Understanding observed skewed climate statistics as a response to Correlated Additive and Multiplicative noise (CAM noise) forcing

The governing equations for atmospheric and oceanic evolution are obviously nonlinear. However, in many contexts the dynamics of climate anomalies on weekly to millennial scales are found to be nearly indistinguishable from that of a stochastically forced damped linear system, and even very simple low-order prediction models based on this approximation are competitive with comprehensive numerical weather and climate models. In this paradigm, natural variability is viewed simply as a linear superposition of randomly excited linear modes (RELM), and forced variability is the response to time-varying forcing given by the associated linear response operator. The stochastic noise forcing of the RELM represents an approximate accounting of incoherent (i.e., unpredictable) multi-scale nonlinear interactions in the system. There are several features of observed climate variability, most notably the often substantially asymmetric behavior of positive and negative anomalies, that are apparently inconsistent with this simple paradigm, with implications for the role of coherent nonlinear scale interactions. We have recently shown that this apparent inconsistency can be resolved by allowing the stochastic noise forcing in the paradigm to have correlated additive and multiplicative noise (CAM noise) components. In this talk we will also show how CAM noise forcing, whose amplitude is different for positive and negative state anomalies, can explain the very different temporal persistence characteristics of positive and negative atmospheric circulation anomalies over the "blocking" regions of the north Pacific and north Atlantic oceans that are usually attributed to coherent nonlinear dynamics.