Patterns of grassland invasions by trees: insights from demographic and genetic spatial analyses

Aims Woody invasions into grasslands have increased globally due to changing land use,climate and introduced woody species,but spatial processes generating and sustaining these invasions are not well understood.To gain insight into the patterns of spread of tree populations within grasslands,and to propose a full spatial analytical toolbox for studying native and non-native woody spe-cies spread when long-term data are not available,we tested if 50 years of grassland invasion in Western Carpathians by Norway spruce(Picea abies Karst.)proceeded by one of the two tradi-tionally competing hypotheses of species spread:(i)by frontier expansion,or(ii)by advanced groups established ahead of the population frontier.We also tested whether the pattern of invasion changed over time.Methods We analyzed the spatial demographic and genetic patterns of a Norway spruce population invading a Western Carpathian grass-land using ripley’s L(t)and genetic kinship coefficients(Fij).We mapped and genotyped spruce trees across the invasion front(from the invasion leading edge to fully colonized grassland near the source population)using three demographic classes(adults,juve-niles and seedlings)to approximate the temporal aspects of the invasion.We studied how the spatial patterns of invasion by indi-vidual demographic classes and their genetic kinship varied among adjacent plots established at different distances from the source population(ranging from 0 to 160 m,in 40-m distance increments).Important Findings Juveniles were positively genetically related to adults on fine scales(<4 m),suggesting that adults within the grassland acted as a seed source and accelerated early invasion.However,adults did not act as nucleation centers for the formation of advanced juvenile groups.Instead,geneti-cally unrelated juveniles formed groups independently of adults.These groups were small and separate at the leading edge but they increased in size and graded into a continuous zone near the source population.Thus,juvenile recruitment occurred as a frontier expansion near the source population and as advanced groups controlled by environmental variation at the leading edge.unlike juveniles,seedlings were clustered on all scales across the invasion front and formed groups around adult crowns at the invasion leading edge.The bulk of seedling establishment occurred at intermediate distances from the source population,indepen-dently from the adults,suggesting that the invasion front continued to expand as a frontier,gradually coalescing with the advanced groups at the leading edge.Thus,the grassland invasion was driven by a gradual frontier expansion of the original population during the first 50 years,with advanced groups enhancing but not driving the invasion process.Frontier expansion appeared more important as a mechanism of woody species spread early in the invasion process in this study,while advanced groups may play a larger role over longer temporal scales.