skip to content

Cambridge NERC Doctoral Training Partnerships

Graduate Research Opportunities
 
Brief summary: 
Using machine learning and ground-truthed methods to estimate the amount of carbon stored in marginal marine ecosystems.
Importance of the area of research concerned: 
The amount of carbon stored in marginal marine environments such as salt and tidal marshes is thought to be globally significant. Anthropogenic climate change and sea-level rise are likely to significantly impact these ecosystems and quantifying the salt marsh carbon budget will help with planning for the future. The aim of this project is to create a unified machine learning model to quantify the carbon stored across global salt marshes, based on the hypothesis that creek density impacts carbon storage in marginal marine environments. Given easily accessible satellite imagery the model pipeline will automatically recognise creek structures and ponds. Subsequentially, it will classify said pond sediment and in-between vegetation, and use this data to predict the overall amount of carbon stored. This approach has been verified through research in the Turchyn group.
Project summary : 
The amount of carbon stored in marginal marine environments is thought to be globally significant, although the total carbon stored in highly dynamic salt marsh sediments, in particular, is poorly constrained (Mcleod et al. 2011; Seyfferth et al. 2020). Anthropogenic climate change and sea-level rise are likely to have a significant impact on these ecosystems and quantifying the salt marsh carbon budget thus will help with planning for the future. The aim of this project is to create a grand unified machine learning model to quantify the carbon stored across global salt marshes, allowing us to accurately predict how the carbon budget may change under anthropogenic climate change.
What will the student do?: 
The primary hypothesis is that creeks that transect the salt marsh influence the flushing of oxygenated water through the salt marsh sediment and thus take away carbon; salt marshes with higher creek density will have lower total storage of carbon in the salt marsh environment. A summer REP in the summer of 2021, and a MSci project in 2021-2022 helped demonstrate this in the East Anglian salt marshes, and demonstrated that the same phenomenon may be true in salt marshes on the East Coast of the US, suggesting a global control on carbon storage. This PhD project will take this to a global scale through a combination of GIS (satellite) work, machine learning algorithms, and fieldwork. Given easily accessible satellite imagery, the model pipeline will automatically recognise creek structures and ponds. Subsequently, the model can be used to classify these pond sediments and in-between vegetation, and use this data to compute and predict the overall amount of carbon stored.
References - references should provide further reading about the project: 
Mcleod et al. 2011; A blueprint for blue carbon: toward an improved understanding of the role of vegetated coastal habitats in sequestering CO2, Frontiers in Ecology and the Environment, https://doi.org/10.1890/110004
Seyfferth et al. 2020, Spatial and temporal heterogeneity of geochemical controls on carbon cycling in a tidal salt marsh, Geochimica et Cosmochimica Acta Vol 282, https://doi.org/10.1016/j.gca.2020.05.013
Hutchings et al., 2019, Creek dynamics determine pond subsurface geochemical heterogeneity in East Anglian (UK) salt marshes, Frontiers in Earth Sciences, Front. Earth Sci., 14 March 2019 Sec. Biogeoscience https://doi.org/10.3389/feart.2019.00041
Applying
You can find out about applying for this project on the Department of Earth Sciences page.
Dr Alexandra (Sasha) Turchyn