Earth science processes and observational data often span 6 orders of magnitude in spatial scale, in addition to varying through time. Difficulties in correlating data taken at different scales and representing different properties commonly limits the integrated interpretation of multiple datasets. Correlation, particularly of noisy data, can be greatly enhanced using visual pattern recognition rather than relying on computational methods, which become difficult to interpret for complex features. Interactive, immersive visualization can substantially aid correlation of data and evaluation of the uncertainties in correlations.