Dr Mahdi (Andre) Nakhavali is a Research Scholar in the Biodiversity and Natural Resources Program at the International Institute for Applied Systems Analysis (IIASA), Austria. His research focuses on terrestrial carbon and nutrient cycling, forest ecosystem dynamics and Earth system modelling, with particular emphasis on the processes that regulate the land carbon sink under global change. He holds a PhD in Mathematics from the University of Exeter, United Kingdom, where he also held research and teaching positions. His work has advanced the representation of dissolved organic carbon transport and phosphorus cycling in land surface models, and integrates process based forest models, land use frameworks and machine learning methods to assess climate impacts and mitigation pathways. He has contributed to several European research programmes on the carbon cycle, forests and climate policy. He serves as Associate Editor of Carbon Balance and Management and has convened scientific sessions at EGU and AGU.
The terrestrial biosphere is a major sink for anthropogenic CO2, but the size, spatial distribution and persistence of this sink remain highly uncertain. Much of this uncertainty is structural, reflecting an incomplete representation of the processes that control carbon uptake, storage and export in land surface and Earth system models. Credible projections therefore require not only improve parameterisation, but a more complete representation of the biogeochemical, ecological and human dimensions of terrestrial carbon dynamics. In this lecture, we will discuss how process based modelling and integrated land system assessment can help us to address the structural uncertainty. We will first look at the terrestrial aquatic continuum, the production, retention, leaching and lateral export of dissolved organic carbon from soils to inland waters and their contribution to the closure of terrestrial carbon budgets. We will also discuss the nutrient limitation, and how the coupled dynamics of carbon nitrogen phosphorus control the ecosystem productivity and buffer the vegetation response to elevated CO2, especially over highly weathered tropical soils. We will discuss how climate change, natural disturbance, forest management and competition for land and biomass together control the carbon dynamics and mitigation potential of forests at regional and global scales. Finally, we will discuss the role of machine learning emulators in complementing process based models to allow efficient exploration of high dimensional environmental and policy scenario space, while being constrained by the process representation of the parent models.Taken together, this lecture points to a central challenge for terrestrial carbon cycle science that the robustness of projections depends less on computational power than on the ecological realism, process completeness and integration of the systems being represented. Linking fundamental biogeochemistry with forest and land use dynamics offers a route to more complete and policy relevant assessments of the terrestrial carbon sink under global change.