Evaluation of biodiversity and structure indices in cultivated forests.
R. Salas1*, B. Fidalgo2, J. Gaspar1 and P. Morais1
1 Instituto Politécnico de Coimbra; Centro de Estudos de Recursos Naturais, Ambiente e Sociedade, Escola Superior Agrária, 3040-316, Bencanta, Coimbra, Portugal. Tel: 00 351 239802940; Fax: 00 351 239802289, *; jgaspar@esac;
2 Universidade Técnica de Lisboa, Instituto Superior de Agronomia, Centro de Estudos Florestais, Tapada da Ajuda 1349-017, Lisboa, Portugal. Tel: 00 351 239802940; Fax 00 351 239802289,
Introduction
The Pan-European Forestry Certification system ensures that purchased wood comes from a forest managed with sustainable methods. The pilot project FORSEE was proposed to consolidate this process of certification by realising a ‘life size’ test of its criteria and indicators in pilot zones. As Atlantic partners, this study, aimed to carry out inventories and studies to assess the relevance of the indicators used in the system of certification. Among these indicators, biodiversity, stand structure and deadwood in these cultivated landscapes are essential aspects in regard to forest certification programs.
Material and Methods
Study area
The study area (40º03’N-40º12’N latitude; 08º09’W-08º18’W longitude) is located in the Municipality of Lousã, in the Central Region of Portugal. The landscape of the Municipality comprises valleys and mountains with steep to moderate slopes spreads over 13840 ha with an elevation ranging between 65 and 1200 m above mean sea level, with a temperate Mediterranean climate. The Pristine forest was basically composed of broadleaved species (oaks and beech), which were intensively exploited in the past. In the first half of the twentieth century, extended plantations of Pinus pinaster were carried out in order to produce material for wood industries. From 1950 onwards, with the growth of the cellulose industry, the Eucalyptus globulus plantations proliferated occupying a significant previous pine-covered area. Consequently, native broadleaved species lost almost all their natural distribution area. More than 60% of forest land in the studied area is private and consists of small dimensioned proprieties (1 ha average). In the past, intensive agriculture and forestry were dominant as land use, however the local population (settled in the lowland), has progressively abandoned these activities for the past two decades.
Photo-interpretation and land cover map
The land cover/use type information of the Municipality of Lousã was derived from false colour aerial photographs, scale 1:25000m provided by CELPA (Association of Paper Production). The classification had two levels, the first level consisted in: agriculture, forest, uncultivated land, settlement; the second level was a sub-classification of agriculture and forest uses.
Sampling and data recording
In 2005, the national forest inventory in Portugal (NFI) was carried out using a systematic plot grid (1 x 1 km). From these plots we randomly selected 73 circular plots of 500 m2. The recorded data from the NFI plot type was: 1) common biometrical measurements of a forest inventory (tree layer); 2) biometrical characteristics of snags (deadwood); 3) shrub species and cover (shrub layer); 4) herbs species and cover using the minimal area method, assessed on a variable plot area, placed in the centre of the forest inventory plot (herb layer). Additionally, in a circular satellite plot of 500 m2 -Sfix- (100m distance from NFI plot) data of snags were also registered.
Data treatment
Biodiversity and structure indices were examined for each plot (stand) by landscape type: 1) we estimated diversity indices within-sample or plot (alpha indices, Whittaker 1972), i.e. Richness (S) (total richness of each landscape type were compared with one non-parametric richness estimator –bootstrap-). Other indices concerning the proportional abundance of species, like Shannon (H’) and Pielou (J’) were calculated, both are considered evenness indices. Simpson (D) index dealing with species dominance were also estimated. We assessed the variation in species composition, examining floristic similarity within landscape types (beta indices, Whittaker 1972). We calculated Jaccard´s similarity coefficient, and we, then, carried out an agglomerative hierarchical cluster analysis between the plots of all landscape types. We used the average method with the Euclidean distance metric option. The variation of the species among landscape types was also evaluated by means of Whittaker’s index; 2) as a measure of landscape type complexity, we calculated the visual cover index –VCI- and the coefficient of variation (CV) of the diameter at breast height (DBH); 3) basal area (G) of snags were estimated, deadwood is acknowledge to be important for conservation biology, particularly large snags (DBH>= 40cm) are usually associated to the creation of different niches.
Results
Land cover/uses map
Forest land dominated landscape, followed by uncultivated areas; agriculture and settlement showed a reduced land cover. Pinus pinaster and Eucalyptus globulus dominated forest land cover with 67% and 23%, respectively; these landscape types presented extended continuous areas. Native broadleaved and exotic conifer were fragmented and dispersed stands of small dimension, respectively occupied 6% and 4% of the forest land (Figure 1).
Figure 1. Landscape types cover in the Municipality of Lousã.
a-diversity
This investigation recorded 189 species including vascular plants and bryophytes. Total Richness (observed and estimated by bootstrap method), together with Shannon, Pielou and Simpson’s indices revealed higher diversity and evenness in native broadleaved landscape type, followed by P. pinaster stands, uncultivated areas, and E. globulus stands; the lowest diversity was observed in the exotic conifer landscape type. As is showed in Figure 2, diversity decreased from herb layer to tree layer in all landscape types. An ANOVA confirmed that Richness and Shannon’s index were significantly different among landscape type (d.f. 4, F=3,31, P< 0,01111 and d.f. 4, F= 2,95, P < 0,0127, respectively). The ANOVA of Simpson’s index did not revealed significant difference among landscape type. This ANOVA showed that Richness in broadleaved stands was higher than that observed in eucalyptus and exotic conifer stands. Moreover, the ANOVA of Shannon’s index also showed significant differences between broadleaved stands and eucalyptus plantations.
Figure 2. Alpha diversity indices.
b-diversity
Jaccard’s similarity confirmed that the floristic composition within landscape types showed variation. By analysing herbs, shrubs and trees layer it was possible to establish shortest distances between the plots belonging to the same landscape type, ─ merging clusters ─ (Figure 3). Whittaker’s index confirmed that tree layer showed the highest dissimilarity of floristic composition within landscape types (2.22), a lower variation in the floristic composition of the herb layer (2.06) and finally the lowest dissimilarity in shrub layer (2.04) within landscape types.
Figure 3. Agglomerative hierarchical cluster analysis, similarity in species composition.
Structure indices and deadwood
Vertical structure index, VCI presented highest values with low variation for all the landscape types (1000-1150) except for uncultivated land, where the scarce tree layer gave the lowest values in this index (VCI = 650). Horizontal structure analysis (CV of DBH) revealed a highest complexity in broadleaved stands, an intermediate complexity (DBH variation) in P pinaster stands, in E globulus stands and in uncultivated areas. The lowest values were observed in exotic conifer due to regular structure of these stands. E globulus and P pinaster presented highest values of deadwood in young trees; in contrast, native broadleaved species also showed some mortality in adult trees, including those with a sixty-centimetre DBH.
Discussion
The Richness recorded in this study was similar or higher compared to other temperate forests in Europe (Estevan et al. 2007), i.e. Richness in oak forests was higher than that in conifer plantations, (Coroi et al. 2004). Shannon and Simpson’s indices showed: the highest diversity values and the best evenness values in old native broadleaved stands. P. pinaster and E. globulus stands together with uncultivated areas, where shortest period of clear-cut or recent fires produced disturbance, revealed lower evenness due to the number of species and individuals present in these landscape types. The worst situation was detected in exotic conifer stands; this general behaviour was also accounted by Kalacska et al. (2004).
Jaccard’s index showed greatest similarity of floristic composition within plots of the same landscape type, but in general, dissimilarity was observed within all landscape types. This could be interpreted as three different succession stages of development: early, intermediate and late stages. Therefore, some similarity was detected in herb, shrub and tree layers, especially among P pinaster and E. globulus stands (early and intermediate stage). Exotic conifer stands (late stage), uncultivated areas (early stage) and broadleaved stands (late stage) showed higher dissimilarity within all landscape types, as observed in other studies (Kalacska et al. 2004). VCI is effectively influenced by the importance given to tree layers, consequently, regular stands presented lower values than irregular stands (Ferris et al. 2000). Nevertheless, it would be relevant to confer the appropriate weight to shrub and herb layers.
The VCI could be complemented with information coming from horizontal structure assessment, like CV of DBH, this calculation allow us to better attest the relevance of the nature of the stand, being that the broadleaved stand shows more variation than conifer plantations as pointed by Coroi et al. (2004). Finally, Peterken (1996) observed that mortality rate and deadwood recruitment increased as mature trees aged and became more vulnerable to diseases and physical instability. He also noted that the lowest mortality rates corresponded to mature vigorous trees of around 40–70 cm DBH. In our study, the scarce management practices produced higher mortality in young trees (DBH < 15 cm) in E. globulus and P pinaster stands.
Conclusions
This study assessed the usefulness of alpha and beta indicators biodiversity among forest types. Significant differences among forest types were found within forest plantations holding lower biodiversity than native broadleaved stands. It also revealed that the behaviour of some biodiversity indicators, such as deadwood, had a different meaning in this type of landscape. Structure indices should be applied complementarily to assess biodiversity.
Bibliography
Coroi, M., Skeffington, S.M., Giller, P., Smith, C., M. Gormally and G. O’Donovan. 2004. Vegetation diversity and stand structure in streamside forests in the south of Ireland. For. Ecol. and Manage. 202(1-3): 39–57.
Estevan, H., Lloreta, F., J. Vayreda and J. Terradas. 2007. Determinants of woody species richness in Scot pine and beech forests: climate, forest patch size and forest structure. Acta Oecologica. 31: 325-331.
Ferris, R., Peace, A.J., J.W. Humphrey and A.C. Broome. 2000. Relationships between vegetation, site type and stand structure in coniferous plantations in Britain. For. Ecol. and Manage. 136(1-3): 35-51.
Kalacska, M., Sanchez-Azofeifa, G.A., Calvo-Alvarado, JC, Quesada, M., B. Rivard and D.H. Jansen. 2004. Species composition, similarity and diversity in three successional stages of a seasonally dry tropical forest. For. Ecol. and Manage. 200 (1-3): 227-247.
Peterken, G.F. 1996. Natural woodland: ecology and conservation in northern temperate regions. Cambridge: Cambridge University Press, 522 pp.
Whittaker, R.H. 1972. Evolution and measurement of species diversity. Taxon 21(2-3): 213-251.