Honduras: Tree Islands as a Tool for Tropical Forest Restoration


This study examined how the creation of tree islands can serve as seed and seedling recruitment foci to help accelerate the reforestation of abandoned tropical pastures. At three experimental sites in northern Honduras, two species commonly used in the region to create living fences–Gliricidia sepium and Bursera simaruba–were planted in grids to create tree islands. Many of the islands had complete canopy cover in two years and were popular stopovers for roosting birds. The increased seed rain resulting from the birds’ presence, along with the protection afforded by the canopies, produced seedling establishment rates three times higher in the islands than on the open pastures.

Quick Facts

Project Location:
Unnamed Road, Honduras, 15.496644495257938, -86.08809640624997

Geographic Region:
Latin America

Country or Territory:

Tropical Forest

Tropical Forest - Seasonal Broadleaf

Area being restored:
6 acres

Organization Type:
University / Academic Institution


Project Stage:

Start Date:

End Date:

Primary Causes of Degradation

Agriculture & Livestock

Degradation Description

Estimates of tropical deforestation in Central America range from 10,000 to 90,000 ha/year, depending upon the source and country in question (Kaimowitz 1997; Utting 1997). Most of this loss results from lands being converted to pasture and seeded with exotic, highly competitive grass species (Uhl et al. 1988; Nepstad et al. 1990; Guariguata et al. 1995; Humphries 1998; Rhoades et al. 1998).

Project Goals

– Increase the introduction of naturally dispersed seeds into pastures
– Improve the microenvironment within tree islands
– Reduce competition from grasses
– Accelerate the process of succession within these areas


The project does not have a monitoring plan.

Description of Project Activities:
The buffer zone at Pico Bonito National Park is heavily deforested. Within the northern buffer zone, three sites were established for this study. These sites are characterized as follows: 1) an abandoned but well established pasture, which had been seeded with the exotic aggressive forage grasses Urochloa brizantha and Pennisetum purpureum; 2) a ridge that originally was a native pine plantation that had been deforested and was now dominated by the bracken fern Pteridium aquilinum; and 3) an abandoned pasture that had been poorly established 15 years ago and was now dominated by a variety of native grasses and forbs. The first project activity was a common garden experiment that evaluated the establishment and survival of 11 species locally thought to establish vegetatively. Of these 11 species, two were chosen-- Gliricidia sepium and Bursera simaruba --for further testing of their ability to establish vegetatively and develop cover at the three deforested study sites. These species were chosen because of their prevalence as living fences in tropical America and their adaptation to a broad range of climatic conditions. The artificial islands were first established between April and June 2000. At each site, aboveground vegetation was initially cleared using machetes in order to create four blocks, each 20-m wide and 100-m long. Four replicates of each of three island sizes (small [4 m2], medium [16 m2], and large [64 m2]) were created for each of the two species, B. simaruba and G. sepium (n = 12 islands per species per site). One island from each size category, and from each species, was established in each block (n = 4 blocks/ site). Five, 13, and 41 stakes were planted in each small, medium, and large island, respectively (n = 236 stakes per species per site). Stakes were planted at 2-m intervals in a grid format with a fifth stake inserted in each center, equidistant to each of the four corner stakes. All stakes were planted within 1 - 3 days of harvesting. Lastly, all blocks were individually fenced prior to the planting of stakes, to prevent accidental cattle incursion into the study area. Extensive monitoring of the tree-island study sites followed, and comparisons were made to open pasture areas.

Ecological Outcomes Achieved

Eliminate existing threats to the ecosystem:
Establishment for Gliricidia sepium was nearly 100% at all sites, whereas B. simaruba ranged from 30 - 50%. Islands comprised of G. sepium stakes developed rapidly, and after nearly two years, a majority of the islands had complete canopy cover, with some canopies reaching heights of 6 metres. Islands of Bursera, on the other hand, failed to develop a uniform canopy, resulting in a higher mortality of stakes and only partial development of canopies. However, the stakes did still provide perch sites for birds. Twenty-three species of birds were recorded in the study, and they showed an overwhelming preference for tree islands over open pasture. The presence of birds increased seed rain at the tree island sites, thereby accelerating vegetative succession by increasing seedling establishment. Indeed, the dispersal of zoochorous tree seeds was significantly greater (2-fold) in islands than in open pasture, and by the end of the study, seedling establishment in islands was three times that found in open pasture.

Factors limiting recovery of the ecosystem:
B. simaruba's weak establishment is probably more a consequence of the environmental conditions in the region rather than a reflection of the establishment capabilities of the species per se. In a study in Nayarit, Mexico, Villanueva-Avalos et al. (1996) found higher survival for B. simaruba stakes (approximately 85%) than several other species, including G. sepium. That region is considerably drier (rainfall approximately 800 mm/year), suggesting that B. simaruba establishes better in more arid regions, as indicated in a review by Sauer (1979). It is also possible that B. simaruba merely requires a longer dry period after staking in which to establish. The species is highly sensitive to moisture right after planting when the underground portion of the stake is susceptible to rot (Messenger et al. 1997; R. A. Zahawi 2001, personal observation). Although G. sepium tree islands exhibited favorable results, the study identified several factors influencing their development. First, the size of the island was important for G. sepium plots, as the dispersal of zoochorous tree species was greater in large islands compared to small islands. Also, greater initial stake height and diameter at breast height (dbh) each resulted in greater crown development for G. sepium. Stakes with thinner stems had a lower survival rate. Lastly, damage to the bark of G. sepium as a result of transportation to the site was observed to affect the growth and development of the tree. Thus, the stakes must be handled with care as they are moved from one locale to another.

Socio-Economic & Community Outcomes Achieved

Key Lessons Learned

Living fences are common through out the tropics, but particularly in southern Mexico and Central America. Many species used in fencing are also found as shade trees in coffee plantations and pastures, and they supply fuel wood, fodder for livestock, flood control, soil enrichment (N-fixation) and lumber. Their widespread use, and the ability of these species to establish themselves under the harsh conditions found in pastures, makes them attractive tools for restoration.

Restoration studies that have attempted vegetative propagation of forest species not typically used by farmers have yielded poor results in field trials (Aide & Eusse 1995; Ray & Brown 1995) unless chemical treatment is employed (Granzow-de la Cerda 1999). These projects also require cumbersome and time-consuming studies of the environmental conditions in an area prior to commencement of any restoration work.

The use of living fence species, on the other hand, alleviates these concerns. Living fence species are cheap and widely available, and ready-to-plant stakes can easily be purchased from local farmers (Budowski 1987; Budowski & Russo 1993). Furthermore, local farmers can be consulted to determine which species grow best in the area, the size of the stakes to be planted, and how deep they should be planted. This liaising eliminates the need for extensive pre-restoration studies and allows plantation to begin immediately.

The technique of using live fence species to create tree islands for restoration of abandoned pastures could have broad application throughout tropical America by offering a cheaper alternative to otherwise expensive, labor-intensive projects.

Sources and Amounts of Funding

This study was supported by a Doctoral Dissertation Improvement Grant from the National Science Foundation, a fellowship from the Organization for Tropical Studies, a Francis M. and Harlie M. Clark Summer Grant, and grants from the Research Board and the Graduate College of the University of Illinois.

Other Resources

Rakan (Zak) Zahawi
Organization for Tropical Studies
Apdo 676-2050
San Pedro, Costa Rica
Email: zahawi@ots.ac.cr

Zahawi, R.A. and Augspurger, C.K. (2006) Tropical forest restoration: tree islands as recruitment foci in degraded lands of Honduras. Ecological Applications 16(2):464-478.

Zahawi, R. A. (2005) Establishment and growth of living fence species: an overlooked tool for the restoration of degraded areas in the tropics. Restoration Ecology 13(1):92-102.

Zahawi, R. A. (2003) Island biogeography and restoration: the role of living fence islands as regeneration foci in the rehabilitation of degraded lands in Honduras. Ph.D. Dissertation. University of Illinois, Urbana, Illinois, USA.

Primary Contact

Organizational Contact