System identification techniques for detection of teleconnections within climate models

Abstract. Quantifying teleconnections and discovering new ones is a complex, difficult process. Using transfer functions, we introduce a new method of identifying teleconnections in climate models on arbitrary timescales. We validate this method by perturbing temperature in the Nino3.4 region in a climate model. Temperature and precipitation responses in the model match known El Nino-Southern Oscillation (ENSO)-like teleconnection features, consistent with modes of tropical variability. Perturbing the Nino3.4 region results in temperature responses consistent with the Pacific Meridional Mode, the Pacific Decadal Oscillation, and the Indian Ocean Dipole, all of which have strong ties to ENSO. Some precipitation features are also consistent with these modes of variability, although because precipitation is noisier than temperature, obtaining robust responses is more difficult. While much work remains to develop this method further, transfer functions show promise in quantifying teleconnections or, perhaps, identifying new ones.