Direct evidence of drainage divide migration reveals intermittent dynamics linked to 100 kyr climate oscillations

It is common to assume that when there are erosion rates and slope gradients across a drainage divide, the divide is prone to migrate and change the drainage area distribution of its bounding catchments. However, direct records of divide migration are exceptionally rare. This raises the following questions: Could the assumption be wrong, and can divides sustain topographic and erosion rate asymmetry over geomorphic and geologic (10⁴– 10⁶ yrs) timescales? And when divides eventually migrate, is the migration driven by endogenic feedback within the basin, or by exogenic forcing, such as climate change?

To address these issues, we study a field area along the escarpments of the Dead Sea plate boundary, Israel, where direct records for divide migration are present in the form of terraces that grade opposite to the channel flow direction. These terraces are interpreted as a record of the divide’s paleo-locations, such that terraces are formed when the divide migrates inland from the edge of an escarpment, inducing drainage reversal and gradually extending the reversed channel that drains toward the escarpment.

Absolute dating of these terraces using luminesces techniques and relative dating using soil chronosequence markers reveal that the terraces become older from the present locations of the divide toward the escarpment, consistent with the interpreted process of their formation. Terrace ages show an average divide migration rate of ~1100 m myr^-1 over the past ~230 kyr, supporting active divide migration over timescales of 10⁵ yrs or shorter.

Terrace groups with similar ages indicate ~100 kyr cycles of periods of divide stalling and episodes of rapid divide migration with rates up to fourfold relative to the average rate. We use numerical simulations to explore possible drivers for the inferred divide intermittent dynamics. Simulations show that the dynamics are inconsistent with landscape evolution under uniform environmental conditions due only to internal basin dynamics. Instead, the inferred intermittency is best explained with time-dependent erosional efficiency that is sensitive to global climate change and correlates with regional paleoclimate proxies.

This study provides the first detection of divide migration rate intermittencies at timescales of 10⁴-10⁵ yrs, and the association between divide dynamics and changing climatic conditions. This highlights the potentially significant impact of climate changes on the plan-form evolution of drainage basins.