REMOTE ESTIMATES OF BANKFULL DISCHARGE FOR A MARTIAN-TERRESTRIAL ANALOG CHANNEL SUGGEST THE NEED FOR NEW METHODS OF ESTIMATING PALEODISCHARGE

Empirical formulas for estimating fluvial channel discharge at ungauged sites have been repeatedly extrapolated to Mars in order to estimate the paleodischarge of ancient fluvial channels and infer paleoclimates. These terrestrial-based form-discharge relationships are bivariate correlations between bankfull channel width and the 2-year flood discharge. The form-discharge relationship most commonly extrapolated to Mars (Q=1.9W^1.22 Standard Error=79%) was derived by regression of measurements from 252 sites in the Missouri River Basin (Osterkamp & Hedman 1982). In contrast to these purely correlative relationships, hydraulic geometry provides a causal relationship, describing the influence of discharge on channel width. Algebraic manipulation of a hydraulic geometry relationship, derived from a meta-analysis of many hydraulic geometry studies (Eaton 2013), yields coefficients and exponents (Q=0.106W^1.86 SE=29%) that substantially differ from those of the form-discharge relationship. To elucidate the more correct formula for Martian applications, we compare discharge estimates from each formula with the known bankfull discharge of a terrestrial analog channel. The Quinn River, NV (bankfull Q = 12 m³s^-1) is a valid terrestrial analog because its banks are not reinforced by rooted vegetation, but instead are stabilized by Late Pleistocene lacustrine silt and clay (Matsubara et al. 2015). We remotely measured bankfull width in 1-m airborne LiDAR topography and plan view visual images as the distance between topographic breaks in slope on opposing banks. Measurement results gave an average width of 15 m (n=44, s=3.6), which, when coupled with the form-discharge relationship and the manipulated hydraulic geometry relationship, yielded discharge estimates of 48 ±38 m³s^-1 and 16 ±4.6 m³s^-1, respectively. While the margins of error for the two discharge estimates cover the Quinn’s bankfull discharge (12 m³s^-1), the discharge estimate from the form-discharge relationship is less precise and less accurate than the discharge estimate from the manipulated hydraulic geometry relationship. These findings suggest that hydraulic geometric relationships are better than the commonly used form-discharge relationships from the Missouri River Basin for estimating paleodischarge on Mars.