Most emphysema is believed to arise from small airways disease, but recent data suggest emphysema begets more emphysema and that its progression may be due to the mechanical stress experienced by normal lung regions adjacent to existing emphysema. To determine whether new emphysema arises from this penumbra of mechanically affected lung (MAL) or from small airways disease. We co-registered inspiratory chest computed tomography (CT) scans acquired at enrollment and 5 years later in 4,972 participants enrolled in a multicenter cohort. Using adaptive Gaussian smoothing, we quantified the 3D effect of all emphysema clusters on adjacent normal voxels by assuming that the mechanical effect of each emphysema cluster on surrounding voxels depends on cluster size and decays with increasing distance. The cumulative mechanical effect on each voxel was used to calculate MAL. Based on the probability distribution of normal voxels progressing to emphysema, we classified voxels into High (≥10.5), Intermediate (>0 to <10.5), and Zero MAL. We co-registered baseline inspiratory and expiratory CT images to quantify functional small airways disease. We quantified the proportion of new emphysema arising from each risk region. In adjusted analyses, higher MAL was associated with faster FEV1 decline (-2.2 ml/year, 95%CI -2.6 to -1.7; p<0.001) and emphysema progression (-0.14 g/L/year, 95%CI -0.16 to -0.12; p<0.001), and greater all-cause mortality (adjusted hazards ratio = 1.07, 95%CI 1.05-1.09; p<0.001). The relative mean contributions of high and intermediate MAL to new emphysema were 60.5% and 37.1%, respectively, in contrast to zero MAL (2.4%) and small airways disease (4.8%). Most new emphysema arises from areas of high MAL and in substantially higher proportion than areas of small airways disease.