CEEDS operates across a range of scientific areas (air, land/soil, water/ice and biodiversity) representing the breadth of environmental science in Lancaster University and the Centre for Ecology & Hydrology.
We seek innovation at the interface between environmental science and data science, in particular seeking to develop novel data science methods in support of novel scientific outcomes.
We build on three important areas of innovation in data science.
We are inspired by the unprecedented amount of data available around the natural environment; from remote sensing by satellites, aircraft or drones through to low-cost sensors widely deployed in the natural environment (cf. an environmental Internet of Things), and also including data provided by emerging areas such as citizen science and data mining. This is balanced by the unprecedented storage and processing capacities offered by contemporary cloud solutions providing an explosion in data and the elastic capacity to manage this data.
We seek data science solutions that support making sense of this highly complex and heterogeneous data and take a deliberately broad perspective on such methods drawing on contemporary statistics (time series analyses, changepoint detection methods, extreme value theory, spatial statistics) through to statistical/machine learning and deep learning. We note the unique challenges of environmental data science and seek methods and combinations of methods that address such challenges. We see uncertainty as crucial in terms of quantifying, propagating and bounding uncertainties and draw heavily of Bayesian methods.
Ultimately our goal is to communicate to a range of stakeholders and to support decision-making and policy. We complement our study of data science methods, with exploration and visualisation techniques to support well-founded mitigation and adaptation strategies in each of our science areas.
We are particularly motivated by environmental challenges that embrace each of these aspects, and that demand integrated, end-to-end solutions from data acquisition through to policy creation.
Ultimately, we are dealing with highly complex systems and hence we place ourselves firmly at the intersection of environmental science, data science and complexity science.
© 2019 LANCASTER UNIVERSITY