Long-term spatial dynamics of Succisa pratensis in a changing rural landscape: linking dynamical modelling with historical maps

1. We attempt to explain the current distribution of a long-lived perennial plant, Succisa pratensis, in a rural landscape in southern Sweden by linking its population biology with documented changes in the landscape, using a dynamical, spatially explicit model incorporating population dynamics and spatial spreading of the plant. Changes in the landscape were inferred from historical maps (1850 and 1900) and aerial photographs (1945 and 2001). 2. We tested whether predictions for the current species distribution are affected by assumptions about its early 19th century distribution, to determine whether recent history and current processes are dominant, and how past landscape changes determine current distributions. 3. Initial conditions influence predictions of current distribution, suggesting that the current distribution still partly reflects the distribution of the species in the early 19th century. A period of 150 years is too short for Succisa to have spread extensively if dispersal parameters are given realistic values. 4. Simulations in which present-day land-use patterns were imposed at earlier dates showed that changes in landscape structure over the past 175 years also had a strong effect on the present-day habitat occupancy and population sizes of Succisa. 5. The dominant process for Succisa now is extinction from marginal habitats. It is therefore likely that the (relatively) high present-day occupation patterns are still due to much larger areas having been available in the past rather than to successful dispersal. Although the species has responded to landscape changes, there is little evidence of population sizes reaching equilibrium. 6. Our approach shows that the wealth of landscape information available from historical maps can be linked with data on population biology by means of dynamical models that can make predictions about species dynamics.