Oxygen minimum zones in the tropical Pacific across CMIP5 models: mean state differences and climate change trends

A Cabré,R Bernardello,D Bianchi,I Marinov

doi:10.5194/bg-12-5429-2015

Abstract

Abstract. We analyse simulations of the Pacific Ocean oxygen minimum zones (OMZs) from 11 Earth system model contributions to the Coupled Model Intercomparison Project Phase 5, focusing on the mean state and climate change projections. The simulations tend to overestimate the volume of the OMZs, especially in the tropics and Southern Hemisphere. Compared to observations, five models introduce incorrect meridional asymmetries in the distribution of oxygen including larger southern OMZ and weaker northern OMZ, due to interhemispheric biases in intermediate water mass ventilation. Seven models show too deep an extent of the tropical hypoxia compared to observations, stemming from a deficient equatorial ventilation in the upper ocean, combined with too large a biologically driven downward flux of particulate organic carbon at depth, caused by particle export from the euphotic layer that is too high and remineralization in the upper ocean that is too weak. At interannual timescales, the dynamics of oxygen in the eastern tropical Pacific OMZ is dominated by biological consumption and linked to natural variability in the Walker circulation. However, under the climate change scenario RCP8.5, all simulations yield small and discrepant changes in oxygen concentration at mid depths in the tropical Pacific by the end of the 21st century due to an almost perfect compensation between warming-related decrease in oxygen saturation and decrease in biological oxygen utilization. Climate change projections are at odds with recent observations that show decreasing oxygen levels at mid depths in the tropical Pacific. Out of the OMZs, all the CMIP5 models predict a decrease of oxygen over most of the surface and deep ocean at low latitudes and over all depths at high latitudes due to an overall slow-down of ventilation and increased temperature.

Highlights

Most marine organisms suffer and might die in hypoxic conditions, i.e. when the oxygen concentration falls below 60– 80 mmol m−3 (Gray et al, 2002; Stramma et al, 2008)
The comparison between CMIP5 model oxygen simulations and WOA09 observations highlights several consistent biases (Fig. 1 and 2a). (a) The majority of the models join the northern and southern oxygen minimum zones (OMZs) regions into a single large tropical OMZ (Fig. 1a), such that the modelled OMZ reaches anoxic conditions at shallower depths (Fig. 1c–d), and expands more westward and deeper than observed (Fig. 1c– d). (b) In observations, the northern OMZ area and volume are much larger than the southern counterpart, while in models, the northern and southern OMZ areas and volumes are much more symmetric with respect to the equator (Fig. 1a– b)
In some models the Northern Hemisphere OMZ is often too small in horizontal extent or does not extend as deep as in observations (Fig. 2). (c) A subset of models (GFDL-ESM2G, GFDL-ESM2M, MPI-ESM, and NorESM1-ME) produce anoxia to a depth of 2500 m or more in the eastern tropical Pacific, which is deeper than observations by more than 1500 m (Fig. 1c–d and 2)

Summary

Introduction

Most marine organisms suffer and might die in hypoxic conditions, i.e. when the oxygen concentration falls below 60– 80 mmol m−3 (Gray et al, 2002; Stramma et al, 2008). Note that this is a limited definition, since the specific survival and performance of organisms depend on the species, temperature, oxygen partial pressure, and CO2 levels (Seibel, 2011). The locations of the tropical Pacific OMZs are mostly determined by sluggish ventilation, coinciding with a maximum in water age rather than a maximum in biological productivity

Methods

Results

Conclusion