Sensitivity of Malaria Potential Transmission to Climatic Changes

The area of potential transmission of malaria is controlled by climatic factors such as temperature, humidity, and rainfall, which regulate the biology of development of both mosquito and parasite. Temperature affects the survival of the parasite only during its life-cycle in the Anopheles vector. All species have the shortest development cycle around 27-31C, from 8 to 15-21 days depending on species, the lower the temperature, the longer the cycle. Below 19C for P. falciparum, 15-16C for the other species, the parasites are unlikely to complete their cycle and hence to further propagate the disease. Temperature also modifies the vectorial capacity of the Anopheles. Optimal temperature values, ranging from 22 to 30C, lengthen the life-span of the mosquitoes and increase the frequency of blood meals taken by the females, to up to one meal every 48 hours. Higher temperatures also shorten the aquatic life cycle of the mosquitoes from 20 to 7 days and reduce the time between emergence and oviposition, as well as the time between successive ovipositions.

Rainfall generally means new breeding places. However rainfall can also destroy existing breeding places: heavy rains can change breeding pools into streams, impede the development of mosquito eggs or larvae, or simply flush the eggs or larvae out of the pools. Conversely exceptional drought conditions can turn streams into pools. The appearance of such opportunistic mosquito breeding sites sometimes precede epidemics.

The interaction between rainfall, evaporation, runoff, and temperature modulates the ambient air humidity which in turn affects the survival and activity of Anopheles mosquitoes. To survive, they need at least 50 or 60 % relative humidity. Higher values lengthen the life-span of the mosquitoes and enable them to infect more people.

Within the limits set by climate, decisive factors are responsible for the distribution of disease and its level of transmission. Among the main factors are the local biogeography, the vector behaviour and distribution, and the distribution and activities of humans.

In perennial zones of malaria, favourable climatic conditions exist all year for both vector and parasite to complete their life-cycle. There are no significant interannual fluctuations, some seasonal fluctuations may exist. The infection rate is high and can be maintained by small populations of mosquitoes, making malaria virtually impossible to eradicate in perennial zones. The indigenous populations build up some form of immunity which protects them from high mortality levels. In perennial zones, most deaths resulting from malaria are infants and children under five, because they have not yet built up their immunity to the disease, and pregnant women.

Malaria is seasonal in places where climatic conditions allow the periodic or occasional development of parasites and vectors. The indigenous populations do not have enough time to develop a proper immunity. Risk groups are less defined and the mortality rate during epidemics can be very high. Usually, epidemics are linked to the absence, decrease or loss of the collective immunity. Hence, eradication campaigns, biogeographical modifications such as deforestation, unusual weather events such as excessive rainfalls or droughts, or the settlement of non-immune populations in malaria zones might be followed by epidemics.

Despite international eradication and control campaigns, the malaria situation has not improved since the late 60s. This can be partly explained by the adaptation of the vectors to new environmental conditions and to the growing resistance of the Anopheles mosquito and of the Plasmodium parasite to pesticides and to anti-malaria drugs, respectively. Human factors like political instability, migration, and socio-economical development also greatly impede the control of the disease.

Climatic changes of the past have greatly affected the distribution of malaria. A global climate change is therefore likely to modify malaria geography again. However making predictions regarding the geographical extent and intensity of malaria is difficult: the relationship between malaria and climate is complex.

A simple, physically and biologically-based, model, the Malaria Potential Occurrence Zone (MOZ) model, has been designed as a first attempt to investigate the sensitivity of the potential transmission of malaria to climatic perturbations, in the absence of human interference, such as those expected from an enhanced greenhouse effect. MOZ is not an epidemiological model. It only focuses on environmental determinants of malaria and on how climate affects the development, growth, and reproduction of malaria parasites and vectors. MOZ operates globally with grid steps of 0.5 in latitude and longitude.

Malaria potential transmission is here defined as occurring when environmental conditions are favourable, at the same time and place, to both the malaria parasites and vectors. Where potential transmission occurs, actual transmission depends on the physical presence of malaria parasites and vectors, as well as on epidemiological conditions. For malaria to be potentially transmitted, environmental conditions should be such that parasites can complete each phase of their life cycle and vector can transmit the disease. Minimum and maximum air temperatures are used to estimate the potential distribution of parasites. From the above the minimum temperature is set to 15 to 19C depending upon
species, the maximum tolerable temperature for the four Plasmodia is assumed to be 32C.

Mosquitoes are present only when breeding sites are available. Therefore, the potential distribution of vectors is determined on the basis of a critical value of the moisture index, which is the ratio of precipitation to potential evapotranspiration. At regional scale, savannas are among the driest mosquito habitats. They occur for values of the moisture index between 0.6 and 0.8. A mean value of 0.7 has been chosen for the critical moisture index.

For the control climate, monthly temperature and precipitation present-day climatological data were used. Data for climate change conditions are based on scenarios produced for a doubling of the present atmospheric CO2 concentration by five atmospheric general circulation models (GCMs), which are documented in the reports of the IPCC.

White : no malaria Light grey : malaria, with or without climate change Dark grey : malaria appears in all five scenarios Black : malaria appears in at least one scenario

The map produced by MOZ for the present climate has been compared to a pre-eradication map, and to a World Health Organisation (WHO) map of the status of malaria for 1990, both by visual inspection and statistical analysis. The MOZ and 19th-century maps correspond well. Differences with the WHO survey may be attributed to the quality of meteorological data, weaknesses in model formulation, epidemiology, eradication and control efforts, and inaccuracies in the reference maps. One may conlude that MOZ provides reliable and useful first-order estimates of malaria potential transmission and its response to climate.

The simulations performed using the five climate change scenarios show significant differences with the simulation done with the present climatology: areas where malaria can potentially be transmitted increase by 7 to 28 % globally. Geographically malaria could spread to temperate countries of the Northern Hemisphere. The increase of the seasonal zones range from 16 to 55 %, while perennial zones might decrease. Seasonal expansion into areas formerly free of malaria is most likely to foster epidemics, causing increased mortality and high morbidity among unprepared or non-immune populations.

A caveat is that the potential transmission of malaria strongly depends on the climate change scenario. Regional differences in malaria potential transmission are attributable to intrinsic GCM known biases, and to model limitations. To go beyond first-order estimates MOZ should be improved with respect to physiology, biogeography, evaporation calculation, and epidemiology. Forth-coming versions of MOZ could indeed consider the time to maturity of parasites and the longevity of mosquitoes. The effects on the parasites and vectors of other environmental variables such as humidity could be taken into account. Dormancy of parasites in their hosts as well as aestivation and hibernation of mosquitoes could also be included. A more precise estimation of malaria transmission could incorporate an explicit treatment of the biogeographical distribution of individual malaria parasites and vectors. Potential evaporation is overestimated by the present calculation. Finally, MOZ could be coupled to epidemiological models to incorporate the human dimension of malaria.

Numerical experiments with various climate change scenarios consistently indicate vast changes in both the intensity and extent of malaria potential transmission. Increases in the areas where malaria could be potentially transmitted range from 4,879,000 km2 to 8,223,000 km2 (an area larger than that of the United States of America including Alaska). Most of the new potential transmission zones are located in the Northern Hemisphere: Europe, Siberia, Asia, and North America.

Climate change could force 150 million people to abandon their homeland. Such a massive influx of " environmental refugees " would likely overwhelm existing public health systems. Today, malaria is a developing country issue. With climate change it could become a public health problem for developed countries as well in a few decades.

Full paper published in Ambio, Vol. 24, N4, June 1995.