• Energy Research

In Belgium, the future response of the climate to increasing greenhouse gas concentration is not clear, especially with regard to the perturbations of the precipitation regime, snow cover, and global radiation. On the one hand, existing studies show results which differ strongly either according to the future scenario, or from one model to another. On the other hand, there is even an absence of studies focussing on Belgium regarding future changes in snow cover and global radiation. Given their potential impacts on the society (water management, energy supply, biodiversity, tourism), future changes in precipitation, snow cover, and global radiation require further research. As the orography, the exposition to the dominant winds, and the proximity of the North Sea determine a large spatial variability in the Belgian climate, the latter requires a fine representation of these features to be properly simulated. Compared to global climate models (GCM), regional climate models (RCM) are recognized for their ability to represent climatic phenomena with higher spatial resolutions. In the framework of this doctoral thesis, the RCM MAR (for "Modèle Atmosphérique Régional" in French), which is developed at the Laboratory of Climatology and Topoclimatology of the University of Liège, was applied for the first time to Belgium. The aim was first to assess the performances of MAR over Belgium and then to study the current and future evolution of hydroclimatic conditions favouring floods, and also the current and future evolution of global radiation. For this purpose, historical simulations were performed over 1959-2014. Future projections (2006-2100) were then performed under the most pessimist IPCC future scenario (RCP8.5). The horizontal resolution used for both historical and future simulations is 5 km. By comparing the MAR outputs to ground-based measurements from 20 weather stations over 2008-2014, the results show that MAR successfully simulates the spatial and temporal variability of the Belgian climate. In fact, the biases found in the MAR results are non-significant and the correlation coefficients are satisfying with regard to temperature, precipitation, snow height, global radiation and cloudiness. The MAR results are particularly satisfying during the winter months and in High Belgium where the climate is the coldest. Regarding hydroclimatic conditions favouring floods, we focused on the Ourthe catchment. In this river, about 70 % of floods occur during the winter months and result from either the rapid melting of the snow pack covering the Ardennes eventually combined with rainfall or abundant rainfall alone. The current evolution of hydroclimatic conditions favouring floods was first assessed for the period 1959-2010. Conditions favouring floods in the Ourthe River present a negative trend over 1959–2010 as a result of a decrease in snow accumulation and a shortening of the snow season. Regarding the impact of the evolution of extreme precipitation events on hydroclimatic conditions favouring floods, the signal is less clear because the trends depend on the data used to force the MAR model. By the end of the 21st century, under the most pessimist scenario, the results show an acceleration of the snow cover depletion resulting in a decrease in conditions favouring floods. Further, the impact of the evolution of extreme precipitation events on hydroclimatic conditions favouring floods, no significant change was found although these trends are subject to uncertainties due to the deficiencies of the convective scheme of MAR. Regarding global radiation, its current evolution was first assessed for the period 1959-2010. In addition, we consider two distinct periods in our analysis: 1959-1979 (dimming) and 1980-2010 (brightening). For both the dimming and the brightening periods, our results show that the annual global radiation trends are mainly driven by global radiation changes in spring and summer. The increase in global radiation observed in Belgium since the 1980s and especially since the 2000s could mainly be explained by a decrease in low and medium cloud cover. This would strengthen the effect of the decrease in aerosol load on global radiation that has been observed in Europe since the 1980s. The origin of these changes in cloudiness is not clear and could result from changes in both aerosol-cloud interactions and atmospheric-circulation, such as more frequent tropical air advections and more frequent anticyclonic conditions over Western Europe due to the poleward shift of extratropical storm tracks. These changes in the atmospheric circulation may result from global warming and may persist in the future. In fact, by the end of the 21st century, under the most pessimist scenario, the models simulate an increase in the blocking regime frequency in summer over Europe. For Belgium, this implies more frequent anticyclonic conditions favouring cloudless conditions. The future projections performed with MAR exhibit significant decreasing total cloud cover, and particularly decreasing low and medium cloud cover. However, this declining cloud cover leads to contrasting changes in global radiation depending on the data used to force MAR. This thesis was funded by the Fonds pour la formation à la Recherche dans l'Industrie et dans l'Agriculture (Communauté française de Belgique) - FRIA (BE)

The first aim of this study is to determine if changes in precipitation and more specifically in convective precipitation are projected in a warmer climate over Belgium. The second aim is to evaluate if these changes are dependent on the convective scheme used. For this purpose, the regional climate model Modèle Atmosphérique Régional (MAR) was forced by two general circulation models (NorESM1-M and MIROC5) with five convective schemes (namely: two versions of the Bechtold schemes, the Betts–Miller–Janjić scheme, the Kain–Fritsch scheme, and the modified Tiedtke scheme) in order to assess changes in future precipitation quantities/distributions and associated uncertainties. In a warmer climate (using RCP8.5), our model simulates a small increase of convective precipitation, but lower than the anomalies and the interannual variability over the current climate, since all MAR experiments simulate a stronger warming in the upper troposphere than in the lower atmospheric layers, favoring more stable conditions. No change is also projected in extreme precipitation nor in the ratio of convective precipitation. While MAR is more sensitive to the convective scheme when forced by GCMs than when forced by ERA-Interim over the current climate, projected changes from all MAR experiments compare well.

The CORDEX.be project created the foundations for Belgian climate services by producing high-resolution Belgian climate information that (a) incorporates the expertise of the different Belgian climate modeling groups and that (b) is consistent with the outcomes of the international CORDEX (“COordinated Regional Climate Downscaling Experiment”) project. The key practical tasks for the project were the coordination of activities among different Belgian climate groups, fostering the links to specific international initiatives and the creation of a stakeholder dialogue. Scientifically, the CORDEX.be project contributed to the EURO-CORDEX project, created a small ensemble of High-Resolution (H-Res) future projections over Belgium at convection-permitting resolutions and coupled these to seven Local Impact Models. Several impact studies have been carried out. The project also addressed some aspects of climate change uncertainties. The interactions and feedback from the stakeholder dialogue led to different practical applications at the Belgian national level. Keywords: Regional downscaling, Climate impact modeling, Statistical downscaling, Dynamical downscaling, Local Impact Models, Climate change, EURO-CORDEX, Uncertainty estimation, Regional Climate Model, Climate Belgium, Water vapour observations