Large grain angles in timber can have a negative effect on wood quality by reducing dimensional stability, strength properties and performance, and causing twisting and warping in sawn timber and poles. In this study, we assessed the impact of tree spacing and site exposure on grain angle, based on the measurements from 360 discs cut from Sitka spruce (Picea sitchensis (Bong.) Carr.) trees from Northern Britain (Scotland and Northern England). The results show that planting on sites exposed to strong prevailing winds, planting at wider spacing or undertaking heavy thinning can significantly increase spiral grain angle. Using the mixed-effect model, we found that between-tree variation ranged from 3.6 to 33.3 per cent, variation due to disc position within the tree ranged from 5.7 to 47.8 per cent; and between-ring variations within discs ranged from 0 to 38.5 per cent. Residual variation ranged from 25.7 to 64.8 per cent. Grain angles are greater at the juvenile stage than at the mature stage of wood development. A non-linear model, developed to predict grain angle, explained only 16 per cent of total variation in grain angle, despite the inclusion of several stand covariates into the model. Although the mixed-effect model improved the root mean square error (RMSE) by 26 per cent, and the coefficient of determination (R2) rose to 53 per cent, its usage requires that a sample of grain angle measurements be made on the particular site for model calibration. The silvicultural response to reduced grain angle would be to plant at closer spacing and delay thinning or thin lightly, or to avoid planting Sitka spruce on very exposed sites. Given that grain angle is highly heritable, a further option is to reduce grain angle through selective breeding.