Practical Outcomes of the First Year of IAPS Monitoring of the LIFE medCLIFFS LmVN

The regions of the Costa Brava with the highest ratio of observations per transect were CS and C, whereas S, CN, and N exhibited comparatively lower ratios (Table 2). The heat map in Figure 3 illustrates the ratio of observations per transect for the nine alien species recorded across the five Costa Brava sampling regions. Overall, this diagram reveals that nearly all species are present in all regions, with the exception of C. nutans, O. ficus-indica, and S. angulatus. Based on this ratio, three regions (CS, C, and N) contain species with more than three observations per transect (Figure 3).

Table 2. Number of transects, observations, and the rate of observations per transect for each region and the total count for the study area.

Figure 3. Heat map of the rate of observations per transect for species in each Costa Brava region: S, Costa Brava South; CS, Costa Brava Center-South; C, Costa Brava Center; CN, Costa Brava Center-North; N, Costa Brava North. The rightmost column contains the values for the Costa Brava as a whole.

It should be noted that while the N region, encompassing the Cap de Creus Natural Park, had the highest number of monitored transects, it did not yield a correspondingly high number of observations. The observed ratios in this region were generally low, except for Carpobrotus spp. and O. ficus-indica (2.44 and 3.88, respectively), both of which are targeted for control measures within the Cap de Creus Natural Park under the LIFE medCLIFFS project. These results indicate that the selection of species for control measures within the project has been appropriate. This is the first time that Opuntia spp. (including O. ficus-indica but also erect prickly pear [O. stricta (Haw.) Haw.] and desert prickly pear [O. engelmannii Salm-Dyck ex Engelm.]) will be managed for control purposes using chemical and biological methods, while control programs (manual/mechanical) aimed at limiting Carpobrotus spp. populations within the Natural Park have been ongoing since 2009 (Gómez et al. Reference Gómez, Prunell, Sabaté and Salvadó2010).

Surprisingly, the most problematic species according to Figure 3 is P. tobira (with a ratio of nearly three), making it the most abundant species in all regions except N. Despite this, P. tobira is not listed as invasive in any of the official catalogs relevant to the Costa Brava (i.e., neither in the Spanish invasive alien species regulation [METDC 2024] nor in the European one [EC 2022]).

As can be seen in Figure 4A, the vast majority of the species are showing a trend that can be cautiously defined as of certain “demographic growth,” as the category with the highest percentage is mixed behavior, followed by propagative behavior and, with much lower percentages, by senescent behavior (Table 3). The only species showing a considerable proportion of senescent populations (which is larger than that of propagative ones) are O. ficus-indica (propagative: 15%; mixed: 53%; senescent: 33%) and Carpobrotus spp. (propagative: 18%; mixed: 62%; senescent: 19%). In the case of O. ficus-indica, the high mortality rates could be linked to the fact that the species has been infested by Dactylopius opuntiae in recent years, and in the N region, this cochineal has been used for the biological control of this cacti since 2023 (Figure 5A). In contrast, we believe that the relatively high percentage of senescent populations of Carpobrotus spp. may be due to the extremely high temperatures experienced in the last two summers in the study area, which withered large parts of the individuals (Figure 5B). The relatively high percentage of senescent populations observed in A. americana (20%) is likely due to a pest, as in the case of O. ficus-indica. The agave weevil (Scyphophorus acupunctatus) has been severely affecting this species in the Costa Brava (and throughout much of the Iberian Peninsula) since at least the 2010s (CG-B and JL-P, personal observations; Figure 5C).

Figure 4. Heat map with the proportion of species observations within the three categories defined, for the Costa Brava as a whole and for each of the five defined regions. Numbers inside each colored box represent the number of associated observations.

Table 3. Proportions for each of the categories of population dynamics considered (propagative behavior, mixed behavior, and senescent behavior) for the nine monitored plant species in the Costa Brava.

Figure 5. Examples of senescent observations. (A) Opuntia ficus-indica intentionally infected by Dactylopius opuntiae as part of the eradication measures implemented in the LIFE medCLIFFS project. Photo: AB-G. (B) Live Carpobrotus spp. with a withered part. Photo: AB-G. (C) Agave americana infected Scyphophorus acupunctatus. Photo: JL-P.

Finally, it is worth noting that there are two cases where the percentage of populations with propagative behavior is higher than that of populations with mixed behavior (C. nutans and S. angulatus). Notably, these two species are the only ones introduced in the Costa Brava after 1970 (Gómez-Bellver Reference Gómez-Bellver2023), and one could therefore hypothesize that their population dynamics are related to the condition of recent neophytes. These two species are likely still in their early stages of invasion, when higher rates of population growth are necessary for a successful establishment (Sakai et al. Reference Sakai, Allendorf, Holt, Lodge, Molofsky, With, Baughman, Cabin, Cohen, Ellstrand, McCauley, O’Neil, Parker, Thompson and Weller2001), and have not yet reached equilibrium with their environment (e.g., Foster et al. Reference Foster, Kharouba and Smith2022).

Just as the results have been analyzed and discussed for the heat map of the entire Costa Brava, the same analysis can be extended to more restricted areas, as the regions delimited in this study (Figure 4B–F), or even at the level of individual transects if necessary. This will depend on the manager’s needs for identifying problematic zones for specific species.

Although the results discussed in this article represent the initial findings from the first year of monitoring carried out by the LmVN, it is expected that the knowledge of IAPS in the region—and the management derived from it—will substantially improve in the coming years of the project. In conclusion, the presented work introduces a novel methodological approach to simplify and present large amounts of citizen science data intended to improve the efficiency of management decisions.

Advantages and Limitations of Our Method to Detect and Monitor Alien Plant Species with Invasive Potential

Biological invasions are among the most pressing ecological and conservation issues nowadays. In addition to raising public awareness of environmental concerns and reducing labor costs, citizen science projects have become essential tools for the detection and management of many exotic species (Crall et al. Reference Crall, Newman, Jarnevich, Stohlgren, Waller and Graham2010; Dickinson et al. Reference Dickinson, Shirk, Bonter, Bonney, Crain, Martin, Phillips and Purcell2012; Encarnação et al. Reference Encarnação, Teodósio and Morais2021; Gallo and Waitt Reference Gallo and Waitt2011; Larson et al. Reference Larson, Graham, Achury, Coon, Daniels, Gambrell, Jonasen, King, LaRacuente, Perrin-Stowe, Reed, Rice, Ruzi, Thairu and Wilson2020; Wallace et al. Reference Wallace, Bargeron, LaForest, Carroll, Pullaiah and Ielmini2021) in different regions and at varying scales. Indeed, the trade-off between cost and effectiveness of citizen science data offers the potential for scientists to tackle research questions at large spatial and/or temporal scales (Belt and Krausman Reference Belt and Krausman2012; Brossard et al. Reference Brossard, Lewenstein and Bonney2005; Holck Reference Holck2007; Levrel et al. Reference Levrel, Fontaine, Henry, Jiguet, Julliard, Kerbiriou and Couvet2010; Szabo et al. Reference Szabo, Vesk, Baxter and Possingham2010). Despite this, part of the scientific community remains concerned about the accuracy of citizen science data, as it has been reported that the quality of these data has historically been more variable than professionally collected data (Belt and Krausman Reference Belt and Krausman2012; Harvey et al. Reference Harvey, Fletcher and Shortis2002; Moyer-Horner et al. Reference Moyer-Horner, Smith and Belt2012; Uychiaoco et al. Reference Uychiaoco, Arceo, Green, Margarita, Gaite and Aliño2005). However, and as will be discussed further, there is evidence that citizen performance can be comparable to that of professionals or scientists (Canfield et al. Reference Canfield, Brown, Bachmann and Hoyer2002; Hoyer et al. Reference Hoyer, Winn and Canfield2001, Reference Hoyer, Wellendorf, Frydenborg, Bartlett and Canfield2012; Oldekop et al. Reference Oldekop, Bebbington, Berdel, Truelove, Wiersberg and Preziosi2011), provided that they receive proper supervision, volunteer training, and feedback, among other recommendations (Aceves-Bueno et al. Reference Aceves-Bueno, Adeleye, Feraud, Huang, Tao, Yang and Anderson2017; López-Guillén et al. Reference López-Guillén, Herrera, Bensid, Gómez-Bellver, Ibáñez, Jiménez-Mejías, Mairal, Mena-García, Nualart, Utjés-Mascó and López-Pujol2024).

In the specific case of the LIFE medCLIFFS project, the volunteer network was complemented by experts who provided supervision, references, training, and acknowledgment. Additionally, the mandatory fields were established as categorical factors to be selected within a form setting on iNaturalist, using a multiple-choice format that has two primary repercussions. The first is the simplicity of the data produced, which is based on the presence or absence of features and does not take into consideration the abundances linked to each developmental stage of each population. The second one is the reduction in the difficulty of sampling for citizen volunteers, as well as the acquisition of a more consistent, homogeneous, and manageable database for researchers and managers. In contrast to projects focused on sampling endangered species or populations (Soroye et al. Reference Soroye, Edwards, Buxton, Ethier, Frempong-Manso, Keefe, Berberi, Roach-Krajewski, Binley, Vincent, Lin, Cooke and Bennett2022), the level of detail of data on invasive alien species generally does not have legal implications (although there would be legal effects if data are used as a basis for listing these species as invasive and prohibiting their sale). Moreover, the sampling procedure does not provide risks to the species that are being investigated, which makes the task of designing the sampling and collecting data for citizen volunteers much easier.

Despite the simplicity of using categorical choices for most features that volunteers must record, our citizen science project (and the underlying methodology) is one of the few that goes beyond merely detecting new locations of alien plants, as is the case with most of the abovementioned citizen science projects. Until recently, most of the citizen science projects on invasive species were exclusively focused on reporting their presence. Crall et al. (Reference Crall, Newman, Jarnevich, Stohlgren, Waller and Graham2010) surveyed 128 citizen science programs of alien species of the United States, and while 95% of them collected data on species presence, only one-fifth recorded data on life stages. Although presence-focused projects are valuable, they typically do not provide outputs with direct control or management implications. Fortunately, there has been an emergence of citizen science projects in recent years that are specifically designed to provide data with direct applications on control and management measures. For example, the citizens using the BambApp recorded a series of traits (maximum height, maximum stem diameter, and size of the clump, as well as habitat type) aimed at helping public authorities to reevaluate the invasive potential of bamboo species (i.e., those included within the subfamily Bambusoideae) and regulate their commerce in northwest Italy (Pittarello et al. Reference Pittarello, Ravetto Enri, Nota, Lombardi, Mezzasalma, Frigerio and Lonati2021).

The method is also applicable to species with a low number of observations (i.e., <20), which makes it useful for EDRR. For this purpose, the data obtained by the LmON are of great importance, as this network is composed of hundreds of observers on the ground (about 600 as of September 2024). New populations of the predefined list of 33 species, or even new species detected in the area by the LmON (as this network includes 184 species), can be incorporated into the annual surveys carried out by the LmVN. However, it must be acknowledged that interpretation problems can arise with species or regions with a low number of observations when generating heat maps due to reliance on proportion matrices. For example, consider a region where 100% of the populations exhibit propagative behavior, but this 100% is actually attributed to only two populations in the area. In contrast, another region may have 70% of the populations displaying this behavior of a total of 10 surveyed populations. In this regard, it is necessary to present the information as completely as possible, because the methodology can serve different purposes (such as selecting the most widespread species—case 2—or, alternatively, selecting species with a low number of populations to prevent their spread—case 1). If the count of observations linked to each square is shown in the diagram, it is possible to see the actual number of associated observations when calculating the percentage (as shown in Figure 4). Alternatively, tables can be incorporated to display the count of partial and total observations along with the proportions represented (as shown in Table 3), providing a more accurate interpretation of the obtained result and avoiding management decisions that are not fully informed.

The approach we developed enables us to effectively organize a large amount of data (in this case, 695 observations) into heat maps. These heat maps provide a concise and visually intuitive representation, facilitating the identification of alien species and/or areas that require monitoring and intervention. Even so, as mentioned earlier, a limitation of this methodology is the inability to determine abundances of each developmental stage within populations exhibiting mixed behavior. For example, a population might have a high density of young plants and a very low density of senescent ones. It should be taken into account that our methodology is not a demographic study sensu stricto, which would require a much more complex approach that includes monitoring individuals throughout the life cycle to estimate population growth rates and identify those life-cycle stages that most influence population trends. Nonetheless, the relatively simple concept of propagative behavior serves as a quick resource to assist in the initial prioritization of species and regions based on their ability to establish growing populations, which is associated with a greater risk of invasion compared with populations where some level of natural decay of individuals is noticeable. Although we do not have data proving that the presence of seedlings/juveniles (i.e., populations with higher propagule pressure) increases the likelihood of establishment (as our observations span only 1 yr), the link between propagule pressure and establishment success is widely acknowledged (Carr et al. Reference Carr, Hooper and Dukes2019; Colautti et al. Reference Colautti, Grigorovich and MacIsaac2006; Simberloff Reference Simberloff2009). Our method, unfortunately, can only be applied to perennial plants, as it might not be feasible to classify a given population of an annual species into any of the three behavioral categories. However, this limitation does not notably reduce the method’s usefulness, as invasive alien plants are typically perennial compared with noninvasive ones (e.g., Huang et al. Reference Huang, Wu, Bai, Zhou and Wang2009; Korpelainen and Pietiläinen Reference Korpelainen and Pietiläinen2023; Sutherland Reference Sutherland2004). In fact, all 33 plant species under monitoring (chosen based on catalogs, scientific evidence, and expert opinion; Materials and Methods), which were regarded as the most invasive or having the most invasive potential in the study area (https://lifemedcliffs.org/en/context/invasive-flora), have a perennial life cycle. This does not mean that there are no annual alien plants in the region (Aymerich and Sáez Reference Aymerich and Sáez2019).

Given the limitations, it is important to keep in mind that this kind of initiative enables the collection of a substantial volume of data through collaboration with residents of many communities in the region. Beyond challenges associated with data quality and methodology, there is also the need to maintain participant motivation, raise awareness, and continuously recruit citizen volunteers to avoid information gaps. For instance, within the LIFE medCLIFFS project, several strategies are designed to keep volunteers engaged. First, there is an incentive system that rewards those who make the most complete and accurate observations (e.g., including high-resolution photographs and with all fields correctly filled in). In addition, periodic meetings between volunteers, project managers, and technicians are held (mostly on the ground but also in the form of dissemination workshops), and feedback is provided through social media (e.g., there is a Telegram channel for answering volunteers’ queries).

Ensuring the completion of a monitoring program—and avoiding information gaps—is of particular relevance, as for most cases it seems much more reliable to recruit a team of citizen scientists rather than depend solely on professional researchers for invasive species detection. There are at least two reasons that make volunteer participation essential, especially for plant research and management. The first one is the sharp decline in the number of professional botanists globally in recent years, as clearly described in Crisci et al. (Reference Crisci, Katinas, Apodaca and Hoch2020). As an illustrative example, the Department of Botany of the Faculty of Pharmacy at Barcelona University has lost over three-quarters of its scientific staff in the past two decades. The second reason is the lack of attractiveness of most fieldwork-based investigations compared with other types of studies (e.g., genetics) for the professional botanists, mostly due to the pressure to publish in high-impact factor journals (Ríos-Saldaña et al. Reference Ríos-Saldaña, Delibes-Mateos and Ferreira2018).

In the context of a shortage (or lack of interest) of professional botanists, the simplicity of the collected data regarding population dynamics becomes an advantage in a citizen science project such as the one presented here. Moreover, the accuracy of the data generated by volunteers is often comparable to that of professional botanists, even for relatively complex measures. For example, volunteers correctly identified plant phenophases (emerging leaves, unfolded leaves, open flowers, full flowers, and ripe fruits) in most cases (91%) in a citizen science project within the USA National Phenology Network (Fuccillo et al. Reference Fuccillo, Crimmins, Rivera and Elder2015). However, it is important to note that the participants in this network attended a 6-h training. Certainly, previous training is crucial for the good performance of the citizen scientists and for ensuring the reliability of the collected data. Salomé-Díaz et al. (Reference Salomé-Díaz, Golubov, Díaz-Segura, Ramírez-Gutiérrez, Sifuentes de la Torre, Koleff, Quintero and Martínez2023) compared the degree of reliability in the phenological classification (initial growth, immature flower, mature flower, and dry fruit) between trained and untrained volunteers for two invasive species [Leonotis nepetifolia (L.) R. Br. and Nicotiana glauca Graham] in Mexico; the level of reliability was significantly higher among trained volunteers compared with untrained volunteers, and for most phenophases the performance of trained volunteers was nearly equal to that of expert scientists.

Broadening Our Proposal: Expanded Categories for Sexually Reproducing Species

A last advantage of our method is that the observations can be made year-round, as the construction of heat maps does not distinguish between vegetative and reproductive adult individuals. Nevertheless, for strictly or partially sexually reproducing species (e.g., P. tobira), populations with flowers and fruits have a higher reproductive potential than those with only vegetative adults (which cannot reproduce). In such cases, the previously mentioned categorization based on the presence of young and senescent plants can be expanded to gain a more detailed knowledge on the population dynamics of the observed populations. By adding a second subdivision according to the presence/absence of sexually reproductive structures (flowers or fruits) in adult plants, one can categorize a given plant population in six subcategories (Figure 6). In contrast, applying this subclassification for species with exclusively asexual reproduction is not always feasible, as it is not possible to distinguish between reproductive and non-reproductive individuals in some cases (e.g., rhizome fragmentation). For example, A. donax does not reproduce sexually outside its native area (Jiménez-Ruiz et al. Reference Jiménez-Ruiz, Hardion, Del Monte, Vila and Santín-Montanyá2021), making it impossible to visually ascertain whether a given population is undergoing asexual reproduction. In other cases of asexual reproduction, this subcategorization can be applied (e.g., for asexual reproduction through buds, bulbils, or plantlets); however, as with sexually reproducing plants, the observations should be made during the reproductive season, which is when the volunteers of the LIFE medCLIFFS project are encouraged to make their observations.

Figure 6. Population subcategory types: type 1, propagative behavior; type 2, senescent behavior; type 3, mixed behavior. The three categories defined in Figure 2 could be separated into two subcategories according to the presence of reproductive adults (with flowers and/or fruits) when necessary.

Figure 7 provides a specific example of the species P. tobira to demonstrate the use of subcategories in cases where a species exhibits sexual reproduction in its invasive range. It also illustrates a detailed analysis of a particular species of interest with a high number of observations (N = 173). The heat map displays the proportion of observations across subcategories and regions, with each colored square in a zone representing a percentage of one of the six possible subcategories that adds up to 100% across each row. Thus, it is evident that P. tobira is present in all five regions of the Costa Brava, albeit with contrasting percentages across the different subcategories of population dynamics. Notably, the S and N regions of Costa Brava are the only ones lacking propagative behavior. Despite this, both regions contain reproductive adults (with flowers and fruits), indicating that these populations are not exempt from the (theoretical) production of offspring by seeds. In addition, there is a higher concentration of populations in the reproductive mixed behavior subcategory in the southern zone.

Figure 7. Heat map with the proportion of Pittosporum tobira observations within each subcategory in the five defined regions of the Costa Brava: C, Costa Brava Center; CN, Costa Brava Center-North; CS, Costa Brava Center-South; N, Costa Brava North; S, Costa Brava South. Each colored square in a region represents the percentage of one of the six possible subcategories of population dynamics, adding up to 100% across each row. Numbers inside each square represent the number of associated observations.