A large fraction of ocean variability on interannual and longer timescales is energized by random atmospheric weather, also referred to as climate "noise". Although the noise is random in time, spatially the atmospheric noise exhibits recurrent patterns, some of which are more efficient in triggering positive feedbacks between the ocean-atmosphere system or more generally amplifying the response of the ocean system. Noise patterns such as these, can trigger resonance in the climate system. By combining the massive data output of long-term IPCC-class climate model simulations with stochastic diagnostic models and machine learning algorithms, this project aims at (1) isolating the patterns of noise that are most effective in energizing the ocean climate variance, (2) quantifying how these patterns may change under greenhouse radiative forcing, and (3) identifying the mechanisms that allows the transfer and amplification of the noise into ocean variance. This project will interact with faculty from the new Georgia Tech Institute for Data Science and Engineering (http://bigdata.gatech.edu/ideas).