Recent empirical research has confirmed the importance of green infrastructure and outdoor recreation to urban people’s well-being during the COVID-19 pandemic. However, only a few studies provide cross-city analyses. We analyse outdoor recreation behaviour across four Nordic cities ranging from metropolitan areas to a middle-sized city. We collected map-based survey data from residents (n = 469–4992) in spring 2020 and spatially analyse green infrastructure near mapped outdoor recreation sites and respondents’ places of residence. Our statistical examination reveals how the interplay among access to green infrastructure across cities and at respondents’ residential location, together with respondents’ socio-demographic profiles and lockdown policies or pandemic restrictions, affects outdoor recreation behaviour. The results highlight that for pandemic resilience, the history of Nordic spatial planning is important. To support well-being in exceptional situations as well as in the long term, green infrastructure planning should prioritise nature wedges in and close to cities and support small-scale green infrastructure.

Background: Context-free outcome measures, such as overall leisure-time physical activity (LTPA), are habitually applied to study the neighborhood built environment correlates of physical activity. This cross sectional study identifies and empirically tests potential methodological limitations related to the use of context-free measures and discusses how these may help in the interpretation of inconsistent associations between participation in moderate-to-vigorous LTPA and objectively measured neighborhood-level built environment attributes.Methods: We employ a public participation geographic information system (PPGIS), an advanced participatory mapping method, to study the spatial distribution of moderate-to-vigorous LTPA among adult urban Finnish residents (n1322). Secondary sources of GIS land-use and sport facility data were used to disaggregate respondent-mapped LTPA by the behavioral context, such as indoor and outdoor sport facilities, green spaces, and other public open spaces. Associations between the use of the identified LTPA settings and a range of objectively measured neighborhood built environment attributes were studied with multilevel logistic regression models.Results: Disaggregated by behavioral context, we observed varied and partly opposite built environment correlates for LTPA. The use of indoor and outdoor sport facilities showed no significant associations with their neighborhood availability, but were significantly associated with personal-level attributes. By contrast, participation in LTPA in green and built public open space shared significant associations with access to and availability of neighborhood green space that persisted after controlling for personal-level covariates. Moreover, neighborhood distances up to 1600 m poorly captured participation in moderate-to-vigorous LTPA, as, on average, 40% of visits were located further from home. However, we found the immediate home environment to be an important LTPA setting for the least active participants.Conclusions: This study demonstrates that LTPA can be a highly heterogeneous measure regarding both the spatial distribution and the environmental correlates of behavioral contexts. The results show that context-free LTPA outcome measures yield inconsistent associations with built environment exposure variables, challenging the applicability of such measures in designing neighborhood-level built environment interventions.