Topographic, lithospheric and lithologic controls on the transient landscape evolution after the opening of internally-drained basins. Modelling the North Iberian Neogene drainage

Abstract

The opening of internally-drained (endorheic) sedimentary basins often leads to a major drainage change, re-excavation of the basin sedimentary infill, and transient landscape. The timing of such basin openings can be dated only in exceptional cases in which the youngest sedimentary infill remains preserved. For this reason, the processes and timing involved in their transient landscape evolution are poorly known. We explore the role of erodibility, basin geometry and flexural isostasy during the capture of internally-drained basins by means of numerical modelling techniques constrained by recent terrace cosmogenic dating and geomorphological analysis, addressing the issue as to why the Duero and Ebro rivers, draining two Cenozoic sedimentary basins in N Iberia with similar geographical dimensions and drainage histories, have undergone a markedly different erosion evolution leading to distinctly different present morphology. To evaluate how these intrinsic parameters affect the transient landscape evolution, we design a synthetic scenario inspired by those basins. The results show that, once a basin becomes externally drained, its drainage integration and erosion rates are strongly dependent on 1) the basin elevation above the base level; 2) the width of the topographic barrier, 3) its erodibility; and 4) the rigidity of the lithosphere. The results show that transient landscape evolution can last for tens of millions of years even in absence of tectonic activity and changes in base level or climate. Basins isolated by wide and resistant barriers such as the Duero Basin may undergo a multi-million-year time lag between drainage opening and basin-wide incision. In the case of the Duero Basin, this delay may explain the paradoxical time lag between the last lacustrine sedimentation dated at 9.6 Ma and the onset of widespread basin incision variously estimated at 3.5 to 1 Ma.

Flooding of a carbonate platform: The Sian Kaʼan Wetlands, Yucatán, Mexico—A model for the formation and evolution of palustrine carbonate factories around the modern Caribbean Sea and in the depositional record

“…The Miocene succession comprises a series of fining upward sequences from alluvial clastics into calcretes and palustrine limestones (Figure 24C,D), interpreted as reflecting successive cycles of groundwater rise (Armenteros & Huerta, 2006; Huerta & Amenteros, 2005). The Miocene Lower and Upper Páramo limestones are separated by an intra‐Vallesian discontinuity (Krijgsman et al, 1996), with the upper unit representing the youngest deposits in the Duero Basin before the establishment of external drainage (Struth et al, 2021).…”

CHONK 1.0: landscape evolution framework: cellular automata meets graph theory

“…This method is close to what we aim to achieve but considers lake as unconditionally filled and outletting, thus not designed for endhoreic basins. TISC (formerly described in Garcia-Castellanos and Jiménez-Munt, 2015) is a pioneer in term of integrating endorheism to LEMs and recognising its impact on landscape evolution (Garcia-Castellanos et al, 2003;Garcia-Castellanos, 2006;Struth et al, 2021). TISC calculates the topography of the depressions and fill them gradually with the available sediment and water.…”

Autonomous generation of alluvial fans in landscape evolution models

“…In that study, our models cover a large geographical area and long intervals of geological time. The selected cell size and timesteps in this paper are those being used by the Great‐Basin modelling team, as well as those used in previous related work (e.g., Berry et al, 2019; Garcia‐Castellanos et al, 2002; Han & Wilson, 2022; Struth et al, 2021).…”

“…The numerical timestep for fluvial processes is assigned as 1 kyr, and the default tectonic and plotting time step is set to be 0:5 Myr (Table 1). The same model mesh (same number of cells) is used in additional sensitivity runs that vary the size of the model domain and as well as those used in previous related work (e.g., Berry et al, 2019;Garcia-Castellanos et al, 2002;Han & Wilson, 2022;Struth et al, 2021).…”