Summary of the plant invasions in the Northern Taiwan
A total of 475 native plants and 139 naturalized plant species were recorded. Richness of native species was higher than that of naturalized species before and after standardization. Average native species numbers per plot before and after standardization were 31.34
± 13.79 (mean ± SD) and 20.54 ± 8.76 (mean ± SD) respectively. On the other hand, average naturalized species numbers per plot before and after standardization were 13.56 ± 6.09 (mean ± SD) and 9.25 ± 3.47 (mean ± SD) accordingly. The mean cover of native plants is 51.26%, while that of naturalized plants is 48.74%.
According to the importance value index (IVI), the most dominant species were Bidens pilosa var. radiata Sch. Bip., Paspalum conjugatum Bergius, Ageratum houstonianum Mill., Ageratum conyzoides L., and Ipomoea cairica (L.) Sweet. Most
habitat types accommodated a single most dominant species: Bidens pilosa var. radiata Sch. Bip. (Table 4). The species composition of
forest was different from that of other habitat types. There were five naturalized species have importance value index higher than 10. The most dominant naturalized plant in forest is Setaria palmifolia (J.
Konig) Stapf., followed by Ageratum houstonianum Mill., Bidens pilosa var. radiata Sch. Bip., Elephantopus mollis Kunth, and Impatiens walleriana Hook. f..
TABLE 4. Dominant naturalized plants of different habitat types.
Rank Cemetery IVI Crop field IVI
1 Bidens pilosa 77.1 Bidens pilosa 48.8
2 Paspalum conjugatum 6.8 Ageratum conyzoides 15.5 3 Ipomoea cairica 6.8 Paspalum conjugatum 13.6 4 Wedelia trilobata 6.2 Ageratum houstonianum 10.4 5 Spermacoce latifolia 5.5 Spermacoce latifolia 9.0 6 Conyza sumatrensis 4.7 Alternanthera sessilis 8.2 7 Chamaesyce hirta 4.4 Alternanthera philoxeroides 7.0
Rank Roadside IVI Forest IVI
1 Bidens pilosa 65.9 Setaria palmifolia 35.3
2 Ageratum houstonianum 8.7 Ageratum houstonianum 25.2
3 Paspalum conjugatum 8.2 Bidens pilosa 15.8
4 Ageratum conyzoides 7.0 Elephantopus mollis 13.2 5 Ipomoea cairica 5.7 Impatiens walleriana 11.7
6 Setaria palmifolia 5.5 Lantana camara 9.9
7 Panicum maximum 5.5 Ageratum conyzoides 9.4
Rank Abandoned field IVI Riparian IVI
1 Bidens pilosa 52.0 Bidens pilosa 70.0
2 Paspalum conjugatum 14.3 Pennisetum purpureum 11.6 3 Ipomoea cairica 8.5 Paspalum conjugatum 10.3 4 Ageratum houstonianum 6.8 Wedelia trilobata 9.1 5 Paspalum dilatatum 6.2 Conyza sumatrensis 7.6
6 Ageratum conyzoides 4.9 Rumex crispus 7.5
7 Conyza sumatrensis 4.9 Paspalum dilatatum 6.8
Biodiversity of habitat types
Species numbers of native species were higher than those of naturalized species in all habitat types; however, different patterns of plant invasions were found among habitat types (Table 5). Roadside habitats have the largest number of naturalized plants, followed by abandoned field (79), crop field (75), cemetery (62), riparian (43), and forest (28). However, cemetery was the most infested habitat according to the percentage of naturalized species to all plants (45%), followed by abandoned field (38%), crop field (34%), riparian (30%), roadside (25%), and forest (10%).
TABLE 5. Numerical summary of six different habitats. Frequency: percentage of plots contain naturalized species (%), Cover: average
percentage of naturalized cover (%), Richness: naturalized species number/total species number (%).
Habitat types
Number of plots
sampled
Native
richness
Naturalized
richness
Frequency (%) Cover (%) Richness (%)
Roadside 936 322 110 94 54 25
Crop field 441 147 75 97 58 34 Abandoned field 332 127 79 98 51 38
Forest 223 253 28 30 11 10
Riparian 190 98 43 90 47 30
Cemetery 120 77 62 99 74 45
Forests harbored significantly more native species, while riparian, abandoned field and cemetery had similar richness of native species and naturalized species (Fig. 6). As for cover, forest has significantly higher native species cover, while crop fields have generally higher naturalized species cover than native species (Fig. 7).
The native—exotic richness relationships were different at fine and broad scales. At fine scale, crop field, abandoned field, and cemetery have significantly positive native—exotic relationships (Table 6). At broad scale, roadside, crop field, abandoned field and riparian have significantly positive native—exotic relationships (Table 7).
FIG. 6. The relationships between native richness and naturalized richness in
1km2 scale.
FIG. 7. The relationships between the cover of native and naturalized species
in 1km2 scale.
TABLE 6. Correlation coefficients of native—exotic richness relationships in different habitat types at fine scale (1m2). N: sample size.
Habitat Correlation coefficients P value N
Roadside 0.05 0.17 936
Crop field 0.12* 0.01 441
Abandoned field 0.14* 0.01 332
Forest 0.02 0.73 223
Riparian 0.11 0.11 190
Cemetery 0.19* 0.04 120
*Correlation is significant at the 0.05 level (2-tailed).
TABLE 7. Correlation coefficients of native—exotic richness relationships in different habitat types at broad scale (1km2). N: sample size.
Land use types Correlation coefficients P value N
Roadside 0.32* 0.01 94
Crop field 0.38* 0.01 45
Abandoned field 0.63* 0.01 34
Forest 0.19 0.34 28
Riparian 0.72* 0.01 19
Cemetery 0.1 0.76 12
*Correlation is significant at the 0.05 level (2-tailed).
Grouping
According to factor analysis, three groups of anthropogenic factors can be categorized (Table 8), and they were named landscape heterogeneity, exploitation intensity, and agriculture groups.
Landscape heterogeneity group contained factors, such as land use richness, land use SHDI, edge density, and patch density (38.10% of total variance). Exploitation intensity group, which included built-up area percentage, naturalness, and road length factor, explained
30.65% of total variance. Agriculture group explained 13.86% of total variance, and it mainly comprised of crop field percentage, also edge density and patch density.
TABLE 8. Groups of factor analysis on anthropogenic activities.
Variables
Landscape
loading
Exploitation
loading
Agriculture
loading Communalities
Land use richness 0.88 -0.17 -0.20 0.84
Land use SHDI 0.87 0.21 0.10 0.81
Edge density 0.71 -0.05 0.61 0.88
Patch density 0.71 -0.10 0.54 0.81
Built-up area percentage -0.21 0.90 -0.21 0.89
Crop field percentage -0.02 0.20 0.89 0.84
Naturalness -0.07 -0.87 -0.18 0.80
Road length 0.07 0.83 0.21 0.73
Eigen value 3.05 2.45 1.11
Total variance explained 38.10% 30.65% 13.86%
Cumulative variance explained 38.10% 68.75% 82.61%
Environmental factors were classified into two groups (Table 9):
elevation-temperature and precipitation. Elevation-temperature group contained both of elevation and temperature variables and explained 57.28% of total variance. Precipitation group were mainly correlated with annual average rainfall and rainfall standard deviation, and explained 26.18% total variance.
TABLE 9. Groups of factor analysis on environmental factors.
Variables
Elevation-temperature
group loading
Precipitation
group loading Communalities
Annual average temperature -0.91 -0.05 0.83
Temperature standard deviation 0.91 0.13 0.84
Average elevation 0.92 0.00 0.85
Elevation standard deviation 0.91 0.12 0.84
Annual average rainfall 0.11 0.91 0.83
Rainfall standard deviation 0.04 0.90 0.81
Eigenvalue 3.44 1.57 5.01
Total variance explained 57.28% 26.18%
Cumulative variance explained 57.28% 83.46%
The effects of factors
Biodiversity and dominance of naturalized species were positively correlated with anthropogenic factors and negatively correlated with environmental factors (Table 10). On the contrary, biodiversity and dominance of native species were negatively
correlated with anthropogenic factors and positively correlated with environmental factors. Among anthropogenic activity factors, only exploitation intensity was correlated with native diversity. The higher correlation coefficients were also appeared when examined
relationships between exploitation intensity and native richness, or between elevation and native richness.
TABLE 10. Correlation analysis of factors and biodiversity indices of native and
naturalized plants. Landscape: landscape heterogeneity factor, Exploitation: exploitation
intensity factor, Elevation: elevation-temperature factor. Species %: species percentage,
frequency %: frequency percentage, cover %: cover percentage. Values demonstrate
Spearman correlation coefficients.
Landscape Exploitation Agriculture Elevation Precipitation
Exotic richness 0.29* 0.26* 0.26* -0.31* -0.16
Native richness -0.13 -0.48* -0.01 0.51* 0.33*
Exotic SHDI 0.28* 0.23* 0.22* -0.20* -0.21*
Native SHDI 0.05 -0.24* 0.09 0.26* 0.21*
Exotic species % 0.25* 0.53* 0.15 -0.60* -0.36*
Native species % -0.25* -0.53* -0.15 0.60* 0.36*
Exotic frequency % 0.30* 0.23* 0.12 -0.34* -0.32*
Native frequency % -0.17 -0.22* -0.17 0.29* 0.33*
Exotic cover % 0.25* 0.28* 0.19 -0.35* -0.29*
Native cover % -0.25* -0.28* -0.19 0.35* 0.29*
*Correlation is significant at the 0.05 level (2-tailed).
Naturalized species richness slightly increased with increasing landscape heterogeneity (Fig. 8), exploitation intensity (Fig. 9), and agriculture factor (Fig. 10). Environmental variables, elevation and annual rainfall, had negative correlation with naturalized species richness (Fig. 11 & 12).
FIG. 8. Scatter plot of landscape heterogeneity and naturalized species
richness. Correlation coefficient = 0.29, p < 0.05.
FIG. 9. Scatter plot of exploitation intensity and naturalized species richness. Correlation coefficient = 0.26, p < 0.05.
FIG. 10. Scatter plot of agriculture factor and naturalized species richness. Correlation coefficient = 0.26, p < 0.05.
FIG. 11. Scatter plot of elevation andnaturalized species richness. Correlation coefficient = -0.24, p < 0.05.
FIG. 12. Scatter plot of annual rainfall and naturalized species richness. Correlation coefficient = -0.21, p < 0.05.
Significant correlation was found between groups of factors (Table 11). Elevation-temperature and exploitation intensity were highly and negatively correlated. Precipitation and exploitation intensity were slightly and negatively correlated.
TABLE 11. Coefficient between groups of factors. Landscape: landscape heterogeneity factor, Exploitation: exploitation intensity factor, Elevation:
elevation-temperature factor. Values demonstrate Spearman correlation coefficients.
Landscape Exploitation Agriculture Elevation Precipitation
Landscape 1 0.16 0.05 -0.09 -0.13
Exploitation 1 0.09 -0.73* -0.22*
Agriculture 1 -0.11 -0.11
Elevation 1 0.12
Precipitation 1
*Correlation is significant at the 0.05 level (2-tailed).
Native—exotic richness relationships
Before cluster analysis, native species richness and naturalized species richness was not correlated (Table 12).
TABLE 12. Native—exotic richness relationships before cluster analysis.
Native indices Correlation coefficient of corresponding naturalized indices
Richness 1m -0.03
Richness 1km -0.01
*Correlation is significant at the 0.05 level (2-tailed).
According to cluster analysis, the plots were separated into four groups (Fig. 13), and these groups were characterized with different anthropogenic factors. These four groups included high exploitation group, high heterogeneity group, low anthropogenic activity group, and high agriculture group. They were highly associated with high exploitation intensity, high landscape heterogeneity, low
anthropogenic activities, and high agriculture activities respectively.
Native—exotic relationships moved from negative to positive along with increasing anthropogenic activities at fine scale (Table 13).
Low anthropogenic activity group demonstrated negative
native—exotic richness relationship, while positive relationships were found for higher anthropogenic activity groups, including high agriculture group and high landscape heterogeneity group. However, similar trend was not found at broader scale.
‐2
FIG. 13. Cluster analysis of plots according to their anthropogenic activities. The
values in Y-axis had been standardized in factor analysis. Error bars indicate means ±
standard deviation.
TABLE 13. Native—exotic richness relationships of groups with different
anthropogenic activity groups. Correlation coefficients of native vs. naturalized
plants including native richness 1m2 vs. naturalized richness 1m2, and native
richness 1km2 vs. naturalized richness 1km2 at different groups were shown.
Naturalized indices
Richness 1km2 0.34 0.32 -0.18 0.24
*Correlation is significant at the 0.05 level (2-tailed).
According to cluster analysis, the plots were separated into three groups (Fig. 14), and these groups were characterized with different environmental factors. These three groups include low elevation and precipitation group, high precipitation group, and high elevation group. They were highly associated with low elevation and precipitation, high precipitation, and high elevation respectively.
Native—exotic relationships moved from negative to positive along with decreasing precipitation at fine scale (Table 14). High precipitation group demonstrated negative native—exotic richness relationship, while positive relationship was found in low elevation and precipitation group. Positive native—exotic relationship was also found at broad scale in low elevation and precipitation group.
‐1.5
FIG. 14. Cluster analysis of plots according to their environmental factors. The values
in Y-axis had been standardized in factor analysis. Error bars indicate means ± standard
deviation.
TABLE 14. Native—exotic richness relationships of groups with different
environmental factor groups. Correlation coefficients of native vs. naturalized
plants including native richness 1m2 vs. naturalized richness 1m2, and native
richness 1km2 vs. naturalized richness 1km2 at different groups were shown.
Naturalized indices
Native indices
Low elevation &
precipitation
group
High
precipitation
group
High
elevation
group
Richness 1m2 0.10* -0.09* -0.08
Richness 1km2 0.36* -0.18 -0.03
*Correlation is significant at the 0.05 level (2-tailed).
Two groups of plots were categorized (Fig. 15) by employing both of the anthropogenic and environmental factors, integrated high anthropogenic activities with low elevation group and integrated low anthropogenic activities with high elevation group. Native—exotic relationships moved from negative to positive along with increasing anthropogenic activities and decreasing elevation at fine scale (Table 15). Integrated low anthropogenic activities with high elevation group demonstrated negative native—exotic richness relationship, while positive relationship was found in integrated high
anthropogenic activities with low elevation group. Positive native—exotic relationship was also found at broad scale in
integrated high anthropogenic activities with low elevation group.
‐1.5
FIG. 15. Cluster analysis of plots according to both of the anthropogenic and
environmental factors. The values in Y-axis had been standardized in factor analysis.
Error bars indicate means ± standard deviation.
TABLE 15. Native—exotic richness relationships of groups with different
anthropogenic and environmental factors. Correlation coefficients of native vs.
naturalized plants including native richness 1m2 vs. naturalized richness 1m2, and
native richness 1km2 vs. naturalized richness 1km2 at different groups were shown.
Naturalized indices
Native indices
Integrated
high anthropogenic
activities with low
elevation group
Integrated
low anthropogenic
activities with high
elevation group
Richness 1m2 0.09* -0.18*
Richness 1km2 0.36* -0.18
*Correlation is significant at the 0.05 level (2-tailed).
Integrated high anthropogenic activities group had higher naturalized species cover percentage than integrated low
anthropogenic activities group, but native species cover percentage and total cover percentage were higher in integrated low
anthropogenic activities group than high anthropogenic activities group (Table 16).
TABLE 16. The difference of cover percentage in two integrated plot groups. The average total cover percentage, average native cover percentage, average naturalized cover percentage at integrated high anthropogenic activities with low elevation group and integrated low anthropogenic activities with high elevation group. Using ANOVA to examine the significant level between two groups. Demonstrate average ± standard error. * indicates the significant differences at 0.05 significant level.
Total cover Native cover Naturalized cover
High anthropogenic activities 89.1 ± 2.5 40.3 ± 2.9 48.8 ± 2.5 Low anthropogenic activities 98.4 ± 3.1 60.5 ± 3.6 38 ± 3
P values 0.02* <0.01* <0.01*