PhD Scholarship in Deep Learning for Turbulence - New Zealand
The Centre for Atmospheric Research at the School of Earth and Environment, University of Canterbury, New Zealand is pleased to announce the availability of 1 PhD student scholarship funded by Royal Society of New Zealand. This PhD project is part of a larger research program that is aiming to better understand and model near-surface atmospheric turbulence.
PhD Project:Deep learning applications to study turbulent structures in the atmospheric boundary-layer
Starting date: Successful candidates to start their PhD projects in 2022 or earlier
Scholarship: NZ $27,000 stipend per annum plus a domestic tuition fees waiver for 36 months (excludes student services fee and insurance)
Conference travel and registration costs are supported for at least one international trip
Journal paper publications from student research outputs are strongly encouraged and financially supported
We are looking for a highly motivated and research driven candidate who has experience in machine learning, preferably with deep learning architectures and applications (examples on page 2). The candidate will be working with a multidisciplinary team to develop scientific concepts in the area of deep learning applications for atmospheric turbulence. Candidates with good atmospheric boundary-layer meteorology and/or data science backgrounds are strongly encouraged to apply.
Research program scope: Our research group is expanding in new areas of physics-informed data science. Specifically, we are interested in advancing our spatiotemporal understanding of atmospheric surface-layer turbulence and how it interacts with roughness boundary layers like urban canopies, vegetation and forests. Large Eddy Simulations (LES) have been the state-of-the-art for the simulation of ABL turbulence, and our group has extensive experience in this area along with field-scale observations of turbulence. More recently, we have developed high speed thermography as a spatiotemporal measurement technique to scan and derive surface turbulence associated with wind-surface interactions. Our objective is to combine our knowledge of the atmospheric boundary layer and the state-of-the-art deep learning models through physics-informed deep learning, and learn/predict/enrich near-surface turbulence information from the observational and LES data.
The PhD candidate will work with a team of atmospheric scientists, data scientists, and experts in machine learning. The University of Canterbury is collaborating with international partners from the Institute of Earth and Environmental Sciences at the University of Freiburg, Germany, and the Department of Environmental Sciences, at the University of Basel, Switzerland. As a result, we have accrued a large data library of atmospheric turbulence observations using high speed time sequential thermography of various landscape surfaces, data from traditional sonic anemometry, and data from large eddy simulation for urban and vegetated canopies.