Research led by the University of Glasgow used the Isle of Jura’s rivers as a natural laboratory to show that the contrast in hardness between a river’s bedload sediment and the underlying bedrock strongly influences how quickly rivers erode landscapes, often more than bedrock strength alone.

The study, published in Earth Surface Processes and Landforms (paper titled “How does lithology influence fluvial erodibility? A study of knickpoint retreat in small bedrock rivers on the Isle of Jura”), was led by Dr. Adam G. G. Smith of the University of Glasgow’s School of Geographical & Earth Sciences, with contributions from University College London and the University of Southampton.

Around 13,600 years ago, as the last ice age ended, rapid glacio- isostatic rebound on Jura caused base-level drops. These created knickpoints (small waterfalls or steep reaches) in the island’s bedrock rivers. Knickpoints migrate upstream over time as the lip erodes and collapses, providing a measurable record of erosion rates since their formation.

Jura’s geology offers a natural experiment: rivers flow over extremely hard Jura Quartzite (one of the hardest rocks on Earth) and a much softer band of slate (Easdale Subgroup). Some rivers stay on quartzite; others cross from hard quartzite into softer slate. Researchers took 832 Schmidt hammer readings (a tool that measures rock hardness via rebound) across 35 outcrops. The quartzite averaged roughly three times harder than the slate (mean uniaxial compressive strength ~307 MPa vs. ~123 MPa).

The team built computer models of knickpoint migration under three assumptions about erodibility (K) and compared them to the actual positions of ~18 coeval knickpoints linked to the 13.6 ka base-level fall:

• Uniform erodibility (rock type ignored).
• Lithology-specific erodibility (harder rock simply more resistant).
• Contrast model: erodibility depends on the difference in strength between the bedrock and the sediment (bedload) the river carries.

The contrast model performed best, producing tighter clustering of predicted knickpoint response times around the known 13.6 ka age. Simply assigning different erosion rates by lithology performed slightly worse than assuming uniform rates. Hard sediment carried over softer bedrock can act like sandpaper and accelerate scour far beyond what bedrock strength alone would predict.

River erosion is still dominated by water discharge (rainfall and catchment area) and channel steepness; rock-hardness effects are secondary but significant. At landscape scales (tens to hundreds of kilometres), incorporating bedload, bedrock contrast improves models that currently treat hard and soft rock as simple separate blocks and miss sediment mixing.

Better constraints on fluvial erodibility help reconstruct how rivers shaped continents in the past and improve predictions of how climate change will affect future landscapes. The work also highlights Scotland’s value as a geological laboratory due to its exposed, rugged terrain.

In short, the research shows that what a river carries can matter as much as, or more than, the rock it flows over when determining how landscapes are carved.

_____________________________________________________________________________________

The Schmidt hammer (also called a rebound hammer) is a portable, non-destructive field and laboratory instrument widely used in rock mechanics, engineering geology, and geomorphology to estimate surface hardness and, via empirical correlations, uniaxial compressive strength (UCS) and elastic modulus of intact rock.

It was originally developed in the late 1940s for testing concrete and adapted for rocks from the 1960s onward. In the Isle of Jura study (Smith et al., 2026), an N- type Schmidt hammer provided the rock-strength data used to evaluate fluvial erodibility models.

A spring-loaded mass (piston) is released when the plunger is pressed perpendicularly against a rock surface. The mass impacts the plunger, transferring energy into the rock. The rebound distance of the mass is measured as a percentage of the initial spring extension and reported as the rebound number ( R ) (or

R_L / $R_N$)

depending on hammer type). Harder surfaces produce higher rebound values because less energy is absorbed by deformation or fracturing.

Hammer Types

• N-type: Impact energy 2.207 Nm. Preferred for most field work and stronger rocks (typically UCS ~20-150 MPa or higher). Produces less data scatter.
• L-type: Impact energy 0.735 Nm (one-third of N- type). Better suited to weaker, porous, or weathered rocks and laboratory testing of smaller cores/samples.

Key standards and guidelines:

• ISRM Suggested Method (revised 2008/2009 by Aydin): Detailed recommendations for apparatus, specimen requirements, testing, and data reduction. sciencedirect.com
• ASTM D5873: Standard Test Method for Determination of Rock Hardness by Rebound Hammer Method (applicable roughly for UCS 1- 100 MPa).

Common field/lab protocol steps:

1. Surface preparation — Clean and smooth the test surface if possible (remove loose weathering crust, but avoid artificial polishing that alters results). Test on representative, intact rock free of visible cracks or discontinuities.
2. Orientation — Prefer horizontal impacts. Correct non-horizontal readings using manufacturer charts or analytical formulas because gravity affects the rebound.
3. Number of impacts — Varies by standard and study:
   • ISRM: Typically 20 readings per test location/site. Discard the lowest 50% (or specific outliers) and average the remainder. A series can be stopped early if 10 successive readings differ by ≤ ±2.
   • ASTM D5873: Often 10 impacts; discard those differing by more than 7 from the mean and recalculate.
   • Other practical variants: 6–10 readings averaging all or discarding extremes; some studies recommend ≥15–30 depending on rock strength and heterogeneity.
4. Spacing — Impacts should be spaced sufficiently (e.g., several centimetres apart) to avoid overlapping influence zones.
5. Calibration — Regularly check the hammer on a standard steel anvil. Modern digital models store calibration factors.
6. Environmental factors — Moisture content, temperature, weathering grade, and surface roughness influence results. Dry conditions are preferred for consistency.

In the Jura study, researchers collected a minimum of 20 rebound values per outcrop (total 832 measurements across 35 outcrops), discarded the lowest 50% following ISRM guidelines, and converted the retained average ( R ) values to UCS.

Advantages:

• Fast, inexpensive, portable, and non- destructive.
• Suitable for both in- situ outcrops and laboratory specimens.
• Useful for large spatial coverage (as on Jura) and relative ranking of rock strength/erodibility.

Limitations and sources of variability:

• Measures near- surface properties; may not represent bulk rock if a weathering rind or thin hard crust is present.
• Sensitive to surface conditions, moisture, discontinuities, and grain- scale heterogeneity.
• Data scatter increases outside the optimal UCS range (~20- 150 MPa).
• Correlations to absolute UCS have inherent uncertainty; best used for relative comparisons or when calibrated locally.
• Weak rocks may crush under impact, requiring the L-type hammer or alternative methods.
• Minimum sample size recommendations vary (often 15- 30 readings per site for statistical reliability).

Best practices for geomorphic/fluvial studies:

• Collect many readings across multiple outcrops to capture spatial variability.
• Report hammer type, number of readings, data- reduction method, and conversion equation used.
• Combine with other proxies (e.g., density, fracture density) when possible.
• In erodibility studies, Schmidt- derived strength is often treated as proportional to resistance or used in contrast models (bedrock vs. bedload strength), as done effectively in the Jura research.

Overall, when applied consistently following ISRM or ASTM guidance and with appropriate local calibration, the Schmidt hammer remains one of the most practical tools for rapid rock- strength characterisation in both engineering and geomorphological fieldwork.

_____________________________________________________________________________________

How does lithology influence fluvial erodibility? A study of knickpoint retreat in small bedrock rivers on the Isle of Jura

The study investigates how best to incorporate rock- strength measurements into estimates of fluvial erodibility (K). It uses a suite of coeval knickpoints on the Isle of Jura (SW Scotland) that formed due to base- level fall from glacio-isostatic rebound at ~13.6 ka. These knickpoints have migrated across two lithologies of contrasting strength: hard Jura Quartzite and softer Easdale Subgroup slate.

Rock strength was quantified with a Schmidt hammer. Knickpoint response times were then predicted under different models of erodibility informed by the strength data and compared against the observed knickpoint positions.

Key Methods and Data

• Schmidt hammer measurements across outcrops.
• Mean uniaxial compressive strength (σUCS): ~307 MPa (1σ = 80 MPa) for Jura Quartzite vs. ~123 MPa (1σ = 86 MPa) for the Easdale Subgroup (statistically significant difference, p < 0.001). Quartzite is roughly three times stronger.
• Identification of ~18 knickpoints linked to the 13.6 ka event (first knickpoint upstream of ~35 m elevation, corresponding to the highest raised beaches).
• Three erodibility models tested:
  • Kinvariable — no lithologic variation.
  • Klitho — different erodibility assigned simply by lithologic unit (based on bedrock strength).
  • Kcontrast — accounts for the contrast between bedrock strength and bed-load sediment strength.

The Kcontrast model (bedrock- bedload strength contrast) produced the tightest and most accurate clustering of predicted knickpoint ages around 13.6 ka. It outperformed the other two models.

Notably, the simple lithology- specific model (Klitho) performed marginally worse than the uniform- erodibility model. This indicates that incorrectly accounting for lithology can be worse than ignoring it.

Fluvial erodibility is not controlled solely by bedrock strength. The contrast between the strength of the impacting bedload and the underlying bedrock is a better predictor (consistent with prior theoretical and field work on abrasion).

Simply mapping discrete hard/soft rock units and assigning fixed erodibilities can miss important sediment-mixing effects and may degrade model performance. Incorporating Schmidt- hammer- derived strength data via a contrast formulation offers a practical improvement for landscape- evolution models.

The authors note that discharge and channel steepness remain the dominant controls on erosion; lithologic effects are secondary but significant and worth refining, especially for larger-scale or higher-resolution models.

This work builds on earlier studies of Jura knickpoints (e.g., Castillo et al.) and related research on bedload–bedrock interactions.

Full- text access is typically behind the Wiley paywall (institutional login, purchase, or open-access options may apply depending on your access).

The University of Glasgow press release provides a good public summary of the findings.

Published: Earth Surface Processes and Landforms

DOI: 10.1002/esp.70387

Provided: University of Glasgow

Authors: Adam G. G. Smith (corresponding author, University of Glasgow), Matthew Fox, Oreoluwa Akinwale, Elias J. Rugen University College London and the University of Southampton

Abstract

Bedrock rivers are a key driver of landscape evolution across large parts of Earth’s continents. Though empirical, theoretical and numerical modelling efforts predict that bedrock erosion should be influenced by lithology, and more specifically, rock strength, relating rock strength to fluvial erodibility (K) in the field, has proven challenging. One recent suggestion as to why this may be the case is that fluvial erodibility is not controlled solely by bedrock strength, but instead the contrast in strength between the underlying bedrock and the strength of the impacting sediment carried by the river (bed load). The goal of this study is to examine how best rock strength measurements can be incorporated into estimates of fluvial erodibility. We do this by studying knickpoint migration, which is related to fluvial erodibility, on the Isle of Jura, SW Scotland. The Isle of Jura is an ideal field site as it hosts a suite of coeval knickpoints, associated with base level fall following glacio-isostatic adjustment, that have migrated over two separate rock types of differing strengths. We quantify rock strength on the island using the Schmidt hammer. We then predict knickpoint response times based on different models of fluvial erodibility informed by our rock strength data. The best fitting model of fluvial erodibility accounts for the contrast between bedrock and bed load strength. Interestingly, the model that simply assigns a different fluvial erodibility value for each lithological unit performs worse than a model that does not account for lithology. This finding challenges present practices for incorporating lithology in landscape evolution models.