University of Leeds
Using perturbed parameter ensembles to explore the behaviour of mixed-phase cold-air outbreak clouds
Abstract
dc:description.abstractCloud feedback remains one of the largest uncertainties in estimating climate sensitivity and future climate change, with the representation of mixed-phase clouds playing a key role in recent assessments. Mixed-phase clouds contain both liquid droplets and ice particles, and their properties are strongly controlled by microphysical processes that determine the formation and evolution of these hydrometeors. Aerosol-cloud interactions through cloud condensation nuclei (CCN), primary ice production through ice-nucleating particles (INPs), and secondary ice production (SIP) can strongly influence mixed-phase cloud properties. However, these processes are often investigated and evaluated individually, and their effects on the responses of mixed-phase clouds to warming remain poorly quantified. This thesis aims to advance understanding of how aerosol-cloud interactions and ice formation processes influence mixed-phase clouds and their responses to warming using perturbed parameter ensemble (PPE) simulations with a high-resolution regional model. Rather than assessing mixed-phase clouds globally, this thesis focuses on marine cold-air outbreaks (CAOs), which provide strongly forced meteorological conditions with extensive mixed-phase cloud formation. In Chapter 2, following the introduction, the thesis first investigates the sensitivities of mixed-phase CAO clouds to droplet number concentration, INP concentration, and the efficiency of the Hallett-Mossop (HM) SIP process using a one-at-a-time approach in two contrasting CAO cases. The results show that these sensitivities depend strongly on environmental conditions, especially cloud temperature, and liquid-ice partitioning in the clouds. Increasing INP concentration reduces cloud cover and albedo in a warm, liquid-dominated case, but increases cloud cover and albedo in a cold, ice-dominated case. Changes in droplet number concentration and HM process efficiency exert strong influences only in the liquid-dominated case. Building on these findings, in Chapter 3, PPE simulations and model emulation are applied to the liquid-dominated case to investigate the joint effects of droplet number concentration, INP concentration, three SIP parameters, and a parameter controlling liquid-ice overlap within a model grid box. The results reveal strong non-linear interactions among these parameters, indicating that the cloud response to one parameter depends on the values of others. Different parameter combinations can also produce similar cloud properties through compensating effects, leading to equifinality when modelled cloud properties are evaluated against observations. In Chapter 4, the same PPE simulations are evaluated against multiple aircraft observations from the M-Phase aircraft campaign to determine whether observations are able to constrain uncertainties in the model. It is found that different observational constraints identify inconsistent subsets of plausible ensemble members, and constraining one cloud property does not necessarily improve agreement with others. This demonstrates that agreement with a single observed variable does not guarantee that the underlying mixed-phase cloud microphysics are correctly represented. The identified model inconsistencies also point to potential structural weaknesses in the model. Finally, in Chapter 5, emulators trained on idealised high-resolution PPE simulations are used to sample thousands of present-day and future CAO cases, extending the analysis beyond individual present-day case studies. These results show that the processes controlling present-day mixed-phase CAO cloud properties are not the same as those controlling their responses to warming. A key finding is that ice formation processes and future changes in aerosols can strongly affect cloud radiative responses to warming.
Degree
thesis:*- Name dc:type.qualificationname
- Ph.D
- Level dc:type.qualificationlevel
- doctoral
- Grantor dc:publisher.institution
- University of Leeds
- Year dc:date.issued
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Huang, Xinyi
- Advisors dc:contributor.advisor
-
- Carslaw, Kenneth
- Murray, Benjamin
- Grosvenor, Daniel
- van den Heuvel, Floortje
- ORCID person:identifier.orcid
Rights
- Language dc:language.dcterms_iso639_2
- eng
Identifiers
dc:identifier.*- Dc Identifier Grantnumber
- NE/T00939X/1