Use of artificial intelligence to understand mountain weather and climate processes

Trapped lee waves, and resultant turbulent rotors downstream, present a hazard for aviation and land-based transport. While high resolution numerical weather prediction (NWP) models can represent such phenomena, there is currently no simple and reliable automated method for detecting the extent and characteristics of these waves in model output. Spectral transform methods can be used to characterise regions of wave activity in model and observational data, however these methods can be slow and have their limitations. Machine learning techniques offer potentially fruitful methods for tackling this problem.
This thesis presents the development of deep learning models to detect and characterise trapped lee waves from the characteristic patterns made by trapped lee waves in NWP model output, performing well against hand-labels and spectral-derived characteristics. The deep learning models are applied to a large archive of high resolution NWP model data to produce climatology information for both the present-day climate and a future climate projection. The climatology is interrogated, showing that there is no diurnal cycle of lee waves, but there is a seasonal cycle and influence of synoptic weather effects on lee waves and their characteristics. The future climate projections (under Representative Concentration Pathway 8.5) show little headline change in occurrence or characteristics of lee waves, but imply changes to wave occurrence within different weather patterns, and a risk of more high amplitude (>3 ms⁻¹) waves in the future.
These deep learning models could prove useful for forecasting in the development of a computationally cheap post-processing tool for operational meteorologists, to be able to more easily visualise the effects of lee waves and the potential hazards involved. The climatology information explored in this thesis has informed further understanding about lee waves, such as the weather conditions that result in the strongest lee wave amplitudes, as well as potential changes to lee wave activity under climate change.