Research
Research Portfolio
The Ocean Biogeochemistry group at MPI-M (2022; photos by B. Dialo)
Selected Research Accomplishments
My working group is jointly affiliated at Universität Hamburg and its Cluster of Excellence Climate, Climatic Change, and Society (CLICCS), Helmholtz-Zentrum Hereon, and Max Planck Institute for Meteorology. With these joint forces, the group follows the ambition to bring the modeling of ocean biogeochemical processes to new frontiers. The focus of my research is on the role of oceanic cycles of carbon, nitrogen and oxygen under global change, and the group’s philosophy has been to address these cycles as an interactive component of the Earth system (Ilyina, 2022). Our main working tool is the model HAMOCC (Hamburg Ocean Carbon Cycle; e.g. Ilyina et al., 2013). We advance HAMOCC as a component of the Earth system models MPI-ESM and ICON (ICOsahedral Non-hydrostatic modeling framework) within model configurations spanning climate-relevant time scales, as well as resovling meso- and submeso-scale processes in ocean biogeochemistry. We aim to enhance model realism by bringing together increased process complexity and increased spatial resolution. This becomes possible by applying optimization approaches (such as component concurrency and porting code to GPUs) that improve the scalability and efficiency of model simulations, as well as with our passion for ocean biogeochemistry.
Our approach enables addressing novel exciting scientific questions related to variability and predictability of processes regulating the ocean carbon sink and carbon-climate feedbacks. These research topics are studied in concert with variability and changes in Earth's climate and ocean physics in past, present, and future in the framework of the Earth system models. The group particularly focuses on the role of biological processes, such as functioning of the biological pump, evolution of the oxygen minimum zones and related transformations in the marine nitrogen cycle. The group is networking on many aspects of the ocean biogeochemical cycles within numerous research projects.
Below is an incomplete (always evolving) list of our recent research accomplishments. More information can be found on the institutional webpages at Uni Hamburg: Modeling the carbon cycle in the Earth system, Hereon: Earth System Modeling, and MPI-Meteorology: Ocean Biogeochemistry.
Predictability of carbon sinks:
Predicting near-term variations in atmospheric CO2 growth, ocean and land carbon sinks under changing emissions remains a major challenge requiring ESMs with interactive carbon cycle. We built such a prediction system based on MPI-ESM. First, this enabled an improved reconstruction of the observed evolution of carbon sinks and atmospheric CO2 over the reanalysis period due to assimilation of observational products. Second, we established a predictive skill of up to 5 years for the air–sea and 2 years for the air–land CO2 fluxes and atmospheric CO2 growth rate, respectively (Li et al., 2023). Thereby we found that the main limit on predictability of atmospheric CO2 growth rate are land CO2 fluxes, while the ocean CO2 fluxes enable predictability (Spring and Ilyina, 2020). In a first benchmarking study of ESM-based prediction systems enhanced with land and ocean carbon cycle components(Ilyina et al., 2021) we showed that these predictive horizons appear robust across a number of prediction systems assessed. This furthermore implies that predictive skills of land and ocean C sinks are similar between models driven by CO2 emissions and concentrations, albeit the latter lack predictive atmospheric CO2 capacity. In an idealized perfect model framework, we demosntrated that initialization of ocean biogeochemical variables on top of the physics, seem to not substantially add predictive capacity (Spring et al., 2021).
Such ESMs with interactive carbon cycle using data assimilation establish themselves as new tools filling the gap between stand-alone models (with inconsistencies due to prescribed forcings) and ESMs (with unresolved timing of climate modes). Our predictions of CO2 variation for next year together with the assimilation reconstructions in the past decades for the first time were added as a new line of evidence to the annual assessment of the Global Carbon Budget (Friedlingstein et al. 2023).
Schematic illustration of an initialized by observations prediction system based on an Earth system model (from Hongmei Li et al., 2023).
Enhancing the realism of biogeochemical processes by added complexity:
Processes related to marine biology determine the biological carbon pump efficiency, as well as elemental cycling and food-web structures in the ocean. Global models have persistent and large uncertainty in the representation of these processes, as well as their response to climate change. The group works on incorporating such missing or insufficiently represented processes in HAMOCC that are crucial in regulating ocean biogeochemical cycles in the Earth system. Thereby we aim to fill gaps in our understanding of ocean biogeochemical feedbacks under global change. The most important model developments include:
Incorporating an explicit representation of cyanobacteria as an additional tracer in HAMOCC (based on the physiological properties of the wide spread cyanobacterium Trichodesmium; Paulsen et al. 2017), enabled us to improve the representation of N2 fixation rates in comparison to observations. N2 fixation evolves in response to the environmental conditions shaping cyanobacteria's ecological niche. This newly added feedback mechanism (via light absorption) has a regulative role for tropical ocean sea surface temperature and its variability (Paulsen at al., 2018). This makes HAMOCC better equipped for climate change simulations, as the warmer and more stratified future ocean might favor N2 fixing cyanobacteria, as opposed to other phytoplankton species, leading to a shift in community composition.
We incorporated a new scheme of particulate organic carbon (POC) transport from the surface into the deep ocean in HAMOCC (Maerz et al., 2020) that explicitly represents marine aggregates and their measurable properties (e.g. size, microstructure, porosity, excess density) and includes temperature-dependent remineralization. As a result, the representation of the POC exported out of the euphotic zone was substantially improved. It made it possible for the first time to investigate how opal and calcium carbonate fluxes co-determine higher transfer efficiencies of POC in high latitudes and lower ones in the subtropical gyres. Temperature was thereby identified as a driving factor for remineralization, with implications for projections of the future biological carbon pump.
Within the BMBF project PalMod, Liu et al. (2021) extended HAMOCC by developing a new comprehensive parameterization of 13C accounting for changes due to biological processes and the 13C air-sea gas exchange. The good agreement between model results and observations for present-day 13C distributions and their evolution over the last century provided the ground, for the first time, to quantify the ocean 13C Suess effect and interpret some earlier unexplained uncertainties in observational 13C products.
Extending HAMOCC to include riverine carbon, nutrients and alkalinity inputs into the ocean (Lacroix et al., 2020) and explicitly representing the fate of terrestrially-derived organic carbon as a tracer in HAMOCC led to novel findings. We showed that pronounced human-induced perturbation of the ocean primary production are driven by counteracting effects of increased terrestrial inputs of nutrients and increased upper ocean thermal stratification (Lacroix et al., 2021). Interestingely, due to a relatively short residence time of terrestrial organic carbon, the global shelf ocean acted as a weak carbon sink already in the preindustrial age. These findings provide a basis for addressing the role of the changing carbon cycle in the global coastal ocean.
Enhancing the realism of biogeochemical processes in high resolution simulations:
Due to a too coarse spatial resolution of ESMs global ocean biogeochemical models suffer from uncertainties in representing patterns of ocean circulation that are decisive for biogeochemistry. We found that increasing spatial resolutions globally (in 10km and 5km configurations; Hohenegger et al., 2023) lead to a smaller bias in representing regional patterns of nutrients and primary production. In our first 5 km global coupled ICON simulations, the footprint of eddies in shaping vertical and horizontal distributions of biogeochemical variables (i.e. phytoplankton concentration) is unraveled. For the first time, we can discern the effect of hurricanes and storms on ocean biogeochemical variables and air-sea CO2 fluxes. Higher resolutions enhance the realism of processes mediated by mesoscale variability, i.e. air-sea gas fluxes, vertical transports of carbon, heat and oxygen (assessed by comparing simulations with different ICON-O resolutions).
Phytoplankton (left) and air-sea CO2 flux with surface wind speed (right) in a 5 km coupled ocean-atmosphere-HAMOCC simulation with the ICON model.
See visualizations of these model simulation:
Air-sea CO2 flux and surface wind speed simulated with high-resolution HAMOCC@ICON: https://www.youtube.com/watch?v=63NZSPjxv6w
Phytoplankton dynamics simulated with high-resolution HAMOCC@ICON: https://www.youtube.com/watch?v=vZ2P6N57oa0
