MOPITT PhD Theses

The bachelor thesis deals with the mapping of air pollution in connection with transport in the Czech Republic. Air quality was derived from satellite images taken by the TROPOMI senzor, which were compared with data from the ECCAD portal. Furthemore a graphical comparison of time series values from the examined months and correlation relations were created. The final visualization should show a higher level of air pollution around the busiest roads, in localities of large cities or in places with permanently bad air conditions. A correlation between satellite image data and data from the ECCAD portal should be demonstrated. However, it is possible that the expected results will not be of such quality that it can be unequivocally declared that the initial hypotheses have been fully proven. The results showed only a weak correlation between satellite images and data from the ECCAD portal. However we can identify places with bad air conditions, such as big cities, or the bussiest roads.

Methane (CH4) is the second most powerful greenhouse gas (GHG) behind carbon dioxide (CO2), and is able to trap a large amount of long-wave radiation, leading to surface warming. Carbon monoxide (CO) plays an important role in controlling the oxidizing capacity of the atmosphere by reacting with OH radicals that affect atmospheric CH4 dynamics. Terrestrial ecosystems play an important role in determining the amount of these gases into the atmosphere. However, global quantifications of CH4 emissions from wetlands and its sinks from uplands, and CO exchanges between land and the atmosphere are still fraught with large uncertainties, presenting a big challenge to interpret complex atmospheric CH4 dynamics in recent decades. In this dissertation, I apply modeling approaches to estimate the global CH4 and CO exchanges between land ecosystems and the atmosphere and analyze how they respond to contemporary and future climate change. Firstly, I develop a process-based biogeochemistry model embedded in Terrestrial Ecosystem Model (TEM) to quantify the CO exchange between soils and the atmosphere at the global scale (Chapter 2). Parameterizations were conducted by using the CO in situ data for eleven representative ecosystem types. The model is then extrapolated to global terrestrial ecosystems. Globally soils act as a sink of atmospheric CO. Areas near the equator, Eastern US, Europe and eastern Asia will be the largest sink regions due to their optimum soil moisture and high temperature. The annual global soil net flux of atmospheric CO is primarily controlled by air temperature, soil temperature, SOC and atmospheric CO concentrations, while its monthly variation is mainly determined by air temperature, precipitation, soil temperature and soil moisture. Secondly, to better quantify the global CH4 emissions from wetlands and their uncertainties, I revise, parameterize and verify a process-based biogeochemical model for methane for various wetland ecosystems (Chapter 3). The model is then extrapolated to the global scale to quantify the uncertainty induced from four different types of uncertainty sources including parameterization, wetland type distribution, wetland area distribution and meteorological input. Spatially, the northeast US and Amazon are two hotspots of CH4 emissions, while consumption hotspots are in the eastern US and eastern China. The relationships between both wetland emissions and upland consumption and El Niño and La Niña events are analyzed. This study highlights the need for more in situ methane flux data, more accurate wetland type and area distribution information to better constrain the model uncertainty. Thirdly, to further constrain the global wetland CH4 emissions, I develop a predictive model of CH4 emissions using an artificial neural network (ANN) approach and available field observations of CH4 fluxes (Chapter 4). Eleven explanatory variables including three transient climate variables (precipitation, air temperature and solar radiation) and eight static soil property variables are considered in developing the ANN models. The models are then extrapolated to the global scale to estimate monthly CH4 emissions from 1979 to 2099. Significant interannual and seasonal variations of wetland CH4 emissions exist in the past four decades, and the emissions in this period are most sensitive to variations in solar radiation and air temperature. This study reduced the uncertainty in global CH4 emissions from wetlands and called for better characterizing variations of wetland areas and water table position and more long-term observations of CH4 fluxes in tropical regions. Finally, in order to study a new pathway of CH4 emissions from palm tree stem, I develop a two-dimensional diffusion model. The model is optimized using field data of methane emissions from palm tree stems (Chapter 5). The model is then extrapolated to Pastaza-Marañón foreland basin (PMFB) in Peru by using a process-based biogeochemical model. To our knowledge, this is among the first efforts to quantify regional CH4 emissions through this pathway. The estimates can be improved by considering the effects of changes in temperature, precipitation and radiation and using long-period continuous flux observations. Regional and global estimates of CH4 emissions through this pathway can be further constrained using more accurate palm swamp classification and spatial distribution data of palm trees at the global scale.

This thesis is motivated by the Earth’s environment taking serious hits, due to the propagation of varying atmospheric pollution across the cities of the globe. This propagation is driven mainly by the discharge of anthropogenic emissions such as; vehicular emissions, fossil fuel burning, industrial and biomass activities. Human exposure to alarming atmospheric pollution levels continues to be associated with a great number of adverse health effects. This challenge has resulted in air quality policymakers introducing globally adopted indices for air quality monitoring. The first task of this thesis was to review the existing air pollution status within the Nigerian territory. From the review of the available literature, the challenges were identified. A Nigerian city was adopted with the prospect of conducting a multi-technique appraisal of city-scale atmospheric pollution monitoring. Northern Nigeria’s educational hub, Zaria, was spatially and temporally analysed for the period of December 2015 November 2016. Considering Nigeria’s limited technological resources, portable pollutant monitors were procured, validated and utilized for outdoor pollution measurement in this study. The results revealed alarming pollution levels, enough to rank Zaria amongst the World Health Organization’s list of polluted cities. With the alarming pollution level in Zaria, respiratory wellness from population exposure was assessed. Certified health markers revealed that outdoor pollution contributed significantly to the respiratory well-being of the studied population. Satellite datasets have been a distinctive means for air quality monitoring over developing countries like Nigeria with limited ground pollution information. Using collocating ground measurements, we appraised the city-scale suitability of multi-satellite pollution measurements. The performance indicators adopted revealed that the Measurements of Pollution in the Troposphere Carbon Monoxide (MOPITT CO) estimates and the Moderate Resolution Imaging Spectroradiometer aerosol optical depth (MODIS AOD), as the suited satellite measurements. Another fundamental issue of this study was to identify if the existing Nigerian Global Navigation Satellite System Reference Framework (NIGNET) could be utilized for aerosol/particulate monitoring. This was achieved by appraising the effect of atmospheric aerosol optical depth (AOD) and ground particulate matter (PM) on the NIGNET derivedprecipitable water vapour (PWV). A dual-comparison of the PWV estimates showed a significant correlation between the GPS PWV and MODIS PWV estimates. Analysing the ground particulate datasets, MODIS AOD measurements and GPS derived precipitable water vapour (PWV) estimates, the result of the collocating datasets showed good agreement. The findings of this study contribute immensely to policies on efficient air pollution monitoring in Nigeria.

Understanding the quality of the air we breathe is critical in quantifying the impact  that atmospheric chemistry has on health. Poor air quality increases the risk of heart  and lung diseases as well as having a detrimental effect on climate, ecology and the  built environment. The burning of fossil fuels and plant matter (biomass burning)  creates large quantities of gases and particulate matter that impact air quality and  the air we breathe. Biomass burning is estimated to contribute 400 Tg of non-methane  organic compounds, 40 Tg of methane and 7.1 Tg of nitrogen oxides to the atmosphere  each year. This thesis aims to better understand the role of biomass burning on air  quality and tropospheric chemistry. The in depth analysis presented here addresses  of the impact of boreal biomass burning in North America on air quality, in particular,  carbon monoxide (CO) and ozone (O3). By using a number of different modelling  techniques along with data collected from a field campaign and satellites the transport  and chemistry of biomass burning emissions were analysed and quantified.  The first research chapter of the thesis used the GEOS-Chem atmospheric  chemistry transport model to interpret aircraft measurements of CO in biomass  burning outflow taken during the 2011 BORTAS-B campaign over Canada. The  model has some skill reproducing the observed variability, but has a positive bias  for observations <100 ppb and a negative bias for observations >300 ppb. It was  found that observed CO variations are largely due to fires over Ontario, with smaller  and less variable contributions from fossil fuel combustion from eastern Asia and NE  North America. To help interpret observed variations of CO an effective physical  age of emissions (¯A) metric was developed. It was found that during BORTAS-B the  age of emissions intercepted over Halifax, Nova Scotia is typically 411 days, and  on occasion as young as two days.The analysis shows that ¯A is typically 15 days  older than the associated photochemical ages inferred from co-located measurements  of different hydrocarbons.It is argued that a robust observed relationship between  CO and black carbon aerosol during BORTAS-B (r² >0.7), form the basis of indirect  evidence that aerosols co-emitted with gases during pyrolysis markedly slowed down  the plume photochemistry during BORTAS-B with respect to photochemistry at the  same latitude and altitude in clear skies.  The second research chapter focuses on O3 production downwind from boreal  biomass burning. Using the GEOS-Chem model, the O3 chemistry within a biomass  burning plume from a fire on 17 July 2011 in mid-Canada was examined. The model  shows a significant positive bias (~20 ppb) in reproducingO3 mixing ratios over North  America for July 2011 when compared to observations. Reducing NO emissions from  lightning and fossil fuel by 50% and 54% respectively reduced this bias to ~10 ppb.  The cause of the remaining bias is uncertain. Using a novel technique with the model,  the centre of the biomass burning plume was tracked and O3 concentrations and  chemistry was extracted from the centre of the plume. The biomass burning enhanced  O3 concentrations throughout the plume by between 1 20 ppb when compared with  the same plume path with no biomass burning. The plume was characterised as being  NOx-rich for the initial four days of transport. The sensitivity of the O3 chemistry  to different emissions was calculated and it was found that the O3 is initially highly  sensitive to NO emissions from biomass burning and then to NO emissions from  fossil fuels as it travels across an urban area surrounding Quebec City. The O3 net  production was found to initially decrease with an increase in NO but increase further  downwind.  The final research chapter of the thesis uses long-term satellite observations to  evaluate natural variability in CO concentrations over the North Atlantic. 15 years  of MOPITT CO column observations were used along with modelled CO from the  GEOS-Chem model. The model was evaluated against the MOPITT overpass and  shows a negative bias of between -8% and -24% over the northern mid-latitudes with  the largest bias seen in spring. The model has a large positive bias (8% 40%) over the  Amazon,West Africa and Indonesia through all seasons. Using Empirical Orthogonal  Function (EOF) analysis on the MOPITT and GEOS-Chem CO columns shows the  largest mode of variability seen in the North Atlantic to be the oxidation of methane  for winter and spring, biomass burning during summer and fossil fuel combustion  from East Asia during autumn.

Observé quotidiennement par des instruments à bord de satellites depuis les années 2000, le monoxyde de carbone (CO) est utilisé comme un traceur du transport de la pollution du fait de sa longue durée de vie (1 à 2 mois) dans la troposphère. Ce gaz est émis principalement par les activités anthropiques (transport, chauffage, industrie) et par les feux (naturels ou provoqués pour le brûlage de la biomasse). Le CO joue un rôle central dans la chimie troposphérique car il est détruit par le radical OH, et régule dès lors la capacité oxydante de l’atmosphère. La première partie de mon travail de thèse a consisté à valider les mesures de CO obtenues par l’instrument en occultation solaire ACE-FTS, en utilisant des observations indépendantes d’instruments sol, avion et satellite. Le deuxième volet de mon travail repose sur les données obtenues par la mission IASI, lancée en 2006 à bord du satellite MetOp-A. Nous disposons de 6 ans de données CO IASI qui sont utilisées afin d’étudier les épisodes de feux, de suivre les panaches de pollution lors de campagnes de mesures, d’établir des tendances, d’élaborer des prévisions de pollution (projet MACC), ou encore d’améliorer les inventaires d’émission. Mes travaux de thèse ont surtout concerné l’étude des variabilités temporelles et spatiales, et j’ai contribué à la validation des restitutions qui sont distribuées à une large communauté d’utilisateurs scientifiques. Enfin, le dernier volet de mon travail a consisté à quantifier et expliquer les différences entre les produits CO de MOPITT et de IASI, qui se place comme son successeur. Le poids du choix des hypothèses a priori inhérent aux algorithmes d’inversion est étudié : un nouveau produit MOPITT utilisant le profil et la matrice de variance-covariance a priori de IASI a été élaboré. Alors que les colonnes totales de ce nouveau produit se comparent bien au produit initial, l’influence du profil mais surtout de la matrice de variance-covariance a priori (qui autorise plus ou moins de variabilité lors de l’inversion) est démontrée lors de l’étude des profils.

Atmospheric carbon monoxide (CO) is a product of incomplete combustion and a byproduct of the oxidation of hydrocarbons. It plays a key role in controlling the oxidative capacity of the atmosphere since it is the main sink for the hydroxyl radical (OH), the primary tropospheric oxidant. As a result of its lifetime, CO is a useful tracer of long-range transport in models. However, estimates of the regional sources of CO are uncertain. Inverse modeling has become a widely used approach for better quantifying the sources, but a fundamental assumption in these inversions, which is typically not valid, is that the observations and models are unbiased.  In this thesis, the GEOS-Chem model and observations of CO from the Measurement Of Pollution In The Troposphere (MOPITT) instrument are employed to study the impact of systematic model errors on inversion analyses of CO. The impact of the treatment of biogenic non-methane volatile organic compounds (NMVOCs), aggregation errors, and discrepancies in the meteorological fields and OH distribution on the CO source estimates are examined. The influence of vertical transport errors on the source estimates is assessed using newly available MOPITT version 5 (V5) retrievals in a comparative inversion analysis employing surface level, profile, and column data.                                                                                                                            To quantify the potential impact of discrepancies in long-range transport on the source estimates, a high-resolution, regional inversion over North America, with optimized lateral boundary conditions, was conducted and compared with the results of a global inversion. The influence of the spatial-temporal distribution of the observations on the source estimates was also assessed through a comparison of the inversion analyses of MOPITT data and aircraft data from the Intercontinental Transport Experiment <96> North America, Phase A (INTEX-A) aircraft campaign.                                                                                                                    The results presented in the thesis provide a more comprehensive understanding of the potential impact of system model errors on inversion analyses of CO. This work also represents the first inverse modeling analysis of the MOPITT v5 retrievals. The results demonstrate the potential utility of these new data for characterizing vertical transport errors in models and they reveal that the new data can provide reliable constraints in regional CO source estimates.

L’assimilation de données permet de combiner d’une manière optimale un modèle  numérique décrivant l’évolution de la composition chimique de l’atmosphère et les mesures disponibles. Dans cette thèse, l’assimilation de données est utilisée afin de caractériser les distributions troposphériques et stratosphériques de l’ozone (O3) et du monoxyde de carbone (CO). Le Modèle de Chimie Transport (CTM) MOCAGE (MOdèle de Chimie Atmosphérique à Grande échelle) est utilisé dans une configuration à deux domaines imbriqués avec les résolutions de 2◦ (global) et de 0.2◦ (régional). La technique variationnelle du 3D-FGAT est utilisée pour toutes les études que constituent cette thèse. Nous avons évalué la complémentarité des mesures satellites au limbe et au nadir aujourd’hui disponibles pour la caractérisation de l’UTLS (Haute Troposphère Basse Stratosphère) en assimilant ces deux types de mesures simultanément. Nous nous sommes en particulier intéressé à la propagation de l’information provenant des mesures assimilés dans le modèle et plus particulièrement, aux impacts de l’assimilation de mesures stratosphérique d’ozone en troposphère aux moyennes latitudes. Les principaux objectifs de cette thèse ont été de montrer la valeur ajoutée de l’augmentation de la résolution modèle pour l’assimilation de données et les effets synergiques de l’assimilation combinée d’un sondeur au limbe et au nadir. Des développements au niveau du système d’assimilation en domaine imbriqué à 0.2◦ ont été effectués. L’assimilation des données dans le domaine global est maintenant prise en compte et les conditions aux bords provenant des champs assimilés montre un impact significatif sur le domaine imbriqué. Dans un premier temps, nous avons assimilé les profils d’ozone stratosphériques mesurés au limbe provenant de MLS (Microwave Limb Sounder) afin d’étudier deux cas d’échange entre la Stratosphère et la Troposphère (STE). L’étude compare les résultats obtenus deux résolution horizontales 2◦ et 0.2◦. L’assimilation de MLS montre une meilleure description des champs d’ozone à l’UTLS que celle obtenue avec le modèle seul en particulier à haute résolution où les filaments et les structures stratifiées dans les profils verticaux sont fidèlement représentés. Les résultats des rétro-trajectoires et des prévisions d’ozone démontrent que l’assimilation des profils stratosphériques MLS à haute résolution ont un impact sur les champs d’ozone troposphérique. L’étude montre aussi l’intérêt de disposer de tels instruments pour les applications troposphériques comme la prévision de la qualité de l’air et les études du bilan de l’ozone troposphérique. L’étude est complétée par l’estimation du flux d’ozone à la tropopause et sa sensibilité par rapport à la résolution et à l’assimilation de données. Les profils d’ozone de MLS et les colonnes troposphériques de IASI (Infrared Atmospheric Sounding Interferometer) ont ensuite été assimilés à haute résolution horizontale (0.2◦). Cette étude compare les assimilations séparés et combinées des colonnes troposphériques IASI et profils stratosphériques MLS, afin de mettre en évidence les effets synergiques de l’assimilation combinée. Les comparaisons avec des données indépendantes montrent que l’assimilation combinée de MLS et IASI donne les champs d’ozone particulièrement réalistes au niveau de la stratosphère l’UTLS et la troposphère. L’assimilation combinée montre aussi d’excellents résultats en comparaison avec les mesures indépendantes de colonnes totales par OMI (Ozone Monitoring Instrument). Enfin la dernière partie de ce travail à été consacrée à l’étude du comportement de la couche de mélange à l’UTLS en termes d’épaisseur et d’altitude. L’étude se situe à l’échelle locale et globale en assimilant les mesures d’O3 de MLS les mesures de CO de MOPITT (Measurements Of Pollution In The Troposphere). Des diagnostiques utilisant les corrélations entre O3 et CO ont permis de quantifier l’effet de l’assimilation sur l’altitude et l’épaisseur de la couche de mélange. Nous nous sommes d’abord intéressé à un cas d’étude STE ; l’assimilation de MOPITT permet d’affiner la couche de mélange alors que l’assimilation de MLS étale la couche de mélange vers la troposphère. Quand les champs issu de l’assimilation de MOPITT et de MLS sont utilisés en même temps dans les diagnostiques, la couche de mélange montre un comportement beaucoup plus réaliste et se trouve localisée au niveau de la tropopause thermique. A l’échelle globale, les mêmes effets sont observés. Quand les champs issu de l’assimilation de MOPITT et de MLS sont utilisés en même temps dans les diagnostiques, la couche de mélange correspond à la tropopause thermique au delà de 50 ◦ . Cette étude montre que l’assimilation de données permet de fournir des analyses chimiques de bonne qualité malgré la résolution verticale relativement faible du modèle (et des modèles chimiques de l’état de l’art en général). ABSTRACT : Data assimilation combines in an optimal way a numerical model describing the evolution of the atmospheric chemical composition and the available trace gases measurements. In this thesis, data assimilation is used to characterize the ozone (O3) and the carbon monoxide (CO) distributions in the stratosphere and in the troposphere. The Chemistry Transport Model (CTM) MOCAGE (MOdèle de Chimie Atmosphérique à Grande Echelle) is used in a configuration with two nested domains at 2◦ (global) and at 0.2◦ (regional). To perform the assimilation experiments a 3D-FGAT variational method is used. We evaluate the complementarity of limb and nadir measurements available at the present day at characterizing the UTLS (Upper Troposphere Lower Stratosphere) region by assimilating simultaneously the two type of measurements. We particularly focus on the impacts of data assimilation of stratospheric ozone measurements on troposphere and conversely of tropospheric data assimilation on stratosphere. Showing the added value of the increased horizontal resolution in the UTLS assimilated fields and the synergistic effects of limb and nadir assimilation were the main objectives of this work. Development in the assimilation system have been made in the assimilation system with the nested domain. Data assimilation in the global domain is now taken in account and the boundary condition from the assimilated fields show significant impacts on the regional domain. Firstly, we assimilate stratospheric ozone profiles from MLS (Microwave Limb Sounder) to investigate two Stratosphere-Troposphere Exchange (STE) case studies. This study compares the results at two horizontal grid resolutions 2◦ and 0.2◦. MLS ozone analyses show a better description of the UTLS region than the free model run. In particular, at high horizontal resolution the MLS ozone analyses present realistic filamentary ozone structures in the UTLS and laminae structures in the ozone profile. Results from backward trajectories and ozone forecasts show that assimilation at high horizontal resolution of MLS stratospheric ozone profiles has an impact on tropospheric ozone fields. This study also shows the interest of such instruments for tropospheric applications as air quality forecasts and tropospheric ozone budget studies. Secondly, stratospheric ozone profiles from MLS and tropospheric ozone columns from IASI (Infrared Atmospheric Sounding Interferometer) have been assimilated at high horizontal resolution (0.2◦). This study compares the separate and combined analysis of IASI tropospheric columns and MLS stratospheric profiles, in order to investigate possible synergistic effects. These comparisons with independent data show that the IASI and MLS combined assimilation analyses provides the most realistic ozone fields of the stratosphere, UTLS and troposphere. Through neither instrument covers the entire atmospheric column alone, the combined MLS and IASI analyses also show the best agreement with the independent total column data provided by OMI (Ozone Monitoring Instrument). Lastly, the behavior in terms of height and depth of the extra-tropical mixing layer in the lowermost stratosphere is investigated at local and global scale using MLS O3 and MOPITT (Measurements Of Pollution In The Troposphere) CO assimilation. We apply diagnostics using O3-CO correlations to quantify the effect of the assimilation on the height and depth of the mixing layer. We firstly focused on a STE case study where MOPITT CO analyses have the capability to sharpen the mixing layer distribution whereas MLS O3 analyses provide a tropospheric expansion of the mixing layer distribution with its maximum close to the thermal tropopause. When MLS O3 analyses and MOPITT CO analyses are used together for the diagnostics, the mixing layer shows more realistic results and matches the thermal tropopause. At global scale, MOPITT CO analyses still shows a sharper chemical transition between stratosphere and troposphere than the free model run. MLS O3 analyses lower and spread the mixing layer toward the troposphere. When MLS O3 analyses and MOPITT CO analyses are used together the mixing layer matches the thermal tropopause poleward to 50◦. This study shows that data assimilation can provide good quality chemical analyses despite a relatively low vertical model resolution.

We apply a four-dimensional variational (4D-Var) data assimilation system to optimize carbon monoxide (CO) emissions and to reduce the uncertainty of emission estimates from individual sources using the chemistry transport model TM5.    In the first study only a limited amount of surface network observations from the National Oceanic and Atmospheric Administration Earth System Research Laboratory (NOAA/ESRL) Global Monitoring Division (GMD) is used to test the 4D-Var system. Uncertainty reduction up to 60% in yearly emissions is observed over well-constrained regions and the inferred emissions compare well with recent studies for 2004. However, since the observations only constrain total CO emissions, the 4D-Var system has difficulties separating anthropogenic and biogenic sources in particular. The inferred emissions are validated with NOAA aircraft data over North America and the agreement is significantly improved from the prior to posterior simulation. Validation with the Measurements Of Pollution In The Troposphere (MOPITT) instrument shows a slight improved agreement over the well-constrained Northern Hemisphere and in the tropics (except for the African continent). However, the model simulation with posterior emissions underestimates MOPITT CO total columns on the remote Southern Hemisphere (SH) by about 10%. This is caused by a reduction in SH CO sources mainly due to surface stations on the high southern latitudes.  In the second study, we compare two global inversions to estimate carbon monoxide (CO) emissions for 2004. Either surface flask observations from NOAA or CO total columns from the MOPITT instrument are assimilated in a 4D-Var framework. In the Southern Hemisphere (SH) three important findings are reported. First, due to their different vertical sensitivity, the stations-only inversion increases SH biomass burning emissions by 108 Tg CO/yr more than the MOPITT-only inversion. Conversely, the MOPITT-only inversion results in SH natural emissions (mainly CO from oxidation of NMVOCs) that are 185 Tg CO/yr higher compared to the stations-only inversion.  Second, MOPITT-only derived biomass burning emissions are reduced with respect to the prior which is in contrast to previous (inverse) modeling studies.  Finally, MOPITT derived total emissions are significantly higher for South America and Africa compared to the stations-only inversion. This is likely due to a positive bias in the MOPITT V4 product. This bias is also apparent from validation with surface stations and ground-truth FTIR columns.  In the final study we present the first inverse modeling study to estimate CO emissions constrained by both surface (NOAA) and satellite (MOPITT) observations using a bias correction scheme. This approach leads to the identification of a positive bias of maximum 5 ppb in MOPITT column-averaged CO mixing ratios in the remote Southern Hemisphere (SH). The 4D-Var system is used to estimate CO emissions over South America in the period 2006-2010 and to analyze the interannual variability (IAV) of these emissions. We infer robust, high spatial resolution CO emission estimates that show slightly smaller IAV due to fires compared to the Global Fire Emissions Database (GFED3) prior emissions. Moreover, CO emissions probably associated with pre-harvest burning of sugar cane plantations are underestimated in current inventories by 50-100%.

L’objectif de cette thèse porte sur la définition d’un capteur géostationnaire infrarouge pour l’observation de la composition chimique de la basse troposphère et l’évaluation de la valeur ajoutée de cet instrument afin de caractériser la variabilité de la moyenne et basse troposphère des principaux polluants et d’améliorer l’observation et les prévisions de la qualité de l’air. Nous nous sommes intéressés à deux polluants importants: l’ozone troposphérique en raison de son impact sur la santé humaine, les écosystèmes et le climat, et le monoxyde de carbone (CO) qui est un traceur de pollution nous renseignant sur les sources d’émissions et les processus de transport. Dans un premier temps, une évaluation d’un schéma linéaire pour la chimie du CO a été effectuée sur une période d’un an et demi en comparaison avec un schéma chimique détaillé (RACMOBUS) et différents types d’observations troposphériques et stratosphériques (satellitaires, aéroportées). L’intérêt principal de ce schéma est son faible coût en temps de calcul qui permet une assimilation sur de longues périodes de jeux de données de CO. L’assimilation de données MOPITT (Measurements Of Pollution In The Troposphere) dans ce schéma a d’ailleurs permis d’évaluer la valeur ajoutée de données d’observations infrarouges à l’échelle globale. Ensuite, les caractéristiques optimales du capteur géostationnaire infrarouge ont été définies en réalisant des études d’inversion de spectres atmosphériques pour sonder l’ozone et le CO pour la qualité de l’air, le but étant d’avoir un capteur techniquement et économiquement faisable, capable de sonder la basse troposphère. Le contenu en information de cet instrument a été comparé, en période estivale, à l’information apportée par un autre instrument infrarouge géostationnaire similaire à MTG-IRS (Meteosat Third Generation - Infrared Sounder), optimisé pour la mesure de la vapeur d’eau et de la température mais capable d’avoir une information sur la composition chimique de l’atmosphère. Enfin dans une dernière partie, la valeur ajoutée de ces deux instruments dans le modèle de qualité de l’air MOCAGE, a été quantifiée en utilisant des expériences de simulation de système d’observations sur une période de deux mois d’été (juillet - août 2009). La capacité de ces deux instruments à corriger différentes sources d’erreurs (les forçages atmosphériques, les émissions, l’état initial et les trois paramètres réunis) qui affectent les prévisions et simulations de qualité de l’air, a été quantifiées. Au final, l’instrument que nous avons défini s’avère effectivement capable d’apporter une contrainte efficace sur les champs d’ozone et de CO dans la moyenne et basse troposphère. The objective of this thesis is to define a geostationary infrared sensor to observe the atmospheric composition of the lowermost troposphere. We evaluate the potential added value of such an instrument at characterizing the variability of the main pollutants and improving air quality observations and forecasts. We focus on two air quality key pollutants: tropospheric ozone, because of its impact on human health, ecosystems and climate; carbon monoxide (CO), which is a tracer of pollutants emissions. Firstly, an evaluation of a linear scheme for the CO chemistry during one year and a half has been performed in comparison with a detailed chemical scheme (RACMOBUS) and different tropospheric and stratospheric observations (satellite and aircraft data). The advantage of such a scheme is its low computational cost which allows data assimilation of CO during long periods. Assimilation of CO data from the Measurements Of Pollution In The Troposphere (MOPITT) instrument allows us to evaluate the information brought by such infrared observations at the global scale. Secondly, the optimal configuration of a new infrared geostationary sensor has been defined using retrieval studies of atmospheric spectra with the objectives to contribute to the monitoring of ozone and CO for air quality purposes; our constraint also set the ground for a sensor with technically feasible and affordable characteristics. For reference, the information content of this instrument has been compared during summer to the information content from another infrared geostationary instrument similar to MTG-IRS (Meteosat Third Generation - Infrared Sounder), optimized to monitor water vapour and temperature but with monitoring atmospheric composition as Lastly, the potential added value of both instruments for air quality prognoses has been compared using observing system simulation experiments (OSSEs) over two summer months (July - August 2009). The skill of the two instruments to correct different error sources (atmospheric forcing, emission, initial state and the three conditions together) affecting air quality simulations and forecasts, has been characterised. In the end, it is concluded that the instrument configuration proposed is effectively able to bring a constraint on ozone and CO fields in the mid-to-low troposphere.

The Air Indicator Report for Public Awareness and Community Tracking (AIRPACT) air quality model simulates EPA criteria pollutants and their precursors, providing quantitative descriptions of regional air quality in the Pacific Northwest.  This study focuses on comparing AIRPACT simulations to satellite-based retrievals of atmospheric pollutants made by select instruments flying in NASA’s Earth Observing System.  Products from the Ozone Monitoring Instrument (OMI), Atmospheric Infra-Red Sounder (AIRS), Measurement of Pollution in the Troposphere (MOPITT), Moderate Resolution Imaging Spectroradiometer (MODIS), and the Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO) are used in an effort to validate AIRPACT and serve as a basis to implement near-real time satellite resources for AIRPACT users in the future.  Focus is given to satellite-derived nitrogen dioxide, aerosol subtypes, aerosol optical depth, and carbon monoxide.  The horizontal and vertical distribution of pollutants in the AIRPACT is shown to be quite reasonable, but signs of overestimation of wildfire emissions are observed.  The broader impacts of using engineering sciences on culturally relevant education is also discussed.

This thesis presents a methodological approach to developing the capability of the Infrared Atmospheric Sounding Interferometer (IASI) instrument to inform on the atmospheric concentrations of carbon monoxide (CO), focussing on three key studies:  1) an assessment of the radiometric accuracy of the instrument; 2) the development of the University of Leicester IASI Retrieval Scheme (ULIRS) to convert measured radiances into a CO product; and 3) an investigation into the reliability and possible use of the ULIRS product.  An intercomparison between the radiances as measured by the IASI and Advanced Along Track Scanning Radiometer (AATSR) instruments is performed, and absolute differences at 11 µm of less than 0:1K are observed. Given the radiometric behaviour across the IASI instrument as a whole, it is also concluded that the IASI instrument is radiometrically accurate to < 0.3K in the 12 and 4.7 µm spectral regions.  A retrieval scheme, the ULIRS, is developed with explicit digital elevation and emissivity information, and a correction for solar surface reflection with a high resolution solar spectrum. Typical random errors over the African region relating to the profiles are found to be ~10% at 5 and 12 km, and on the total columns to be ~12 %.  The ULIRS dataset and the operational CO products from the Measurements Of Pollution In The Troposphere (MOPITT) are inter-compared. A methodology which uses the same a priori statistics, and which reduces the smoothing bias between the two sets of data shows that there is only a small bias between the ULIRS and MOPITT V4 products. A simplified top-down approach to estimating CO emissions from fires is also presented, highlighting the need for a better understanding of the correct detection of burnt area from space-based measurements.

Ridge restricted maximum likelihood (RREML) is a new method for regression analysis in linear models with dependent errors. Assume the linear model where the stochastic error terms are not independent, and the covariance structure is a function of some covariance parameter, in this case a spatial covariance parameter. Restricted maximum likelihood (REML) could be used to estimate this covariance parameter, but REML has no built-in methods for when multicollinearity exists in the design matrix. RREML takes the Bayesian analog of the ridge regression model, but modifies the context in order to incorporate the estimation of the variance parameter. The motivation behind such an approach is that by introducing a bit of bias in the estimator we will stabilize the variance of the estimates. By weighting the covariance of the prior distribution appropriately, the analysis should be able to both incorporate the information from the prior distribution and control the influence it has on the posterior estimates of the model. This work involves an approach that will be used in order to confront the inherent multi- collinearity of the design matrix obtained in inverse modeling as discussed in Kasibhatla et al. A Bayesian linear regression approach is commonly used in the atmospheric chemistry community in order to deal with the instability of the linear model, but it is found that these predetermined prior distributions can be too influential on the final results of the estimates. The goal of the proposed work is to control this sensitivity to the prior distribution while also incorporating a covariance structure on the error terms.

In this thesis the TOMCAT chemical transport model is used to investigate the processes which control the concentrations of CO and O3 in the Arctic troposphere. Particular focus is on understanding the main sources of CO, O3 and NOy species in the Arctic, distinguishing between natural and anthropogenic sources and the current drivers of interannual variability (IAV).  First results from a new version of TOMCAT, with extended hydrocarbon chemistry and heterogeneous uptake of N2O5, shows better agreement with observed CO from MOPITT, surface stations and aircraft. Changes in simulated burdens demonstrate the importance of NMHC as a source of CO, O3 and PAN in the troposphere and show that the complexity of chemical schemes may have contributed to previously reported inter-model differences. The high PAN sensitivity to additional NMHC is particularly important in the Arctic as it is the dominant source of NOx in the Arctic lower troposphere, producing up to 30% of total O3 in the summer.  This thesis contains the first source contribution analysis to consider impacts of fire emissions throughout the year in comparison to anthropogenic sources. Anthropogenic emissions are found to be the largest source of Arctic CO (48%), followed by methane (25%) and fires (13%). In summer, fire and anthropogenic sources contribute equally to the total CO burden. Boreal fires are the dominant source of O3 and NOx compared to anthropogenic emissions. North America contributes the largest amount (30%) to the total anthropogenic CO burden, followed by East Asia (26%), Europe (23%) and South Asia (9%). In contrast, North America makes the largest contribution (9%) to the Arctic O3 burden, followed by Europe (7%) and then Asia (6%). Overall, CO shows that the Arctic is most sensitive to emissions changes in Europe, then North America and then Asia.  Fire emissions are the dominant driver of current Arctic CO IAV, causing 84-93% of observed variability. A statistically significant correlation is found between observed CO and the El Nino 3.4 index due to a link with fires. El Nino is strongly associated with increased fire emissions in regions of North, Central and South America, Africa, and Asia. In contrast, El Nino is associated with reduced fire emissions in eastern North America, Europe, southern Asia and Australia. The temperature dependence of fires in several regions indicates that fire activity will increase in a warmer climate.

In the frame of atmospheric composition change related to human activities, homogeneous andcontinuous atmospheric measurements have to be performed. This is why a new observatory (calledOASIS for Observation of the Atmosphere by Solar occultation Infrared Spectroscopy) has beeninstalled on the roof of the University of Paris-Est in order to analyze concentrations and variations ofkey atmospheric trace species, such as ozone and carbon monoxide. Development of experimentalmethodology and adaptation of inversion code (PROFFIT) allow to retrieve total and partial columnsof these gases. Results are compared to different data from ground-based (SAOZ for O3), satellites(IASI, GOME-2, OMI, for O3, and IASI and MOPITT for CO) and from atmospheric modelcalculations (REPROBUS for O3 and MOCAGE for CO). Very good correlations were found,showing that a middle resolution instrument, such as OASIS is able to monitor atmospheric trace gaseswith a good accuracy. Then, thanks to a measurements campaign (April-May 2009), realized at Izaña(Tenerife, clear atmospheric conditions) observatory which is equipped with high resolutionspectrometer, we provide valuable information about the performance of various total column ozonemeasuring instruments from ground (Brewer) and space (IASI, OMI, GOME-2). All temporal O3 andCO variabilities (seasonal, day-to-day, diurnal and extreme events) observed at both sites, such asrespective high and middle resolution instruments capabilities were discussed in details.

Pour quantifier l’influence des activités humaines sur le climat et en particulier sur la compositionchimique de l’atmosphère, il est primordial de disposer de mesures continues et homogènes. C’estpourquoi l’objet de ce travail de thèse a consisté, dans un premier temps, à installer une nouvellestation d’observation (OASIS pour Observations Atmosphériques par Spectrométrie InfrarougeSolaire) sur le toit de l’Université Paris-Est à Créteil, pour mesurer des spectres d’absorption solairedans l’infrarouge, en vue d’analyser les concentrations de gaz traces et leurs différentes variabilitéstemporelles. Cette étude est centrée sur deux molécules clés de l’atmosphère impliquées, à diverseséchelles, dans les problèmes environnementaux actuels : l’ozone (O3) et le monoxyde de carbone(CO). Le développement de la méthode expérimentale et l’adaptation du code d’inversion (PROFFIT)a permis de restituer les colonnes totales et partielles de ces deux espèces. Pour valider lesperformances de cet instrument et de la méthode employée, les résultats des analyses de OASIS ontété comparés avec diverses données corrélatives provenant de mesures au sol (SAOZ pour l’O3), desatellites (IASI, GOME-2, OMI, pour l’O3, et MOPITT et IASI pour le CO) et de modèles globaux(REPROBUS pour l’O3 et MOCAGE pour le CO). Les résultats ont montré de très bons accords etprouvent la qualité du spectromètre à moyenne résolution spectrale, de la station OASIS, pour mesurerles concentrations de l’O3 et du CO atmosphérique.Une partie de ce travail a également concerné la validation des performances des instruments spatiauxactuels de détection de l’ozone atmosphérique, grâce à la campagne de mesures (avril-mai 2009) quej’ai réalisée à Izaña (Ténérife). Cette station, située en région sub-tropicale, à 2370 m d’altitude et enmilieu non pollué, est intégrée dans le réseau NDACC, et possède un spectromètre à haute résolutionspectrale. Les concentrations d’ozone issues des inversions des spectres mesurés ont été comparéesaux données fournies par un instrument au sol (Brewer, situé sur le même site) et par plusieursinstruments satellitaires (IASI, OMI, GOME-2). Les résultats ont présenté d’excellents accords.Les diverses variabilités temporelles observées à Izaña et à Créteil (saisonnière, inter-journalière,diurne et événements extrêmes) ont été interprétées et les capacités respectives. In the frame of atmospheric composition change related to human activities, homogeneous andcontinuous atmospheric measurements have to be performed. This is why a new observatory (calledOASIS for Observation of the Atmosphere by Solar occultation Infrared Spectroscopy) has beeninstalled on the roof of the University of Paris-Est in order to analyze concentrations and variations ofkey atmospheric trace species, such as ozone and carbon monoxide. Development of experimentalmethodology and adaptation of inversion code (PROFFIT) allow to retrieve total and partial columnsof these gases. Results are compared to different data from ground-based (SAOZ for O3), satellites(IASI, GOME-2, OMI, for O3, and IASI and MOPITT for CO) and from atmospheric modelcalculations (REPROBUS for O3 and MOCAGE for CO). Very good correlations were found,showing that a middle resolution instrument, such as OASIS is able to monitor atmospheric trace gaseswith a good accuracy. Then, thanks to a measurements campaign (April-May 2009), realized at Izaña(Tenerife, clear atmospheric conditions) observatory which is equipped with high resolutionspectrometer, we provide valuable information about the performance of various total column ozonemeasuring instruments from ground (Brewer) and space (IASI, OMI, GOME-2). All temporal O3 andCO variabilities (seasonal, day-to-day, diurnal and extreme events) observed at both sites, such asrespective high and middle resolution instruments capabilities were discussed in details.

In this thesis, the measurements of carbon monoxide (CO) obtained from an nadir sounding Measurement Of Pollution In The troposphere (MOPITT) instrument are used. Atmospheric CO is one of the most abundant and widely distributed air pollutant and is a very important indirect greenhouse gas via its reaction with the OH radical which in turn controls the oxidising capacity of the troposphere. In this thesis, the MOPITT Level 2 Version 3 (L2V3) retrieved CO data is primarily used and compared with recently released (April 2009) Level 2 Version 4 (L2V4) retrieved CO data to examine the potential of the MOPITT instrument to differentiate emission features in the lowermost troposphere including mega-cities. This study develops a novel robust methodology using day-night difference profile simulations to examine the ability of the instrument to identify CO enhancements in the lowermost layer of the atmosphere using ‘typical’ averaging kernels. More realistic CO profiles from the TOMCAT model are then used to validate this methodology. The day-night difference simulations are performed for the Indian subcontinent. It is shown that for L2V3, the daytime and nighttime degrees of freedom for a signal (DOFS) exhibit a bi-modal distribution for all selected Indian regions. The L2V3 simulation study clearly demonstrates, for higher DOFS, that day700-night700 differences give a closer differentiation of lowermost CO than other measures for MOPITT data, the first time that this has been processed. For L2V4, similar DOFS distributions are observed for the Indian subcontinent. The L2V4 simulation study also demonstrates for the first time that day850-night700 CO differences give a closer differentiation of lowermost CO by taking account of L2V4 day and night a priori mixing ratios. Finally, the methodologies developed in chapter 3 and 4 are applied to identify spatially isolated signals of lowermost CO for one year of data i.e. 2007. Features associated with nearly 100 cities are identified, the use of thresholds for higher DOFS retrievals and the use of non-surface retrieval levels with less tie to a priori. The significant step forward being consistent day-night differences for two different analyses (L2V3, L2V4).

This thesis developed a new CO vertical column density product from near IR observations of the SCIAMACHY instrument aboard ENVISAT. To correct the effects of the ice layer on the SCIAMACHY near-IR detectors a normalization procedure based on collocated MOPITT observations over the oceans was developed.  Although in this thesis only SCIAMACHY observations for effective cloud fractions below 20% are used, the remaining effects of clouds can still be large due to the higher cloud top albedo. Thus, a newly developed correction scheme is applied, which explicitly considers the cloud fraction, cloud top height and surface albedo of individual observations. The resu lting SCIAMACHY CO data is we ll suited for the investigation of the CO distribution over the continents, where important emission sources are located. This thesis compared the new SCIAMACHY CO data set, and also observations from the MOPITT instrument, to the results from three chemistry climate models (MATCH, EMAC, and GEOS-CHEM); the focus of this comparison is on regions with strong CO emissions (from biomass burning or anthropogenic sources). The comparison indicates that over most of these regions the simulated CO vertical column densities are systematically smaller than the satellite observations with the largest differences compared to the SCIAMAC HY observations. Because of the high sensitivity of the SCIAMACHY observations for the lowest part of the atmosphere, this indicates that especially close to the surface the model simulations systematically underestimate the true atmospheric CO concentrations, probably caused by an underestimation of the true CO emissions by current emission inventories.

Carbon monoxide (CO) is present in the troposphere as a product of fossil fuel combustion, biomass burning and the oxidation of volatile hydrocarbons. It is the principal sink of the hydroxyl radical (OH), thereby affecting the concentrations of greenhouse gases such as CH4 and O3. Consequently, CO has an atmospheric lifetime of 1-3 months, making it a good tracer for studying the long range transport of pollution. Satellite observations present a valuable tool to investigate tropospheric CO. The Atmospheric InfraRed Sounder (AIRS), onboard the Aqua satellite, is sensitive to tropospheric CO in ~50 of its 2378 channels. This sensitivity to CO, combined with the daily global coverage provided by AIRS, makes AIRS a potentially useful instrument for observing CO sources and transport. An optimal estimation retrieval scheme has been developed for AIRS, to provide CO profiles from near-surface altitudes to 150 hPa. Through a validation study, using CO profiles from in-situ aircraft measurements, this retrieval scheme has been shown to provide CO observations with strong correlations to in situ measurements. Compared to the operational AIRS v4 CO product this retrieval scheme is shown to provide total column CO retrievals with a reduced bias relative to the in situ measurements (~ -10% to ~ -1%). In addition, the optimal estimation retrieval is shown to provide improved estimation and characterization of the retrieval errors. Further validation work has been carried out through comparison with the established CO observations from the MOPITT instrument, onboard the Terra satellite. Good agreement (correlation coefficient > 0.9, and bias < 1.0 ppbv) between the instruments is observed in the mid-troposphere. At this level, the optimal estimation scheme is shown to remove a positive bias of ~10 ppbv, relative to MOPITT, that is present in the AIRS v4 CO product. The AIRS instrument is also shown to be less sensitive to CO in the lower troposphere than MOPITT. AIRS is also demonstrated to provide fewer pieces of independent information about the vertical structure of CO at tropical latitudes, where higher thermal contrast increases the sensitivity of MOPITT. Through time series analysis, the capability of AIRS to detect seasonal trends in CO is demonstrated. The potential of AIRS to be used to track, both horizontal and vertical, CO transport is explored. AIRS is shown to be capable of tracking horizontal transport, and to have potential to track vertical transport when combined with another satellite sensor.

This thesis uses a global 3-D chemical transport model (GEOS-Chem) and its adjoint, in conjunction with multiple global satellite datasets (from MOPITT, AIRS, SCIAMACHY and TES) to better understand and quantify the sources of carbon monoxide. Adjoint inverse model dramatically improves the resolution of the CO source constraints and overcomes the aggregation error of the low resolution analytical estimates. The study aimed to estimate Asian CO sources using MOPITT satellite measurements obtained during Spring 2001 TRACE-P campaign. The two inverse methods, adjoint and analytical, generally give consistent source constraints when averaged over large regions. The adjoint solution reveals fine-scale variability (cities, political boundaries) that the analytical inversion cannot resolve, for example, in the Indian subcontinent, and some of that variability is of opposite sign which points to large aggregation errors in the analytical solution. Upward correction factors to Chinese emissions from the prior inventory are largest in central and eastern China, consistent with a recent bottom-up revision of that inventory. MOPITT, AIRS, TES and SCIAMACHY CO satellite datasets all provide potentially complementary information about CO sources. MOPITT measurements have a long record of validation, AIRS provides unprecedented daily global dataset and SCIAMACHY instrument has unique vertical sensitivity extending all the way to the surface. Previous source inversion studies have mostly used individual datasets, while I investigated the benefit of using multiple measurements of varying vertical sensitivity, data density and data quality. Large uncertainties exist in the source estimates, and modeled concentrations show large disagreements with observations, particularly in matching the amplitude of the observed seasonal cycle. After establishing consistency among MOPITT, AIRS and SCIAMACHY Bremen datasets, I estimated monthly CO sources globally at 4° x 5° degree resolution over the whole year (May 2004April 2005) in an adjoint inversion. CO source constraints benefit from the multiple datasets where the data are consistent (Northern Hemisphere and Australia) and remain difficult where the data is not consistent and where there are additional biases in the model (S. America, southern Africa). I find large northern hemispheric seasonal correction in the middle latitudes, with fall-winter-spring emissions much larger than in the summer. Annual global CO emission estimate is 1350 Tg.

La thèse s’est inscrite dans le développement du service d’observation PAES (Pollution Atmosphérique à l’Échelle Synoptique), réseau de stations d’altitude, dédié à la mesure à long terme de l’ozone et de ses principaux précurseurs gazeux. La représentativité spatiale et temporelle a été évaluée grâce à l’analyse des données de surface comparées à des mesures aéroportées (MOZAIC) et satellitaires (MOPITT) ; les stations d’altitude s’avèrent capables de suivre les changements globaux de la composition de l’atmosphère et, dans une certaine mesure, sont représentatives de la variabilité de la troposphère libre malgré l’influence de la surface. Un bilan de la qualité de l’air en Europe de l’Ouest, depuis les années 1990 a été dressé à partir des données d’ozone et de CO de PAES et de différents réseaux français et européens. Enfin, l’analyse couplée des données d’observation obtenues pendant la campagne de terrain PIC 2005 (13 juin - 7 juillet 2005 autour du Pic du Midi) et de simulations (Méso-NH) a permis notamment de comprendre le cycle diurne inversé de l’ozone observé dans des climatologies issues de certaines stations d’altitude. Le rôle du transport à l’échelle synoptique et des échanges entre couche limite et troposphère libre sur la mesure en altitude a été montrée dans l’étude détaillée de l’épisode de pollution du 23 juin 2005.

The horizontal and vertical distributions of ozone and its precursors over North America during the spring and the summer are frequently determined by several factors: cloud convection, lightning NOX production, mixing depth, and long-range transport. The critical factors that contribute to the spatial distribution of air pollutants are studied using the Regional chEmical trAnsport Model (REAM) with diverse satellite measurements as well as in-situ surface and aircraft measurements. Among the space-borne measurements, GOME and OMI NO2 column measurements show enhanced lightning NO X over the continent and the western North Atlantic. Concurrent convective transport-causing CO column peaks and high CO enhancements in the upper troposphere (UT) over the ocean are shown from the modeling analysis of the CO column by MOPITT and UT CO by TES. Likewise, TOMS-SAGE II and OMI-MLS O3 column peaks and TES UT O3 enhancements due to convective outflow and lightning NOX are also observed. Lightning NOX production in REAM is much larger than that in GEOS-CHEM, resulting in better simulations of GOME NO2 columns over the western North Atlantic. Consequently, REAM simulates larger O3 increasing trends in better agreement with TOMS-SAGE II and OMI-MLS O3 columns over the southern United States and the western North Atlantic than GEOS-CHEM. Another factor, mixing depth, is a key parameter for the boundary layer structure of the model. Simulated spring to summer transitions of O3 and its precursors over North America indicate that the simulated boundary layer structure plays a key role in differentiating REAM from GEOS-CHEM. Large enhancements of columns and upper tropospheric O3 comparable to those over the eastern United States are found over the western North Atlantic in the satellite measurements and REAM simulations. The O3 enhancement region migrates northward from the spring to the summer. A model analysis indicates that the northward shift is driven by O3 in the stratospheric flux, convective outflow and production from lightning NOX. In addition, long-range transport affects the spatial distributions of air pollutants, particularly during the spring. During the late spring, large enhancements of NOX, PAN, O3, CO, CFCs, and Halon-1211 in UT are found over North America due to a surge of trans-Pacific pollutant transport from observations during the TOPSE 2000 experiment. The transition occurs later than that of the typical low-altitude trans-Pacific transport, which peaks around March or April. Surface and aircraft measurements show a large amount of reactive nitrogen tracers over the Antarctic plateau during the summer. These enhanced measurements are investigated, and their photochemical impact is assessed by 1-D CTM and 3-D CTM, REAM. The 1-D model and REAM reasonably simulate the surface measurements of NO, HNO3, HNO4, and balloon NO measurements at the South Pole. However, compared with the Twin Otter NO measurements, REAM underestimates NO concentrations over plateau regions because parameterization based on surface measurements at the South Pole underestimates emissions in higher-elevation plateau regions. After all, around 50% of reactive nitrogen is scavenged by deposition, and the other is lost by transport. Thus, a shallow but highly active oxidizing canopy surrounds the Antarctic plateau due to snow NO X emissions.

This thesis presents the characterization and calibration of the MOPITT-A instrument which uses the technique of correlation spectroscopy to ensure carbon monoxide in the atmosphere. A theoretical model is developed for the instrument and compared to MOPITT-A measurements collected under controlled laboratory conditions, which were designed to emulate atmospheric signals. It is shown that the model and measurements are in very good agreement with each other and that the MOPITT-A instrument behaves as expected. It was found that the gain of the instrument varies with time. The cause of the gain variation is not known but it is suggested that frosting inside the detector nest would be consistent with the observed nature of the variation.

This dissertation consists of two complementary major sections. The first is an examination of summertime cyclone transport processes during NASA’s Intercontinental Chemical Transport Experiment (INTEX-A) field campaign during Summer 2004. Warm conveyor belts (WCBs) are important mechanisms for transporting pollution during the cool season. These airstreams distribute surface emissions throughout the troposphere, playing a major role in the long range transport of chemical species. Previous efforts to understand the lofting of WCBs have not investigated the relative importance of vertical forcing. In this study, we use fine resolution model-derived meteorological data, air parcel trajectories, flux calculations, and a diagnostic package for weather systems to perform a focused investigation of WCBs during the warm season INTEX-A period. Lifting and transport mechanisms during INTEX-A are compared to a well documented cool season WCB case in the literature. Results show that weak, mid-latitude cyclones are capable of producing vertical transport as great or greater than much stronger cyclones. An analysis of forcing terms contributing to vertical motion reveals that the Laplacian of latent heat release is the primary contributor to vertical motion during some cases of INTEX-A. The latent heating term is found to be greatest in areas of deep convection. This convection allows weak cyclones to produce WCB-like transport. WCB pathways are similar for the cases studied. In each example, air which originates far south of the low in the warm sector, ascends to the north, and joins the upper-level westerly flow northeast of the low center. Although the transport pathways are similar, the forcing mechanism and location of maximum vertical transport are found to exhibit strong case-to-case variability. When cyclone scale dynamics are relatively weak, widespread deep convection, especially south of the cyclone’s center, is necessary to produce transport resembling a WCB. The second major component of the research is an investigation of warm season carbon monoxide (CO) transport episodes during INTEX-A using synthetic data to simulate CO retrievals from MOPITT’s gas correlation radiometers. This was done as if MOPITT was in geosynchronous orbit, providing simultaneous views of CO rather than the much sparser view currently provided by operational MOPITT aboard the polar orbiting Terra satellite. Pollution and its transport are global problems that require space-derived measurements for diagnosis and research. Since CO has a median lifetime of approximately two months, it is a good tracer of tropospheric circulations. This study determines the extent to which space-based MOPITT retrievals describe CO during several meteorological scenarios. Our procedure creates synthetic MOPITT retrievals using CO output from the Sulfur Transport Eulerian Model (STEM) regional scale chemical transport model. That is, STEM-derived CO vertical profiles are imported into a radiative transfer code. The calculated radiation spectra then are input to the MOPITT CO retrieval algorithm to create a synthetic version of MOPITT CO. MOPITT is assumed to be in geostationary orbit, and the effects of clouds are not considered. Simulated imagery is shown to be a valuable tool for understanding the capabilities of current sensors and the potential for new sensors to be placed on different platforms. This type of study would not have been possible using operational retrievals. We examine three phenomena observed during INTEX-A: Alaskan fires, urban plumes, and a warm conveyor belt. The evolution of thick, broad STEM CO patterns in the mid-troposphere is well represented by the synthetic MOPITT CO retrievals. Due to the MOPITT retrievals having coarse vertical resolution, as well as being constrained by a priori information, differences in the magnitudes of STEM and MOPITT CO were often seen. Neglecting differences in magnitude, MOPITT is shown to be useful at describing CO during several meteorological scenarios, specifically in the upper troposphere.

Subject of this thesis is the derivation of time series, variability and trendstropospheric trace gases from ground-based FTIR measurements (FTIR =rier Transform InfraRed). Hereby, in Kiruna (Northern Sweden) and Izaona (Te-neriffa) long-term measurements under the NDACC (Network for the Detectionof Atmospheric Composition Change). From these measurements will beProfiles and zeniths of many trace gases, including CO, O3, N2O, CH4, C2H6and HCFC-22 derived. These measurements have been in Kiruna since 1996, inIzaona since 1999, continuously performed. Therefore, they are particularly suitable for theDeterminations of long-term trends.First of all, the evaluation strategy in the infrared spectral range, ie thetion of the gases to be tested, explained and a calculation of the error is made. in theFramework of the UFTIR project (Time Series of Upper Free Troposphere observationsfrom a European ground-based FTIR network) were tested on six European FTIRMeasuring stations with a coordinated evaluation strategy Time series of profiles andPart of the target gases CO, O3, N2O, CH4, C2H6 and HCFC-22 created. A heavyThis paper focuses on the discussion of CO time series. For this gasco-ordination for a joint publication within the framework of theJektes UFTIR taken over. The annual quantity of the individual gases andpeculiarities of the time series, such as the breadth of dependence of CO andCO anomalies studied in the Nordhemisph`are 1998 and 2002/2003 due toof biomass combustion have occurred.Another part of the thesis deals with the validation of the satellite instrumentMOPITT (Measurements of Pollution in the Troposphere). There will be wholesalaries and profiles of CO for the stations Kiruna and Izaona compared. Zus`atzlicha comparison between in situ and FTIR CO measurements is shown. CO is not puretroposph`arisches gas, but also occurs in the upper stratosph`are and lowerMesosph`are ago. Due to the improved evaluation methodology in the context of the workseparation of stratospheric and tropospheric CO has become possible.The Sauls of strato-mesosph`arischem CO show a clear annual course, fromthe entry of mesosph`arischer air into the stratosphere in the polar vortex detectedbe k`onnen. In addition, the time series already discussed with modelresults from the Oslo CTM2 model (chemical transport model).A statistical trend analysis method, the so-called “B ootstrap resampling”method is applied to all measurements. The advantages of this method are that nostatistical a priori distribution of data is assumed and next to the resultthe uncertainties associated with the depreciation of the trend,be true k. The trend determination was tropospheric on time seriesGases were carried out at all six UFTIR stations for the years 1995-2004. It showsFor N2O and CH4 a quasi-linearly increasing trend, for CO a slightly negativeTrend with some special increases in years with increased biomasscombustion and a negative trend for C2H6. The ozone trend seems to be increasing be, with a slightly negative trend in the troposph`are and a positive trendin the stratosphere. HCFC-22 shows a clearly positive trend of about 4 бper year in Kiruna.

A new high-resolution (0.004 cm-1) research-grade Fourier transform spectrometer was installed at the University of Toronto Atmospheric Observatory (TAO), and commissioned in October 2001. TAO fills a coverage gap in North America in the Network for the Detection of Atmospheric Composition Change (NDACC), and was accepted as a Complementary Station in March 2004 as a direct result of work presented in this thesis. Trace gas concentrations are derived using an existing Optimal Estimation Method (OEM) retrieval algorithm implemented at TAO as part of this work. A set of prior constraints necessary for retrievals of vertical profiles of trace gases using the OEM approach was optimized, and includes a priori volume mixing ratio profiles suited to TAO’s location, and the prior and measurement noise covariances. A three-year time series of O3, NO. NO2, HCl, HF, N2O and CH4 (exhibiting the known seasonal cycles) was submitted to the NDACC database for long-terra monitoring, trend detection, and satellite validation purposes. A quantitative retrieval characterization was performed whereby the averaging kernels, weighting functions, measurement, smoothing and temperature error covariances of the retrievals of these seven gases were calculated. The degrees of freedom for signal ranged from 1.10 for NO2 to 4.12 for N2O. A systematic study showed that the effect of several different layering schemes (implicit retrieval constraints) on the retrieved vertical profile is less than 5% for N2O and HF, and less than 2% for NO, NO 2, CH4 and HCl; however, retrieved profiles of H2O showed differences up to 13% when the layering scheme was varied. A maximum tropospheric layer thickness of 0.8 km is inferred for accurate H2O retrievals. Finally, the usefulness of Fourier Transform InfraRed (FTIR) observations for studying different regions of the atmosphere was demonstrated by (1) documenting a limited sensitivity of ground-based measurements to mesospheric-lower thermospheric concentrations of NO, (2) correlating anomalously high stratospheric HCl, HF and O3 concentrations derived at TAO with the passage of a polar air filament above Toronto, and (3) showing very good agreement between monthly mean abundances of tropospheric CO between January 2002September 2003 and spatially coincident MOPITT satellite measurements.

The overarching goal of this dissertation is to develop robust, spatially and temporally resolved CO sources, using global chemical transport modeling, CO measurements from Climate Monitoring and Diagnostic Laboratory (CMDL) and Measurement of Pollution In The Troposphere (MOPITT), under the framework of Bayesian synthesis inversion. To rigorously quantify the CO sources, I conducted five sets of inverse analyses, with each set investigating specific methodological and scientific issues. The first two inverse analyses separately explored two different CO observations to estimate CO sources by region and sector. Under a range of scenarios relating to inverse methodology and data quality issues, top-down estimates using CMDL CO surface and MOPITT CO remote-sensed measurements show consistent results particularly on a significantly large fossil fuel/biofuel (FFBF) emission in East Asia than present bottom-up estimates. The robustness of this estimate is strongly supported by forward and inverse modeling studies in the region particularly from TRansport and Chemical Evolution over the Pacific (TRACE-P) campaign. The use of high-resolution measurement for the first time in CO inversion also draws attention to a methodology issue that the range of estimates from the scenarios is larger than posterior uncertainties, suggesting that estimate uncertainties may be underestimated. My analyses highlight the utility of top-down approach to provide additional constraints on present global estimates by also pointing to other discrepancies including apparent underestimation of FFBF from Africa/Latin America and biomass burning (BIOM) sources in Africa, southeast Asia and north-Latin America, indicating inconsistencies on our current understanding of fuel use and land-use patterns in these regions. Inverse analysis using MOPITT is extended to determine the extent of MOPITT information and estimate monthly regional CO sources. A major finding, which is consistent with other atmospheric observations but differ with satellite area-burned observations, is a significant overestimation in southern Africa for June/July relative to satellite-and-model-constrained BIOM emissions of CO. Sensitivity inverse analyses on observation error covariance and structure, and sequential inversion using NOAA CMDL to fully exploit available information, confirm the robustness of the estimates and further recognize the limitations of the approach, implying the need to further improve the methodology and to reconcile discrepancies.

Quantifying the continental outflow and intercontinental transport of air pollutants is one of the greatest challenges of atmospheric chemistry today. It is expected that the industrialization of Asia will be one of the major drivers for changes in atmospheric composition in the coming decades. Recent studies have suggested that transpacific transport of Asian pollution has significant implications for ozone and aerosol air quality in the United States. In my Ph.D. work, tropospheric observations from space are linked with in situ observations and 3-D models to examine the mechanisms and impact of the intercontinental transport of these pollutants as well as the magnitude of their emission.  Satellite observations of carbon monoxide (CO) from the Measurement of Pollution in the Troposphere (MOPITT) instrument were combined with measurements from the TRACE-P aircraft mission over the northwest Pacific, and with a global 3-D chemical transport model (GEOS-CHEM), to examine Asian pollution outflow and its transpacific transport during spring 2001. Four major events of transpacific transport of Asian pollution were seen by MOPITT, in-situ platforms, and GEOS-CHEM, indicating that satellites can be used to successfully monitor long-range pollution transport. One of these events was sampled by TRACE-P over the NE Pacific and this analysis demonstrates the first field observation of PAN decomposition driving ozone production in polluted plumes transported to the remote troposphere.  Inverse modeling techniques were employed to compare the constraints on Asian sources of carbon monoxide from MOPITT satellite observations, and aircraft observations from the TRACE-P mission. MOPITT observations provide greater information towards geographically disaggregating source regions within Asia in comparison to the aircraft observations, reflecting the ability of the satellite to observe all outflow and source regions. The MOPITT and TRACE-P observations are independently consistent in the constraints that they provide on Asian CO sources, where biomass burning emissions are much less than previously thought and emissions from small industrial coal facilities in China have been underestimated.  Recent observations of Asian aerosol pollution at sites in western North America have raised concern about possible air quality implications. The value of MODIS satellite observations for quantifying transpacific aerosol pollution transport, forecasting Asian pollution events at sites in the United States, and testing a CTM (GEOS-CHEM) simulation of this process was explored. Most transpacific aerosol pollution events observed by MODIS in spring 2001 are associated with significant Asian sulfate enhancements simulated by the model at sites in the NW United States, occasionally exceeding 1 micro-g/m3.

Boreal forest fires are a significant contributor to atmospheric composition in the high northern hemisphere, and are highly variable both spatially and temporally. This study uses a new emissions model [Kasischke et al. , 2005] to generate input to the University of Maryland Chemical Transport Model [Allen et al., 1996], with the goal of examining and constraining the key uncertainties in current understanding of boreal forest fire behavior. Model outputs are compared with data from the MOPITT instrument as well as in situ measurements of CO. A case study of CO transport during the summer of 2000 is used to examine several key uncertainties in the emissions estimates, describing how current levels of uncertainty affect atmospheric composition and applying atmospheric measurements can be applied to constrain uncertainty. Source magnitudes determined by inverse methods were shown to be highly sensitive to the assumed injection properties. For the boreal forest in 2000, the best agreement with observations was obtained with a pressure-weighted profile of injection throughout the tropospheric column, but detailed examination of the results makes clear that any uniform parameterization of injection will be a significant source of error when applied globally. Comparison of simulated CO distributions from daily, weekly, and monthly aggregate emissions sources demonstrated that while model data sources produced a valid representation of emissions at weekly resolution, the atmospheric distribution outside the source region has very little sensitivity to temporal variability at scales finer than 30 days. Different estimates of burned area produced large differences in simulated patterns of atmospheric CO. The GBA-2000 global product and the data sources used by Kasischke et al. [2005] gave better agreement with atmospheric observations compared to the GLOBSCAR product. Comparison of different estimates of fuel consumption indicated that atmospheric measurements of CO have limited sensitivity to spatial variability in fuels, but that current fuels maps can improve agreement with atmospheric measurements. These results provide a clear indication of how atmospheric measurements can be used to test hypotheses generated by emissions models.

MOPITT-A (Measurements Of Pollution In The Troposphere-Airborne) is a new correlation radiometer that operates from the superpod of an ER-2 aircraft. The instrument is equipped with four channels, three of which are sensitive to tropospheric carbon monoxide. The fourth channel is sensitive to methane. In 2000, the instrument participated in the SAFARI 2000 field campaign in southern Africa, whose goal was to quantify atmospheric and land changes associated with biomass burning processes. Case study retrievals were performed for the September 7th flight, using radiances from a single CO channel. The method employed is a regression scheme that relies on the fact that radiances are mainly a function of the surface temperature and the CO column. The product of the retrieval is a tropospheric CO column based on a constant vertical profile. The instrument is able to distinguish between clean air and polluted air belonging to the so called “river of smoke”, laden with biomass burning material. The results compare favorably with two CO products independently derived from spectra recorded by the Scanning High Resolution Interferometer Sounder (S-HIS) on the ER-2. The precision of the retrievals is of the order of 5 ppbv.

In this thesis, measurements of tropospheric carbon monoxide (CO) obtained by the Measurements of Pollution in the Troposphere (MOPITT) instrument are investigated.  The MOPITT instrument and measurement techniques are discussed, and recent results from the MOPITT validation campaign are presented.  MOPITT phase I retrieved CO profile measurements are compared to in situ data obtained during the ACTO campaign.  These comparison suggest that MOPITT CO profiles may possess a positive bias throughout the troposphere.  The results also demonstrate the need to take into account the effects of vertical resolution and the influence of the a priori on MOPITT retrieved profiles through the use of the retrieval averaging kernels.  Comparisons of MOPITT and TOMCAT model CO profiles are also presented.  The results suggest that MOPITT profiles are biased high compared to the model, and that these biases vary greatly from location to location.  Simulations conducted using MOPITT operational averaging kernels to investigate the ability of the MOPITT instrument to measure layer enhancements of CO are then discussed.  It is shown that MOPITT is sensitive to layer enhancements, although it is not always possible to determine the altitude of such enhancements.  The seasonal cycle in MOPITT CO profile data is derived and compared to that of six CMDL surface measuring sites.  A method for detecting long range transport events with anomalous CO using a regional analysis, which achieves limited success, is presented.  Finally, an investigation of the combined use of the differing sensitivity of MOPITT between day and night in order to gain extra vertical information from MOPITT data is discussed.  This is the first such study of its kind and the results are encouraging.  Analysis shows that by examining the difference between day and night MOPITT retrieved ‘surface’ data it is possible, in a number of regions and seasons, to obtain further information on the vertical structure of CO.  The results are validated using the TOMCAT model to represent the ‘real’ atmosphere.  Examinations of day-night differences in phase I MOPITT data suggest that it is possible to use these differences to identify CO source regions such as biomass burning and industrialised regions.

Inverse modeling of carbon monoxide emissions

The importance of understanding the MOPITT instrument transmittance is discussed. Measurements were made of the MOPITT Flight Model instrument transmittance using a Difference Frequency Laser System. The data suffered from alignment problems, which were largely corrected in later tests on the Engineering Qualification Model. Despite these problems, the Flight Model data indicated that the methane channel filters were found to differ significantly from piece-part measurements of the MOPITT narrow band-pass filter transmittances. The Flight Model data motivated adjustments to the piece-part data for all channels to produce nominal filter profiles used for current MOPITT retrievals. These profiles are shown to be consistent with laboratory and in-flight data.  A radiative transfer model of MOPITT and the Earth’s atmosphere was developed and validated. The model was used to establish sensitivity levels of MOPITT to uncertainty in calculation parameters, the filter profile and other instrument parameters. The calculation parameters employed by earlier radiative transfer models are shown to be of insufficient accuracy for MOPITT retrieval purposes. MOPITT calibration events are shown to reduce the impact of instrument parameter uncertainties on target gas retrievals. The model was also used to retrieve filter shift information for the methane channels.

An aircraft version, MOPITT Airborne Test Radiometer (MATR), was developed to provide measurements to test the forward model and cloud clearing algorithms used in the retrieval of carbon monoxide and methane from satellite measurements taken by the instrument Measurement Of Pollution In The Troposphere (MOPITT). A description of the instrument MATR and the gas correlation technique used on-board MATR are presented here. Inverse methods were used to retrieve information about tropospheric carbon monoxide and the surface temperature from MATR flight measurements. The results from the retrievals are presented along with an intercomparison between the retrieved carbon monoxide quantity and available in-situ measurements. An advantage of MATR is its capability of recording high-rate data. Engineering data contains information about the optical and electronic characteristics of the instrument, and the stability of the length and pressure modulator. The temperature cycle within the pressure modulator cell is an unknown factor. I have used three independent approaches to gain an insight into the complex structure of the temperature cycle within the cell. I used two available pressure modulator models and used inverse methods to retrieve temperature information about the gas cycle within the cell from cell emission measurements taken by MATR and from high spectral spectroscopic transmission measurements taken through the pressure modulator cell. The results of the modelled and retrieved cell temperature cycles are presented here.