Identification of the Invasive Species in This Region

We identified 347 non-native plant species, that were naturalized or invasive across the Indian subcontinent (ESM 2). Of these, 282 species were present in the NWGK, and 79 species (approximately 25% of the naturalized species) were identified as invasive. Despite its limited geographic extent (approximately 70,000 km²), the NWGK contains more than half of the subcontinent’s non-native, naturalized flora, and approximately 25% of these are invasive (Inderjit et al. Reference Inderjit, van Kleunen, Hejda, Babu, Majumdar, Singh, Singh, Salamma, Rao and Pyšek2018; Sankaran et al. Reference Sankaran, Khuroo, Raghavan, Molur, Kumar, Wong and Pagad2021). This exceeds expectations based on the “tens rule” (Williamson and Fitter Reference Williamson and Fitter1996) but aligns with recent global syntheses that suggest that the tens rule may oversimplify invasion processes and underestimate the magnitude of impacts (Jeschke and Heger Reference Jeschke and Heger2018).

The discovery frequency of invasive plant species in this region has remained relatively constant since 1840, with temporal fluctuations likely reflecting variation in sampling effort (Figure 1A). More than half of the invasive species were annual herbs, whereas trees were the least represented growth form (Figure 1B). The biogeographic origin of species was highly skewed, with the large majority originating from South America, and with comparatively minor contributions from Africa, other parts of Asia, Australia, and North America (Figure 1C). These skewed distributions are congruent with trends observed in South Asia and are likely driven by climatic similarity and historical trade linkages (Gulzar et al. Reference Gulzar, Wani, Hassan, Reddy, Shrestha, Mukul, Shabbir, Iqbal, Ranwala, Sujanapal, Rashid and Khuroo2024). Taxonomic representation was also skewed, with most species belonging to the families Asteraceae and Fabaceae, followed by smaller contributions from Euphorbiaceae, Solanaceae, and Amaranthaceae (Figure 1D). Other families were represented by only one to three species. The domination by Asteraceae and Fabaceae is consistent with regional and global patterns that likely reflect the large family sizes and life-history traits adapted to seasonal climates and high levels of disturbance (Gulzar et al. Reference Gulzar, Wani, Hassan, Reddy, Shrestha, Mukul, Shabbir, Iqbal, Ranwala, Sujanapal, Rashid and Khuroo2024; Stohlgren et al. Reference Stohlgren, Pyšek, Kartesz, Nishino, Pauchard, Winter, Pino, Richardson, Wilson, Murray, Phillips, Ming-yang, Celesti-Grapow and Font2011).

Figure 1.

Invasive plants of the Northern Western Ghats and Konkan region (NWGK). (A) Discovery of invasive species based on the date of first report. The 95% confidence band was estimated using nonparametric bootstrapping where first-report dates were resampled 1,000 times with replacement. (B) Distribution of growth form: annual herbs, shrubs, perennial herbs, trees. (C) Biogeographic origin: continent. (D) Taxonomic affiliation: family. The labeled version of the phylogenetic tree is available in Supplementary Figure S5.

Figure 1. Long description

Panel A: A line graph shows the cumulative number of invasive species over time, with the x-axis representing the year of first report from 1840 to 2000 and the y-axis representing the cumulative number of invasive species. The graph includes a 95% confidence band estimated using nonparametric bootstrapping. Panel B: A vertical bar graph displays the distribution of growth forms, with the x-axis listing categories (annual herbs, shrubs, perennial herbs, trees) and the y-axis showing the number of species. Annual herbs have the highest count, followed by shrubs, perennial herbs, and trees. Panel C: A vertical bar graph illustrates the biogeographic origin of invasive species by continent, with the x-axis listing continents (Tropical America, Africa, Asia, Australia, North America) and the y-axis showing the number of species. Tropical America has the highest count, followed by Asia, Australia, North America, and Africa. Panel D: A vertical bar graph presents the taxonomic affiliation of invasive species by family, with the x-axis listing families and the y-axis showing the number of species. The Asteraceae family has the highest count, followed by other families in descending order. A phylogenetic tree diagram is also included to show the taxonomic relationships.

Range Size and Distribution of the Invasive Species

Range sizes of invasive plant species in the NWGK estimated from the herbarium-based approach varied by an order of magnitude. Approximately 15% of species were highly widespread, occupying around 20% of the surveyed area, but more than 50% of the species had limited distributions (Figure 2). The distribution of invasive plants mirrored the proportional area of land-use categories, with croplands and deciduous forests supporting the highest numbers. However, after we accounted for relative coverage of LULC categories, mangroves and human-dominated landscapes harbored a disproportionately larger number of invasive species than expected by chance. No differences between expected and observed numbers were detected for the other LULC categories (Supplementary Figure S6). The prevalence in areas of human habitation is consistent with expectations based on high introduction frequencies, greater disturbance, and underutilized niches (Hierro et al. Reference Hierro, Villarreal, Eren, Graham and Callaway2006). Disturbance in mangrove ecosystems is known to facilitate introduced species (Biswas et al. Reference Biswas, Biswas, Limon, Yan, Xu and Khan2018), and is consistent with the higher number of invasive plants found in mangroves in this region. Given the important ecosystem services provided by mangroves (Bouillon et al. Reference Bouillon, Borges, Castañeda-Moya, Diele, Dittmar, Duke, Kristensen, Lee, Marchand, Middelburg, Rivera-Monroy, Smith and Twilley2008; Donato et al. Reference Donato, Kauffman, Murdiyarso, Kurnianto, Stidham and Kanninen2011), the high incidence of invasive plants in these sensitive ecosystems warrants increased monitoring and proactive management.

Figure 2.

Range size estimates of invasive plants of Northern Western Ghats and Konkan (NWGK) using herbarium collections. (A) Histogram of the number of grids occupied by species. Representative examples of species with (B) low: Cosmos bipinnatus Cav.; (C) medium: Acacia auriculiformis A. Cunn. ex Benth.; and (D) high: Alternanthera sessilis (L.) DC. geographic extent.

Temporal Dynamics

For the 38 species with sufficient sample sizes to examine temporal dynamics of invasion, we observed distinct invasion trajectories (Supplementary Figure S7). Most species (>90 %) had lag durations that ranged between 0 and 40 yr, but the majority of these showed no detectable lag phase (Supplementary Figure S8). In examining the relationship between invasion success and the invasion dynamics, we saw that residence time was positively associated with range size (P < 0.001), and lag duration was negatively associated with range size (P < 0.05). Thus, shorter lag durations likely reflect rapid establishment and spatial expansion following introduction, potentially due to early release from biotic resistance (Larkin Reference Larkin2012). Life span was marginally related to range size (P = 0.06), with annuals occupying larger ranges than perennials (Supplementary Table S4). Identifying and quantifying these temporal patterns can help anticipate future spread and control species in early stages.

Expert-based Assessment of Demographic and Impact Dimensions

Because herbarium records can be subject to temporal lags and sampling biases, we complemented this approach with expert assessments to quantify recent distribution, spread, local abundance, and impact for 56 invasive plants. According to expert assessments, approximately half of the species exhibited medium to high range sizes (Figure 3A). With respect to recent change in range size, experts suggest that the majority showed no change over the past three decades, and only around 15% of species have increased in range during this period (Figure 3B). Experts did not identify declining trends for any species over the last 30 yr, with the exception of Parthenium hysterophorus L., which they suggested is possibly due to displacement by other emerging invasive species such as [Senna uniflora (Mill.) H.S. Irwin & Barneby] and [Mesosphaerum suaveolens (L.) Kuntze]. Additionally, most species occurred across multiple habitat types, with relatively few habitat specialists (Figure 3C).

Figure 3.

Distribution of invasive plants along different dimensions of invasion obtained through expert-based surveys. (A) Range size; (B) temporal trend: change in range size; (C) habitat breadth: number of habitats occupied; (D) local abundance; (E) PA Presence: presence in protected areas (PA) qualified by disturbance levels; and (F) biodiversity impact. Abbreviations used: V.low, very low; Med, medium; Mod, moderate; Abs, absent; S.dist, sometimes disturbed; F.dist, frequently disturbed; Undist, undisturbed.

According to experts, local abundance was generally low, with most species classified as having very low to low abundance (Figure 3D). Nonetheless, around 30% of species were reported to have moderate to high local abundance. Because local abundance, range size, and per capita effects jointly determine invasion impact (Parker et al. Reference Parker, Simberloff, Lonsdale, Goodell, Wonham, Kareiva, Williamson, von Holle, Moyle, Byers and Goldwasser1999), these moderate- to high-abundance species are likely to drive disproportionate ecological effects. Approximately half of the invasive plants were reported to occur within protected areas, but predominantly in disturbed habitats at the fringes of these protected areas (Figure 3E). Notably, [Chromolaena odorata (L.) R.M.King & H.Rob]. was the only species that was consistently reported from undisturbed areas within protected sites. These findings support evidence that protected areas can restrict invasive species establishment and impacts (Mungi et al. Reference Mungi, Qureshi and Jhala2021). Around 45% of the invasive species negatively affected local biodiversity, but according to experts, most of the others did not have any discernible impacts on biodiversity (Figure 3F). The experts did not identify any species (except for C. sulphureus) that were not already included in our initial list, which validates that our list of invasive species for this region was comprehensive. The experts identified the following six species that were not included in our list as potentially emerging invaders of NWGK: S. uniflora, [Gliricidia sepium (Jacq.) Kunth], [Tecoma stans (L.) Juss. ex. Kunth], [Ipomoea parasitica (Kunth) G. Don], Mikania micrantha Kunth, and Physalis peruviana L. Finally, experts did not identify any species that were formerly invasive but should no longer be regarded as problematic.

Identifying the Major Axes of Variation in Dimensions of Invasiveness

The first three principal component axes explained approximately 74% of the total variation across the different dimensions of invasiveness. This indicates coordination among these different dimensions and suggests that much of the multidimensional nature of invasiveness can be captured by a reduced number of axes. PC1 accounted for approximately 45% of the variation and was primarily associated with high loadings from range size (expert survey), presence in protected areas, and local abundance. PC2 explained 17% of the variation and was characterized by higher loadings for range size (herbarium-derived) and impact on biodiversity, while PC3 was dominated by habitat breadth obtained from the expert survey (Supplementary Figure S9). Overall, dimensions derived from the expert-based survey exhibited stronger loadings on PC1, except for impact on biodiversity, which loaded on PC2, and habitat breadth, which loaded most strongly on PC3. This indicated that ecological impacts vary among species and are only partially explained by distribution patterns. These results emphasize the need for management prioritization frameworks that integrate both spread potential and ecological impact, rather than assuming impacts scale directly with range size. The herbarium-derived dimension of range size showed higher loadings on PC2 (Figure 4; Supplementary Figure S10) and, interestingly, was only weakly correlated with range size from the expert-based survey (Supplementary Figure S11).

Figure 4.

The major axes of variation in dimensions of invasiveness. Principal component analysis (PCA) biplot for the 55 species based on scores for different dimensions of invasiveness obtained from the herbarium- and expert-based approaches. Range H, range size herbaria; Range Q, range size questionnaire; Habitat, habitat breadth; PA-prs, presence in protected areas; Abundance, local abundance; Impact_bio, impact on biodiversity.

Identifying High-Priority Invasive Plants

The cumulative scores obtained across all dimensions of invasiveness from expert responses were used to generate a ranked list of invasive plants of NWGK (ESM 2). Based on these scores, five species—C. odorata, M. suaveolens, C. sulphureus, [Alternanthera ficoidea (L.) P. Beauv.], and Lantana camara L.—had exceptionally high scores (Figure 5). Among the remaining species, 19 exhibited very low scores (<10), while 29 showed moderate scores (10–16). Of the five species with high impact scores, C. odorata and L. camara are also nationally prioritized (Mungi et al. Reference Mungi, Kaushik, Mohanty, Rastogi, Johnson and Qureshi2019; Sandilyan Reference Sandilyan2019), whereas C. sulphureus and A. ficoidea have received comparatively little attention, resulting in limited knowledge for effective control and management strategies. Given their high local priority, further work is needed to develop and evaluate appropriate control measures for these species.

Figure 5.

Distribution of the final composite scores for the invasive species in the Northern Western Ghats and Konkan region (NWGK) based on expert evaluation across multiple dimensions of invasion.

Comparing Invasive Species at the Regional, National, and Global Scales

Comparison of the 79 regionally problematic invasive species with national and global invasive species revealed substantial differences (Figure 6). Eighty-two percent of the species identified as problematic in NWGK were also identified as invasive at the national level, whereas only 50% were classified as invasive at the global scale. Fourteen regionally problematic species were absent from the national database: six of these appeared in the global database, while eight were missing from both. Thirty-four species were identified as invasive at all the three scales, which represents approximately half of the problematic species in the NWGK list. In addition, 122 species listed as invasive in India were not invasive at the global (GIDIAS) or the regional (NWGK) scale. The lack of agreement in species presence across lists developed at different spatial scales and using different methods highlights the need for regional-scale assessments to guide management.

Figure 6.

Euler diagram illustrating the overlap of species. Global, 1188 species (green); national (India), 265 species (blue); regional (Northern Western Ghats and Konkan [NWGK], this study), 79 species (peach).

For species recognized as invasive by more than one list (global, national, regional), we conducted a rank-correlation analysis to compare their ranks across lists. Surprisingly, no significant correlations in species rankings were detected across most pairwise comparisons, with the exception of the expert-derived regional list and the global list (r = 0.28, P < 0.05) (Supplementary Figure S12). Sample sizes in some of these pairwise comparisons were limited, and thus lack of relationships should be viewed cautiously. Nevertheless, these results indicate limited concordance in prioritization outcomes across the different lists and highlight the importance of spatial scale in invasion impact assessments. Global databases often assign high-impact rankings to species with large range sizes that occupy multiple ecosystems, even if their per capita or context-specific impacts are moderate (Nentwig et al. Reference Nentwig, Bacher, Kumschick, Pyšek and Vilà2018). In contrast, assessments conducted at smaller spatial scales with narrower environmental envelopes (Davies et al. Reference Davies, Chesson, Harrison, Inouye, Melbourne and Rice2005) may identify species with narrower range size but high local impacts.

Advantages and Limitations of Herbarium- and Expert-derived Data

We combined two complementary approaches, each with distinct advantages and limitations, to assess demographic and impact dimensions of invasive plants in the NWGK. Herbarium records provided quantitative estimates of range size, spread rates, and temporal dynamics, while expert assessments captured recent distributions, local abundance, and ecological impacts. While both approaches are commonly used, both have their inherent limitations. Herbarium data can be taxonomically, spatially, and temporally biased, and collection lags may obscure current distributions (Hyndman et al. Reference Hyndman, Mesgaran and Cousens2015). For example, collection effort is often biased toward species with medicinal, economic, or horticultural value, while common weedy species tend to be neglected. Our results showed that herbarium data overestimated species associated with human use or habitation (e.g., Eclipta prostrata (L.) L., Oxalis corniculata L.) and underestimated the range size for recent invaders (e.g., P. hysterophorus, Ipomoea carnea Jacq.). However, some of these biases can be controlled for, and herbarium collections represent a readily available quantitative data source that can be effectively used to provide insight into invasion dynamics and impact for a large number of species at large spatial and temporal scales.

Expert assessments can also be subject to observational and perception bias and limited by lack of empirical evidence. For instance, our assessment included only a single question addressing direct ecological impacts, because experts could not reliably evaluate more complex dimensions such as impacts on nutrient cycling or hydrology, likely due to the lack of sufficient empirical data. Additionally, data from expert assessments are typically qualitative, coarser in spatial resolution, and limited in temporal scale (White et al. Reference White, Jennings, Renwick and Barker2005). Expert assessments primarily reflect contemporary knowledge and are less suited to reconstructing long-term invasion dynamics. However, expert assessments are better suited for evaluating local abundance and ecological impacts that cannot be inferred from herbarium data. These contrasting advantages and limitations highlight the value of integrating multiple sources of evidence for invasive species prioritization. Herbarium data provide fine-scale estimates of range size and temporal trends, whereas expert assessments, though typically at coarser spatial scales, enable prioritization across additional dimensions beyond range size. Together, they provided a more comprehensive assessment of invasion dynamics and impacts in the NWGK region.