Malaria is one of the greatest health threats in tropical regions like the Amazon, but predicting its spread is difficult. While it's no secret that mosquitoes prefer warm air and standing pools of water, how do you translate that awareness to a large scale? By getting a little help from orbit, apparently. Researchers are using data from NASA satellites (such as the Landsat series) to predict malaria outbreaks by identifying areas where the soil moisture creates prime breeding grounds for the mosquitoes that transmit malaria. They know that floods and deforestation tend to create mosquito-friendly pools of water -- compare that with fine-grained models of human behavior (say, loggers or miners who work in wet conditions) and you have a unified system that can anticipate outbreaks about 3 months in advance, right down to individual households.