| Summary: | Observations and Earth System Model (ESM) projections indicate that a reduction in the oxygen inventory of the global ocean, termed ocean deoxygenation, is a likely consequence of on-going anthropogenic warming. The contribution of external forcing factors to observed changes in dissolved oxygen concentration ([O2]) relative to natural internal variability is examined using statistical methods which synthesise historical measurements and ocean biogeochemistry model output. Using a formal optimal fingerprinting method, an externally forced signal, derived from ESM response patterns, is detected within the observational record of [O2] between �1970 and �1992 at the 90% confidence level. Positive detection results in response to external forcing are robust for depth-averaged (100–3000 m) and depth-resolving zonal mean patterns globally and for the Pacific basin, however [O2] changes in the Atlantic basin are indistinguishable from internal variability as characterised by unforced ESM integrations. Current ESMs are also shown, using optimal detection techniques, to consistently underestimate the magnitude of observed [O2] change by a factor of �2 – 4. Accordingly, targeted hindcast experiments are conducted using the PlankTOM10-NEMO3.1 model, quantifying the impact of physical and biogeochemical processes on the spatiotemporal distribution of O2. The largest magnitude of uncertainty is shown to be entrained into [O2] response patterns due to model parameterisation of pCO2-sensitive C:N ratios in carbon fixation and imposed atmospheric forcing data. Historical trends and variability in Bottom Mixed Layer (BML) [O2] for the North Sea region are also investigated.
|
|---|
