Relationships between variations in peak Canadian forest fire season (JJA) severity and previous winter (DJF) global sea surface temperature (SST) variations are examined for the period 1953 to 1999. Coupled modes of variability between the seasonal severity rating (SSR) index and the previous winter global SSTs are analyzed using singular value decomposition (SVD) analysis. The robustness of the relationship is established by the Monte Carlo technique. The importance of the leading three SVD modes, accounting for approximately 90% of the squared covariance, to Canadian summer forest fire severity is identified. The first mode relates strongly to the global long-term trend, especially evident in the warming of the Southern Hemisphere oceans, and shows significant positive correlation in the forested regions of northwestern, western and central Canada, while southern B.C., the extreme northwest coastal regions of B.C. and Yukon, and the Great Lakes region are identified as having significant negative correlation. The second mode relates to the multidecadal variation of Atlantic SST (AMO) and shows highly significant negative correlation extending from the western NWT and Canadian Prairie Provinces across northern Ontario and Quebec. The third mode is related to Pacific Ocean processes and the interrelationship between El Ni¿o--Southern Oscillation (ENSO) and the Pacific Decadal Oscillation (PDO) and shows strong positive correlation in western Canada and negative correlation in the lower Great Lakes region of southern Ontario and southern Quebec. A 6-month lag relationship between Canadian forest fire variability and large-scale SSTs may provide the basis for developing long-range forecasting schemes for fire severity in Canada.