Malaria in Kenya
    Gore error list                                                                  Cool It, British                        Cool It, American        

In his film, Al Gore shows a diagrammatic picture of mosquitoes swarming around at the slopes of a mountain range, first at relatively low altitudes, and later also at higher altitudes. The text is:
"There are cities that were founded because they were just above the mosquito line. Nairobi is one. Harare is another. There are plenty of others. Now the mosquitoes with warming are climbing to higher altitudes."

There seems to exist no official definition of what is meant by a "mosquito line". You might compare with the term "tree line", which is a line above which no trees can grow. By analogy, you would assume that a "mosquito line" is a line above which no mosquitoes can live. But this interpretation would not fit in relation to malaria in Kenya.
Instead, you would have to understand the term as a line above which malaria is not present every year, but only in certain years, when epidemics occur.

There are several malaria parasites, the most widespread of which are Plasmodium vivax and Plasmodium falciparum. The former occurs or has occurred also at rather high latitudes on the northern hemisphere, and is not the most dangerous one. The latter is restricted to the tropics, and is the most deadly species.

The number of days required for P. falciparum parasites to complete that part of its development which takes place inside the mosquito depends on temperature. The lower the temperature, the longer the duration of parasite development, and the lower the probability that the mosquito is still alive when the parasites are ready to infect humans. The relationship is as follows (source:this link):

Duration of development
Parasite´s chance of
survival in mosquito
16° C
No development
            0 %
18° C
56 days
            0.28 %
20° C
28 days
            5.9 %
22° C
19 days
           15 %

Below 18° C parasite survival is so low that a population of parasites cannot survive. At 22° C and above, survival is so high that a successful life cycle from mosquitoes to humans and back to mosquitoes will permanently function. In regions with such temperatures, there will be "stable transmission" of malaria, i.e. there is a permanent risk of infection. When temperatures range between 18° and 22° C, the risk of infection is variable.

In addition, rainfall is also important, determining the abundance of mosquitoes.

Accordingly, a country like Kenya may be divided into four different zones with respect to malaria: 1) zones with stable malaria. 2) zones with seasonal malaria, i.e. only at those times of the year  when conditions are favourable. 3) zones prone to malaria epidemics, i.e. areas where malaria is not present every year, but with outburst of epidemics in certain years when favourable weather conditions prevail for at least c. three months. 4) zones with no malaria (mainly zones where malaria never has entered, or has been eradicated).

Temperatures depend on altitude. Every 100 m increase in altitude corresponds to a 0.5° C decline in temperature. In Kenya, regular occurrence of malaria is roughly restricted to altitudes below 1600 m, but epidemics occur in some years at 1700-2300 m, and malaria has occasionally been recorded as high as 2400 m.

Nairobi is situated at 1660 m and thus may be said to be just above the line delimitating the zones of regular occurrence. The coldest month, July, has average maximum temperatures of 20.6° C, i.e. temperatures less than optimal for the P. falciparum parasite. Thus, it is true that Nairobi is just inside the zone where malaria occurs only as epidemics, not as a permanent or seasonal penomenon. If this is what is understood by "the mosquito line", Al Gore is approximately right.

The Nandi and Kericho districts in the western Kenyan highlands were originally free of malaria. For this reason, they were the preferred districts for European settlers in the years around 1900. A sanatorium for colonial officials was built in Nandi because malaria did not exist there. The construction of a road connecting Uganda with the sea led to increases in transport, trade and settlement. The first malaria epidemic occurred in 1918, when troops were being resettled after the First World War. Apparently, infected people brought the disease to the district. Over the following twenty years, malaria spread up onto the high plateaus.

Treatment of infected persons seems to have broken the infectious cycles - i.e. too few parasites survived in humans to efficiently infect the next generations of mosquitoes. In any case, epidemics became rarer, and without any special efforts against malaria, the Nandi and Kericho districts had very little malaria during the years 1950 to 1970. After 1970, malaria became gradually more frequent, and a series of severe epidemics occurred during the 1990s. One of the causes may have been the increasing human population - the more people, the greater the chance that an infected mosquito actually finds a person to bite. In addition, parasites that had become resistant to the anti-malarial drug chloroquine began to spread in the 1980s (Björkman & Bhattarai (2005): Acta tropica 94: 163-169).

The question is if changing climate may also have contributed to this pattern. For a country in the tropics, we would expect that average temperatures have increased by less than c. 0.5° C, and such a slight increase would be barely noticeable. It would correspond to temperature zones climbing 100 m upwards on the mountain slopes. There exists a series of weather data for a tea estate in the Kericho district, see this link, figure 4.12. The text accompanying the figure states that there are no obvious effects of "warming" in this area since 1967. This is not quite true, however. The maximum temperatures show a slightly rising trend (non-significant); the trend appears much more clearly if you count how many consecutive months each year have temperatures above 22° C. Only from 1987 onwards do we have years where the maximum temperatures are 22° or more the whole year round, that is, only from this year onwards could malaria occur permanently throughout the year. Thus, the slight rise in coherent periods with relatively high temperatures, especially during the 1990s, coincides rather precisely with the increasing permenanence of malaria in the district.

Also, others have analysed this temperature series. A study group led by S. I. Hay has analysed weather data from the region; they have found that although there is a significant upward trend in temperatures from 1970 onwards, this trend is what mathematicians call "a random walk" and thus does not represent a true upward trend (see this link). In another paper (see this link), the authors write: "In essence, climate and climate variability had not changed, so could not have caused the malaria resurgences". However, others dispute their conclusions (link), and a repeated statistical study of temperature changes in the region, by M. Pascual and others (link), finds that temperature trends in the region parallel global trends: a temperature decline from 1950 to 1970, and a rise of about 0.5° C from 1970 onwards. This trend is significant beyond what is explained by "red noise", approximately the same as "random walk". Thus, they directly reject the conclusion of Hay et al. Furthermore, they find that the temperature increase is within an interval where the effect on the malaria parasite could be disproportionately large.

Thus, several factors that might favour the malaria parasite have acted simultaneously: increasing human populations, increasing resistance to drugs, and increasing temperatures. It is not easy to tell what factor has been most important. It is quite possible that the factors have interacted: if climate change and population increase create better conditions for the malaria parasite, then the chance that the parasite will develop drug resistance is increased. Another study in a district in Uganda likewise shows a temperature increase and a concomitant increase in malaria in 1987 (Loevinsohn (1994): Lancet 343: 714-718). In this study, variations in malaria incidence could be explained very precisely by variations in temperature. And, as a contrast to the situation in the Kericho district in Kenya, there were no confounding factors: no road building, no increase in population, no local increase in drug resistance etc. This points to temperature as a decisive factor.

There is also a study of malaria on Mount Kenya (Chen et al 2006) which indicates that malaria and its mosquito host have spread there recently at rather high altitudes.

Furthermore, a relationship between temperature changes and malaria is demonstrated by a study performed in Madagascar, see this link. Here, the author found that when analysing temperatures of just those months in the year when the infectious cycle would normally start, these temperatures are found to explain a very large proportion of the variation in malaria incidence. In the author´s words: "This review suggests that the rejection of climate-disease associations in studies so far published may not have used biologically relevant climate parameters."

That a relationship between altitude and the rate of infection with the falciparum malaria parasite does exist, is well demonstrated in this paper: Drakeley et al. (2005): Journal of infectious diseases 191: 1589-1598. Here, the incidence is found to decrease by 20 % for every 100 m increase in altitude.

In conclusion, we see just what we would expect to see when trends are just emerging: an increasing prevalence of malaria, especially at higher altitudes, and a relationship between this and the variations in temperatures. But the climate changes are still so small that they are barely significant, and only very precise analyses, analysing exactly those months that are most important for the life cycle of the malaria parasite, give significant results. There are several other factors which very likely contribute to the increase in malaria incidence, and this further obfuscates the relationship with temperature. So, if you believe that climate change is underway, then you can say that the "signal" (the malaria-temperature relationship) is only just beginning to appear above the "noice" (other factors affecting malaria). If temperatures indeed continue to increase, the malaria "signal" should become stronger in future.