David Sursham, Plymouth Marine Laboratory
Improving Marine Ecosystem Understanding and Predictions using a Novel Data Assimilation Technique
CoauthorsStefano Ciavatta, Peter Jan van Leeuwen, Phil Browne
Abstract: The carbon cycle between the ocean and the atmosphere is an important consideration for climate change studies. This cycle is closely linked to the concentration of chlorophyll at the sea surface, as it is a proxy for phytoplankton biomass. Observations of sea surface chlorophyll concentration are available via satellite, which can be assimilated into a biogeochemical marine ecosystem model. The aim of this project is to investigate how to best perform this assimilation.
In the context of marine biogeochemistry, we are most interested in ensemble-based data assimilation techniques. The effectiveness of these techniques has been investigated by performing twin experiments using the Equivalent Weights Particle Filter (Van Leeuwen, 2010), which is ideal for the treatment of non-linear models, and the Localised Ensemble Transform Kalman Filter (Hunt, Kostelich and Szunyogh, 2007), which performs well in high dimensional systems. For this experiment, these methods have been applied to a non-linear, high-dimensional marine ecosystem model (ERSEM) coupled to a physical model (GOTM).
The outcome of this experiment is to understand the how these data assimilation techniques perform in practice when used with our model. Beyond making a step towards improving the assimilation of surface chlorophyll concentration, these findings may help to inform other modellers about how to perform data assimilation for similar non-linear models.