Chaotic chasms: Canyon evolution governed by autogenic channel-hillslope feedbacks

Abstract:

The causes and timing of river canyon formation are frequently debated, as is the extent to which canyon landscapes record climatic and tectonic signals. Current models of both the channel and adjacent hillslopes do not fully capture the mechanics of canyon evolution. In particular, they neglect the two-way feedbacks associated with delivery of large blocks of rock from hillslopes to channels. Here we develop the first model of canyon evolution that includes the role of blocks in both hillslope and fluvial processes. Blocks on the hillslopes reduce soil fluxes and bedrock weathering rates, whereas blocks in the channel inhibit incision through the effects of bed cover and hydraulic drag. Blocks weather on the hillslopes, and continue to decline in size in the channel. We numerically simulate canyon formation in a 2-D landscape containing a block-releasing caprock subjected to a steady baselevel lowering rate.

While canyon walls are near the channel, hillslopes are short and blocks released from canyon walls are still large upon reaching the channel. This lowers the rate of drop of the boundary condition experienced by the adjacent hillslopes and reduces the block delivery rate, which then leads to a period of temporarily rapid river incision.

We find that such feedbacks cause 1000-year average channel incision rates to oscillate between 0 and 5 times the imposed baselevel lowering rate, decoupling both the channel and the adjacent hillslopes from baselevel on million-year timescales. This unsteadiness of fluvial erosion rate persists until the canyon walls retreat far enough that blocks of significant size no longer reach the channel. Our model captures both the planview and cross-section form of many river canyons observed on Earth. Canyon form and erosion rates differ substantially between model runs incorporating block feedbacks and control runs without blocks, indicating that delivery of large blocks from canyon walls to channels exerts a fundamental control on both form and pace of canyon evolution. Our results demonstrate that channel-hillslope block delivery feedbacks during canyon formation strongly alter the propagation of baselevel signals through river canyons, which complicates the interpretation of canyon morphology and erosion rates.

Plain language summary:

River canyons are some of Earth’s most spectacular landforms. To understand the timing and causes of canyon formation, we need realistic models for how rivers and canyon walls erode through time. We present a computer model for canyon formation that incorporates the delivery of large rocks from the canyon walls to the river. These large rocks not only change how quickly the channel erodes, but also affect the shape of the canyon walls. Early in canyon formation when the walls are near the channel, large rocks cause significant variation in time in the efficiency of river erosion. Once the canyon walls are far from the channel, large rocks from the walls have enough time to degrade en route to the channel that they no longer influence river erosion. Canyons in our model are distinct from those produced by simpler models in both their shape and long-term erosion rates. Our results show that river canyons may not be reliable indicators of environmental history because their shape at any point in time is strongly controlled by the history of delivery of large rocks from the adjacent cliffs.