Abstract
French Guiana is a French territory located in the inter-tropical zone (ITZ). The ITZ is an area with highly variable dynamic in which we encounter significant amounts of convective clouds. Consequently the solar energy available at the ground is highly variable. This variability causes economical and technical challenges to fully exploit this resource. This thesis dissertation aims to answer the following scientific issue: How could the solar irradiance be assessed and forecast accurately in the ITZ to increase the penetration rate of this intermittent renewable energy into the electricity grid? To answer this scientific issue, we use two tools: Heliosat-II (H-II) and Weather and research forecast (WRF). We used these tools in order to produce improved GHI estimates in the inter-tropical zone. The first chapter introduces the thesis and the research issue. The second chapter presents a modification to H-II; with this modification H-II can account for cloud absorption. The GHI estimates from modified H-II provide therefore tools for decision making in the ITZ. These tools allow one identifying the most suitable locations to install solar facilities in the ITZ with respect to both solar potential and surrounding facilities that favor grid stability. In the third chapter we study first the accuracy in the ITZ of the GHI forecasts from integrated forecast system (IFS) and global forecast system (GFS) numerical weather prediction model (NWP). We validate the accuracy of these downloaded products by comparison with ground measurements from three countries located in the ITZ that have tropical climate. This study aims to fill the gap with regard to the accuracy of global NWP model in the ITZ. Second we propose a methodology to calibrate WRF to produce improved GHI forecasts in the ITZ. The goal is to restrain and select the minimum number of simulations to run, to obtain improved GHI forecasts compared to a non-calibrated model. This methodology to calibrate WRF is validated in French Guiana by comparison with the GHI forecasts of AROME NWP model and ground measurements. The fourth chapter deals with the use of an hybrid 3D variational (3D-Var) ensemble transform Kalman filter (ENTKF) to further improve the GHI forecasts of calibrated WRF in the ITZ. This methodology originally used in the tracking of extreme convection events such as cyclones is applied for the first time for GHI forecasts. This methodology applied to the ITZ therefore allows obtaining improved GHI forecasts which makes easier monitoring the electricity production from solar facilities.