PhD studentship, ecological impacts of future fire activity, Reading, UK

Apostolos Vougarakis and Sandy Harrison have a fully-funded PhD on "Predicting the ecological impacts of future fire activity on a global scale" currently being advertised as part of the QMEE (Quantitative Methods in Ecology and Evolution) Centre for Doctoral Training at the University of Reading, UK.

Description

Wildfire is the most important type of natural disturbance impacting the terrestrial biosphere. The nature of the fire regime affects vegetation on multiple space and time scales: small fires can trigger gap dynamics at a local scale, while the expression of fire-related trait syndromes is strongly controlled by the frequency and intensity of large fires. The prevalence of wildfire is also implicated in the maintenance of specific vegetation types, notably savannas and grasslands.

The incidence of wildfire is influenced by multiple factors, including long-term climate conditions, short-term weather, vegetation type and productivity, and human factors affecting land-use. These factors may have different effects on different aspects of the fire regime: human ignitions affect the number of fires and their seasonal distribution, for example, but have little impact on fire spread and the total area burned. This complexity means it is difficult to predict the impact of future changes in climate, climate-induced changes in vegetation, and human activities on fire regimes solely using empirical evidence, and the situation is further complicated because there are multiple feedbacks between these different controls.

Coupled fire-vegetation models are the only way of predicting future changes in large-scale fire regimes and exploring how these will affect regional vegetation. However, although several such models have recently been developed, there are large differences in their past and future predictions1 and uncertainties in model structure and parameterisation that make it hard to know whether modern fire regimes are correctly simulated for the right reasons.

The overarching goal of this project is to improve modelling capacity to address the ecological impacts of future change in wildfire. You will use advanced statistical techniques, including generalized linear and mixed-effect modelling, with remote-sensing observations (e.g.2) to explore what controls different aspects of wildfire regimes (numbers of fire starts, fire type, fire seasonality, frequency and intensity, burned area, and emissions) and whether these controls vary regionally. You will use these analyses to inform the design of "perturbed-parameter" experiments with a simple global fire model (INFERNO3). In these experiments, key model processes are specified using a plausible range of possible parameter values to investigate which uncertainties have the largest impact on wildfire regimes. These analyses will help quantify uncertainties in projections but will also lead to improvements to the fire model. In the final step of the project you will couple INFERNO to a simple model of vegetation productivity (P4) to investigate how future change in climate and other fire controls might affect fire regimes, the expression of fire-related recovery traits (resprouting, serotiny, re-seeding) and the balance between grass and woody cover (and thus the extent of savannas, forests and grasslands).

References

1) Andela et al. (2017), Science, doi: 10.1126/science.aal4108.
2) Bistinas, I, Harrison, SP, Prentice, IC et al. (2014). Biogeosci., doi:10.5194/bg-11-5087-2014
3) Mangeon, S, Voulgarakis, A et al. (2016), Geosci. Model Dev., doi:10.5194/gmd-9-2685-2016
4) Wang H et al. (2016), bioRxiv, doi: 10.1101/040246

Logistics

Pi Name: Apostolos Voulgarakis; Pi2 Name: Sandy Harrison; Pi3 Name: Prof. Colin Prentice, Department of Life Sciences, Imperial College London, This email address is being protected from spambots. You need JavaScript enabled to view it.

Expertise

Development of mathematical theory, Computing, Quantitative data analysis, Ecological observations / data collection.

Role requirements

Manipulating EO/model data requiring data management skills; multivariate analysis (ordination, PCA, RDA), statistical techniques (GLMs, mixed-effect models), perturbed-parameter analysis, applying model diagnostics (e.g. Gleckler & Taylor diagrams); process modelling (Fortran, Python). It involves developing new theory and quantitative methods, and new applications for existing theory/methods. Perturbed parameter ensembles have never been designed and performed in fire modelling to systematically explore uncertainty and model diversity. Furthermore, the range of the statistical tools proposed to be used in conjunction with a suite of state-of-the-art satellite and ground-based observations is unique and will provide valuable new insights.

The project addresses key questions in fire ecology: how fire influences reproductive trait expression and vegetation distribution. We will study fire as a process structuring plant adaptation and vegetation competition.

Improved understanding of fire impacts on ecosystems will lead to better ecosystem management. Since fire algorithms are used in Earth system models, our work will have implications for climate modelling, as well as for studying air quality/health issues, all relevant to policy. Better fire modelling will improve predictions of hazards, with implications for the insurance industry.

Current fire models perform poorly and have limited predictive power and this work will develop a new-generation fire model with a strong empirical basis. It will lead to both a better understanding of what controls fire and its ecological impacts and better representation of fire in Earth system models. It will pave the way for including fire-related feedbacks in future climate predictions.

The project applies Earth observation and advanced statistical tools to study fire regimes (all supervisors), perturbed parameter modelling to explore uncertainty (Voulgarakis), and models of fire (Voulgarakis) and vegetation (Prentice) to explore ecological impacts (Harrison). It links process physics, climatology, environmental science and statistics with ecology, at trait and ecosystem levels.

Nerc Relevance

Atmospheric physics & chemistry, Climate and climate change, Ecosystem-scale processes and land use.

Training

The student will receive training in environmental physics and quantitative methods at the Department of Physics (Voulgarakis). Training on fire and Earth system modelling will be provided by Voulgarakis and the Met Office. Harrison will provide training in fire ecology and management and analysis of ecosystem data in her Department. Prentice will provide training in ecosystem processes/modelling. The location will be Imperial College Department of Physics; University of Reading Department of Geography & Environmental Science; Imperial College Department of Life Sciences; Met Office Hadley Centre.

Applications

Read the full position description here: https://mhasoba.pythonanywhere.com/qmee_cdt/default/view_proposals/view/proposals/26

For more information, contact Sandy Harrison: This email address is being protected from spambots. You need JavaScript enabled to view it., Centre for Past Climate Change, School of Archaeology, Geography and Environmental Science, University of Reading, Reading, UK.