gambiae s.l, and the different prevalence distributions of An. gambiae M and S across the country were mapped. Mosquito collection methods were also compared, to highlight differences in sampling. Second, spatial analysis of the An. gambiae M and S prevalences were examined using ArcGIS Spatial Analyst and Statistics tools (ESRI, Redland, CA). The Moran's I statistic was used to determine spatial autocorrelation patterns i.e. clustered, dispersed, random, and the Getis-Ord Gi* statistic was used to identify the specific locations where high and low prevalences were clustered (Z scores, 95% confidence levels (CI) +1.96 and −1.96 standard deviations). In addition, the kernel density estimation (KDE) method, non-parametric way of estimating the probability density function, was used to create a continuous surface representing the high to low density distributions of each molecular form.
Environmental data and analysis
To examine environmental factors associated with the prevalence distributions of An. gambiae M and S in Ghana, specific data on elevation, vegetation, precipitation, temperature and humidity were obtained for each location (i.e. collection site), and compiled into a database, for descriptive and statistical analyses in SPSS 16.0 (SPSS, Inc, Chicago, IL).
Elevation data were derived from the U.S Geological Survey's ETOPO2 Digital Elevation Model available from ESRI (Redlands, CA). The elevation at each collection site was determined by importing the digital elevation map into ArcGIS 9.2 and extracting the underlying value (metres). Vegetation and climate data were based on the mean annual values for the specific year of study, obtained from the best available sources via the Climate Data Library of the International Research Institute for Climate and Society [18]. Vegetation cover was based on Normalized Difference Vegetation Index (NDVI) satellite data extracted from the LandDAAC MODIS version 005 West Africa from USGS [19], [20]. Precipitation (mm), temperature (C°) and specific humidity (qa) measures were obtained from satellite data from the National Oceanic and Atmospheric Administration (NOAA) and based on daily mean readings taken 2 meters above the ground [21]–[25].
First, the relationship between An. gambiae M and S, and each environmental variable was examined using bivariate correlations, Pearson's correlation coefficient (2 tailed P values≤0.05 significance). Stepwise multiple linear regression analysis was then used to identify the environmental factor that would best predict the distribution of each molecular form. To account for environmental variables that may be highly correlated with each other, the level of colinearity tolerance in the stepwise regression procedure was set at ≥0.8 and only variables above this threshold were accepted in the models. Second, to better understand the environmental parameters associated with the An. gambiae M and S forms, mean environmental measures between high and low prevalence sites were compared using the Mann-Whitney U test with Bonferroni correction for multiple comparisons.
Relationship with disease prevalence distributions
To examine the relationship between the An. gambiae M and S prevalence and the distribution of disease, maps on the LF prevalence [13] and P. falciparum prevalence [14] for West Africa were imported into ArcGIS 9.2 and geo-referenced. The LF map was modeled from the W. bancrofti seroprevalence data collected in 2000 from 401 villages throughout Benin, Burkina Faso, Ghana and Togo [13]. The P. falciparum malaria prevalence map was modelled on extensive data obtained from children aged 2–10 years in non-epidemic periods, using a generalized linear mixed model [14].
The An. gambiae s.s collection sites were used as focal points, whereby the underlying disease prevalence data could be compared with the entomological data. The LF and malaria prevalence data, corresponding to the latitude and longitude of each mosquito collection site, were extracted and exported for descriptive and statistical analyses, which included bivariate correlations, and comparison of means between high and low prevalence sites of the An. gambiae M and S forms.
Further, it was of particular interest to explore the entomological and environmental characteristics in different LF transmission zones based on the prevalence data map in Gyapong et al. 2002 [13]. The prevalence distributions ranged from 0 to 30%, and were classified into distinct transmission zones, which were digitized in ArcGIS 9.2. The entomological and environmental data from each of collection sites within each zone were summarized, and significant differences identified by comparing the standard errors (+2SE) of the means.
Results
Entomological mapping
The collated An. gambiae species complex database contained 143 records with a total of 12,607 mosquitoes, reflecting both larval and adult catches. From this, the distribution of An. gambiae s.l was mapped (Figure 1). The most dominant species was An. gambiae s.s, which was found at 114 sites (total n = 10,028), followed by An. melas (6 sites, total n = 469) and An. arabiensis (8 sites, total n = 240). The map indicates that the distribution of An. gambiae s.s was widespread, while An. melas was primarily found along the coastal Savanna zones, a predominantly marshy environment, and An. arabiensis mainly in the northern savanna zone. Anopheles funestus, the second most important vector of malaria and lymphatic filariasis in Ghana, was also recorded in 9 sites with a total of 1,825 mosquitoes. Refer to supplementary data File S1 “Anopheles distribution records in Ghana” for further information on An. funestus and other Anopheles mosquito species, collection sites and coordinates, year and month of collection (where available), collection methods, numbers collected and identified, and the data sources.
Of the An. gambiae s.s data, a total of 70 collection sites had information on the M form (n = 2,826) and S form (n = 4,098). The collection sites were located predominately in three geographical regions i.e. the south eastern, central western and north central. The distribution of An. gambiae M and S forms varied across the country in different proportions (Figure 2). Figure 2 indicates the distribution of each species according to the various collection methods. Overall, An. gambiae M and S were sympatric in most locations. However, An. gambiae M form was more prevalent in the northern savanna, and coastal savanna areas of the country, and in four sites it was the only species observed. In contrast, An. gambiae S form was more prevalent in the middle region of the country, and in seven locations it was the only species collected. Interestingly, bivariate correlation analysis between each species indicated that their prevalences were negatively correlated (−0.763).
While, it is possible that the geographical grouping of collection sites may influence trends, the spatial analyses carried out in this study indicated a positive spatial autocorrelation or clustering for both the An. gambiae M (MI = 0.19, Z score = 4.2, P≤0.01) and An. gambiae S (MI = 0.19, Z score = 4.2, P≤0.01) forms. The resultant Z scores of the Getis-Ord Gi* hot spot analyses (using inverse-distance weighting), indicated similar trends with significantly different clustering of high and low prevalences of the An. gambiae M and S forms. These spatial trends were overlaid a density distribution surface map shown in Figure 3, further highlighting the high to low patterns of each species.
Environmental analysis
The relationship between the prevalence of An. gambiae M and S forms, and the environmental variables are shown in Table 1. Overall, bivariate correlation analysis indicated that the An. gambiae M form was significantly positively associated with temperature (r = 0.51), and negatively with elevation (r = −0.28), precipitation (r = −0.33), and humidity (r = −0.26). This contrasts to the An. gambiae S form, which was found to be significantly negatively associated with temperature (r = −0.58), and positively with elevation (r = 0.30) and rainfall (r = 0.41). Interestingly, elevation, precipitation and temperature correlations increased when data were stratified by the two main collection methods, HLC and larval collections (Table 1). Multiple regression analyses of all data (n = 70), indicated that temperature was an important variable for both molecular forms, explaining for An. gambiae M, 28% (R2 = 0.28, F = 25.8, P≤0.001) and for An. gambiae S, 36% (R2 = 0.36, F = 37.9, P≤0.001) of the variance in the model (Table 2).
For each molecular form, comparisons of environmental measures between locations with significantly high and low prevalences, defined by positive Z scores (≥+1.96) and negative Z scores (≤−1.96) respectively, are shown in Table 3. Overall, locations with high An. gambiae M prevalences had higher NDVI and temperatures, but lower elevation, precipitation and humidity measures than those locations with lower prevalences by these species and/or where the prevalence of An. gambiae S form was higher. Statistical comparisons indicated significant differences (P value<0.004 Bonferroni corrected) between elevation, precipitation and temperature for both An. gambiae M and S forms between the high and low prevalence areas.
To further explore the differences in elevation, precipitation and temperature, the mean prevalence of An. gambiae M and S forms was plotted across a range of environmental groupings (Figure 4). An. gambiae M prevalences were found to be higher at elevations of 0–200 m, and where mean daily precipitation ranged between 1.0–2.5 mm, and mean daily temperatures ranged between 26.1–27.6°C. In contrast, An. gambiae S prevalences were found to be higher at elevations >200 m, and where mean daily precipitation ranged between 2.6–3.8 mm, and mean daily temperatures ranged between 24.5–26.0°C.
Disease association
The relationship between the An. gambiae M and S, and LF and malaria prevalences were first examined using bivariate correlation analysis. Results shown in Table 1, indicate that An. gambiae M was significantly positively associated with LF (r = 0.46), while An. gambiae S form was significantly negatively associated with LF (r = −0.48), but positively with malaria (r = 0.26). Interestingly, correlation analysis between each mosquito species, and each disease indicated significant negative associations between An. gambiae M and An. gambiae S (r = −0.76), and between LF and malaria prevalence (r = −0.41).
Second, we compared disease prevalences between high and low An. gambiae M and S sites (as described above). Overall, locations with high An. gambiae M prevalences (Z scores≥+1.96), were found to have significantly higher LF prevalences (20%) than those locations with low prevalences (Z scores≤−1.96), by these species (2.4%) and/or where the prevalence of An. gambiae S form was significantly high (2.2%). No significant differences were found between malaria prevalence and each mosquito species (Table 3).
Finally, we examined the LF data in Gyapong et al. 2002 [13], and identified three main transmission zones i.e. zero/low (<1%, n = 19), medium (1–10%, n = 32) and medium/high (10–30%, n = 19). Graphical presentation of these three zones is included in Figure 5, which summarizes mean entomological and environmental measures for the sites within each zone. In the medium/high LF transmission zone, the mean An. gambiae M prevalence (53.2%,) and temperatures (26.8°C) were found to be significantly higher (±2SE), and An. gambiae S prevalence (25.6%) significantly lower, than those found in medium transmission (21.1%; 25.5°C; 64.7%) and low transmission (21.4; 25.3°C; 64.4%) zones. In the zero/low transmission zone, precipitation measures (2.83 mm) were found to be significantly higher (+2SE), than those in the medium transmission (2.34 mm) to high transmission (1.99 mm) zones (Figure 5). Refer to the supplementary File S2 for further details on Figure 5.
Note: The maps represent the LF transmission zones in Ghana. The graphs below summarize the mean entomological and environmental measures for the sites within each zone. NDVI stands for Normalized Difference Vegetation Index.
Discussion
Most entomological studies on An. gambiae s.l in Ghana [3], [4] have focused on small area/district based collections except the study by Yawson and colleagues [12]. Although the study by Yawson et al. attempted a broader coverage, this study represents the first nationwide review aimed at identifying the distribution of the various member species of the An. gambiae s.l, the molecular forms of the An. gambiae s.s, together with key environmental drivers and how they may relate to diseases. This study confirms previous observations that An. gambiae s.l is the major human biting mosquito species in Ghana [3], [4], [12], [26] and within the An. gambiae s.l, An. gambiae s.s is the predominant species [3], [4]. Other members of the An. gambiae s.l found in Ghana are An. arabiensis and An. melas.
The distribution of the An. gambiae M and S forms varied significantly across the country. The two molecular forms were found sympatrically in most locations, except in some areas in the middle region of the country where only the An. gambiae S form was observed, and in certain areas in the northern savanna and coastal savanna areas where only the An. gambiae M form was observed. This is confirmed by the spatial analysis and high Z score values in An. gambiae M/S form dominant areas. Also, the clustering remained relatively consistent irrespective of the different collection methods. Along the coast and in the northern savanna, the An. gambiae M form was predominant and clustered, while the S molecular form was most common and clustered in the middle belt. This positive spatial autocorrelation indicates that An. gambiae M/S distributions are geographically defined, and nearby areas are likely to comprise the same or similar species compositions, than those further away. The distribution of each species is also influenced by distinct and geographical related environmental factors and habitat characteristics. For example, the dominance of the An. gambiae M form in the northern and coastal savanna areas may be due to the wide presence of permanent breeding conditions provided by irrigation facilities [27] and ponds of water resulting from rivers run-offs since the M form is known to be associated with flooded areas, while the S form is heavily dependent on rainfall [12], [28]. The dominance of the An. gambiae S form in the middle region of the country may be explained by the fact that this region is mountainous, forested, with lower mean temperatures and the highest recorded rainfall in Ghana, which supports the findings of our study. In sub-Saharan Africa the abundance and distribution of Anopheles mosquito species is dependent on environmental factors and ecological zones [29]–[32] as well as on human population changes and anthropological effects, which may lead to land-use changes ultimately affecting vector distribution and abundance [33], [34] and as shown in a recent paper by Costantini and colleagues [35]. This current study did not focus on human factors but was able to show the wide variability in abundance and distribution of the An. gambiae M and S forms, which appear to be driven by a range of environmental factors. The relation between human population density and vector distribution is an indirect one and difficult to measure. However, it may be inferred that the NDVI, which is a measure of vegetation greenness and density, is somehow affected by the degree of land occupancy and exploitation by humans through the action of occupation and clearance for agricultural and developmental purposes.
Variations in the vector population densities of these two molecular forms have been observed in populations in Mali and Cameroon as well as between the various chromosomal forms in Mali [32], [35], [36]. Observations on the distributions and the predominance of the An. gambiae S form in a larger part of the country from this study confirm suggestions that the An. gambiae S form, has broader environmental ranges, and therefore is found in more locations than the M form [37]. Our environmental analyses suggest that elevation, precipitation and temperature are important variables driving the spatial distribution of each mosquito species, and the differences between them. In particular, temperature appears to be a key factor distinguishing the two species; probably due to influences on their production as reported for An. arabiensis and An. gambiae s.s [38]. The An. gambiae M form was more prevalent and seemingly better adapted to higher temperatures, than the S form. This is in agreement with the suggestion that the An. gambiae M form shows a more latitudinal range in West Africa than the S form [39], being the most dominant form encountered in hot, arid regions of the Sudan-savanna or Sahelian zones [27], [35], [40]–[42].
The mosquito vectors' association with each disease might explain the negative association between LF and malaria described by Kelly-Hope and colleagues [43]. In this current study, malaria prevalence was positively associated with the An. gambiae S form, whilst high LF prevalence and high transmission zones were associated with high temperatures and significantly high An. gambiae M prevalences. This relationship between An. gambiae M and S forms with LF and malaria supports previous studies that suggest the Mopti form of An. gambiae s.s is more associated with W. bancrofti than malaria transmission [44], [45] and also that it is a relatively poor vector of malaria compared with other species such as the Savanna form of An. gambiae [46], [47]. The LF distribution map by Gyapong and colleagues [13], indicates that the disease distribution in West Africa has the highest prevalence in the hotter Sudan/Sahel savanna areas, which are also An. gambiae Mopti chromosomal form dominant areas [41], [43].
The results from this study provide useful information on the distribution of the An. gambiae M and S forms in Ghana, highlighting the environmental factors that may play a role in determining their distributions. The information provided also marks a beginning in understanding the LF disease distribution pattern in Ghana relative to the forms of the An. gambiae s.s. Despite the limitations of this study in using previously modeled LF and malaria data, these results are very useful for disease control and allocation of resources, especially for LF, which together with its vectors appear to be restricted to hotter, less elevated regions of the country. Another limitation to this study is the small number of sites involved in the analysis of the An. gambiae M and S distribution. This coupled with their uneven distribution between the southern half and the northern half of the country could introduce potential bias in the analyses. Undertaking spatial statistical analysis with a larger dataset in the future, and further modeling their distributions through ecologic niche modeling [11], [35], [48], could further enhance our understanding of the distribution of the disease and its vectors, as well as defining the spatial limits of the vectors' distribution which is crucial for disease control. This would, however, require a constant update of the vector database generated in this study as well as collection of data from remote areas of the country where no available data exists.
Supporting Information
File S1.
Data file for Anopheles mosquito species, collection sites and coordinates, year and month of collection (where available), collection methods, numbers collected and identified, and the data sources.
https://doi.org/10.1371/journal.pone.0009927.s001
(0.04 MB XLS)
File S2.
Summary of entomological and environmental variables in different LF transmission zones.
https://doi.org/10.1371/journal.pone.0009927.s002
(0.08 MB DOC)
Acknowledgments
We thank Dr Michael Bell, The Earth Institute, Columbia University, USA for his advice on the best environmental data sources. We are grateful to the Director of the Noguchi Memorial Institute for Medical Research, University of Ghana for the permission to publish this work.
Author Contributions
Conceived and designed the experiments: DKdS LAKH MDW BL DAB. Analyzed the data: DKdS LAKH. Contributed reagents/materials/analysis tools: DAB. Wrote the paper: DKdS LAKH MDW BL DAB.
References
- 1. Larivière M, Diallo S, Picot H (1966) Aperçus épidémiologiques sur la filariose lymphatique en Afrique de l'Ouest. Bull Soc Med Afr Noire de Langue Francaise 11(3): 670–680.
- 2. Lenhart A, Eigege A, Kal A, Pam D, Miri ES, et al. (2007) Contributions of different mosquito species to the transmission of lymphatic filariasis in central Nigeria: implications for monitoring infection by PCR in mosquito pools. Filaria J.
- 3. Appawu MA, Baffoe-Wilmot A, Afari EA, Nkrumah FK, Petrarca V (1994) Species composition and inversion polymorphism of the Anopheles gambiae complex in some sites of Ghana, West Africa. Acta Trop 56: 15–23.
- 4. Appawu MA, Dadzie S, Baffoe-Wilmot A, Wilson MD (2001) Lymphatic filariasis in Ghana: entomological investigation of transmission dynamics and intensity in communities served by irrigation systems in the Upper East Region of Ghana. Trop Med Int Health 6(7): 511–516.
- 5. Boakye DA, Wilson MD, Appawu MA, Gyapong J (2004) Vector competence of Anopheles mosquito populations in the Bongo district of Ghana to Wuchereria bancrofti. Ann Trop Med Parasitol 98: 501–508.
- 6. Dunyo SK, Appawu M, Nkrumah FK, Baffoe-Wilmot A, Pedersen EM, et al. (1996) Lymphatic filariasis on the coast of Ghana. Trans R Soc Trop Med Hyg 90(6): 634–638.
- 7. Afari EA, Appawu M, Dunyo S, Baffoe-Wilmot A, Nkrumah FK (1995) Malaria infection, morbidity and transmission in two ecological zones Southern Ghana. Afr J Health Sci 2(2): 312–315.
- 8.
World Health Organization [WHO] (2004) Global Strategic Framework for Integrated Vector Management. Available: http://whqlibdoc.who.int/hq/2004/WHO_CDS_CPE_PVC_2004_10.pdf Accessed 18 October 2008.
- 9. Coetzee M, Craig M, le Sueur D (2000) Distribution of African malaria mosquitoes belonging to the Anopheles gambiae Complex. Parasitol Today 16(2): 74–77.
- 10. Kelly-Hope L, Ranson H, Hemingway J (2008) Lessons from the past: managing insecticide resistance in malaria control and eradication programmes. Lancet Infect Dis 8(6): 387–389.
- 11. Levine RS, Peterson AT, Benedict MQ (2004) Geographic and ecologic distributions of the Anopheles gambiae complex predicted using a genetic algorithm. Am J Trop Med Hyg 70(2): 105–109.
- 12. Yawson AE, McCall PJ, Wilson MD, Donnelly MJ (2004) Species abundance and insecticide resistance of Anopheles gambiae in selected areas of Ghana and Burkina Faso. Med Vet Entomol 18: 372–377.
- 13. Gyapong JO, Kyelem D, Klienschmidt I, Agbo K, Ahouandogbo F, et al. (2002) The use of spatial analysis in mapping the distribution of bancroftian filariasis in four West African countries. Ann Trop Med Parasitol 96: 695–705.
- 14. Kleinschmidt I, Omumbo J, Briet O, van de Giessen N, Sogoba N, et al. (2001) An empirical malaria distribution map for West Africa. Trop Med Int Health 6: 779–786.
- 15.
Government of Ghana Official Website (2009) Available: http://www.ghana.gov.gh/ Accessed 11 March 2009.
- 16.
NGA GEOnet Names Server (GNS) (2008) Available: (http://earth-info.nga.mil/gns/html/) Accessed 2 October 2008.
- 17.
Directory of Cities and Towns in the World, Global Gazetteer Version 2.1. Available: (http://www.fallingrain.com/world/). Copyright 1996–2006 by Falling Rain Genomics, Inc. Accessed 2 October 2008.
- 18.
IRI (2008) IRI/LDEO Climate Data Library of the International Research Institute for Climate and Society, Columbia University. Available: (http://iridl.ldeo.columbia.edu/index.html) Accessed 20 October 2008.
- 19.
USGS 2008: United States Geological Survey (2008) Normalized Difference Vegetation Index: LandDAAC MODIS version_005 WAF NDVI. Available: (http://iridl.ldeo.columbia.edu/SOURCES/.USGS/.LandDAAC/.MODIS/.version_005/.WAF/.NDVI/). Accessed 20 October 2008.
- 20. Huete A, Didan K, Miura T, Rodriguez EP, Gao X (2002) Overview of the radiometric and biophysical performance of the MODIS vegetation indices. Remote Sens Environ 83 (1–2): 195–21.
- 21.
NOAA (2008a) DAILY Diagnostic above_ground temp Temperature from NOAA NCEP-NCAR CDAS-1: Climate Data Assimilation System I; NCEP-NCAR Reanalysis Project. Available: http://iridl.ldeo.columbia.edu/SOURCES/.NOAA/.NCEP-NCAR/.CDAS-1/.DAILY/.Diagnostic/.above_ground/.temp/). Accessed 20 October 2008.
- 22.
NOAA (2008b) NCEP CPC FEWS Africa DAILY RFEv2 Estimated Precipitation: CPC/Famine Early Warning System Daily Estimates. Available: (http://iridl.ldeo.columbia.edu/SOURCES/.NOAA/.NCEP/.CPC/.FEWS/.Africa/.DAILY/.RFEv2/.est_prcp/). Accessed 20 October 2008.
- 23.
NOAA (2008c) DAILY Diagnostic above_ground qa Specific humidity from NOAA NCEP-NCAR CDAS-1: Climate Data Assimilation System I; NCEP-NCAR Reanalysis Project. Available: http://iridl.ldeo.columbia.edu/SOURCES/.NOAA/.NCEP-NCAR/.CDAS-1/.DAILY/.Diagnostic/.above_ground/.qa/). Accessed 20 October 2008.
- 24.
Love T (2002) The Climate Prediction Center Rainfall Algorithm Version 2, NOAA.
- 25. Kalnay E, Kanamitsu M, Kistler R, Collins W, Deaven D, et al. (1996) The NCEP/NCAR 40-Year Reanalysis Project. Bull Am Meteorol Soc 77: 437–471.
- 26. Appawu MA, Baffoe-Wilmot A, Afari EA, Dunyo S, Koram KA, et al. (2001) Malaria vector studies in two ecological zones in southern Ghana. Afr Entomol 9: 59–65.
- 27. Touré YT, Petrarca V, Traoré SF, Coulibaly A, Maiga HM, et al. (1998) Distribution and inversion polymorphism of chromosomally recognized taxa of the Anopheles gambiae complex in Mali, West Africa. Parassitologia 40: 477–511.
- 28. Diabaté A, Baldet T, Chandre C, Dabire KR, Kengne P, et al. (2003) KDR mutation, a genetic marker to assess events of introgression between the molecular M and S forms of Anopheles gambiae (Diptera: Culicidae) in the tropical savanna area of West Africa. J Med Entomol 40: 195–198.
- 29. Zhou G, Munga S, Minakawa N, Githeko AK, Yan G (2007) Spatial Relationship between Adult Malaria Vector Abundance and Environmental Factors in Western Kenya Highlands. Am J Trop Med Hyg 77(1): 29–35.
- 30. Muturi EJ, Mwangangi J, Shililu J, Muriu S, Jacob B, et al. (2007) Mosquito Species Succession and Physicochemical Factors Affecting Their Abundance in Rice Fields in Mwea, Kenya. J Med Entomol 44(2): 336–344.
- 31. Yawson AE, Weetman D, Wilson MD, Donnelly J (2007) Ecological zones rather than molecular forms predict genetic differentiation in the malaria vector Anopheles gambiae s.s. in Ghana. Genetics 175: 751–761.
- 32. Sogoba N, Vounatsou P, Bagayoko MM, Doumbia S, Dolo G, et al. (2008) Spatial distribution of the chromosomal forms of anopheles gambiae in Mali. Malaria J.
- 33. Walsh JF, Molyneux DH, Birley MH (1993) Deforestation: effects on vector-borne disease. Parasitology 106: S55–S75.
- 34. Molyneux DH (2003) Climate change and tropical disease: Common themes in changing vector-borne disease scenarios. Trans R Soc Trop Med Hyg 97: 129–132.
- 35. Costantini C, Ayala D, Guelbeogo WM, Pombi M, Some CY, et al. (2009) Living at the edge: biogeographic patterns of habitat segregation conform to speciation by niche expansion in Anopheles gambiae. BMC Ecology 9: 16.
- 36. Slotman MA, Tripet F, Cornel AJ, Meneses CR, Lee Y, et al. (2007) Evidence for subdivision within the M molecular form of Anopheles gambiae. Mol Ecol 16(3): 639–649.
- 37. Esnault C, Boulesteix M, Duchemin JB, Koffi AA, Chandre F, et al. (2008) High genetic differentiation between the M and S molecular forms of Anopheles gambiae in Africa. PLoS ONE 3(4): e1968.
- 38. Kirby MJ, Lindsay SW (2008) Effect of temperature and inter-specific competition on the development and survival of Anopheles gambiae sensu stricto and An. arabiensis larvae. Acta Trop 109: 118–123.
- 39. Ndiath MO, Brengues C, Konate L, Sokhna C, Boudin C, et al. (2008) Dynamics of transmission of Plasmodium falciparum by Anopheles arabiensis and the molecular forms M and S of Anopheles gambiae in Dielmo, Senegal. Malar J.
- 40. Calzetta M, Santolamazza F, Carrara GC, Cani PJ, Fortes F, et al. (2008) Distribution and Chromosomal Characterization of the Anopheles gambiae Complex in Angola. Am J Trop Med Hyg 78(1): 169–175.
- 41. Bayoh MN, Thomas CJ, Lindsay SW (2001) Mapping distributions of chromosomal forms of Anopheles gambiae in West Africa using climate data. Med Vet Entomol 15: 267–274.
- 42. della Torre A, Tu Z, Petrarca V (2005) On the distribution and genetic differentiation of Anopheles gambiae s.s. molecular forms. Insect Biochem Mol Biol 35: 754–769.
- 43. Kelly-Hope LA, Diggle PJ Rowlingson BS, Gyapong JO, Kyelem D, et al. (2006) Negative spatial association between lymphatic filariasis and malaria in West Africa. Trop Med Int Health 11(2): 129–135.
- 44. Hunter J (1992) Elephantiasis: a disease of development in north-east Ghana. Social Sci Med 35: 627–649.
- 45. Ijumba JN, Lindsay SW (2001) Impact of irrigation on malaria in Africa: paddies paradox. Med Vet Entomol 15: 1–11.
- 46. Coluzzi M (1993) Advances in the study of afrotropical malaria vectors. Parassitologia 35: 23–29.
- 47. Carnevale P, Guillet P, Robert V, Fontenille D, Doannio J, et al. (1999) Diversity of malaria vectors in rice growing areas of the Afrotropical region. Parassitologia 41: 273–276.
- 48. Stockwell DRB, Peters D (1999) The GARP modelling system: problems and solutions to automated spatial prediction. Int J Geogr Inf Sci 13: 143–158.