The Atlantic Multidecadal Oscillation (AMO), a sea surface temperature mode of natural variability with dominant timescales of 30 -70 years and largest variations centered on the northern North Atlantic latitudes is one of the principal climate signals that have earned considerable attention in the recent decades, due to its multilateral impact on both local and remote weather and climate and its importance in predicting extreme events, such as drought development over North America. A 3-dimensional structure of the AMO is constructed based on observations and coupled, ocean-atmosphere 20th century climate simulations. The evolution of modeled, decadal-to-multidecadal variability and its hydroclimate impact is also investigated between two successive model versions participating in the CMIP3 and CMIP5 projects. It is found that both model versions underestimate low frequency variability in the 70-80 and 30-40 year ranges, while overestimating variability in higher frequencies (10-20 year range). In addition, no significant improvements are noted in the simulation of AMO's hydroclimate impact. A subsurface, vertically integrated heat content index (0-1000m) is proposed in an effort to capture the thermal state of the ocean and to understand the origin of AMO variability, especially its surface-subsurface link on decadal- to- multidecadal timescales in the North Atlantic basin. The AMO-HC index exhibits stronger oscillatory behavior and shorter timescales in comparison to the AMO-SST index, while leading the latter by about 5 years. A cooling of the North Atlantic subsurface is discernible in the recent years (mid-2000s -present), a feature that is almost absent at the ocean surface and could have tremendous implications in predicting future North Atlantic climate and in relation to the recent hiatus in the rise of global surface temperatures that was noted in the latest Intergovernmental Panel on Climate Change assessment report. Finally, AMO's decadal variability is shown linked to Gulf Stream's northward surges and the low-frequency NAO, as envisioned by Vinhelm Bjerknes in 1964. A cycle encompassing the low-frequency NAO, Gulf Stream's poleward excursions and the associated shifts in surface winds and SSTs over the subpolar North Atlantic is proposed as a possible mechanism for AMO's origin and a principal target for future research.