We investigate relevant scales and cost-efficient methods for measurements of habitat for saproxylic, aspen-associated beetles, in two boreal forest landscapes in south-eastern Norway. We sampled saproxylic beetles with window traps on fresh aspen dead wood and conducted field surveys of dead wood in the surroundings (1, 2 and 3 km radius). In addition, we used maps derived from satellite imagery to extract data on forest age and volume for the same surroundings. We found that species richness of saproxylic beetles was related to dead wood volume estimated by field survey within 1 or 2 km radius from the trapping points. However, the map-derived variable describing area of forest with high levels of deciduous wood volume was the overall best predictor of species richness. The scale of response to this variable differed; species richness of all saproxylic beetles was best predicted by estimates for the 2 km radius whereas richness of species more strongly associated with aspen was best predicted by estimates for the 3 km radius. This might indicate that aspen-associated beetles disperse over larger distances than many other beetles; possibly an adaptation to the scattered distribution and ephemeral nature of aspen dead wood in boreal forests. Our results show that substrate variables must be measured at the scale that is ecologically relevant for the study organism to elicit a response. We also show that remote sensing data -often easy to obtain for large scales—can be used to complement studies of scale-specific responses and thus improve conservation measures for saproxylic beetle species.