The goal of this study is to improve seasonal tornado outbreak forecasting by creating a statistical model that forecasts tornado outbreak frequency in the US using teleconnection indices as predictors. For this study, a tornado outbreak is defined as more than 6 tornado reports associated with a single synoptic system and an event N15 rating index of 0.5 or higher. The tornado outbreak season is confined to all months after February for a given calendar year. Monthly teleconnection indices are derived from a rotated principal component analysis (RPCA) of the geopotential height fields. Various regression techniques were trained with a sample of monthly teleconnection indices, tested on new data, and optimized to achieve the highest predictive skill. The outcome of this study could potentially allow forecasters the ability to predict tornado outbreak potential on a climatological scale with months of lead-time, allowing for better preparation strategies for tornado outbreak seasons.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / In order to obtain an increase of quality seasonal climate forecast of precipitation, for the state of Rio Grande do Sul, were implemented and evaluated nine types of simulations that uses different cumulus parameterization schemes available in the Regional Climate Model version 4 RegCM4.The tested parameterizations were Grell with closure Arakawa and Schubert - AS and Fritsch and Chappell - FC, MIT-Emanuel and mixed convection that is the use of different convection schemes on the land and the sea. The evaluation method consisted of analysis qualitative and quantitative statistics of seasonal precipitation climate forecasts of five regions of Rio Grande do Sul, from August 2013 to August 2014.The statistics applied were Taylor diagram, random and systematic error analysis, concordance index and contingency table. The forecasts were evaluated using observed data from meteorological stations of the Instituto Nacional de Meteorogia (INMET). The analysis showed the RegCM4 had higher correlations and lower errors compared to the Global Model. The best results were observed in the northern and western part of the state with the parameterizations Grell FC, Grell AS and the combination of Emanuel simulated ocean and Grell AS on land. Although some regions were not adequately represented by the Regional Climate Model RegCM4, yet it performed well reducing the overestimation of precipitation observed in the simulation of the Global Model and improved temporal distribution of the same. / Com a finalidade de se obter uma previsão climática sazonal de precipitação de qualidade para o Estado do Rio Grande do Sul foram realizadas e avaliadas nove tipos de simulações que utilizam diferentes esquemas de parametrização cumulus, disponíveis no Modelo Regional Climático RegCM versão 4. As parametrizações testadas foram Grell com fechamento Arakawa e Schubert - AS e Fritsch e Chappell FC, MIT-Emanuel e convecção mista que é a utilização dos diferentes esquemas de convecção sobre a terra e o mar. O método de avaliação consistiu de análises estatísticas qualitativas e quantitativas das previsões climáticas de precipitação sobre cinco regiões do Rio Grande do Sul no período de agosto de 2013 a agosto de 2014. As estatísticas aplicadas foram Diagrama de Taylor, análise de erro aleatório e sistemático, índice de concordância e tabela de contingência. As previsões foram avaliadas através dos dados observados de estações meteorológicas do Instituto Nacional de Meteorologia. As análises mostraram que o RegCM4 apresentou correlações maiores e erros menores em comparação com o Modelo Global. Os melhores resultados foram observados na metade norte e oeste do Estado com as parametrizações Grell FC, Grell AS e a combinação de Emanuel simulado sobre o oceano e Grell AS sobre a terra. Apesar de algumas regiões não terem sido representadas adequadamente pelo Modelo Regional Climático RegCM4, ainda assim ele teve um bom desempenho em reduzir as superestimações de precipitação observadas na simulação do Modelo Global e melhora na distribuição temporal da mesma.

Les sécheresses des années 1970 accentuées dans les années 1980 et les aménagements hydroagricoles ont eu pour conséquence une certaine dégradation des ressources naturelles du delta intérieur du fleuve Niger (DIN). Cette dégradation des ressources naturelles associées à la pression croissante de la population sur le milieu a conduit à une gestion conflictuelle de ces ressources. De plus, le système d’exploitation actuel des ressources et le fonctionnement hydrologique font du DIN un écosystème complexe à gérer. Dans ces travaux, nous adoptons une approche intégrée et pluridisciplinaire pour aborder la problématique de gestion des ressources du DIN. L’approche mobilise l’hydrologie, la sociologie, l’agronomie et l’économie. L’analyse des données hydrologiques a permis de mettre en évidence une différence significative entre les dates de passage des maximums des fleuves Niger et Bani qui alimentent le DIN. Cette différence impacte les calendriers des principaux systèmes de production contribuant ainsi à exacerber les conflits. Les stratégies de gestion élaborées avec les parties prenantes, allient des pratiques de sécurisation et de maîtrise de l’eau, d’intensification, de régénération des ressources, de renforcement des capacités et des incitations économiques. Par ailleurs, dans un contexte de variabilité climatique accrue, nous avons analysé les perceptions d’un échantillon de pêcheurs, d’éleveurs et de cultivateurs sur la prévision du climat et sur l’intérêt des prévisions climatiques et des crues. Il ressort que les besoins en information climatique des usagers d’une plaine inondable comme le DIN se rapportent surtout aux dates de début de saison, de passage des maximums de crue, de la date d’arrivée des crues et des hauteurs maximales de crue. L’intérêt économique simulé de l’information climatique pour un système agraire de riziculture inondée présente un gain moyen de 10%. Par contre le coût des éventuelles erreurs de prévision serait particulièrement élevé pour les producteurs avec un déficit moyen sur le revenu de 24%. / The droughts of the 1970s intensified in the 1980s and the development of irrigation schemes upstream of the Inner Niger Delta in Mali (IND) have resulted in the degradation of natural resources. This degradation of natural resources associated with the increasing population pressure on the environment has led to a more conflictive management of these resources. In addition, the current operating system of resources and the hydrological functioning of the IND make the management of the ecosystem more complex.In this work, we adopt an integrated and multidisciplinary approach to address the IND resource management issues. The approach mobilizes hydrology, sociology, agronomy and economy.The analysis of hydrological data has highlighted a significant difference between the maximum passing dates of the two rivers that supply the IND. This difference impacts the schedules of the main production systems thus contributing to exacerbate conflict. The management strategies developed with stakeholders combine the practices of security and water control, intensification, resource regeneration, capacity building and economic incentives. Moreover, in a context of increased climate variability, we analyzed the perceptions of a sample of fishermen, herders and farmers on climate prediction and their interest of climate and flood forecasts.. It appears that the need for climate information of floodplain users as IND relate mainly to the onset date of rainy season, flood maximum passaging date, the arrival of floods and flood peak heights.The simulated economic value of climate information for an agrarian system of flooded rice obtains an average gain of 10%. As against the cost of possible prediction errors is particularly high for producers with an average deficit on income of 24%.