Le système d'upwelling péruvien est l'un des systèmes marins côtiers les plus productifs de l’océan mondial. Le vent de surface le long de la côte est le principal moteur de l'upwelling. Cette thèse vise à étudier la variabilité du vent côtier et ses processus lors du réchauffement de la couche de surface, à différentes échelles de temps: (1) des échelles de temps interannuelles, correspondant aux événements El Niño, et (2) des échelles de temps multi-décadaires résultant du changement climatique régional. Une série de domaines emboités d’un modèle atmosphérique régional est utilisée pour simuler le vent de surface. Dans la première partie de la thèse, on étudie les processus responsables de l'augmentation, contre-intuitive, du vent observée au large du Pérou au cours de la période El Niño 1997-1998. Des expériences de sensibilité montrent que le réchauffement inh de la omogène des eaux de surface, plus important dans le nord, entraîne un gradient de pression accru le long côte, accélérant le vent. Dans une seconde partie de la thèse, l’évolution des vents côtiers est étudiée dans le scénario du «pire cas» du changement climatique RCP8.5. Forcés par le gradient de pression le long de la côte, les vents diminuent en été, tandis qu’ils s’accroissent en hiver, renforçant ainsi légèrement le cycle saisonnier. / The Peruvian upwelling system is one of the most productive coastal marine systems of the world ocean. As in other upwelling systems, alongshore surface wind is the main driver of the coastal upwelling. This thesis aims to study the coastal wind variability and the processes responsible for it during the ocean surface layer warming conditions, at different time scales: (1) interannual time scales, corresponding to El Niño events and (2) multi decadal time scales resulting from regional climate change. A suite of regional atmospheric model embedded domains is used to simulate the surface winds. In the first part of the thesis, the counter-intuitive wind increase observed off Peru during the 1997-1998 El Niño is studied. Sensitivity experiments show that the inhomogenous alongshore surface warming, larger in the north, drives an enhanced alongshore pressure gradient that accelerates the alongshore wind. In the second part of the thesis, the evolution of coastal wind changes is investigated under the “worst case” RCP8.5 climate change scenario. Mainly driven by the alongshore pressure gradient, summer winds decrease whereas winter winds increase, thus slightly reinforcing the seasonal cycle.

Dissertation submitted to the Faculty of Science, University of the Witwatersrand, for completion of the Degree of' Master of Science / The atmospheric response of the Colorado State University Regional Atmospheric Modelling System (RAMS) to sea-surface temperature anomaliesis investigated. A period of four days was chosen from 21 to 24 January 1981, where focus was placed on the development and dissipation of a tropical-temperate trough across Southern Africa. Previous experimenting this mesoscalenumerical model have detemined the kinematic, moisture, and thermodynamic nature of these synoptic features. The research in this dissertation focuses specifically on the sensitivity of the numerical model's simulated responses to positive sea-surface temperature anomalies. Three separate experiments were devised, in which positive anomalous temperatures were added to the ocean surface north of Madagascar (in the tropical Indian Ocean), at the region of the Agulhas Current retroflection, and along the tropical African west coast (in the Northern Benguela and Angola currents). The circulation aspects of each sensitivity test were investigated through the comparison of simulated variables such as vapour and cloud mixing ratios, temperature, streamlines and vertical velocity, with the same variables created by a control simulation. The results indicate that for the first sensitivity test, (the Madagascar anomaly), cyclogenesis was initiated over the area of modified sea temperatures which resulted in a marginal decrease in continental precipitation. The second sensitivity test (over the Agulhas retroflection) produced a much smaller simulated response to the addition of anomalously warm sea temperatures than the tropical Indian Ocean anomaly. Instability and precipitation values increased over the anomalously warm retroflection region, and were slowly transferred along the westerly wave perturbation and the South African east coast. The third sensitivity experiment showed a predominantly localised simulated increase in precipitation over Gabon and the Congo, with the slow southward progression of other simulated circulation differences taking place. The small perturbations in each of the simulated meteorological responses are consistent with the expected climate response to anomalously warm sea-surface temperatures in those areas. / AC 2018