Introduction

Carpobrotus edulis (Aizoaceae) is a perennial herb native to South Africa that has been introduced to numerous temperate coastal dunes worldwide for ornamental and soil fixing purposes (Portela et al. 2023). This resilient, self-fertile plant showcases remarkable adaptability to a range of challenging conditions, including sandy soils, saline environments, intense solar radiation, and other environmental stressors (Campoy et al. 2018). It also displays extensive vegetative growth, enabling efficient colonization by its propagules (Souza-Alonso and González 2017; Roiloa 2019).

Carpobrotus edulis exhibits a highly invasive behavior, forming dense mats that significantly disrupt various components of dune ecosystems. Noteworthy consequences include impeding the regeneration of native species and inducing alterations in soil pH and organic matter levels (Portela et al. 2023). The presence of this species in invaded soils has been associated with diminished germination, survival, growth, and reproductive success of native plant species (Conser and Connor 2009).

In Uruguay, the date and location of introduction of C. edulis remains unknown, and it is absent from early floristic reference lists for the country (e.g., Herter 1930; Legrand 1959). However, historical accounts suggest its presence in Punta del Este (Maldonado) around 75 years ago (E. Marchesi, pers. comm.). Partial surveys of coastal vegetation have mentioned its occurrence since the 2000s (Delfino and Masciadri 2005; Castineira et al. 2013; Mai et al. 2022). In fact, local private and public stakeholders introduced Carpobrotus edulis for coastal resort ornamentation and dune stabilization (e.g., Puppo et al. 2020). Furthermore, due to its exceptional tolerance, its use in "green roof" installations is promoted (Rosatto et al. 2015), creating novel pathways for its expansion.

Despite the recognized risks, the majority of invasive species in Uruguay, including Carpobrotus edulis, lack comprehensive assessments of their extent (MVOTMA 2016). The lack of records in national herbarium collections hinders the study of its distribution and abundance. Only three records of the species have been collected in the Rocha Department (i.e., administrative regions in the country), in the localities of Cabo Polonio and La Pedrera (F. Haretche, pers. comm.; E. Marchesi, pers. comm.; M. Valdivia, pers. comm.). Given the vulnerable state of the coastal dune system in Uruguay, threatened by urbanization, climate change, and non-native species afforestation (Panario and Gutiérrez 2005), understanding the invasion status is crucial for effective management. This study aims to report the extent of C. edulis invasion in Uruguay including, besides its geographic distribution, the density patterns, using data gathered through citizen science efforts. We also aim to assess the number of records per year of the species, its phenological patterns and the correlation of its occurrence with the levels of anthropization.

Methods

Data download

We used data sourced from iNaturalist (https://www.inaturalist.org/), a comprehensive global platform for documenting species observations. This popular citizen science platform enables people to contribute valuable data by uploading photographs and audio recordings captured using their mobile devices, cameras or tablets. Records undergo a rigorous community-based identification process. An observation is classified as 'Research Grade' when it includes location, date, media evidence, and the community agrees on species-level identification or lower, i.e. when more than 2/3 of identifiers agree on a taxon. iNaturalist has become a robust and influential platform for biodiversity monitoring on a global scale, gaining significant recognition and widespread adoption within the scientific community (Hochmair et al. 2020; Aristeidou et al. 2021). Notably, in the case of Uruguay, the coastal zone where C. edulis occurs, has been extensively sampled by citizen efforts (Grattarola et al. 2020), offering valuable data for studying the invasion dynamics of this species. Acknowledging that Carpobrotus’ taxonomy is problematic (Novoa et al. 2023), we chose to work with C. edulis as this is the main identified species in the country. Only two records have been identified as C. chilensis, yet they have not been confirmed. Research grade records of Carpobrotus edulis in Uruguay were downloaded on 2023-05-10 through the iNaturalist API (https://api.inaturalist.org/v1/docs/) using R (R Core Team 2023). The code and data can be accessed at: https://github.com/bienflorencia/carpobrotus-uruguay.

Spatial and temporal analysis

We quantified the number of C. edulis records per year, department and season. To assess the accumulation of records in space, we divided the coastal zone into grid cells of 10 × 10 km and downloaded all the records identified to the species level available on iNaturalist for this area. To quantify the level of sampling effort for the entire database and compare it to the C. edulis numbers, we used the curvilinearity of smoothed species accumulation curves at the grid cell level (following Grattarola et al. 2020). Given extensively surveyed regions predominantly align with the vicinities of urbanized or protected areas, we used a corrected Pearson’s correlation for spatial autocorrelation (following Grattarola et al. 2020), to first assess the spatial association between the total number of records in the database and the number of records of C. edulis per grid cell and then between the number of total human population and the number of records of C. edulis per grid cell. All maps were prepared using WGS 84/UTM zone 21S projection.

Image categorization

We analyzed the photographic evidence of each record. First, we classified the records based on the density of C. edulis plants, considering only the photos taken at a minimum distance of approximately 1 m. We distinguished between low density, characterized by isolated individuals; medium density, where the species formed spaced branches; high density, with dense mats and overlapping branches; and unable to detect, when the photo did not meet the distance criterion. Then, we examined all the photos again (regardless of the minimum distance) to identify reproductive structures such as buds, flowers, and fruits. To classify the records, we considered the categories established on iNaturalist: plants with visible but not open flower buds (flower budding), plants with visible flowers (flowering), plants with fruit (fruiting), combinations of these stages (flowering/flower budding and flowering/fruiting) or records that did not provide any evidence of reproductive structures (no evidence of flowering/fruiting). Buds were recognized by their small and closed appearance, while flowers displayed vibrant yellow, white or pink petals. Fruits, resembling “figs,” were identified by their ellipsoid shape and fleshy appearance. Finally, for each iNaturalist record, we examined the presence of urban structures such as streets, curbs, central flowerbeds, or parking lots on the sandy habitat. We categorized these observations using a binary criterion: "absent" for records that lacked any evidence and “present” for records that indicated the presence of the aforementioned infrastructure within a 5-m distance radius.

Results

We observed 204 records of C. edulis in Uruguay, made by 78 users from 2008-03-16 to 2023-05-10. The species occurs within 10 km of the coastline, occupying 22 grid cells, equivalent to ca. 220 km of invaded coast. The invaded area is associated with coastal dune ecosystems (Fig. 1a) and covers a portion of the Atlantic and the outer Rio de la Plata coasts. The records spanned four departments: Montevideo (12 records), Canelones (46), Maldonado (98), and Rocha (48) (Fig. 1b). Nineteen records (9.3% of the total) were reported within Laguna Garzón and Cabo Polonio protected areas (Fig. 1b, protected areas shown in green). We did not observe records of C. edulis in Humedales del Santa Lucía, Isla de Flores, Laguna de Rocha, or Cerro Verde, the other coastal protected areas in Uruguay.

Fig. 1
figure 1

Records of Carpobrotus edulis in Uruguay downloaded from the iNaturalist platform. a Photo of the species forming a carpet over the coastal dunes in Maldonado Department (laurillac, January 25, 2023, iNaturalist observation, http://www.naturalista.uy/observations/147333000). b Point occurrences (black dots) of the species in Uruguay (dark grey polygons delimit the departments of Montevideo, Canelones, Maldonado and Rocha, from west to east). Protected coastal areas are shown in green and urbanized areas, with an area of more than 5 km2, are shown in blue. c Number of records per year showing a steep increase over the last three years

Full size image

Records of C. edulis have increased strongly in the last three years, following the increment of plant records in the platform (Fig. 1c). Given the species presence has been poorly reported in the last 75 years (e.g., in national herbariums), we are unaware of the moment when the expansion occurred.

Spatial distribution

The coastline of Uruguay has been unevenly sampled on iNaturalist (Fig. 2a). The spatial pattern of the national sampling effort (i.e., the total number of records) was significantly positively correlated with that of the number of C. edulis records (r = 0.665, p < 0.001). We observed that the western part of the C. edulis distribution (continuous along the coast) is located within areas of larger sampling efforts, while the gaps observed in Rocha correspond to less sampled areas (Fig. 2a, b). However, this is not the case in Montevideo, an area with abundant sampling on iNaturalist but with a low number of records of C. edulis. This is consistent with finding very low levels of spatial correlation between the number of people living in the area and the number of C. edulis observations recorded (r = 0.299, p < 0.05). In this regard, we must consider that a significant portion of the dune system in the Montevideo Department has been removed or replaced (Gudynas 2000).

Fig. 2
figure 2

Spatial analysis of the records of Carpobrotus edulis in Uruguay downloaded from iNaturalist. a National sampling effort per grid cell, calculated as the degree of curvilinearity (mean slope) of the last 10% of the species accumulation curve for all records on iNaturalist. Poorly sampled grids are shown in orange while better-sampled ones are shown in blue. b Accumulation of records of Carpobrotus edulis per grid cell in the departments of Montevideo, Canelones, Maldonado and Rocha is shown with a green scale (i.e., dark green grids represent higher accumulation). Grid cells are 10 × 10 km in size. Grid cells with no data are shown in dark grey

Full size image

Carpobrotus density and human infrastructure

Most of the records showed a medium (32.8%, N = 67) or high (36.8%, N = 75) plant density. Only 14.2% (N = 29) of the observations presented low density. The rest of the records (16.2%, N = 33) could not be evaluated as they did not meet the minimum distance criteria. Infrastructure near the C. edulis plants was observed in 22.5% (N = 46) of the records. The majority of observations had no infrastructure (58.8%, N = 120) and 18.6% (N = 38) could not be assessed because they did not meet the minimum distance criterion.

Phenology

The species has been reported across all seasons, but mainly in the summer (52.9% of the total) when more people circulate through the shores and dunes (Fig. 3). More than half of the C. edulis records did not show evidence of flowering or fruiting (57.4%). The rest of the records were divided into plants with flowering buds (1.5%), flowers (13.2%), or visible fruit (20.6%), and combinations of these phenological stages (5.4% of records were flowering and flower budding and 1.9% were flowering and fruiting). Carpobrotus edulis exhibited flowering primarily during the spring months, followed by fruiting in the summer. We also observed a smaller peak in flowering during the autumn, with fruits starting to appear in winter and increasing towards spring. This secondary peak in flowering during autumn suggests a potential extended reproductive period for C. edulis, which may contribute to its successful establishment and spread (Fig. 3).

Fig. 3
figure 3

Phenological state of the records of Carpobrotus edulis in Uruguay is shown as the number of records per month. Plants with visible but not open flower buds (flower budding), plants with visible flowers (flowering), plants with fruiting (fruiting), combinations of these stages (lowering/flower budding and flowering/fruiting) or records that did not provide any evidence of reproductive structures (no evidence of flowering/fruiting). The number of records per season is shown in the top right. Summer (January, February, and March), autumn (April, May, and June), winter (July, August, and September), and spring (October, November, and December)

Full size image

Discussion

Carpobrotus edulis is widely distributed along the coastal dune belt of Uruguay, encompassing the Rio de la Plata estuary (Montevideo, Canelones, and parts of Maldonado Department) as well as the oceanic coast (parts of Maldonado and Rocha Departments). It is concerning that this species is invading two coastal protected areas, especially considering the ecological value of coastal dunes, which are objects of conservation (MVOTMA 2016). The process of real estate development and urbanization has led to the removal of significant patches of dune vegetation, while the afforestation with pines (Pinus spp.) has caused further loss of these fragile ecosystems (Puppo et al. 2020). Our findings suggest that Carpobrotus is strongly invading the remaining coastal dune vegetation. A similar process is occurring with Acacia longifolia, which is a highly prevalent non-native species in the dune ecosystems of Uruguay (Castiñeira et al. 2013). This alarming situation requires immediate attention and urgent actions, as it could lead to irreversible losses of important native biodiversity (Puppo et al. 2020). This invasion may affect native communities of the Uruguayan coast by replacing the coastal vegetation (Mai et al. 2022). For instance, Altuna et al. (1999) suggested that the invasion of Carpobrotus can lead to reduced availability of plants that serve as food resources for endemic coastal fauna.

The observed rise in Carpobrotus observations since 2019 is likely attributable to the increasing use of the iNaturalist app (Grattarola et al. 2020). The popularity of iNaturalist as a citizen science platform has greatly contributed to valuable species occurrence data collection, underscoring the importance of utilizing citizen science initiatives for monitoring and studying invasive species dynamics (Dimson et al. 2023). However, our analysis of iNaturalist records reveals that there are still areas with low or no occurrences, particularly in the eastern coast (i.e. Rocha Department). Further exploration is needed in these and other less sampled regions to assess the presence of Carpobrotus. If absent, preventive measures should be implemented to avoid its introduction.

Despite we did not detect a strong association between the presence and density of C. edulis and total human population or human infrastructure, the species was abundant in urbanized areas (except for Montevideo, see blue polygons in Fig. 1b). Local governments commonly include C. edulis in their ornamental planting plans and may even provide artificial irrigation to ensure its growth (MZ, LRT and GL personal observations). These practices are counterproductive to Uruguay's national biodiversity strategy (MVOTMA 2016), and we strongly advocate for their urgent reassessment. Given the current awareness of the risks associated with biological invasions (Xu et al. 2014), the Uruguayan government should critically review all activities carried out by national institutions that promote, directly or indirectly, the introduction and expansion of invasive species. Furthermore, despite the challenges posed by its removal and prevention of recolonization, as well as the significant soil alterations it causes, it is imperative to implement eradication and control of C. edulis in coastal dune ecosystems (Souza-Alonso and González 2017; Campoy et al. 2018; Roiloa et al. 2022). Along with this, the actions of the local authorities must be adjusted to stop using invasive species in public ornamentation and use instead native herbaceous plants that provide the same services that are attributed to C. edulis.

The extensive engagement of people in coastal areas, where individuals spend considerable amounts of time, has facilitated robust data collection surpassing that found in scientific literature and collections. The information available in scientific collections tends to be biased, with little attention to non-native species (Delisle et al. 2003). Moving forward, we recommend prioritizing sampling efforts in underrepresented regions, ensuring comprehensive coverage and filling data gaps. We also encourage iNaturalist users to capture reproductive structures in the photos, and to include both close-up and wider-range images to assess C. edulis local abundance and extension for a more comprehensive understanding of its invasion dynamics. While acknowledging the potential biases associated with citizen science data, iNaturalist remains a powerful tool for studying invasive species, and its continued use and engagement within the academic community and park rangers are vital for advancing our knowledge in this field (Pyšek et al. 2020; Hausdorf 2023). Finally, we also want to emphasize the need to survey and monitor the threats to the coastal dunes, a very valuable and severely threatened ecosystem in Uruguay. Citizen science can help generate social awareness of the problem.