Abstract :
[en] Sensible (H) and latent (LE) heat are frequently studied forms of energy exchanges between an ecosystem and the atmosphere. The distribution of available energy between them reflects many underlying eco-physiological processes that may or not be linked. In order to identify these processes, this study investigates the evolution of both mentioned surface turbulent heat fluxes on an annual and multi-annual scale while taking their interactions into account. Its purpose is, firstly, to bring new insights on well-known H/LE patterns and, secondly, to derive underlying correlations between processes governing heat exchanges over an ecosystem. To achieve such objectives, a time-frequency analysis tool, namely the continuous wavelet transform, is implemented owing to its compatibility with the investigation of non-stationary phenomena. This methodology is applied on a fourteen-year dataset of half-hourly fluxes of H and LE obtained by eddy-covariance over a young beech forest at ICOS Hesse site, North-eastern France. Both the Morlet and Mexican Hat wavelets are used, due to their respective characteristics and their suitability for investigating environmental fluxes. At this stage, many wavelet-based results are available along with preliminary conclusions on the ecosystem’s behavior. In particular, this study highlights several periodic components at intermediate scales, ranging from days to years, throughout the growing season and links them to environmental drivers. Among these drivers, incident radiation and soil water content seem to play a key role in the energy partitioning of the forest ecosystem. Ultimately, this work explores the opportunities of long-term analyses of microclimatic measurements offered by the use of the continuous wavelet transform, which is expected to become an increasingly popular tool at the dawn of current global climate change.