Previous PageTable Of ContentsNext Page

Trees Outside Forests: Costa Rica

David Morales Hidalgo and Christopher Kleinn
Tropical Agricultural Research and Training Centre (CATIE)

Photo 48. Growing coffee under shade. Costa Rica (© Harmand/Cirad)


The latest land use survey, done by the National Meteorological Institute in 1992, showed that 54.08 percent of the national territory is covered by primary and secondary forest; 32.32 percent is used for livestock production; and agriculture, infrastructure and water account for the remaining 13.6 percent (NMI et al, 1992).

The international market for meat brought about considerable change in forest cover over time as livestock production activities steadily pushed back the forest frontier. When the market for meat fell in the nineties, secondary forest regenerated on rangelands no longer used for grazing.

Tree systems outside forests, in addition to their ecological importance in carbon sequestration, account for over half of wood production in Costa Rica. It is therefore imperative to conserve and enhance them. With this in mind, the regional project TROF: Tree Resources outside Forests worked out a methodology to inventory and monitor these resources.


Forests (and hence Trees outside forests) are defined differently in Costa Rica depending on the sector involved. The definition may be legal, technical, environmental, or other, which makes it harder to compare the various studies of forest and crown cover. There are also no explicit criteria for the definition of terms such as natural forest, secondary forest, and unlogged forest.

Forest law, which comes under the Ministry of Energy and Environment (MINAE: Ministerio del Ambiente y Energia responsible for the application of forest laws and regulations, defines the forest as "an original or autochthonous ecosystem, logged or not, regenerated naturally or silviculturally, covering two or more ha, and characterized by the presence of uneven-aged mature trees of various species and sizes, with one or more stories, a crown cover of at least 70 percent, and including at least 701 trees/ha with a minimum dbh of 15 cm" (Forest Law 7575, Article 3).

A typology has been devised for tree systems outside the forest (Kleinn, 1999; Morales, 1999). It makes a distinction between tree formations lying outside natural forests and under two ha in area, and tree formations that are the result of human activity. Eight categories have been classified as follows: agroforestry systems, scattered trees in pastures, line planting, trees growing among permanent crops, trees growing among annual crops, windbreaks, living fences, and trees growing in built-up areas. Silvopastoral systems are in the statistical forefront with coverage of 30.22 percent of the territory, followed by seasonal crops at 2.57 percent, coffee plantations at 2.11 percent, and palm groves at 0.56 percent, orange groves at 0.48 percent and mango at 0.12 percent (NMI, 1996; MAG, 2000).

Patterns and extent of change

The pattern of change in tree systems outside forests is closely linked to a land privatization process which fostered deforestation. Under decolonization, title could be granted only to land acknowledged as productive land, which meant that it had to be cleared prior to use. Successive governments upheld this provision, maintaining the rule that land title was to be conferred only on land free of forests. It is also true that with the arrival of the first Europeans, forests were cut back to provide land for livestock production - first to meet local demand and, subsequently, to supply foreign markets. The 1961-1995 data on land use show that forest land shrank from some 3 240 00 ha to 1 569 000 ha during that period, whereas rangeland rose from 915 000 ha to 2 330 000 ha (FAOSTAT). Grasslands, which in l961 represented only 18 percent of the territory in 1961, covered over 45 percent by 1990.

Other causes of deforestation, in addition to the expansion of the agricultural and livestock frontiers, were unauthorized logging, forest fires, credit policies and urbanization. According to González and Lobo, 1999, deforestation rose to an annual 50 000 ha in the years between 1950 and 1990. Studies done in 1967 and in 1977 provided estimates of the area of the various tree formations and the density of tree cover. In 1967, 48 percent of the country was forested with a crown cover of 90 percent, whereas the same parameters were 33 and 81 percent by the year 1977. Regression of the cover of virtually all tree resources was also noted (Table 9).

The literature on the economic, ecological and social scope of Trees outside forests is rather scant. It is known, however, that 43.43 percent of the total volume of wood harvested in 1990 came from this resource. By 1998, this figure had risen to 51.58 percent (Gonzàlez, Lobo, 1999). The trend is bound to accentuate in that mature forest resources are increasingly limited. Wood resources currently supply 18.4 percent of the energy produced (Central Bank of Costa Rica, 2000). In economic terms, trees not growing in forests contribute 0.38 percent of GDP. Coffee production represents 26.93 percent of the aggregate value of the agricultural sector and bananas 18.49 percent, with the forest sector contribution estimated at 8.92 percent (McKenzie, 2000). In the 1980s, the agricultural sector accounted for 20 percent of GDP, of which one-fifth to one-quarter from the meat sub-sector (Pérez, 1995).

Various tree species have been used in non-forest tree systems to improve soil fertility, inter alia in silvopastoral schemes (Russo, 1981; Canet, 1986). Trees outside forests also play a role in biodiversity: 79 different species have been identified in an area of 25 000 ha of pasture (COSEFORMA, 1995). Line planting, windbreaks, living fences and other pasture borders are introduced systems which nature often alters over time. The species composition of these systems depends partly on the environmental circumstances and farmer preferences, but mainly on the availability of nursery seeds and seedlings. Line planting in corridors influences herd movements and seedling distribution (Burel, 1996, cited in Camero et al, 2000). Such systems act as biological corridors, which are essential in an agricultural landscape characterized by fragmented ecosystems.

Table 9: Comparison of forest cover in 1967 and 1977





Tree cover


Crown cover

Tree cover


Crown cover













3 628



2 03,2


5 220






13 337



2 814,1


18 947



2 349,4



7 716



3 881,1


7 857



3 543,5



24 528



22 148,8


16 806



13 629,7



1 891





2 270






51 100





51 100




Source: Sylvander, 1981, in Kleinn, 1999
Cat. I Zones with few or no trees, used for agriculture and/or livestock;
Cat. II Zones with scattered trees and small forest formations used primarily for agriculture or livestock;
Cat. III Zones with large forest formations and farmland or grasslands;
Cat. IV Zones entirely or mostly covered by forest;
Cat. V Zones covered by mangrove, palm or mountain shrub.
Categories I, II and III include mostly Trees outside forests.

Institutional and management aspects

Natural resources are regulated by legislation covering wildlife, national parks and biodiversity, in addition to forest law. However, these legal instruments are not designed to regulate the management of trees outside forests - though forest law does devote the most attention to the issue, as it specifically regulates the management and use of agroforestry systems. It is important to point out that the Government supports tree resource management through financial incentives to forest and forest plantation owners. Trees not growing in forest areas are neither specifically nor adequately covered by these measures, and so they tend to be sidelined, and, over time, degraded.

A certain amount of data has been produced on trees outside forests in recent years, particularly in agroforestry systems. CATIE, the Tropical Agricultural Research and Training Centre in Turrialba, has been working on the topic for years. In collaboration with universities and with the Ministry of Agriculture and Livestock, CATIE has given priority to the development of agroforestry. The emergence of a national development and management policy for this invaluable but as yet little-documented resource would be greatly favoured by increased and accessible information on the subject.

Farmers and livestock producers, who improve the countryside by planting trees along rivers and streams, living fences around their holdings, windbreaks, trees in vegetable gardens, and the like, are aware that this is an economically profitable and ecologically viable resource. It is worth pointing out that the economic and cultural background of farmers and stockowners has a bearing on the importance they attach to these trees.

Assessment and Planning

Most of the completed inventories of tree systems outside forests have focussed on agroforestry systems at the smallholder level. They cover small areas, use random or systematic sampling systems, and gather data on circular or rectangular sampling plots.

There have been a few inventories covering larger areas, such as a 1981 inventory based on aerial photographs which systematically sampled 1 km2 plots throughout the country (Sylvander, 1981). In 1995, natural forests in the Huetar Norte region and rangelands with six or more trees/hectare were assessed using clusters of sample plots forming a scheme of sampling points covering this region (COSEFORMA, 1995). Another inventory covered on-farm trees, based on a sampling of holdings (Van Leeuwen and Hofslede, 1995; Harvey and Haver, 1999). In 2000, the TROF project inventoried 10 000 representative hectares in the Pacìfico Norte region, systematically selecting 5 100 ha sub-plots on which all trees were counted.

The results of these studies remained in the technical domain, the data untapped by policy makers, who need to be made aware of the significance of these resources and provided with the country-wide numerical data they would need for planning.

TROF Inventory Project in Central America

The TROF Project covering the Central American region has been in operation since 1998. Implemented by the MAG in collaboration with CATIE and the University of Fribourg in Germany, its dual objective is to formalize an inventory and monitoring methodology and to disseminate the ensuing data.

The inventory is now completed. The two-stage sample was the method deemed most appropriate for trees growing outside forests. The recommended application of this sampling method is to pre-select the sampling sites (primary units of 1 km x 1 km) based on existing data, followed by sample plots on the ground as secondary units, in proportion to the tree cover of the resource, assuming the presence of Trees outside forests as the most relevant variable. The ideal situation would be pre-stratification based on the segmentation and fusion of a LandSat image with an IRS image, (a process to be developed by the TROF Project), enabling a stratification of densities. Subsequently, in accordance with the rules concerning sample size, the primary units would be selected, and, within each, the secondary units.

The methodology was tested on silvopastoral systems because they are so extensive in Costa Rica. The inventory was done near Cañas in Guanacaste in the dry northern Pacific region on an overall area of 294.33 ha (cf. table 10). All trees with a dbh of 10 cm or more and a mean height of at least 5 m. were counted. The height limit selected for living fences was 1.3 m.

In 2000 the TROF Project worked on the development of the final stage of the sample. The job involved analyzing and simulating different probability sampling methods. Some results are already available and the method is now being tested. The experience acquired by the TROF Project should help identify the necessary methodological and operational components for a successful inventory of trees outside forests.

Table 10: Data summary from the inventory on pasture in Cañas, Guanacaste

Type of formation

Number of

Ground area


Number of species

a. Undisturbed primary forest





b. Older secondary forest





c. High `Tacotal'





d. Pasture not bordered by trees





e. Pasture bordered by trees





f. All pasture (d+e)





Ratio of pasture/forest (f/a x 100)

7,00 %

8,68 %

4,19 %


Source : TROF  Project, 2000

Plot size must be considered regardless of plot shape, and the plots have to be compact. In Central America, as indeed elsewhere, there tend to be many fairly small holdings. This implies a time-consuming search for authorization to take ground measurements. One option is to use easily plotted, 50m x 50 m (or 100m x 100 m) square plots. All trees growing on these plots are measured and land use noted using a dot grid, or a point in the plot centre.

Where remote sensing is the chosen method, high resolution is essential due to frequent perceptual problems with satellite images of trees growing outside forests. Where aerial photographs are used, the preferred scale is 1:50 000 - 1:10 000 (though cost may be a constraint, as the TROF Project found). Due to the rapid pattern of change in these systems, the photographs should be recent. And for work on the ground, access to trees growing on privately owned land can be a further constraint. The heterogeneity of this kind of resource, its variable configuration and the diversification of its physiognomy further complicate the sampling process, necessitating different sampling intensities and different types of plots. Lastly, mindful that the definition of forests varies from one country to the next, the objective of the study needs to be constantly borne in mind.

The data from the inventory methodology tested in Costa Rica should make it easier for users and policy makers to include Trees outside forests in their development and management policies. The data are not only useful for regional interpretation, they might also serve to map non-forest tree resources, estimate stored carbon and volumes of wood, and identifiy tree species and biological corridors.


Trees outside forests are increasingly on the agenda. People are learning that, over time, this resource alone can meet wood and non-wood requirements in many countries. In Costa Rica these trees should be included on a more systematic and specific basis in both management and incentive policies for tree conservation conservation and regeneration, and in felling and harvesting regulations. Policies need to be devised to promote them, regulate their use and management, and establish cooperative links among all stakeholders affected by the future of trees outside forested areas.


BCCR. 2000. Cuadros estadísticos. Central Bank of Costa Rica. San José. Consulted on 20 Aug. 2000. Website

Camero, J.; Camargo, M.; Ibrahim, A. and Schlonvoigt. 2000. Agroforestería y sistemas de producción animal en Centro América. In Pomareda, C. and Steinfeld, H. 2000. Intensificación de la Ganadería en Centroamérica: Beneficios Económicos y Ambientales. CATIE-FAO-SIDE. pp 177-198.

Canet, B. 1986. Características del sistema silvopastoral jaul (Alnus acuminata) con lechería de altura en Costa Rica. In Symposium sobre Técnicas de producción de leña en fincas pequeñas y recuperación de sitios degradados por medio de silvicultura intensiva. Costa Rica.

COSEFORMA. Cooperación en los sectores forestal y maderero). 1995. Inventario Forestal de la Región Huetar Norte. Resumen de resultados. IIe edition. Based on the GWB Consultation (Gesellschaft für Walderhaltung and Waldbewirtschftung, mbH) and COSEFORMA document no 43: Evaluación de la calidad de plantaciones forestales en la Región Huetar Norte de Costa Rica. 26 pp.

González, F. and Lobo, S. 1999. Breve descripción del estado de los recursos forestales de Costa Rica. In Kleinn, C and Davis, R (eds) 1999. Proceedings of the Workshop on Forest Resources in Eleven Countries of Latin America. Turrialba, Costa Rica 17-20 May 1999. pp 65-81.

Harvey, C. and Harber, W. 1999. Remnant trees and the conservation of biodiversity in Costa Rica Pastures. Agroforestry Systems 44: pp 37-68.

IMN et al. 1992. Mapa de cobertura de la tierra de Costa Rica. Atelier MAG, IGN, DGF, IMN. San José.

IMN. 1996. Proyecto Inventario nacional de Emisiones de Gases con efecto de invernadero. National Meteorological Institute. San José.

Kleinn, C. 1999. Compilation of information on Trees outside forests:Regional Special Study for Latin América, including also Haiti. Contribution to Forest Resources Assessment 2000, FAO. Tropical Agicultural Research and Training Centre (CATIE). 87 pp.

Kleinn, C. 2000. Inventory of trees outside forests: how to survey a scattered heterogeneous resource. Document presented at the 21st IUFRO World Congress. Kuala Lumpur. Malaysia.

MAG. 2000. Estadísticas del sector agrícola. Ministry of Agriculture and Livestock. San José. Consulted 24 Aug. 2000. Website

Mckenzie, T. 2000. Actualización de la metodología estadística para el sector forestal industrial de Costa Rica. MINAE, COSEFORMA, GTZ, FONAFIFO, ONF. San José. 38 pp.

Morales, D. 1999. Recopilación de información sobre árboles fuera de bosque en Costa Rica. Tropical Agricultural Research and Training Centre (CATIE). Statistics Unit. [unpublished].

Perez, E. 1995. Competitividad del sector ganadería de carne en Costa Rica. Summary of the study «Cadena Agropecuaria de Carne». Consultation for IICA. San José.

Russo, R. 1981. Árboles con pastos, justificación y descripción de un caso en Costa Rica. Tropical Agricultural Research and Training Centre (CATIE). Turrialba. Costa Rica.

Sylvander, R. 1981. Los Bosques del país y su distribución por provincias. Ed. EUNA.San José.

TROF. 2000. Descripción y objetivos del proyecto. Consulted 12 Aug. 2000. Website  

Van Leeuwen, A. and Hofslede, A. 1995. Forests, Trees and farming in the Atlantic zone of Costa Rica. An evaluation of the current and future integration of trees and forest in farming systems in the Atlantic Zone of Costa Rica. Technical Series. CATIE Technical Paper no 257.

1 Freely translated

Previous PageTop Of PageNext Page