A Maximum Entropy (MaxEnt) algorithm was calibrated with ground data to generate living aboveground biomass (AGB), its associated uncertainty, and forest probability maps for Mexico. The input predictor layers were extracted from Optical and Synthetic Aperture Radar (SAR) imagery, as well as from a digital elevation model. The combination of the three spatial datasets showed superior accuracy and lower relative error (0.31 and 58%) than the use of single dataset (0.12 - 0.19, and 62% - 74%) or two datasets (0.25 - 0.28, and 58% - 59%). The AGB map showed a root mean square error (RMSE) of 17.3 t C ha-1 and R2 = 0.31 when validated with inventory plots. The total carbon stored in forests was estimated to be 1.69 Gt C ± 1%, which agrees with the total national estimations. This new map proved to have similar accuracy as previous AGB maps of Mexico, but to be more representative of the shape of the probability distribution function of AGB in the national forest inventory data. Different forest area masks with similar forest definitions but originating from different sensors are widely-used to constrain AGB retrievals. The use of different forest masks yielded differences of about 24.1 million ha in forest cover extent and 0.36 Gt C in total carbon stocks for Mexico. A forest cover mask derived from the combination of spatial datasets showed higher accuracy (κ=0.83) than alternative masks derived from SAR (0.78) or optical datasets (κ=0.66). This work found an increasing AGB trend with elevation in Mexico, and that the allometric relationship between AGB and canopy height (H) at plot level significantly varies within biomes and across the topographic gradient (p-value < 0.001). As a result, the amount of AGB per unit of H at higher altitudes is higher than at lower altitudes. This has implications in the use of generalised models across large areas such as those seen in the tropical carbon maps (TCMs) (Saatchi et al., 2011b, Baccini et al., 2012). TCMs show large discrepancies when compared to in-situ observations and regionally calibrated maps. The use of a single allometry (vs. regional allometry), and the calibration of the algorithm without taking into account regional variations are the main sources of the discrepancies. Errors up to 74% are found in this thesis when using the continental allometry from Saatchi et al. (2011b) over Mexico. The results show that the variability on forest ecosystems play a key role when mapping AGB at larger scales. Thus, approaches that take into account these regional variations, are the way forward to improve these products.

Immersion mode heterogeneous ice nucleation is the crucial first step in the glaciation of mixed-phase clouds, which have an important but poorly understood influence on global climate. Additionally, immersion mode ice nucleation plays an important role in the cryopreservation of biological material. At present, this important process is not well understood, hindering progress in these fields. In particular, it is not clear what physical and chemical properties cause a substance to nucleate ice well. The first section of this project describes the development and testing of a new droplet 1 μL volume droplet freezing assay. This instrument is a fast and effective tool for evaluating the ice nucleating efficacy of relatively large quantities of a given nucleator compared to instruments that use droplets of a size that are typically present in clouds. The rest of the thesis describes the characterisation of a series of nucleators using this instrument, with the aim of improving understanding of immersion mode ice nucleation. Four types of carbon nanomaterial were investigated; all were found to nucleate ice in the immersion mode. This included graphene nanoflakes, which are among the smallest entities that have been found to nucleate ice. Surprisingly, more oxidised nanomaterials did not nucleate ice more efficiently than less oxidised ones. Following on from previous work which found that feldspars nucleate ice more efficiently than other minerals, it was shown that alkali feldspars nucleate ice much more efficiently than plagioclase feldspars. The structures of alkali and plagioclase feldspars are similar so the large difference observed is surprising. In order to probe the reasons behind these observations we tested a range of alkali feldspar of known microtexture. It was found that those lacking microtexture nucleate ice similarly to plagioclase feldspars showing that a feature associated with microtexture (and therefore not directly related to the chemical or crystallographic structure of alkali feldspar) plays an important role in ice nucleation. Finally, it was shown that ice nucleation by feldspars and quartz is significantly enhanced by the presence of ammonium salts and deactivated by several alkali halides. Some other nucleators were found to be unaffected. This provides a possible route for learning more about the mechanism of ice nucleation by different nucleators.

As the atmosphere warms under climate change, it will hold more moisture. Latent heat released as water vapour condenses provides an important contribution to the atmospheric heat budget, affecting stability and providing complex feedbacks. Consequently, theories for the general circulation of the atmosphere proposed based on dry dynamics may not apply in moist simulations. In order to understand the possible changes to the Earth's atmospheric circulation as the climate warms, a deeper understanding of these feedbacks is required. Changes to the atmospheric thermal structure and circulation as humidity is increased have been explored in an intermediate complexity general circulation model. To provide a reference climate more comparable with that of previous studies, and of the real world, a simple parameterisation of shortwave and longwave radiative transfer has been developed, which compares favourably with existing simple radiation schemes. Experiments have then been performed with fixed optical depths in which the moisture content of the model is varied. In the zonal mean, increasing moisture content results in an increase in static stability throughout the atmosphere. Consequent changes to the Hadley cell, zonal jets, and storm track have been analysed using simple theories, and by comparison with an experiment in which the sea surface temperature in the tropics is increased. This reveals that the majority of the effects of increased moisture content on the circulation are generated by low latitude warming. The simulations further reveal stronger midlatitude poleward transport of moist static energy as saturation vapour pressure is increased, and an unexpected increase in sensible heat transport in the cold sector of storms. A mechanism for the latter is proposed related to the environmental static stability against which the system develops. The experiments also suggest changes to the rate of conversion of available potential energy to eddy kinetic energy as moisture content increases.

The breaking of ocean surface gravity waves is an important phenomenon that affects the dynamics of the upper ocean, development of the wave field, and air-sea exchange processes. As the surface expression of this process, whitecaps provide a visible signature of wave breaking; their areal extent per unit area sea surface - known as the whitecap fraction, W - can be used to quantify the amount and scale of wave breaking. W is traditionally estimated using digital images of the ocean surface and is widely used to represent whitecaps in remote sensing applications, and in the parameterisation of a host of air-sea processes in models. These parameterisations - generally functions of wind speed alone - are based on limited amounts of data, and fail to take into account the known influence of secondary factors on whitecaps. A novel approach to estimating W using satellite observations has recently been developed, based on passive radiometric measurements of brightness temperature at microwave frequencies. The satellite-based approach enables measurement of W on a global scale, and in a variety of conditions. In this work, the basic characteristics of W estimates at two different radiometric frequencies, W10 (10 GHz) and W37 (37 GHz), is investigated. The wind speed dependence, global distribution, and seasonal dependence of the estimates are investigated. Comparison is made against estimates obtained from simple, but widely used, wind speed only parameterisations formulated from in situ data. A more direct comparison of radiometric and photographic W estimates, based on ship-satellite matchups, is also made. Both comparisons indicate that satellite-based W has a different wind speed dependence, resulting in estimates that are, on average, higher at low wind speeds and lower at higher wind speeds than parameterisations formulated from in situ, photographic measurements. On a global scale, this results in satellite-based W being more uniform latitudinally than predictions from traditional formulations. A dataset comprising estimates of W10 and W37, together with collocated and concurrent estimates for a variety of forcing parameters, is used to investigate the the influence on W10 and W37 of secondary forcings, such as the wave field and environmental factors. It is found that on a global scale wind speed describes much of the variability in both W10 and W37 though the influence of secondary factors on W can be appreciable (especially for W37). Based on the magnitude of the influence of secondary forcing factors on W10 and W37, it is concluded that much of the variability in whitecap fraction is likely due to the behavior of the thinner, decaying foam patches, variability that is not captured by the retrieval using the 10 GHz channel. Though whitecap fraction offers a pragmatic approach to inferring the magnitude of processes associated with breaking surface waves, it remains an indirect measure with inherent limitations. More fundamental questions regarding the interpretation and use of W are considered. A dynamical model that relates whitecap fraction to breaking wave statistics is used to illustrate the contribution to whitecap fraction due to whitecaps in different lifetime stages. Such a model provides a framework for better relating whitecap fraction to the dynamic, active part of the wave breaking process which is likely more closely linked to processes such as breaking-induced energy dissipation, turbulent mixing, and bubble-mediated gas exchange. Finally, the implications of use of radiometric estimates for quantifying air-sea processes - specifically production of sea spray aerosol and bubble-mediated gas exchange - is discussed. It is shown that difference between the satellite-based W estimates and those predicted using traditional parameterisations provides an explanation for the consistent geographical biases in sea spray aerosol concentration found in a number of large scale models. The benefit of these novel observations will also extend to predictions of other air-sea processes, and remote sensing applications, that require estimation of W; these benefits will be enhanced if whitecaps and their radiometric signature are more closely related to the physical processes which they are used to quantify.

Recent and future changes to Arctic climate have the potential to impact the region's wildlife, vegetation and the local indigenous communities. In addition, changes in the Arctic impact lower latitudes through the modification of weather patterns and ocean circulation. There is a need to accurately represent the Arctic region on various timescales to predict future climate changes and to produce improved seasonal and sub-seasonal mid-latitude weather forecasts. Both climate and numerical weather prediction models currently perform poorly over the Arctic region, especially in their representation of cloud occurrence, cloud radiative and microphysical properties and the surface turbulent fluxes. The Arctic Ocean Experiment (AOE) 2001 and the Arctic Summer Cloud-Ocean Study (ASCOS) 2008 took place in the central Arctic Ocean during the late summer/early freeze-up period. The aim of both campaigns was to improve the understanding of processes relating to the formation and persistence of low-level Arctic clouds. This study uses data from both campaigns to gain an insight into surface exchange, the structure of the lower atmosphere and cloud formation and then uses this knowledge to evaluate the performance of the Met office Unifed Model (MetUM) over the central Arctic region. The air temperature away from the surface, pressure and wind speed fields are generally well reproduced by the model, suggesting it captures the large-scale circulation with good accuracy. A significant problem is however, found in the model's temperature dependent albedo parameterisation scheme. Due to an underestimation of the model ice surface albedo, too much radiation is absorbed at the surface, which causes the surface temperature to be too high. This causes a feedback of errors that locks the albedo at its minimum value of 0.5 and the surface temperature at 0 C for most of the observation period. The model also significantly overestimates the magnitude of the surface turbulent fluxes. This is shown to be due to the use of a value for the roughness length for momentum, z0 that is too large and the application of Monin-Obukhov similarity theory under the observed conditions. The measurements show that the boundary layer was almost always less than 200 m deep; this means that the constant flux layer was always less than 20 m deep and often extended to only a few metres above the surface. Spectral analysis of the turbulence measurements shows that turbulent properties differ between the upper (30.60 and 15.40 m) and lower measurement levels and that the observed boundary-layer depths are a likely explanation for this. The third main error involves the model's representation of the low-level layer of stratus cloud. The modelled clouds are too thin and too low in the model, which was at least partly due to the overestimation of boundary-layer depth and inaccuracies in the structure of the lower atmosphere. A number of sensitivity tests involving the surface albedo, roughness length for momentum and vertical grid resolution are performed to refine these conclusions and investigate possible solutions. Several recommendations for improvements to the MetUM and for further research are also presented.

Clouds are a critical component of Earth’s climate and hydrological cycle. The formation of ice in the atmosphere, especially at lower altitudes, can substantially impact the evolution of clouds and their radiative properties, and represents the initiation of the cold rain precipitation process. In mixed-phase clouds quantitatively understanding the interactions between ice and liquid, and the subsequent impact on the cloud development, is fundamentally dependent on the process of ice formation and its representation within cloud models. Experiments show that ice nucleating particles (INPs) exhibit variability in both freezing efficiency and time-dependent behaviour. The variability in freezing efficiency is currently well characterised and represented, but variability in time-dependence is poorly characterised and rarely represented in models and parameterisations. The primary aim of this thesis is to understand the role that time-dependence plays in the freezing behaviour of droplets, and secondly to examine the sensitivity of mixed-phase clouds to time-dependence in immersion mode freezing. It is initially found that CNT-based models are unable to reproduce the observed time-dependent behaviour. A new model is therefore presented that uniquely incorporates the variability in both freezing efficiency and time-dependent behaviour; this is applied to experimental data to understand the manifestation of time-dependence in experiments. The model is then used to derive a new theoretical framework for use in experimental analysis and cloud modelling studies. The framework is underpinned by the finding that the temperature dependence (named λ) of the nucleation rate coefficient solely determines the time-dependent behaviour observed in droplet freezing experiments. New and existing experimental data is used to demonstrate the ability for the framework to reconcile data obtained on different timescales with different experimental methods. Finally, an efficient and representative parameterisation is used to explore the sensitivity of mixed-phase clouds to time-dependence. Using a series of increasingly complex models (0D to 2D) it is shown that the inclusion of time-dependence impacts cloud properties in regimes where the updraught speed is relatively low.

Narrow potential vorticity (PV) intrusions across the tropopause (PV streamers) are interconnected with surface high- and low- pressure systems and frequently present during high-impact weather in the mid-latitudes. They often form during the breaking of Rossby waves (observed as large meanders of the jet stream), and occur at the end of the wave life cycle. Extreme weather events are likely to increase in the future due to global warming and climate change, so, enhanced knowledge about the main infuences on these events is crucial for early warning systems. This thesis explores the location and relationship between upper- (troughs and PV streamers) and lower- (cyclones and anticyclones) level features impacting the United Kingdom during heavy precipitation days. It then investigates the genesis of preceding Rossby waves and initiation of PV streamer formation related to heavy precipitation cases. Finally, the predictability of these features is analysed through an assessment of forecasting skill in the European Centre for Medium-Range Weather Forecasts (ECMWF deterministic models as well as the THORPEX Interactive Grand Global Ensemble (TIGGE) prediction system through the use of a novel feature-based error method. Stratospheric PV streamers were found to be present during UK heavy precipitation cases 88% of the time in summer, 85% in autumn, 72% in spring and 63% in winter. They are the dominant in influence on heavy precipitation for the UK in summer as well as in autumn in combination with cyclones. Cyclones are located to the north-west of the UK during all of the seasons with highest frequency in winter and least in summer. Anticyclones also have an impact on UK precipitation by steering systems further north in the winter. The results vary regionally, with western and northern areas characterized by orographic in influences. Eastern Scotland has the most consistent pattern of stratospheric streamer involvement, and the combination of upper-level and orographic effects create uplift of moist air leading to heavy precipitation events. Remarkably, spring has the most variance in the distribution of upper- and lower-level features with evidence of a distinct east/west split across the country. Cyclone and anticyclone pairings dominate in the west, while stratospheric and tropospheric streamer coupling enhances precipitation in the east. Rossby waves preceding the events are triggered from 3 to 7 days in advance, with some seasonal variations. The trigger points range from the Pacific Ocean basin, throughout North America to the western Atlantic Ocean. The waves then proceed across the North Atlantic, where PV streamers are initiated. Triggers located to the west (or behind) the PV streamer lead to enhanced ridges and LC1 or anticyclonic streamer types, while triggers to the east (ahead of the streamer) increase the likelihood of LC2 or cyclonic streamer types. Influences to the east are the most common form of trigger closely followed by streamers forming from recirculated stratospheric air (for example when a parcel of air re-attaches to the stratosphere). PV streamers are generally represented well in short forecast lead times (1-2 days) with a growth in structural and location errors as lead time increases. An interesting result of the feature error method was the identification of significantly lower PV mean and maximum amplitudes (by as much as -3.5PVU) especially in the upper eastern flank of the streamer. This could be due to insfficient influx of high PV air into the streamer and would benefitt from further investigation. In associated heavy precipitation forecasts, TIGGE ensemble members with more accurate rainfall prediction have consistently better PV streamer representation than those who under-predicted the precipitation. The evidence indicates that improved understanding and prediction of PV streamers can lead to better predictability of heavy precipitation and thus an enhancement in early warning systems.

Traditionally, hydrologists had a relatively minor role in rainfall data processing; they usually simply took data from meteorologists. However, meteorological organisations usually provide weather service over a larger area and scale (i.e. country level); the applicability of this large-scale information to urban hydrological applications is therefore questionable. This work tries to provide a local view on rainfall processing, aiming to improve the suitability (in terms of accuracy and resolution) of operational rainfall data for urban hydrological uses. This work explores advanced downscaling and adjustment techniques to address the identified issues in urban hydrology: accuracy and resolution. On the basis of a a review and the testing of state of the art techniques, the Bayesian-based adjustment technique and the newly-developed cascade-based downscaling techniques are found to be suitable tools to improve respectively the accuracy, and the resolution of operational radar (and raingauge) rainfall estimates. In addition, a combined application of these two techniques is tested; the results suggested that, although extra uncertainty may appear, this combination demonstrates a clear potential for providing accurate and high-resolution (street-scale and 5-min) rainfall estimates.

This thesis aims to develop an understanding of the relationship between climate, topography and glacier fluctuations in Iceland. A mass balance/glacier flow model is applied to the Holocene fluctuations of Sólheimajökull, an outlet glacier in southern Iceland. The model is also used to predict the response of Sólheimajökull to future climatic warming. The findings provide insight into the spatial variability of glacier fluctuations in Iceland, and the dynamics of Holocene climatic changes in the North Atlantic. The results from the model suggest that the response of Icelandic glaciers to climatic change can be related to glacier area-altitude distribution. Outlet valley glaciers located in high precipitation areas descend to elevations of 0-100 m where air temperature is mild. Ablation occurs throughout the year and glacier mass balance has a large amplitude response to temperature variations. Furthermore, outlet valley glaciers experience dynamic length variations in response to climatic change. This is a geometric effect where small changes in ice cap volume result in significant fluctuations in glacier lengths. In contrast, wide ice cap lobes in central Iceland exhibit a different response to climatic change. Precipitation levels are lower and glaciers terminate at altitudes of 600-800 m. Ablation is restricted to the summer months, and glacier mass-balance has a lower amplitude response to temperature variations. In addition, ice cap lobes experience smaller dynamic length fluctuations in response to climatic change. This is because ice cap margins undergo small changes in extent to response to changes in glacier volume. Finally, where ice cap lobes terminate on sandur plains, further advance leads to glacier widening and an unsustainable increase in ablation. The numerical model is used to successfully reconstruct Holocene climatic changes over the last 5000 years from the record of glacier length variations in Sólheimajökull.

Prediction is one of the hardest things in ecological science. Predicting the weather is one of the hardest things of all. This is what makes predicting the ecological consequences of climate change so exceptionally demanding. As a first step, we would like to understand the effects of weather variation on the behaviour of those ecological systems for which we have the best long-term data. The Park Grass Experiment at Rothamsted allows us to model the effects of the timing of rainfall and the accumulation of day-degrees in spring on primary productivity in an ungrazed grassland. I use the insights gained from this model to interpret the effects of weather variation in two classic long-term studies of plant-herbivore interactions: the Red Deer on Rum and the Soay Sheep on St Kilda. In both cases, direct effects of extreme weather on animal populations (“killing weather”) turn out to be much less important than weather-driven changes in plant production. Because most of the important effects of weather on animal population dynamics act via changes in food availability, it is the interaction between weather and population density that matters more than anything else, rather that weather effects alone. The same weather that would lead to mass starvation at high population densities, might have no measurable impact on animal performance when numbers were low. The analysis is focused on the following questions: which weather variables are most important; when do they have their most important effects; what effect sizes do they generate; and what is the shape of the relationship between the weather variable and the ecological response variable? The answers to these questions will help to guide subsequent analyses of demography and genetics on these two Hebridean Island systems.