ArticlesPotential effect of climate change on malaria transmission in Africa
Introduction
90% of malaria cases occur in Africa.1 In the past decade, the incidence of malaria has been escalating at an alarming rate. There is an increasing interest in the mapping and predictive modelling of the distribution, intensity, and seasonality of malaria transmission.2, 3, 4, 5, 6 Climate change is likely to have various effects on health, including changes in distribution and seasonal transmission of vector-borne diseases.7 The extent of these effects, however, continues to generate intense debate,8, 9 especially in the projected effect of climate change on the global distribution of malaria, in which different approaches have resulted in widely varying estimates. A general issue facing all researchers has, however, been the absence of comprehensive, good-quality empirical data to validate the models used. The link between climate and malaria distribution has long been established. Sustained transmission depends on favourable environmental conditions for both vector and parasite. The effect of temperature on the duration of the sporogonic cycle of the malaria parasite and vector survival10, 11 is particularly important.
Several methods have been used to estimate changes in the worldwide distribution of malaria in scenarios of global climate change. One approach relies on a biological model that predicts a large increase in global malaria potential.12, 13 Some have criticised biological models on the basis that crucial parameters and their relations with environmental factors have not yet been quantified.14 Thus, biological models have used only a limited number of covariates, and doubts have been raised about the qualitative validity of some results.15 An alternative approach, based on a statistical model derived from the current malaria distribution projects little change in distribution.14 The use of current malaria distribution to derive the model resulted in areas that are climatically suitable for transmission but in which malaria has been eradicated (eg, northern parts of Australia), skewing the results. Generic disadvantages of worldwide or continent-wide statistically-driven models are that data sets used to statistically develop the models are often of uncertain accuracy, models are not easily reproducible (ie, results vary with training data and methods used), and the results are often applicable only to national or subregional scales. We use a large set of parasite surveys done throughout Africa to produce a spatiotemporally validated model of malaria transmission and project the effect of three climate scenarios by Hadley Centre global climate model (HadCM3) climate experiments.
Section snippets
Data
We used mean long-term monthly rainfall and temperature data as the basis for the seasonality model.16 The gridded surfaces were based on weather station data from 1920 to 1980 and have a spatial resolution of 0·05°. The temperature data have SEs of 0·5°C and monthly mean precipitation data have errors of 10–30%.
The population data we used was an interpolated gridded surface17 of resolution 0·042° with 1995 population estimates. Data were interpolated with a spatial interaction model that
Results
Our seasonality model estimates that on average there are 3·1 billion person-months of exposure to malaria (445 million people exposed) in Africa every year (table 2). The spatial and temporal validation of the predicted current malaria distribution (figure 2) was undertaken with positive (n=3199) and negative (n=592) parasite surveys of 1 month duration. The model showed a sensitivity (ie, the ability of the model to accurately predict areas of transmission to within a month) of 63% (95% CI
Discussion
We have produced a spatiotemporally validated malaria transmission model for Africa and projected changes in transmission patterns in differing climate scenarios. Our findings have important implications for malaria control in Africa since both the duration and timing of malaria transmission season are important to inform efforts in malaria control. The duration of the season will affect the dynamics of transmission, with longer seasons allowing heightened transmission and high levels of
References (30)
- et al.
A climate-based distribution model of malaria transmission in sub-Saharan Africa
Parasitol Today
(1999) - et al.
Predicting malaria seasons in Kenya using multitemporal meteorological satellite sensor data
Trans R Soc Trop Med Hyg
(1998) - et al.
Models to predict the intensity of Plasmodium falciparum transmission: applications to the burden of disease in Kenya
Trans R Soc Trop Med Hyg
(1998) Global-warming and vector-borne disease in temperate regions and at high altitude
Lancet
(1998)- et al.
Effect of HIV-1 and increasing immunosuppression on malaria parasitaemia and clinical episodes in adults in rural Uganda: a cohort study
Lancet
(2000) The world health report 1999: making a difference
(1999)- et al.
A spatial statistical approach to malaria mapping
Int J Epidemiol
(2000) - et al.
Satellite imagery in the study and forecast of malaria
Nature
(2002) - et al.
Human Health
- et al.
Malaria in the African highlands: past, present and future
Bull World Health Organ
(1998)