Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction: Ecosystem Services Related to Plant Species Composition and Diversity

Biodiversity is rapidly declining in tropical ecosystems, and this decline has important implications for the functioning of the ecosystems and the services that they provide (Balvanera et al. 2006; Isbell et al. 2011; Millenium Ecosystem Assessment 2005). Tropical forests can provide a vast array of products, and they regulate many natural processes, e.g., climate regulation, maintenance of air and water quality, nutrient cycling, soil formation, and the sequestration and storage of carbon. They are also a repository for genetic resources, and in addition provide recreational value. Logging and pressures from agriculture or infrastructure development are resulting in the large-scale loss of forested land, and as a result a decline in the provision of ecosystem services (Millenium Ecosystem Assessment 2005).

Ecosystem services are in most cases not related to single plant species, but to the ecosystem as a whole (Lyons et al. 2005). Complex interactions between the components of biodiversity and physical and chemical environmental factors determine the quantity, quality, and stability of ecosystem services (Mace et al. 2012). However, these interactions are commonly poorly understood, especially with regard to diverse tropical ecosystems. The first step towards understanding them is to elucidate how environmental parameters and disturbance regimes drive species composition and diversity dynamics in tropical ecosystems. In addition, it is important to explore functional properties of species or species groups and their contributions to ecosystem functioning (de Bello et al. 2010; Cadotte et al. 2011).

This chapter summarizes how plant species composition and richness in the San Francisco Reserve and its surroundings are driven by elevation, topography, and different types of natural disturbances (landslides, canopy gaps). In addition, the effects of human land use (active and abandoned pastures) and forest fragmentation are discussed.

2 Materials and Methods

Plant species composition was studied according to a variety of different methods depending on the type of vegetation and research question to be addressed. To study the composition and species richness of trees along an altitudinal gradient, 17 permanent plots of 400 m2 size each were established at altitudes ranging from 1,850 to 2,570 m a.s.l. in the mature natural forest of the Reserva Biologica San Francisco (RBSF, Homeier et al. 2010). In addition, 18 plots (matrix plots) of the same size were installed at 2,000 m a.s.l. to study the effects of topography and related soil parameters on tree species composition and richness. These plots encompassed three different topographic positions (lower slopes, mid-slopes, and upper slopes), with six replicates located on each position (see Chap. 10). All plots were covered by mature, closed-canopy forest representative of the respective topographic situation. Tree regeneration in natural forest canopy gaps was studied in the lower montane forest by comparing the influence of the topographic position (upper slopes vs. lower slopes) on the regeneration (see Homeier and Breckle 2008 for details).

The natural vegetation of the lower montane forests (2,000–2,100 m a.s.l.) and the upper montane forests (2,400–2,500 m a.s.l.) was compared with anthropogenically modified vegetation found at the same altitudes. Terrestrial vascular plant species and their ground coverage (in %) were recorded within eight transects of 100 m2 size each (50 m × 2 m) at each site and elevation level (32 transects in total). Eight additional transects, two each, were laid out on active pastures, abandoned pastures, bracken fern areas, and pine afforestations within an altitudinal range of 1,900–2,200 m a.s.l. (see Diertl 2010 for detailed inventory methods).

In addition to terrestrial plants, vascular epiphytes were studied by recording the occurrence of plant individuals on entire remnant trees in pastures and on canopy trees of similar size and the same species in closed forests. Access to the canopy was achieved by single-rope climbing technique (see Werner et al. 2005 for details).

3 Results and Discussion

3.1 Altitudinal and Topographical Gradients as Drivers of Species Richness in the Rio San Francisco Valley

Angiosperm species richness in the San Francisco Valley generally decreases with altitude (Fig. 8.1). Although the existing inventory of approximately 1,200 angiosperm species (Homeier and Werner 2007) is far from being complete, general altitudinal patterns of species composition are not expected to change significantly with additional species records. The patterns of life form composition along the altitudinal gradient are rather similar at each altitude, although an absolute reduction in the number of species is evident. The decrease in species richness of shrubs and herbs that are the dominant life forms of the uppermost elevations is lower than that of epiphytes or trees.

Fig. 8.1
figure 00081

Angiosperm species richness of different plant life forms in natural forest habitats along an altitudinal gradient (drawn with data from Homeier and Werner 2007). Forest types according to Homeier et al. (2008)

Full size image

Ravine and lower slope forests are generally more productive (as expressed by tree basal area growth or fine litter production) and richer in tree species than are upper slopes and ridges (Homeier et al. 2010). Soil nutrient concentration appears to be a major factor in determining differences in forest structure and tree species composition within the study area (Homeier et al. 2010; Wolf et al. 2011). The differences in stand structure are principally related to tree species composition. Ravine forest tree species appear to be unable to recruit on the poor, acidic soils covering the upper slopes and ridges, whereas the slow-growing ridge specialists appear to be poorly able to compete for light with the faster-growing ravine species (Homeier 2008). There is a surprisingly small number of tree species shared between lower slopes and upper slopes, indicating that most species are restricted in their occurrence to specific sites.

Topographical gradients with their array of edaphic, climate, and landform parameters thus contribute to the variety of (micro-) habitats and increase vegetation heterogeneity as well as plant species richness of the tropical montane forests being investigated in the present study, as has also been shown for other montane forests (e.g., Webb et al. 1999). The changes in species composition and forest structure detected along topographical gradients often resemble those observed along altitudinal gradients (e.g., Webb et al. 1999; Takyu et al. 2002).

3.2 Forest Dynamics in Response to Natural Disturbance

The highly diverse natural forests of the study area harbor a vast number of plant species adapted to specific ecological niches within the spatially heterogeneous and temporally variable site conditions. Natural disturbance events are followed by distinct successional stages, each of which is characterized by a specific plant composition ranging from early pioneers on recent landslides or in canopy gaps to climax species in mature closed-canopy forest.

Two major natural disturbance regimes affect the study area, and each has long-lasting effects on the structure and the species composition of the vegetation. Tree and branch falls result in gaps in the forest canopy that change the light regime on the forest floor and the microclimate in the understory. The tree species composition of different successional stages on the ridges and upper slopes of the lower montane forest is rather uniform, and typical pioneer species are completely lacking (Homeier 2008; Homeier and Breckle 2008). In contrast, the recovering forest on lower slopes and ravine sites is characterized by the presence of fast-growing pioneer taxa such as Cecropia spp. (Urticaceae), Heliocarpus americanus (Malvaceae), and Piptocoma discolor (Asteraceae) which appear to be restricted to more fertile soils. Hence, the absence of pioneer trees on the upper slopes and ridges is likely to be due to unfavorable soil conditions, since the immediate vicinity of the two forest types makes dispersal constraints unlikely. Differences in tree composition and population dynamics in these two neighboring forest types are maintained during the succession from young towards mature forest.

While the stunted upper montane forests exhibit fewer canopy gaps than do lower altitude forests, they are susceptible to the second natural disturbance factor of landslides that have a severe impact on these forests (Restrepo et al. 2009; Muenchow et al. 2012). The duration required for forest recovery and the reestablishment of the original plant species composition after a landslide in the study area is highly dependent on the elevation at which the landslide took place and on the size and aspect of the landslide (Bussmann et al. 2008; Ohl and Bussmann 2004). However, successions setting on landslide sites always start with mosses and lichens covering the bare soil. The first vascular plants establish a foothold after 5–10 years, with Lycopodiaceae and Baccharis genistelloides (Asteraceae) being common early pioneers. Later, various fern species (especially Gleicheniaceae), orchids, grasses, and shrubs (Asteraceae, Ericaceae, and Melastomataceae) appear before first tree species become established.

Both canopy gaps and landslides contribute to the spatial heterogeneity of environmental conditions in the study area and thus drive ecosystem dynamics and species richness. It takes no longer than a few decades for open areas to have been reclaimed by closed-canopy forest. The large local species pool and the presence of species well adapted to the natural disturbance events constitute the basis for the resilience of this unique ecosystem and also for its resistance to invading alien plant species.

3.3 Land Use Patterns as Determinants of Secondary Vegetation

Ecuador is characterized by one of the highest deforestation rates in South America (FAO 2005; Mosandl et al. 2008). Today approximately 48 % of the montane forests below 2,200 m a.s.l. and at least 6 % of the higher altitude natural vegetation in the Rio San Francisco valley (Göttlicher et al. 2009) have been replaced by man-made ecosystems (Fig. 8.2; see also Chap. 2).

Fig. 8.2
figure 00082

Tropical montane forest and cattle pastures in the Rio San Francisco valley

Full size image

Topographic gradients are of great relevance for understanding the regeneration of vegetation after human disturbance. Accessible terrains on the lower slopes rising from the Loja-Zamora road (see Fig. 1.1c) into the mountains have been transformed into pastures by seeding or planting forage plant species. Agricultural crops (maize) play a subordinate role on deforested areas and are cultivated at most for 1–2 growing seasons subsequent to slash-and-burn practices. Forests on the upper slopes and ridges are burned frequently and repeatedly, but rarely taken under cultivation. The common forage species in the RBSF are Setaria sphacelata, Melinis minutiflora, Pennisetum clandestinum, Holcus lanatus, and Axonopus compressus (all Poaceae). Except for the native A. compressus, all of the forage grasses have been introduced; S. sphacelata being the most recently introduced African forage grass in the area. Due to its vigorous growth under a wide range of environmental conditions, this tussock grass has increased its share of ground cover over the past 20 years despite its low nutritional value (e.g., protein content).

Current land use practices have a severe impact on the natural ecosystem. This is clearly shown by a plant composition that is completely different to that characteristic of the natural forests and lacks the vast majority of the forest species. Figure 8.3 gives an overview of the family composition of the natural forest vegetation compared to secondary site vegetation at two different altitudes.

Fig. 8.3
figure 00083

Plant family composition of natural forest vegetation and secondary vegetation at two different elevations (redrawn after Diertl 2010 and Peters et al. 2010). Angiosperm families: AQUifoliaceae, ARACeae, ASTeraceae, BROmeliaceae, CARyophyllaceae, CYPeraceae, ERICaceae, LAUraceae, MELAstomataceae, MYRTaceae, ORChidaceae, PIPeraceae, POAceae, ROSaceae, RUBiaceae, SOLanaceae. Pteridophyte families: DRYopteridaceae, POLypodiaceae

Full size image

Only 9.9 % of the natural forest species are found on sites currently or previously under human use. Not more than 2.5 % of the forest species are found on pastures being currently grazed, where species richness depends strongly on the density of the dominant forage grass species. The dominance of the highly competitive S. sphacelata results in a low variety of accompanying herbaceous species (an average of 13 species) compared to the more traditional Melinis-, Pennisetum-, Holcus-, and Axonopus pastures (average of 33 species). All together, 245 species of vascular plants from 169 genera and 73 families were recorded for pastures situated between 2,000 and 2,200 m a.s.l. The families exhibiting the most species were the Asteraceae with 27 genera and 45 species, the Poaceae (16 genera, 24 species), and the Melastomataceae (6 genera, 15 species). Forty six percent of the species were herbs, 27 % shrubs, 14 % grasses, 7 % climbers and 6 % trees.

Common woody pioneer species are Tibouchina laxa, T. lepidota (both Melastomataceae), and Piptocoma discolor. Asteraceae shrub species such as Ageratina dendroides and Baccharis genistelloides occur very frequently across all secondary vegetation types; whereas Melastomataceae such as Monochaetum lineatum and Brachyotum benthamianum, as well as the Ericaceae shrubs Gaultheria erecta and Bejaria aestuans, are characteristic for bracken fern-dominated areas on previously burned sites. The highest average species density per 100 m2 was found in a pine afforestation (47 species), the lowest being detected on abandoned pastures (22) (Table 8.1). Because of the high competitiveness of the forage grass S. sphacelata, the numbers of species found on the abandoned pastures are consistently very low.

Table 8.1 Mean numbers of vascular plant species found in various land use situations at 1,900–2,200 m a.s.l. in comparison with the natural forest of the study area, together with the corresponding effective Shannon diversity D (Jost 2006) and evenness (natural forest data in the last column are derived from Diertl 2010)
Full size table

Floristically, the vast majority of the plant species growing on man-made sites are Andean elements, and only few are aliens. Of the latter, the most important species are cultivated for forage (of African origin: S. sphacelata, M. minutiflora, P. clandestinum, of European origin: H. lanatus, Trifolium repens) or for forestry (Pinus patula from Mexico). While none of these species is invasive in the study area, the cosmopolitan bracken fern Pteridium arachnoideum covers large areas that had previously been burned. However, the fern is restricted to open landscapes, as it does not invade natural forest stands.

3.4 Secondary Succession Processes in Anthropogenic Vegetation

The course of secondary successions on pastures is strongly influenced by the abundance of fodder grass species. The dense canopy and compact root network of the dominant species S. sphacelata prevent rapid succession from taking place upon pasture abandonment.

Figure 8.4 shows the number of species found per 10 m2 on Setaria pastures under different land use conditions. The already very small number of herb species present under active pasture conditions decreases continuously during later stages of abandonment, while S. sphacelata attains a ground coverage of almost 100 % and reaches a mean height of 1–1.5 m. The vegetation of abandoned Setaria pastures is still very poor in species number even after 18 years of fallow. The most frequent shrub species are then Munnozia senecionides, Austroeupatorium inulifolium, Ageratina dendroides, and Lepidaploa canescens (all Asteraceae). Species of the genus Rubus are also common, while bracken fern is not very abundant (mean coverage of about 3 %). In contrast, silvipastures with remnant forest trees exhibit more species than do the other pasture sites (Fig. 8.4). These pastures were not established by slash-and-burn practices, but by lumbering trees without previous burning.

Fig. 8.4
figure 00084

Number of species per 10 m2 found on Setaria pastures under different land use conditions (active use, abandoned, and silvipasture). The boxes represent the medial 50 % (lower to upper quartile) of each sample, bold vertical lines show the median, and the bars illustrate the minimum and maximum values of the respective samples

Full size image

This lumbering leads to a completely different pasture species composition including woody elements such as Miconia, Hieronyma, and various Lauraceae that do not occur on previously burned sites. The higher species numbers on silvipastures are not due to herbs, but rather to tree seedlings. These sites accordingly provide better initial conditions for a rapid succession than do the common pastures.

The fires set for slash-and-burn practices do not restrict themselves to their target areas. As a consequence, a bracken fern stage characterized by the high abundance of P. arachnoideum establishes itself on burned-out ridges and steep slopes that are not suitable for agricultural use due to their low soil nutrient content. Repeated burning stabilizes the dominance of bracken, since the subterranean rhizomes of this fern tolerate high temperatures. In addition, fire removes the poorly decomposable dead fronds of bracken, and thus removes any restrictions to the propagation of the fern by self-shading. Increasing invasion of bracken fern similarly takes place on active pastures after repeated burning (Roos et al. 2010).

3.5 Fragmentation Effects: Epiphytes on Remnant Trees as a Model System

Forest fragmentation is a common phenomenon in most tropical montane regions. During the last decades, the size of intact forest fragments tended to decline rapidly, dividing populations and creating distinct edge habitats. In the face of global climate change, even narrow bands of agricultural matrix may threaten the biodiversity of forests by hindering compensatory range shifts, such as the upslope migration of forest species on the deforested Andean foothills (Bush 2002). Understanding how forest organisms are affected by habitat fragmentation thus requires information as to how these organisms respond to all of the factors that affect landscapes (Gascon et al. 1999).

Isolated trees (ITs) are a common element of anthropogenic land use matrices throughout the tropics. They constitute keystone structures by offering refuge, enhancing connectivity, and providing nuclei for regeneration (Wolf 2005; Manning et al. 2006; Zahawi and Augspurger 2006). ITs can be viewed as the “ultimate,” smallest possible forest fragment (Gove et al. 2009), and they thus provide an easily replicable model system for fragmentation studies. ITs represent small habitat islands for forest organisms that are likely to experience constrained dispersal. Being immediately surrounded by open land, they experience pronounced physical edge effects (elevated light levels, wind turbulence, temperatures, and enhanced evapotranspiration rates; Laurance 2004).

With well over 1,200 species of lichens, bryophytes, and vascular plants (Liede and Breckle 2007), epiphytes (including epiphylls) constitute the majority of plant species in the RBSF area. Patterns of IT epiphyte diversity have been studied for all of the three principal groups of epiphytes (see above; Nöske et al. 2008). It is noteworthy that lichen species richness found on ITs was not lower than that found on trees of the closed forest. The species composition on the ITs was also only moderately distinct from that on the forest trees.

In contrast, the epiphytic bryophyte and particularly vascular plant species found on ITs were much smaller in number than those found on forest trees and showed strong compositional differences in the two habitats (Werner et al. 2005; Nöske et al. 2008; unpublished data). A common pattern on isolated trees was a spatial homogenization of epiphyte species. Most epiphyte species in closed-canopy forests occupy narrow vertical niches that offer a suitable trade-off between light and humidity conditions. The microclimate gradient driving this stratification is nonexistent on isolated trees (Nöske 2005; Werner and Gradstein 2008), and typical upper canopy species are widely distributed among the crowns of isolated trees. The distance of ITs to the forest as a source of diaspores was not significantly correlated with species richness per isolated tree for any of the three taxonomic groups of epiphytes (Nöske et al. 2008), suggesting that dispersal limitations were of minor importance.

The processes that lead to such pronounced changes in species assemblage structure and diversity have been studied for vascular epiphytes. Well-established individual plants (late juvenile to adult stages) on the trunks of isolated trees in a fresh clear-cut were tagged and observed over a period of 3 years. More than 50 % of the individuals died within the year following the isolation of their host trees. The mean cumulative mortality of the individuals on the ITs was 72 %, which is significantly higher than that observed in undisturbed forest (11 %). While the mortality of nearly all of the major epiphyte families was significantly higher on the IT than on the forest trees, it was most dramatic with regard to ferns. Plants surviving on ITs generally showed a reduced maximum leaf length (Werner 2011). A parallel experimental study on recruitment indicated that rates of early epiphyte establishment on isolated trunks of Piptocoma discolor were 90 % lower than those observed in forest. Even after statistically removing the effect of low plant abundance by rarefaction, the number of epiphyte species that were able to newly establish was significantly lower on IT trunks than it was in the closed forest. Although a broad range of taxa were able to establish in the forest, only few xerotolerant groups (especially tank bromeliads and a desiccation-tolerant polypodioid fern) managed to establish on IT trunks.

These results prove that increased mortality of resident epiphytes as well as impaired recruitment of new epiphytes affect the composition and species richness of vascular epiphyte assemblages on isolated trees. The poor recruitment in terms of both species and individual numbers is hence unlikely to compensate for the strongly enhanced mortality of resident species, and thus results in long-term impoverishment (Fig. 8.5, see also Nöske et al. 2008; Köster et al. 2009). The high rates of mortality and the taxonomically skewed recruitment suggest that increased exposure to light and wind and the resulting increase in desiccation stress are the key drivers of epiphyte diversity on ITs rather than dispersal constraints.

Fig. 8.5
figure 00085

A conceptual model of the extinction and (re-)colonization dynamics of epiphyte species on isolated trees remaining upon the conversion of closed forest into agricultural matrices. Since only a limited number of local species may be adapted well enough to dwell under a high level of exposure, recolonization will often not be able to compensate the loss of the original epiphyte flora

Full size image

Caution should, however, be taken in generalizing these results. The resilience of epiphyte communities may vary widely with local climate conditions (Werner et al. 2012), and can be substantially higher in perhumid or strongly arid areas (e.g., Werner and Gradstein 2009; Larrea and Werner 2010). On the other hand, most shade- or moisture-demanding taxa of any local community are unlikely to flourish in the multiple edge environment of ITs (Poltz and Zotz 2011). This is commonly indicated by the reduced β-diversity of IT assemblages (e.g., Hietz 2005; Larrea and Werner 2010). Plant responses similar to those documented for ITs in the San Francisco Valley can also be expected to occur along forest edges and in degraded forest. Consequently, the retention of scattered trees and narrow strips or small fragments of forest on clear-cuts are unlikely to be sufficient tools for the conservation of tropical epiphyte diversity.

3.6 Expected Effects of Reduced Plant Functional Diversity

Changes in biodiversity are accompanied by changes in plant functional diversity (Diaz et al. 2007). Our knowledge about plant functional traits and the role of individual species in the montane forests of the study area is still very limited. We can therefore only assume that the immense loss of species richness associated with land use change results in a drastic deterioration of ecosystem services on the landscape level. Recent studies (e.g., Isbell et al. 2011) suggest that a high level of plant diversity is needed to maintain ecosystem services, and that even the loss of less common species can have significant effects on the ecosystem level (Lyons et al. 2005). Because of the large species pool existing in the study area, one might expect a high complementarity of functions over the range of species, and that insignificant species losses would have no noticeable short-term effects on ecosystem functioning. However, it is beyond question that the drastic species loss resulting from forest fragmentation and current unsustainable land use practices is accompanied by a decreasing resistance to invasive species (e.g., the example of bracken fern in frequently burned areas) and a diminution of ecosystem services.

The average annual total value of ecosystem services provided by tropical forests is estimated to amount to 6,120 $ ha−1 (TEEB 2009). According to Göttlicher et al. (2009), almost 48 % of the forests in the study area situated below 2,200 m a.s.l. (lower montane forest, originally amounting to a total of ~7,480 ha) have already been anthropologically transformed. Of this area, 15.4 % is being actively used as pasture, 10.6 % is covered by bracken fern and 21.7 % is overgrown by shrub-dominated succession. Assuming that the average value of the ecosystem services generated by this transformed area is reduced by approximately 50 % (according to Portela and Rademacher 2001), one may estimate an annual loss of 10.9 million $ for that relatively small area. This calculation is only of theoretical value since there is no market for ecosystem services. However, it is hard to imagine that current agricultural incomes (see Chap. 17) will be able to compensate for this loss.

4 Conclusion: Implications for Conservation

Managing the montane ecosystems of South Ecuador in a sustainable fashion is a key challenge for the future. Basic knowledge about ecosystem functioning is still scarce for the region, and the study of plant functional traits should receive increasing attention, since these traits are the main attribute by which plants influence ecosystem functioning and thus ecosystem services.

The extent to which the locations that are most valuable for ecosystem services coincide with those that support the highest biodiversity is of critical importance for designing conservation and management strategies. In the study region, high biodiversity commonly coincides with relatively poor soils that are not appropriate for agriculture. This means that the chances for conserving the ecosystems of interest in the study area are good, since the benefits of converting areas of high biodiversity to agricultural use are low. On the other hand, conservation can only be achieved by increasing the value of forest ecosystems to the local communities. It is therefore necessary to define property rights and to establish suitable “Payment for ecosystem services” (PES) compensation schemes.