Este trabalho teve como objetivo investigar a relação entre produtos de atividades antropogênicas na Amazônia e sua influência no efeito indireto dos aerossóis no clima. Para isso, foi feita uma caracterização físico-química detalhada dos aerossóis naturais e de queimadas na Amazônia e procurou-se compreender como estes diferentes tipos de aerossóis se comportam como Núcleos de Condensação de Nuvens. Foi estudado também a influência dos aerossóis de queimadas na composição química da precipitação e no transporte de nutrientes. Visando atingir estes objetivos, foram feitas medidas em regiões distintas da Amazônia com relação ao impacto por atividades antropogênicas, principalmente queimadas. Foi possível observar em várias circunstâncias uma relação entre a composição do material particulado e da precipitação, o que nos permitiu concluir que as emissões antropogênicas influenciam significativamente a composição da precipitação. Foram identificadas as principais componentes que afetam a composição do material particulado em suspensão na Amazônia, e concluimos que o material particulado originado de emissões biogênicas é predominante em regiões preservadas, com pequena contribuição também de poeira de solo e transporte de aerossóis marinhos. Em regiões sob influência de atividades antropogênicas, observou-se que a composição dos aerossóis e da precipitação é afetada mesmo na estação úmida. No estudo das propriedades físicas e químicas das partículas de aerossol que são relevantes para o seu papel como Núcleos de Condensação de Nuvens, concluiu-se que a distribuição de tamanho é mais importante do que a composição química das partículas, devido ao fato das emissões de novas partículas por queimadas ocorrer predominantemente acima do diâmetro seco de ativação. A composição química só foi importante em valores de supersaturação baixos (< 0.2%), o que significa que esse efeito pode ser importante para nuvens estratiformes, onde o valor máximo de supersaturação é baixo, devido a baixa velocidade de ascensão das parcelas. A exportação de nutrientes devido ao transporte em larga escala de aerossóis de emissões de queimadas se mostrou particularmente crítica com relação às quantidades de fósforo que estão sendo perdidas irreversivelmente pela floresta amazônica, que foi cerca de 7 vezes maior do que a quantidade reposta por deposição úmida. Essa perda de fósforo pode ser crítica para o ecossistema em longo prazo. / The main objective of this study was to investigate the relationship between anthropogenic emissions in the Amazon basin and the indirect aerosol effect on climate. A detailed study of physical and chemical properties of natural and biomass burning aerosols was conducted, in order to understand how these completely different aerosols behave as Cloud Condensation Nuclei (CCN). It was also investigated the influence of biomass burning aerosols in chemical composition of precipitation, and transport of nutrients. The measurements were carried out in completely different regions respect to the impact of anthropogenic activities, especially biomass burning emissions. The analysis of aerosols and rainwater chemistry showed that anthropogenic emissions have a significant influence in the composition of precipitation. Factor analysis was applied to perform source identification, and the conclusion is that at remote and free of anthropogenic emission areas, the most important contribution was from biogenic emissions, with a small contribution of soil dust and marine aerosols advection. It was quite different at regions under influence of anthropogenic activities, where measurements showed a clear anthropogenic influence even during wet season both in aerosols and precipitation chemistry. In the study of hygroscopic properties of aerosol particles, the main conclusion was that size distribution of particles is the most important parameter to determine the ability of aerosols to act as CCN, because most of biomass burning emission are particles bigger than the activation diameter. Chemical composition was an important factor only if supersaturation is below 0.2%, because in this supersaturation range the activation diameter is extremely sensible to small changes in supersaturation. Transport of nutrients due to largescale transport of biomass burning aerosols was specially critical concerning phosphorus exportation, estimated as 7 times the apportionment through wet deposition. Continuous exportation of phosphorus can be a long term limitation to the forest ecosystem, if biomass burning activity maintain its present levels.

Este trabalho teve como objetivo investigar a relação entre produtos de atividades antropogênicas na Amazônia e sua influência no efeito indireto dos aerossóis no clima. Para isso, foi feita uma caracterização físico-química detalhada dos aerossóis naturais e de queimadas na Amazônia e procurou-se compreender como estes diferentes tipos de aerossóis se comportam como Núcleos de Condensação de Nuvens. Foi estudado também a influência dos aerossóis de queimadas na composição química da precipitação e no transporte de nutrientes. Visando atingir estes objetivos, foram feitas medidas em regiões distintas da Amazônia com relação ao impacto por atividades antropogênicas, principalmente queimadas. Foi possível observar em várias circunstâncias uma relação entre a composição do material particulado e da precipitação, o que nos permitiu concluir que as emissões antropogênicas influenciam significativamente a composição da precipitação. Foram identificadas as principais componentes que afetam a composição do material particulado em suspensão na Amazônia, e concluimos que o material particulado originado de emissões biogênicas é predominante em regiões preservadas, com pequena contribuição também de poeira de solo e transporte de aerossóis marinhos. Em regiões sob influência de atividades antropogênicas, observou-se que a composição dos aerossóis e da precipitação é afetada mesmo na estação úmida. No estudo das propriedades físicas e químicas das partículas de aerossol que são relevantes para o seu papel como Núcleos de Condensação de Nuvens, concluiu-se que a distribuição de tamanho é mais importante do que a composição química das partículas, devido ao fato das emissões de novas partículas por queimadas ocorrer predominantemente acima do diâmetro seco de ativação. A composição química só foi importante em valores de supersaturação baixos (< 0.2%), o que significa que esse efeito pode ser importante para nuvens estratiformes, onde o valor máximo de supersaturação é baixo, devido a baixa velocidade de ascensão das parcelas. A exportação de nutrientes devido ao transporte em larga escala de aerossóis de emissões de queimadas se mostrou particularmente crítica com relação às quantidades de fósforo que estão sendo perdidas irreversivelmente pela floresta amazônica, que foi cerca de 7 vezes maior do que a quantidade reposta por deposição úmida. Essa perda de fósforo pode ser crítica para o ecossistema em longo prazo. / The main objective of this study was to investigate the relationship between anthropogenic emissions in the Amazon basin and the indirect aerosol effect on climate. A detailed study of physical and chemical properties of natural and biomass burning aerosols was conducted, in order to understand how these completely different aerosols behave as Cloud Condensation Nuclei (CCN). It was also investigated the influence of biomass burning aerosols in chemical composition of precipitation, and transport of nutrients. The measurements were carried out in completely different regions respect to the impact of anthropogenic activities, especially biomass burning emissions. The analysis of aerosols and rainwater chemistry showed that anthropogenic emissions have a significant influence in the composition of precipitation. Factor analysis was applied to perform source identification, and the conclusion is that at remote and free of anthropogenic emission areas, the most important contribution was from biogenic emissions, with a small contribution of soil dust and marine aerosols advection. It was quite different at regions under influence of anthropogenic activities, where measurements showed a clear anthropogenic influence even during wet season both in aerosols and precipitation chemistry. In the study of hygroscopic properties of aerosol particles, the main conclusion was that size distribution of particles is the most important parameter to determine the ability of aerosols to act as CCN, because most of biomass burning emission are particles bigger than the activation diameter. Chemical composition was an important factor only if supersaturation is below 0.2%, because in this supersaturation range the activation diameter is extremely sensible to small changes in supersaturation. Transport of nutrients due to largescale transport of biomass burning aerosols was specially critical concerning phosphorus exportation, estimated as 7 times the apportionment through wet deposition. Continuous exportation of phosphorus can be a long term limitation to the forest ecosystem, if biomass burning activity maintain its present levels.

La région arctique est particulièrement sensible au changement climatique. Aux latitudes polaires, les nuages arctiques ont un effet important sur le bilan radiatif à la surface. La première partie de ce travail est constitué de l’intercomparaison instrumentale au sol à la station PUY en Mai 2013. Les mesures ont montré une bonne corrélation entre les diamètres effectifs et les distributions en taille des gouttelettes d’eau obtenus par les instruments, mais avec des biais systématiques sur les concentrations. Ces biais ont été reliés à l’estimation du volume d’échantillonnage et nous avons donc proposé une méthode consistant à normaliser les données par rapport à un instrument qui réalise des mesures intégrées. D’autre part, le FSSP et le FM ont fait l’objet d’expériences visant à évaluer l’influence de l’angle de déviation par rapport au vent extérieur et de la vitesse du vent. La seconde partie de ce travail a pour objet la campagne de mesure qui s’est déroulée à la station du Mont-Zeppelin, Ny-Alesund, Svalbard, de Mars à Mai 2012 dans le cadre du projet CLIMSLIP. Une comparaison a été effectuée entre un cas « pollué », avec des masses d’air provenant d’Asie de l’Est et d’Europe, et un cas « propre », dont les sources d’aérosols sont majoritairement locales et ne dépassent pas l’Europe du Nord. Les résultats ont montré que le cas pollué possède des concentrations en BC, aérosols et gouttes plus élevées, un mode accumulation plus important, un diamètre de gouttes plus faible et une fraction d’activation plus élevée. Enfin, le premier et le second effet indirect des aérosols ont pu être quantifiés. / The arctic region is especially sensitive to climate change. At high latitudes, arctic clouds have an important effect on the surface radiative budget. The first part of this work consists in a ground based cloud instrumentation intercomparison in the PUY station in May 2013. The measurements showed a good correlation between the effective diameters and the droplet size distributions obtained by the instruments, but with a systematical bias on the concentrations. These biases have been relied to the assessment of the sampling volume and we thus proposed a methodology to standardize the data according to an ensemble of particles probe. Moreover, the FSSP and the FM have been the subject of experiments to assess the influence of the deflection angle according to exterior wind and the wind speed. The second part of this work is about the measurement campaign at the Mount-Zeppelin station, Ny-Alesund, Svalbard, from March to May 2012 in the frame of the CLIMSLIP project. A comparison has been performed between a « polluted » case, with air masses coming from East Asia and Europe, and a « clean » case, where the aerosol sources are predominantly local and do not exceed the northern Europe. The results showed that the polluted case possessed higher concentrations in BC, aerosols and drops, an accumulation mode more important, weaker droplet diameters and higher activation fraction. Finally, the first and second aerosol indirect effects have been quantified.

Die Eisbildung in Wolken wirkt sich auf die Niederschlagsbildung, die optischen Eigenschaften, und die Persistenz der Wolken aus und beeinflusst somit das Wetter und das Klima. Sogenannte eisnukleierende Partikel (ice nucleating particle; INP), katalysieren den Gefrierprozess von Wolkentröpfchen und tragen so zur primären Eisbildung in Wolken bei. In dieser Arbeit wurden die Häufigkeit und die Eigenschaften von INP in der Arktis untersucht. Hierzu wurde ein Vielzahl an Proben analysiert: Proben zweier Eisbohrkerne (aus Spitzbergen und Grönland); Filterproben von Aerosolpartikeln, die an Bord eines Flugzeuges über dem arktischen Ozean nordöstlich von Grönland gesammelt wurden; Filterproben von Aerosolpartikeln, die an Bord eines Schiffes in der Nähe von Spitzbergen gesammelt wurden. Zusätzlich wurden auch Meeresoberflächenfilm-, Meerwasser- und Nebelwasserproben gesammelt. Es wurde festgestellt, dass die INP-Konzentrationen in der Arktis im Allgemeinen niedriger sind als in den mittleren Breiten. Und obwohl die INP-Konzentrationen bei einer Temperatur von Probe zu Probe eine hohe Variabilität aufweisen, bewegen sie sich seit der Kleinen Eiszeit im 16. Jahrhundert auf einem ähnlichen Niveau und zeigen keinen langfristigen Trend. Außergewöhnlich eisaktive Proben zeichnen sich durch hohe INP-Konzentrationen bei wärmeren Temperaturen (ca. über -15°C) aus. Die in diesen Fällen aktiven INP können auf einen biogenen Ursprung zurückgeführt werden. Ferner wurden eindeutige Hinweise auf das Vorhandensein lokaler mariner INP-Quellen gefunden in der Arktis gefunden. Dies ist ein interessantes Ergebnis, da auch gezeigt wurde, dass ohne signifikante Anreicherung während des Transfers vom Ozean in die Aerosolphase, die vorhandenen INP im Meerwasser die INP-Konzentration in der Luft nicht erklären können. Die INP-Konzentrationen Temperaturbereich unterhalb von -26°C, hingegen scheinen eher durch Ferntransport von Staub aus den mittleren Breiten und/oder terrestrischer Quellen in der Arktis bestimmt zu sein.:1 Introduction 2 Experimental 2.1 Campaign Overviews 2.1.1 Arctic Ice Cores 2.1.2 PAMARCMiP 2.1.3 PASCAL 2.2 Instrumentation 2.2.1 Droplet Freezing Assays 2.2.2 HERA 2.2.3 Low Volume Filter Sampler 2.2.4 SPIN 2.2.5 Sea and fog water sampling 2.2.6 Other Aerosol Instrumentation 2.3 Data analysis 2.3.1 INP concentration 2.3.2 Back trajectories 2.3.3 Sea ice fraction and thickness 2.3.4 Transmission Electron Microscopy 3 Results 3.1 INP measurements on Arctic ice core samples 3.1.1 Results & Discussion 3.1.2 Summary 3.2 Airborne INP measurements during PAMARCMiP 3.2.1 Results & Discussion 3.2.2 Summary 3.3 Ship-borne INP measurements during PASCAL 3.3.1 Results & Discussion 3.3.2 Summary 4 Summary and Conclusion / Ice formation in clouds impacts precipitation initiation, cloud optical properties, and cloud persistence, and hence influences weather and climate. At the base of the primary ice formation in clouds stands the ice nucleating particle (INP), which catalyzes the freezing process of cloud droplets. In this thesis, the abundance and properties of Arctic INP were investigated in samples from two ice cores (Svalbard and Greenland), in samples of aerosol particles collected on an aircraft over the Arctic ocean northeast of Greenland, and in ship-borne aerosol filter samples, as well as sea surface microlayer, bulk sea water and fog water samples collected in the vicinity of Svalbard. It was found that INP concentrations in Arctic are generally lower than in mid-latitudes. And while they show a high inter-sample variability, INP concentrations have been on similar levels since the Little Ice Age in the 16th century and show no long-term trend. Exceptionally ice-active samples are characterized by high INP concentrations at warmer temperatures (approximately above -15°C). The INP active in these cases were attributed to a biogenic origin. Furthermore, clear evidence for the presence of local marine INP sources was found in the Arctic. This is an interesting finding as it was also shown that without significant enrichment during the transfer from the ocean to the aerosol phase, the INP in the sea water can not explain the INP concentration in the air. INP concentrations temperature range below -26°C, on the other hand, appear to be determined more by long-range transport of dust from mid-latitudes and/or terrestrial sources in the Arctic.:1 Introduction 2 Experimental 2.1 Campaign Overviews 2.1.1 Arctic Ice Cores 2.1.2 PAMARCMiP 2.1.3 PASCAL 2.2 Instrumentation 2.2.1 Droplet Freezing Assays 2.2.2 HERA 2.2.3 Low Volume Filter Sampler 2.2.4 SPIN 2.2.5 Sea and fog water sampling 2.2.6 Other Aerosol Instrumentation 2.3 Data analysis 2.3.1 INP concentration 2.3.2 Back trajectories 2.3.3 Sea ice fraction and thickness 2.3.4 Transmission Electron Microscopy 3 Results 3.1 INP measurements on Arctic ice core samples 3.1.1 Results & Discussion 3.1.2 Summary 3.2 Airborne INP measurements during PAMARCMiP 3.2.1 Results & Discussion 3.2.2 Summary 3.3 Ship-borne INP measurements during PASCAL 3.3.1 Results & Discussion 3.3.2 Summary 4 Summary and Conclusion

Important aspects of the adjustments to aerosol-cloud interactions can be examined using the relationship between cloud droplet number concentration (Nd) and liquid water path (LWP). Specifically, this relation can constrain the role of aerosols in leading to thicker or thinner clouds in response to adjustment mechanisms. This study investigates the satellite retrieved relationship between Nd and LWP for a selected case of mid-latitude continental clouds using high-resolution Large-eddy simulations (LES) over a large domain in weather prediction mode. Since the satellite retrieval uses the adiabatic assumption to derive the Nd, we have also considered adiabatic Nd (NAd) from the LES model for comparison. The joint histogram analysis shows that the NAd-LWP relationship in the LES model and the satellite is in approximate agreement. In both cases, the peak conditional probability (CP) is confined to lower NAd and LWP; the corresponding mean LWP (LWP) shows a weak relation with NAd. The CP shows a larger spread at higher NAd (>50 cm–3), and the LWP increases non-monotonically with increasing NAd in both cases. Nevertheless, both lack the negative NAd-LWP relationship at higher NAd, the entrainment effect on cloud droplets. In contrast, the model simulated Nd-LWP clearly illustrates a much more nonlinear (an increase in LWP with increasing Nd and a decrease in LWP at higher Nd) relationship, which clearly depicts the cloud lifetime and the entrainment effect. Additionally, our analysis demonstrates a regime dependency (marine and continental) in the NAd-LWP relation from the satellite retrievals. Comparing local vs large-scale statistics from satellite data shows that continental clouds exhibit only a weak nonlinear NAd-LWP relationship. Hence a regime-based Nd-LWP analysis is even more relevant when it comes to warm continental clouds and their comparison to satellite retrievals.

In this thesis, we have presented a systematic analysis of the change of cloud properties due to variation in aerosol concentration over Indian region using satellite observations, and Weather Research and Forecasting Model coupled with Chemistry (WRF-Chem) simulations. The Tropical Rainfall Measurement Mission (TRMM) based Microwave Imager (TMI) estimates (2A12) have been used to compare and contrast the characteristics of cloud liquid water and ice over the Indian land region and the surrounding oceans, during the pre-monsoon (May) and monsoon (June–September) seasons. Based on the spatial homogeneity of rainfall, we have selected five regions for our study (three over ocean, two over land). In general, we find that the mean cloud liquid water and cloud ice content of land and oceanic regions are different, with the ocean regions showing higher amount of CLW. A comparison across the ocean regions suggests that the cloud liquid water over the or graphically influenced Arabian Sea (close to the Indian west coast) behaves differently from the cloud liquid water over a trapped ocean (Bay of Bengal) or an open ocean (Equatorial Indian Ocean). Specifically, the Arabian Sea region shows higher liquid water for a lower range of rainfall, whereas the Bay of Bengal and the Equatorial Indian Ocean show higher liquid water for a higher range of rainfall. Apart from geographic differences, we also documented seasonal differences by comparing cloud liquid water profiles between monsoon and pre-monsoon periods, as well as between early and peak phases of the monsoon. We find that the cloud liquid water during the lean periods of rainfall (May or June) is higher than during the peak and late monsoon season (July-September) for raining clouds over central India. However, this is not true over the ocean. As active and break phases are important signatures of the monsoon progression, we also analyzed the differences in cloud liquid water during various phases of the monsoon, namely, active, break, active-to-break (a2b) and break-to-active (b2a) transition phases. We find that the cloud liquid water content during the b2a transition phase is significantly higher than that during the a2b transition phase over central India. We speculate that this could be attributed to higher amount of aerosol loading over this region during the break phase. We lend credence to this aerosol-liquid water/rain association by comparing the central Indian cloud liquid water with Southeast Asia (where the aerosol loading is significantly smaller) and find that in the latter region, there are no significant differences in cloud liquid water during the different phases of their monsoon. The second part of our study involves evaluating the ability of the Weather Research and Forecasting Model coupled with Chemistry (WRF-Chem) to simulate the observed variation of cloud liquid water and rain efficiency. We have used no chemistry option, and the model was run with constant aerosol concentration. The model simulations (at 4.5 km resolution) are done for the month of June–July 2004 since this period was particularly favorable for the study of an active–break cycle of the monsoon. We first evaluate the sensitivity of the model to different parameterizations (microphysical, boundary layer, land surface) on the simulation of rain over central India and Bay of Bengal. This is done to identify an “optimal” combination of parameterizations which reproduces the best correlation with observed rain over these regions. In this default configuration (control run), where the aerosol concentration is kept constant throughout the simulation period, the model is not able to reproduce the observed variations of cloud liquid water during the different phases of an active-break cycle. To this end, we proceeded to modify the model by developing an aerosol-rain relation, using Aerosol Robotic Network (AERONET) and TRMM 3B42 data that realistically captures the variation of aerosol with rain. It is worth highlighting here that our goal was to primarily isolate the indirect effect of aerosols in determining the observed changes in cloud liquid water (CLW) during the active-break phases of the Indian monsoon, without getting into the complexity of a full chemistry model such as that incorporated in WRF-Chem. Moreover, the proposed modification (modified run) is necessitated by the lack of realistic emission estimates over the Indian region as well as the presence of inherent biases in monsoon simulation in WRF. The main differences we find between the modified and control simulations is in the mean as well as spatial variability of CLW. We find that the proposed modification (i.e., rate of change of aerosol concentration as a function of rain rate) leads to a realistic variation in the CLW during the active-break cycle of Indian monsoon. Specifically, the peak value of CLW in the b2a (a2b) phase is larger (smaller) in the modified as compared to the control run. These results indicate a stronger change in CLW amount in the upper levels between the two transition phases in the modified scheme as compared to the control simulation. More significantly, we also observe a change in sign at the lower levels of the atmosphere, i.e., from a strong positive difference in the control run to a negative difference in the modified simulation, similar to that observed. Additionally, we investigated the impact of the proposed modification, via CLW changes, on cloud coverage, size of clouds and their spatial variability. We find that the transformation of optically thin clouds to thick clouds during the break phase was associated with larger cloud size in modified compared to the control simulation. Moreover, the higher rate of decay of the spatial variability of CLW with grid resolution, using the modified scheme, suggests that clusters of larger clouds are more in the modified compared to control simulation. Taken together, the interactive aerosol loading proposed in this thesis yields model simulations that better mimic the observed CLW variability between the transition phases.