Long-term observational studies conducted at large (regional) spatial scales contribute to better understanding of landscape effects on population and evolutionary dynamics, including the conditions that affect long-term viability of species, but large-scale studies are expensive and logistically challenging to keep running for a long time. Here, we describe the long-term metapopulation study of the Glanville fritillary butterfly (Melitaea cinxia) that has been conducted since 1991 in a large network of 4000 habitat patches (dry meadows) within a study area of 50 by 70 km in the Åland Islands in Finland. We explain how the landscape structure has been described, including definition, delimitation, and mapping of the habitat patches; methods of field survey, including the logistics, cost, and reliability of the survey; and data management using the EarthCape biodiversity platform. We describe the long-term metapopulation dynamics of the Glanville fritillary based on the survey. There has been no long-term change in the overall size of the metapopulation, but the level of spatial synchrony and hence the amplitude of fluctuations in year-to-year metapopulation dynamics have increased over the years, possibly due to increasing frequency of exceptional weather conditions. We discuss the added value of large-scale and long-term population studies, but also emphasize the need to integrate more targeted experimental studies in the context of long-term observational studies. For instance, in the case of the Glanville fritillary project, the long-term study has produced an opportunity to sample individuals for experiments from local populations with a known demographic history. These studies have demonstrated striking differences in dispersal rate and other life-history traits of individuals from newly established local populations (the offspring of colonizers) versus individuals from old, established local populations. The long-term observational study has stimulated the development of metapopulation models and provided an opportunity to test model predictions. This combination of empirical studies and modeling has facilitated the study of key phenomena in spatial dynamics, such as extinction threshold and extinction debt.