Soutenance de thèse de Sarah Coffinet

Validation and application of lipid biomarkers to reconstruct environmental changes in East Africa

Le jury sera composé de:
Vincent GROSSI, Directeur de recherche CNRS, France, Rapporteur
Rich D. PANCOST, Professeur, Université de Bristol, Royaume-Uni, Rapporteur
Jérémy JACOB, Chargé de recherche CNRS, France, Examinateur
André-Jean FRANCEZ, Maître de conférences, Université de Rennes 1, France, Examinateur
François BAUDIN, Professeur, UPMC, France, Examinateur
Amos E. MAJULE, Professeur, Université de Dar es Salaam, Tanzanie, Invité
Sylvie DERENNE, Directeur de recherche CNRS, France, Directrice de thèse
Arnaud HUGUET, Chargé de recherche CNRS, France, co-encadrant de thèse    

 The Rungwe Volcanic Province (RVP) in East Africa offers a wide diversity of continental archives (loess-paleosols sequences, peatlands, lakes), allowing the investigation of past environmental changes in tropical continents. This work focused on the validation of glycerol dialkyl glycerol tetraether- (GDGT) and n-alkane-derived proxies. Analyses of these two biomarkers along altitudinal transects allowed assessing their ability to record temperature changes in the region. GDGTs were found to be robust temperature proxies in East Africa, and a regional calibration was established to improve temperature reconstructions from soil archives. In contrast, investigation of n-alkane hydrogen isotopic composition (δ²Hwax) in surface soils revealed that its variations seemed to be impacted by additional biological and physical parameters than temperature, preventing its generalization at a regional scale. Application of GDGT and n-alkane proxies to continental archives revealed past environmental changes over the Late Holocene in the RVP. A late Holocene synchronous temperature maximum was especially observed in a peat core and sediment core from two sites of the RVP. A multi-proxy approach was applied to the peat core, revealing that an ecosystem change from a lake to a peatland occurred ca. 2000 yrs BP ago. The present work showed that GDGTs and n-alkanes are promising biomarkers because of their sensitivity to slight climate variations, but that they need to be combined with other proxies to accurately reconstruct environmental changes.