Alejandro Velázquez, Alejandra Fregoso, Gerardo Bocco y Gonzalo Cortez

Alejandro Velázquez Montes. Ph.D. in Landscape Ecology, University of Amsterdam, The Netherlands. Researcher, Institute of Geography, Universidad Nacional Autónoma de México (UNAM), Morelia. Address: Aquiles Serdán Nº 382; Colonia Centro, C.P. 58000, Morelia, Michoacán, México. e-mail: avmontes@igiris.igeograf.unam.mx

Alejandra Fregoso Domínguez. M.Sc. in Geo-Information Science and Earth Observation, International Institute for Geo-Information Science and Earth Observation (ITC), The Netherlands. Researcher, Instituto Nacional de Ecología, SEMARNAT, México.

Gerardo Bocco Verdinelli. Ph.D. in Landscape Ecology, University of Amsterdam, The Netherlands. Researcher, Centro de Investigaciones en Ecosistemas and Instituto Nacional de Ecología, SEMARNAT, Mexico.

Gonzalo Cortez Jaramillo. M.Sc. in Forest Management, Colegio de Postgraduados, Chapingo, Mexico. Lecturer, Instituto Tecnológico Agropecuario plantel 7, México.

Este estudio compara de manera cuantitativa (ventajas y desventajas) entre los enfoques de cartografía de la vegetación y rodales forestales. Los resultados intentan conciliar el uso y la conservación del mismo objeto de estudio, "el bosque". El enfoque para el estudio de la vegetación siguió los métodos de descripción y clasificación de la escuela Europea. En total se realizaron 177 relevés siguiendo un diseño estratificado. Se distinguieron 13 comunidades de plantas e identificaron las especies características de cada comunidad. El enfoque para el estudio forestal siguió un diseño estratificado sistemático de muestreo; en donde se muestrearon 136 rodales forestales subdivididos en 1271 subrodales. El estudio incluyó la colecta de datos dasométricos en 4662 sitios de aproximadamente 1000 m² por sitio. Para la comparación se consideraron tres componentes: información florística, dinámica de la vegetación y representación espacial. El estudio de vegetación incluyó 609 especies de plantas vasculares, mientras que el forestal 11 taxa (sólo aquellos de relevancia comercial para productos maderables). Las 13 comunidades de plantas representan diversos estadios de la dinámica de la vegetación, mientras que el forestal solo incluye la dinámica de las poblaciones de especies seleccionadas. En el contexto espacial, ambos enfoques resultaron muy afines (70%). Las diferencias se explican por rodales forestales que albergan mosaicos heterogéneos de comunidades de plantas pero con fisonomías relativamente homogéneas. Se recomiendo fuertemente integrar una estrategia de muestreo complementaria de ambos enfoques para obtener de manera simultanea la información relevante para acciones de uso y conservación del bosque.

Este estudo compara de maneira quantitativa (vantagens e desvantagens) entre os enfoques de cartografia da vegetação e rodales florestais. Os resultados tentam conciliar o uso e a conservação do mesmo objeto de estudo, "o bosque". O enfoque para o estudo da vegetação seguiu os métodos de descrição e classificação da escola Européia. Em total se realizaram 177 releves seguindo um desenho estratificado. Se distinguiram 13 comunidades de plantas e identificaram as espécies características de cada comunidade. O enfoque para o estudo florestal seguiu um desenho estratificado sistemático de amostragem; aonde se observaram 136 rodales florestais subdivididos em 1271 subrodales. O estudo incluiu a coleta de dados dasométricos em 4662 lugares de aproximadamente 1000 m² por lugar. Para a comparação se consideraram três componentes: informação florística, dinâmica da vegetação e representação espacial. O estudo de vegetação incluiu 609 espécies de plantas vasculares, enquanto que o florestal 11 taxa (somente aqueles de relevância comercial para produtos madeiráveis). As 13 comunidades de plantas representam diversos estados da dinâmica da vegetação, enquanto que o florestal só inclui a dinâmica das populações de espécies selecionadas. No contexto espacial, ambos enfoques resultaram muito afins (70%). As diferenças se explicam por rodales florestais que abrigam mosaicos heterogêneos de comunidades de plantas, mas com fisonomias relativamente homogêneas. Se recomendou fortemente integrar uma estratégia de amostragem complementaria de ambos enfoques para obter de maneira simultânea a informação relevante para ações de uso e conservação do bosque.

KEYWORDS / Conservation / Indigenous Communities / Landscape Approach / Mexican Forest / Vegetation Mapping /

Received: 06/06/2003. Modified: 10/02/2003. Accepted: 10/14/2003

Developing inter-tropical countries are subjected to severe forest degradation and conversion processes (Myers, 2000). In these countries, where most biodiversity occurs, high human population densities and ill-planned development programs exert a strong pressure over the forests (Wahlberg et al., 1996) As a consequence, natural resource depletion processes are dramatic (Myers, 2000). During the last decade, long-term forest use and conservation has become a key issue. Contemporary management (timer management), forest resources as soils, water, biodiversity and timber, rely upon management schemes determined by human demands so that their natural dynamics is rarely taken into account (Vogt et al., 1997). The goal of meeting present human needs without compromising the availability of forest resources for future generations has been addressed by the Brundtland Commission (CED, 1997). Currently, forest management encompasses profitable economic use, soil, water and wildlife conservation, and eventually the maintenance of climatic conditions, simultaneously (Daily et al., 1996; Oliver et al., 1992; Sist et al., 1998). Finding compromises between forest use and conservation where anthropogenic activities are seen as key yardsticks has become a cornerstone of environmental scientists (Seymour and Hunter, 1999). Under this view, alternative paths based upon robust scientific methods need to be undertaken in order to strengthen current forest use plans (Velázquez et al., 2001).

Contemporary forest management plans promoted wood extraction of profitable tree species (Wolf, 1998; Seymour and Hunter, 1999); alternative forest uses were not economically attractive (Daily et al., 1996). Timbering schemes simulated natural forest disturbances such as fires, plagues or hurricanes, to determine the amount of extractable wood (Brokaw and Lent, 1999). The potential available wood volume was related to the intensity of the disturbance without considering the inherent forest dynamics (succession and evolution) and its spatial heterogeneity (Spies and Turner, 1999). In other words, static and homogeneous forest patterns are assumed, regardless of temporal or spatial scales.

A landscape approach may, to some extent, serve as a basis for developing ecologically sound forest use schemes (Mummery et al., 1999; Velázquez et al., 2001). Landscape ecology deals with the totality of physical, ecological and geographical entities, integrating all natural and human patterns and processes (Farina, 1998). Furthermore, the analysis of structure, composition and function allows prediction of landscape dynamics (Pitkänen, 1998; Palik and Engstrom, 1999; Neave and Norton, 1998). Natural geographic entities and their inherent heterogeneity across spatial units, and homogeneity within the unit, may be considered in conducting rapid forest use and conservation actions (Spies and Turner, 1999; Mummery et al., 1999). In this perspective, forest stands can be understood as ecological as well as productive bodies. Thus, timber and non-timber alternative uses can be evaluated simultaneously.

This paper discusses the potential contribution of an integrated forest and landscape approach to developing long-term forest management and conservation schemes. The research was conducted at a forest indigenous community in central Mexico, where both economic capital efficiency and conservation of biological carrying capacity are demanded simultaneously (Velázquez et al., 2001).

Nuevo San Juan Parangaricutiro is an indigenous (Purepecha) community located 15km east of Uruapan, state of Michoacan (Figure 1). Climate is temperate and seasonal with a mean annual precipitation of 1200mm and mean annual temperature of 15ºC (García, 1981); soils are derived from young and recent volcanic materials (Rees, 1970; Inbar et al., 1994). The main land cover is characteristic of temperate forest (Rzedowski, 1981). Land use includes subsistence agriculture, cattle grazing, avocado orchards and forestry. A thorough description is provided by Bocco et al. (2000).

Currently, 1300 comuneros (family heads that conform the community) who have granted rights on the communal land and their families, inhabit the Area. The major economic activity is the Community’s forestry enterprise, with some 850 indigenous employees earning wages above the minimum salary, an unusual fact in rural Mexico (Bocco et al., 2000). The Community is well known for its sustained use of forest and the integrated management of derived goods (Alvarez-Icaza, 1993). Manufactured products, including wooden floors, furniture and resin, are commercialized at the national and international markets. The Community was granted the right to administer its own forest technical services in 1988, thus receiving the complete control of the resource by the government (Velázquez et al., 2001). In 1998, Nuevo San Juan received the green certification by the Smartwood World Forest Council. This recognition implied both economic and ecological benefits, and promoted the search for alternative forest uses by the general assembly of the community. Further productive diversification may strengthen this social enterprise (Kolosvary and Corbley 1998).

The community area under forest cover was stratified using 1996 panchromatic black and white aerial photographs at a scale of approximately 1:25000. The photo interpretation was carried out on the basis of standard photographic image elements (tone, texture, pattern, shape and location). Delineation of 1271 homogeneous forest stands for timer management purposes was based on similarities in forest cover (Velázquez et al., 2001), topography and tree density (Figure 2). The units were digitized, geometrically corrected in a vector-format mosaic and labeled according to the legend as a forest stand map. For this procedure a geographic information system (GIS; ILWIS, 1997) was used. Figure 2 describes the processes followed to obtain the maps from forestry and ecological approaches.

Once the stands were defined, each was evaluated in terms of its exploitable wood volume and classified in terms of its quality for management plans purposes using the Site Index, which is an important component in growth and yields models and reflects site productivity as the average height of the dominant tree. The index age was set at 50 years. For that purpose forestry data were collected under a systematically sampling scheme on 4662 sample plots. These circular plots were approximately 36m in diameter (1000m²). In every plot, 30 variables were measured including elevation, aspect, slope, tree species, and forest stand parameters such as DBH (1.30m), height and basal area (Bocco et al., 2000). Emphasis was placed on commercially profitable tree species (Pinus pseudostrobus, P. montezumae, Abies religiosa, Quercus spp and Cupressus lindleyi).

Volume models for each of the profitable tree species were developed. A multiple regression model where volume is a function of stem diameter and height was used. The best model was the combined variable and the equation was adjusted to a log lineal function for each species (Eq. 1). Estimation of height growth patterns for the profitable tree species were developed. The Schumacher growth algorithm was selected as the most robust model for stand height prediction with the aid of Statistics Analysis Software (Cody and Smith, 1987).

The forest variables were handled in a relational database, and linked consistently to the spatial database in the GIS; the relational key was the identifier of every polygon (stand). Once the stands were characterized according to its productivity (site index and Schumacher model) these were then regrouped on the basis of their quality status and represented spatially using the GIS as a Forest quality map.

Stratification of the forest area was accomplished on the same set of aerial photographs as for the previous approach. Delineation of homogenous vegetation units was based on the same photographic elements as above (Figure 2). The discriminated vegetation types on the photographs were coniferous forest (Abies, Pinus), broad-leaf forest (Quercus, Alnus, Salix, Clethra, Arbutus), non-forest vegetation cover (Baccharis), and reforestation stands. The units were also digitized, geometrically corrected in a vector-format mosaic that matched the previous forest mosaic geometry and labelled according to the legend, as a preliminary vegetation map in the GIS (sensu Velázquez, 1993).

Vegetation was described following the Zürich-Montpellier approach (Werger, 1974) on 177 vegetation sample sites (relevés in the original terminology). The vegetation scheme was carried out on the vegetation units defined under a stratified random sampling strategy. Sites were homogeneous and representative of the vegetation types; at least 3 relevés were surveyed per vegetation mapped polygon. Both size and shape of sampling units were defined according to the concept of minimum area, on the basis of ecological homogeneity and the relationship species-area (Werger, 1974; Braun-Blanquet, 1979).

For every sampling site the following data were recorded: physiognomic and physiographic site description, geographic coordinates, relief and micro-relief, altitude, slope gradient and aspect, soil depth (including depth of litter), disturbance characteristics and a complete floristic census of all vascular plants. The floristic description was accompanied by a quantification of cover abundance per species (Velázquez, 1993) and per stratum (tree, shrub, grass and herb layers). Cover was estimated, per species, as the total projection on the ground of all of the foliage of individuals of the same species (Werger, 1974). The variables were handled in a second relational database, and linked consistently to the corresponding spatial database (preliminary vegetation map) in the GIS; the relational key was the identifier of every polygon (vegetation type).

Vegetation data were integrated and analyzed through a numerical classification method using two way indicator species analysis program (TWINSPAN; Hill, 1979). This procedure allowed the recognition of all vegetation communities and their species affinities. In order to typify plant communities and to identify characteristic species, the degree of presence and average cover value per species were used (Mueller-Dumbois and Ellenberg, 1974).

To compare the two approaches, both relational databases were normalized in terms of comparable elements for the wooden taxa surveyed by the forest approach and two data matrices were built. Forest stands as well as plant communities were characterized on the basis of their species composition, and relative and absolute frequencies of tree species per stand were calculated (Fregoso, 2000). For the vegetation approach, only the commercially relevant tree taxa were considered on this vegetation data matrix including

where Fr: Stand/Plant community relative frequency, and f_ai: tree species (i) relative frequency.

To select the plant community to which each forest stand fits best, both matrices described above were compared and tree operations were conducted for the integration analysis: 1) selection of plant communities that shared the same plant taxa with a specific forest stand, 2) comparison of the tree species relative frequencies per stand and per plant communities and 3) selection of the plant community that presents the highest similarity of tree species relative frequencies per stand.

Once every stand was assigned to a unique plant community, the vegetation information was used to re-label the forest stands of the forest stand map with the name of the plant community assigned. For that purpose, the forest map and the vegetation-stand matrix described above were handled digitally through a geographic information system (ILWIS, 1997). For the spatial analysis five GIS operations were conducted: 1) detection of non-forested polygons and their exclusion from the spatial model, 2) re-labelling of forest stands according to the plant community they fitted best from the vegetation-stand matrix, 3) re-grouping of forest stand polygons comprising the same plant community label, 4) data display, and 5) designing the cartographic legend and printing.

The forest approach focused on wooden species allowed recognition of 11 different plant species, included in 4 categories: pine, fir, oaks and broad-leaved trees. These results contrast significantly with the 609 vascular plant species registered during the vegetation approach. From these, 422 species clustered into 189 genera and 77 families were depicted as characteristics of plant communities. The other 187 species excluded were considered rare (recorded £5 times in the 177 relevés). In brief, the over 600 vascular species represent a large potential for alternative uses, whereas contemporary forest management only uses about 2% of the total species richness recorded.

The characterization of the forest in terms of productivity stand quality for management purposes resulted in 4 classes (very high, high, medium and un-forested areas; Figure 3). The characterization of the forest in terms of its vegetation distinguished 13 plant communities; five typifying pioneer conditions and the rest representing mature forest structures (Figure 4). A complete list of preferential species depicting all forested plant communities is given in Table I. A thorough phytosociological description of these plant communities is provided by Gimenez et al., (1997) and Fregoso (2000).

This section includes the results obtained from the comparison of the two approaches and the spatial analysis that links plant communities and forest stands in a map (Figure 5). Results regarding the spatial distribution are presented in Table II. The integration approach shows that the vegetative community Pinus leiophylla-Piptochaetium virescens is best represented in 388 forest stands. This plant community is distributed on an area of 3533ha, on 85 polygons (units) covering 31% of the total forest mass coverage. The community of P. pseudostrobus-Ternstroemia pringlei was related to 433 forest stands covering 25% of the total forest mass coverage, on 136 polygons. The Abies religiosa-Galium mexicanum community is distributed on 2046ha, composed of 187 forest stands and 50 polygons. Pinus montezumae- Dryopteris sp. covers 1601ha represented by 20 forest stands on 16 polygons. The rest of the forest stands comprised plant communities covering surfaces from 10 to 100ha (Table II).

The vegetation approach included 609 vascular plant species, whereas the forest one only 11, those of importance for wooden products. Vegetation heterogeneity was well represented by the 13 plant communities depicted by the vegetation approach. In contrast, the forest approach only regards physiognomic heterogeneity of a few selected plant population species. In the spatial context, a substantial percentage of both approaches was successfully linked (70%). The rest of the forest stands harbor heterogeneous conditions that restricts linking plant communities and forest stands.

The contemporary forest approach (sensu Smith, 1962) and vegetation analysis under the landscape approach (sensu Zonneveld, 1995) provide substantially different information. The first refers, exclusively, to commercial tree life forms, giving most weight to forest density and forest structure. The second relies upon plant strategies and leading environmental factors involved in their distribution, where species composition, structure and physiognomy are therefore important. Thus, the overall impression of forest species richness differs significantly among approaches (see Table I). In addition, both consider geomorphologic features to delineate forest stand and landscape units respectively. The second, however, is regarded as the major geographic attribute to delineate landscape units (Velázquez et al., 2001), whereas delineation of forest stands depend mostly on density and height of the tree layer (Hunter, 1999). Furthermore, the landscape approach considers ecological processes such as succession, so that all vascular plant species play a role; therefore vegetation is seen as a dynamic attribute of the landscape where its distribution and development is determined mainly by climate, soils, relive and management activities. Whereas, the forests approach indirectly considers these factors as causes of the forest productive capacity, this analysis is mainly done at individual trees within the production unit area, regarding other ecological aggregation forms as vegetation communities.

The transitional areas (ecotypes, sensu Seymour and Hunter, 1999) were the most difficult areas to describe and to map (Werger, 1974). These ecotypes are usually avoided by the forest approach by sampling what is supposed to be homogeneous stands. These areas, nonetheless, include most disagreement between both approaches. As a consequence, forest stands considered homogeneous harbor large ecological heterogeneity, contrary to landscape units (Fregoso, 2000). To illustrate this further, 5% of the forest stands included a combination of three plant communities (Pinus leiophylla-Piptochaetium virescens, Abies religiosa-Galium mexicanum, P. montezumae-Dryopteris sp.). As seen in Figure 4, these plant communities are grouped into significantly different clusters. Ecological processes (e.g., succession and growth rate) as well as environmental processes (e.g., humidity, soils) also vary substantially among these communities.

The method developed appears to be an accurate way to join together these two approaches, where forest stands and vegetation units matched over 85% in their limits. This suggests that a complementary approach to link information is feasible. This is relevant since the sampling strategy (time-cost) in both approaches also differs significantly. The total forest volume estimation implied over 4500 sampling sites located along the transect (about US$ 80000). This contrasts drastically with the landscape approach since only 177 sampling units (relevés) were needed to typify all plant communities (about US$ 40000). The complete list of species and their analysis required over two years and three botanists to be completed.

To conclude, to ensure long term forestry use, a tied combination of forest (commercial woody species) and landscape (relief-soils-vegetation) approaches ought to be complemented (IUCN, 1996). This is meant to fulfill ecologically sound forest management (Giménez et al., 1997; Velázquez et al., 2000); and to favor natural landscape evolution (Hunter, 1999; Spies and Turner, 1999).

The authors acknowledge the staff of the indigenous community of Nuevo San Juan, especially Luis Toral and his team, for logistic and academic support. Field research was sponsored by DGPA-UNAM (IN- 210599), CONABIO (R092), and FMCN (B1-007/2).

1. Álvarez Icaza P (1993) Forestry as a social enterprise. Cultural Survival 17: 45-47.        [ Links ]

2. Bocco G, Velázquez A, Torres A (2000) Comunidades indígenas y manejo de recursos naturales. Un caso de investigación participativa en México. Interciencia 25: 9-19.        [ Links ]

3. Braun-Blanquet JJ (1979) Fitosociología: bases para el estudio de comunidades vegetales. Blume. Madrid, España. 820 pp.        [ Links ]

4. Brokaw N, Lent R (1999) Vertical structure. In Hunter MJr (Ed) Maintaining Biodiversity in Forest Ecosystems. Cambridge University Press. Cambridge, UK. pp. 373-399.        [ Links ]

5. CED (1997) Our common future. Internal report. Commission of Environment and Development. Oxford University Press. New York, USA.        [ Links ]

6. Cody RP, Smith JK (1987) Applied statistics and the SAS programming language. 2nd ed. SAS Institute Inc. North Carolina, USA. 280 pp.        [ Links ]

7. Daily CG, Alexander S, Ehrlich PR., Goulder L, Lubchenco J, Matson PA, Mooney HA, Postel S, Shneider SH, Tilman D, Woodwell GM (1996) Ecosystem services: benefits supplied to human societies by natural ecosystems. Issues in Ecology 2: 1-16.        [ Links ]

8. Farina L (1998) Principles and Methods in Landscape Ecology. Chapman and Hall. London, UK. 235 pp.        [ Links ]

9. Fregoso A (2000) La vegetación como herramienta base para la planeación, aprovechamiento y conservación de los recursos forestales: El caso de la comunidad indígena de Nuevo San Juan Parangaricutiro, Michoacán, México. Tesis. UNAM, México. 67pp.        [ Links ]

10. García E (1981) Modificaciones al sistema de clasificación climática de Köppen. Enriqueta García de Miranda. 4ª Ed. México, DF. 221 pp.        [ Links ]

11. Giménez J, Escamilla M, Velázquez A (1997) Fitosocología y sucesión en el volcán Paricutín (Michoacán México). Caldasia 19: 487-505.        [ Links ]

12. Hill M (1979) TWINSPAN A FORTRAN program for the detrendend correspondence analysis and reciprocal averaging. Cornell University. Ithaca, New york, USA. 52 pp.        [ Links ]

13. Hunter MLJr (1999) Maintaining biodiversity in forest ecosystems. Cambridge University Press. Cambridge, UK. 698 pp.        [ Links ]

14. ILWIS (1997) Application and reference guides. Integrated Land and Water Information System. Version 2.0. ITC, Enschede, Netherlands. 352 pp.        [ Links ]

15. Inbar M, Lugo J, Villers L (1994) The geomorphological evolution of the Paricutín cone and lava flow, México, 1943-1990. Geomorphol. 9: 57-76        [ Links ]

16. IUCN (1996) Communities and Forest Management. International Union for Conservation of Nature. Cambridge, UK.        [ Links ]

17. Kolosvary R, Corbley KP (1998) Forest management with GIS. Industry taps image processing and GIS to earn green certification. GIM-Feature 12: 27-29.        [ Links ]

18. Mueller-Dombois D, Ellenberg H (1974) Aims and methods of vegetation ecology. Wiley. New York, USA. 547 pp.        [ Links ]

19. Mummery D, Battaglia MC, Beadle L, Turnbull CRA, McLeod R (1999) An application of terrain and environmental modeling in a large-scale forestry experiment. Forest Ecol. Manag. 118: 149-159.        [ Links ]

20. Myers N (2000) Sustainable Consumption. Science 287: 2419.        [ Links ]

21. Neave H, Norton T (1998) Biological inventory for conservation evaluation IV. Composition, distribution and spatial prediction of vegetation assemblages in southern Australia. Forest Ecol. Manag. 106: 259-281.        [ Links ]

22. Oliver CD, Berg DR, Larsen DR, O´Hara KL (1992) Integrating management tools, ecological knowledge, and silviculture. In Naiman R, Sedell J (Eds.) New Perspective for Watershed Management. Springer. New York, USA. pp. 361-382.        [ Links ]

23. Palik B, Engstrom T (1999) Species composition. In Hunter RJr (Ed.) Maintaining Biodiversity in Forest Ecosystems. Cambridge University Press. Cambridge, UK. pp. 65-94.        [ Links ]

24. Pitkänen S (1998) The use of diversity indices to assess the diversity of vegetation in managed boreal forest. Forest Ecol. Manag. 112: 121-137.        [ Links ]

25. Rees JD (1970) Paricutin revisited: A review of man´s attempt to adapt to ecological changes resulting from volcanic catastrophe. Geoforum 4: 7-25.        [ Links ]

26. Rzedowski J (1981) Vegetación de México. Limusa, México. 432 pp.        [ Links ]

27. Seymour R, Hunter M (1999) Principles of ecological forestry. In Hunter RJr (Ed.) Maintaining Biodiversity in Forest Ecosystems. Cambridge University Press. Cambridge, UK. pp. 22-64.        [ Links ]

28. Sist P, NolanT, Bertault J, Dykstra D (1998) Harvesting intensity versus sustainability in Indonesia. Forest Ecol. Manag. 108: 251-260.        [ Links ]

29. Smith DM (1962) The practice of Silviculture. 8th ed. Wiley. New York, USA. 578 pp.        [ Links ]

30. Spies T, Turner M (1999) Dynamic forest mosaics. In Hunter RJr (Ed.) Maintaining Biodiversity in Forest Ecosystems. Cambridge University Press. Cambridge, UK. pp. 95-160.        [ Links ]

31. Velázquez A (1993) Landscape ecology of Tláloc and Pelado volcanoes, México. ITC publication Nº16. 151 pp.        [ Links ]

32. Velázquez A, Giménez J, Escamilla M, Bocco G (2000) Vegetation Dynamics on Recent Mexican Volcanic Landscapes. Acta Phytogeog. Suecica 85: 71-88.        [ Links ]

33. Velázquez A, Bocco G, Torres A (2001) Turning scientific approaches in to practical conservation actions. Environ. Manag. 27: 655-665.        [ Links ]

34. Vogt K, Gordon JC, Wargo JP, Vogt DJ, Asbjornsen H, Palmiotto PA, Clark HJ, O´Hara JL, Keaton WS, Patel-Weynand T, Witten E (1997) Ecosystems. Springer. New York, USA. 470 pp.        [ Links ]

35. Wahlberg N, Moilanen A, Hanski A (1996) Predicting the occurrence of endangered species in fragmented landscapes. Science 273: 1536-1538.        [ Links ]

36. Werger MJA (1974) On concepts and techniques applied in the Zurich-Montpellier Method. Of vegetation survey. Bothalia 11: 309-323.        [ Links ]

37. Wolf J (1998) Species composition and structure of the woody vegetation of the Middle Casamance region (Senegal). Forest Ecol. Manag. 111: 249-264.        [ Links ]

38. Zonneveld IS (1995) Landscape Ecology. An Introduction to Landscape Ecology as a base for Land Evaluation, Land Management and Conservation. SPB. Amsterdam, The Netherlands. 199 pp.        [ Links ] ]]> 1 1993 17 45-47 2 2000 25 9-19 3 1979 820 4 373-399 5 CED 1997 6 1987 2 280 7 1996 2 1-16 8 1998 235 9 2000 67 10 1981 4 221 11 1997 19 487-505 12 1979 52 13 1999 698 14 ILWIS 1997 352 15 1994 9 57-76 16 IUCN 1996 17 1998 12 27-29 18 1974 547 19 1999 118 149-159 20 21 1998 106 259-281 22 361-382 23 65-94 24 1998 112 121-137 25 1970 4 7-25 26 1981 432 27 22-64 28 1998 108 251-260 29 1962 8 578 30 95-160 31 1993 151 32 2000 85 71-88 33 2001 27 655-665 34 1997 470 35 1996 273 1536-1538 36 1974 11 309-323 37 1998 111 249-264 38 1995 199