Representation of Temperate Shelf Seas in an Intermediate-Complexity Global Climate Model

Abstract

Whilst accounting for only 7% of the global ocean surface area, shelf seas contain unique regimes that enable them to contribute disproportionally to the global carbon cycle. Their shallow nature allows competition between buoyancy forcing and tidal mixing, producing seasonal stratification. In spring, this separates a warm, nutrient-rich surface mixed layer from a colder, nutrient-rich subsurface layer about a thermocline. The spring phytoplankton bloom efficiently uses surface nutrients, such that the duration of the bloom
depends on diapycnal mixing supplying subsurface nutrients across the thermocline. The thermocline also separates phytoplankton production at the surface from respiration in the subsurface layers, enabling the accumulation and eventual export of CO2. This continental shelf pump contributes to the global carbon cycle, such that shelf seas account for up to 50% of the total open ocean CO2 storage.
This thesis examines the representation of temperate shelf seas in the intermediate-complexity UVic Earth System Climate Model, using the North Sea as a model example.
By comparing model predictions to local observations, we identify key shortcomings in the model’s skill for replicating shelf seas; specifically river basin configuration, predictions of freshwater discharge, and mixing parameterisation. These model limitations produce a flushing time 219% of the observed 3.11 years, and a failure to reproduce any periods of full mixing during the seasonal cycle of stratification. These discrepancies have the potential to impact the local primary production rates, and by extension, bias the carbon cycle.
Through a series of targeted model modifications, we increase agreement between model predictions and observations, such that the modelled flushing time was 97% of the observational estimate, periods of winter mixing occurred during the seasonal stratification cycle, and modelled primary production rates increased by 16% compared to the control simulation. We conclude throughout that improving the representation of freshwater has a more significant impact compared to updating local mixing.
Our results highlight a challenge for coarse resolution models, to accurately represent shelf seas such that their pivotal role in the global carbon cycle is replicated, without compromising computational efficiency.

Date of Award	20 Jun 2025
Original language	English
Awarding Institution	Bangor University
Sponsors	Natural Environment Research Council (NERC) & Envision DTP
Supervisor	Tom Rippeth (Supervisor), Mattias Green (Supervisor), Andreas Schmittner (Supervisor) & Sophie-Berenice Wilmes (Supervisor)