Marine boundary layer (MBL) cloud and fog are ubiquitous over the global oceans, overlying more than 30% of the ocean surface. Due primarily to their large spatial extent, relat ively speaking minor changes in cloud radiative properties have the potential to sub-stantially affect the global radiation budget, and so future climate. Changes to a particular cloud property are often a small residual effect due to many competing processes. Therefore, modelled cloud changes can be associated with large uncertainties, and :MEL clouds were identified in the most recent IPCC report as the largest contributor to the spread in modelled cloud feedbacks . This thesis exploits the existence of long-term (30yr+ ), consistent, global satellite records of cloud extent and properties to evaluate the representation of MBL clouds in the recent Hadley Centre HadGEM-2A global climate model. In general, the position and fraction of MBL clouds is much improved over previous versions of the model. The representation of cloud liquid water path (L\VP), however, still requires improvement, with the model consistently underestimating LWP by as much as 50% relative to microwave-derived observations. Large differences between the retrieval of L\VP from different satellite platforms may contribute to the poorly constrained model representation of L\VP. The final part of this thesis constructs a well-constrained observational climatology of LWP, using microwave and visible-spectrwn derived satellite retrievals, for the West African Stratocwnulus region. For overcast, single layer cloud, latent heat flux (LHF) is found to be the best indicator of observed cloud L\VP compared to other surface and atmospheric measurements of heat and moisture. The advantages of well characterised and accurate measurements from multiple observational platforms for the purposes of model evaluation are emphasised throughout this thesis. ii

No description available.

Infrared cameras combine the high speed imaging power of a digital camera with the ability to acquire data during the day or the night, making them a promising new tool for ground-based atmospheric research. Previous work shows that images from these cameras, with the help of external measurements, can be used to automatically identify cloud over a limited sky area. In this study, cloud detection ability is extended to all parts of the sky hemisphere and the need for any external data is eliminated.

Cloud droplets in the atmosphere have an organic component which has been shown to form a monolayer film at the air-water interface of the cloud droplet and the atmosphere. The process by which a cloud droplet film will oxidise and the persistence or loss of an oxidised organic film from the air-liquid interface of a cloud droplet is not well quantified. To determine the surface properties of a cloud droplet film during atmospheric oxidation a measurement of a kinetic variable, the concentration of material comprising an organic film, during reaction with atmospheric radicals was required. The coupling of a Langmuir trough with a neutron reflectometer allows the measurement of the surface coverage of a monolayer in unison with measurement of the monolayer surface pressure. The technique of coupling a Langmuir trough with neutron reflectometry is used extensively for research into the properties of surfactants for industrial and medicinal use. This thesis builds on the work of King et al., (2009) and King et al., (2010), whom produced the first neutron reflectivity measurements of atmospheric proxy monolayers reacting with ozone. This is the first thesis detailing the neutron reflectometry measurement from an atmospheric perspective. Measurements were taken of representative fatty acid molecules which have atmospheric relevance (stearic acid, oleic acid and methyl oleate) as well as measurements of phospholipid molecules which are potential parent species for the fatty acids found in atmospheric waters (I ,2-dipalmitoyl-sn-glycero-3-phosphocholine). The mono layers were reacted with aqueous phase OH radical and with gas-phase ozone to assess the kinetics of the oxidation of the monolayers at the air-water interface.

The cloud climate feedback has long been recognised as one of the greatest sources of uncertainty of future climate predictions. In order to narrow this uncertainty models should be evaluated in terms of cloud properties and their radiative sensitivity to those properties. Over recent years methods have been developed to retrieve cloud properties (ice and liquid water content), therefore allowing this evaluation. This thesis focuses on model validation by using retrieved cloud property information, from active remote sensing instruments at Lindenberg in Germany, as input to a radiative transfer scheme. The radiative budget of the atmosphere is simulated and can be used as a tool to assess model cloud representation and its radiative response to cloud properties. First, the radiation scheme is assessed under cloud-free conditions. The clear sky top of atmosphere (TOA) radiation budget can be modelled with an average difference of 15 W m -2 in the shortwave (SW) and 5 W m -2 in the longwave (LW) of observations. The surface radiative ftuxes are simulated with an average difference of 30 W m -2 in the SW and 12 W m-2 in the LW. After testing the sensitivity of the radiation simulations to cloud property retrieval errors, the radiation budget in the presence of cloud is calculated. Cloudy sky simulations agree with observations with an average difference of 10 W m-2 at the TOA in both the SW and LW, and at the surface, 1 W m -2 in the SW and 6 W m -2 in the LW, although the variability is larger than this indicates. A positive bias is found in the OLR simulations which is partly attributed to un-observed thin ice clouds, which consist of small particles that are beyond the sensitivity of the radar. Simulations show that these clouds reduce the OLR by -4 W m -2 , explaining a portion of this discrepancy. The cloud property data and the radiative simulations are used to assess the cloud representation, and cloud radiative response, in the ECMWF model. The temperature dependant parameterisation of mixed phase clouds is compared with observations and found to be unrealistic. The radiative impact of this error reduces the magnitude of the SW and LW simulated radiative effect by approximately 20%.

For simulations of future climate to be reliable, clouds must be represented as realistically as possible. However, most radiation schemes in climate models currently make a number of simplifications when accounting for clouds, one of the most important being the removal of horizontal inhomogeneity (the "plane-parallel" approximation). This has significant effects on the interactions of the clouds with radiation. In this thesis, a new scheme called "Tripleclouds" is presented that attempts to account for the neglected inhomogeneity by using two regions of cloud in each vertical layer of the model as opposed to one. One of these regions represent the optically thinner cloud in the layer, and the other represents the optically thicker cloud. Furthermore, the scheme uses a ''decorrelation'' overlap method that improves on more conventional "maximum-random'' overlap by overlapping vertically continuous cloud with increasing randomness for pairs of layers with larger vertical separation.

Thin and sub-visual cirrus cloud reside at or near the tropical tropopause layer (TTL) with high frequency and have a significant impact on climate due to their effect on the radiation balance relating to the microphysical properties. There is a significant lack of information about cirrus clouds including the conditions that form them as well as how these clouds evolve in time and difficulties in representing these clouds in models with high resolution. TTL observations show an atmospheric phenomenon of very low ice number concentration and substantial in-cloud super-saturations in cirrus cloud (0.005-0.2 cm-3 and RHj up to 130% respectively) which is at odds with homogeneous nucleation only. In this work, we have made use of recent experimental laboratory results (conducted at the AIDA chamber in Germany) which have indicated that aerosol solution particles can become glassy and therefore solid particles when the temperatures are low and similar to TTL temperatures. Glassy aerosols have the potential to form ice particles with much lower freezing thresholds than liquid aerosols need and there is likely an abundance of these particles in the TTL region, unlike typical IN particles. By incorporating the experimental results into the Advanced Particle Simulation Code (APSC), it has been shown that cirrus formed by only homogeneous freezing agrees with observations only for weak updrafts (up to 3 or 4 cm/s) and only for a deposition coefficient, (1, equal to 1.0. The simulations with glassy particles have shown good agreement of modelled cirrus cloud properties with TTL observation when forced by gravity waves whereas liquid aerosol nucleation is at odds with the observations. In this work, it has been shown that using the latest experimental results of glassy nucleation incorporated in a cirrus model shows that glassy aerosol nucleation is a likely explanation for TTL observations.

A new 3D model of light scattering applicable to dielectric faceted objects is presented. The model combines Geometric Optics with diffraction on individual facets yet maintains the low computational expense of standard Geometric Optics. The current implementation of the model is explained and then applied to the problem of light scattering by ice crystals in cirrus clouds. Accurate modelling of the scattering properties of such crystals is crucial to better understanding of cirrus radiative properties and hence to climate modelling and weather forecasting. Calculations using the new model are compared to a separation of variables method and the Improved Geometric Optics method with encouraging results. The model shows significant improvements over standard Geometric Optics. The size applicability of the new model is discussed. The model is applied to a range of crystal geometries that have been observed in cirrus including the hexagonal column, the hollow column, the droxtal and the bullet rosette. For each geometry the phase function and degree of linear polarization are presented and discussed. Ice analogue crystals grown at the University of Hertfordshire have optical properties very close to ice but are stable at room temperature. The geometries of three ice analogue crystals are reconstructed and the single scattering properties of the reconstructions are presented. 2D scattering patterns calculated using the model are compared to laboratory photographs of scattering patterns on a screen created by an ice analogue hexagonal column. The agreement is shown to be very good. By applying the model to a range of geometries, it is shown that the results in the form of 2D scattering patterns can potentially be used to aid particle characterization. By combining the model with a Monte Carlo radiative transfer code, comparisons are made with aircraft radiance measurements of cirrus provided by the Met Office. The improvements over standard Geometric Optics are found to persist following a radiative transfer treatment.

La rétroaction des nuages demeure l’une des incertitudes majeures des modèles de prévision climatique, en particulier les interactions entre aérosols, nuages et rayonnement (IPCC - Boucher et al., 2013). Les nuages sont en effet difficiles à prendre en compte car ils présentent des variabilités spatiales et temporelles importantes. Les mesures de télédétection aéroportées avec une résolution de quelques dizaines de mètres sont très appropriées pour améliorer et affiner nos connaissances sur les propriétés des nuages et leurs variabilités à haute résolution spatiale. Dans ce contexte, nous exploitons les mesures multi-angulaires du nouveau radiomètre aéroporté OSIRIS (Observing System Including PolaRization in the Solar Infrared Spectrum), développé par le Laboratoire d'Optique Atmosphérique. Il est basé sur le concept POLDER et est un prototype du futur instrument spatial 3MI sur les plates-formes MetOp-SG de l’EUMETSAT-ESA à partir de 2022. En télédétection, les nuages sont généralement caractérisés par deux propriétés optiques: l'épaisseur optique des nuages (COT) et le rayon effectif des particules d'eau / de glace formant le nuage (Reff). Actuellement, la plupart des algorithmes de télédétection opérationnels utilisés pour extraire ces propriétés de nuage à partir de mesures passives sont basés sur la construction de tables pré-calculées (LUT) sous l'hypothèse d'une couche de nuage plan-parallèle. Cette méthode est très dépendante des conditions de simulations choisies pour la construction des LUT et rend difficile l'estimation des incertitudes qui en découlent. Au cours de cette thèse, nous utilisons le formalisme de la méthode d’estimation optimale (Rodgers, 2000) pour mettre au point une méthode d’inversion flexible permettant de restituer COT et Reff en utilisant les mesures multi-angulaires visibles et proche-infrarouges d’OSIRIS. Nous montrons que cela permet l'exploitation de l'ensemble des informations disponibles pour chaque pixel afin de s'affranchir des effets angulaires des radiances et d’inverser des propriétés plus cohérente avec l'ensemble des mesures. Nous avons, d’autre part, appliqué le cadre mathématique fourni par la méthode d’estimation optimale pour quantifier les incertitudes sur les paramètres restitués. Trois types d’erreurs ont été évaluées: (1) Les erreurs liées aux incertitudes de mesure, qui atteignent 10% pour les valeurs élevées de COT et de Reff. (2) Les erreurs de modèle liées à une estimation incorrecte des paramètres fixes du modèle (vent de surface de l'océan, altitude des nuages et variance effective de la distribution en taille des gouttelettes d'eau) qui restent inférieures à 0,5% quelles que soient les valeurs de COT et Reff restituées. (3) Les erreurs liées au modèle physique simplifié qui ne prend pas en compte les profils verticaux hétérogènes et utilise l'hypothèse du nuage plan-parallèle homogène et l'approximation du pixel indépendant. Ces deux dernières incertitudes s'avèrent être les plus importantes. / Cloud feedbacks remain one of the major uncertainties of climate prediction models, particularly the interactions between aerosols, clouds and radiation (IPCC - Boucher et al., 2013). Clouds are indeed difficult to account for because they have significant spatial and temporal variability depending on a lot of meteorological variables and aerosol concentration. Airborne remote sensing measurements with tens of meters resolution are very suitable for improving and refining our knowledge of cloud properties and their high spatial variability. In this context, we exploit the multi-angular measurements of the new airborne radiometer OSIRIS (Observing System Including PolaRization in the Solar Infrared Spectrum), developed by the Laboratoire d'Optique Atmosphérique. It is based on the POLDER concept as a prototype of the future 3MI space instrument planned to be launched on the EUMETSAT-ESA MetOp-SG platform in 2022.In remote sensing applications, clouds are generally characterized by two optical properties: the Cloud Optical Thickness (COT) and the effective radius of the water/ice particles forming the cloud (Reff). Currently, most operational remote sensing algorithms used to extract these cloud properties from passive measurements are based on the construction of pre-computed lookup tables (LUT) under the assumption of a homogeneous plane-parallel cloud layer. The LUT method is very dependent on the simulation conditions chosen for their constructions and it is difficult to estimate the resulting uncertainties. In this thesis, we use the formalism of the optimal estimation method (Rodgers, 2000) to develop a flexible inversion method to retrieve COT and Reff using the visible and near-infrared multi-angular measurements of OSIRIS. We show that this method allows the exploitation of all available information for each pixel to overcome the angular effects of radiances and retrieve cloud properties more consistently using all measurements. We also applied the mathematical framework provided by the optimal estimation method to quantify the uncertainties on the retrieved parameters. Three types of errors were evaluated: (1) Errors related to measurement uncertainties, which reach 10% for high values of COT and Reff. (2) Model errors related to an incorrect estimation of the fixed parameters of the model (ocean surface wind, cloud altitude and effective variance of water droplet size distribution) that remain below 0.5% regardless of the values of retrieved COT and Reff. (3) Errors related to the simplified physical model that uses the classical homogeneous plan-parallel cloud assumption and the independent pixel approximation and hence does not take into account the heterogeneous vertical profiles and the 3D radiative transfer effects. These last two uncertainties turn out to be the most important.

La formation et l’intensification des précipitations sont le résultat de différents processus microphysiques menant au grossissement des hydrométéores nuageux. Les caractéristiques macrophysiques de formation et de développement des nuages convectifs fournissent un cadre environnemental qui influence et contraint ces processus microphysiques. L’observation de l’évolution des nuages à haute résolution temporelle permet de mettre en évidence ces processus microphysiques dont la compréhension est indispensable à la prévision à courte échéance des fortes précipitations. Dans ce travail, nous avons mis au point une méthode originale de suivi à haute résolution temporelle (cinq minutes) de systèmes convectifs isolés et associés à de fortes précipitations à partir de l’instrument SEVIRI (Spinning Enhanced Visible and Infrared Imager) embarqué sur le satellite géostationnaire MSG (Météosat Seconde Génération). À partir de plusieurs cas d’étude, une analyse combinée de l’évolution temporelle des propriétés microphysiques (phase thermodynamique, rayons effectifs des hydrométéores), optiques (épaisseur optique) et macrophysiques (ratio périmètre/surface, température moyenne) au sommet des nuages ainsi que l’observation de l’évolution des précipitations au sol ont permis d’identifier sur une période d’intensification des précipitations allant de 30 minutes à 2h, un comportement typique des propriétés au sommet des nuages. Ce résultat fournit une base d'analyse pour la détermination future d’indicateurs précurseurs des fortes précipitations. / Formation of precipitation and its intensity are the result of the microphysical processes that contribute to the growth of precipitating hydrometeors. Macrophysical features of cloud formation and growth provide a framework of environmental conditions constraining microphysical processes. Cloud observations at high temporal resolution can highlight microphysical processes to better understand them which is essential for short-term prediction of heavy rainfall events within the context of nowcasting. A new methodology to track convective cloud systems with a time resolution of five minutes was developed from SEVIRI (Spinning Enhanced Visible and Infrared Imager) on board Meteosat Second Generation (MSG) geostationary satellite. A temporal analysis of cloud top microphysical properties (cloud top phase, cloud effective radius), cloud optical properties (cloud optical thickness), cloud top macrophysical properties (perimeter to area ratio, mean temperature) and ground based precipitation estimation was conducted. Particular temporal trends of cloud top properties was observed over a 30 minutes to 2h rainfall enhancement period. This result provides an analytical basis to determine future precursors of heavy rainfall events.