| description abstract | This study examines the primary atmospheric controls over winter precipitation variability in the Great Lakes basin and the potential for seasonal prediction. We employ partial least squares (PLS) regression to identify the primary modes of joint variability between winter precipitation over each of the Great Lakes and concurrent anomalies in midlevel atmospheric flow. We find that the first identified pattern (PLS1) is related to El Niño?Southern Oscillation (ENSO), while the other patterns represent unique anomalies in atmospheric flow that govern precipitation gradients over the basin, with limited seasonal predictability. Nonlinearities are found in the relationship between a sea surface temperature (SST)-based index for ENSO and PLS1 with respect to the phase, strength, and type of ENSO event. An examination of the ENSO-related propagating wave train that drives variability of PLS1 precipitation reveals that seasonally lagged tropical Pacific convection, as measured by remotely sensed outgoing longwave radiation (OLR), is more strongly and linearly related to Great Lakes winter precipitation than SST-based ENSO indices. Cross-validated linear regressions based on October OLR signals explain 20%?32% of the out-of-sample precipitation variability in the Great Lakes basin. We conclude with a deeper assessment of the underlying relationship between patterns of OLR anomalies in the western equatorial Pacific and Great Lakes winter precipitation. Results show that precipitation response to El Niño is similar regardless of OLR intensity in the tropical Pacific, but for La Niña events, the precipitation response is stronger under weak tropical OLR anomalies. The potential for further improvements in ENSO-based seasonal forecasts are discussed. | |
