Magíster en Meteorología y Climatología / La temperatura en terrenos complejos está controlada por elementos fisiográficos y por flujos que modifican los balances locales de energía, generando fuertes contrastes de temperatura. En esta investigación se analizó un contraste de temperatura entre estaciones ubicadas a similar altitud y un contraste entre laderas, observados sobre la cordillera de Nahuelbuta, centro-sur de Chile. Para esto, se utilizaron datos de temperatura registrados por la red de estaciones AFEX, imágenes MODIS y reanálisis. Además se modelaron las condiciones de mesoescala con el modelo WRF. Los resultados indican que WRF no captura totalmente la diferencia de temperatura entre estaciones ubicadas a similar altitud, pero sí el contraste de temperatura entre el sector sur y norte de la cordillera. Este contraste se originaría por la actividad ondulatoria, resultado de la interacción entre el viento sureste y la montaña. Esta onda está asociada a subsidencia, que sumado al contraste térmico tierra-mar, produce una diferencia de temperatura del aire entre el sector norte, más cálido y el sector sur, más frío. Esta diferencia de temperatura en altura se manifiesta en superficie por el flujo de calor desde estas capas superiores, que se ve favorecido por una intensificación del viento producido por la actividad ondulatoria.

O objetivo deste trabalho foi o analisar o impacto da resolução dos conjuntos de dados topográficos nas simulações das concentrações de dióxido de enxofre (SO2) emitido por uma fonte localizada no Sul do Brasil. Para isso foram realizadas duas simulações aplicando o modelo regional Weather Research and Forecasting acoplado com a química (WRF/Chem), configurado com duas representações do terreno de diferentes resoluções espaciais. Foram utilizados os dados padrão do modelo com melhor resolução, Global 30 Arc-Second Elevation (GTOPO), com aproximadamente 1 km, e inserido no bancos de dados do modelo as informações de terreno em alta-resolução do Radar Shuttle Topography Mission (SRTM) (30 metros). Para as emissões antrópicas do modelo foi elaborado um programa capaz inserir os volumes do poluente SO2 de forma horária expelidos pela chaminé, de acordo com as taxas de emissão medidos diretamente na fonte. O programa representou a emissão do poluente no ponto de grade correspondente a localização e a altura acima da superfície da chaminé da fonte. As simulações foram configuradas com os seguintes esquemas de parametrização: para microfísica de nuvens foi utilizado o Goddard Cumulus Ensemble; os esquemas de radiação de onda longa e curta foram o Goddard e o Rapid Radiative Transfer Model para modelos de circulação geral da atmosfera (MCGA); para a parametrização de cumulus o esquema utilizado foi o Grell 3D Ensemble Scheme; e para os esquemas de camada superficial e camada limite planetária foram utilizados os da teoria da similaridade do Fifth-Generation National Center for Atmospheric Research/Penn State Mesoscale Model (MM5) e o Yonsey University, respectivamente. A escolha desta combinação de esquemas foi definida a partir de um estudo inicial da sensibilidade do modelo à mudança das parametrizações. Os resultados dos experimentos numéricos alterando a topografia foram validados a partir dos dados de monitoramento das estações meteorológica e da qualidade do ar pertencentes à empresa responsável pelo empreendimento associado à fonte. Foi observado que as simulações com os dados SRTM expressaram o terreno da região de estudo mais próximo à realidade, representando o aspecto heterogêneo do relevo, ressaltando os picos e os vales. Os resultados das validações meteorológicas utilizando os dados topográficos indicaram melhoras nas simulações das variáveis meteorológicas: temperatura, umidade relativa, velocidade do vento e precipitação. Os experimentos com os dados topográficos GTOPO e SRTM no modelo WRF/Chem, configurado com as emissões horárias da fonte de Candiota, reproduziram o comportamento dos ventos para transporte de SO2 até as estações de monitoramento conforme os dados observados. Porém foram identificados padrões diferentes na representação das concentrações do poluente entre as duas simulações do modelo, associados aos escoamentos dos ventos representados pelos experimentos. A resolução da topografia afetou na simulação de SO2 devido ao aumento da forçante superficial induzida pelo terreno. Este aumento na forçante, influenciou a advecção da pluma de SO2, resultando em diferentes padrões das concentrações de SO2 no ponto de grade correspondente às estações de monitoramento. Contudo, os resultados das simulações das concentrações de SO2, tanto de forma horária quanto na abordagem das médias diárias, não indicaram uma relação linear entre a utilização de dados em alta resolução e a melhora na representação do SO2 pelo modelo WRF/Chem. / The objective of this work was to analyze the impact of the higher resolution topographic data sets in the simulations of the Sulfur dioxide (SO2) concentrations emitted by a source located Southern Brazil. Two simulations were performed applying the Weather Research and Forecasting model coupled with Chemistry – WRF/Chem, configured with two representations of the terrain with different spatial resolutions. The standard data of the model with the best resolution (approximately 1 km), Global 30 Arc-Second Elevation (GTOPO), and was inserted in the model databases the high-resolution (30 meters) terrain information of the Radar Shuttle Topography Mission (SRTM). For the anthropic emissions of the model, a program was developed capable of inserting the hourly SO2 pollutant volumes expelled by the chimney, according to the emission rates measured directly at the source. The program inserted these emissions into the grid point corresponding to the location and height above the surface of the emission source. The simulations were configured with the following parameterization schemes: for cloud microphysics Goddard Cumulus Ensemble; for the long and short wave radiation treatment it was used the Goddard and the Rapid Radiative Transfer Model for general circulation models; for the cumulus parameterization the scheme it was used the Grell 3D Ensemble Scheme; and for the surface layer and planetary boundary layer schemes, the similarity theory of the Fifth-Generation National Center for Atmospheric Research/Penn State Mesoscale Model (MM5) and the Yonsey University, respectively. The choice of this combination of schemes was defined from an initial study of the sensitivity of the model to the change of parametrizations. The results of the numerical experiments altering the topography were validated from the monitoring data of the meteorological stations and the air quality belonging to the company responsible for the enterprise associated to the source. It was observed that the simulations with the SRTM data expressed the terrain of the region of study closest to reality, representing the heterogeneous aspect of the terrain, highlighting the peaks and valleys. The results of the meteorological validations using the new topographic data indicated an improvement in the simulations of the meteorological variables: temperature, relative humidity, wind speed and precipitation. The experiments with the GTOPO and SRTM topographic data in WRF/Chem model, configured with the hourly emissions of the Candiota source, reproduced the winds behavior that transported the SO2 to the monitoring stations according to the observed data. However, different patterns were identified in the pollutant concentrations between the two simulations of the model, associated to the wind flows represented by the experiments. The topography resolution affected in the simulation of SO2 due to the increase of the surface forcing induced by the terrain. This increase in the forcing influenced the advection of the SO2 plume, resulting in different patterns of SO2 concentrations at the grid point corresponding to the monitoring stations. However, the results of simulations of SO2 concentrations, both hourly and in the approach of daily averages, did not indicate a linear relationship between the use of high resolution data and the improvement in the representation of SO2 by WRF/Chem model.

The objective of this thesis is to assess the optimal power production and flight trajectories of crosswind, ground-generation or pumping-mode airborne wind energy systems (AWES), subject to realistic onshore and offshore, mesoscale-modeled wind data as well as LiDAR wind resource assessment. The investigation ranges from small scale AWES with an aircraft wing area of 10 m^2 to utility scale systems of 150 m^2. In depth knowledge of the wind resource is the basis for the development and deployment of any wind energy generator. Design and investment choices are made based on this information, which determine instantaneous power, annual energy production and cost of electricity. In the case of AWES, many preliminary and current analyses of AWES rely on oversimplified analytical or coarsely resolved wind models, which can not represent the complex wind regime within the lower-troposphere. Furthermore, commonly used, simplified steady state models do not accurately predict AWES power production, which is intrinsically linked to the aircraft's flight dynamics, as the AWES never reaches a steady state over the course of a power cycle. Therefore, leading to false assumption and unrealistic predictions. In this work, we try to expand our knowledge of the wind resource at altitudes beyond the commonly investigated lowest hundreds of meters. The so derived horizontal wind velocity profiles are then implemented in to an optimal control framework to compute power-optimal, dynamically feasible flight trajectories that satisfy operation constraints and structural system limitations. The so derived trajectories describe an ideal, or at least a local optimum, and not necessarily realistic solution. It is unlikely that such power generation can be reached in practice, given that disturbances, model assumptions, misalignment with the wind direction, control limitations and estimation errors, will reduce actual performance. We first analyze wind light detection and ranging (LiDAR) measurements at a potential onshore AWES deployment site in northern Germany. To complement these measurements we generate and analyze onshore and offshore, mesoscale weather research and forecasting (WRF) simulations. Using observation nudging, we assimilate onshore LiDAR measurements into the WRF model, to improve wind resource assessment. We implement representative onshore and offshore wind velocity profiles into the awebox optimization framework, a Python toolbox for modelling and optimal control of AWES, to derive power-optimal trajectories and estimate AWES power curves. Based on a simplified scaling law, we explore the design space and set mass targets for small to utility-scale, ground-generation, crosswind AWESs. / Graduate

Magíster en Meteorología y Climatología / La Cordillera de Nahuelbuta (CN, 37,7°S-73,0°O) es un macizo de 100 km de ancho y 180 km de largo, en donde la precipitación orográfica realiza una importante contribución durante el paso de sistemas frontales conectados a un flujo húmedo del NO. Las observaciones in-situ del proyecto CCOPE (2015), mostraron que más del 30% de precipitación fue generada por colisión coalescencia (lluvia cálida) y asociada a nubes orográficas someras. El resto, ocurrió en nubes profundas, originadas por ascenso a gran escala y mecanismos de crecimiento de hielo (lluvia iniciada por hielo). Dentro de una tormenta frontal, la precipitación puede transitar entre ambas fases y su incremento en la ladera de barlovento de la CN parece ser más marcado en los casos cálidos. Para verificar si dicha transición ocurre en toda la CN y evaluar la distribución de la precipitación instantánea y acumulada en cada fase para un mismo sistema frontal se simularon 5 eventos ocurridos el 2015 con WRF a alta resolución. Estas simulaciones contemplaron 4 configuraciones del modelo, las cuales consistieron en un aumento de los niveles verticales con respecto a la corrida de control, incorporación de un dominio externo y cambio del esquema de microfísica Thompson al WSM5. En 3 de los 5 eventos simulados WRF logró simular la transición entre ambos regímenes de precipitación en toda la CN. En uno de los eventos precipitó nieve y graupel en las cumbres de la CN, por lo cual no fue posible realizar la clasificación, mientras que en otro el modelo no simuló la fase de lluvia cálida que sí se observó en los datos MRR, posiblemente debido a una deficiencia en los datos de entrada del modelo. Los resultados indican que en ambas fases de lluvia existe un control orográfico, pero este es más marcado en la fase de lluvia cálida. En este régimen la precipitación se concentra sobre la CN organizada en bandas a lo largo del viento distribuidas heterogéneamente, con los montos de lluvia incrementándose hacia su ladera de barlovento y zonas más altas, ausentándose en la costa y mar adentro y sugiriendo el desarrollo de nubes orográficas. Dicho incremento (razón de ~1:3 entre la costa y la montaña) en el acumulado total de precipitación se asocia a una mayor duración efectiva de esta (razón ~1:2). Además, las nubes convectivas en esta fase se estructuran en bandas organizadas a lo largo del flujo húmedo. En la fase de lluvia iniciada por hielo en cambio, la precipitación se distribuye de manera homogénea a lo largo de una banda que cruza la CN con orientación NO-SE (asociada al frente frío). La precipitación tiene mayor simultaneidad en toda la CN y el mayor acumulado en la ladera de barlovento de esta se vincula a una mayor intensidad de la precipitación. Estos resultados son comunes a las diferentes configuraciones utilizadas en el modelo WRF, aun considerando variaciones en el esquema de microfísica, demostrando que los hidrometeoros y procesos considerados por aquellos utilizados en esta tesis, no alteran los rasgos distintivos de cada régimen de lluvia. Dichos procesos podrían ser considerados en el futuro como punto de partida para un nuevo método de clasificación de la precipitación líquida.

Atmospheric aerosol particles play an important role in global climate change, via direct and indirect radiative forcing. Elemental carbon (EC) and nitrate are important contributors to anthropogenic aerosol radiative forcing over Europe, since they strongly absorb and/or scatter solar radiation, respectively. However, the evaluation of their climate effects remains highly uncertain. Improvements on the simulation of particle number/mass size distribution (PSD) in modelling will help us to refine model assessments of climate change. The simulations were performed over Europe with a fully online-coupled regional air quality model (WRF-Chem) for the time period of September 10-20th, 2013. Measurements in the HOPE-Melpitz campaign and other datasets in Europe were adopted to evaluate the model uncertainties. The meteorological conditions were well reproduced by the simulations. However, a remarkable overestimation of coarse mode PSD was found in the simulations. The overestimation was mainly contributed by EC, sodium nitrate and sea salt (SSA), stemming from the inadequate emission of EC and SSA. The EC inventory overestimates EC point sources in Germany and the fractions of coarse mode EC emissions in Eastern Europe and Russia. Allocating too much EC emission into the coarse mode could shorten EC lifetime and reduce its long-range transport, thus partly (~20-40%) explaining the underestimation of EC in Germany, when air masses came from eastern direction in previous studies. Furthermore, WRF-Chem overestimated coarse mode SSA mass concentrations by factors of about 8-20 over northwestern and central Europe in this study, due to the shortcoming of its emission scheme. This could facilitate the coarse mode sodium nitrate formation and lead to ~140% overestimation of coarse mode nitrate. Under such circumstances, nitric acid was exhausted, and fine mode ammonium nitrate formation was inhibited. The overestimated SSA shaped the PSD of nitrate towards larger sizes, which might influence the optical properties, lifetime and climate effect of nitrate accordingly. A transport mechanism would broaden the influence of SSA on nitrate PSD to central Europe, where a considerable amount of nitrate precursors and ammonium nitrate is present.:Table of Contents List of Figures List of Tables Abbreviations 1. Introduction 1.1 Particle size distribution 1.2 Elemental carbon particle size distribution simulation 1.3 Chemical pathways for particulate nitrate 1.4 Influence of sea salt on nitrate particle mass size distribution 1.5 Objectives 2. Methodology 2.1. WRF-Chem model 2.1.1. General description 2.1.2. Model configuration 2.1.3 Anthropogenic source emissions 2.1.4 Natural source emissions 2.2 HOPE-Melpitz campaign 2.3 GUAN network over Germany 2.4 Other datasets 3. Results and Discussion 3.1 First publication 3.1.1 Evaluation of the size segregation of elemental carbon (EC) emission in Europe: influence on the simulation of EC long-range transportation 3.1.2 Supporting information 3.2 Second publication 3.2.1 Sea salt emission, transport and influence on size-segregated nitrate simulation: a case study in northwestern Europe by WRF-Chem 3.2.2 Supporting information 4. Summary and Conclusions 5. Outlook Appendix A Bibliography Acknowledgements

Air pollution is the main environmental health issues that affects all the regions and causes millions premature deaths every year. In order to take any preventive measures, we need the ability to predict pollution level and air quality. This task is conventionally solved using deterministic models. However, those models fail to capture complex non-linear dependencies in erratic data. Lately machine learning models gained popularity as a very promising alternative to deterministic models. The purpose of this thesis is to conduct a comparative study between ChemicalTransport Model (WRF-Chem) and a Statistical Model built from machine learning algorithms in order to understand which one is advantageous predicting the air quality and the meteorological conditions using data from Cuenca, Ecuador. The study aims to compare the two methods and conclude on which of them is better in forecasting the concentration of fine particulate matter (PM2.5) in an urban complex terrain. I concluded that even though WRF-Chem has the biggest advantage of forecasting all the data of interest for broader time horizon machine learning algorithms provide better accuracy for middle-term period. Machine learning models also require much less computational power but lack ability to predict meteorological conditions along with pollution level.

As the wind industry is developing, it is asking for more reliable short-term wind forecasts to better manage the wind farms’ operations and electricity production. Developing new wind farms also requires correct assessments of the long-term wind potentials to decide whether to install a wind farm at a specific location. This thesis is studying a new generation of numerical weather forecasting models, named mesoscale models, to see how they could answer those needs. It is held at the company Maïa Eolis which operates several wind farms in France. A mesoscale model, the Weather Research and Forecasting model (WRF), was chosen and used to generate high resolution forecasts based on lower resolution forecasts from NCEP’s Global Forecasting System. The stages for implementation of daily forecasts for the company’s wind farms were: explore and configure the model, automate the runs, develop post-processing tools and forecasts visualization software which was intended to be used by the management team. WRF was also used to downscale wind archives of NCEP’s Final Analysis and determine the possibility to use these in assessing wind potentials. Finally the precision of the model in both cases and for each wind farm was assessed by comparing attained data from the model with real power production.

Windblown dust events can be defined as windblown dust emitted from the Earth's surface to the atmosphere. These events have significant impact on local air quality. Predicting the location and magnitude of these events is vital for Utah air quality assessment and planning. Previous modeling studies have focused only on past dust events. This work utilized a state-of-the-science software framework based on the Community Multiscale Air Quality (CMAQ) v5.3.1 modeling system to predict dust events in Utah. The framework was verified using previous studies for dust events in April 2017 and March 2010. Once verified, the framework was used to predict the impact of future land use properties on dust events. Two scenarios were studied – shrinking of the Great Salt Lake and the addition of large-scale solar farms west of the Wasatch Front. Both showed increases in dust concentrations overpopulated areas using the meteorological conditions from the April 2017 dust event. Such information from future impact studies can assess potential impacts from climate change and can guide government water and land use policies to mitigate dust event impacts.

Air quality forecasts are of great value since several pollutants in our environment effect both human health, global climate, vegetation, crop yields, animals, materials and acidification of forests and lakes. Air-quality forecasts help to make people aware of the presence and the quantity of pollutants, and give them a chance to protect themselves, their business and the Earth. Many different air-quality models are in daily use all over the world, providing citizens with forecasts of air quality and warnings of unhealthy air quality if recommended highest concentrations are exceeded. This study adapts the WRF meteorological model (Weather research and Forecasting model) to be a driver of the CMAQ air-quality model (models-3 Community Multiscale Air Quality model). Forecasts of ozone and nitrogen dioxide concentrations from this coupled WRF/CMAQ modelling system are tested against observed data during a four-day period in May, 2006. The Lower Fraser Valley study area is a fertile valley surrounded by mountain chains in southwest British Columbia, Canada. The valley stretches from the Pacific coast eastwards towards the Rocky Mountains. This valley hosts more than 2 million people and it is west Canada’s fastest growing region. The Lower Fraser Valley holds a big city, Vancouver, several suburbs, numerous industries and a widespread agricultural production. During the analysed four-day period in May, a synoptic high-pressure built over the region, favoring high concentrations of pollutants as ozone and nitrogen dioxide. The created WRF/CMAQ model forecasted an acceptable magnitude of nitrogen dioxide but the daily variations are not recreated properly by the model. The WRF/CMAQ model forecasts the daily variation of ozone in a satisfying way, but the forecasted concentrations are overestimated by between 20 and 30 ppb throughout the study. Factors that could contribute to the elevated ozone concentrations were investigated, and it was found that the weather forecasting model WRF was not generating fully reliable meteorological values, which in turn hurt the air-quality forecasts. As the WRF model usually is a good weather forecasting model, the short spin-up time for the model could be a probable cause for its poor performance. / Prognoser över luftkvaliteten är mycket värdefulla, då flera luftföroreningar i vår närmiljö påverkar människans hälsa, det globala klimatet, vegetation, djur, material och bidrar till försurning av skog och vattendrag. Luftkvalitetsprognoser gör människan mer medveten om närvaron av luftföroreningar och i vilken mängd de finns. De ger människan en chans att vidta skyddsåtgärder för att skydda sig själv, sitt eventuella levebröd, och Jorden. Många olika luftkvalitetsmodeller används idag dagligdags över hela världen och förser invånare med prognoser för luftkvaliteten och varningar om koncentrationerna av föroreningar överstiger rekommenderade värden. I denna studie används väderprognosmodellen WRF (Weather Research and Forecasting model) för att driva luftkvalitetsmodellen CMAQ (models-3 Community Multiscale Air Quality model). Prognoser av ozon- och kvävedioxidhalterna i luften från den kopplade WRF/CMAQ modellen analyseras mot observerade data under en fyra dagars period i maj, 2006. Studieområdet Lower Fraser Valley är en bördig dalgång som är omgiven av bergskedjor i sydvästra British Columbia, Kanada. Dalen sträcker sig från Stilla havskusten och österut mot Klippiga bergen. I denna dalgång bor mer än 2 miljoner människor och det är västra Kanadas snabbast växande region. Lower Fraser Valley rymmer en storstad, Vancouver, flera förorter, många industrier och även stora jordbruksområden. Den fyra dagars period i maj som analyseras karaktäriseras av ett högtrycksbetonat synoptiskt väderläge med lokala variationer, vilka tillsammans är gynnsamma för att uppmäta höga koncentrationer av luftföroreningar som ozon och kvävedioxid. Den skapade WRF/CMAQ modellen prognostiserar godtagbar magnitud hos kvävedioxid men den dagliga variationen återskapas inte av modellen. Modellen prognostiserar den dagliga variationen av ozonkoncentration på ett tillfredsställande sätt, men storleksmässigt ligger koncentrationerna en faktor 20-30 ppb för högt rakt av under hela studien. Kringliggande faktorer som kan påverka koncentrationen ozon studeras närmare och det framkommer att den meteorologiska prognosmodellen WRF inte genererar fullt tillförlitliga värden för en rättvisande luftkvalitetsprognos. Då WRF modellen vanligtvis är en bra prognosmodell kan den korta initialiseringstiden för modellen vara en trolig orsak till dess otillräckliga prestation.

A methodology is presented to evaluate the accuracy of cloud cover fraction (CCf) forecasts generated by numerical weather prediction (NWP) and climate models. It is demonstrated with a case study consisting of simulations from theWeather Research and Forecasting (WRF) model. In this study, since the WRF CCf forecasts were initialized with reanalysis fields from the North American Mesoscale (NAM) Forecast System, the characteristics of the NAM CCf products were also evaluated. The procedures relied extensively upon manually-generated, binary cloud masks created from VIIRS (Visible Infrared Imager Radiometry Suite) imagery, which were subsequently converted into CCf truth at the resolution of the NAM and WRF gridded data. The initial results from the case study revealed biases toward under-clouding in the NAM CCf analyses and biases toward over-clouding in the WRF CCf products. These biases were evident in images created from the gridded NWP products when compared to VIIRS imagery and CCf truth data. Thus, additional simulations were completed to help assess the internal procedures used in the WRF model to translate moisture forecast fields into layered CCf products. Two additional sets of WRF CCf 24 h forecasts were generated for the region of interest using WRF restart files. One restart file was updated with CCf truth data and another was not changed. Over-clouded areas in the updated WRF restart file that were reduced with an update of the CCf truth data became over-clouded again in the WRF 24 h forecast, and were nearly identical to those from the unchanged restart file. It was concluded that the conversion of WRF forecast fields into layers of CCf products deserves closer examination in a future study.