Messinian Salinity Crisis
Isolation of the Mediterranean by competing tectonic uplift and erosion in the Gibraltar Arc
[related publication: Garcia-Castellanos & Villaseñor, 2011, Nature]
[see also Urgeles et al., 2010, Basin Res.]
[related poster at EGU Meeting in Vienna, 2011]
We have constructed a simple mechanistic model for the closure of the last corridor that connected the Atlantic and the Mediterranean before the Messinian Salinity Crisis (MSC). The depth of the corridor is not prescribed as in previous studies, but it results from the competition between a prescribed uplift rate and an erosion rate calculated from the water flow and head loss along the corridor.
The results obtained for previous estimations of the erosional and uplift parameters show that these two processes may have reached a dynamic equilibrium along the last corridor crossing the Betic-Rif orogen, explaining the long initial connection between both seas during the early stage of the MSC. This long connection is needed to explain the large amount of salt deposited at the bottom of the Mediterranean. The uplift rate required for this competition is comparable to the one predicted by geodynamic models of slab detachment, a model that is supported by seismic tomography imaging the mantle underneath the Betics.
For the calculations we wrote a code in C under Linux. The program is called asalted, is open source, and is available for download here.
Our geodynamic interpretation is based on seismic tomography from an updated earthquake database. This tomography confirms the existence of a slab attached to the Iberian lithosphere in the western Betics and the Gulf of Cádiz, but detached from Iberia at the eastern Betics. This suggests a slab tear propagation towards the west underneath the Betics that may explain the uplift of the Betic corridors as well as the subsidence leading to the Zanclean flood.
Code developed for this research.
This C program by Daniel Garcia-Castellanos is open source and licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License.
A simplified version (spreadsheet calculator, for teaching purposes) of outburst flood evolution is available here.
asalted_Mediterranean_closure_spreadsheet.xlsxPost-Messinian flood of the Mediterranean
[related scientific publications: Micallef et al., 2018, Sci. Reports]
[related scientific publications: Garcia-Castellanos et al., 2009, Nature]
[related poster at AGU Fall Meeting in SF, 2009]
[related outreach publications: ScienceNews; Science; La Recherche]
C code developed for this research available on GitHub.
The program spillover.c by Daniel Garcia-Castellanos is licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License.
The flood that may have put an end to the Messinian Salinity Crisis in the Mediterranean is the largest documented in Earth's history, and still its nature has remained poorly constrained. The record in the Mediterranean sediments indicates an abrupt change from Late Messinian (evaporitic) to normal (open marine) environments at those times (5.33 Myr ago). But in geoscience "abrupt" could mean tens of thousands of years. To address this question, we used geophysical observations in the Alboran Sea and the Gulf of Cadiz and computer modeling techniques to investigate the feedback between water flow and erosion during the Zanclean flood.
Micallef et al., 2018, Sci.Reports.pdf
The model incorporates feedback between incision and water flow using finite difference techniques and an explicit solving scheme at regular time steps. We have developed a simple formulation that allows calculating the evolution of floods produced by overspill of a water basin into another, incorporating feedback between sill incision and water flow velocity. We used this formulation to model the geometry of an erosion channel crossing the Gibraltar Strait from west to east. The 300-650 m-deep erosion seen in borehole and seismic data (previously interpreted as the result of river erosion during the Messinian desiccation) are found to be consistent with erosion parameters obtained from independent river-incision studies.
According to the model predictions, feedback between incision and water flow implied that the flood necessarily ended as a catastrophic event reaching peak discharges of about 1000 times the present Amazon River, producing incision rates above 0.4 m/day and a sea level rise in the Mediterranean faster than 10 m/day. Although the whole flooding process may have lasted up to several thousand years, 90% of the water was transferred in a short period ranging between months and a few years. The figure shows an example run (this one adopting the stream unit power approach): In the initial stage of our model, water starts seeping through an arbitrarily small water gate located in a bathymetric sill separating the Atlantic and the dry Mediterranean basins. As basal shear-stress incises the sill, water flow increases and so does the incision rate, in a feedback that leads to exponential increase in discharge during the early model stages. All models show a long first phase with very little incision due to the reduced amount of water flow allowed by the initial sill depth of 1 m. As the Gibraltar gate grows deeper and wider, water flow and incision rate increase exponentially. This situation persists until the flow reduction due to the rising level of the Western Mediterranean becomes more important than the growth of the water gate. This event is labeled as Stage 1. Later, the effective slope S is progressively reduced and so does the flow velocity V, the water discharge Q and the erosion rate dzs/dt. As the Sicily Sill is reached (Stage 2), the level of the western Mediterranean remains constant while all water crossing the Gibraltar gate is transferred to the eastern basin. After the eastern basin also fills up to the Sicily sill (Stage 3), the whole Mediterranean will rise synchronously. Headloss across Gibraltar reduces gradually to zero, towards an asymptotic equilibrium (Stage 4) where there is no significant level difference between both oceans.
Water flow erosion
In our model, water flow causes rock incision, following either a shear stress or a stream unit power formulation, both tested in mountain rivers (e.g. Whipple & Tucker, 1999; Lavé & Avouac, 2001; Wobus et al., 2005; Attal et al., 2008). The shear stress approach is written as
where kb and a are positive constants. An analytical solution of this equation coupled to slope-driven water flow shows that the sill is incised exponentially along time in the early stages of flooding. The speed of this growth is dependent on the lithological erodability kb and the slope S in the Mediterranean side of the sill.
Slope- and head-loss-controlled water discharge
For the velocity of water flow, we adopt the Manning's formulation:
where V is the average velocity (m/s), n=0.05 is the roughness coefficient, and Rh is the hydraulic radius (m) of the strait connecting the Atlantic and the Mediterranean.
Source code used for the modeling
My public-domain code spillover is developed for Linux platforms in C language.