The linked reference discounts the observed positive trend in the SAM to conclude that natural variability overwhelms the anthropogenic radiative forcing response in the high-latitude Southern Hemisphere. To me discounting the observed positive trend in the SAM is a very bad idea as the first attached image from Fogt et. al. (2011) shows that during a La Nina event (as NOAA just called for 2016-17) combined with a positive SAM leads to atmospheric telecommunication of energy from the Tropical Pacific to West Antarctica as is supported by the atypically low current Antarctic sea ice extent in West Antarctica as indicated in the second attachment.
Jones, J.M., Gille, S.T., Goosse, H., Abram, N.J., Canziani, P, Charman, D., Clem, K., Crosta, X., de Lavergne, C., Eisenman, I., M. H. England, Fogt, R., Frankcombe, L., Marshall, G.J., Masson-Delmotte, V., Morrison, A.K., Orsi, A., Raphael, M., Renwick, J., Schneider, D., Simpkins, G., Steig, E., Stenni, B., Swingedouw, D., and Vance, T.R., 2016: Assessing recent trends in high-latitude Southern Hemisphere surface climate. Nature Climate Change, 6, 917-926http://web.science.unsw.edu.au/~matthew/nclimate3103.pdf
Extract: "Understanding the causes of recent climatic trends and variability in the high-latitude Southern Hemisphere is hampered by a short instrumental record. Here, we analyse recent atmosphere, surface ocean and sea-ice observations in this region and assess their trends in the context of palaeoclimate records and climate model simulations. Over the 36-year satellite era, significant linear trends in annual mean sea-ice extent, surface temperature and sea-level pressure are superimposed on large interannual to decadal variability. Most observed trends, however, are not unusual when compared with Antarctic palaeoclimate records of the past two centuries. With the exception of the positive trend in the Southern Annular Mode, climate model simulations that include anthropogenic forcing are not compatible with the observed trends. This suggests that natural variability overwhelms the forced response in the observations, but the models may not fully represent this natural variability or may overestimate the magnitude of the forced response."