Human health is frequently omitted from household-level studies on agricultural productivity, land-use choices, and forest degradation and deforestation. Intuition, however, suggests that it could be an extremely important factor. This dissertation is built on three essays that use household survey data from the Brazilian Amazon to examine the conditions under which human health and other critical market conditions are important factors in determining household agriculture production choices and efficiency. Essay I (Chapter 2) examines how health affects the labor allocation and production choices of migrant smallholders in the Brazilian Amazon. We show that the impacts of illness on household decisions depend critically on labor market function in the rural areas of the tropics. Furthermore, results from a formal statistical test of the labor markets shows that they do not work well, in other words are incomplete or thin, in the study area. These results are important both in specification of future smallholder household economic models and in targeting policies to better alleviate poverty and encourage more sustainable use of forest and land resources in similar tropical regions. Essay II (Chapter 3) investigates the role of health as a productive input and non-input factor of production. By using a non-neutral stochastic production approach, the impact of health is decomposed into direct effect on the production function and indirect effects on technical efficiency. The finding of the essay suggests that poor health has significant negative impacts on rural household production. The most important policy implication is that careful designing of agriculture development and rural settlements programs is important, and the provision of health care should be tied to these development projects. Essay III (Chapter 4) examines the demand for labor applied to land clearing, staple food production, livestock, working off-farm, and time taking care of sick people in the household. Specifically the empirical application examines the impact of disease on labor allocation, accounting for time lost by households taking care of sick members as a non-productive activity. Disease plays an important role in household decisions because farm activities are performed inefficiently by sick households and changes in household labor efficiency brings about a change in the relative price of competing uses for a household's time. Chapter 5 provides a summary and general conclusion of the work, and then provides comments on policy design and recommendations for further studies. In summary, the combined results of these studies show that both health condition and the quality of labor markets have significant interacting impacts on the labor allocation decisions by smallholders with accompanying welfare and deforestation implications. / Doctor of Philosophy / Most of the rural population of the Brazilian Amazon is made up of small-scale farmers – the so-called 'smallholders' – who are characterized by a lack of access to formal credit, a disconnection from social services, poor access to markets, and a dependency on their own labor as the main input in agricultural production, and thus survival. Since labor is the main input used in smallholder activities, albeit to different extents, anything that changes total household labor or labor efficiency adjusts the relative returns of competing uses, and thus labor allocation decisions. This PhD dissertation is an effort to understand whether markets, family health, and seasonality affect labor allocation decisions, and furthermore, whether those allocation decisions vary depending on productive activity. Based on cross-sectional farm data from the Brazilian Amazon, I find that the impacts of illness on household decisions depend critically on how well labor market function in rural areas. The results from a statistical test of the labor markets shows that they do not work well in the study area. These results are important both in specification of future smallholder household economic models and in targeting policies to better alleviate poverty and encourage more sustainable use of forests and land resources in similar settings. Also, I find that poor health has a significant negative impact on technical efficiency of rural household farm production. The most important policy implication is that careful designing of agriculture development and rural settlements programs is important, and the provision of health care care should be tied to these development projects. Finally, in the context of the region of study, where labor markets are thin, disease plays an important role on in household decisions because farm activities are performed inefficiently by sick households and change in labor efficiency brings about a change in relative prices of competing uses of household's time. My empirical work supports the hypothesis that health influence labor allocation decisions. In conclusion, the combined results of these studies show that both health conditions and the quality of labor markets have significant interacting impacts on the labor allocation decisions by smallholders with accompanying welfare and deforestation implications.

Le changement climatique est relié à l’évolution de la composition chimique de l’atmosphère et de sa capacité oxydante, impliquant le système COV-NOy-HOx-O3. La troposphère tropicale, de l’Afrique de l’Ouest en particulier, joue un rôle critique sur la composition atmosphérique globale pour trois raisons majeures : (1) l’existence d’importantes sources de précurseurs d’espèces photooxydantes, (2) une photochimie active, (3) une activité convective intense en période de mousson. Pour évaluer son rôle, il est nécessaire de bien caractériser ces différents processus et leur interaction. Cette question est au coeur du programme international AMMA (Analyse Multidisciplinaire de la Mousson Africaine) dans lequel s’inscrit cette thèse. Ce travail a pour objectif de caractériser et d’évaluer l’impact de la convection nuageuse profonde sur la chimie photooxydante de la troposphère libre en Afrique de l’Ouest, en particulier pour les composés organiques volatils (COV), qui sont d’importants précurseurs d’ozone. Ce travail s’appuie sur les données physico-chimiques recueillies sur les deux avions de recherche français au cours de la campagne d’observation intensive de l’été 2006 de AMMA. Afin de compléter le dispositif instrumental embarqué, une nouvelle instrumentation de mesure indirecte des COV a été d’abord développée. Puis, l’utilisation de traceurs physico-chimiques et la mise en place d’outils diagnostiques appliqués aux COV (profils verticaux de concentrations, rapport de concentration de COV ad hoc, horloge photochimique, réactivité totale vis-à-vis de OH) ont montré que la convection profonde assure un transport vertical rapide et efficace des espèces gazeuses réactives émises près de la surface vers la haute troposphère. Enfin, un modèle photochimique de boîte 0D a permis de renseigner l’évolution de la composition chimique des masses d’air post-convectives. Les simulations montrent que les espèces transportées par la convection participent activement à la chimie et conduisent à une production nette et significative d’ozone dans la haute troposphère. La sensibilité de la production d’ozone aux précurseurs gazeux (COV et NOx) a été également évaluée. / Recent climatic change is tightly linked to the evolution of the chemical composition of the atmosphere and its oxidizing capacity through VOC-NOy-HOx-O3 system. The tropical troposphere, in particular of West Africa, plays a major role in the global atmospheric composition for three major reasons: (1) the existence of important ozone precursor sources, (2) an active photochemistry, (3) an intense convective activity during the monsoon period. To evaluate its role, it is necessary to characterize these processes and their interactions. This is one of the main objectives of the AMMA (African Monsoon Multidisciplinary Analyzes) international program. The present work goes through the frame of AMMA. Its main objective is to characterize and evaluate deep convection impact on the upper troposphere chemistry of West Africa, in particular for volatile organic compounds (VOC). This work is based on the data collected on the two French research aircrafts during the special observation period of AMMA in summer 2006. In order to enhance the instrumental device deployed onboard, a new offline instrumentation for non-methane hydrocarbons (NMHC) measurement was developed. Then, various physical and chemical tracers and several diagnostic tools applied to VOC data (vertical profiles, concentration ratios, photochemical clock, OH reactivity) showed that deep convection provides a fast and effective vertical transfer of reactive species emitted near the surface to the upper troposphere. At last, a photochemical box model 0D was used to simulate the chemical evolution of the composition of postconvective air masses. Simulations showed that reactive species transported by deep convection participate actively to the upper troposphere chemistry and lead to a significant and net ozone production. Ozone production sensitivity to VOC and NOx was also evaluated in the model.