Thèse – Projecting the evolution of soil due to global change

Saba Keyvanshokouhi

Mercredi 7/03/2018 9h00-11h30

CEREGE, Grand Amphithéâtre, Technopôle de l'Arbois-Méditerranée, BP80, 13545 Aix-enProvence

Publication Keyvanshokouhi, S., Cornu, S., Samouëlian, A., & Finke, P. (2016). Evaluating SoilGen2 as a tool for projecting soil evolution induced by global change. The Science of the Total Environment, 571, 110-123.

saba.keyvanshokouhi@ugent.be saba.keyvanshokouhi@inra.fr

Prof. Dr. Peter Finke (encadrant) - Ghent University, Belgium Dr. Sophie Cornu (encadrant) - CEREGE-INRA, Aix-en-Provence, France Dr. François Lafolie (encadrant) - EMMAH-Sol, INRA, Avignon, France Dr. Isabelle Cousin (examiner) - INRA – Orléans, France Prof. Dr. Jean-Thomas Cornélis (examanteur) - Université de Liège, Belgium Prof. Dr. Daniela Sauer (examinateur) - Georg-August-Universität Göttingen, Göttingen, Germany Prof. Dr. Jan Vanderborght (rapporteur) - Forschungszentrum Jülich, Jülich, Germany Prof. Dr. Michael Sommer (rapporteur) - Leibniz Centre - Agricultural Landscape Research (ZALF), Müncheberg, Germany

In this study, we are discussing the use of soil evolution modelling in projecting the evolution of soil under climate, land use and agricultural practices change at the horizon of 2100. We first identified a suitable model, among the existing pedogenetic models, according to their process coverage and the sensitivity to variation of climate, land use and agricultural practices. The chosen model, SoilGen2.24, was demonstrated to be sensitive to those forcing factors and its main limitations were identified in three categories of over-simplified processes, missing processes and a simplifying assumption of constant soil volume. Additionally, a modular structure was proved to be fundamental to overcome these limitations. To overcome part of these limitations, we thus 1) built up the first fully modular soil evolution model, OC-VGEN, by using the process definitions of SoilGen2.24 model in a flexible environment, i.e. a modelling platform, VSoil; 2) tested different formalisms for some of the key processes responsible for the OC depth distribution, namely the root depth distribution, bioturbation and the depth evolution of the OC decomposition rate; 3) proposed a first, semimechanistic approach to account for soil volume change in a short to medium time scale soil evolution modelling. This latter is still a challenge for most of the soil evolution models. This work demonstrated the necessity of 1) a pedogenetic model when estimating soil response to forcing factors such as climate, land use and agricultural practices at the century scale; 2) a better definition and calibration of some still insufficiently known soil processes. We concluded that the newly developed OCVGEN model demonstrated a high potential to be set as a corner stone for the future of soil evolution modelling.