Diagnosing North Pacific Blocking Using a Linear Inverse Model

The impacts of midlatitude blocking on lives and property through its association with flooding, cold snaps and drought are well-known, but understanding what dynamically leads to blocking remains elusive. North Pacific blocking is particularly difficult for numerical weather prediction models to anticipate, and it is a circulation feature routinely associated with reduced forecast skill. In this study we employ a linear inverse model (LIM) to determine the optimal initial structures that most rapidly amplify into a central Pacific block two weeks later. The LIM infers the dynamics of a system from the lagged covariance statistics between model variables, which in this study includes 850 and 200 hPa midlatitude stream function and tropical outgoing longwave radiation (OLR). While the LIM assumes that variance is stochastically forced, deterministic, nonmodal growth can occur over a finite period of time. Prior studies have demonstrated the LIM’s utility for forecasting at subseasonal-to-seasonal timescales, but investigating blocking with a LIM has not previously been undertaken. The optimal initial structures that produce a block through nonmodal interactions are presented. The relative contributions from the midlatitude and tropical components are isolated, and it is shown that the central Pacific blocking pattern is a blend of internal midlatitude interactions and an influence on the midlatitude structure associated with tropical OLR. The former produces a pattern resembling the negative phase of the Pacific-North American pattern, while the tropical component produces a La Niña-type teleconnection. We conclude with a comparison of the skill of the LIM at the subseasonal forecast range during blocked periods and un-blocked periods and discussion of the limits of forecasting blocking onset.