The factors that shape vertical root distributions in different soils and under different climates and vegetation are poorly understood. This makes it difficult to parameterize root profiles in vegetation-, hydrology, biogeochemistry-, or global circulation models. Recently, it has been proposed that vertical root distributions in the vadose zone could be predicted from soil water infiltration and extraction patterns as a function of climatic variability, soil, and vegetation characteristics. A number of ecological factors favor shallow over deep roots, suggesting that root profiles of plant communities may tend to be as shallow as possible and as deep as needed to fulfill evapotranspirational demands. To test this hypothesis, a stochastic, one-dimensional soil water infiltration and extraction model (SWIEM) was developed that simulates soil water infiltration through 600 discrete soil layers to a depth of 6 m. Water input is simulated in Monte Carlo fashion based on site-specific long-term precipitation data. Water extraction proceeds from the top down, with extraction depths determined by potential evapotranspiration (PET) and the vertical distribution of soil water. The resulting shallowest possible water extraction profile was tested against nine measured root profiles from long-term ecological research sites in different biomes. Two other approaches, based on mean root distributions for biomes and an empirical regression model, were also compared to the observed root distributions. Soil water extraction patterns predicted by the SWIEM model matched observed vertical root distributions better than the other two approaches. These findings show that vertical root distributions in different biomes tend to approach the shallowest possible shape, thereby creating a useful null model for future research on root distributions and a promising tool for parameterization of global models.