The relative influence of habitat loss, fragmentation and matrix heterogeneity on the viability of populations is a critical area of conservation research that remains unresolved. Using simulation modelling, we provide an analysis of the influence both patch size and patch isolation have on abundance, effective population size (Ne) and FST. An individual-based, spatially explicit population model based on 15 years of field work on the red-cockaded woodpecker (Picoides borealis) was applied to different landscape configurations. The variation in landscape patterns was summarized using spatial statistics based on O-ring statistics. By regressing demographic and genetics attributes that emerged across the landscape treatments against proportion of total habitat and O-ring statistics, we show that O-ring statistics provide an explicit link between population processes, habitat area, and critical thresholds of fragmentation that affect those processes. Spatial distances among land cover classes that affect biological processes translated into critical scales at which the measures of landscape structure correlated best with genetic indices. Therefore our study infers pattern from process, which contrasts with past studies of landscape genetics. We found that population genetic structure was more strongly affected by fragmentation than population size, which suggests that examining only population size may limit recognition of fragmentation effects that erode genetic variation. If effective population size is used to set recovery goals for endangered species, then habitat fragmentation effects may be sufficiently strong to prevent evaluation of recovery based on the ratio of census:effective population size alone.