The effects of variations of drainage basin morphometry and relief characteristics on flood peak magnitude and time-to-peak are investigated using simulated stream networks. The networks are produced by three models: headward growth, systematic capture, and minimum power relaxation. Translational and kinematic wave flood routing were used to generate synthetic hydrographs. Peak discharge and time-to-peak are predictable to a high degree by five different sets of morphometric-relief parameters. In order of decreasing order of importance in predictive ability the parameters characterize basin size, relative relief, basin concavity, and basin shape. Both simulated and natural stream networks exhibit strong dependence of planimetric morphometry upon basin concavity. The effect of this dependency is to increase the effect of basin concavity upon flood hydrographs.