Climate change and the distribution and intensity of infectious diseases.

Many infectious diseases of humans, including malaria, dengue, cholera, and schistosomiasis, are restricted to, or more prevalent in, tropical and subtropical zones. Within the tropics and subtropics, they are more prevalent at lower than at higher altitudes. Warmer temperatures characteristic of lower latitudes and altitudes generally increase rates of survival, development, and replication of parasites and of blood-feeding vectors such as mosquitoes. Warmer conditions also increase activity (including biting) rates of vectors, resulting in higher rates of parasite transmission (reviewed by Harvell et al. 2002). A series of papers in the 1990s (e.g., Shope 1991, Martens et al. 1995, Colwell 1996, McMichael et al. 1996, Patz et al. 1996) contended that recent and future trends in climate warming were likely to increase the incidence and geographic distribution of infectious diseases, particularly those caused by vector-borne and water-borne parasites and pathogens. Owing to media attention and popular concern, the spread of infectious diseases was featured in the growing list of negative outcomes known or anticipated to arise from anthropogenic climate change (e.g., Intergovernmental Panel on Climate Change 2001; the 2006 documentary film, An Inconvenient Truth). However, unequivocal demonstrations of a causal link between climate change and human infectious diseases are rare (albeit increasing). Some diseases are likely to decrease in incidence and range with climate warming (Harvell et al. 2002), and others are likely to respond to precipitation or humidity more than to temperature, leading to poor predictive power under warming scenarios. Many diseases are strongly influenced by other ecological, sociological, economic, and evolutionary factors besides climate change. These latter observations have stimulated the emergence of critics of a climate-change– infectious-disease linkage. Lafferty (2009) provides an overview of recent criticisms, emphasizing three major categories: (1) In many cases, we should expect diseases to shift geographically without net expansion under climate change; (2) non-climatic factors are more important than climate; and (3) models that predict increasing disease transmission with climate warming are flawed if transmission rates fail to exceed a specific threshold (R0 . 1) that allows disease persistence. Here, I discuss these criticisms in turn. Diseases might shift without expanding.—The geographic range of diseases might fail to expand under climate warming scenarios due to complex relationships between temperature and vital rates of parasites or vectors. Given an optimal temperature range for parasite or vector fitness, temperature increases beyond the optimum will reduce disease transmission. The critical issue for the climate-change–disease debate is whether local temperatures will exceed the optimal range for specific parasites or vectors. The Intergovernmental Panel on Climate Change (2007) Synthesis Report synthesizes the evidence that local maximum temperatures will increase only modestly, while minimum temperatures will increase dramatically under climate change scenarios. The stability of temperature maxima compared to minima extends to both daily and seasonal variation. Consequently, when temperature optima exist, we can reasonably expect that climate warming will affect parasite and vector performance asymmetrically, with temperatures more likely to move from below optimum to optimum, as compared to moving from optimum to above optimum. Therefore, experiments in which temperatures raised above the range of natural variation cause declines in parasite performance (e.g., Dawson et al. 2005) do not necessarily mean that real climate warming will reduce parasitism. Of course, these expectations should be addressed empirically. An influential study by Rogers and Randolph (2000) created a statistical model to predict future changes in the distribution of falciparum malaria under climate warming scenarios. Rogers and Randolph mapped the recorded, present-day geographic distribution of falciparum malaria cases. They then ascertained the boundary conditions, based on the mean, maximum, and minimum of temperature, precipitation, and saturation vapor pressure, within which falciparum malaria is currently reported, and outside of which it is not. Using general circulation model scenarios of climate change, they projected where those boundary conditions are likely to occur in 2050. One model, the HadCM2 ‘‘medium–high’’ scenario, predicted 23 million additional people living in malarious areas, whereas another, the HadCM2 ‘‘high’’ scenario predicted 25 million fewer people living in malarious areas (Rogers and Randolph 2000). In essence, they predicted that falciparum malaria will shift without expanding. However, these models are Manuscript received 8 April 2008; revised 22 May 2008; accepted 29 May 2008. Corresponding Editor: K. Wilson. For reprints of this Forum, see footnote 1, p. 901. 1 E-mail: [email protected]