We report that human walk patterns closely follow Levy walk patterns commonly observed in animals such as monkeys, birds and jackals. Our study is based on about one thousand hours of GPS traces involving 44 volunteers in various outdoor settings including two different college campuses, a metropolitan area, a theme park and a state fair. Important implications of this finding include that many statistical features of human walks are scale-invariant and bursty, and do not conform to the central limit theorem. None of commonly used mobility models for mobile networks captures these properties. Levy walks are more diffusive than Brownian motion (BM) while less diffusive than random way point (RWP). Based on these findings, we construct a simple Levy walk mobility model that emulates human walk patterns expected in outdoor mobile network environments. We demonstrate that the Levy walk model can be used to recreate the statistical patterns commonly observed in previous mobility studies such as the power-law distributions of human inter-contact times and that the simulation performance of mobile network routing protocols under the Levy walk model exhibits distinctive performance features unexplored under existing mobility models.