Tropical dry forests are the most threatened tropical terrestrial ecosystem. However, few studies have been conducted on the natural regeneration necessary to restore these forests. We reviewed the ecology of regeneration of tropical dry forests as a tool to restore disturbed lands. Dry forests are characterized by a relatively high number of tree species with small, dry, wind-dispersed seeds. Over small scales, wind-dispersed seeds are better able to colonize degraded areas than vertebrate-dispersed plants. Small seeds and those with low water content are less susceptible to desiccation, which is a major barrier for establishment in open areas. Seeds are available in the soil in the early rainy season to maximize the time to grow. However, highly variable precipitation and frequent dry spells are important sources of mortality in seeds and seedlings. Collecting seeds at the end of the dry season and planting them when soil has sufficient moisture may increase seedling establishment and reduce the time they are exposed to seed predators. Germination and early establishment in the field are favored in shaded sites, which have milder environment and moister soil than open sites during low rainfall periods. Growth of established seedlings, however, is favored in open areas. Therefore, clipping plants around established seedlings may be a good management option to improve growth and survival. Although dry forests have species either resistant to fire or that benefit from it, frequent fires simplify community species composition. Resprouting ability is a noticeable mechanism of regeneration in dry forests and must be considered for restoration. The approach to dry-forest restoration should be tailored to this ecosystem instead of merely following approaches developed for moister forests.

With the increasing need to restore former agricultural lands worldwide and in the tropics, in particular, it is critical to explore different models for how to restore these lands in a cost-effective manner which best simulates natural forest recovery and provides for human livelihoods. We propose that agro-successional restoration, which we define as incorporating a range of agroecology and agroforestry techniques as a transition phase early in forest restoration, could be used more widely to overcome socioeconomic and ecological obstacles to restoring these lands. Over centuries, farmers and scientists have developed various agroforestry techniques that aim to cultivate crops and trees, in a range of crop types, time periods of cultivation (a few years to several decades), and complexity of species planted. The management practices used in these systems, such as weeding and increasing soil fertility, parallel those used in many forest restoration efforts. The synergism between these approaches is evidenced by many existing agro-successional examples currently used by smallholders in the tropics. Benefits of the agro-successional model include extending the management period of restoration, offsetting some management costs, providing food security for small landholders, and involving small landholders in the restoration process.

No abstract

The coexistence of grasses, herbs, shrubs and trees characterizes savannas; therefore, to restore such ecosystems one should consider reintroducing all these growth forms. Currently, little is known about field establishment of most ''Cerrado'' (Brazilian savanna) species that could be used for restoration purposes. Most knowledge on restoration is focused on planting seedlings of tree species from forest physiognomies. Alternatively, direct seeding can be an appropriate method to reintroduce plants of different life forms to restore savannas. We evaluated the initial establishment success under field conditions of 75 ''Cerrado'' native species (50 trees, 13 shrubs, and 12 grasses) in direct seeding experiments in four sites in Central Brazil for 2.5 years. For that, we tagged and measured tree and larger shrub species and estimated ground cover by small shrub and grass species. Sixty-two species became established (42 trees, 11 shrubs and 9 grasses) under field conditions. Thirty-eight of the 48 tagged species had relatively high emergence rates ([10%) and 41 had high seedling survival ([60%) in the first year. Among grasses and small shrub species, Andropogon fastigiatus Sw., Aristida riparia Trin., Schizachyrium sanguineum (Retz.) Alston, Lepidaploa aurea (Mart. ex DC.) H.Rob., Stylosanthes capitata Vogel, S. macrocephala M.B.Ferreira & Sousa Costa, Achyrocline satureioides (Lam.) DC. and Trachypogon spicatus (L.f.) Kuntze had the greatest initial establishment success (up to 30% soil cover). The data on harvesting period, processing mode and field establishment for these 75 species can be readily used in restoration efforts in the ''Cerrado''.

Large-scale restoration programmes in the tropics require large volumes of high quality, genetically diverse and locally adapted seeds from a large number of species. However, scarcity of native seeds is a critical restriction to achieve restoration targets. In this paper, we analyse three successful community-based networks that supply native seeds and seedlings for Brazilian Amazon and Cerrado restoration projects. In addition, we propose directions to promote local participation, legal, technical and commercialisation issues for up-scaling the market of native seeds for restoration with high quality and social justice. We argue that effective community-based restoration arrangements should follow some principles: (i) seed production must be based on real market demand; (ii) non-governmental and governmental organisations have a key role in supporting local organisation, legal requirements and selling processes; (iii) local ecological knowledge and labour should be valued, enabling local communities to promote large-scale seed production; (iv) applied research can help develop appropriate techniques and solve technical issues. The case studies from Brazil and principles presented here can be useful for the up-scaling restoration ecology efforts in many other parts of the world and especially in tropical countries where improving rural community income is a strategy for biodiversity conservation and restoration.

Direct seeding is a promising technique for ecological restoration, but it has been poorly studied in neotropical savannas. Different types of plant cover (no cover, crops, or green manure) and fertilization (unfertilized, synthetic fertilizer, or poultry litter) were used to verify if survival and growth of different tree species after direct seeding could be enhanced by the use of any combination of these techniques. Seedling emergence, establishment, and growth were followed for 2 years for six savanna tree species sown in an agricultural field in Central Brazil. Germination was high (52%, on average) for Anacardium occidentale, Aspidosperma macrocarpon, Hymenaea stigonocarpa, Dipteryx alata, Eugenia dysenterica, and Magonia pubescens. Six additional species were planted, but less than 5% of these seeds germinated. Crops (60% shade) did not affect seedling survival and biomass compared with the control treatment, supporting the use of this strategy during the initial phase of restoration to involve farmers in the process. In contrast, the excessive shading (95%) from the green manure treatment decreased the survival of two species and the growth in biomass and diameter of five species, especially when combined with fertilization. Seedling growth was very slow throughout the experiment, requiring extended weed management. This study supports the use of direct seeding of the studied species for savanna restoration, but methods could be improved to include a larger number of species.

No abstract

No abstract