ESHAW: Eco-System and Health Analysis Workshop
Bouaké, Côte d'Ivoire, 8-26 November 1999 

With financial support from the IDRC "Ecosystem Approaches to Human Health" Program Inititative

The overall objective of the workshop is to examine correlations between environmental and malaria-related factors at macro level (West Africa) and at micro level (key sites). The basic approach is multiple regression between geo-referenced malaria data and a set of potential environmental factors. The most widely available reliable malaria data are the parasite ratios, measuring the percentage of children with detectable parasites in their blood.

It is expected that differences found over shorter distances are not explained by variation in the same environmental factors which explain differences found over larger distances. In addition to the macro-level regressions, we will, therefore, also undertake in-depth studies of four key sites (two in the forest zones of Cote d'Ivoire and Cameroon, and two in the savanna zones of Cote d'Ivoire/Mali and Ghana/Burkina Faso), using more detailed malaria data as well as high-resolution Landsat and SPOT satellite imagery.
 

Expected products

• A cleaned database of malaria data collected throughout West and Central Africa, will be the first product of the workshop.
• A set of regional maps indicating distribution of physical and functional environmental explanatory variables.
• A statistical model that links malaria endemicity to environmental explanatory factors.
• A predictive malaria endemicity map considering environmental risks factors at macro-scale.
• A working hypothesis on the relationship between environment and malaria related factors (vector densities, transmission, disease) at micro level.

MARA / ARMA - an International Collaboration (http://www.mara.org.za)(http://www.mara.org.za)

• Olivier Briët, West Africa Rice Development Association (WARDA), 01 BP 2551 Bouaké 01, Bouaké, Côte d'Ivoire. Email: o.briet@cgiar.org 
• Immo Kleinschmidt, South African Medical Research Council, 771 Umbilo Road, Durban 4001, Republic of South Africa. Email: kleinsci@mrc.ac.za 
• Nathan Kumasenu Mensah, Navrongo Health Research Center, Navrongo, POBox 114 UE/R, Ghana. Email: navrongo@africaonline.com.gh ; Accra office: Navrongo Health Research Center, Navrongo, Mamprobi Polyclinic, Accra E-mail: nhrc@africaonline.com.gh ;
• Judy Omumbo, Kenya Medical Research Institute/Wellcome Trust Collaborative Programme, Nairobi, Kenya. PO Box 43640, Nairobi, Kenya. Tel: 254 2 710672/715160. Fax: 254 2 711673. Email: Judy@wtrl.or.ke 
• Nafomon Sogoba, Malaria Research & Training Center, Faculté de Medecine de Pharmacie et d’ Odonto-stomatologie, Université du Mali Bamako BP. 1805 Tel. (223) 22 52 77, Fax. (223) 22 49 87 E-mail nafomon@mrtcbko.malinet.ml 
• Thomas Teuscher, West Africa Rice Development Association (WARDA), Bouaké, Côte d'Ivoire. Tel: (+225) 63 45 14; Fax: (+225) 63 47 14; Email: t.teuscher@cgiar.org 

• Moussa Mahaman, West Africa Rice Development Association (WARDA), Bouaké, Côte d'Ivoire, Email: m.moussa@cgiar.org 
• Nicole Teuwsen, GIS consultant, Bouaké, Côte d'Ivoire. Tel: +(225)(31)655202 Email: briet@africaonline.co.ci 
• Nick van de Giesen, Center for Development Studies, Walter_Flex_Str 3, 53113 Bonn University, Bonn, Germany. Tel: +49 228 731720; E-mail: nick@uni_bonn.de 
• Pieter Windmeijer, ALTERRA, P.O.Box 125, 6700 AC, Wageningen, The Netherlands. Tel: + 31 317 474268; Email: p.n.windmeijer@sc.dlo.nl 

• Asaph Anyamba, National Aeronautics and Space Agency (NASA), Washington DC, USA

Malaria-related data

The MARA / ARMA collaboration database consists of over 5,000 reports on malaria endemicity in sub-Saharan Africa. For ESHAW, a sub-sample for West and Central Africa was used. The limitation of the ESHAW’s objectives to West and Central Africa is justified by 1) the fact that more than 50% of populations exposed to a high transmission probability for more than 7 months per year live in the sub-region (see 1^st MARA / ARMA Technical Report) and 2) the availability of environmental and agricultural databases (WARDA, Inland Valley Consortium – IVC). The malaria dataset used for ESHAW contains over 800 geo-referenced data points (discrete survey locations) relating to community based surveys in which at least 50 children under fourteen years of age were examined. This represents over a quarter of a million children surveyed for malaria parasites.

The map below summarizes the malaria data points used for the ecological analysis. During the workshop, we will develop analytical and predictive malaria models based on environmental factors.

Environmental data

Macro scale (approximately 1:5,000,000)

Previous MARA modeling efforts produced country specific models using in the case of Mali distance to water bodies, NDVI rainy season, average maximum temperature March to May and length of rainy season. In Kenya, NDVI and average rainfall and temperature were found to be the best determinants of parasite ratio distribution. In the case of South Africa, early winter rainfall, average maximum winter temperature, closeness to perennial water bodies and distance to the coast explained 64% of observed malaria incidence. ESHAW will evaluate the impact of the following macro-scale variables on the parasite ratio distribution in West and Central Africa:

 1) Normalized Difference Vegetation Index (NDVI)
 

NDVI in February

NDVI in August

The NDVI has proven to be a powerful predictor for malaria transmission in Kenya. The predictive value is not so much due to the direct effect of the vegetation but to the fact that the NDVI reflects general biological growth conditions integrated over a longer period of time. The NOAA-AVHRR satellite collects the NDVI data. The processed images above show the average maximum values for the respective months.

2) Temperature

Minimum temperature in December from 5°C in the North to 23°C in the South. These data are derived from the Hutchinson Climate Dataset for Africa. In contrast to the highland areas of Eastern and Southern Africa, minimum temperatures are not expected to have much effect on malaria vectors in West Africa.

3) Precipitation
 

Precipitation in March

Precipitation in August

The Inter-Tropical Convergence Zone (ITCZ) moves between these months from well below the equator to the savanna zone of West Africa after which it descends in a sinusoidal fashion, giving coastal areas its typical bimodal rainfall pattern. The data reflect monthly averages according to the Hutchinson climate database.

4) Agro-Ecological Zone

Four agro-ecological zones have been distinguished on the basis of the lenght of the growing period, i.e. the period that water is available for vegatative production on well drained soils. It is a function of precipitation, evaporation, and a fixed amount of availble water in the soil. The zones are from south to north i.e. bottom to top: the Equatorial Forest zone (> 270 days) Guinea Savanna zone (165 – 270 days), Sudan Savanna zone (90 –165 days) and the Sahel zone (< 90 days).

5) Distance to Water

In the map below the most important perennial rivers are show in a high resolution image. This image will be used to calculate for every entry of the MARA data base the distance of the sample site to the nearest permanent water body. The first image provides a regional overview of this distance with rivers in black and the colour changing from blue thru green to red with increasing distance. The second view shows a detailed segment where rivers are depicted in blue colour and lighter shades of gray with increasing distance from water bodies.

6) Lithology

Lithology, in combination with the rain fall, will give an indication of the density of wetlands for the different areas. The information on lithology is summarized and five lithology classes described. There are: 1) recent alluvial and marine unconsolidated deposits (aquamarine), 2) medium to coarse-textured sedimentary deposits (blue), 3) cristalline rocks like granites and gneisses (yellow), 4) finer textured rocks like schists (red), and 5) mixed lithologies like quartzites, conglomerates, etc. (dark blue). Dark green indicates unavailability of information.

7) Morphology

The West African region is divided in four morphology classes, ranging from the recent coastal plains and floodplains (aquamarine), the interior plains (blue), the plateaux with a higher altitude (yellow), and the mountainuous areas with high relief (red).

8) Population

Although the workshop’s main focus is on the physical environment, population density will be included in the modeling exercise in a first attempt to study the undoubtedly important impact of socio-economic factors. The data are derived from Uwe Deichmann`s population database.
 

Micro scale (approximately 1:50,000)

One purpose of the micro-scale studies is to develop a tool to quantify wetland densities (including open water bodies) using remote sensing techniques. More detailed malaria prevalence information will be used to evaluate the influence of these environmental determinants on parasite ratio distribution.

Another purpose of the studies at the four micro-level key sites is the development of a ‘drainage density index’ applicable at the macro scale. The proposed index contains information on agro-ecological zone, lithology and precipitation and will be validated against existing location specific information on wetland density.

Left hand image: Enhanced image combining bands 3,4 and 7 in a false color composite: Dark green indicates deteriorated forests, light green indicates healthy forest, pink represents open, barren surfaces including villages, dark blue indicates open water bodies and light blue shows floating vegetation (window depicts the Sassandra river; window width corresponds to approximately 25 km). Right hand image: Thermal band (channel 6) indicates surface temperature thus visualizing the drainage network under forest cover (see upper right hand part of the TM image). This information is used to map wetlands in the forest zone.

Modeling Results and Risk Maps

Watch this page for updated final products early in the new year!