AbstractDownscaling aims to link the behaviour of the atmosphere at fine scales to properties measurable at coarser scales, and has the potential to provide high resolution information at a lower computational and storage cost than numerical simulation alone. This is especially appealing for targeting convective scales, which are at the edge of what is possible to simulate operationally. Since convective scale weather has a high degree of independence from larger scales, a generative approach is essential. I will describe a statistical method for downscaling moist variables to convective scales using conditional Gaussian random fields, with an application to wet bulb potential temperature (WBPT) data over the UK. This model uses an adaptive covariance estimation to capture the variable spatial properties at convective scales.
BioRachel is a senior scientist in the Informatics Lab, working on ways to combine physics-based numerical simulation with machine learning. Her research spans several projects with collaborators in the Met Office and beyond - active projects include emulation of gravity wave parameterisations with ML models; applications of causal inference to climate indices; using spatial statistical models to understand convective-scale variability; and new methods for spatial verification. She is currently undertaking a PhD at the University of Exeter, with a focus on using Gaussian random fields for stochastic super-resolution of convective-scale fields.