Spatial and temporal variability of the subduction rates in the Southern Ocean using CMIP models

Lucas Almeida, Federal University of Rio Grande, Institute of Oceanography, Rio Grande, Brazil, Mauricio M Mata, Federal University of Rio Grande, Instituto de Oceanografia, Rio Grande, Brazil and Matthew R Mazloff, University of California San Diego, Scripps Institution of Oceanography, La Jolla, United States
Abstract:
The subduction rates in the circumpolar region are responsible for the formation of the Subantarctic Mode Water and Intermediate Antarctic Water in the Southern Ocean, which consist of water masses part of the meridional circulation, highly important in the global climate system. The objective of this work is to evaluate the representation of this process in coupled climate models (CMIP models, generation 5 and 6). The subduction rates were calculated by the sum of the Ekman pumping (Wek) and the geostrophic terms throughout the entire Southern Ocean. Most models unveiled a high performance in the formation of subduction hotspots in the circumpolar region. However, these hotspots are placed in different areas of the region with high levels of variability. The natural variability, with about 5 years of period, was computed using the PIcontrol experiment (Taylor et al., 2012; Eyring et al., 2016). These values do not have a relevant difference when compared to the ones in the historical experiment, which means that no large changes in these values are observed post the Industrial Revolution (i.e. 1850 AD). Responding to the westerlies variability in the circumpolar region, the Wek variations has a major influence in the variability of the subduction process. To evaluate the models performance, we computed the Wek average for the last decade and compared with ERA-Interim Reanalysis. The results showed that most of the models have a good representation of the negative Wek values in the circumpolar region (i.e. subduction), but the upwelling in the southernmost areas was poorly represented. The relatively coarse atmospheric models resolution and the difficulty with the sea-ice volume are two possible reasons of these results. The next step of this study is to calculate the subduction rates in the future scenarios and quantify the impact of the climate change in this process.