Assimilation of ocean colour data to constrain the state and parameters of marine biogeochemical models.

Marine biogeochemical models are run as part of Earth system and climate models, to assess the contribution of marine biology to the carbon cycle, and the response of marine ecosystems to a changing climate. Physical-biogeochemical ocean models are also used to provide short-range forecasts and reanalyses of the ocean state. This provides the opportunity to use observations to constrain the models through data assimilation (DA), with satellite ocean colour providing daily global observations of the concentration of chlorophyll, an index of marine phytoplankton biomass, in the surface ocean. DA into biogeochemical models generally use chlorophyll concentration based on satellites to update the biogeochemical model states [1-2]. But, the growing complexity of biogeochemical models, e.g., for the plankton pool, demands the inclusion of plankton functional types (PFTs) into DA. Using satellite algorithms it is now possible to obtain estimates of the chlorophyll in PFTs, e.g. size classes of phytoplankton [3]. Moreover, currently, the parameters of the biogeochemical models are set to be globally and temporally constant, whereas the model dynamics may be sensitive to spatio-temporal variations of the biological parameters, which can be captured through DA [4]. So, it is imperative to transition our capability for assimilating satellite-derived PFTs into advanced models such as MEDUSA - the biogeochemical model that forms a part of the new UK Earth System Model, UKESM1. The student will:
(1) Develop a DA scheme for assimilating the outputs of satellite algorithms that split ocean colour signals into phytoplankton size classes into MEDUSA.
(2) Develop improved methods for using ocean colour data to update the model nutrient, carbon and zooplankton variables;
(3) Implement a state-parameter estimation scheme of DA to capture simultaneous changes in the state variables and the biological parameters of MEDUSA on regional and global scales;
(4) Incorporate in situ observations of chlorophyll, nutrients and oxygen and investigate their relationships with physical variables.

This project will develop a computationally efficient DA scheme suitable for implementation in operational forecasting or reanalysis, whilst making fullest use of the empirical information from the observations combined with knowledge of model processes. It promises an effective platform to interface the advanced satellite algorithms, and state-of-the-art ocean biogeochemical models, through DA, which may greatly improve reanalyses and forecasts of marine biogeochemistry, and contribute to our understanding of modelling marine ecosystems and the Earth’s climate system.
Training opportunities:
The student will benefit from in house training on remote sensing, ocean modelling from experts in Meteorology, National Centre for Earth Observation and Department of Geography and Environmental Science within UoR; and from various training opportunities through the SCENARIO DTP and Reading University’s Researcher Development Programme. The student will visit the Ocean Forecasting R&D team at the Met Office (CASE partner), for an overall duration of about 3 months during the project, where a number of experienced scientists will be available to provide support and advice on physical and biogeochemical ocean modelling and data assimilation systems for forecasting and reanalysis. The student will be given access to the supercomputing system (currently a Cray XC40) used by Met Office researchers and will be provided with training on how to use the systems, and on scientific software packages for analysis and visualisation of data, as well as archiving systems to access past ocean and weather data.
https://www.reading.ac.uk/ges/Aboutus/Staff/shovonlal-roy.aspx; http://www.met.reading.ac.uk/userpages/sgs00kh.php