The hypothesis that climate change may enhance the intensity, duration, and frequency of precipitation extremes needs to be carefully studied for future adaptation policies or flood disaster mitigation strategies. Several aspects of this hypothesis are examined here in the context of precipitation extremes in South America. The volatility of precipitation extremes is quantified from observations to obtain the natural variability and geospatial-temporal trends of the extremes. The geospatial-temporal extremes dependence structures estimated from observed data and climate model simulations are compared with each other to develop an understanding of the ability of the climate models to simulate precipitation extremes. The ability to utilize the statistical properties of the precipitation extremes, in conjunction with precise estimates of population counts, to obtain objective and high resolution risk maps at continental scales for disasters potentially caused by the precipitation extremes, is illustrated. The limitations of this study and the outstanding challenges are discussed. New opportunities for future research are suggested, with a particular emphasis on novel and emerging tools in high-performance computing and knowledge discovery from massive geographic data. The results presented here are based on prior and ongoing work at the Computational Sciences and Engineering Division of the Oak Ridge National Laboratory.