Restoring ecosystem function to millions of hectares of deforested land is a global conservation priority. Second-growth forests can provide benefits including carbon storage, biodiversity conservation and human livelihood, especially in poor areas where the negative externalities from deforestation have a greater effect. As a consequence, forest restoration at landscape scales has increasing prominence in international and national environmental policies. These large-scale goals are complicated by uncertainty in forest recovery rates across heterogeneous landscapes. For example, some sites may require active land management to avoid arrested succession, while others may not. Unpredictable successional trajectories mean that restoration is often an expensive investment with high failure rates. Predictive models could potentially enable ecologists to forecast successional trajectories over large areas and long time periods and increase interest of restoration stakeholders. However, most models for forest dynamics have failed to address the dynamics of early succession that are critical for determining whether sites will remain in a state of arrested succession or regrow into diverse, native forest. Thus, considerable uncertainty regarding the mechanisms driving biomass accumulation and species richness during forest succession remains.
The first few years of succession are critical for forest recovery and depend on both local and landscape-scale factors. However, the lack of data on landscape-scale succession, a due to the necessarily limited spatial extent of field studies, is one of the main sources of uncertainties when modeling early forest succession. While field studies have demonstrated the impact of local-scale factors including remnant forest structure and past land use on succession at single sites, whether these results are scalable to landscapes remains unclear. Accounting for heterogeneity in landscape-scale succession will require accounting for landscape-scale factors, including both biophysical variables (e.g. topography) and socioeconomic variables (e.g. human population density). Also, gathering information of the very first stages of succession will increase models accuracy and help to understand better the processes that determine the successional path.
These challenges are exacerbated in dry tropical forest, which remains understudied compared to humid tropical forest. Specially there is a lack of studies about the first year of dry tropical secondary succession. Tropical dry forest is one of the most threatened habitats on earth, and landscape-scale restoration could have major benefits for this habitat. One striking case is Central America, where dry tropical forest was once dominant ecosystem, but now comprises only 1.7% of its former area. In addition, there is a lack of dry tropical forest protected areas because of the association of dry secondary forest areas with provisioning ecosystem services. This association is leading to under appreciation of dry tropical forest, an ecosystem that is becoming scarce and contains a high amount of endemic species. This ecosystem is capable not only of providing provisioning services, but also provides regulation, and cultural ecosystem services. At the same time, there is tremendous potential for restoration of dry tropical ecosystems in Central America, as economic development has resulted in the abandonment of land traditionally used for cattle pasture. Therefore, it is necessary to persuade land owners to manage their land in a sustainable way that helps to restore and conserve dry tropical forests. Modeling secondary succession using long-term and landscape-scale datasets on dry secondary tropical forest looks like a valuable tool for this task as they can increase the effectiveness of land management and predict restoration success.