Tahoe Abstracts
Discharge variability controls on erosion by overland flow
Kelin Whipple*, Roman DiBiase, and Matt Rossi
Erosion and sediment transport by overland flow only occur when a threshold bed shear stress is surpassed. As a consequence of this non-linearity erosion and sediment transport are dominantly accomplished by relatively extreme events. In theory the integral of all floods exceeding this threshold will contribute to erosion and sediment transport so the shape of the flood probability distribution is critically important. Flood distributions (whether measured using mean daily flows or peak annual flows) are heavy-tailed, often with an approximately power-law tail. The “heaviness” of flood distributions varies with climatic setting, however. As has long been recognized at the inter-annual scale, Molnar et al. (2006) showed that mean daily flow distributions are heavier tailed in arid settings. We follow the work of Lague et al. (2005) to illustrate the influence of both mean daily flows (for a given distribution) and flow variability (heaviness of the tail for a given mean flow) on erosional effectiveness, and to explore the geomorphic consequences of the inverse correlation between mean flow and flow variability. This analysis reveals a rich behavior fundamental to our understanding of the linkages between climate, climate change, and erosion. Important open questions include why there is an inverse correlation between mean flow and flow variability in the continental US, and whether this holds in settings with a broader range of climatologies. The processes involved in the transformation of precipitation into runoff are well known and numerous factors that could dampen or amplify the input variability of precipitation are readily identified, but much remains unknown about the controls on runoff variability as a function of mean climate. New data resources and new methods of analysis of the spatio-temporal variability of precipitation and runoff could potentially lead to new understanding.
Climatic controls on discharge variability
Matt Rossi*, Kelin Whipple, and Enrique Vivoni
The transformation of the probability distribution of precipitation into the probability distribution of runoff is foundational to hydrology and carries important implications for the relationships among climate, climate change, and erosion. Here we present an empirical analysis of the extensive network of meterologic and river gauging stations across the continental US with the aim of revealing some of the climatic controls on runoff variability. First we establish that the stretched exponential is the most effective parametric model for quantifying the variability of both daily wet-day precipitation and daily discharge for extreme events (tails of the distributions) – consistent with the idea that precipitation statistics should be the underlying control on runoff statistics. However, we find that runoff variability is greater that the variability of wet-day precipitation at a station. Nevertheless, the inverse relationship between mean annual runoff (MAR) and runoff variability in the continental U.S. suggested by prior work (Molnar et al., 2006) is also observed in precipitation statistics. Interestingly, the precipitation metric that exerts the least control on rare precipitation events, storm frequency, largely explains this correspondence. This may follow because it is highly correlated with ET efficiencywhich can significantly amplify runoff variability by changing the relative frequency of low flood stages to high ones. Recognizing that this empirical analysis ignores other processes that must modify runoff generation like antecedent moisture, vegetation dynamics, the spatial extent of storms, and the routing of water, we use these observations to propose a framework for future studies that attempt to isolate these variables.