Man has long observed the forms and movement of clouds and attempted to account for them. Aristophenes and Lucretius among others have reported contemporary ideas, but a more detailed knowledge of cloud structure was required before much progress could be made towards understanding them. A certain amount can indeed be learned by visual observations The ice or water phases can be recognised and the lifetime and development, in particular of cumulus clouds, can be observed. Time lapse photography and theodolite measurements of cloud tops (Scorer and Ludlam, 1953) can be used to determine the updraughts inside them. Radar can be used to follow the development of clouds over a large area. Such studies have shown that cumulonimbus exhibits a cellular structure. These cells have a characteristic life history (Byers and Braham, 19^9) and orographic features are often found to influence their initiation . The influence of orography can also be studied by means of a fine network of rai n gauges(Bergeron, I96O). Ice and water have different dielectric, constants so the same intensity of precipitation gives rise to a much stronger radar echo if it is in liquid form. The melting zone gives an especially strong echo since the just melted drops have not yet reached their terminal velocity and spread out. Thus examination of the radar characteristics of clouds can reveal their gross structure. However, such important factors as temperature and humidity cannot be found and the detailed structure of the precipitation and a i r movements only with some ambiguity and in limited circumstances. Since the radar echo exhibits a doppler shift if the target is moving a vertical pointing radar can be used to determine the fall speeds of precipitation elements. In a convective storm the terminal velocities of the particles are less than the updraughts, and Atlas (1965) has made use of t h i s to determine the pattern of convection i n a storm. Such an investigation requires assumptions to be made about the s i z e distributions of the precipitation p a r t i c l e s , but radar can be used to determine these i f the updraughts can be neglected as i n nimbostratus conditions (Caton, 1966). Since radar gives the s i z e distributions at a l l heights i t becomes possible to study the growth of the particles as they fall through the cloud. Similar studies can be made at the ground if a number of observing stations are situated at various heights, for example on a mountain side (Magono, 1960). While this kind of investigation i s limited to only a few levels and these below or only i n the lower part of the cloud, it does not suffer from the uncertainties that are attendant on similar radar work and there is also the advantage of being able to determine the local meteorological conditions. Some studies can be made at ground stations that are impossible for radar, for example the size distributions of cloud droplets (Okita, 1962) and studies in the region where ice crystals are melting and their velocities varying (Ohtake, 1965).

Cumulonimbus clouds perturb their local environment in such a way that subsequent clouds may experience dramatically different fields of gas and aerosol species. A full understanding of the behaviour of these species in a convective environment therefore requires examination beyond single-cloud studies. This study uses a cloud-resolving model to examine the evolution of species typical of the remote marine atmosphere over a period of several days and across a domain size that encompasses several cloud systems· at once. Using a series of gas and aerosol sub-models that have been developed in the cloud-resolving model, the full effect of an evolving field of convective clouds on gas and aerosol formation, transport, and eventual removal is examined. The knowledge gained from these simulations will ultimately aid representation in global models where such processes are not explicitly resolved. Sea salt aerosol and dimethyl sulphide gas, or DMS, both exhibit large temporal and spatial variability in the convective environment. This variability is inextricably linked to the 'cold pool' outflow from convective systems. Resolving the interaction between the cold pool and the chemical species is shown to be important for various aspects of the species behaviour. An inability to resolve such features, such as in a global model, leads to inaccuracies in both species concentration and vertical transport. In the case of DMS, the abundance as well as the amount vented to the free troposphere is enhanced significantly in comparison to a global model estimate. Emissions into the domain are greater in the CRM because it resolves the localised intense winds embedded in cold pools. Additionally, secondary convection occurs through dynamical lifting at the leading edges of cold pools. Because DMS concentrations are high in these regions, the ability to resolve such features enhances the transport to the free troposphere by up to 50% in comparison to using a mean gas field, typical of a global model. In the case of sea salt aerosol, it is found that precipitation is co-located with regions of maximum and minimum sea salt. This feature arises due to a build up of sea salt inside the cold pool head, a region wh~re precipitation from secondary cono vection tends to occur. Aerosol is also found to be scavenged in localised regions faster than the precipitation lifecycle itself. An inability to resolve these features is shown to have a significant influence on the amount of aerosol scavenged from the domain. The examination of the marine sulphur cycle around convective clouds shows that an anti-correlation exists between 802 and its precursor gas, DMS. Although concentrations of DMS are high around convection due to enhanced surface fluxes, the efficient scavenging of S02 dominates over its formation. Also, it is found that the concentration of accumulation mode sulphate particles falls dramatically during convection. The reason for this is the dominance of nucleation scavenging events. Additionally, it is shown that although aqueous-phase oxidation of S02 is significant, a large majority of the sulphate generated in this way is precipitated to the ground in rain drops.

Cirrus clouds are thought to have a significant role in atmospheric processes: specifically; their heating/cooling contribution to the Earth's radiative balance, and the consumption of water substance due to their formation. Their presence in the upper troposphere I lower stratosphere (UTLS) also provides a surface for heterogeneous chemistry. The SLIM CAT-Cirrus model is developed to provide a tool to investigate aspects of these properties. SLIM CAT-Cirrus is based upon the existing SLIM CAT chemistry transport model and a parameterisation of the formation of cirrus ice by homogeneous nucleation. The advantages and drawbacks of the use of legacy models are discussed especially issues regarding the loss of the underlying decision-making regarding design approach, approximations, and assumptions. Techniques adopted and adapted from the software engineering and QA disciplines are used to mitigate these problems and maintain future traceabilty; this takes the form of examples of practical measures that small groups or individuals researchers can use. The difficulty in validating a complex global model in the absence of a definitive reference has been addressed by using diverse measurement data sources, and a suite of statistical merries. Model verification testing is also used to characterise processes that are difficult [0 validate. Validation of the modelled frequency of cirrus occurrence against satellite data showed an initial under-prognosis by the model. To address this a statistical scheme has been devised to reproduce some of the effects of phenomena such as gravity waves that are not resolved by the model grid. The modelled effects of the formation of cirrus on the water budget in the UTLS are comparable with measurements from the HALOE (HALogen Occultation Experiment), and are also in-line with the drying effect cirrus are thought to have on air entering the stratosphere. The radiative effects of cirrus have been represented using specific cirrus radiative parameterisations. The cirrus heating shows positive feedback into vertical transport causing meso-scale uplift of the kind thought to be responsible for part of the BrewerDobson atmospheric circulation.

It is generally accepted that ice clouds are of significant importance when considering the Earth's atmosphere with regards to the radiation budget and precipitation processes. This thesis explores three aspects of ice cloud processes: ice nucleation, ice aggregation and ice development inside growing cumulus congestus.

This thesis presents novel uses of routinely measured dual polarisation radar variables to improve our understanding of the microphysics of mixed-phase clouds and rainfall. Fundamentally, a new variable L = - log10(1-Ρhv) is defined, which has preferable statistical properties to the co-polar correlation coefficient (Ρhv) and allows rigorous confidence intervals on Ρhv to be derived. The use of this variable also removes biases introduced by averaging many Ρhv samples and allows, for the first time, Ρhv to be used quantitatively. An emphasis is placed on how these variables can be used to retrieve microphysical information in embedded mixed-phase regions, which are particularly poorly understood at present. Using a combination of differential reflectivity (ZDR) and differential Doppler velocity (DDV), new statistics of the frequency of occurrence of mixed-phase clouds are presented. During a 3 month observational campaign, it is estimated that embedded mixed-phase clouds occur 26% of the time. A technique to remove the ambiguity of interpreting ZDR measurements when pristine oriented crystals are present amongst larger aggregate crystals is also presented. By combining L and ZDR, the contribution to the radar signal from pristine oriented crystals (C) and their \intrinsic" ZDR (ZPDRI) that would otherwise be hidden is retrieved. The results show that elevated ZDR above the melting layer was typically the result of pristine oriented crystals with ZPD RI between 3 and 7 dB, with varying contributions to the radar reflectivity. The retrieval provides an insight into the microphysics of embedded mixed-phase clouds using dual polarisation radar not before possible. Finally, the possibility of using L to measure the shape parameter (µ) in the gamma drop size distribution to improve rain rate retrievals is also investigated. It is shown that including drop oscillations is essential for this application. In a convective rain case study, µ appears to be substantially larger than 0 (an exponential DSD), and would result in an overestimated rain rate by up to 50% compared to if a simple exponential DSD is assumed. The potential to retrieve µ with operational radars is also discussed.

Changes to the hydrological cycle under global warming are expected to involve both intensification of the hydrological cycle and changes in the location of key features. Thus, differences between General Circulation Models (GCMs) in simulated local precipitation changes may arise in part from location biases in the models' mean climates. Combined with the large internal variability and short spatial scales of precipitation, as well as the limited observational record, such biases would exacerbate the difficulty of robustly evaluating forced regional precipitation changes. While many techniques exist to correct biases in local intensity distributions, few have dealt with errors in location. In this thesis, therefore, we make use of a technique from medical image registration to remove location biases from simulated precipitation fields, and demonstrate that this reduces disagreement on projected changes. We therefore proceed to introduce a tool tailored towards the removal of location biases from GCMs. We use this tool not only to investigate historical precipitation changes - showing that it marginally increases the detectability of the historical anthropogenic forcing on precipitation - but also to investigate the physical origin of precipitation biases. In particular, we compare the transformations found for different GCMs, and investigate how location biases in the West African Monsoon vary across a perturbed physics ensemble. This study enables us to identify a potential mechanism relating the location of the convergence zone to the parameterisation of the GCM. By demonstrating the utility of such techniques, we hope that they will continue to be developed and applied in the Atmospheric Sciences, and that other applications may prove valuable.

In this thesis, the potential to observe the distribution of water vapour (H2O) and ozone (O3) in the tropical upper troposphere and lower stratosphere (UTLS) from an atmospheric limb sounding spectrometer is assessed. Vertical profile data retrieved operationally from observations from the Michelson Interferometer for Passive Atmospheric Sounding (MIPAS) are examined under "clear sky" and thin cloud conditions. In the tropical UTLS, the distributions of these two naturally occurring important climate gases are characterised by sharp gradients through a cold tropopause. This in combination with the high occurrence of cirrus clouds in this region, makes accurate "clear sky" satellite observations difficult to achieve. The impact of clouds on MIPAS measurements is quantified by simulating the effects of tropical UTLS cirrus clouds into MIPAS H2O and O 3 retrievals. It is found that clouds in the MIPAS line of sight can lead to a systematic positive bias in MIPAS H2O measurements and an increase in random retrieval error for both H2O and O3 . It was also found that cloud errors can propagate into the "clear sky" profile above the cloud altitude. Recommendation for improvements to the operational cloud filtering methods of tropical MIPAS H2O and O3 are proposed. Comparisons of MIPAS H2O and O3 screened with the proposed cloud filtering schemes, to correlative in situ and solar occultation data suggest that improved filtering improves the agreement. However, MIPAS H2O and O3 data quality remain limited by the increased retrieval uncertainty near the tropopause. Finally, the seasonal variability of the tropical cloud filtered MIPAS H2O and O3 through the tropical UTLS is examined. Regional analysis reveals possible evidence of upper tropospheric O3 enhancements and monsoon effects. The results show that MIPAS can sufficiently sound the tropical UTLS region.

Quantifying the ice nucleation activity of combustion aerosols is crucial in understanding their impact on cloud properties, and consequently, on climate. This study investigates the ice nucleation abilities of soot and combustion ashes in the immersion mode at conditions relevant to mixed-phase clouds. Some physical and chemical properties of these aerosols, which aided the interpretation of their ice nucleation activities, were also explored. Soot generated from eugenol and n-decane compounds were used as proxies for atmospheric soot from biomass and hydrocarbon combustion, respectively. Combustion ashes studied were coal fly ash (CFA), wood bottom ash, domestic bottom ash, and coal bottom ash. The ice nucleation experiments were performed with droplet freezing assay instruments; the results obtained are summarised below: (1) Eugenol and n-decane soot particles in suspension were found to nucleate ice between -16.5 and -28 ºC. To describe the ice nucleation efficiency of these particles, a singular model of ice nucleation that yields the ice active nucleation sites density (ns) was applied. Both soot types showed ns in the range: 1 - ~ 107 cm-2 at a temperature range of -16 to ~ -28 °C. An estimation of potential ice nuclei (IN) number based on the ice nucleation efficiency of these soot types indicated that soot substantially contributes to primary ice formation in mixed-phase clouds from temperatures below -22 ºC. At ~ - 25 ºC, soot showed a competition with mineral dust IN. This study suggests that soot is an important IN in mixed-phase clouds especially in regions where mineral dust is not a dominant IN. (2) The freezing temperatures for combustion ashes were between -15 to -36 °C. The fraction of droplets frozen showed that the freezing temperatures were in this order of significance: CFA >> wood ash > domestic ash > coal ash. The ns values estimated for all ashes were between 10-2 and 107 cm-2 for freezing temperatures between -15 and -36 °C. The best estimate suggests that combustion ashes can account for global primary ice nuclei number up to 1 cm-3, and this could impact on primary ice formation budget in mixed-phase clouds.

Rain-triggered lahars are a significant secondary hazard at volcanoes where unconsolidated pyroclastic material is exposed to intense rainfall, frequently occurring for years to decades after the initial eruptive activity affects proximal areas or primary hazard zones. Mechanisms of rain-triggered lahar initiation are often inferred from downstream flow observations, whilst rain-triggered lahar risk mitigation typically relies on ground-based flow detection. As a result, increasing knowledge of the physical processes involved in rain-triggered lahar initiation and enhancing the use of underutilised instrumental networks are key areas in the development of lahar risk mitigation techniques. This thesis examined rain-triggered lahar initiation in three primary ways: (i) The field-based examination of factors influencing the nature of the rain-triggered lahar hazard following the April 2015 eruption of Calbuco; (ii) The development of quantitative rainfall simulation experiments examining the effects of grain size distribution and antecedent rainfall; and (iii) The analysis of rainfall data and instrumental lahar records to devise new methods of rain-triggered lahar prediction and forecasting. Parameters identified at Calbuco as dictating the spatial variability and magnitude of the post-eruption rain-triggered lahar hazard included the volume and grain-size of emplaced pyroclastic material, vegetation coverage, pre-eruption ice and snow cover, topography and rainfall characteristics. Subsequent laboratory-based rainfall simulation experiments featuring man-made tephra beds examined the quantitative effects of factors including the grain-size of surface tephra and antecedent rainfall upon rain-triggered lahar initiation processes. Increased surface runoff was demonstrated during periods of heightened antecedent rainfall and as a result of reduced surface grain size. Reduced surface grain size also induced the formation of surface crusts, further enhancing runoff. Real-time telemetered rainfall data have been utilised as an effective basis for the creation and development of lahar forecasting models, with peak rainfall intensity acting as the optimal rainfall parameter for predicting lahar occurrence. The demonstrated increased warning times provided by such real-time predictive models illustrates their value both alongside existing lahar detection networks and as an alternative where such resources are unavailable. The incorporation of antecedent rainfall data has been shown to increase model performance, as has the integration of catchment recovery proxies at locations in periods of eruptive quiescence. The probabilistic models developed within this thesis also facilitate the continuous temporal calibration and adjustment of predictions as the databases used to generate lahar forecasts expand and evolve. The quantitative examination of factors driving rain-triggered lahar initiation processes within this thesis and the development of new techniques of lahar forecasting and prediction provide a platform for enhanced lahar risk mitigation. Further research should aim to integrate quantitative lahar magnitude thresholds into the developed lahar forecasting models, facilitating probabilistic inundation modelling and enhanced inter-location lahar comparisons.

Accurate, stochastic representations of rainfall structures and weather patterns in the space-time dimension are a challenging task. Recently, efforts have been focused on the simulation of large spatial fields, representation of higher-order statistics, simulation of spatial extremes and overcoming the problem of overdispersion - an underrepresentation of inter- and intraannual variance in weather generator simulations. In this dissertation, these issues are adressed by presenting three different multisite methodologies - a 'conventional' rainfall generator using orthogonal Markov chains with Richardson-type separation in event-amount generation (multisite, PXEOF-enhanced orthogonal Markov chain model methodology), a more novel approach using multivariate EOFs to express precipitation in the region as a two-component combination of deterministic evolution patterns and corresponding stochastic amplitude coupled with an autoregressive moving average model (multisite, ARMA-enhanced PXEOF model methodology) and, finally, a multivariable extension for the simulation of four meteorological variables with improved interannual variability on the station level (multivariable, multisite PXEOF-EEOF model methodology). Based on above methodologies, 1,000 to 10,000 years of daily simulated weather for 196 stations (20 stations in the case of the multisite, multivariable framework) in Peninsular Malaysia were generated. Statistical characteristics of the synthetic dataset are examined in comparison with the observational record and comparisons between models are made. Regarding the ARMA-enhanced PXEOF model methodology, the need for an autoregressive model component to improve short-term rainfall dependence is demonstrated and model evaluation is focused on a slightly 'neglected' topic, often missing from model evaluations in the literature - spatial rainfall footprints and areal statistics. For the multivariable, multisite PXEOF-EEOF model methodology, the spatial and cross-variable correlation structure as well as the effect of introducing interannual correlations is investigated in further detail. The thesis concludes by summarizing benefits and challenges of using multivariate EOFs in weather generators and a recommendation for the shift towards a more parsimonious model framework with modular structure is made.