Stratified ice flow explains the lack of deformation on Mars' North Polar Layered Deposit (NPLD)

Introduction: 

The North Polar Layered Deposit (NPDL) is a massive ice deposit located on Mars' north polar region [1,2,3,4]. Radar observations from the SHAllow RADar (SHARAD), reveal a layered internal structure consisting of stratified ice (>95% ) and dust [3], extending to a thickness of 2-3 km. Because of the connection between north polar ice stability and obliquity-modulated climate cycles, the NPLD is a true time capsule of Mars' climate during the NPLD timespan of 4 Myr [5,3,4].

However, the lack of evidence for NPLD flow poses a problem: an ice sheet of this size should actively deform. Previous studies predict ice deformation rates that should be observable [1,6], in stark contrast with the lack of evidence for surface deformation (HiRISE, ~cm/yr). A similar lack of evidence for flow is found by SHARAD, which should record substantial deformation accrued during the NPLD 4  Myr lifetime [7,4].

While numerous studies have investigated NPLD dynamics to understand its flow history and in response to climate changes [7,5,6,4], the question of the lack of observable flow in the NPLD remains unresolved [4]. Here we hypothesize that the lack of accrued NPLD deformation is caused by the rheology of its stiffer layers. We develop and implement a stratified ice flow model coupling vertically variable rheologies derived from ice-dust mixture experiments [10,11], and realistic NPLD stratigraphies from SHARAD. Our results solve the quantitative conundrum by showing that dusty ice layers play a key role preventing NPLD long-term deformation. [4].

Methods:   

We develop a 1D stratified ice flow model with realistic ice and dust mixture rheology, including the experimental observation that dust-ice mixtures become jammed after a threshold dust content (f) of 67%, hence producing rigid layers [10,11]. We model realistic NPLD stratigraphic profiles from SHARAD radargrams [8], and model dust-depth distributions from observations and mass balance modeling [13], deriving layer compositions ranging from pure ice to pure dust.

We address the uncertainty in the parameters and the robustness of the model results with a Monte Carlo approach with over 150 scenarios including `cold' (160 K) and `warm' (172 K) NPLD runs, to retrieve the vertical velocity profile, the maximum and bulk velocities, and the surface deformation rates. We vary the dust distribution with a Bayesian approach, using a Poisson distribution centered around 5% dust, with content increasing with depth.

Model setup. We numerically integrate the Stokes equations for viscous flow at all deformable layers, with a dust-dependent rheology. Velocity is considered continuous across all deformable layers (f < 0.67), whereas a zero velocity condition is applied at the bottom and top of rigid layers with f > 0.67. At the base, the NPLD is assumed to be cold-based, whereas the surface is a free boundary.

Results:

Our results highlight two aspects: (1) Thick, pure ice layers located near the base host the fastest flowing strata of the NPLD. This is a fundamental difference with terrestrial ice sheets, and any type of deformable dome without internal cohesive layers, where the maximum velocity is located at the surface and the maximum deformation occurs at the base. Here the base still hosts the largest strain rates, but cohesive dust-rich layers include equally important deceleration, balancing out the basal velocity gain. Therefore, observations searching for deformation in the NPLD should focus on internal, deep, thick pure ice layers and not on surface motion.

Previous experimental work on martian as well as terrestrial ice-dust mixtures [10,11,12] shows that this flow transition occurs at or near f ~ 0.67. However, the temperature range of these studies is significantly larger from the range of application we consider in this paper, and while the terrestrial studies show that this transition occurs at yet even warmer temperatures, neither these works nor other studies we know of have explored the ductile to brittle transition at temperature and pressure relevant to the NPLD. The details of the rheology of dust-ice mixtures under uniaxial compression, shear, and temperatures relevant to the NPLD is therefore a knowledge gap with the potential to alter our conclusions of NPLD flow dynamics, but for which we cannot mitigate.

Conclusions:

The lack of visible flow on the North Polar Layered Deposits (NPLD), both in observations of the ice margin and ice surface and considering the lack of deformation in the interior layering, via SHARAD radar observations, have long challenged the results of ice flow models that aimed to understand this unique behavior.  This study considers the internal stratified nature of the North Polar Layered Deposits (NPLD) in developing a novel model for layered ice flow, considering the existence of ductile layers with varying dust content as well as the presence of undeformable 'dust-rich' layers with dust fractions over 67%.

 

 Our results yield consistent ice flow rates of order cm/Myr (figure 2) for most scenarios at a temperature of T = 162 K. We find that the main controls on NPLD deformation rates are the existence and distribution of dust-rich layers, which slow down the bulk flow and introduce locally static layers. The concentration of such layers near the base of the NPLD, where most glacial deformation should occur, dramatically slows down NPLD flow and prevents the surface to deform observably. Our results reconcile for the first time observations and glacial flow modeling, and highlight the importance of considering ice-dust stratification in future martian ice flow models, including beyond the polar regions.

Acknowledgments:

This project has received funding from the European Union H2020 program under MSCA grant agreement MGFR – 101027900, and from the RPL-PULSAR program GELMARS to A.G.G.