Gareth G Roberts 

Much understanding of continental topographic evolution is rooted in measuring and predicting rates at which rivers erode. Flume tank and field observations indicate that stochasticity and local conditions play important roles in determining rates at small scales (e.g. < 10 km, thousands of years). Obversely, preserved river profiles and common shapes of rivers atop uplifting topography indicate that erosion rates are predictable at larger scales. These observations indicate that the response of rivers to forcing can be scale dependent. Here I demonstrate that erosional thresholds can provide an explanation for why profile evolution can be very complicated and unique at small scales yet simple and predictable at large scales.