Hurricanes Frances and Ivan

Peeks Creek (2004)

Peeks Creek (2004)

Peeks Creek (2004)

Peeks Creek (2004)

Far from a one-off

We need to understand:

• Dynamics of hillslope sediment transport
• Where these landslides occur
• What triggers landslides
• How we can predict them

Colluvial hollows

• In upland, humid landscapes we know that the majority of landslides initiate in colluvial hollows
• But these features are difficult to map

Theory

Theory

• Theory predicts uniform morphology

Reality

Colluvial hollows

No systematic study of more than a handful

Two Questions:

1. Can we identify hollows from topographic data?

2. How accurate are our models?

Coweeta, NC

Coweeta, NC

• Long-Term Ecological Research site covering $50~km^2$
• Daily rain gauge data going back to the 1930s
• Persistent soil mantle through much of the Quaternary
• Type location for colluvial hollows
• High resolution LiDAR data available

Digital deforestation

Digital deforestation

Extracting hollows from topography

1. Digitally remove forests


2. Identify channel initiation point


3. Identify upstream contributing area

Slope-area plots

Slope-area plots

Slope-area plots

Slope-area plots

Slope-area plots

Slope-area plots

Slope-area plots

Slope-area plots

Slope-area plots

Slope-area plots

Slope-area plots

$\chi$-analysis

A coordinate transformation which allows us to represent river long profiles as stright lines, normalising by drainage area

$\chi$-analysis

Extracting hollows

Extracting hollows

Measuring hollows

How can we compare shapes quantitatively?

The shapes of things are very important to some people

The shapes of things are very important to some people

But how does the EU measure our fruit?

But how does the EU measure our fruit?

But how does the EU measure our fruit?

But how does the EU measure our fruit?

Elliptical Fourier Analysis

• Decompose a polygon as a series of ellipses (harmonics)
• Shapes are normalised with respect to area, rotation and point spacing
• Allows quantitative comparisons of hollows from across a landscape

Elliptical Fourier Analysis

Elliptical Fourier Analysis

Elliptical Fourier Analysis

Our data shows that the theory is wrong!

The theory is wrong?

Hollows are transitional landforms between hillslopes and channels:

Channel Erosion: $E = K_{f}A^{m}S^{n}$

Hillslope Erosion: $E = K_{h}S\frac{\rho_s}{\rho_r}$

In hollows, these two forces should be approximately equal, yielding uniform hollows

The theory is wrong?

What processes are missing from our understanding?

• Subsurface morphology
• The role of biota, especially vegetation
• Stochastic landsliding

Models of slope stability make implicit or explicit assumptions about the uniformity of hollows

Modelling hollow soil depth

The limiting factor on landslide hazard in hollows is the rate of soil production, not the intensity of storms!

Future steps

Introduce more complex hydrology and more spatial variability to model hollows...

Summary

• Understanding small scale hillslope morphology is vital to understanding landsliding
• The next generation of landslide models need to consider source area morphology
• I wasn't wasting my time reading about how to measure the shape of fruit...

Thank You!