This dissertation seeks to document observations, statistics, and physical processes relating to warm, salty mesoscale anticyclonic eddies in the Irminger Sea and their impact on the properties of the region. Anticyclones are studied with a variety of platforms including a mooring, a glider, Argo floats, satellite altimetry, and a regional model. They are found to be widespread throughout the central Irminger gyre, and are similar to anticyclones documented in the neighboring Labrador Sea and Lofoten Basin. Two anticyclone formation regions are found along the boundary current, producing two size classes of anticylones. Over the period between 2002 and 2009, anticyclone transport makes an important contribution to the heat and salt budgets of the gyre upper (<460 m) layer. Along with surface fluxes, they dominate the heat and salt balance on this time scale. The eddy contribution is relatively steady over the 7-year period and presumably over longer timescales as well. Changes in the temperature and salinity in the central gyre can be explained by anomalies in surface forcing. Over yearly time scales, the eddy contribution plays a minimal role in the heat and salt budgets, and the dominant terms include surface forcing and non-eddy horizontal advection. The assumption was made that anticyclones decay in the basin interior, motivated by the observed absence of eddies in the winter, and we explore the hypothesis that winter mixing can lead to eddy decay. To study this, a 1D mixing model is used to show that the barotropic component of velocity can be weakened during a period of sustained surface heat loss, with little change in the baroclinic component. This can push the eddy toward a counter-rotating state, which has been shown by theoretical studies to be unstable for similar anticyclonic vortices, and may thus be a mechanism for eddy decay. Observations of anticyclones are compared to a 1/20 degree resolution regional model (VIKING20). The model reproduces many features of the circulation and water masses, and similar salty anticyclones as observed. However, it cannot resolve the smallest eddies that we detected, and the long-term evolution of the eddies in the model appears to be dominated by numerical diffusivity and viscosity. Therefore, the model cannot be used for studying the processes of eddy decay