Abstract/Summary

This thesis examines the behaviour of Indian monsoon depressions (MDs), synopticscale systems that frequent the monsoon trough throughout the boreal summer. The original contribution to knowledge is a substantial advancement of our understanding on the structure, variability, and dynamics of MDs. A feature-tracking algorithm is developed to emulate the subjective methods used by the India Meteorological Department, based on surface wind speed and surface pressure. This is applied to reanalysis data spanning from 1979 to the present day (2016), extracting 106 events. These are used to build a statistical composite, with which we interrogate the structure and variability subject to external forcings. It is found, for example, that MDs significantly intensify during La Ni˜na and active monsoon spells, and have different spatial structure over ocean than land. These events are also examined in satellite data, where we discover and characterise a bimodal, diurnal cycle in surface precipitation; the structure of cloud type with dominating deep convection south of the centre; and present the first composite picture of vertical hydrometeor structure in MDs. A propagation mechanism whereby the MD behaves as a vortex in the presence of a wall (i.e. the Himalayan massif) is presented and shown to perform better than competing theories at predicting the velocity and heading of MDs. It is also shown, using a case study in the Met Office Unified Model (MetUM), that varying antecedent soil moisture conditions can significantly change the length of a MD track but not its direction. The sensitivity of MDs to changes in horizontal resolution in the MetUM are also explored across seven case studies at eight resolutions. It is found that the intensity is typically slightly overpredicted, correlating with too great a latent heat release in the mid-troposphere. Spatial structure is shown to improve with resolution but improvement saturates beyond N512 (40 km) resolution.