|
Intellectual Merit
|
Techniques for surface dating using cosmogenic nuclides (CNs), such as Beryllium-10 (10Be), have developed significantly over the past ∼30 years (Lal, 1991; Morris, 1991; Anderson et al., 1996; Repka et al., 1997; Gosse and Phillips, 2001; Balco et al., 2008; Hidy et al., 2010). Despite the in- creasingly common use of CN dating methods, standardized approaches for quantifying the effects of inheritance remain elusive. Common approaches for determining inheritance include excavating pits to constrain the exponential falloff in CN production with depth (Anderson et al., 1996) and sampling sediments in an active channel to obtain catchment-wide estimates of inheritance (Repka et al., 1997). Treatment of inheritance in datasets composed of individual clasts collected from a surface commonly involves removing outliers, either by visual inspection (e.g. Frankel et al., 2011) or using a statistical rationale, such as Chauvenet’s criterion (e.g. Me ́riaux et al., 2012). Numerous studies acknowledge that assumptions of Gaussian behavior for the statistical distribution of such datasets, evaluated using a variety of goodness-of-fit statistics, are not valid (e.g. Hidy et al., 2010). Landslides have been suggested as one factor contributing to the skewed statistical distribution of these datasets (e.g. McPhillips et al., 2014; Niemi et al., 2005; Yanites et al., 2009). We have devel- oped a model that builds on previous work (Niemi et al., 2005; Yanites et al., 2009) to predict the distribution of inherited 10Be concentrations in cobble and boulder samples from a surface deposit sourced from a landslide-dominated catchment. Our analytical model suggests that a generalized pareto distribution (GPD), rather than a normal distribution, provides a more appropriate fit to the distribution of cobble ages from geomorphic surfaces reported in many publications. This model will allow for more accurate determinations of surfaces age, and will have implications for any geomorphic study of surface age using in-situ produced CNs. These results are particularly impor- tant for determinations of fault slip rate, where an over- or underestimate of surface age will result in a slip rate biased towards lower or higher values, respectively. Incorrect interpretation of these results can have far-reaching consequences, particularly for seismic hazard, which is determined in large part on estimates of fault slip rate over Quaternary timescales. We are continuing to work on developing what will be an open-source algorithm for determining the error bounds on surface age using our model. |
