The overall objective of this dissertation research is to understand the spatiotemporal dynamics of water quality parameters in different water bodies of South Florida. Two major approaches (multivariate statistical techniques and remote sensing) were used in this study. Multivariate statistical techniques include cluster analysis (CA), principal component analysis (PCA), factor analysis (FA), discriminant analysis (DA), absolute principal component score-multiple linear regression (APCS-MLR) and PMF receptor modeling techniques were used to assess the water quality and identify and quantify the potential pollution sources affecting the water quality of three major rivers of South Florida. For this purpose, a 15-year (2000–2014) data set of 12 water quality variables, and about 35,000 observations were used. Agglomerative hierarchical CA grouped 16 monitoring sites into three groups (low pollution, moderate pollution, and high pollution) based on their similarity of water quality characteristics. DA, as an important data reduction method, was used to assess the water pollution status and analysis of its spatiotemporal variation. PCA/FA identified potential pollution sources in wet and dry seasons, respectively, and the effective mechanisms, rules, and causes were explained. The APCS-MLR and PMF models apportioned their contributions to each water quality variable. Also, the bio-physical parameters associated with the water quality of the two important water bodies of Lake Okeechobee and Florida Bay were investigated based on remotely sensed data. The principal objective of this part of the study is to monitor and assess the spatial and temporal changes of water quality using the application of integrated remote sensing, GIS data, and statistical techniques. The optical bands in the region from blue to near infrared and all the possible band ratios were used to explore the relation between the reflectance of a waterbody and observed data. The developed MLR models appeared to be promising for monitoring and predicting the spatiotemporal dynamics of optically active and inactive water quality characteristics in Lake Okeechobee and Florida Bay. It is believed that the results of this study could be very useful to local authorities for the control and management of pollution and better protection of water quality in the most important water bodies of South Florida.

Contexte : La dengue est une fièvre ayant pris davantage d’ampleur dans plusieurs régions du monde à climat tropical, dont le département Méta en Colombie, un lieu endémique dans la dernière décennie. Sachant que la propagation de cette maladie est entièrement dépendante des sites de reproduction du vecteur de transmission Aedes, l’investigation de prédicteurs environnementaux les affectant nous semblait pertinente d’un point de vue épidémiologique. Notre étude vise des retombées sur un plan de surveillance de santé publique et l’élaboration de programmes basés sur des observations temporelles et spatiales. Objectifs : Décrire la distribution spatio-temporelle de la dengue dans les 29 municipalités de Méta selon des prédicteurs environnementaux, soit les précipitations, la température ambiante, la température du point de rosée et la pression au niveau de la mer dans le cadre d’une étude écologique. Méthodes : Des données de surveillance SIVIGILA, de télédétection ainsi que de recensement DANE des 29 municipalités du département de Méta entre 2011 et 2019 ont été utilisées pour étudier l’association entre la dengue et les prédicteurs environnementaux. La modélisation de ces associations a été effectuée à l’aide d’un modèle de Markov cachés à changement de régime. Résultats : Des associations positives importantes ont été détectées entre les précipitations et l’incidence de la dengue atteignant des risques relatifs élevés (RR=1.97 à El Calvario à un décalage de 5 semaines ; RR=1.61 à San Juanito à un décalage de 20 semaines). La température ambiante et la température du point de rosée ont suivi en matière de pertinence n’excédant pas des RR de 1.11 pour la maladie. Or, la pression au niveau de la mer n’a pas présenté des résultats concluants. Selon les différents aspects étudiés dans les nombreuses cartes de ce mémoire, nous avons observé une dépendance géographique répétée à l’ouest du département. Conclusion : Nos conclusions recommandent qu’une attention particulière de la santé publique doit s’imposer quant à la prévention de la dengue en saison pluvieuse dans le Meta. Les hausses des valeurs de température ambiante et de la température du point de rosée peuvent expliquer certaines augmentations de l’incidence et incitent donc à la surveillance de ces paramètres également. À noter que nos résultats ont possiblement pu être influencés par un biais d’information et un biais de confusion. / Context: Dengue is a type of fever that has significantly spread in several regions of the world characterized by a tropical climate, including the Meta department in Colombia, which seemed to be an endemic place in the last decade. Knowing that the spread of this disease is entirely dependent on the reproduction sites of the Aedes transmission vector, the study of environmental predictors affecting those sites seemed relevant from an epidemiological point of view. Our study aims to add to the knowledge that can be used for public health surveillance and program elaboration based on temporal and spatial observations of dengue patterns within the region. Objectives: To describe the spatiotemporal distribution of dengue fever in the 29 municipalities of Meta according to its environmental predictors (i.e. precipitation, ambient temperature, dew point temperature and pressure at sea level) within the framework of an ecological study. Methods: SIVIGILA surveillance, remote sensing and DANE census data from 29 municipalities in the Meta department between 2011 and 2019 were used to study the association between dengue fever and environmental predictors. These associations were modeled using a regime-switching hidden Markov model. Results: Significant positive associations were detected between rainfall and dengue incidence reaching high relative risks (RR=1.97 in El Calvario at a 5-week lag; RR=1.61 in San Juanito at a 20-week lag). Ambient temperature and dew point temperature followed in terms of relevance not exceeding RRs of 1.11 for disease. However, pressure at sea level did not show conclusive results. According to the different aspects studied in the many geographical maps of this thesis, we observed a repeated geographical dependence in the west of the department. Conclusion: We recommend that special public health attention should be given to the prevention of dengue fever in the rainy season in Meta. Increases in ambient temperature values and dew point temperature may explain some increases in incidence and therefore prompt monitoring of these parameters as well. It should be noted that our results may have been influenced by an information bias and confusion bias.