- Top of page
- Literature Cited
- Supporting Information
In this study, we aimed to assess the processes controlling compositional change in a Northern Andean páramo highly affected by human-induced disturbances over the last few decades (La Rusia, Colombia). Along the 3000–3800 m asl altitudinal range, we randomly sampled fifty 10 × 10 m plots. Therein, we measured altitude and variables related to soil conditions (i.e., moisture, nutrient contents, bulk density, and texture), occurrence of human-induced disturbances (i.e., fire, vegetation clearing, potato cultivation, and cattle grazing), and land-use history. We also recorded richness and abundance of plant species, identifying them as exotic or native. We differentiated four groups of plots according to their species composition. The groups had significant differences in altitude, soil conditions, land-use history, and particularly, in richness of exotic species and exotic/native cover ratio. They could be ascribed to shrub- and grass-páramo vegetation types based on their relative dominance of woody and herbaceous species; however, these groups were not arranged according to the hypothetical composition of altitudinal belts, but rather formed a mosaic of patches. This mosaic was determined not only by altitude but also by soil conditions and disturbance history of sites. Our results corroborate recent findings which highlight shrub- and grass-páramo vegetation types as patches of contrasting species composition and structure that depend on local environmental variables, as well as human-induced disturbances as a major determinant of compositional discontinuities in these ‘high mountain’ tropical ecosystems.
Este trabajo fue realizado con el objetivo de evaluar los procesos que controlan los cambios composicionales en un páramo fuertemente afectado por las perturbaciones de origen humano durante las últimas décadas (La Rusia, Colombia). A lo largo del gradiente altitudinal entre los 3000 y los 3800 m s.n.m., se muestrearon de forma aleatoria 50 parcelas de 10 × 10 m. En estas se midió la altitud y las variables ambientales relacionadas con la humedad, el contenido en nutrientes, la densidad aparente y la textura del suelo, la ocurrencia de perturbaciones humanas como el fuego, la tala, el cultivo de papa y el pastoreo, y la historia de uso humano. También se registró el número y la abundancia de especies vegetales, determinando el origen exótico o nativo de las mismas. Se diferenciaron cuatro grupos de parcelas en función de su composición específica. Estos mostraron diferencias significativas en altitud, condiciones del suelo, historia de uso humano, y especialmente, en riqueza de especies exóticas y cociente especies exóticas/nativas. Dichos grupos podían estar relacionados con los tipos de vegetación asociados habitualmente al sub-páramo y al páramo en función de la dominancia relativa de leñosas y herbáceas en los mismos, sin embargo, no parecían estar distribuidos según el esquema hipotético de los cinturones vegetales, sino formando un mosaico de parches. Dicho mosaico estaba determinado no sólo por la altitud, sino también por las condiciones del suelo y la historia de uso humano de las diferentes zonas del páramo. Nuestros resultados corroboran los de estudios recientes que sugieren que los tipos de vegetación asociados al sub-páramo y al páramo normalmente no aparecen formando cinturones, sino mosaicos de parches en función de variables ambientales locales, y también los de aquellos que señalan a las perturbaciones humanas como el principal determinante de las discontinuidades composicionales en los ecosistemas de alta montaña tropical.
In tropical mountains, plant communities usually change along the altitudinal gradient, forming vegetation belts characterized by different species compositions and abundances (Gentry 1995, Keating 1999). These belts are primarily determined by changes in temperature and precipitation regimes (Troll 1968). In tropical climate regions, temperature varies little through the year, but it does decline with altitude resulting in pronounced temperature discontinuities at different altitudes (Janzen 1967, Ohsawa 1995). Above 3000 m asl, extremely low temperatures and frost cause large changes in species composition from cloud or upper montane forests to high tropical mountain plant communities dominated by shrubs and grasses, known as páramos and punas (Northern and Southern Andes, respectively), or as afro-alpine (East Africa) and tropical-alpine (East Asia) vegetation belts (Ashton 2003). The páramos are characterized by their high species richness and degree of endemicity, and large regional variations in species composition (Sturm & Rangel 1985, Luteyn et al. 1992). These ecosystems account for a wide array of growth forms that are highly adapted to the low seasonal and strong daily oscillations of temperatures they experience and which include tussock grasses, giant caulescent rosettes, sclerophyllous shrubs, and mat-forming and cushion plants (Körner 2003). Depending on the relative dominance of these growth forms at different altitudinal ranges, páramos are typically divided into three vegetation belts (Cuatrecasas 1968): the shrub-páramo (sub-páramo), dominated by shrubs (3200–3500 m asl); the grass-páramo (páramo), dominated by tussock grasses (3500–4100 m asl); and the super-páramo, dominated by mat-forming and cushion plants, and found immediately below the permanent snow line (>4100 m asl). The altitudinal range occupied by vegetation belts and the more or less discontinuous character of the boundaries between them are strongly determined by changes in precipitation regimes between the different sections and aspects of a specific mountain range (Körner 2003, 2007). This applies particularly to the páramos of the Northern Andes. In the dryer western aspects of these mountains, frost starts at lower altitudes and the snow line is at higher altitudes than in the wetter eastern aspects. This determines the formation of a broader páramo vegetation belt which shows a less abrupt boundary with the upper montane forests located immediately below, in the dryer western aspects (Sarmiento 1986).
In high tropical mountains, the boundaries of the vegetation belts are not inflexible, but rather tend to intermingle with each other, forming vegetation mosaics. These mosaics are usually determined by environmental variables related to cloudiness (Körner 2003, Sklenar et al. 2008), topography (Young & Keating 2001), presence of rocky outcrops, and soil conditions (Arellano & Rangel 2008). Soil conditions vary markedly within the páramos, even at local scales, and are particularly important. These include, among others, soil texture, which condition the relative dominance of shrubs and grasses (Keating 1999), and soil moisture, which exerts tight control over seedling establishment and survival (Wesche 2003), and determine total percent of vegetation cover (Young & Peacock 1992). Variables associated with soil fertility, such as soil organic matter, pH, cation exchange capacity, and amounts of exchangeable nutrients, are also closely related to changes in species composition (Fariñas & Monasterio 1980). Vegetation mosaics can be also caused by disturbances such as fire (Wesche et al. 2000, Hemp 2006, Martin et al. 2007) and grazing (Molinillo & Monasterio 1997, Suárez & Medina 2001). Despite the increasing evidence that natural fires played an important role in the evolution of páramo vegetation (Di Pascuale et al. 2008), these ecosystems were not extensively burned, cultivated, or grazed until the arrival of the first European settlers (Eckholm 1975, Sarmiento 1986). During the last few decades, most páramos have been exposed to high levels of grazing (Molinillo & Monasterio 1997) and to processes of agricultural expansion and intensification (Etter & Villa 2000, Suárez & Medina 2001) and have therefore become severely degraded (Sarmiento 2000, Sarmiento et al. 2003). In particular, the widespread use of fire for opening new grazing areas has resulted in a general decrease in species richness (Hofstede et al. 1995), and changes in species composition, such as the disappearance of most shrub species (Williamson et al. 1986, Keating 2000), and an increase in the number and abundance of exotic species (Lozano et al. 2009, Olivera & Cleef 2009). These processes have been particularly intense at the lowest altitudinal limits of the páramos, and in the boundary between the shrub- and grass-páramo vegetation belts (Verweij et al. 1995, Etter & Villa 2000). Indeed, the patchy appearance of the shrub-páramo might be related to the existence of areas in which soil conditions vary or where human-induced disturbances have occurred with different frequencies and intensities (Sarmiento et al. 2003).
Given the extent of degradation processes that páramos have recently experienced and their role as suppliers of ecosystem services, such as climate and hydrologic regulation and soil conservation (Etter 1998), it is of crucial importance to advance our knowledge of processes determining vegetation changes at the altitudinal ranges in which these ecosystems are being radically transformed. In this study, we aim to assess the processes controlling changes in species richness and composition at the lowest altitudinal limits of a Northern Andean páramo (La Rusia, Colombia), highly affected by human-induced disturbances over the last few decades. Specifically, we addressed the following questions: (1) Can we characterize different plant communities throughout the study area according to their species composition? (2) Can we classify these plant communities into the scheme of shrub- and grass-páramo vegetation types according to their relative dominance of woody and herbaceous species? (3) To what extent are compositional differences among these plant communities determined by altitude, soil conditions, occurrence of human-induced disturbances or by land-use history? (4) How does species richness and abundance of exotic species vary among these plant communities?
- Top of page
- Literature Cited
- Supporting Information
We identified a total number of 304 plant species belonging to 82 families. The cluster analysis differentiated four groups of plots at Relativized Manhattan distance = 10.96, according to their species composition: G1, G2, G3, and G4, comprising 13, 6, 10, and 21 plots, respectively (Fig. S2). The INDVAL analysis identified 29 significant indicator species (P < 0.05), most of them with high indicator values (Table S1). The indicator species were different in each of the four groups considered. The most meaningful components of the PCA ordination were PC1 and PC2 (Table 1). Principal component 1 explained 42.9 percent of variance and represented a gradient of variation between plots with high values of soil moisture, effective cation exchange capacity and cations such as Ca2+, K+, and Mg2+ (negative extreme) and plots with high values of bulk density, density of solids and clay (positive extreme). This indicates a gradient ranging from plots with humid and fertile soils to plots with clayey and compacted soils. Principal component 2 explained 19.9 percent of variance and described a gradient ranging from plots with a high Al3+ concentration in its negative extreme and plots with a higher pH (less acidic soils) in its positive extreme. The four groups of plots differentiated by the cluster analysis differed significantly in their woody/herbaceous cover ratio (KW-H3,50 = 12.49, P < 0.05). Pair-wise Mann–Whitney U comparisons revealed that the plots of group G4 had a significantly higher woody/herbaceous cover ratio than plots of groups G1 and G2 (Fig. 1).
Figure 1. Differences in woody/herbaceous cover ratio among the four groups of plots (G1, G2, G3, and G4). Different letters indicate significant differences (Pair-wise Mann–Whitney U post hoc comparisons).
Download figure to PowerPoint
Table 1. Results of the principal components analysis performed with soil variables; proportion of clay (Clay), soil moisture, percent of organic matter, effective cation exchange capacity (CEC), contents in cations (Ca2+, K+ and Mg2+, Na+) and Aluminum (Al3+), pH, soil bulk density, and density of solids. Only the variables highly correlated with the components (factor loading >0.5) are shown
|Variable||Factor loading||Variable||Factor loading|
|Component 1 (42.9% Variance explained)|
|Soil moisture||−0.89||Bulk density||0.89|
|CEC||−0.77||Density of solids||0.77|
|K+||−0.67|| || |
|Mg2+||−0.6|| || |
|Component 2 (19.9% Variance explained)|
The NMDS ordination for species occurring in over 10 percent of the plots yielded a minimum Kruskal's stress of 0.22 (P < 0.001, Monte Carlo test of 50 permutations). The ordination was determined primarily by altitude (r2 = 0.612, P < 0.001) and secondly by PC1 (r2 = 0.296, P < 0.001), potato cultivation (r2 = 0.278, P < 0.001), and age of abandonment > 15 yr (r2 = 0.257, P < 0.05) (Fig. 2). Groups of plots differed significantly in altitude (F3, 46 = 8.3, P < 0.001). The Tukey HSD post hoc test revealed that the plots from group G3 were significantly located at a lower mean altitude (3356 m asl) than the plots from G1 and G4 (3565 and 3588 m asl, respectively). Group G2 occupied an altitudinal range wider than G1 and G4 and was located between these groups and G3. There was also a ‘coincidence strip’ (3435–3555 m asl) at which plots of all groups appeared (Fig. 3A). Groups of plots also differed significantly in soil conditions (F3,46 = 6.9, P < 0.001). Group G4 presented lower values on the first PCA axis than any other group, which indicates that it had the wettest and most fertile soils. On the contrary, the Tukey HSD post hoc test showed that G1, G2, and G3 did not differ significantly in their values on the first PCA axis (Fig. 3B). No significant differences were found in the ANOVA performed with the second PCA axis as the continuous variable (results not shown). Groups of plots did not differ significantly in occurrence of human-induced disturbances (pseudo-F3,46 = 1.11, permutation-P = 0.18), but they did so in land-use history (pseudo-F3,46 = 2.44, permutation-P < 0.05). They also differed significantly in the occurrence of potato cultivation and in age of abandonment >15 yr (Table 2). DISTLM v.5 (University of Auckland, Auckland, New Zealand) does not enable pair-wise a-posteriori comparisons between groups to be performed, but examination of the proportion of plots affected by human-induced disturbances and with ages of abandonment higher and lower than 15 yr revealed that all groups presented a high percentage of plots affected by human-induced disturbances. Groups G2 and G3 had a larger proportion of cultivated plots than G1 and G4, and G1 had a larger proportion of burned plots than the other groups. Groups G2 and G3 have a larger proportion of plots with age of abandonment higher than 15 yr, than G1 and G4 (Table 2).
Figure 2. Patterns of compositional change across the study area and their relationships with environmental variables; non-metric multidimensional scaling ordination (NMDS) of species composition in 50 plots for species occurring in over 10 percent of them, showing the ordination dimensions 1 and 2, their interpretation with respect to environmental variables. The smooth surface generated for altitude is also shown. Groups of plots differentiated by the Hierarchical Bray–Curtis Analysis (i.e., G1, G2, G3, and G4) are indicated with different symbols. Species are indicated with gray X-shaped crosses. Color version is available in the online version of the article.
Download figure to PowerPoint
Figure 3. Differences in (A) altitude (m asl) and (B) soil variables (PCA Component 1) among the four groups of plots (G1, G2, G3, and G4). Different letters indicate significant differences (Tukey HSD post hoc test). The 3500 m asl border between the shrub- and grass-páramo vegetation belts and soil conditions associated with positive and negative sides of PCA Component 1 are indicated.
Download figure to PowerPoint
Table 2. Proportion of plots affected by the different human-induced disturbances and with age of abandonment higher and lower than 15 yr in the groups of plots differentiated by the hierarchical cluster analysis (G1, G2, G3, and G4). Pseudo-F statistics and permutation-P values obtained in the distance-based analyses of variance performed for each one of these variables are also indicated
|Group||Human-induced disturbances||Age of abandonment|
|Fire||Clearing||Cultivation||Grazing||>15 yr||<15 yr|
We identified 11 exotic species in the plots (Table S2), all of which were herbaceous and most of which were perennial. They came from Europe, North Africa, and Western Asia. Within their respective geographical areas of origin, most of these species (i.e., Anthoxanthum odoratum, Holcus lanatus, Hypochaeris radicata, Trifolium repens, Taraxacum officinale, and Rumex acetosella) showed a clear preference for meadows and wet pastures. Only in six (of 50) plots no exotic species were found. Groups of plots differed significantly in richness of exotic species (KW-H3,50 = 23.5, P < 0.001; Fig. 4A). G2 and G3 presented the highest richness of exotic species (5.5 and 4.5 median values, respectively), whereas G1 and G4 presented the lowest richness of exotic species (1 and 3 median values, respectively). With regard to exotic/native cover ratio, we also found significant differences between groups of plots (KW-H3,50 = 29.2, P < 0.001; Fig. 4B). Groups G3 and G2 presented the highest and the second highest exotic/native cover ratio, respectively. Remarkably, two of the exotic species presented in group G2 (Trifolium repens and Taraxacum officinale) and two of the exotic species presented in group G3 (Anthoxanthum odoratum and Hypochaeris radicata) were identified as Indicator Species in the INDVAL analysis, most of them with high indicator values (Table S1).
Figure 4. Differences in (A) richness of exotic species and (B) exotic/native cover ratio among the four groups of plots (G1, G2, G3, and G4). Different letters indicate significant difference (Pair-wise Mann–Whitney U post hoc comparisons).
Download figure to PowerPoint
- Top of page
- Literature Cited
- Supporting Information
Changes in species composition throughout the Páramo de la Rusia were primarily determined by altitude (Fig. 2) and despite the existence of a ‘coincidence strip’ in which plots of all groups appeared (3435–3555 m asl), groups were clearly arranged in an altitudinal pattern (i.e., G4, G1, G2, and G3 from higher to lower mean altitudes) (Fig. 3A). Significant differences in the relative dominance of woody versus herbaceous plants among the groups of plots identified in the cluster analysis, however, did not match the scheme of shrub- and grass-páramo vegetation belts. For instance, group G4 presented the highest mean woody/herbaceous cover ratio, but the majority of its plots were located within the grass-páramo belt, whereas G3, with a lower woody/herbaceous cover ratio, exhibited a large number of plots located within the shrub-páramo belt. This prevents us from considering the existence of strict vegetation belts in our study area and suggests that other factors apart from those associated with altitude (i.e., changes in temperature and precipitation regime) exert a crucial control over compositional changes throughout this Andean páramo.
Changes in species composition among groups of plots identified in the cluster analysis were also strongly determined by soil conditions (Fig. 2). In particular, in spite of the fact that groups G1 and G4 both covered the altitudinal range of the grass-páramo vegetation belt (Fig. 3A), they showed highly significant differences in their soil conditions (Figs. S1D and 3B). Indeed, G4 exhibited the highest values of soil moisture and effective cation exchange capacity across the study area. Within well-preserved and relatively undisturbed páramos, soils in erosive areas (rocky outcrops and steep slopes) show compacted and clay-textured soils, whereas those in sedimentary areas (valley bottoms) tend to present high soil moisture and nutrient contents (Abadín et al. 2003, Sarmiento et al. 2003). These differences in soil conditions may strongly affect vegetation structure, promoting a high level of dominance of tussock grasses on the rocky outcrops and steep slopes, and a high dominance of shrubs in the valley bottoms (Arellano & Rangel 2008). Our results corroborate these observations. Among the groups of plots differentiated at the Páramo de La Rusia, group G1 was located in the most erosive zone, had the second lowest woody/herbaceous cover ratio, and nine of its ten indicator species were herbaceous. In contrast, G4 was located in the most sedimentary zone, showed the highest wood/herbaceous cover ratio, and two of its five indicator species were woody (Figs. S1D and 1; Table S1).
Compositional changes across the Páramo de La Rusia were also strongly influenced by land-use history (Fig. 2). Land-use history varied significantly among groups of plots (Table 2). Groups G3 and G2 were affected by human-induced disturbances until more recently than G4 and G1. The G3 plots appeared in the lowest altitudinal range of the study area (3080–3560 m asl), corresponding mostly to the shrub-páramo vegetation belt (Fig. 3A). Although some plots of this sector showed high woody/herbaceous cover ratios, most were dominated by herbaceous species (Fig. 1). Grasslands with low representation of woody species at relatively low altitudes could constitute a type of secondary vegetation, originated at the early stages of post-disturbance succession in the páramos (Suárez & Medina 2001). Group G3 did not show significant differences in the first PCA axis with G1 and G2 (Fig. 3B), yet comprised the plots with the lowest soil moisture, the lowest effective cation exchange capacity, and the lowest nutrient concentrations within the study area (results not shown). The lowermost areas of the North Andean páramos, such as those occupied by G3, are more accessible and therefore more disturbed than the uppermost ones (Bader & Ruitgen 2008). Human-induced disturbances are quicker to disrupt soil physical properties than a decline in nutrient status (Hofstede et al. 1995), and low levels of soil fertility therefore usually indicate intense and long-lasting degradation processes (Jaimes & Sarmiento 2002, Montilla et al. 2002, Abadín et al. 2003).
Our results, together with all these facts, indicate that the vegetation structure and soil conditions of G3 plots could be related to their long-lasting land-use history. This possibility is also supported by the fact that G3 presented the highest exotic/native cover ratio (Fig. 4B). Presence of exotic species is commonly associated with human activity (Seabloom et al. 2003). The invasion process of any area depends on conditions of enrichment of limited resources and the availability of invading propagules (Davis et al. 2000). Human-induced disturbances such as cultivation and grazing usually involve the addition of otherwise limited resources, which contributes to the invasion success of alien grass species (Burke & Grime 1996). The lowest areas of the Northern Andean páramos, in which G3 is located, are closer to the upper montane forests (in some cases, in their potential area of occurrence) and to the human settlements downslope. Consequently, they have been more affected by human-induced disturbances than the uppermost ones and usually present the greatest abundance of exotic species (Mora-Osejo & Sturm 1994, Ramsay & Oxley 1996). Furthermore, these areas show environmental conditions equivalent to those found in other ecosystems around the world (e.g., European Atlantic forests and meadows) and are consequently more suitable for colonization by exotic species (Keating 2000, Márquez et al. 2004). The two species with the highest I.V. in this G3, Anthoxanthum odoratum and Hypochaeris radicata, were exotics (Table S1). In their respective areas of origin, these species are commonly found in natural meadows and wet pastures (Table S2). In the Colombian Andes, they appear within the upper montane forest vegetation belt and in artificially created meadows and pastures, where they have become widespread due to farming activities such as grazing, reaping, and fertilization (Verweij et al. 1995).
In contrast with G3, G2 was located between the shrub- and grass-páramo vegetation belts (Fig. 3A), and showed the second highest exotic/native cover ratio (Fig. 4B). Plots of G2 had, however, the highest richness of exotic species (Fig. 4A). These facts might be related to the shorter history of anthropogenic activity in the area in which G2 is located, in comparison with G3, which had promoted a more recent invasion of exotic species in the first group. Occurrence of disturbances in a local community usually gives rise to the rapid spread of alien species, which results in increased diversity (Sax & Gaines 2003). Over the following years, however, community assembly proceeds, and diversity steadily decreases by means of processes of competitive exclusion (Davis et al. 2005). Our results point toward the existence of disturbance-intensity and land-use history gradients from G3 to G2. An endemic species, Lachemilla orbiculata, associated with low and mean grazing intensities in Northern Andean páramos (Cárdenas-Arévalo & Vargas-Ríos 2008) presented the highest I.V. for G2, which supports the hypothesis that invasion of exotic species in G2 has been more recent that in G3 due to the shorter history of anthropogenic activity in the area in which G2 is located, in comparison with G3.
With regard to the G1 and G4 plots, most of them had remained unaffected by human-induced disturbances longer than plots G3 and G2 (Table 2). In these groups, both exotic/native cover ratio and richness of exotic species were low (Figs. 4A and B). Their indicator species were native herbaceous plants (i.e., Paepalanthus columbiensis and Halenia asclepiadea) and shrubs (i.e., Pentacalia vaccinioides and Arcytophyllum nitidum) (Table S1) typical of well-preserved páramos (Morales et al. 2007). These results coincide with observations made in Venezuelan páramos, in which 12 yr after the occurrence of the last human-induced disturbance event, richness of exotic species starts to decrease and over 90 percent of native species richness can be recovered (Jaimes & Sarmiento 2002, Sarmiento et al. 2003). Moreover, absence of burning and grazing enhances the natural development of a dense shrub layer (Suárez & Medina 2001). This fact suggests that the high woody/herbaceous cover ratio presented by G4 (Fig. 1) might be related not only to its wet and fertile soils but also to the fact that this group presented the highest percentage of plots unaffected by disturbances for over 15 yr (Table 2).
Interestingly, the groups did not show significant differences in the occurrence of human-induced disturbances in general terms. This might be related to the fact that the majority of plots have been affected by a sequence of disturbances (i.e., fire and/or clearing, potato cultivation, and cattle grazing), which makes it difficult to find truly unaffected areas. The importance of potato cultivation in determining compositional changes among plots (Fig. 2), and the significant differences among groups in relation to this variable (Table 2), might be related to the higher percentage of cultivated plots in groups G2 and G3 compared with G1 and G4 (Table 2). Likewise, the high percentage of plots affected by fire in G1 (Table 2) could explain the high I.V. (Table S1) shown by the tussock grass Calamagrostis effusa in this group. This species is very common in the grass-páramos of the Eastern Andean range, recovers rapidly, and becomes clearly dominant during early succession after fire (Vargas-Ríos 1997).
In short, the four groups of plots differentiated in the Páramo de La Rusia correspond to plant communities that differ significantly in species composition and could be ascribed to the shrub- and grass-páramo vegetation types. These plant communities form a mosaic of patches across the study area, whose arrangement is controlled not only by altitude but also by local factors such as soil conditions and the disturbance history of sites. Similar findings have been addressed in Ecuadorian páramos (Olivera & Cleef 2009), the ‘ericaceous belt’ of the Eastern African ranges (Wesche et al. 2000, Wesche 2003), and the ‘campos de altitude’ in Southeast Brazil (Safford 2001). Our results point at the crucial role played by human-induced disturbances in determining compositional discontinuities between vegetation types in tropical mountains (Martin et al. 2007, Cierjacks et al. 2008). They also provide strong support for the alternate view of Coblentz and Keating (2008), who argue that shrub-páramo rarely occurs as a transition community between the upper montane forests and the grass-páramo, but as patches of contrasting species composition and structure which appear in certain areas depending on differences in topography, soil conditions, and land-use history. Our results clearly indicate that within the Páramo de La Rusia, the abundance of exotic species is greater in zones that were altered earlier and, to a greater extent, by human-induced disturbances. Richness of exotic species, however, is higher in more recently and less altered zones.