The relationships between global pressure patterns have been the source of much research, primarily due to their impacts on global climate. However, such atmospheric teleconnections identified using highly simplistic compositing methodologies that assume linearity (e.g. the mean, principal component analysis – PCA, etc.). It is also known that atmospheric patterns follow highly nonlinear relationships, but have been described using such linear techniques. To resolve this discrepancy, the authors provide a nonlinear approach for the formulation of atmospheric teleconnectivity . Geopotential height data at 700 mb from 1948 – 2010 from the NCEP/NCAR reanalysis dataset were used in the reformulation of the teleconnections. The NCEP/NCAR reanalysis data reside on a 2.5° latitude-longitude global grid with 17 vertical levels. However, these data were interpolated to an equally spaced Fibonacci grid to ensure that no artificial influences on the inner product matrix come about as a result of poleward converging longitude lines. Monthly and seasonal patterns of heights were obtained using a nonlinear technique known as kernel principal component analysis (KPCA). KPCA utilizes a kernel function to map nonlinear data into a higher dimensional Hilbert space where the data have a linear relationship, after which a traditional PCA is implemented. The resulting seasonal and monthly KPCA height patterns will be compared to updated PCA height patterns to identify similarities and differences in the approaches. Future research will involve verifying these patterns against output from climate model simulations to identify which compositing technique better represents the climate model simulations.