Developing quantitative predictive models of ignimbrite sheet architecture: implications for hazards at modern volcanoes

Ignimbrites are the deposits of pyroclastic density currents (PDCs) (Figure 1). These are amongst the most catastrophic flows on Earth and are extremely hazardous to humans, livestock and infrastructure. Our understanding of PDCs comes primarily from observations of modern eruptions and studies of ancient deposits, in combination with numerical and analogue modelling to understand the flow dynamics of the current and the sedimentation of ignimbrites.

One of the great challenges in volcanology is attempting to understand the behaviour of the currents during eruption and how they interact with the landscape on and beyond the volcano. Forecasting the flow paths of the PDCs is incredibly difficult. Our understanding of these phenomena is typically based on observations from paroxysmal eruptions at modern volcanoes (e.g. recent eruptions in Guatemala). In such scenarios these small volume eruptions typically focus PDCs down valleys proximal to the volcano, although some of these valleys can be overtopped and the currents emplaced laterally. However, we have a much more limited understanding of PDC behaviour in larger Plinian and catastrophic caldera-forming eruptions.

There have been numerous recent advances on ignimbrite architecture (Figure 2) and the sedimentation of ignimbrites in these eruption scenarios (e.g. Branney et al., 2008; Brown and Bell, 2013; Brown and Branney 2013; Brown et al., 2003, 2004, 2023; Drake et al., 2022; Giordano and Doronzo, 2017; Gooday et al., 2018); however, we know very little about PDC behaviour and ignimbrite architecture at the “basin-scale”. What happens when PDCs move beyond proximal valley systems? Where and how are the ignimbrites deposited and how do the currents vary both spatially and temporally? Addressing such questions is difficult due to the complex flow dynamics of PDCs (e.g. Breard et al., 2016); however, quantitative basin-scale predictive models in combination with advanced deposit architecture schemes used in sedimentology (e.g. Martin et al, 2021; Miall, 1985; Owen et al, 2015, 2017, 2018), as well as stratigraphic forward modelling (e.g. Snieder et al, 2021) represent an innovative method to answer these questions.

Basin-scale models are essential in interpreting ancient sedimentary successions, and reducing uncertainty in assessing geological resources in basins. By studying the deposits in these basins we can gain insights into past terrestrial environments, and changes in climate, tectonics and base level. Due to the interaction of these processes, the resultant deposits vary significantly in time and space and have a multitude of different characteristics. Predictive models enable us to identify common facies and architectures and understand how major geological process changes affect the deposits. Quantitative basin-scale predictive models are in their relative infancy in continental sedimentary successions (e.g. Owen at al., 2018), but these innovative studies have successfully provided a predictive framework that can now be applied to other basins. These models use a ‘systems-based’ approach whereby palaeogeographic models of the basin are developed based on statistical information on key characteristics such as palaeocurrent trends, grain size, channel-body and storey thickness. These quantitative observations provide powerful information on downstream trends compared to more traditional lithostratigraphic approaches.

The architecture of the pyroclastic (and associated sedimentary) units deposited in large-volume eruptions is extremely complex. However, by employing a quantitative systems-based approach to recent to ancient large-volume ignimbrite sheets, we can identify spatial and temporal trends and develop quantitative predictive models of distal “basin-scale” processes, which is unprecedented in these settings. This approach goes beyond more traditional lithostratigraphic and mapping approaches by quantitatively evaluating numerous ignimbrites and effectively considering their interaction with accommodation space and response to changing base level, rather than the flow dynamics and sedimentation of individual ignimbrite sheets. From these data we can then evaluate downstream predictive trends in the deposits. These analogues can then be applied to understanding the hazards posed by volcanism and the response to such catastrophic events at modern volcanoes.

N.B. This is not a means of “predicting” eruptions or accurately forecasting PDCs at active volcanoes, but it may be used to inform potential eruption scenarios.

Click on an image to expand

Image Captions

Figure 1: Small pyroclastic density current, Sakurajima, Japan, Figure 2: Examples of complex ignimbrite architecture, Tenerife


This project will use quantitative field based studies to develop basin/system-scale models of ignimbrite architecture. The work will use ignimbrites on Tenerife and Gran Canaria as case studies to develop the models. Over two field seasons extensive quantitative data will be collected using logs and architectural panels. These data will include measurements of key characteristics such as palaeocurrent trends, grain size, channel body thickness, and channel percentage. A key stage will include the identification of large-scale sedimentary architectures in the ignimbrites and classification using an established hierarchal approach (e.g. Owen et al., 2017). Stratigraphical correlations will be determined where possible. These data will be statistically analysed to determine relationships between the different field criteria and to identify spatial and temporal trends. Field data will be supported by outcrop models collected using imaging software to aid quantitative observations, but also to develop models of deposit architecture. Numerical modelling will be applied to develop forward models of sheet architecture. Some petrographic and geochemical analyses will also be undertaken to aid correlations.

Project Timeline

Year 1

Year 1: Literature review; field data collection; outcrop model development; preliminary data analysis; petrographic and geochemical analyses

Year 2

Year 2: Field data collection; outcrop model refinement; advanced data and statistical analysis; numerical model development; paper 1 (on architecture of ignimbrite systems)

Year 3

Year 3: Final data and quantitative analysis; paper 2 on basin-scale predictive models; thesis writing

Year 3.5

Year 3-3.5 (6 months only): Thesis completion; further papers if relevant.

& Skills

The student will receive expert training, from leaders in the field, in:

1. The identification of a variety of pyroclastic and volcaniclastic rocks in the field, using a rigorous lithofacies approach.
2. Quantification of lithofacies architecture of these bodies through detailed logging and outcrop measurements.
3. Developing outcrop models using these data.
4. Statistical analysis of quantitative data.
5. Optical microscopy and geochemical analysis (XRF), including sample preparation.
6. Presentation and writing skills.
7. Expedition skills (working in extreme environments).

The student will be joining an innovative and multi-disciplinary geology group at the University of Glasgow. The School’s students and academic staff meet regularly for research seminars and discussions, and so the student will be involved in a research active environment. The student will join the Geosphere research cluster at Glasgow, which forms part of the Earth Systems Science Research Group in Geographical and Earth Sciences.

Excellent employability skill training will be provided by IAPETUS2 and the University of Glasgow College of Science and Engineering Graduate School. The project will be of interest to those considering careers in the resource or hazards industries.

References & further reading

Branney et al., (2008): https://doi.org/10.1007/s00445-007-0140-7
Breard et al., (2016): https://doi.org/10.1038/ngeo2794
Brown and Bell (2013): https://doi.org/10.1144/jgs2012-147
Brown and Branney (2013): https://doi.org/10.1007/s00445-013-0727-0
Brown et al., (2003): https://doi.org/10.1017/S0016756802007252
Brown et al., (2004): https://doi.org/10.1007/s00445-003-0321-y
Brown et al., (2023): https://doi.org/10.1016/j.jvolgeores.2023.107845
Drake et al., (2022): https://doi.org/10.30909/vol.05.02.397432
Giordano and Doronzo (2017): https://doi.org/10.1038/s41598-017-04880-6
Gooday et al., (2018): https://doi.org/10.1007/s00445-018-1243-z
Martin et al., (2021): https://doi.org/10.3389/feart.2020.564017
Miall (1985): https://doi.org/10.1016/0012-8252(85)90001-7
Owen et al., (2015): https://doi.org/10.2110/jsr.2015.35
Owen et al., (2017): https://doi.org/10.1111/sed.12364
Owen et al., (2018): https://doi.org/10.1111/sed.12515
Snieder et al, (2021): https://doi.org/10.1111/bre.12597

Apply Now