Summer precipitation variability is investigated over the England and Wales region. The England and Wales Precipitation dataset (EWP) is an area average estimate derived from rain gauges, taking into account temporal stability and representativeness of the gang<' stations used. The observed daily precipitation distribution is well represented by a Weibull fit. Regional daily precipitation accumulations show large variability over the region, hilt no robust trends are found in mean or extremes over the last 80 years. Comparing the observations with the ERA-Interim short term forecast shows that the ECMWF reanalysis system represents the variability in EWP observations (correlation of 0.91). The shape parameter of the Weibull fit to ERA-Interim precipitation is indistiuguishable from the EWP estimate, but the ERA-Interim estimates are scaled down by 22%. Light precipitation events are overestimated by the ERA-Interim reanalysis. A novel Lagrangian approach is used to investigate monthly precipitation variability during summer for England and Wales. The ROTRAJ trajectory model calculates back trajectories to link precipitation to the origins of water and identifies the evolution of th(' moistun' content of trajectories along their path towards England and 'Vales. Precipitation estimates in the ROTRAJ trajectory model are calculated using specific humidity changes along trajectories in the last 6 hours prior to arrival. The ROTRA'} estimate has similar skill to the ECWMF short term forecast for moderate and extn'llle events (4mm/day), but tends to overestimate the observed light precipitation events (4mm/day) more strongly than the ECI\IWF forecast when compared to EWP.

A vertical wind tunnel has been constructed inside a cold room to simulate the fall of a frozen water drop from the 0 C level in a thundercloud. It was possible to freely support an ice sphere clear of the sides but the particle crashed early in the melting process. The ice spheres were frozen onto a 120 pm diameter platinum wire and during the final stages of melting the particle hung from the wire and was free to rotate about all 3 degrees of freedom. The spheres melted on this type of support produced the same amount of electrification as those melted on the wire loop support used by DRAKE (1968).The charge on the meltwater was found to be always positive and to be highly dependent on the freezing rate, and water drops frozen in still air at between -10 and -15 C produced an order of magnitude less charging than drops frozen in an airstream flowing at 11 ms(^1) at similar temperatures. Examination of the ice particles under a microscope suggested that this effect was due to air escaping from the ice at low freezing rates and smaller air bubbles being formed at high freezing rates. Evidence was found for the enhancement of electrification at high melting rates which DRAKE attributed to the onset of vigorous convection in the meltwater. The effect of carbon dioxide on melting electification was also discussed. The electrification due to melting precipitation under ideal conditions in a thundercloud was estimated as 4 C km(^3) which can be compared to 8 C km(^3) found by SIMPSON and ROBINSON (l94l). It was suggested that the importance of melting ice in thundercloud electrification cannot be established until more information is available on the nature of the solid precipitation, the environment in which it melted and the location and magnitude of the lower positive charge.

This work describes and demonstrates a novel numerical framework suitable for simulating the behaviour of freely falling liquid droplets. The specific case studied is designed such that the properties of the system are similar to those of raindrops falling through air. The study of raindrops is interesting from both an engineering standpoint and from a standpoint of pure curiosity. As a natural phenomenon, rainfall is something which is experienced by everybody, yet its properties are often misunderstood. The primary engineering application is in improving the ability of radar to determine the characteristics of rainfall for meteorological purposes. The significant original contributions to knowledge within this work come from several areas. The numerical methods used are a unique combination of a high order incompressible implicit large eddy simulation method, a conservative level set method, and a pressure projection method. These methods have all been implemented on a highly parallel GPU architecture, with a resulting performance increase of approximately ten times when a single GPU was compared to a single CPU core. The water droplets were simulated in a regime not previously studied by three dimensional methods. The results of these simulations confirmed the validity of the numerical model by reproducing several important experimental results. New insight was then gained regarding the behaviour of droplet wakes, an area with little previous research. The results of the test simulations show great promise for future use of the numerical framework developed. While the simulations todate have been of air-water interactions, there is little reason the model should be constrained to such a system. In theory almost any low speed isothermal interaction of immiscible Newtonain fluids, with length scales of greater than 1mm, could be modeled accurately by these methods.

Numerical weather prediction centres have started assimilating directly cloudy radiances from hyperspectral infrared sounders (e.g. Infrared Atmospheric Sounding Interferometer, IASI), by using channels sensitive to the atmosphere above the clouds. Radiative effects are usually modelled with a simple cloud scheme in which clouds are assumed to be single-layer grey bodies of negligible depth. In particular, one of those techniques makes use of a onedimensional variational retrieval (lD-Var) cloud analysis to provide the cloud parameters that are then utilised to constrain the forward calculations within the global assimilation. In this thesis we investigate an extension of this technique by increasing the complexity of the cloud model to allow a more effective use of the sounding measurements. A new cloud scheme enabling an additional cloud layer to be treated in a ID-Var setting has been proposed. In this scheme the infrared effects are modelled by means of four cloud parameters and are retrieved simultaneously with temperature and humidity profiles. A validation of the new scheme using both simulated and observed IASI radiances showed that the two-layer-cloud representation reduces significantly the bias in the mean profiles of retrieved minus background temperature differences, particularly in multilayer cloud formations. However, the bias is still too large to allow useful assimilation of channels below the cloud. Nevertheless, providing a better estimate of cloud position is valuable as it helps to prevent the assimilation of channels sensitive to the atmosphere below the cloud. Furthermore, the statistical analysis of the retrievals from IASI measurements - complemented by the Advanced Very High Resolution Radiometer cluster information - showed that the model is more accurate in less heterogeneous scenes. This suggests that an a priori assessment of cloud layers present in a scene may be used to constrain appropriately the data leading to a more effective use of cloudy radiances within the full assimilation.

Motivated by the importance of clouds for weather and climate due to their radiative impact, this thesis addresses the current poor representation of thin, stratiform mixed-phase clouds by state-of-the-art numerical weather prediction (NWP) models. Due to the supercooled liquid water present at cold temperatures near the cloud top, mixed-phase clouds strongly influence the amount of solar radiation reaching the surface and have a net cooling effect on the climate. Supercooled liquid water content is underestimated, by a factor of 2 or more, in all 5 NWP models tested and ERA-Interim when compared with ground-based remote-sensing observations of mixed-phase clouds. The ice water content is better predicted, but ice cloud fraction is underestimated. A new ice cloud fraction parameterization is developed to correct this bias, based on radar observations. EMPIRE, a new high-resolution single column model is developed and used to determine the most important processes for maintaining mixed-phase clouds. It is found that altering the model specification of ice particles (size, fall speed, concentration or habit) affected the liquid water content and most also affect the ice water content. A key reason why models underestimate liquid is an overestimate of ice growth rate but parameterizing N0 as a function of ice water content based on aircraft measurements leads to a significant improvement. A strong sensitivity to the model vertical resolution is identified. At coarse resolutions EMPIRE produces less than 2% of the liquid water content of high resolution simulations. This is because the coarse resolution model does not resolve the vertical profile of temperature, liquid and ice near the cloud top. By adding a parameterization of the vertical structure of the upper part of the cloud, the resolution sensitivity is largely removed suggesting that the implementation of such a parameterization in NWP models could improve their simulation of mixed-phase clouds.

No description available.

Many hydrological applications require high-resolution rainfall data under scenarios of climate change. This thesis uses numerical climate model output at a coarse spatial resolution to condition simulations of sub-daily rainfall sequences at individual sites. Downscaling techniques based on generalised linear models are employed, along with stochastic models based on Poisson cluster processes. The two model classes are coupled using stable relationships between the properties of observed rainfall sequences at different time scales. It is recognised that projections of future climate can differ widely between climate models and it is therefore necessary to account for climate model uncertainty. A hierarchical statistical model is proposed, and implemented in a Bayesian framework, which provides a logically coherent and interpretable way to describe uncertainty in multivariate sequences of climate model output. A way of dramatically reducing the computing time needed to fit such a model, based on condensing the data via the use of maximum likelihood estimates, is also discussed. The ideas are illustrated by considering the generation of future daily rainfall sequences at sites in the UK, using climate model outputs under the SRES A2 emissions scenario.

A study on rainfall attenuation at 11 and 36GHz and on prediction of rain-attenuation statistics at 11, 20 and 36 GHz from point rainfall measurements is presented. Emphasis is given on the effects of raindrop size distribution and the application of a "synthetic storm" model, respectively. A "generalised gamma" distribution introduced by Deirmendjian is used for evaluating the ratio of 36.5GHz to 11 GHz attenuation, as a function of the variance of the distribution and some of the theoretical values are compared with those predicted by known drop-size distributions. Attenuation measurements simultaneously recorded at 11 and 36.SGHz (Bradford link) are compared with Laws and Parsons' drop-size distribution and are included between theoretical boundaries obtained from "generalised gamma" distribution. It is shown from this dual frequency experiment that Laws and Parsons' distribution represents on average the rain at moderate attenuations whereas at low attenuation values the actual rain consists of drops smaller than those accounted for by Laws and Parsons. Six months' statistics of rain-attenuation at 11 and 36.5GHz demonstrate that the expected traction of raining time that attenuation exceeds a specified value follow a log-normal distribution. Attenuation statistics measured from the 13.6km Bradford link at 11 and 36.5GHz are found to be in good agreement with statistics predicted from a technique based on a "synthetic storm" model. Attenuation predictions for a number of in-line partially overlapping links in East Anglia indicate the application of this technique in storms of arbitrary orientation and in short and intermediate path lengths. For relatively short links, good estimates of attenuation/time profiles, for individual events, are obtained whereas improved event correlations can be seen on longer links in case of storms moving almost parallel to the link. It is also shown that the use of individual storm velocities is not more advantageous (in terms of prediction accuracy) than the use of an average storm velocity.

A case study of the 7 convective storms in S.E. England on 9 July 1959 as seen by radar indicates that the individual convective cells comprising the most intense storms persisted the longest and were the most highly organized, new cells forming intermittently on the right flank of existing cells. The structure and evolution of one of these cells is analysed in particular detail. This was a large, intense and persistent cell which produced widespread large hail in the Wokingham area. Its longevity indicated the existence of a very persistent circulation, and this is demonstrated to have changed from the usual pulsatory kind of circulation to a quasi-steady one during the 30 minutes for which it was at the height of its intensity. Certain characteristic features exhibited by the radar echo throughout this period form the basis of the derivation of a semi-quantitative three-dimensional model of the air-flow within the cell, in terms of which the growth of the observed hail is satisfactorily explained. The air-flow within a section along the direction of travel of this cell resembles that often drawn to represent an ordinary cold front: it differs from that in previous storm models mainly in that it implies that in the presence of adequate wind shear the up-and-down-draughts can be maintained continuously without serious interference from opposite sides of the storm. The importance of the shear is supported by the fact that on this occasion the latent instability was slight but the shear strong, whereas on an earlier occasion of considerable instability and little shear there were only short-lived and unorganized storms with scattered bail. The growth of hailstones is treated in detail and in particular it is shown that their layer structure is consistent with growth within a steady updraught provided there are small fluctuations in the effective water content. Assuming an exponential spectrum of hail size within the updraught and charge separation by the splintering mechanism it is inferred that the region of strong updraught associated with the growth of the large hail was responsible for a negligible proportion of the total charge separated.

This thesis describes experimental investigations of two mechanisms of charge generation in the light of the requirements of a thunderstorm electrification theory. Attempts to measure the electrification produced by the melting of ice and snow known as the Dinger-Gunn effect have been made with three different experimental arrangements. In one of these, snow crystals were grown and melted under very clean conditions inside a diffusion cloud chamber. The results of the three experiments are consistent in failing to detect a separation of charge and in indicating that any charge produced was 4102 esu/gm or two orders of magnitude smaller than reported by Dinger and Gunn. This result was unaffected by varying the purity of the ice, its air content and the rates of freezing and melting and the pH value of the water. Accordingly, it is concluded that the charging associated with the melting of snow or hail is unlikely to be an important feature in the electrification of clouds and precipitation. An investigation of the distortion and break-up of freely falling water drops in electric fields in the range from 300 V/cm to 1500 V/cm is described. Large drops of water were produced and allowed to break up whilst freely falling in air. The drops were photographed during break-up 5 in an electric field by a high-speed tine camera and the electric charge carried by the fragments was measured. Factors influencing the break-up of water drops in steady and transient air streams are discussed and the critical conditions for rupture and deformation deduced. A theory for the expansion rate of a bubble formed by a freely falling water drop is discussed in the light of the experimental results and tentative suggestions made for the improvement of the theory. Charges separated by water drops breaking up in a field of about 300 V/cm are found to be about 1 esu/gm or two orders of magnitude larger than those used in Simpson's thunderstorm theory. The sign of the charge separated reverses as the field is reversed and the magnitude is proportional to the electric field. The mechanism is not very sensitive to impurity in the water. Accordingly, it is concluded that electrification due to the break-up of raindrops in an electric field may be important as a secondary mechanism in a thundercloud.