First, the following link may lead to a complimentary copy of article by the Washington Post. Second, I attach the first & second figures from the article (see the caption for the first figure at end of the post). Finally, I would like to point-out that the formation of Antarctic Bottom Water, AABW, also contributes to the speed of the AMOC and the rate of production of AABW has slowed in recent years; so the exceptional slow-down of the AMOC (which naturally oscillates) may not be exclusively due to ice melting from Greenland.
Stefan Rahmstorf, Jason E. Box, Georg Feulner, Michael E. Mann, Alexander Robinson, Scott Rutherford and Erik J. Schaffernicht, (2015), "Exceptional twentieth-century slowdown in Atlantic Ocean over turning circulation", Nature Climate Change, DOI: 10.1038/NCLIMATE2554
http://www.nature.com/articles/nclimate2554.epdf?referrer_access_token=djhLLqtPdDNATUwJf44DRdRgN0jAjWel9jnR3ZoTv0NAhBvJD3qQKAFJ5ZYnRB2DfVKqstvbeSrKxpKUhj2SxF7BcI_loegLGlYCV27ok_Njli4FpCNFd520NkNH-gNy_R7BHOTlk8WVlOM-EydqJ_fXB_3x-E3hIshOeW5WWHqcaPgYVH6Ha2paJACMrQS0vL1bzMOuRrJUW7F2fIb6zTOfarfleGahqDJs4nRADLaiLU5g6rQIKxir0Igbm9o6CWHumkVB6-NveR4QQcF04yFUDA2eESQkZFHTbg4BEjL4NRYUqzDNwfzQCuzM7uRDviUJ3rOglH4F30jdZAuf9g%3D%3D&tracking_referrer=www.washingtonpost.comCaption for first figure: "Figure 1 | Linear trends of surface temperature since AD 1901. Based on
the temperature data of NASA GISS (ref. 48). a, Global equal area map (Hammer projection) for 1901–2013; white indicates insufficient data. b, Same analysis for the North Atlantic sector for 1901–2000. In addition to the observed temperature trends b also shows the grid points (black
circles) of the subpolar-gyre region for which time series are shown in Figs 3 and 5, as well as the model-average 2◦C cooling contour (white) from a climate model intercomparison in which the models were subject to a strong AMOC reduction induced by adding a freshwater anomaly to the northern Atlantic. The geographic extent of the model-predicted temperature response to an AMOC reduction coincides well with the region of observed twentieth-century cooling. The models are forced more strongly and cooling extends further west as a result of shutting down Labrador Sea convection, which has only briefly happened in the real world so far. (Note that the second cooling patch in central Africa is in a region of poor data coverage and may be an artifact of data inhomogeneities.)"