Rainfall, temperature, and humidity are the three climate variables that most influence malaria transmission. High rainfall tends to increase the number of mosquitoes, because more surface water breeding sites become available. It also leads to increased atmospheric humidity, which supports mosquito survival. Relative humidity generally needs to be above 60% for mosquito survival and malaria transmission. Temperature affects the development of both the mosquito and the Plasmodium parasite. Under warmer conditions mosquitoes develop faster, and the parasite also multiplies more rapidly. Malaria transmission is complex, being influenced by other factors besides climate, but when trends in non-climate factors were quantified and accounted for, rainfall variability was found to explain more than two-thirds of the variability in incidence. This means that rainfall moni¬toring can give several weeks advance warning of possible epidemics; but the same research went further, linking malaria incidence with sea surface temperatures, which affect continental rainfall and are used for seasonal forecasting. Subsequent analysis confirmed that seasonal forecasts can provide useful indications of the likelihood of an epidemic several months in advance (Thomson et al., 2006). African Heads of WHO, the United Nations Development Programme (UNDP), the United Nations Children’s Fund (UNICEF), and the World Bank launched the Roll Back Malaria Global Partnership, which now has more than 90 partners. Roll Back Malaria aims to identify stakeholders, consolidate research, and deliver concerted support to malaria control through State endorsed the Roll Back Malaria initia¬tive in the Abuja Declaration of 2000. The new Malaria Early Warning and Response System (MEWS) has been devel¬oped by the partners of the Roll Back Malaria initiative, including national ministries of health. It has five components. • Vulnerability assessment and monitoring. Knowledge of vulnerable populations and areas helps control services plan their response ahead of an epidemic. Factors that increase vulnerability include co-infection with other diseases, malnutrition, resistance to anti-malarial drugs, and population migration, whereby non- immune people move into endemic areas or ‘parasite carriers’ move into epidemic-prone areas. • Seasonal climate forecasting. Given the established link between climate and malaria incidence, reliable forecasting can help predict epidemics. Seasonal climate forecasts can give several months lead time, allowing effective control and other measures to be put in place. • Environmental monitoring. This also gives advance warning of possible epidemics, but with a shorter lead time of 1–3 months. Rainfall, temperature, and humidity are monitored, together with vegetation status and flooding. • Sentinel case surveillance. A surveillance system that quickly detects a rise in malaria cases, i.e. the beginning of an epidemic, is crucial. • Planning, preparedness, and response. The four components above allow planning and preparedness for epidemics, so that response activities can be implemented in the right place at the right time. The MEWS has been introduced over the past few years in southern African countries with epidemic-prone areas, includ¬ing Botswana, Madagascar, Mozambique, Namibia, South Africa, Swaziland, and Zimbabwe. Botswana has provided a particu¬larly useful testing ground (WHO, 2004). Straddling the southern margins of Africa’s malaria transmission zone, Botswana has extensive epidemic-prone areas.