Four different heat wave definitions (as outlined by Hajat et al, 2006; D’lppoliti et al, 2010; Anderson and Bell, 2011; Nairn and Fawcett, 2013) were used to characterize heat wave mortality across the United States. The goal was to identify if certain definitions perform better or worse than others. Overall every definition performed poorly, resulting in high False-Alarm Ratios and low Heidke Skill Scores. However, the Nairn and Fawcett (2013) and Anderson and Bell (2011) definitions performed consistently better than the other definitions. Despite several limitations, the results of this study indicate that the heat wave definitions need refinement. Additionally, the Nairn and Fawcett (2013) definition could be one of the best definitions for assessing heat waves and heat-related mortality.

Introduction: More frequent and intensive heat waves have been generally considered as a consequence of global climate change. In spite of the fact that there is a lack of generally accepted definition of heat waves in terms of intensity and duration, the adverse impacts of heat wave are undeniable. East Tennessee has experienced noticeably higher summer temperatures in the recent past, including possible deadly heat waves. This study aims at measuring the frequency and duration of heat waves in East Tennessee over the last ten years using various heat definitions extracted from epidemiological literature. Methods: A literature search was conducted in PubMed using key words: Heat wave, mortality and morbidity. The search located 311 peer-reviewed articles that were screened and 108 were reviewed for heat wave definitions. The definitions used in the 108 articles were then summarized and seven most commonly used definitions were selected for this study. Daily temperature (mean, maximum and minimum) data from three weather stations namely Tri-City Airport, Kingsport and Elizabethton in East Tennessee were collected for the period from 2008 to 2017 to test the seven different definitions of heat waves. The World Meteorological Organization (WMO)’s definition “daily maximum temperature for more than five consecutive days exceeding the average maximum temperature (for the 30-year normals period) by at least 9º Fahrenheit (5º Celsius)” was also tested and compared with the seven definitions extracted from epidemiological literature. Results: Using the WMO definition and temperature data from the Tri-City Airport Weather Station located in Bristol, TN, it was estimated that the total number of heat wave in the 10-year period was 58, which spanned a total of 1343 days, or 134.3 days per year on average. The results further indicate that the onset of heat wave days has been earlier within the past four years than the previous years with the earliest occurrence on February 24, 2017. Year 2016 is regarded as the hottest year in the period as virtually the whole of 2016 summer was a heat wave period. It is also worthy of note that the end-date of the heat wave episodes using the WMO definition is observed to have extended for the first time to the month of November as seen in 2016, which was not so for the other years. The heat wave episodes observed with the 97th percentile mean and 90th percentile minimum temperatures in Tri-City Airport Weather Station were lower than the WMO definition. Similar results were observed in Kingsport weather station. But in Elizabethton weather station, the 97th percentile mean and 90th percentile minimum temperatures produced higher episodes of heat waves but shorter days than the WMO definition. Conclusion: Heat wave is deadly and evidences suggest that its manifestation may become severe as more damages are being done to the atmosphere through pollution and deforestation. A reversal of this trend through early warning systems and preparedness is very important and urgent. Further research will look into the mortality health outcome associated with heat waves in this region.

A developed seamless extreme heat validation approach (Ford et al. 2018) is applied to three Subseasonl Experiment’s (SubX’s) medium-range forecast models, which arethe U.S. National Oceanic and Atmospheric Administration’s Earth System Research Laboratory FIM-iHYCOM (ESRL), the U.S. National Aeronautics and Space Administration’s Earth System Research Laboratory’s Goddard Earth Observing System Atmosphere-Ocean General Circulation Model, Version 5 (GMAO), and the U.S. National Centers for Environmental Prediction’s Global Ensemble Forecast System, version 11 (GEFS), for evaluating their heat wave predictability. Moreover, two land surface initializations, green vegetation fraction (GVF) and heat fluxes (LE/H), of each model are evaluated for understanding the interaction between heat wave predictability and the inconsistencies in the terrestrial segment of land-atmosphere feedbacks. The validation approach shows the overestimated autocorrelation of maximum temperature heat waves causing (1) the lowest reliability and overestimation of heat waves hindcasts, (2) lower heat wave hindcast skill of ensemble mean, and (3) higher discrimination between heat wave hindcast and observations of each ensemble member over lead times for all three models. Both ESRL and GEFS present the relationship between GVF and heat wave hindcast is positive, but negative relationship is shown on the GMAO. In addition, both ESRL and GEFS modelsunderestimate latent heat flux, but overestimate sensible heat flux in the Midwest. Therefore, for both ESRL and GEFS models, the relationship between heat wave and sensible heat fluxes (or GVF) is positive, and negative for the relationship between heat wave and latent heat flux (or evapotranspiration). In contrast, the GMAO model overestimates both latent and sensible heat fluxes in the Midwest. Therefore, for the GMAO model, the relationship between heat wave and latent/sensible heat fluxes (or GVF) is positive, and negative for the relationship between heat wave and evapotranspiration.

Alterations to the land surface can be attributed to both human activity and natural variability. Human activities, such as urbanization and irrigation, can change the conditions of the land surface by altering albedo, soil moisture, aerodynamic roughness length, the partitioning of net radiation into sensible and latent heat, and other surface characteristics. On the other hand, natural variability, manifested through changes in atmospheric circulation, can also induce land surface changes. These regional scale land surface changes, induced either by humans or natural variability, can effectively modify atmospheric conditions through land-atmosphere interactions. However, only in recent decades have numerical models begun to include representations of the critical processes driving changes at the land surface, and their associated effects on the overlying atmosphere. In this work we explore three mechanisms by which changes to the land surface–both anthropogenic and naturally induced–impact the overlying atmosphere and affect regional hydroclimate. The first land-atmosphere interaction mechanism explored here is land-use and land-cover change (LULCC) due to urban expansion. Such changes alter the surface albedo, heat capacity, and thermal conductivity of the surface. Consequently, the energy balance in urban regions is different from that of natural surfaces. To evaluate the changes in regional hydroclimate that could arise due to projected urbanization in the Phoenix–Tucson corridor, Arizona, my first study applied the Weather Research and Forecasting (WRF) with an Urban Canopy Model (UCM; which includes a detailed urban radiation scheme) coupled to the Noah land surface model to this region. Land-cover changes were represented using land-cover data for 2005 and projections to 2050, and historical North American Regional Reanalysis (NARR) data were used to specify the lateral boundary conditions. Results suggest that temperature changes are well defined, reflecting the urban heat island (UHI) effect within areas experiencing LULCC, whereas changes in precipitation are less certain (statistically less robust). However, the study indicates the likelihood of reductions in precipitation over the mountainous regions northeast of Phoenix and decreased evening precipitation over the newly urbanized area. The second land-atmosphere interaction mechanism explored here is irrigation which, while being an important anthropogenic factor affecting the local to regional water cycle, is not typically represented in regional climate models. In this (second) study, I incorporated an irrigation scheme into the Noah land surface scheme coupled to the WRF model. Using a newly developed water vapor tracer package (developed by Miguez-Macho et al. 2013), the study tracks the path of water vapor that evapotranspires from the irrigated regions. To assess the impact of irrigation over the California Central Valley (CCV) on the regional climate of the U.S. Southwest, I ran six simulations (for three dry and three wet years), both with and without the irrigation scheme. Incorporation of the irrigation scheme resulted in simulated surface air temperature and humidity that were closer to observations, decreased the depth of the planetary boundary layer over the CCV, and increased the convective available potential energy. The results indicated an overall increase in precipitation over the Sierra Nevada Range and the Colorado River Basin during the summer, with water vapor rising from the irrigated region moving mainly northeastward and contributing to precipitation in Nevada and Idaho. The results also indicate an increase in precipitation on the windward side of the Sierra Nevada Range and over the Colorado River Basin. The former is possibly linked to a sea-breeze type circulation near the CCV, while the latter is likely associated with a wave pattern related to latent heat release over the moisture transport belt. In the third study, I investigated the role of large-scale and local-scale processes associated with heat waves using the Modern Era-Retrospective Analysis for Research and Applications (MERRA) reanalysis, and evaluate the performance of the regional climate model ensemble used in the North America Regional Climate Change Program (NARCCAP) in reproducing these processes. The Continental US is divided into different climate divisions (following the convention of the National Climate Assessment) to investigate different mechanisms associated with heat waves. At the large scale, warm air advection from terrestrial sources is a driving factor for heat waves in the Northeast and Midwest. Over the western United States, reduced maritime cool air advection results in local warming. At the local scale, an antecedent precipitation deficit leads to the continuous drying of soil moisture, more energy being partitioned into sensible heat flux and acting to warm surface air temperatures, especially over the Great Plains. My analysis indicates that the NARCCAP simulated large-scale meteorological patterns and temporal evolution of antecedent local-scale terrestrial conditions are very similar to those of MERRA. However, NARCCAP overestimates the magnitude and underestimates the frequency of Northeastern and Midwestern US heat waves, partially due to anomalous heat advection through large-scale forcing. Overall, the aforementioned studies show that utilization of new parameterizations in land surface models, such as the urban canopy scheme and the irrigation scheme, allow us to understand the detailed physical mechanisms by which anthropogenic changes in the land surface can affect regional hydroclimate, and may thus help with informed decision making and climate adaptation/mitigation. In addition to anthropogenic changes of the land surface, humans are of course affecting the overlying atmosphere. Currently, NARCCAP is the best available tool we have to help us understand the effects of changes greenhouse gas induced climate change at the regional scale. The regional climate models participating in NARCCAP are able to realistically represent the dominant processes associated with heat waves, including the atmospheric circulation changes and the land-atmosphere interactions that drive heat waves. This lends credibility, when analyzing the projections of these models with increased GHG emissions, to the assessment of changes in heat waves under a future climate.

Background: During August 2003 and July 2006 there were observed records of high temperature and high concentrations of pollutants across Europe. The effect of heat waves led to significant increases in total mortality, respiratory mortality and cardiovascular mortality. This study evaluates the association between exposure to the heat waves and daily non- accidental mortality, respiratory mortality and cardiovascular mortality in Prague, the Czech Republic. Methods: The effect of heat waves in summer 2003 and 2006 on mortality was investigated using the negative binomial regression (type of the Poisson model). Counts of death were regressed on temperature, long-term trends, season, day of week and concentrations of pollutants (O3 levels, PM10 levels, NO2 levels, SO2 levels, CO levels). We used one day lag. Results: We found an association between the heat waves in summer 2003 and 2006 and daily mortality and mortality on respiratory and cardiovascular diseases. No statistically significant association was detected. The effect of the heat wave was more significant in women. Conclusions: The effect of the heat wave in August 2003 and July 2006 caused adverse effect on the mortality in Prague, though lower as compared to many other cities in Europe. Keywords: heat wave, mortality, temperature, ozone,...

No verão do Rio Grande do Sul (RS) são frequentes os registros de temperatura do ar acima de 30°C por alguns dias consecutivos, o que gera um fenômeno denominado de ondas de calor que associadas à atuação da massa de ar tropical continental, afetam o conforto térmico e o bem-estar das pessoas. Essa constante troca de massas de ar ocasiona extremos de temperatura e interfere no conforto térmico. Os estudos sobre o conforto térmico usando dados horários no RS são raros, e, portanto, são desconhecidas as diferentes condições térmicas para a sensação do conforto. O presente estudo analisou a sensação de conforto térmico e sua variabilidade para nove municípios do RS empregando o Índice de Temperatura Efetiva (ITE) para os verões de 2006/2007 a 2013/2014. Essa sensação varia no RS de acordo com a região geográfica, onde se identificou três padrões de conforto térmico na análise do ITE máximo e mínimo. Os municípios de menor altitude, como Porto Alegre, Santa Maria e Uruguaiana, são os que mais registraram desconforto térmico. A faixa térmica do calor moderado ocorre com maior frequência no ITE das máximas, seguido pela faixa de calor. A sensação de calor extremo é registrada nesses três municípios. No ITE das mínimas a faixa do confortável é predominante, seguida pela de ligeiramente quente. A maior continentalidade de Santa Maria e Uruguaiana faz com que esses dois municípios sejam os mais quentes do RS. Bento Gonçalves, Caçapava do Sul e Erechim, com altitudes de 623 m, 421 m e 777 m, respectivamente, registram sensação térmica de calor moderado nas máximas de ITE, seguido pela sensação de ligeiramente quente. Já as mínimas correspondem à faixa do confortável e de ligeiramente fresco. Torres e Rio Grande, que são litorâneos e ao nível do mar, tiveram registros na faixa do calor moderado no ITE máximo e de confortável no ITE mínimo. São José dos Ausentes, que apresenta a maior altitude, registrou ITE mais baixo, com índice na faixa do ligeiramente quente nas máximas e de frescor no registro das mínimas. / In summer of Rio Grande do Sul (RS) are frequent air temperature records above 30°C for a few several days, which generates a phenomenon called heat waves associated with the action of the continental tropical air mass, affecting thermal comfort and the well-being of people. This constant exchange of air masses causes extreme temperature events and interferes in the thermal comfort. Studies about thermal comfort using hourly data of RS are rare, and therefore are unknown the different thermal conditions for the sensation of thermal comfort. This study analyzed the sensation of thermal comfort and variability for nine municipalities in the RS using the Effective Temperature Index (ETI) for the summers of 2006/2007 to 2013/2014. This sensation varies in RS according to the geographical region, where it was identified three patterns of thermal comfort in the analysis of maximum and minimum ETI. The municipalities with the lowest altitude, in Porto Alegre, Santa Maria and Uruguaiana, are the most recorded thermal discomfort. The temperature range of moderate heat occurs most frequently in the maximum ETI, followed by heat range. The sensation of extreme heat have been recorded in these three municipalities. In minimum ETI, the comfortable range is predominant, followed by slightly warm. The continentality effect of Santa Maria and Uruguaiana makes these two cities the hottest places of the RS. Bento Gonçalves, Caçapava do Sul and Erechim, with altitudes of 623 m, 421 m and 777 m, respectively, recorded thermal sensation moderate heat in maximum ETI, followed by sensation slightly warm. Already the minimum correspond to the comfortable and slightly cool range. Torres and Rio Grande, which are coastal places at the sea level, had records in the range of moderate heat maximum ETI and comfortable minimum ETI. São José dos Ausentes, with the highest altitude, recorded lower ETI, with index in the range of slightly warm at maximum freshness in the record minimum.

Doctor of Philosophy / Department of Geography / John A. Harrington Jr / The climate of the Middle East is warming and extreme hot temperature events are becoming more common, as observed by the significant upward trends in mean and extreme temperatures during the last few decades. Climate modeling studies suggest that the frequency, intensity, and duration of extreme temperature events are expected to increase as the global and local climate continues to warm. Existing literature about heat waves (HWs) in Saudi Arabia provides information about HW duration using a single index, without considering the observed effects of climate change and the subtropical arid climate. With that in mind, this dissertation provides a series of three stand-alone papers evaluating temporal, geographic, and atmospheric aspects of the character of warm season (May-September) HWs in Saudi Arabia for 1985 to 2014. Chapter 2 examines the temporal behavior(s) of the frequency, duration, and intensity of HWs under the observed recent climate change. Several issues are addressed including the identification of some improved methodological practices for HW indices. A time-sensitive approach to define and detect HWs is proposed and assessed. HW events and their duration are considered as count data; thus, different Poisson models were used for trend detection. Chapter 3 addresses the spatio-temporal patterns of the frequency and intensity of hot days and nights, and HWs. The chapter reemphasizes the importance of considering the on-goings effects of climate warming and applies a novel time-series clustering approach to recognize hot temperature event behavior through time and space. Chapter 4 explores the atmospheric circulation conditions that are associated with warm season HW event occurrence and how different HWs aspects are related to different circulation types. Further, possible teleconnections between HWs and sea surface temperature (SST) anomalies of nearby large bodies are examined. Results from Chapters 2 and 3 detected systematic upward trends in maximum and minimum temperatures at most of the 25 stations, suggesting an on-going change in the climatology of the upper-tail of the frequency distribution. The analysis demonstrated the value of using a time-sensitive approach in studying extreme thermal events. Different patterns were observed over time and space not only across stations but also among extreme temperature events (i.e., hot days and nights, and HWs). The overall results suggest that not only local and regional factors, such as elevation, latitude, land cover, atmospheric humidity, and distance from a large body of water, but also large-scale factors such as atmospheric circulation patterns are responsible for the observed temporal and spatial patterns. Chapter 4 confirmed that as the Indian Summer Monsoon Trough and the Arabian heat low were key atmospheric features related to HW days. SST anomalies seemed to be a more important factor for HWs intensity. Extreme thermal events in Saudi Arabia tended to occur during regional warming due to atmospheric circulation conditions and SSTs teleconnections. This study documents the value of a time-sensitive approach and should initiate further research as some of temporal and spatial variabilities were not fully explained

No verão do Rio Grande do Sul (RS) são frequentes os registros de temperatura do ar acima de 30°C por alguns dias consecutivos, o que gera um fenômeno denominado de ondas de calor que associadas à atuação da massa de ar tropical continental, afetam o conforto térmico e o bem-estar das pessoas. Essa constante troca de massas de ar ocasiona extremos de temperatura e interfere no conforto térmico. Os estudos sobre o conforto térmico usando dados horários no RS são raros, e, portanto, são desconhecidas as diferentes condições térmicas para a sensação do conforto. O presente estudo analisou a sensação de conforto térmico e sua variabilidade para nove municípios do RS empregando o Índice de Temperatura Efetiva (ITE) para os verões de 2006/2007 a 2013/2014. Essa sensação varia no RS de acordo com a região geográfica, onde se identificou três padrões de conforto térmico na análise do ITE máximo e mínimo. Os municípios de menor altitude, como Porto Alegre, Santa Maria e Uruguaiana, são os que mais registraram desconforto térmico. A faixa térmica do calor moderado ocorre com maior frequência no ITE das máximas, seguido pela faixa de calor. A sensação de calor extremo é registrada nesses três municípios. No ITE das mínimas a faixa do confortável é predominante, seguida pela de ligeiramente quente. A maior continentalidade de Santa Maria e Uruguaiana faz com que esses dois municípios sejam os mais quentes do RS. Bento Gonçalves, Caçapava do Sul e Erechim, com altitudes de 623 m, 421 m e 777 m, respectivamente, registram sensação térmica de calor moderado nas máximas de ITE, seguido pela sensação de ligeiramente quente. Já as mínimas correspondem à faixa do confortável e de ligeiramente fresco. Torres e Rio Grande, que são litorâneos e ao nível do mar, tiveram registros na faixa do calor moderado no ITE máximo e de confortável no ITE mínimo. São José dos Ausentes, que apresenta a maior altitude, registrou ITE mais baixo, com índice na faixa do ligeiramente quente nas máximas e de frescor no registro das mínimas. / In summer of Rio Grande do Sul (RS) are frequent air temperature records above 30°C for a few several days, which generates a phenomenon called heat waves associated with the action of the continental tropical air mass, affecting thermal comfort and the well-being of people. This constant exchange of air masses causes extreme temperature events and interferes in the thermal comfort. Studies about thermal comfort using hourly data of RS are rare, and therefore are unknown the different thermal conditions for the sensation of thermal comfort. This study analyzed the sensation of thermal comfort and variability for nine municipalities in the RS using the Effective Temperature Index (ETI) for the summers of 2006/2007 to 2013/2014. This sensation varies in RS according to the geographical region, where it was identified three patterns of thermal comfort in the analysis of maximum and minimum ETI. The municipalities with the lowest altitude, in Porto Alegre, Santa Maria and Uruguaiana, are the most recorded thermal discomfort. The temperature range of moderate heat occurs most frequently in the maximum ETI, followed by heat range. The sensation of extreme heat have been recorded in these three municipalities. In minimum ETI, the comfortable range is predominant, followed by slightly warm. The continentality effect of Santa Maria and Uruguaiana makes these two cities the hottest places of the RS. Bento Gonçalves, Caçapava do Sul and Erechim, with altitudes of 623 m, 421 m and 777 m, respectively, recorded thermal sensation moderate heat in maximum ETI, followed by sensation slightly warm. Already the minimum correspond to the comfortable and slightly cool range. Torres and Rio Grande, which are coastal places at the sea level, had records in the range of moderate heat maximum ETI and comfortable minimum ETI. São José dos Ausentes, with the highest altitude, recorded lower ETI, with index in the range of slightly warm at maximum freshness in the record minimum.

No verão do Rio Grande do Sul (RS) são frequentes os registros de temperatura do ar acima de 30°C por alguns dias consecutivos, o que gera um fenômeno denominado de ondas de calor que associadas à atuação da massa de ar tropical continental, afetam o conforto térmico e o bem-estar das pessoas. Essa constante troca de massas de ar ocasiona extremos de temperatura e interfere no conforto térmico. Os estudos sobre o conforto térmico usando dados horários no RS são raros, e, portanto, são desconhecidas as diferentes condições térmicas para a sensação do conforto. O presente estudo analisou a sensação de conforto térmico e sua variabilidade para nove municípios do RS empregando o Índice de Temperatura Efetiva (ITE) para os verões de 2006/2007 a 2013/2014. Essa sensação varia no RS de acordo com a região geográfica, onde se identificou três padrões de conforto térmico na análise do ITE máximo e mínimo. Os municípios de menor altitude, como Porto Alegre, Santa Maria e Uruguaiana, são os que mais registraram desconforto térmico. A faixa térmica do calor moderado ocorre com maior frequência no ITE das máximas, seguido pela faixa de calor. A sensação de calor extremo é registrada nesses três municípios. No ITE das mínimas a faixa do confortável é predominante, seguida pela de ligeiramente quente. A maior continentalidade de Santa Maria e Uruguaiana faz com que esses dois municípios sejam os mais quentes do RS. Bento Gonçalves, Caçapava do Sul e Erechim, com altitudes de 623 m, 421 m e 777 m, respectivamente, registram sensação térmica de calor moderado nas máximas de ITE, seguido pela sensação de ligeiramente quente. Já as mínimas correspondem à faixa do confortável e de ligeiramente fresco. Torres e Rio Grande, que são litorâneos e ao nível do mar, tiveram registros na faixa do calor moderado no ITE máximo e de confortável no ITE mínimo. São José dos Ausentes, que apresenta a maior altitude, registrou ITE mais baixo, com índice na faixa do ligeiramente quente nas máximas e de frescor no registro das mínimas. / In summer of Rio Grande do Sul (RS) are frequent air temperature records above 30°C for a few several days, which generates a phenomenon called heat waves associated with the action of the continental tropical air mass, affecting thermal comfort and the well-being of people. This constant exchange of air masses causes extreme temperature events and interferes in the thermal comfort. Studies about thermal comfort using hourly data of RS are rare, and therefore are unknown the different thermal conditions for the sensation of thermal comfort. This study analyzed the sensation of thermal comfort and variability for nine municipalities in the RS using the Effective Temperature Index (ETI) for the summers of 2006/2007 to 2013/2014. This sensation varies in RS according to the geographical region, where it was identified three patterns of thermal comfort in the analysis of maximum and minimum ETI. The municipalities with the lowest altitude, in Porto Alegre, Santa Maria and Uruguaiana, are the most recorded thermal discomfort. The temperature range of moderate heat occurs most frequently in the maximum ETI, followed by heat range. The sensation of extreme heat have been recorded in these three municipalities. In minimum ETI, the comfortable range is predominant, followed by slightly warm. The continentality effect of Santa Maria and Uruguaiana makes these two cities the hottest places of the RS. Bento Gonçalves, Caçapava do Sul and Erechim, with altitudes of 623 m, 421 m and 777 m, respectively, recorded thermal sensation moderate heat in maximum ETI, followed by sensation slightly warm. Already the minimum correspond to the comfortable and slightly cool range. Torres and Rio Grande, which are coastal places at the sea level, had records in the range of moderate heat maximum ETI and comfortable minimum ETI. São José dos Ausentes, with the highest altitude, recorded lower ETI, with index in the range of slightly warm at maximum freshness in the record minimum.

Les vagues de chaleur sont encore peu documentées au Sahel, malgré un impact très fort sur les populations. L'objectif de cette thèse est d'apporter de nouvelles connaissances et une meilleure compréhension des processus et mécanismes mis en jeu dans les vagues de chaleur au Sahel au printemps. L'étude se structure autour de trois axes : (i) la détection des vagues de chaleur, (ii) la documentation de leurs caractéristiques morphologiques, dynamiques et thermodynamiques ainsi que leurs tendances climatiques et (iii) la compréhension des processus et mécanismes en jeu lors de ces événements. Dans un premier temps, une nouvelle méthodologie d’identification des vagues de chaleur, définies ici comme une forte augmentation de la température à des échelles synoptiques à intra-saisonnières, a été développée. Cette définition permet de détecter des vagues de chaleur « météorologiques », associées au temps sensible et qui constituent un véritable enjeu pour la prévision. La détection repose sur trois étapes : (i) un filtrage sélectionne les échelles souhaitées, ici synoptiques à intra-saisonnières ; (ii) les valeurs extrêmes des anomalies de températures ainsi calculées sont conservées, grâce à un seuil limite fixe égal au quantile Q90 ; et (iii) une contrainte morphologique est finalement appliquée pour identifier les événements de grande échelle cohérents dans le temps et l’espace. Par ailleurs, les températures maximales (Tmax) et minimales (Tmin) sont considérées séparément car leurs fluctuations résultent de processus distincts. Cette méthodologie a été appliquée à plusieurs types de jeux de données : un produit basé sur des observations locales (BEST), trois réanalyses météorologiques (ERA-Interim, NCEP2 et MERRA) et des simulations climatiques (CMIP5). Les vagues de chaleur détectées sont ensuite documentées en termes de caractéristiques morphologiques et thermodynamiques, et leurs tendances climatiques sont analysées. L’occurrence de ces événements se concentre particulièrement en début de printemps, et la tendance à la hausse de leurs températures s’explique principalement par le réchauffement moyen au Sahel. Les biais moyens de température dans les simulations, atteignant plus de 5°C dans certaines simulations climatiques, expliquent la plus grande partie des erreurs sur les températures des vagues de chaleur. Les événements détectés dépendent aussi de la méthodologie utilisée, c’est pourquoi nos résultats sont comparés avec ceux obtenus à partir d’autres méthodologies communément utilisées dans la littérature. Les processus et mécanismes mis en jeu lors des vagues de chaleur sont ensuite étudiés. Au printemps, les sols sont secs et le flux de chaleur latente très faible. Les vagues de chaleur détectées sur les Tmax sont anormalement sèches, celles en Tmin anormalement humides. Une dépression est présente lors des deux types d’événements et centrée sur la zone impactée. Les vagues de chaleur diurnes sont associées à un vent de nord-ouest alors que les vagues de chaleur nocturnes correspondent à un vent de sud-ouest, qui favorise l’advection d’humidité par le flux de mousson. La vapeur d’eau joue un rôle majeur sur les vagues de chaleur nocturnes, amplifiant l’effet de serre de l’atmosphère. Une première analyse indique que les vagues de chaleur détectées en Tmax ne sont pas associées à des couches limites plus profondes. Ces résultats soulignent qu’au Sahel, les vagues de chaleur sont associées à des mécanismes différents de ceux observés dans d’autres régions du monde, comme en Europe, où les vagues de chaleur font souvent intervenir des conditions anticycloniques et un assèchement des sols. La méthodologie a été finalement adaptée à la prévision des vagues de chaleur en temps réel et implémentée sur le site Internet http://acasis.sedoo.fr/. Ce travail a permis de suivre et prévoir en temps réel les vagues de chaleur impactant l’Afrique de l’ouest pendant les printemps 2016-2017. / Heat waves are still undocumented over the Sahel, despite great consequences over the population. This PhD aims at improving the knowledge of Spring Sahelian heat waves, together with a better understanding of the processes and mechanisms operating during those events. The work is organized in three parts: the first one on the heat waves detection, the second one on their morphological, dynamical and thermodynamical characteristics as well as their climatological trends and the third one on the understanding of the processes and mechanisms involved in these events. First, a new methodology has been developed to detect heat waves, defined as synoptic to intraseasonal fluctuations of temperature. This definition allows the identification of “meteorological” heat waves, which are an important issue for weather forecast. The new methodology involved three steps: (i) a temperature filtering to select specific temporal scales; (ii) the determination of a temperature thresholds (90th percentile) and (iii) finally the application of a morphological constraint. This methodology was applied separately to the maximum (Tmax) and minimum (Tmin) temperatures, as the mechanisms that drive their fluctuations are distinct. The method was used with several types of datasets: one based on local observations (BEST), three meteorological reanalyses (ERA-Interim, NCEP2 and MERRA) and on several climate model CMIP5 simulations. The detected heat waves are then documented in terms of morphological and thermodynamic characteristics and in terms of trends over the past sixty years. From March to July, the further into the season, the shorter and the less frequent the heat waves become, and from 1950 to 2012, these synoptic to intraseasonal heat waves do not tend to be more frequent but they become warmer. This trend follows the long-term trend of temperature over the Sahel. Simulation temperature biases, which can reach more than 5°C in some climate simulations, explain most of the spread in the temperatures of the detected heat waves. Finally, the method used to define heat waves has an impact on the events, thus a comparison with other methodologies was carried out. Physical mechanisms involved in heat waves are then analyzed. During the Sahelian Spring, the soil is dry and the latent heat flux is very low. Daytime heat waves are abnormally dry, whereas nighttime events are abnormally humid. A depression is involved in both cases, located over the heat wave region for the Tmax heat waves and a larger-scalenorthward depression for the Tmin heat waves. North-easterly winds are predominant for the Tmax events, while south-westerly winds prevail during the Tmin events thus strengthening the moist air advection from the monsoon flow. The water vapor plays a major role on the nighttime heat waves via an enhanced greenhouse effect. Observations also highlight that boundary layers are not deeper during daytime heat waves. In Europe, heat waves are often associated with anticyclonic conditions, stronger sensible heat fluxes and deeper boundary layers: Sahelian heat waves are then associated with different mechanisms. The heat wave detection was adapted to numerical weather forecast and implemented on the website http://acasis.sedoo.fr/. It has allowed to monitor and forecast in real-time West Africa spring heat waves in 2016 and 2017.