Rainfall is one of the most important of the weather parameters on which fire-danger calculations are based. Therefore, a fire weather network which has been optimally designed for measurement of daily rainfall will be adequate for other input parameters as well. Spatial patterns of daily rainfall correlation (r) in southern New Brunswick, Canada, were examined. A non-linear least-squares algorithm was used to fit a geometric function to the spatial distribution of inter-station correlation coefficients in the study area. Using the function, isocorrelation patterns were constructed around individual stations for the arbitrarily selected value of r=0.9 (r2=0.8). All stations showed elliptical patterns, of varying sizes, with eccentricities between 0.3 and 0.6. Correlation field orientations were generally southwest-northeast, approximately parallel to the coastline of the Bay of Fundy, located to the south of the Province. There was also evidence of differential gradients in the correlation patterns, with tighter gradients toward the Bay. The procedure was applied to the local forestry network and areas of data paucity and redundancy were identified. © by the Society of American Foresters. Abstract reproduced by permission.