42nd Lunar and Planetary Science Conference (2011) 2748.pdf

TIMING, DURATION, AND HYDROLOGY OF THE EBERSWALDE CRATER PALEOLAKE, MARS. R. P. Irwin III, Planetary Science Institute, 1700 E. Fort Lowell Rd., Suite 106, Tucson AZ 85719, [email protected], visiting scientist at NASA Goddard Space Flight Center, Code 698, Greenbelt MD 20771.

Introduction: Martian paleolakes offer a conve- the southern one are 140 and 180 m3/s using equations nient setting for mass- and energy-balance calculations 1 and 2, respectively. These results are broadly con- to constrain the past hydrology and water source [1,2]. sistent with previous studies [4,15,16]. Eberswalde crater is particularly useful in this regard due to its relict delta [3–7] that reflects the water level in the crater. Here we use the dominant discharge es- timated from inverted-channel width and meander geometry, the area of the paleolake and contributing watershed, and energy requirements for lake evapora- tion to constrain the mean annual runoff production and water source. Given the maximum age of the delta based on crater counts on Holden ejecta [8], these re- sults suggest a discrete post-Noachian interval of fluvial erosion on Mars. Age of the Eberswalde Delta: Although Ebers- walde crater is older than the adjacent Holden crater, the Eberswalde delta post-dates the Holden impact and superimposes its ejecta [4,8], which provide a larger surface area for crater counts. The superimposed cra- ter population >1 km in diameter indicates a Late Hes- perian age for Holden crater, with error bars that range within the Early and Late Hesperian Epochs [8]. The Eberswalde delta is thus Hesperian or younger and post-dates Noachian valley networks. Hesperian or Fig. 1. Southern meander on Eberswalde delta used for later fluvial activity also occurred at Gale crater, Aeo- discharge calculation. The circles are 340 m across. lis Mensae [9], the Valles Marineris region [10–12], and other areas on Mars [13]. Event Runoff Production: Given a dominant dis- Estimates of Dominant Discharge: Two well- charge for the northern channel of 460 m3/s, the event preserved meander loops on the Eberswalde delta have runoff production would be about 3 mm/day from a relationships between inverted-channel width and topographically defined watershed of 12,500 km2 or measures of meander geometry that are consistent with about 8 mm/day from the 5000 km2 crossed by terrestrial meandering rivers [14]. Previous studies mapped valleys. These values correspond to (perhaps have used empirical relationships between discharge significantly) higher rates of precipitation or melting, and channel dimensions or a model of alluvial fan likely in excess of 1 cm/day at times, given losses to formation, all scaled for Martian gravity, to estimate evaporation and infiltration from the watershed sur- dominant discharge ranging from 410 to 700 m3/s face. The rates for the southern channel are about a [4,15,16]. third of those for the northern one. Lake Evaporation: Evaporation from the 400- 1.22 1.22 2 Q2 = 1.9Wb (0.62) = 1.2Wb (1) km lake would be highest for a perennially ice-free 1.54 1.54 Q1.5 = 0.011λm (0.62) = 0.0068λm (2) lake and lowest for a frozen one. An energy-balance approach can be used to estimate the evaporation rate In Equations 1 and 2, Qx is the dominant discharge E (m/s): (m3/s), which has a recurrence interval of x years in the settings where the relationships were derived; Wb is E = (K + L – G – H + I – ΔQ/Δt) / ρv, (3) bankfull width; and λm is meander wavelength. The northern and southern meanders (the latter is shown in where K is net shortwave radiation input (the variables Fig. 1) reflect different channel-forming discharges, in parentheses are all J/m2s), L is net longwave radia- although their stratigraphic positions on the delta are tion input, G is energy conduction to the ground, H is consistent with a roughly similar age. Results for the sensible heat loss to the atmosphere, I is net input of northern meander are 460 and 390 m3/s, and those for energy from inflows, ΔQ is the net change in stored 42nd Lunar and Planetary Science Conference (2011) 2748.pdf

heat over time Δt, ρ is the density of water (1000 wetter intervals were hyperarid by terrestrial standards kg/m3), and v is the latent heat of vaporization (although perhaps not as dry as modern Mars). Pre- (2,260,000 J/kg) [17]. We assume a constant water vious studies introduced the concept of an intermittent temperature (ΔQ = 0), and energy exchange with the water cycle and multiple generations of valley devel- ground (G) is typically negligible. Thus, without sig- opment on early Mars [e.g., 10,19–21], although nificant heat input from a hot atmosphere (L and –H), means to activate and deactivate the water cycle re- the energy available to evaporate inflowing water is main conceptual. The need to liberate the Martian limited to excess heat in the contributing stream and water inventory from topographic and thermal traps in net shortwave radiation, which is <160 W/m2 averaged the highlands, poles, and subsurface suggest that ideal over 24 hours. The former (I) is small for a cold river conditions for polar melting (high obliquity, perihelion but rises to 76 W/m2 at 25°. Adequate energy was summer) under a thicker early atmosphere may have available to evaporate at most two or three meters of been favorable. water per year from the lake, if the lake remained un- Implications for Mars Science Laboratory frozen and the entire energy budget was used for eva- (MSL) Site Selection: The relatively young age of poration, but lower rates are more likely. fluvial to lacustrine deposits in Eberswalde and Hol- Water Source: Hypotheses for the Eberswalde den craters is likely related to the excellent preserva- water source include impact-generated runoff and tion of their stratigraphy, relative to older Noachian some kind of atmospheric water cycle. Both hypo- materials. Moreover, these deposits record an interval theses imply a limited lifetime for the lake. of time when liquid water was abundant, in contrast to Melting of pre-existing ice by hot ejecta from the drier conditions that prevailed throughout most of Holden impact would produce a single continuous but Martian history, outside of geologically discrete pe- declining discharge, of which the inverted channels on riods of valley network development. the delta surface represent a late-stage minimum. Con- Acknowledgements: Discussions with the MSL tinuous flow at 290 m3/s (the average of four esti- Fluvial Processes Tiger Team and particularly Edwin mates) would require 23 m/year of evaporation from Kite helped to refine this work. the lake surface, an order of magnitude more than the References: [1] Irwin R. P. et al. (2007) 7th Int. energy budget allows. Conf. on Mars, abstract 3400. [2] Fassett C. I. and Precipitation of snow onto hot, fresh ejecta would Head J.W. (2008) Icarus, 198, 37–56. [3] Malin M. C. require relatively high precipitation and melting rates and Edgett K. S. (2000) Science, 302, 1931–1934. [4] to generate runoff, due to the need to offset infiltration Moore J. M. et al. (2003) GRL, 30, 2292. [5] Bhatta- into the ejecta, which typically has a high infiltration charya J. P. et al. (2005) GRL, 32, L10201. [6] Wood capacity [18]. Water that entered the lake as ground- L. J. (2006) GSA Bull., 118, 557–566. [7] Lewis K. W. water rather than through the surface stream would and Aharonson O. (2006) JGR, 111, E06001. [8] Irwin also need to be evaporated, but highly intermittent R. P. and Grant J. A., USGS map, 1:500k, in revision. runoff might offset this excess supply within the ener- [9] Irwin R. P. et al. (2005) JGR, 110, E12S14. [10] gy budget. Mangold N. et al. (2004) Science, 305, 78–81. [11] Another alternative is ephemeral runoff production Fassett C. I. and Head J. W. (2008) Icarus, 195, 61– through rain or snow melted by insolation. To offset 89. [12] Di Achille G. et al. (2007) JGR, 112, E07007. evaporation of 0.1–1 m/yr from the 400-km2 lake, in- [13] Grant J. A. et al. (2009) Icarus, 10.1016/ termittent precipitation of >1 cm/day (some of which j.icarus.2009.04.009. [14] Williams G. P. (1988) in would infiltrate or evaporate from the watershed) and Flood Geomorphology, eds. V. R. Baker et al., runoff production of ~0.8–8 cm/year would be re- 321334, John Wiley. [15] Jerolmack D. J. et al. quired. To form a delta of ~6 km3 [3] at a reasonable (2004) GRL, 31, L21701. [16] Irwin R. P. et al. (2005) water-to-sediment volume ratio of 100:1, these condi- Geology, 336, 489492. [17] Dingman S. L. (2002) tions would have persisted for ~103–104-year time- Physical Hydrology, Prentice-Hall. [18] Grant J. A. scales. The Eberswalde delta did not require a very and Schultz P. H. (1993) JGR, 98, 11,02511,042. wet climate (in fact, it appears to disallow one), but [19] Baker V. R. and Partridge J. (1986) JGR, 91, intermittent runoff production was significant at times. 3561–3572. [20] Grant J. A. Schultz P. H. (1990) Ica- This rough duration falls within the wide range sug- rus, 84, 166195. [21] Baker V. R. et al. (1991) Na- gested by prior workers, <102 years [15] at one ex- ture, 352, 589–594. treme and >105 [5] at the other. Interpretation: These results suggest that peak conditions for Martian runoff may have resembled an arid terrestrial desert, and that conditions outside of