United States Department of Agriculture IN THIS ISSUE • The Science and Prediction of Post-Fire Debris Flows in the Western United States • Notices and Technical Tips • Predicting Erosion Risk during Large Floods The Technical Newsletter of the National Stream and Aquatic Ecology Center Fort Collins, Colorado August 2017 high likelihood of catastrophic wildfires in the western United StreamNotes is an aquatic and riparian systems publication with The Science and States and the encroachment of the objective of facilitating Prediction of Post-Fire human activities into steep fire- knowledge transfer from research prone areas have created the need to & development and field-based Debris Flows in the better understand, predict, and success stories to on-the-ground Western United States mitigate these hazards. This article application, through technical highlights recent advances in articles, case studies, and news Dennis M. Staley understanding post-fire debris-flow articles. Stream related topics Research Physical Scientist generation and provides an include hydrology, fluvial U.S. Geological Survey overview of the evolution of hazard geomorphology, aquatic biology, Denver, Colorado riparian plant ecology, and climate assessments in the western United change. Jason W. Kean States. Specific emphasis is placed upon free, publicly available tools StreamNotes is produced quarterly Research Hydrologist as a service of the U.S. Forest U.S. Geological Survey developed by the U.S. Geological Service National Stream and Denver, Colorado Survey (USGS) and available for Aquatic Ecology Center (NSAEC). analyses in support of Burned Area This technical center is a part of the Debris flows are among the most Emergency Response (BAER) and Washington Office’s Watershed, destructive hydrological Emergency Watershed Protection Fish, Wildlife, and Rare Plants consequences of fires in steep (EWP) program activities. These program. watersheds (Figure 1, Figure 2). The tools consist of empirical models Editor: David Levinson Technical Editors: • Steven Yochum • Brett Roper Layout: Steven Yochum To subscribe to email notifications, please visit the subscription link. If you have ideas regarding specific topics or case studies, please email us at [email protected] Ideas and opinions expressed are not necessarily Forest Service policy. Citations, reviews, and use of trade names do not constitute endorsement by the USDA Forest Service. Click here for our non- discrimination policy. Figure 1: Downstream effects of post-fire debris flows in Camarillo Springs, California, downstream of the area burned by the 2013 Springs Fire (12/12/2014). 1 of 10 that predict the likelihood, potential volume, and the rainfall intensity- duration thresholds for debris flows in recently burned watersheds. Fire and Debris-Flow Generation Wildfire significantly increases the potential for post-fire debris-flow generation in recently burned watersheds by enhancing the discharge and velocity of surface water flow, resulting in higher rates of erosion. While runoff in unburned forested watersheds is typically generated through saturation-excess overland flow processes, runoff in recently burned areas is generated mainly as infiltration-excess (Hortonian) Figure 2: Downstream effects of post-fire debris flows in Mullally Canyon, La overland flow, which occurs when Crescenta, California, downstream of the area burned by the 2009 Station Fire rainfall rates exceed the infiltration (2/6/2010). capacity of the soil. The fire- induced physical and chemical connectivity between hillslopes and changes to soil and vegetation gullies, further permits more rapid systems, such as decreased raindrop Rainfall and Debris-Flow flow concentration and transfer of Generation interception, hyper-dry conditions, water and sediment downslope, enhanced hydrophobicity and which in turn increases discharge, The temporal occurrence of post- infiltrated ash, serve to effectively shear stress, and sediment yield fire debris flows has been found to increase the amount of rainfall (Moody and Kinner, 2006; Neary et closely correlate with pulses of reaching the surface, decrease the al., 2012). high-intensity rainfall and the hydraulic conductivity, and generation of infiltration-excess Unlike debris flows that initiate decrease the infiltration rate of overland flow (Wells, 1987; Gabet, from shallow landslides, there is no burned soils, thereby increasing 2003; Cannon et al., 2008; Kean et discrete initiation point or material runoff discharge and related al., 2011; Staley et al., 2013). source in a majority of post-fire erosion. In addition to increased Antecedent moisture conditions debris flows (Parrett, 1987; Meyer discharge, the runoff velocity in (e.g., after wildfire, either within and Wells, 1997; Cannon, 2001). burned areas is often higher after single storms or seasonal) have been Instead, debris-flow initiation in wildfire. Combustion of surface found to have very little, if any, burned areas often results from litter reduces surface roughness, influence on the likelihood of post- progressive sediment bulking time to ponding (i.e., the filling of fire debris-flow initiation (Cannon processes, where infiltration-excess fine-scale surface depressions) and et al., 2008). For example, in a plot- overland flow produced on time to reach pond capacity (i.e., the scale field experiment, Wells (1987) hillslopes gradually entrains amount of time needed to overtop was able to initiate small debris material, ultimately transforming fine-scale surface depressions), flows after only 3 minutes of rainfall sediment-laden surface water flow allowing runoff to flow at intensities between 12 and 55 to debris flow (Kean et al., 2013). uninterrupted downslope at higher mm/h. More recently, field Severe erosion of hillslopes, gullies velocities (Moody and Ebel, 2014; monitoring in the San Gabriel and channels serves as the primary Moody and Martin, 2015). Mountains of southern California source of material for the debris Increased flow depth produces recorded the occurrence of post-fire flows generated from progressive higher shear stresses, inducing rill debris flow after 16 minutes of sediment bulking (Santi et al., 2008; and gully erosion in areas of moderate intensity rainfall during Smith et al., 2012; Staley et al., concentrated flow. Development the very first rainstorm following 2014). and expansion of well-defined rill wildfire (Kean et al., 2011; Staley et and gully networks, and enhanced al., 2013). Debris flows have also StreamNotes 2 of 10 U.S. Forest Service August 2017 National Stream and Aquatic Ecology Center been generated several months after associated with the passage of a the calculation of debris-flow wildfire after several significant debris flow (Kean et al., 2011), likelihood and volume, and the 15- rainstorms (Kean et al., 2011). rather than a water-dominated flow minute rainfall intensity-duration (i.e., flash flood). threshold are provided in Table 1. Within rainstorms, debris-flow generation has been strongly correlated with short bursts of high- Post-Fire Debris-Flow Post-Fire Debris-Flow intensity rainfall (Kean et al., 2011; Prediction Hazard Assessment Staley et al., 2013). From precise The basic post-fire research findings Prior to January 2014, most post- monitoring of debris-flow timing, outlined above have provided the fire debris-flow hazard assessments Kean et al. (2011) identified a near- foundation for the development of published by the USGS were zero lag time between the tools for the prediction of post-fire produced as hard-copy reports occurrence of short bursts of high- debris-flow hazards in the western accompanied by a series of poster- intensity rainfall and the passage of United States. These multivariate sized digital maps that displayed a debris flow at the San Gabriel statistical models are intended to post-fire debris-flow probability, monitoring site. The best temporal predict (1) the likelihood of a debris expected volume, and combined correlation and shortest lag between flow at a given location in response hazard. Feedback from primary rainfall intensity and debris-flow to a design storm (Staley et al., stakeholders, including USFS initiation was identified for rainfall 2016; USGS, 2017), (2) potential Burned Area Emergency Response intensity measured between 5 and debris-flow volume (Gartner et al., (BAER) teams, the National 30 minute durations (Kean et al., 2014), and (3) rainfall intensity- Weather Service (NWS), and 2011; Staley et al., 2013). In duration thresholds (Cannon et al., numerous other state and local addition, the first indication of flow 2008; Cannon et al., 2011; Staley et agencies suggested that this mode of in a stream channel (as measured by al., 2013; Staley et al., 2015; Staley assessment dissemination was stage) has frequently been et al., 2017). Specific equations for antiquated and relatively ineffective Table 1: Post-fire debris-flow hazard assessment model equations used for predicting likelihood, potential volume, and estimated rainfall intensity-duration threshold. Model Name Equation / Variables Citation Debris-Flow X = -3.63 + (0.41 * PropHM23 *i15) + (0.67 x (dNBR / 1000) * i15) + (0.7 * KFFACT * i15) Staley et Likelihood al., 2016 (L) L = exp(X) / (1 + exp(x)) PropHM23 = Proportion of upslope area burned at high or moderate severity with gradient in excess of 23° dNBR / 1000 = average differenced normalized burn ratio (dNBR) of upslope area, divided by 1000 KFFACT = soil erodibility index of the fine fraction of soils