Debris-flow erosion and deposition dynamics

Abstract

Debris flows are a major natural hazard in mountains world wide, because of their destructive
potential. Prediction of occurrence, magnitude and travel distance is still a scientific challenge,
and thus research into the mechanics of debris flows is still needed. Poor understanding of the
processes of erosion and deposition are partly responsible for the difficulties in predicting debrisflow
magnitude and travel distance. Even less is known about the long-term evolution of debrisflow
fans because the sequential effects of debris-flow erosion and deposition in thousands of flows
are poorly documented and hence models to simulate debris-flow fans do not exist. Here I address
the specific issues of the dynamics of erosion and deposition in single flows and over multiple
flows on debris-flow fans by terrain analysis, channel monitoring and fan evolution modeling.
I documented erosion and deposition dynamics of debris flows at fan scale using the Illgraben
debris-flow fan, Switzerland, as an example. Debris flow activity over the past three millenia in
the Illgraben catchment in south-western Switzerland was documented by geomorphic mapping,
radiocarbon dating of wood and cosmogenic exposure dating of deposits. In this specific case I
also documented the disturbance induced by two rock avalanches in the catchment resulting in
distinct patterns of deposition on the fan surface. Implications of human intervention and the
significance of autogenic forcing of the fan system are also discussed.
Quantification and understanding of erosion and deposition dynamics in debris flows at channel
scale hinges on the ability to detect surface change. But change detection is a fundamental task
in geomorphology in general. Terrestrial laser scanners are increasingly used for monitoring down
to centimeter scale of surface change resulting from a variety of geomorphic processes, as they
allow the rapid generation of high resolution digital elevation models. In this thesis procedures
were developed to measure surface change in complex topography such as a debris-flow channel. From this data high-resolution digital elevation models were generated. But data from laser
scanning contains ambiguous elevation information originating from point cloud matching, surface
roughness and erroneous measurments. This affects the ability to detect change, and results
in spatially variable uncertainties. I hence developed techniques to visualize and quantify these
uncertainties for the specific application of change detection. I demonstrated that use of data filters
(e.g. minimum height filter) on laser scanner data introduces systematic bias in change detection.
Measurement of debris-flow erosion and deposition in single events was performed at Illgraben,
where multiple debris flows are recorded every year. I applied terrestrial laser scanning
and flow hydrograph analysis to quantify erosion and deposition in a series of debris flows. Flow
depth was identified as an important control on the pattern and magnitude of erosion, whereas deposition
is governed more by the geometry of flow margins. The relationship between flow depth
and erosion is visible both at the reach scale and at the scale of the entire fan. Maximum flow depth
is a function of debris flow front discharge and pre-flow channel cross section geometry, and this
dual control gives rise to complex interactions with implications for long-term channel stability,
the use of fan stratigraphy for reconstruction of past debris flow regimes, and the predictability of
debris flow hazards.
Debris-flow fan evolution on time scales of decades up to ten thousands of years is poorly
understood because the cumulative effects of erosion and deposition in subsequent events are
rarely well documented and suitable numerical models are lacking. Enhancing this understanding
is crucial to assess the role of autogenic (internal) and allogenic (external) forcing mechanisms on
building debris-flow fans over long time scales. On short time scales understanding fan evolution
is important for debris-flow hazard assessment. I propose a 2D reduced-complexity model to
assess debris-flow fan evolution. The model is built on a broad range of qualitative and empirical
observations on debris-flow behaviour as well as on monitoring data acquired at Illgraben as part
of this thesis. I have formulated a framework of rules that govern debris-flow behaviour, and that
allows efficient implementation in a numerical simulation. The model is shown to replicate the
general behaviour of alluvial fans in nature and in flume experiments. In three applications it
is demonstrated how fan evolution modeling may improve understanding of inundation patterns,
surface age distribution and surface morphology.