Context <p>Primary forests, with little to no direct human impact, can serve as baselines to assess land-use effects. However, most have been converted for other land-uses and the locations and characteristics of the remaining forests are poorly characterized.</p> Objectives <p>This study presents a map of Swedish primary forests, explores their persistence in a managed landscape, and assesses their suitability as baselines for studying land-use impacts.</p> Methods <p>We mapped 384 primary forests across Sweden and assessed their naturalness using historical records. To evaluate factors influencing their persistence, we analyzed their accessibility by measuring proximity to historical roads and timber floating rivers, and their boulder coverage. Lastly, we assessed the suitability of primary forests as baselines by comparing their topographic characteristics with those of managed secondary forests.</p> Results <p>Most primary forest land (93.6%) remains at higher elevations, associated with low productivity and accessibility, while 6.4% occurs in lowland areas. About one-third of the lowland productive primary forest land is topographically similar to the typical Swedish forest landscape. These forests are located farther from 1939 roads and more frequently in boulder-rich terrain than non-primary forest land.</p> Conclusions <p>Different types of primary forests persist: low-productivity mountain forests and lowland forests with historically limited accessibility. A considerable share of the latter have topographic conditions similar to those of managed secondary forests, indicating that their persistence was likely not driven by unfavorable condition for forest management but by accessibility constrains. These lowland primary forests may be useful baselines for further studying ecosystem impacts of land-use change.</p>