River-deposit dimensions versus stratigraphic elevation in Aeolis Dorsa: resolving the great drying of Mars

Edwin Kite, Antoine Lucas, John C. Armstrong, Oded Aharonson & Michael P. Lamb

Rationale: River-deposit dimensions constrain hydrology and climate on Early Mars, but stratigraphy is essential to build a time series of constraints on climate change Today, use measurements of Early Mars river-deposit dimensions versus stratigraphic elevation to:1)Characterize river-forming episodes2)Constrain river discharge versus time.

wavelength

width

River deposits record constraintssorely needed for Early Mars climate models

Environmental scenarios for precipitation-fed runoff on Early Mars vary widely: e.g. Haberle et al. 2012, Kite et al. Icarus 2013, Mischna et al. 2013, Segura et al. 2012, Urata & Toon 2013, Wordsworth et al. 2013

Andrews-Hanna & Lewis 2011 …

Early Mars rivers constrain magnitude, duration, intermittency, and number of wet events: e.g. Burr et al. 2010, Palucis et al. 2014, Irwin et al. 2005, Hoke et al. 2011, Williams et al. 2011, Morgan et al. 2014, Grant & Wilson 2012 Need error bars on geologic constraints to avert climate model overfitting

Omitted: later-stage and non-climate driven: e.g. Fassett et al. 2010, Hobley et al. 2014, Hauber et al. 2013, Kleinhans et al. 2010, Kraal et al.2008, Mangold et al. 2012, Harrison et al. 2011, Jones et al. 2011.

Uzboi-Ladon

Isidis rim

Gale crater large valleys alluvial fansspace-timecorrelation?

E. Meridiani…

how many episodes of climate-driven river formation?

?

500 m

Advantages of Aeolis Dorsa(a 105 km2 sedimentary-rock basin, ~10°E of Gale)e.g. Burr et al. 2009, Zimbelman & Scheidt 2012,Kite et al. 2013, Kite et al. Nat. Geosci. in press

~ 30

m ra

nge

In e

leva

tion

PSP_

0074

74_1

745

/ ESP

_024

497_

1745

(DTM

)

Basin-scale mapping distinguishes 102 m-thick river-deposit-hosting units

F1

Kite et al., Nature Geoscience, in press

= HiRISE DTMs by Antoine Lucas

smoothly eroding,fine-scale channels,retains many craters

overlies

900 km E of Gale

yardangs, meander belts,retains fewcraters

Howard PNAS 2009

60 km

F2

B20_017548_1739_XI_06S206W

400m

Data reduction

inspired by Howard & Hemberger, Geomorphology 1991Similar approach for channel widths.

color:moderntopographicrange (4m)

102 wavelengths (52 channels)

189 widths(137 channels)

Dramatically different erosional expression, modest change in river-deposit dimensions

38% change in median channel wavelength 38% change in median channel width

no evidence for changing with stratigraphic elevation (n=42) modal is = 10-15

Channel wavelength (m) Channel width (m)

Foreman et al., Nature 2012See also: Ward et al., Science 2000

Amundson et al., GSA-B 2012

Example: During planet-scale hyperthermal: rapid increase in sediment flux and discharge increased precipitation

thick, wide, multistorey channel deposits

Before planet-scale hyperthermal: thin, narrow single-story channel deposits

Fluvial signatures of climate events on Earth

find breakpoints

using nonparametricmethod

A tool to search for abrupt climate change on Mars

bootstrap

F1F2

repeat103 x

CDF of breakpointsfrom 103 trials:

also perturb bootstrapped

points

zstrat

no evidence for

abrupt change

abrupt change

zstrat

# breakpoints

0

nominal breakpoint

breakpoint in bootstrapped

data

Results presented today are similar using planar, quadratic, IDW, and universal kriging methods for structure contour interpolation.

1) Meander wavelengths tighten upwards2) Small meanders rare/absent below contact, common above contact

(4-2

0) M

yr fr

om e

mbe

dded

-cra

ter f

requ

ency

(K

ite e

t al.,

‘Pac

ing

Early

Mar

s flu

vial

…,’

Icar

us 2

013)

strat.errorstrat.error

1. Meander wavelengths

2. River widths

1) Channel widths narrow upwards2) Narrow channels rare/absent below contact, common above contact

(4-2

0) M

yr fr

om e

mbe

dded

-cra

ter f

requ

ency

(Kite

et a

l., ‘P

acin

g Ea

rly M

ars

fluv

ial…

,’ Ic

arus

201

3)

strat.error

• Catchment area is unknown

• Wider channel deposits at higher stratigraphic levels (F3)

• Taphonomy of channels?– e.g. Williams et al. Icarus 2013

• Role of aeolian deposition?– e.g. Milliken et al. GRL 2014,

Kocurek & Ewing SEPM Sp. Pub. 2012, Kite et al. Geology 2013, Bridges & Muhs SEPM Sp. Pub. 2012

Limitations and caveats

This talk

alluvial fans

meander belts

“rhythmite”

river depositsnot observed

snowpack temperature

- Consistent with orbital forcing Evidence for 3 m amplitude cut-and-fill cycles during F2(wet-dry cycle in Burns Fm? Metz et al.

2009) - Also consistent with multiple transient events!

Phas

e of

orb

ital c

ycle

Kite et al., ‘Seasonal melting…,’ Icarus 2013

Variability in river discharge versus time isn’t enough to exclude orbital forcing

Threshold

Mean Nominalbankfull Q from mean

Mean Nominal bankfull Q from mean

interpolated elevation, geologic break-point:

F2 (275±26) m 42 m3 s-1 (21±2) m 20 m3 s-1

F1 (402±18) m 76 m3 s-1 (31±3) m 43 m3 s-1

Change 1.5 x 1.8 x 1.5 x 2.1 x

interpolated elevation, break-point from hydrology:

above (274±16)m TBD m3 s-1 18 m TBD m3 s-1

below (410±19)m TBD m3 s-1 30 m TBD m3 s-1

change 50% 67%

unit assigned from map (no interpolation)

F2 (320±26)m TBD m3 s-1 (23±1) m TBD m3 s-1

F1 (420±18)m TBD m3 s-1 (35±2) m TBD m3 s-1

change 1.31x 1.7x 1.67

Total interval (4-20) Myr from embedded-crater frequency (Kite et al., ‘Pacing fluvial…,’ Icarus 2013)

Results using other interpolation methods & other break-pointsare similar

Eaton, Treatise on Geomophology, 2013Burr et al., JGR-E, 2010

Conclusions• F2 records a distinct river-forming episode - after the big meander

belts and before the alluvial fans.– ~40% reduction in river-deposit dimensions at or near the F1/F2 contact in Aeolis

Dorsa

• Consistent with ~2 x reduction in peak discharge across the contact.– 200m stratigraphy, (4-20) Myr total depositional interval – Goal: relate to quantitative models linking sed. & strat. to climate e.g. Kite et al., ‘Seasonal melting,’ Icarus 2013, Kite et al., ‘Growth and form …’, Geology 2013

With thanks to: Devon Burr, Alan Howard, Rebecca Williams, Robert Jacobsen, Lynn Carter, Bill Dietrich, Laura Kerber, Frederik Simons, Ross Irwin, Bill Dietrich, Alexandra Lefort, & Noah Finnegan for discussions, ideas, and inspiration.

More information: www.astro.princeton.edu/~kite

End of presentation

F2

MSL rover

This Talk

longitude

latit

ude

Thar

sis

Hellas

Early Mars water-availability model output (Kite et al., ‘Seasonal melting …’ Icarus 2013a)

Relevance to Gale Crater

Zimbelman & Scheidt, Science 2012

Correlated in modelsof liquid-water availabillity:

Correlated in lithology(?):