Statistical Network Inference and
Time Series Analysis of Marine Ecosystems

Investigator: Jacob Bien, USC

Understanding ocean ecosystems critically hinges on our ability to uncover the complex relationships and interactions between living organisms and geochemical properties as they vary across space and time. As the scale and richness of marine data sets increase, the research questions being investigated become more ambitious and therefore more complicated to answer in a rigorous way. It becomes essential to address the methodological challenges associated with analyzing such datasets in a statistically sound and computationally efficient way.

In this project, we will develop statistical tools that can be used to infer the temporal and spatial interactions between large numbers of entities (e.g., microbial species) and the effect of other covariates (e.g. salinity and temperature) on these interactions. Recent developments from the high-dimensional statistics literature on graphical modeling and time series analysis will play a central role in this work. Efficient code will be produced and made publicly and freely available for marine researchers.