Spatial and temporal channel changes across the watershed scale following wildfire and floods
Date
2018
Authors
Brogan, Daniel Joseph, author
Nelson, Peter A., advisor
Kampf, Stephanie K., committee member
MacDonald, Lee H., committee member
Niemann, Jeffrey D., committee member
Journal Title
Journal ISSN
Volume Title
Abstract
Fires and floods are important drivers of geomorphic change. The hydrologic and sedimentologic effects of fires have been relatively well studied at the hillslope scale, but we still lack the ability to accurately quantify and predict post-fire flooding and geomorphic changes at larger scales. This lack of understanding stems primarily from two reasons. First, there is generally limited availability of repeat high-resolution topography following fires, and this limits our ability to quantify and explain changes throughout a given channel network. Second and more fundamentally, one cannot simply scale up hillslope processes to the watershed scale, or vice-versa. Since global warming is leading to more wildfires and a higher likelihood of extreme precipitation, understanding downstream flooding and sedimentation is more critical than ever for safeguarding downstream landowners, water users, and aquatic biota. This dissertation investigates these shortcomings by documenting post-fire channel changes across watershed scales and how extreme floods can alter the more typical post-fire geomorphic response. I focus on two ~15 km2 watersheds, Skin Gulch (SG) and Hill Gulch (HG), that burned in the 2012 High Park Fire, Fort Collins, Colorado, U.S.A. Over the subsequent four years I used repeat surveys of 10-11 cross sections and longitudinal profiles along the lower channel network of each watershed, and five sequential airborne laser scanning (ALS) surveys, to quantify erosion and deposition. SG was first subjected to a high-intensity convective storm just days after the fire was contained; the resulting flood caused an exceptionally large peak flow, and extensive downstream deposition of cobbles, boulders and woody debris. Fifteen months later SG and HG experienced catastrophic stripping and bed coarsening due to an unusually rare and widespread mesoscale storm, with much greater changes in SG. These events and the data used to document their effects set up the basis of three separate, yet interdependent comparisons. First, I compare and contrast the peak flows and the effects of the two distinctly different flood disturbances in SG: the short-term peak flow and substantial deposition caused by the convective flood immediately after burning; and the widespread channel and valley bottom erosion caused by the mesoscale flood. Peak flows were estimated using three independent techniques: 1) slope-area method, 2) critical flow, and 3) 2D hydrodynamic modeling. The peak flow estimates for the 2013 flood had a higher relative uncertainty and this stemmed from whether I used pre- or post-flood channel topography. The results document the extent to which a high and moderate severity wildfire can greatly increase peak flows and alter channel morphology, illustrate how indirect peak flow estimates have larger errors than is generally assumed, and indicate that the magnitude of post-fire floods and geomorphic change can be affected by the timing, magnitude, duration, and sequence of rainstorms. Second, I use the repeat surveys of the cross sections and longitudinal profiles to quantify the channel response to the 2012 wildfire, summer thunderstorms, spring snowmelt, and the mesoscale flood in both SG and HG. The varying response between the two watersheds during the mesoscale flood necessitated further investigation. Discussions with a local landowner indicated that a flood in 1976 caused tremendous channel erosion and widening in the lower portion of HG. Geomorphic changes in HG after the fire and the mesoscale flood were much smaller than in SG, and this can be attributed to: greater post-fire, pre-mesoscale flood deposition in SG; reduced sensitivity in HG as a result of the large erosional flood in 1976; and the spatial distribution of burn severity leading to a lower peak flow in HG from the mesoscale flood. These results suggest that fires can trigger significant and dynamic channel changes over sub-decadal timescales, but unusually long or intense rainstorms can cause larger and more persistent watershed-scale changes regardless of whether a catchment has recently burned. I propose a state-and-transition conceptual model to relate landscape sensitivity to geomorphic changes according to its history of fires and floods. Third, I use the repeat ALS data to quantify spatial and temporal patterns of erosion and deposition throughout the channel networks of SG and HG. These volumes of change are related to valley and basin morphology, precipitation amounts and intensities, and burn severity. The results suggest that the amount and location of stored sediment in the valleys is critical for evaluating potential locations of erosion and deposition. Morphometric characteristics, when combined with burn severity and a specified storm, can indicate the relative likelihood and locations of post-fire erosion and deposition risks. Taken together, this body of work demonstrates: 1) how the timing and sequence of different disturbances affect the relative sensitivity of watersheds to downstream channel changes; 2) that the effects of extreme floods are longer lasting and more dominant than the effects of wildfires; and 3) that the amount and location of stored sediment in the valleys is critical for predicting potential geomorphic change. This information can help resource managers assess downstream risks and prioritize areas for post-fire hillslope rehabilitation treatments.
Description
Rights Access
Subject
erosion
geomorphology
wildfire
floods
deposition
sensitivity