Population Parameters as Key Factors for Site-Specific Distribution of Invasive Weed Rhynchosia senna in Semiarid Temperate Agroecosystems

Abstract

The genus Rhynchosia includes more than 550 species, some exhibiting invasive behavior. Rynchosia senna var. senna (RS) is a challenging weed to control in its native range; however, its invasive potential remains unknown. The aim of this study was to evaluate RS demographic parameters to determine its invasive potential, including (i) plant fecundity during the first year of young adult and in adult plants, (ii) seed dispersal, (iii) pre- and post-dispersal predation, (iv) soil seedbank persistence, and (v) field emergence patterns. RS fecundity declined in autumn and mainly in early established cohorts. Fecundity was influenced by pre-dispersal predation (Bruchus spp. 12 ± 2%), and post-dispersal removal by birds (66 ± 4%) and arthropods (37 ± 5%). Seed dispersal decreased with distance. Seedling emergence occurred mainly during early summer (75%), and to a lesser extent during late summer (20%) and autumn (5%). Seed physical dormancy loss (~80% in the first year) defines a short persistent seedbank. Under the evaluated conditions (native environment), RS shows a limited invasive potential. However, in non-native environments, in the absence of natural predators, its prolific fecundity and the occurrence of staggered emergence patterns could easily enhance invasiveness, enabling rapid colonization, as observed in Medicago polymorpha L.

1. Introduction

Weeds present a significant threat to crop yield depending on their biological traits and the soil and climatic conditions of different agroecological regions [1]. If left unmanaged, weeds reduce crop yield, reproduce, and increase future infestations, posing a persistent threat to agricultural production. Consequently, weed control is a critical component of cropping systems. Currently, most cropping systems heavily depend on herbicides due to their efficiency, ease of application, and relatively low cost [2,3,4]. However, increasing efforts are being made to minimize chemical inputs to mitigate environmental impacts and prevent herbicide-resistant weed populations. Public concern over the environmental side effects of herbicide use has grown over the past few decades [4,5]. For attaining effective weed control, knowledge of critical periods of weed-crop competition is necessary, as it plays a decisive role in weed management [6].

A major challenge in herbicide-based weed control is the uncertainty regarding the occurrence of weed emergence patterns. Efficient herbicide application requires accurate predictions of weed seedling emergence timing and density [7]. Weed distribution is directly influenced by dispersal and disturbance patterns, but successful colonization, establishment, and growth depend on demographic processes. Dynamic population models have proved to be very useful to understand weed management under different production systems and environments [8,9,10,11].

In central Argentina, agriculture has expanded westward from the humid Pampas towards the semiarid Espinal region in the austral Pampas, primarily at the expense of converting native, xerophytic forest ecosystems into croplands [12,13,14].

The expansion of agriculture to marginal lands with short histories of continuous farming and the introduction of new crop types provide interesting scenarios for studying weed community assembly [14]. Rynchosia senna var. senna (RS) appears as a difficult-to-control weed in its native range [15,16,17,18]. This genus (Papilionoidea; Fabaceae) includes more than 550 species in the world. Some of these species, such as R. capitata, R. minima and R. phaseoloides, have been identified as invasive weeds [19,20,21,22]. RS is a perennial herbaceous plant with a spring–summer growth cycle, though its invasive potential remains unknown [14]. The plant re-sprouts from the base of the crown in early spring, while seeds germinate simultaneously. RS exhibits highly indeterminate flowering and non-uniform maturity, with seed dispersal occurring through explosive pod dehiscence [23]. The seeds are reniform in shape, exhibiting color dimorphism (black and gray), with a P1000 of 14 g [24]. Due to its growth habit, this species forms dense clumps with twining and climbing stems. It competes with crop plants for water and nutrients, while its vines climb over vegetation, shading crops, causing lodging in small grains, and hindering harvest, as observed in Convolvulus arvensis [25,26].

In order to predict the invasive potential of RS, population demography studies are required. This approach involves identifying, understanding, and quantifying biological and ecological factors that regulate population dynamics. Key demographic parameters include seed production, dispersal, seedbank longevity, dormancy and germination [27]. It is essential to assess not only the effects of mean population parameters on population size but also how spatial and temporal variations in these parameters influence population dynamics [28,29,30]. Several authors have demonstrated through various studies that understanding the demographic parameters of different plant species allows for the identification of those factors that significantly influence population increase or decline [31,32,33]. This knowledge enables timely and efficient management decisions, targeting phenological stages when demographic parameters are most sensitive to control measures [34,35].

This study was conducted to evaluate the demographic parameters of Rhynchosia senna var. senna to assess its potential as an invasive weed species. Within these parameters, the following were quantified: (i) plant fecundity during the first year of young plants and in adult plants, (ii) seed dispersal, (iii) pre- and post-dispersal predation, (iv) soil seedbank persistence, and (v) field emergence patterns over the growing season. Quantifying these parameters will allow us to draw inferences about the species’ invasive potential and develop demographic models in future studies.

2. Materials and Methods

2.1. Field Study Site and Plant Material

The experiments were conducted from 2019 to 2022 in the experimental field of the INTA EEA Ascasubi (39°22′ S, 62°39′ W), located in the semiarid Espinal region. The soil of the experimental site was an entic haplustoll, sandy loam that was slightly alkaline (pH 7.5) and high in P content (30.6–33.5 ppm P Bray and Kurtz), with low organic matter content (1–1.2%). Weather data from each year were registered at the nearby meteorological station (less than 200 m) (Table S1).

A native population of RS from the EEA Ascasubi was selected as the seed source. In this population, young and adult plants were marked for fecundity and seed dispersal tests.

2.2. Demographic Parameter Estimation

2.2.1. Experiment 1: Fecundity

Seeds per plant were estimated during two growing seasons (2019/20, 2020/21) in young adult plants (<1 year old) and adult (>1 year) plants. Yellow pods were hand-collected in young (n = 10) and adult plants (n = 10). Seed harvest was divided into three cohorts: early summer (E.S, December and January), late summer (L.S, February and March), and autumn (A, April and May). RS is highly prone to pod dehiscence as it matures, so the harvest was performed before full maturity. Yellow pods tend to be dehiscent under laboratory conditions after drying.

2.2.2. Experiment 2: Pre-Dispersal Seed Predation

The proportion of Bruchus sp.-infected seed was quantified using seeds collected in Exp. 1 during two growing seasons. For this purpose, in each harvesting time, batches of 50 seeds each were randomly separated (three replicates) and further incubated on Petri dishes containing filter paper imbibed with distilled water at a constant temperature (25 ± 1 °C) in a growing chamber (until no more germination occurred). Ungerminated seeds were opened to check for Bruchus larvae. For each cohort, the number of seeds with Bruchus present (%) was quantified.

2.2.3. Experiment 3: Seed Rain Dispersal Distance

Seed rain dispersal distance was evaluated over two growing seasons (2019/20, 2020/21) by placing trays (0.15 × 0.50 m) at distances of 0, 1, 2, 3, and 4 m (d) from both young and adult plants (n = 4) with a distance between plants of 1 m. The trays had a 0.04 m rim to prevent secondary seed dispersal and were lined with a 1 mm mesh to prevent water accumulation. Trays were placed three times within each cohort for one week. After one week, seeds were collected from each tray and counted. The proportion of dispersed seeds was estimated as a function of distance, as a result of explosive pod opening.

2.2.4. Experiment 4: Post-Dispersal Seed Predation

During two growing seasons (2020/21 and 2021/22), trays were placed at ground level, each containing 50 seeds (n = 3), and subjected to three treatments to assess predation by arthropods and birds. The treatments involved the use of a 0.18 × 0.23 m plastic cage. To evaluate bird predation, two of the treatments consisted of covered and uncovered conditions, where seeds were either exposed or protected by 30 g (7200 kg ha⁻¹) of wheat stubble. For arthropod predation, the third treatment involved trays at ground level covered with a 0.20 × 0.30 m tin roof and lateral protection using a plastic mesh with 0.01 × 0.01 m openings to prevent bird access while allowing insect access. These treatments were applied each year during the seed production period, seven times in the first year and five times in the second year. Observations were made every 15 days, and seeds were counted. After each observation, the number of seeds was quantified and all remaining seeds in each treatment were removed, and the plastic cages were refilled with fifty fresh seeds.

2.2.5. Experiment 5: Seed Bank Viability and Dormancy Loss

Mature pods were collected at the end of summer 2018/19, packed in paper bags, and stored under laboratory conditions for 2–4 weeks prior to the experiment. The seeds used to study the soil seedbank behavior were those without any signs of predation (i.e., no exit holes or depressions) and regular in shape, with a P1000 of 14 g. Batches of 100 randomly selected seeds (n = 6) were placed inside permeable nylon mesh bags (0.10 × 0.10 m) to simulate natural soil conditions. The batches were buried in autumn (April 2019) at a depth of 5 cm in sandy loam soil (pH = 7.5, soil organic matter = 1.2%) devoid of vegetation or leaf litter. According to [36,37], the upper 5–6 cm of the soil layer is primarily involved in germination and emergence, while for some species, the emergence rate is absent or negligible beyond 10 cm. The burial site was located in the experimental field of INTA EEA Ascasubi. Seeds were exhumed after 84, 174, 285, 414, 506, 644, 736, 811, and 1078 days of burial.

The proportion of physically dormant (PY) seeds was assessed before burial (initial) and at different exhumation times. PY seeds (i.e., “hard” or impermeable) were identified using an imbibition test performed at 20 ± 2 °C for 2 days [38]. Swollen and germinating seeds were recorded and removed, while the remaining hard seeds were returned to the mesh pouches in the field. This process was repeated to generate cumulative hard seed breakdown curves [39].

The rate of PY dormancy loss between exhumation periods could be attributed to seed germination or mortality. Therefore, a second trial was conducted to quantify the mortality rate of RS seeds. The difference between the dormancy loss rate and the mortality rate was attributed to the seed germination rate.

Seed bank mortality was assessed by burying seeds in impermeable nylon bags (102 microns) during autumn 2019, following natural seed dispersal. Batches of 200 seeds (n = 3) were placed inside each bag. Seeds were exhumed after 83, 171, 407, 412, 503, 639, 731, and 806 days of burial. Exhumed seeds were mechanically scarified with sandpaper to remove part of the seed coat and then placed on imbibed filter paper in a growth chamber at a constant temperature of 25 °C for fifteen days. Finally, any remaining ungerminated seeds were dissected under a microscope, and those with firm white embryos were considered viable [40]. The proportion of seedbank mortality, corresponding to dead seeds without the presence of bruchids, was calculated relative to the total number of seeds exhumed on each date.

The seed germination potential was estimated as the proportion of non-PY seeds (1 − PY seeds) that remained viable over time. This was determined by calculating the cumulative proportion of seeds that had released dormancy and adjusting for cumulative seed mortality.

2.2.6. Experiment 6: Field Emergence

The experiment was conducted in an undisturbed field with a high population of RS during the 2019/20, 2020/21, and 2021/22 growing seasons, from the 21st of September to the 1st of May. RS seedlings with the first pair of leaves fully expanded were counted every 10 days in four permanent quadrats (0.25 m²) randomly distributed across the field. The seedlings were carefully removed with minimal disturbance to the soil [35]. Weed control was performed manually by hand. At the end of each year of experimentation, the quadrats were removed and randomly redistributed within the remaining area of the 8000 m² experimental field. Thus, seedling emergence was not affected by seedbank depletion.

2.2.7. Experiment 7: Plant Survivorship in Summer and Winter

The survival of RS seedlings during the summer and the persistence of young plants (<1 year-old) after the winter period were evaluated over two growing seasons in the experimental field. Seeds were sown in early spring (September 21st) at a density of 3200 seeds.m² in permanent plots. In each growing season, more than 20 seedlings (n = 3) were tagged in each cohort (ES, LS, and A). The survival of marked seedlings from each cohort was quantified in late summer and early spring of the second year, after the winter period. Seedlings during the summer were considered dead when all above-ground parts lost color, showed no elasticity, and did not recover after precipitation. Winter survival was easily observed by the regrowth of the marked plants. Percent survival was estimated as the ratio of the total number of surviving plants to the initial number of marked seedlings.

2.3. Statistical Analysis

A linear mixed-effects model (LMM) was used to evaluate multiple parameters, using InfoStat software (version 2020e). In the case of fecundity (Experiment 1), the effects of three seasonal cohorts (early summer, late summer, autumn) during the growing season and plant age (young and adult plants) were assessed, with cohorts and plant age as fixed effects, including their interaction (cohorts × plant age). Bruchus predation (Experiment 2) was analyzed with cohorts as a fixed effect. Seed dispersal (Experiment 3) was modelled with cohorts and distance from the source plant (0, 1, 2, 3, 4 m) as fixed effects, including their interaction. Post-dispersal predation (Experiment 4) was assessed by incorporating treatments (without stubble cover, with stubble cover, and cage) and cohorts as fixed effects, enabling the assessment of how different environmental conditions influenced seed predation dynamics. Seedling emergence (Experiment 6) was analyzed with cohorts as a fixed effect, while plant survivorship (Experiment 7) included cohorts and periods (summer and winter) as fixed effects, considering their interaction to determine whether survival rates fluctuated across different seasonal. To account for inter-annual variability, the growing season was incorporated as a random factor in Experiments. 1, 2, 3, 4, 6, and 7, ensuring that year-to-year variations were captured in the analysis. The model was fitted using the restricted maximum likelihood (REML) method with the lme function from the nlme package in R (version 4.3.1). The significance of fixed effects was determined using sequential hypothesis testing, and post-hoc comparisons were conducted at α = 0.05 using the least significant difference (LSD) test.

Additionally, linear and nonlinear regression analyses were performed in Experiments 1, 3, 5, and 6 using GraphPad Prism Software (version 8.0) to further explore relationships between variables.

3. Results

3.1. Fecundity

The results indicate that both cohorts (p < 0.01) and plant age (p < 0.01) influenced fecundity, with a marked reduction observed in autumn and young adult plants (Figure 1). A significant interaction between these factors suggests that the effect of plant age is not consistent across cohorts. In particular, while fecundity is generally higher in summer, younger plants experience a more pronounced reduction compared to adult plants (Figure 1). Additionally, the inclusion of the year as a random effect accounted for inter-annual variability, which was moderate (σ²_year = 34.6%). These findings highlight the importance of considering cohorts and plant development stage when assessing fecundity dynamics in RS.

3.2. Pre-Dispersal Seed Predation

Bruchus predation does not vary across cohorts (F = 1.45; p > 0.05), and inter-annual variability was negligible (σ²_year ≤ 1%). The estimated mean predation rates were similar between cohorts: early summer (16%), autumn (13%), and late summer (7%) (p > 0.05).

3.3. Seed Rain Dispersal Distance

Seed dispersal decreased with increasing distance from the source plant in both adult and young adult individuals (p < 0.01). The cohorts had no effect on seed dispersal (p > 0.05), nor did the interaction between cohorts and distance (p > 0.05) (Figure 2). Additionally, the random effect of year (σ²_year ≤ 1%) indicated that variability in seed dispersal did not significantly impact the results, suggesting that the overall trends observed were consistent across years. The highest probability of seed dispersal was observed near the mother plant, which progressively decreased with increasing distance. These results suggest that dispersal is strongly distance-dependent but not influenced by cohort variation or inter-annual variability.

3.4. Post-Dispersal Seed Predation

Predation rates (p) were significantly affected by treatment types and cohorts. Specifically, predation was highest without stubble cover (71%), compared to stubble cover (58%) and cage treatment (42%) (p < 0.01) (Figure 3). Additionally, predation varied significantly among cohorts, with the highest rates observed in early summer (76%), followed by late summer (58%) and autumn (38%). No significant differences between treatments and cohort interaction were observed. The inclusion of the random effect of year did not significantly influence the overall predation rates.

3.5. Seed Bank Behavior

After burial, the proportion of non-PY dormant seeds in the soil seedbank was well described by an exponential model (Figure 4a). Total seed mortality increased linearly with burial time (Figure 4a). Germination potential increased up to ~0.5 after 290 days of burial (Figure 4b), coinciding with the emergence window observed in RS under field conditions (November to April). During the second window, germination potential slightly declined to ~0.40, and it was further reduced in the third growing season (~0.10).

3.6. Field Emergence Experiment

The results showed differences in the proportion of seedling emergence between cohorts (p < 0.01). The highest emergence occurred in ES (75%), followed by LS (20%) and A (5%). These results indicate a strong seasonal effect on seedling emergence, with ES being the most favorable period (Figure 5). The random effect of the year showed minimal variance, suggesting a consistent seasonal trend across years.

3.7. Survivorship During Summer and Winter

Seedling survival during summer was not influenced by cohort occurrence (p = 0.76) (

Table 1. Plant survival rate of Rhynchosia senna in summer and winter periods for each cohort. Early summer (E.S), late summer (L.S) and autumn (A). Letters indicate significant differences. Inter-annual variability (σ²_year). “*” show significant differences at the level p < 0.05, “**” show significant differences at the level p < 0.01. “ns” represents no significant differences.

Plant Survival	Cohorts			p-Value	σ2year (%)
	E.S	L.S	A
summer period	0.82	0.64	0.84	ns	14.5
winter period	0.98 a	0.74 b	0.32 c	**
p-value	*	ns	**
Cohort*periods		**

). In contrast, significant differences in plant survival were observed during winter, with higher rates in ES compared to A. Lower winter survival was observed for seedlings that emerged later in the growing season. The random effect of year (σ²_year) suggests moderate variability in the survival rates of RS seedlings across different years (Table 1). This indicates that year-specific conditions had an important effect on seedling survival.

4. Discussion

Our findings provide a valuable insight into the general demographic parameters of RS. We observed that adult plants (>1 year old) of RS exhibit a high fecundity (~6000 seeds per plant), with peak production in early and late summer (Figure 1). The seed production levels of RS are similar to those reported for Lespedeza cuneata, a leguminous species native to Asia that was introduced to the United States and has since become an invasive species. L. cuneata is known for its high seed output and the ability to establish persistent seed banks, contributing to its invasive success [41]. Conversely, the fecundity of young adult plants (<1 year old) is 96% lower than during the second year. Therefore, the high seed production of adult plants is expected to contribute to the species’ persistence under natural conditions.

High predation rates were observed during pre- and post-dispersal. Pre-dispersal predation averaged 12% of the seeds and did not vary throughout the growing season. These values are similar to those recorded for other legumes, such as Vicia angustifolia [42] and Astragalus lehmannianus [43], and are lower than those reported for other perennial legume species, such as Neltuma sp., which coexists sympatrically with RS, and Leucaena leucocephala [44,45,46]. Seed beetles, also known as bruchid beetles (Chrysomelidae: Bruchidae), are particularly common predators and parasites of legume seeds in dryland environments, and many exhibit high host specificity [45,47]. In this study, it remains to be determined which species are involved.

In addition, predation by birds (>55%) and arthropods (~40%) is the main factor limiting the formation of a dense soil seedbank, acting as a significant demographic hurdle. The highest predation rate was observed in summer, coinciding with the period of maximum fecundity. In the field experiment, adult feral pigeons (Columba livia) were identified as the primary avian predators of RS seeds. Consistent with Renzi [33], stubble cover was observed to reduce seed predation but did not entirely prevent it. Arthropods, particularly ants (Acromyrmex lobicornis) were identified as the primary invertebrate seed predators. These predation pressures can significantly influence population dynamics and the structure of weed communities [48].

Seedbanks play a crucial role in stabilizing population dynamics by spreading risk and enabling population recovery following disturbances [49]. Based on our results, we infer that RS exhibits a type III persistent seedbank sensu [50]. This type of seedbank has also been observed in other legumes, such as Vicia sp. and Neltuma sp. [32,51]. In a type III seedbank, many seeds germinate in the first growing season after dispersal, while a small reserve of viable seeds remains ungerminated. The pattern of physical dormancy (PY) release in RS exhibited a distinct trend during the first year, with approximately 80% of the dormancy loss. The remaining 20% did so in the following two years. This trend was effectively described by an exponential model. The highest germination potential, based on Non-PY and viable seeds (1 − dead seeds), occurred during the species’ emergence window in summer. In the second and third years, it was likely reduced due to an increase in seed mortality and a decline in seedbank persistence (Figure 4). The number of established plants in wild species is strongly influenced by the proportion of the seedbank that has exited dormancy [52].

The cumulative emergence of RS followed a logistic model (Figure 5), with high emergence occurring in early and late summer. The emergence pattern was protracted, spanning at least five months during the growing season, with a peak observed in ES. In RS, germination often began a few days after rainfall events. Different water potentials and environmental conditions also had a significant influence on seed germination and seedling emergence [53]. However, in unpredictable environments such as the semiarid Espinal region, germination represents a high-risk strategy, particularly when rainfall is sufficient to trigger germination but insufficient to support successful seedling establishment [54]. These conditions often lead to high summer seedling mortality, as precipitation may initiate germination but fail to sustain the necessary conditions for seedling survival. The summer survival rate was independent of cohorts, whereas winter survival depended on cohorts, being low for autumn (32%), possibly due to the smaller plant size during winter dormancy (Table 1). Autumn cohorts, exposed to early frosts without adequate development and sufficient reserves, exhibit reduced winter survival and limited spring regrowth. The ability of perennial legumes to withstand winter conditions and resume growth in spring depends on the availability of endogenous carbon and nitrogen reserves stored in vegetative organs during the cold acclimation period in the preceding autumn [55].

Despite the high fecundity of RS, which could confer an invasive potential, the establishment success of its progeny is likely reduced by 12% due to pre-dispersal seed predation, 57% due to post-dispersal predation, 24% due to summer plant mortality, and 14% due to winter. On average, considering seedbank dynamics, less than 13% of the seeds would reach the adult plant stage in the first year. However, the post-dispersal predation values in this study may be overestimated, as the trial may have experienced higher predation pressure due to the absence of alternative seed sources and the lack of vegetation cover, which might have otherwise mitigated predation [56].

The distance dispersion was variable depending on the age of the plant because young adult plants have a smaller height (0.15–0.25 m) than adult plants (0.30–0.50 m), leading to a shorter seed dispersal [57]. Additionally, some authors [58,59] propose that pod dehiscence in some legume species is predominantly regulated by the structural characteristics of pod tissues. Consequently, no interannual variation is observed [60]. Nonetheless, dispersal distances could be considered short for both developmental stages.

5. Conclusions

Both indigenous and alien species have the potential to alter habitats and ecological processes, which may lead to ecological and economic damage [61]. Rhynchosia senna var. senna does not appear to be a potentially invasive species in its natural environment. This is primarily due to a high level of seed predation, reduced seed viability and rapid dormancy release, leading to a low persistent seed bank (<3 years). In addition, the restricted spatial dispersal of seeds (<4 m) is unlikely to promote invasiveness.

Conversely, the environmental conditions of the study site highlight SR strategy as a within-season bet-hedging strategy, distributing reproductive effort across the growing season to enhance the likelihood of seedling survival under unpredictable conditions [44]. Rather than maximizing fecundity and emergence at once, this strategy buffers against environmental variability by staggering germination events. In more favorable environments, where abiotic stress is lower and resource availability is higher, the establishment success and long-term persistence of the species could be significantly greater, potentially increasing its competitiveness and transforming it into a problematic weed. In addition, in exotic environments where predator pressure is lower, RS could become problematic due to its high fecundity and seedling establishment rates. Globally, the most threatening weed species are alien, highlighting biological invasions as a major driver of environmental change, with the potential to cause both ecological and economic losses.

Future studies should investigate secondary dispersal by birds and rodents to better understand its impact on RS population dynamics. Additionally, population modelling and its validation will help elucidate and predict the species’ behavior in both its native and non-native environments.