Invasive shrub Lantana camara L. alters the flora and soils in tropical dry deciduous forests of Central India

Lantana camara (hereafter Lantana) is a highly noxious invasive weed species of global concern. However, its impacts on floristic and soil properties in tropical dry deciduous forests are elusive and fragmented. We aimed to assess the changes in the flora and soil properties following the invasion by Lantana in Central Indian forest ecosystems. Three study sites were selected, and each site was further divided into two subsites: Lantana‐invaded (LI) and uninvaded (UI). In total, 60 plots of 0.25 ha each (10 plots in each subsite) were laid randomly. Within each plot, floristic structure, composition, diversity, soil organic carbon (SOC), soil total nitrogen (STN), moisture (M%), pH, and bulk density (BD) were assessed. Lantana‐invaded sites showed a significant decrease in density (D), basal area (BA), species richness (SR), and evenness (E) of saplings (<3 cm diameter at breast height [DBH]), juveniles (between 3 and 9.9 cm DBH), and herbs. In LI sites, a reduction of 57% and 25% was observed in lower DBH class of trees (saplings and juveniles). In all the LI sites, significant increase in SOC, STN, and M%, and a significant decrease in pH were recorded. Lantana may greatly impact the vegetation and soil properties, and successively, these strong changes may increase its invasive potential and ability to replace native species by averting their natural regeneration potential. Therefore, a proper management strategy of this noxious weed is imperative to prevent its further expansion and future problems.


| INTRODUC TI ON
Biological invasions are a major threat to global biodiversity and have significant ecological impacts in a wide range of ecosystems (Seebens et al., 2017;Stinca et al., 2020). Invasive species are the second biggest threat to biodiversity loss after habitat fragmentation (Ahmad et al., 2019;Simberloff et al., 2013) as they pose detrimental impacts on the native local biodiversity, ecosystems, economy, and human health (Hejda et al., 2009;Ramírez-Albores et al., 2019).
They are found in almost all terrestrial ecosystems and are important drivers of global change biology (Vitousek, 1994). Invasive species impact ecosystems by altering the fire regimes, geomorphology (Fei et al., 2014;Gaertner et al., 2014) and have substantial impacts on both ecosystem structure and functioning, such as reduction in native species diversity, changes in ecosystem productivity, and alteration of soil nutrient pools (Pysek et al., 2012).
Numerous mechanisms have been identified by which plants alter the physical, chemical, and biological properties of soils | 1413 LONE Et aL. (Binkley & Sollins, 1990). Many involve changes in the quantity, quality, and/or timing of inputs of plant-derived substrate; others may result from changes in microclimate associated with changes in density and height of the vegetation, or changes in water relations (Sharma & Raghubanshi, 2009). In addition to the potential underlying mechanisms of Lantana camara (hereafter Lantana) colonization, the contemporary reports on changes in density and Lantana proliferation (Prasad, 2010;Ramaswami & Sukumar, 2014) indicate the possible role of other mechanisms operating at a larger scale, for example, extended drought or stochastic rainfall. Furthermore, the effect of other local-scale factors, for example, biotic interactions, soil, and topography are also considered as possible mechanisms underlying invasion (Sharma & Raghubanshi, 2011). The success of invasive plants is due to its particular traits such as phenotypic plasticity, short-life spans, pollination by generalists, high fecundity, rapid growth rates, and allelopathy. (Rejmánek, 1995).
Tropical forests are known to have rich species diversity and tropical dry forests are the most endangered and degraded of all ecosystems in the world (Janzen, 1988). There is a growing consensus that invasive plants create "novel" tropical ecosystems with vegetation transitions (Seastedt et al., 2008). Furthermore, shifts in vegetation transitions are often accompanied by alterations in soil physico-chemical properties (Ehrenfeld, 2010) that lead to further degradation. The facilitation of invasion by forest degradation is often overlooked, particularly in the tropical forests of developing countries such as India (Mungi et al., 2020). Globally, Lantana removal experiments have been shown to promote native tree diversity and avian diversity (Lambert et al., 2016;Safari & Byarugaba, 2008).
However, as Lantana is not the only factor shaping forest communities, successful restoration requires that Lantana removal should be accompanied with other suitable measures and long-term monitoring of interactions is necessary (Lambert et al., 2017;Yeates & Schooler, 2011). Despite the numerous studies on the impacts of invasive species, only few have studied the impacts on diversity, vegetation structure, and soil properties in forests (Dobhal et al., 2011;Kumar et al., 2020;Sharma & Raghubanshi, 2009Sundaram & Hiremath, 2012).
Lantana is a vigorously growing shrub that is highly invasive with 650 varieties in over 60 countries (Global Invasive Species Database, 2020). It is widespread covering about 13 million hectares (Goyal et al., 2018;Sharma et al., 2005) and threatens 44% of the total Indian forests (Mungi et al., 2020). It has spread in almost all the dry deciduous forests of India (Sharma & Raghubanshi, 2006

| Study species
Lantana camara L. (Verbenaceae) is a woody shrub native to Central and South America and is regarded as one of the ten worst invasive species in the world (Richardson & Rejmanek, 2011).

It was introduced as an ornamental hedge plant in East India
Company Botanical Gardens in Calcutta in 1809, from where it escaped and became invasive (Kannan et al., 2013). The plant is profusely branched and grows up to 2-4 m high in open unshaded sunny environments (Day et al., 2003), and as a liana up to 15 m when light intensity is low (Lowe et al., 2000). It is shade-tolerant and produces 10,000-12,000 fruits (Kohli et al., 2006) and is very commonly distributed across fragmented dry forests in Central India (Mungi et al., 2020).

| Study area description
The present study was conducted in three forest study sites (Kesli, Deori and Shahgarh ranges) located between 21°17′-26°52′N and 78°08′-82°49′E in the district Sagar of Madhya Pradesh (M.P), Central India (Table 1, Figure 1). The area is covered by Vindhyachal mountain range at an average height of 420 m asl. The forest in the area belongs to group 4b of the Champion and Seth's classification (Champion & Seth, 1968), and the climate is subtropical with hot dry summers (March to mid-June), monsoon season (mid-June to September) and, cool and dry winters (October-February). The area receives an annual average rainfall of 1197.6 mm of which approximately 90% takes place during the southwest monsoon. The mean annual minimum and maximum temperatures vary between 11.6 and 40.7°C in January and May, respectively. The major soil types of the area are clay loam, sandy clay loam, and sandy loam. The vegetation of the area is characterized by tropical dry deciduous forests, dominated by Tectona grandis, Diospyros melanoxylon, Butea monosperma, and Lagerstroemia parviflora. During the last few decades, these forests have been subjected to anthropogenic disturbances such as grazing, felling and lopping for timber, fodder, and fuelwood collection, which lowered the canopy density and increased the light availability. This led the forests to be severely infested by plant invasions, particularly Lantana  which impacted the physical, chemical, and biological aspects of ecosystems (Lone et al., 2019).

| Vegetation sampling design
A reconnaissance survey of the entire region was carried out, three study sites were selected, and each site was divided into two subsites (Lantana-invaded [LI] and uninvaded [UI]). The LI subsites were selected, having Lantana cover/density >50%, whereas UI subsites were selected with no or very less Lantana cover. The phytosociological attributes such as species richness, diversity, density, frequency, basal area, and size class distribution were studied during the peak growing period (August-October) in 2017 and 2018. In each subsite, 10 square plots of 50 m× 50 m were laid randomly. The UI plots were chosen in neighboring localities at >50-100 m away from LI plots with similar site conditions. Each 50 m × 50 m plot was further sub-gridded into 25 (10 m× 10 m) quadrats to enumerate trees. In addition, 10 quadrats (5 m × 5 m) and 10 sub-quadrats (1 m× 1 m) were laid to enumerate the shrubs and herbs, respectively (Kershaw, 1964;Misra, 1968). The tree individuals were classified as follows: saplings (<3 cm DBH, diameter at breast height), juveniles (between 3 and 9.9 cm DBH), and adults (≥10 cm DBH). A total of 60 plots (50 m× 50 m) were laid in the three study sites. In each quadrat, all the tree individuals at 1.37 m above the ground were measured and shrub diameter was recorded at >10 cm above the ground level. Herbaceous individuals were counted and measured with digital Vernier caliper. Vegetation composition was evaluated by analyzing the density (D), basal area (BA), abundance, frequency, and Importance Value Index (IVI) following Curtis and McIntosh (1951) and Misra (1968). Diameter class-wise distributions were calculated for each subsite. The plant specimens were prepared and identified at the Department of Botany, Dr.
Harisingh Gour Vishwavidyalaya (A Central University) with the help of Flora of Bhopal (Oommachan, 1977). The assignment of a species to the family was done as per Angiosperm Phylogeny Group IV (APG IV) system of classification (Stevens, 2017).
The plant names and the corresponding author citations were given following GRIN (Germplasm Resources Information Network) Taxonomy and Global Biodiversity Information Facility (Wiersema, 2019).

| Soil sampling
From each plot, soil samples were collected at five random points to a depth of 10 cm using a soil core sampler of 5 cm internal diameter.
These samples were mixed thoroughly, and 250 g were collected, airdried, stored in airtight polyethylene bags, and sent to the laboratory for further analysis. The composite soil samples were ground using a mortar and pestle, and then sieved through a 2 mm stainless steel sieve. Soil organic carbon (SOC) and soil total nitrogen (STN) were estimated following Kirk (1950) and Walkley and Black (1934).
Another set of soil samples were collected from each study site to estimate bulk density (BD). Three sets of undisturbed soil cores were taken from each plot. The samples were oven-dried at 105 ± 5°C for 72 h and re-weighed. The coarse fragments were separated by sieving, and the samples were re-weighed. Soil BD was calculated following Pearson et al. (2005): where 2.65 was taken as a constant for the density of rock fragments (g cm −3 ).
The total C content of 0-10 cm soil depth was estimated following Pearson et al. (2005): Soil moisture (M%) was measured by the gravimetric method. Soil pH (1:2.5 ratio of soil: water) was measured with digital pH meter. Three replicates were tested for each forest plot (30 per subsite).

| Statistical analysis
A t-test was performed to compare the differences in means of species richness, density, basal area, diversity indices, and soil properties, and Tukey's HSD test (p < .05) was performed using SPSS 20.0. Diversity indices and box plots were computed and drawn using PAST 3.1 (Hammer et al., 2001, Natural History Museum, University of Oslo). Linear correlations were done to assess the relationships of Lantana densities with soil properties and vegetation attributes (species richness, density, and basal area). Bonferroni corrections were done to reduce multiple comparisons.

| Impact on species richness and diversity
A decline in species richness (SR) and diversity was observed with increase in Lantana density. The LI sites had a significantly lower (p < .05) SR with 98 species when compared with UI sites (132). In LI sites, the SR of sapling, juvenile, and adult trees ranged from 1-5, 3-15 and 3-14, whereas in UI sites, it ranged from 1-9, 2-14 and 4-16, respectively. The mean SR of saplings (13), juveniles (30), and adult trees (29) were significantly (p < .05) lower in LI sites than the saplings (21), juveniles (33) and adult trees (40) in UI sites. The herbaceous SR ranged from 16-36 and 12-26 in UI and LI sites, respectively, and the mean SR was significantly (p < .001) lower in LI (53) than UI (72) sites, whereas the SR of shrubs and lianas were also significantly (p < .05) lower in LI sites (7) than UI sites (13). The total SR was 55, 72, and 54 in LI sites and 71, 98, and 76 in UI sites for saplings, juveniles, and adults, respectively (Figures 2-5, Figure S1).
In total, 141 plant species (49 trees, 78 herbs, 14 shrubs, and lianas) from 122 genera and 44 families were documented in all the three study sites (Table S1). Eighty-nine species (63.1%) were found common to both UI and LI sites, while 43 species (30.5%) occur only in UI sites and nine species (6.4%) occur only in LI sites (  Figure S1).

| Impact on density and basal area
A significant impact of Lantana density (No. ha −1 ) was observed on the density (no. ha −1 ) and basal area (m 2 ha −1 ) of tree and herbaceous vegetation (Table S1). The density of tree saplings and juveniles decreased significantly (p < .001) with increase in Lantana density, whereas the adult tree density did not show any significant trend (Figures 2-5, Figure S1). The tree density of saplings and juveniles ranged from 4-308 and 4-100 individuals ha −1 in UI and 104-1336 and 132-732 individuals ha −1 in LI sites, respectively. The mean densities of saplings and juveniles were also reduced significantly The basal area of saplings and juveniles decreased significantly and 6 (LI) species}, and Acanthaceae {6 (UI) and 4 (LI) species} were the most speciose families (Table 2).

| Impact on diameter class distribution
Lantana density was associated with changes in diameter class distribution of trees. Tree density of lower diameter class (<3 DBH) was reduced by 57% in LI sites followed by 25% and 23%, of 3.1-10 and 30.1-40 DBH classes, respectively ( Figure 6). Tree density decreased significantly (p < .001) with increase in diameter class in both UI and LI sites. The highest density of 51.8% was contributed by 3.1-10 cm diameter class, and lowest of 0.7% density was contributed by >50 cm diameter class.

| Impacts on the soil properties
Lantana density was related to changes in soil properties. The values of BD (g cm −3 ) and pH were significantly (F = 12.96, p < .05) lower in LI than in UI sites, whereas M% was significantly (F = 68.55, p < .001) higher in LI than in UI sites (Figure 7). The SOC and STN stocks were significantly (F = 48.3, p < .001; F = 51.7, p < .001) higher in LI than in UI sites (Figure 7).

| Correlations of Lantana density with other variables
Lantana density was significantly negatively correlated with SR, density, and basal area of tree saplings and herb density (p < .001, Table 3). Non-significant negative correlations were observed with SR of juveniles, adults, total trees, and herbs; densities of tree juveniles and total trees; and basal area of tree juveniles, adults, and total trees. The SOC, STN, and M% showed significant positive correlations, whereas soil pH showed a significant negative correlation with Lantana density. Soil BD had a non-significant negative relationship with Lantana density (Table 3).

| DISCUSS ION
Comparing invaded and uninvaded sites helps us to measure the impact of invasive species on the native resident communities (Levine et al., 2003). In the present study, a significant (p < .05) de- where some species are more easily excluded than others in invaded sites (Stinson et al., 2007). About 33%, 8%, 26%, 25%, 43%, and 24% decline in SR were observed in saplings, juveniles, adults, herbs, shrubs and lianas, and total species, respectively in LI sites. Increase in Lantana cover causes pervasive losses in SR across multiple life forms (Sharma & Raghubanshi, 2010). The highest decline in SR was observed in the case of tree saplings, juveniles, and herbs. This could be because Lantana produces copious light-weight seeds with high adaptability that enables them to grow vigorously and suppress the growth of native plant species. Furthermore, the release of allelochemicals from its roots also hampers the growth of native plants (Kumar et al., 2020).
The values of Shannon index were lower in LI sites than UI sites for all the life forms and significantly in case of saplings, juveniles, and herbs. Lantana invasion is often associated with significant decreases in plant species richness, diversity, and evenness in deciduous forest types (Badalamenti et al., 2016). The evenness index increased in LI sites for tree saplings, but decreased for shrub and lianas. Changes in evenness of an ecosystem may impact the productivity, resistance to invasion, and local plant extinction rates (Wilsey & Potvin, 2000). The impact of invasion largely depends on the degree of dominance (Pyšek & Pyšek, 1995). Overall, the values of dominance index were lower in LI sites than UI sites.
The species compositional changes induced by Lantana invasion are primarily driven by gradual changes in vegetation structure (Gooden et al., 2009) (Gentle & Duggin, 1997). For example, if a species is mainly represented by saplings in a particular location, it is easily susceptible and might eventually be displaced, but if it contains several adult individuals, it would resist invasion and is unlikely to get displaced (Gooden et al., 2009). Such variations in species density due to invasion by Lantana gradually alter landscape-level heterogeneity (Vitousek et al., 1996).
Invasion by Lantana has significantly (p < .001) lowered the basal area of tree saplings, juveniles, and herbs in LI sites compared with UI sites. This might be due to the formation of thick Lantana thickets that alter the microenvironment (light and temperature) inhibiting germination or growth (Sharma & Raghubanshi, 2007). Furthermore, at the ground level, there occurs accumulation of Lantana litter which also cause allelopathic suppression of growth and recruitment of saplings, juveniles, and herbs (Gentle & Duggin, 1997). The changes in quantitative ecological parameters such as density, basal area, and IVI by invasion of Lantana lead to alteration in plant assemblage patterns and forest structure, which may eventually create demographic instability (Sharma & Raghubanshi, 2010). Analysis of tree size class distribution reveals the population structure of a forest (Newbery & Gartlan, 1996).

Uninvaded (UI) Lantana-invaded (LI)
good regeneration behavior, while their insufficient numbers denote poor regeneration (Saxena & Singh, 1985). In this study, tree density declined significantly (p < .001) with increase in diameter class in LI sites than in UI sites ( Figure 6). This trend indicates that the tree species in the studied sites possess a good regenerative capacity. Even so, tree saplings and tree juveniles were the most impacted life forms by Lantana invasion and further expansion of this species could affect their regenerative potential. Lantana is known to displace natural scrub communities and prevent natural regeneration of some tree species (Ambika et al., 2003;Sharma & Raghubanshi, 2006). and chemistry significantly alter SOC, STN, and M% in soils (Sharma & Raghubanshi, 2009). The significantly higher M% in the soil of LI sites could be due to the presence of more debris and organic matter from dead Lantana leaves, which decompose slowly (Fan et al., 2010;Singh et al., 2014). Lantana produces a substantial quantity of litter rich in allelochemicals which might react with organic matter and affect soil property (Ruwanza et al., 2013). Allelochemicals are usually released from roots, shoots, leaves, or flowers, which negatively affect the neighboring native species (Rice, 1974). The effects of allelochemicals are usually greater in the introduced range than in the native range (Inderjit et al., 2011). Soil BD and pH were lower in LI sites than in UI sites which could be due to higher litter content and its decomposition by micro-organisms. Similar results have been observed in tropical dry deciduous forests by Sharma and Raghubanshi (2009). The altered soil properties provide favorable conditions for further invasion by Lantana and other invasive species (Niu et al., 2007). Lantana invasion not only affects the native resident plant community structure and composition, but also changes the soil physico-chemical properties that may not be suitable for their growth.
The SR, density, and basal area of tree saplings and herb density showed significant negative relationships with Lantana density, whereas negative non-significant correlations were observed with SR of juveniles, adults, total trees, and herbs; densities of tree juveniles and total trees; and basal area of tree juveniles, adults, and total trees (Table 3). The SOC, STN and M% showed significant positive correlations with Lantana density, whereas pH showed a significant negative relationship (Table 3). Similar findings were recorded by Badalamenti et al. (2016) in Mediterranean ecosystems.
Lantana density was significantly positively correlated with shrub and liana density (p ≤ .01), which is due to its dominance in the shrub and liana category. Invasive plants are known to alter soil pools and ecosystem processes by changing the organic matter inputs, decomposition, and mineralization (Ashton et al., 2005;Ruwanza & Shackleton, 2016). Changes in soil nutrients following Lantana invasion could be a contributor to its successful proliferation. Levine et al. (2006) also opined that such increase in nutrients positively affect the growth and spread of the invader in the form of the "push and pull" theory of invasion.
The studied sites are open forests and are exposed to frequent grazing, illegal felling of trees for timber, fuel, and fodder, have canopy openings, variability in light, drought, and fires (Kumar et al., 2020;Sharma & Raghubanshi, 2006). Thus, these disturbances trigger changes in community structure, composition, and microclimate conditions, which could help in the successful proliferation of Lantana. The pervasive threat posed by Lantana to native vegetation at the scale of individual forest types, as well as at a larger landscape-level has long-term consequences for forest structure and composition (Sundaram & Hiremath, 2012). Some forest types may be more vulnerable to invasion than others, while communities vary in their responses to invasion (Hejda et al., 2009). Therefore, in heterogeneous landscapes, it is necessary to examine the response of community variables to invasive species at both at the scales of landscape-level as well as that of individual forest types (Sundaram & Hiremath, 2012). Based on these findings as well as other studies (Prasad, 2010;Ramaswami & Sukumar, 2011Sharma & Raghubanshi, 2007), it is expected that after accounting for the influence of rainfall, terrain, slope, altitude, fire frequency, and tree density, the SR of tree species would decrease as Lantana cover increased.
The disturbance factors such as fires, livestock grazing, illegal felling of trees, canopy openness, and drought in the low diversity/density study sites will lead to lower diversity/density values and facilitate invasion, which would further lead to decline in diversity/density. Increased species richness might confer resistance to Lantana invasion through greater community stability, resilience to disturbance, and a more complete utilization of light, space, and nutrient resources through niche partitioning (Tilman et al., 2006). This is confirmed by empirical and correlative evidence that Lantana invasion is inhibited by intact undisturbed vegetation (Prasad, 2012). Native species richness remains stable at lower density levels, but declines rapidly above the threshold level, which leads to compositional change. Thus, sparse Lantana cover has little effect on the resident community, with impacts elicited at an advanced stage of invasion. Potentially, broad-scale conservation of species diversity could be achieved by maintaining Lantana infestations below the threshold cover at sites containing regionally common species that are also widely represented in non-invaded vegetation. Furthermore, Lantana removal programs that also consider the local site conditions (Lambert et al., 2016;Yeates & Schooler, 2011) are essential for effective restoration of these forests.

| CON CLUS ION
The present study revealed a significant negative impact of L. camara on diversity and vegetation attributes of native plant communities in the tropical dry deciduous forests of Central India. It is clear that Lantana-invaded sites comprise significantly lower richness, density, and basal area than uninvaded sites. Lantana invasion not only affects the vegetation, but also alters the soil properties favoring its own growth. Furthermore, the alterations in vegetation structure, composition, and soil properties could lead to changes in ecosystem functioning. Appropriate methods and long-term monitoring studies in permanent plots are needed for better understanding, management, and restoration of the invaded landscapes in tropical forests.

AUTH O R S ' CO NTR I B UTI O N
The study was conceptualized and designed by PAL, JAD, SK, and MLK. Material preparation, field work, data collection, and analysis were performed by PAL, JAD, and SK. The first draft of the manuscript was written by PAL, JAD, and SK. MLK provided review and comments. All the authors have read and approved the submitted version of the manuscript.

ACK N OWLED G M ENTS
We are thankful to the Madhya Pradesh State Forest Department and Forest Department of Sagar district for permission and for providing the necessary facilities and staff support during the field work. We also thank Prof. Pramod Kumar Khare, Department of Botany, for the identification of plant specimens.

CO N FLI C T O F I NTE R E S T
No potential conflict of interest was reported by the authors.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are openly available in the Dryad Digital Repository: https://doi.org/10.5061/dryad. sj3tx 966t .