Available PhD studentships and summer internships
DPhil project: Ocean Biogeochemical Optimisation in ESMs (OBOE) – NERC-Met Office Case Industrial Studentship (apply)
Supervisors: Prof. Samar Khatiwala (Earth Sciences) and Prof. Coralia Cartis (Mathematical Institute), University of Oxford; Prof. Colin Jones, NERC/Met Office; Drs. Andrew Yool and Adrian Martin, National Oceanography Centre
Project description: As one of the principal reservoirs of CO2, the ocean plays a crucial role in the carbon cycle and in regulating Earth’s climate. Understanding and modelling the interconnections between the ocean carbon cycle and climate is therefore critical for robust estimates of future climate change. A principal challenge in this regard is the absence of well-established sets of equations governing the behavior of marine ecosystems, which play a key role in ocean carbon dynamics. Consequently, fundamental processes, such as the formation and sinking of organic matter from the surface into the ocean interior are crudely parameterised. Improving the representation of these processes in global ocean biogeochemical models, embedded within Earth System Models (ESMs) used to project future climate change, is thus an important goal of current research and of this project in particular. Specifically, we seek to evaluate and improve the performance of MEDUSA (Model of Ecosystem Dynamics, nutrient Utilisation, Sequestration and Acidification), the ocean biogeochemical model in the next generation Met Office/NERC Earth system model (UKESM), currently under development. MEDUSA models the interaction between macro- and micro-nutrients, phytoplankton and carbon, representing these processes through a range of parameterisations that include a number of key uncertain parameters. We seek to improve the underlying formulation of these parameters to better represent available observational constraints.
To achieve this a number of challenges need to be addressed. First, because of the complex interaction between biogeochemistry and circulation, model sensitivities vary both in space and time, as well as with respect to the model field (e.g., nutrients v primary production). Second, evaluating the performance of global models is prohibitively expensive as every parameter change requires integrating the model for several thousand simulated years to equilibrium before the model can be compared with observations. As a result there have been very few attempts at systematically optimising the performance of models such as MEDUSA. To overcome this, the student will exploit a fast “offline” tracer simulation scheme and recently-developed mathematical optimisation techniques to optimise MEDUSA, a first for a global biogeochemical model of this complexity, especially one used in a state-of-the-art ESM.
Key outcomes of this project include (1) an estimate of MEDUSA’s sensitivity to various parameters and thus the relative importance of key processes that affect the strength of the biological carbon pump; (2) an optimal set of parameters that minimizes the model-observation cost function built on several fields; and (3) a quantitative assessment of the impact of parameter optimisation on key aspects of UKESM1-projected Earth system change, such as global climate sensitivity, marine carbon uptake and the resulting biogeochemical state of the deep ocean.
This project brings together ocean biogeochemists, a mathematician and an Earth system modeller and the student will benefit from working actively with scientists from several disciplines, including the UKESM model development core group. S/he will receive training in not only marine biogeochemical and Earth system modelling, but also in high performance computing, numerical analysis and mathematical optimisation techniques with broad applicability in science and engineering. The student will be affiliated with Oxford’s NERC-funded Environmental Science Doctoral Training Partnership in Environmental Research and will thus benefit from courses offered through the DTP as well as the Mathematical Institute.
UK/EU students with a good first degree in the natural sciences, maths or engineering and strong computing skills are encouraged to apply.
Summer Undergraduate Research Experience Placement: Fast simulation of marine biogenic particles on programmable GPUs
Supervisor: Prof. Samar Khatiwala, Department of Earth Sciences, University of Oxford
Eligibility and stipend: This is a summer research experience placement funded by the Natural Environment Research Council (NERC). It is open to undergraduates enrolled at any UK university in a quantitative course such as maths, physics, statistics, computer science or engineering. There is a bursary of £2,000.
Project description: Marine particles are increasingly recognized to play a fundamental in global biogeochemical cycles. For instance,the sinking of organic matter produced by photosynthesizing phytoplankton at the surface of the ocean transport large amounts of carbon from the atmosphere into the deep ocean, a process known as the “biological carbon pump”. Such particles also scavenge and transport trace elements, thus playing a key role in their geochemical cycling. However, most existing ocean models are Eulerian in nature (i.e., are designed to represent tracer concentrations) and are thus unsuited to simulate the fundamentally discrete and Lagrangian nature of these particles. The objective of this project is to develop and apply novel mathematical and computational approaches to efficiently simulate marine particles with a conventional Eulerian ocean circulation model. The key insight is to interpret the (deterministic) transport of tracers simulated by these models as probabilities that govern where a particle is likely to move. The student will both explore the mathematical underpinnings of this idea as well as develop computational algorithms to make it feasible to apply it to a large number (O(1012)) of particles in a realistic, global ocean circulation model. The algorithms will be designed to exploit many-core architectures such as programmable, general purpose graphics processing units (GPGPUs) that are increasingly used in scientific computing. GPGPUs are not only capable of extreme parallelism but are very efficient at tasks such as random number generation (a critical requirement for the proposed approach). Using the GPU cluster at Oxford’s Advanced Research Computing Centre, the student will test and apply the new algorithms to perform the first global simulation of marine biogenic particles.
Prerequisites: Undergraduate in maths, physics, computer science or similar field with excellent programming skills and willingness to get their hands dirty with code.