Invasive potential of Eucalyptus globulus: Seed dispersal, seedling recruitment and survival in habitats surrounding plantations

• Eucalyptus globulus spread from plantations into surrounding habitats was studied.
• Native forests had lower invasibility than shrublands and pine plantations.
• Strongly leptokurtic seed dispersal curve: 98.6% of seeds recorded within 15 m.
• A safety belt around plantations could reduce eucalypt spread.
• Safety belts covered by native forest vegetation would be more effective.

In many countries, plantation forestry depends mostly on a small number of fast-growing exotic species (mainly eucalypts and pines). These species are highly productive but may also become invasive in the communities surrounding plantations. The spread of alien species depends on characteristics of the species itself (species invasiveness) and of the host community (community invasibility), so that some habitats are more vulnerable than others to the invasion of a given species. Our aim in this study is to assess the ability of Eucalyptus globulus to spread in the absence of fire from plantations into the three most common types of surrounding habitats in the study area (NW Spain; native forests, pine plantations and shrublands) and to make recommendations to minimize the risk of spread. We independently assessed propagule pressure and community invasibility by measuring (1) the amount of seeds of E. globulus reaching different distances from the edge of plantations, and (2) seedling emergence and survival in the three types of habitats mentioned. We then estimated the density of seedling and sapling recruitment at different distances from plantation edges by means of a stochastic model parameterized with field data. Seed dispersal distances up to 80 m were recorded, with the frequency distribution of dispersal distances being strongly leptokurtic. Most E. globulus seeds were recorded in the first 5 m from plantation edges (84.3%), and 98.6% in the first 15 m. Native forests were the most resistant to E. globulus spread, showing lower seedling emergence (0.9% vs. 3.3% and 2.2% in shrublands and pine plantations) and survival (3.7% vs. 17.5% and 19.9% up to 6 months). Despite relatively low establishment success ratio, high seed production leads to considerable seedling densities in surrounding habitats, especially in pine plantations and shrublands. In order to control E. globulus spread, we recommend eliminating all new recruits in a safety belt around plantations (e.g. a 15-m wide belt could reduce recruitment in c. 95%) while they can be easily uprooted (1–2 years) and have not reached sexual maturity. Given the resistance of native forests to eucalypt establishment, fostering this type of vegetation in those belts would facilitate their maintenance and increase their effectiveness.