Active or decaying landscapes?

Hurst Figure 1

New methods to study the shape of Earth’s surface show that it can offer clues about whether a landscape is growing or decaying.

Landscapes may grow in response to movement of the Earth’s crust at plate margins or along faults, where earthquakes often occur.

This discovery came from analysing high quality, 3D, digital representations of the elevation of the Earth’s surface, acquired by laser-scanning technology mounted on an aeroplane (see Figure 1).

The Dragon’s Back pressure ridge formed on an actively uplifting segment of the San Andreas Fault in California. At the fault, two tectonic plates slide past one another, carrying the growing landscape away from the focus of uplift so that the landscape then decays.

New techniques may assist earthquake prediction

New techniques to analyse the digital surface model documented changes in the shape of hills during landscape growth and decay, which were distinct.

This research carried out at the University of Edinburgh, in collaboration with scientists in California, suggests that detailed studies of hillslopes may help identify zones of tectonic activity that might pose earthquake and landslide hazards.

The article describing this work was published in Science — Hurst et al., 2013. Hillslopes Record the Growth and Decay of Landscapes.

How hillslopes change over long timescales

Hurst Figure 2

Generally in a growing landscape, slopes get gradually steeper, the elevation of hills and mountains increases, and ridges/hilltops get narrower.

Conversely when a landscape relaxes, slopes and elevations reduce and hilltops broaden. (Figure 2).

Critically, the timing of these changes is distinct for growing and decaying landscapes.

A hillslope will first begin to grow when a stream starts to erode faster (which in the case of the Dragon’s Back is a response to surface uplift).

The toe of the hillslope will steepen first, and through time the steepened part of the hillslope will grow and extend upward toward the top.

So, the hilltop is the last part of the hillslope to ‘feel’ the change.

Hillslopes do not like to be too steep, and gravitational processes such as landslides and creep (the slow, steady movement of regolith downslope) try to lower the hillslope.

But if the landscape continues to be uplifted and streams continue to erode down, it is all the hillslopes can do just to keep up!

Once uplift stops the channels don’t need to cut down any more, but hillslopes still don’t want to be steep. They gradually lower and their sediment accumulates in the valley floors as the landscape relaxes.

Unlike during landscape growth the hillslopes and hilltops appear to respond in concert, rather than in turn and it is this difference that allows us to distinguish between growing and decaying landscapes (Figure 3).

Hillslope schematic

Uplift and landscape changes at the Dragon’s Back pressure ridge

Numerical models of how landscapes evolve can predict these responses, but we needed to find a real landscape to test.

The Earth’s surface evolves very slowly relative to the lifespan of a human, or even that of human records, so it is rare that we get the chance to observe these sorts of changes in action, beyond getting a snapshot — typically frustrating for geologists and geomorphologists.

‘Living field laboratory’

However, a rather special landscape called the Dragon’s Back pressure ridge in California has allowed a time-series of landscape change to be inferred in response to the occurrence of surface uplift driven by crustal movement along the San Andreas Fault.

Understanding of the tectonic processes at the Dragon’s Back site and their interaction with the Earth’s surface is the result of extensive research by co-author Prof. George Hilley (Stanford University), Prof. Ramon Arrowsmith (Arizona State University) and their colleagues.

Hilley and Arrowsmith's publication, Geomorphic response to uplift along the Dragon's Back pressure ridge..., in the journal Geology in 2008 summarised the nature of tectonic forces acting in the area derived by meticulous geological mapping and interpretation of the required displacements in the Earth’s crust.

Their efforts suggested that uplift is pinned with respect to the North American tectonic plate, whilst the Pacific plate is carried through the zone of uplift and out the other side.

They therefore concluded that a space-for-time substitution was possible, so that the landscape further away from the zone of uplift along the Dragon’s Back landform has had longer to relax since being uplifted.

Using ‘geomorphic signatures’

Hilley and Arrowsmith (2008) use the geomorphic signatures of channel and hillslope processes to infer that channels respond rapidly to the changes in surface uplift.

This is important because the channels set the boundary for hillslopes, so that a hillslope can only respond to uplift if the channels at the bottom of the slope have adjusted their erosion rates first.

This was vital for our goal to document hillslope response to tectonic processes.
Ramon Arrowsmith from Arizona State University recently posted a wealth of additional information about the field site on his Active Tectonics Blog.

This material has been accumulated as part of research on the area spanning two decades, without which this work would not have been possible, or even plausible. You can view Ramon’s post here:

Quantifying hillslope properties

Hurst Figure 4

Some theoretical work, by Prof. Josh Roering (University of Oregon) and colleagues presented in Earth and Planetary Science Letters in 2007, suggested that the average gradient of hills and the curviness of their summits should vary predictably with how rapidly adjacent streams are downcutting, which in turn is controlled by tectonic processes at the Dragon’s Back site.

Hurst et al. (2012) in the Journal of Geophysical Research — Earth Surface developed techniques to extract quantitative properties of hillslopes from digital surface models for comparison to such predictions.

The approach involves a simple model for routing flow across the surface by assuming water would follow the steepest path, in order to identify the likely location of streams and rivers.

From the stream network it is possible to collect which parts of the landscape are connected to which streams to find all the small catchments that make up a landscape.

The margins of these catchments can then be collected together to form a network of hilltops.

Following the steepest path from each hilltop back to an identified stream then allows the calculation of the length of the hillslope, its mean slope and the curvature of its hillcrest (Figure 4).

Growth or decay?

The location of hilltops and the distribution of uplift at the Dragon’s Back pressure ridge are shown in Figure 5(a).

Figure 5 (b) shows the quantities of mean hillslope gradient (S), hilltop curvature (CHT) and hillslope length (LH) and how they change along the length of the landform, and by inference, how they change through time.

Note that hillslopes become steep (high S) relatively quickly but their hilltop curvature does not peak until further along/later on.

This is critical for discriminating growing from decaying landscapes.

Hurst Figure 5

Time series of hillslope form

Roering et al. (2007) described two parameters for exploring hillslope morphology, a dimensionless erosion rate E* which is mainly governed by hilltop curvature, and a dimensionless relief R* which is dictated by hillslope gradient.

These parameters are based on theoretical models that approximate how hillslopes transport sediment.

In theory, if hillslopes are undergoing constant rate of change (i.e. their adjacent channels are downcutting at a fixed and constant rate), the values of E* and R* should follow a theoretical relationship shown by the dashed line in Figure 6.

Hurst Figure 6

Using the data from Figure 5 and colour-coding the results as a function of time/distance along the landform, hillslopes at the Dragon’s Back pressure ridge show morphology that is distinct when the landscape is growing compared to when it is decaying.

This is primarily driven by a lag in the hillslope response, since the hilltop can only respond once the entire hillslope is tightly coupled to erosion rates in the adjacent streams.

Hillslope gradients increase rapidly, but hilltop curvature struggles to increase until much later. Conversely during relaxation of the landscape, slope gradients and hilltop curvature both wane simultaneously.

The result is supported by application of a numerical model for the evolution of a hillslope, from which the history of channel incision can be back calculated by simulating the hillslope many thousands of times, whilst slightly changing the history the model experiences, and looking for the best match to the observed topography (Figure 7).

The model supports the occurrence of hysteresis in hillslope form and reproduces an incision history remarkably similar to that inferred from valley form by Hilley and Arrowsmith (2008).

Hurst Figure 7


Hilley, G E, and Arrowsmith, J R.  2008.  Geomorphic response to uplift along the Dragon's Back pressure ridge, Carrizo Plain, California.  Geology, 36(5), 367-370, doi:10.1130/g24517a.1.

Hurst, M D, Mudd, S M, Attal, M, and Hilley, G.   2013. Hillslopes Record the Growth and Decay of Landscapes  .  Science, 341(6148), 868-871, doi:10.1126/science.1241791.

Hurst, M D, Mudd, S M, Walcott, R, Attal, M, and Yoo, K.  2012. Using hilltop curvature to derive the spatial distribution of erosion rates.  Journal of Geophysical Research-Earth Surface, 117(F2), F02017, doi:10.1029/2011jf002057.

Roering, J J, Perron, J T, and Kirchner, J W.  2007.  Functional relationships between denudation and hillslope form and relief. Earth and Planetary Science Letters, 264(1-2), 245-258, doi:10.1016/j.epsl.2007.09.035


This work was supported by a National Environmental Research Council grant (NE/G524128/1) awarded to Martin D. Hurst and by grant NE/H001174/1 and a grant from the Carnegie Trust for the Universities of Scotland to Simon M. Mudd.


Contact enquiries further information