Climate Impacts on Vector-Borne Disease Transmission: Global and Site-Specific Analyses

The United Nation's Intergovernmental Panel on Climate Change estimates an unprecedented global rise of 2.0°C by the year 2100. Such change can affect serious infectious diseases, including dengue fever and malaria. Both large-scale iterative modeling and site-specific microclimatic analysis of disease ecology are needed in tandem to address health effects of climate change scenarios. In two separate studies of dengue and malaria transmission, both General Circulation Models (GCMs) of global climate change and site-specific climate analysis are used respectively to investigated climate change impacts on dengue fever and malaria transmission risk. For the first study, analysis was conducted using the integrated MIASMA model to link GCM projections of climate with a vectorial capacity model of transmission. Preliminary results indicate climate conditions being more suitable to dengue transmission, given viral introduction. An expansion of potential epidemic risk both geographically and temporally is inferred from this study. In the malaria study, preliminary results from regression analysis show mosquito biting rates to correlate to ambient temperature and rainfall. Parasite development was also shown to relate to temperature and humidity. Further interdisciplinary cooperation and multi-scaled analytical approaches will be required to better assess the potential effect of climate change on malaria and dengue. J Epidemiol, 1996 ; 6 : S145-S148.

While uncertainties always will accompany predictive climate modeling, there is increasing agreement between climate projections arising from varying methodologies in multiple climate centers internationally. The medical community is beginning to examine the consequences that these projections may portend for public health, and the World Health Organization considers global warming as a serious public health challenge for the future 3).
Infectious agents which cycle through cold-blooded insect vectors to complete their development are quite susceptible to subtle climate variations. In temperate regions, climate change would affect vector-borne diseases by altering the vector's range, reproductive and biting rates, as well as pathogen development rate within the vector host4).

Site-Specific Regression Analysis of Malaria and Climate
Weekly man-biting rate entomological data near Kisumu, Kenya collected from 1985 to 1988, were analyzed against meteorologic data over that time period. Variables recorded at Kisumu included: hourly temperature, dew point, wind speed, wind direction, cloud cover, and daily precipitation. From these values, weekly values (maximums, minimums, and averages) were calculated and compared to the weekly mosquito data.
Regression analysis was conducted to determine differences between the primary Anopheline vectors in this region, A. gambiae and A. funestus. Differences in ecological niches are known for these species with regard to breeding sites. To study the effect of rainfall on the man biting rate, MBR was plotted versus the amount of precipitation for these two different species.

Dengue: Global Modeling
Preliminary results show that dengue epidemic potential increases with a relatively small temperature rise, indicating fewer mosquitoes are necessary to maintain or spread dengue in a vulnerable population. The length of seasonal transmission increased in five major cities that differ both in climate and extent of dengue transmission. Expected transmission from baseline climate data matched observed dengue incidence at these sites, giving validity to our model results. Globally, all three climate GCMs showed an increase of epidemic potential, particularly in temperate regions.

Malaria: Species-Specific Analysis
The two species react very diffferently to the change in the precipitation. The slope for A. gambiae is much greater than the slope for A. funestus. The R2) for A. funestus is fairly small indicating that their is very little relationship between precipitation and biting rates of this mosquito. The relationship is stronger for A. gambiae. The key point is that difference in the slopes is significant. Thus emphasizing the fact that each species would react differently to the same climatic (in this case, precipitation) change.

DISCUSSION
To address the health effects of global climate change, studies must both accommodate the global nature of climate projections, and address the biological and ecological influences of climate on diseases at the local level. Also, complex integrated models are only as good as the basic science used to build such models. Root and Schneider have described a research paradigm called "Strategic Cyclical Scaling 17) ,"whereby iterative analysis involves the combined knowledge gained from alternating both large-and small-scale studies. The two studies described above help illustrate this concept. While they unfortunately do not address the same infectious disease, previous studies can be applied to these analyses to highlight key information gaps regarding effective health risk assessment of climate change.
For example, an estimated one million additional fatalities per year could be attributed to climate change by the middle of the next century, according to the global malaria model developed by Martens and colleagues 18).While this study represented a major step forward in climate change/health assessment, from the interspecies dynamics noted in the above Kenyan analysis, new information of mosquito response to precipitation can augment future climate change assessments; many Anophelene mosquito species transmit malaria, and if each species responds uniquely to varying precipitation, such information will be essential to improving validity of model projections.
What effect will global climate change have on dengue fever risk? To start with, an important climate-driven simulation model, DENSiM, has been developed by Focks and colleagues 15) that has been validated in the laboratory, five field studies in Bangkok and Honduras, a field study in New Orleans on larval habitat, and predicted ovipositional activity of Ae. aegypti in seven cities in the southeastern U.S15, 19) As an integrated, sitespecific model, predicted entomological indices closely match observed data. However, projections of climate change are at a resolution of scale approximating grid boxes over 200 kilometers diameter. Therefore, global scale or "top down" modeling, such as described above, can augment the site-specific DENSiM analysis of dengue transmission, since it is not feasible to obtain the numerous variables necessary to run the fullblown model on a global scale. CONCLUSION Scenario-based analysis of longterm future impacts must accompany conventional historical empirical studies 20),to adequately address the environmental health hazard posed by global climate change. Integrated mathematical modelling will be required for simulating complex global-wide climatological and ecological interactions. Accepting inevitable uncertainty, modeling can help in the analysis of these interactions. Yet, more emphasis needs to be placed on the iterative process of modeling, rather than on the numerical outputs which are generated. That is, in the process of model building, a deeper understanding may emerge as researchers begin to identify key gaps in data via heterogeneous, multi-scaled and multidisciplined investigation.