Skip to main content

EO in Malaria Vector Control and Management

Final Report Summary - MALAREO (EO in malaria vector control and management)

Executive summary:

MALAREO is a two year project, started on 1 February 2011. The project is partially funded by the European Commission (EC) and is focusing on the link between malaria and remote sensing in the cross-border region of southern Mozambique, eastern Swaziland and north-eastern South Africa. MALAREO aims to support an efficient and effective vector control management in the project area by introducing earth observation (EO) and geographical information system (GIS) capacities within the malaria control programmes.

The MALAREO project supports the global strategy to have a substantial and sustained reduction in the burden of malaria in the near and mid-term (2015) and the eventual global eradication of malaria in the long term. GIS, EO, global positioning systems (GPS) and spatial statistics play a crucial role to plan, apply and monitor optimal malaria vector control measures. MALAREO wants to promote the use of these techniques by building EO and GIS capacity and providing relevant EO-based products for the malaria research and control community.

The primary objective of MALAREO is to develop technology and implement earth observation (EO) capacities within malaria vector control and management programmes in southern Africa (SA). To achieve this objective, knowledge exchange and capability will occur in two directions, from the European Union (EU) to SA and backwards. By doing this, the project will contribute to the installation of an EO monitoring cell that will support the daily work of the national malaria control programmes (NMCPs).

Project context and objectives:

The World Malaria Report 2011 by the World Health Organisation (WHO) estimated 216 million malaria cases worldwide, of which approximately 81 % or 174 million cases were reported in the African region. There were an estimated 655 000 deaths caused by malaria, of which 91 % were in Africa. Around 86 % of malaria deaths were children under five years of age. In 2008, the roll back nalaria (RBM) partnership prepared a global malaria action plan (GMAP) in line with the 2010 targets of the United Nations (UN) Secretary General. A global strategy outlined the goal to have a substantial and sustained reduction in the burden of malaria in the near and mid-term (2015) and the eventual global eradication of malaria in the long term.

GIS, EO, spatial modelling and spatial statistics play a crucial role to plan malaria vector control. MALAREO is a mixed European-African consortium that aims to build GIS, EO and spatial statistics capacities and implement the use of (new) EO products within the malaria vector control and management programmes in SA.

The project is focusing on the cross-border region of southern Mozambique, eastern Swaziland and north-eastern South Africa. The region is largely undeveloped, this being exacerbated by the fact that it falls within a malaria area. Increased anti-malarial drug resistance and the lack of a malaria control programme in the past (mainly in south Mozambique) have contributed to this impeded development. The three countries in the MALAREO study area are in different stages of malaria control, which causes different conditions for the use of EO products.

There are different epidemiological milestones for malaria elimination as defined by WHO. South Africa is currently in the pre-elimination stage of malaria control. Swaziland is currently in the elimination stage of malaria control and is aiming for elimination of local malaria cases by 2015. Mozambique is currently in the control stage of malaria control.

Project results:

Description of the work performed and the main results achieved

In a first phase, a user survey was conducted to gather information on the current capacity needs and user requirements from malaria researchers and malaria control programmes in the MALAREO study area. Existing capacity and user requirements were analysed and compared with state-of-the-art requirements. A capacity gap analysis identified and prioritised required capacities. This outcome is the basis for defining the EO monitoring solutions and the required capacity building. During the project lifetime, progress should be made in research and development of new EO products as well as in capacity building of basic and advanced skills, where the latter should be linked to the specific EO applications developed in MALAREO.

Identification of GIS and EO requirements

End-user surveys conducted in the study area of MALAREO have shown high interest for linking EO with epidemiology as well as for EO products directly supporting the malaria control programmes (MCP) in their daily work. Following these user requirements, MALAREO is addressing two different applications of earth observation, i.e. EO applications to support malaria epidemiological studies and EO applications to directly support malaria vector control. Epidemiological EO applications mainly address parameters that are suitable to predict the environmental conditions for the vectors. These parameters are used to produce malaria risk maps. EO applications that directly support malaria control aim at providing relevant geo-data that optimise planned vector control measures, which is a new field for EO applications.

The skills of staff were interrogated against the state-of-the-art of EO applications and statistical modelling for malaria vector control. The existing capacity of the surveyed institutions were assessed and categorised into different skills groups. The results showed divergent training needs, but clearly indicate the need to start with introductory training themes. GIS, remote sensing and spatial statistical modelling courses are given at increasing level, focusing on a stable group of trainees following a series of courses going from introductory to advanced level.

EO products monitoring solutions

Based on the user requirements and the state-of-the-art of EO applications for malaria, the products to be developed are categorised into two fields of EO applications, i.e. EO applications for epidemiology and EO applications for supporting the MCPs - as explained below. There is a thematic overlap between these two applications, whereby some products will be used as input for the epidemiological modelling as well as for the direct support of the national malaria control programmes.

MALAREO developed 10 different map types ; household map, image base map, land cover maps, water bodies maps, distance-to-land cover maps, potential vector breeding site maps, habitat foci maps and population density maps. In addition, existing remote sensing derived products are combined with MALAREO products for malaria risk modelling using an Bayesian statistical modelling approach. Such additional existing data were collection-5 moderate resolution imaging spectroradiometer (MODIS), land surface temperature (LST) and emissivity data from the land processes distributed active archive centre (LP DAAC) of the National Aeronautics and Space Administration (NASA) and rainfall estimate (RFE) data with 8 km spatial resolution that was derived from the Famine Early Warning Systems network (FEWS NET) Africa data portal.

Household mapping

Detailed household maps derived from very high resolution (VHR) EO data are required to support NMCPs when preparing their control measures. An approach for an automated house/hut detection based on VHR data is developed and tested by the use of object-based image information and/or textural parameters of panchromatic and multispectral bands.

Beside the free image service Bing maps was tested for its use for house mapping. All houses in the VHR sites were manually digitised for a direct comparison to the results of the automated VHR data analysis.

Image base maps

End-users stated that even the high resolution (HR) base images would add value to the programmes. Since free image services such as Google Earth etc. do not provide high resolution imagery for the whole project area, true colour red green blue (RGB) RapidEye image mosaics were generated and rendered to geo-referenced jpg-files and provided to the NMCPs for being used in a GIS. Hardcopy base maps of the study area can assist the programme managers of the NMCP that are not familiar with GIS. Printed maps have therefore been delivered to the programme managers of the NMCP in Deutsches Institut fuer Normung (DIN) A0 format.

HR land cover mapping

The 25 000 km2 high resolution RapidEye data are classified according to malaria-relevant land covers using an object-oriented approach. Eleven land cover classes have been differentiated, namely flowing water, standing water, wetland, forest/woodland, bush-shrubland, grassland/savanna, bare soil/rock, settlement/infrastructure, roads/tracks, large-scale agriculture and subsistence farming.

Water body mapping and distance to water map

Even small water bodies are playing an important role as larval breeding sites for malaria vectors. The identification of water bodies are thus a direct indicator for malaria risk and the distance to water is a major determinant for Bayesian modelling of malaria incidence. Remote sensing was used to identify water bodies over large rural areas. The higher the resolution of the data, the more small water bodies can be detected.

Population distribution mapping

Depending on the needs of the Bayesian modelling of the malaria risk, modelling of population distribution can be done. A population density map links demographic data from statistical sources with land use information derived from EO data. A combination of the AfriPop dataset and the settlement/Infrastructure class produced an enhancement AfriPop population map. This map will also be very useful for the NMCP programme manager.

Mapping of potential breeding sites

The reference data set of vector breeding sites identified in the entomological survey from Swaziland and the HR water body map was combined to generate probability maps of the presence of breeding sites in larger areas. A statistical approach, whereby multi-temporal LST and RFE data were used in combination with 8 HR distance-to-land cover layers and a digital elevation model (DEM) in the decision tree software See5 (RuleQuest Research Pty. Ltd., NSW, Australia) was applied. The decision tree learning algorithm is a commercial decision tree and rule induction engine and classifies the data by the use of independent variables. The training data was generated from the entomological sample points as derived from field data through zonal statistics in ArcGIS 10.1. In total, following 18 environmental variables were used to analyse parameters that describe the existence of potential vector breeding sites:

1. altitude (DEM, 30m spatial resolution)
2. eight distance-to-land cover layers (5m spatial resolution)
3. four eight day averages of the LST for the four week period prior/during the entomological survey (1km spatial resolution)
4. a 24-day average of the LST for the three week period prior to the entomological survey (1km spatial resolution)
5. four decadal (10-day) RFE for the four week period prior/during the entomological survey (8km spatial resolution).

These five selected relevant features from this training data set, provided a decision tree or rule set for classification and gave the classification accuracy based on training data. In this case, the analysis was performed using all samples as training data. The land cover information on standing water and wetlands from the high resolution land cover map were afterwards used to apply the final ruleset from the See5 analysis, in order to classify the remotely sensed water bodies and wetlands in the whole malarious area of Swaziland according to their potential to be vector breeding sites.

Malaria risk modelling

It is the aim of MALAREO to improve Bayesian modelling results by using improved input data. Major improvements can be achieved by using medium to high resolution data for as well land cover classification and identification of water bodies, as for vegetation indices and elevation. This was achieved by the use of high resolution RapidEye data, medium resolution MODIS data and by the use of the advanced spaceborne thermal emission and reflection radiometre (ASTER) global DEM with 30m spatial resolution.


All these EO products are gathered into the MALAREO MapBook. This map book was distributed to all MALAREO end-users. Based on a variety of thematic maps, the management of integrated vector control, including the planning of indoor residual spraying (IRS), the distribution of insecticide treated nets (ITN) or larvaciding, can be substantially improved and can result in more effective vector control measures. All geo-data were provided to the NMCs to implement the data in their data management systems. The use of these data for the work of the NMCPs and the use of this map book was also part of the capacity building done in MALAREO.

Capacity building

One of the main objectives of the project was also to train the end-user community from countries affected by malaria in the use of GIS and the application of EO products. This will enhance local malaria combating techniques and abilities. Three separate courses were organised from basis to advanced level, spread across the duration of the project. They were incremental in nature, starting with fundamental interface training and theoretical knowledge in GIS and RS, moving onto applied skills in GIS and mapping epidemiology and finally onto analysis, modelling and map production using real world data.

The three training sessions were organised at the Howard College Campus of the University of KwaZulu-Natal in Durban, South-Africa. Participants came from different provincial MCPs in South-Africa, Swaziland, Mozambique and NMCPs from Madagascar, Botswana, Angola, Sudan and the African Centre of Meteorological Applications in Niger.

The ultimate goal was to enable the participants to utilise their own data collected in the field and enable them to produce spatial products in GIS which would be useful to track, monitor and combat malaria more effectively. The envisaged outcome was to train a core group to a moderate/high level of competency in GIS, remote sensing and spatial statistics, giving them sufficient knowledge and ability to train their colleagues afterwards.

At the end of the training sessions, all participants received a digital versatile disc (DVD), as well as hard copies with the training materials; the exercises and the lectures. All these presentations and documents are available on the MALAREO website.

Potential impact:

A final demo event with the end-users from the South African NMCP was held in South Africa in January 2013. This meeting gave a good impression on the actual and potential future impact of the MALAREO work. The developed MALAREO EO products have been presented. The end-users emphasised the benefit of these EO products for malaria control, since these products will greatly improve planning of malaria control measures and will complement the followed approach of linking environmental and epidemiological data, which is a first step towards an early warning system for malaria.

Three training sessions were organised throughout the second period of the project. The feedback on the organised courses was very positive. As the course was using open-source GIS software, it can be expected that the built-up capacity will result in immediate positive impact on the functioning and organisation of the MCPs.

MALAREO is in a good position to positively impact the use of EO and GIS for malaria vector control in the region and contribute to an operational use of spatial products at local level in the fight against malaria. However it is already clear that follow-up after the MALAREO project will be required; both in terms of continuation of the work in the project area as well as extension of the project approach to other areas.

MALAREO will hopefully contribute to a better integration and awareness of EO solutions in local and national MCPs and build on the fundaments of an EO monitoring cell that support the MCPs in their fight against malaria. Additionally, the project results should advance the state-of-the-art on malaria research and build progress towards an operational use of EO products and solutions supporting malaria vector control.