Introduction

Life on Earth is constantly changing, and scientists are unable to discover everything on their own. Bearing this in mind, each year the interaction between scientists and the wider public grows (Follett and Strezov 2015; Theobald et al. 2015; Kosmala et al. 2016; Pocock et al. 2017a, b). This growth in public participation provides many opportunities for scientific research, which generates collaboration between scientists and citizen volunteers – a concept called citizen science (CS) (Dickinson and Bonney 2012; Bonney et al. 2014; Frigerio et al. 2018). Citizen science provides an invaluable opportunity to collect information that would otherwise be challenging for researchers due to time and resource constraints (Cohn 2008; Kobori et al. 2016).

Citizen science bridges science and society by involving members of the general public in scientific research in many areas of study, especially in subjects related to biodiversity, such as species population data (Dennis et al. 2017; Callcutt et al. 2018; Harvey et al. 2018; Crawford et al. 2020; García et al. 2021; Sanderson et al. 2021), general ecology (Washitani et al. 2020; Olsen et al. 2020; Herremans et al. 2021; Barahona-Segovia et al. 2023; Shumskaya et al. 2023), new taxa (Alther et al. 2021; Fagan-Jeffries et al. 2021; Rosa et al. 2021; Mota et al. 2022), invasive species (Pocock et al. 2017b; Ryan et al. 2019; Johnson et al. 2020; Streito et al. 2021; Tran et al. 2022), geographic distribution (La Sorte and Somveille 2020; Marcenò et al. 2021; DiBattista et al. 2021; Diao et al. 2022; Wangyal et al. 2022), and rare or threatened species (Kaminski et al. 2015; Sullivan et al. 2019; Wilson et al. 2020; Mesaglio et al. 2021; Fontaine et al. 2022; Rosa et al. 2023a).

Another valuable source of ecological data is passive citizen science (PCS). Similar to the CS approach, PCS involves the participation of non-experts unconnected to any particular CS program, and datasets are collected without the organisation or oversight of specific campaigns. Data are retrieved from internet resources, particularly social media, where members of the public have uploaded observations such as photos of wildlife (Edwards et al. 2021). The use of internet sources (such as Facebook, Flickr, and iNaturalist) for gathering wildlife-related data in citizen science initiatives has emerged in recent years (Barve 2014; Liberatore et al. 2018; Marcenò et al. 2021; Geller et al. 2021; Mesaglio et al. 2021; Cranswick et al. 2022). In an overview of the impact of internet social networks on traditional biodiversity data collection methods, Di Minin et al. (2015) are optimistic that social media can potentially play an important role in conservation science.

Regardless of the type of citizen science approach (conventional CS or PCS), photographs of specimens (dead or alive) from internet sources represent a type of data that is largely applicable to a range of ecological studies. A photograph of a specimen provides a record for the scientific community to view and discuss, both in the quest for consensus regarding the taxonomic identification of the organism photographed, and as a permanent record for any necessary future revision (Casanovas et al. 2014). This kind of data can provide preliminary information on the distribution and/or biology of the organism, even if studies of this type are still incipient (Jimenez-Valverde et al. 2019; Fontaine et al. 2022). Opportunely, data from citizens (especially distribution records) are used to calculate trends and geographical ranges of species (Maes et al. 2015), which are two important criteria to estimate a species’ extinction risk (Mace et al. 2008; IUCN 2012; Vantieghem et al. 2017; Rosa et al. 2023a).

The butterflies represent one of the biological groups for which citizen scientists have provided reliable and abundant information on geographic distribution (Dennis et al. 2017; Lewandowski and Oberhauser 2017; Ryan et al. 2019; Washitani et al. 2020; Herremans et al. 2021; Sanderson et al. 2021; Plummer et al. 2024). These organisms have great popular acceptance, due to their strikingly colored wings, easy of observation and photography, abundance, wide distribution and taxonomic diversity (Heppner 1991; Boggs et al. 2003; Mitter et al. 2017).

Therefore, the purpose of the present study is to validate the importance of data provided by citizens as a valuable source of information to support conservation assessments, including estimates of the geographical range of threatened Brazilian butterflies.

Materials and methods

Data collection

The target butterflies in this study are those present in the current Brazilian Red List (63 taxa) (MMA 2022); some examples from amateur photographers are shown in Fig. 1. All recent taxonomic changes were followed for the sake of taxonomic stability.

Fig. 1
figure 1

Photos of some threatened Brazilian butterflies (all collaborations from citizen scientists photographers). (A) Parides ascanius (Cramer, 1775) (courtesy of Carlos Gussoni); (B) Petrocerus catiena (Hewitson, 1875) (courtesy of (Bettina Dungs); (C) Eresia erysice erysice (Geyer, 1832) (courtesy of Nilda Borges); (D) Parides bunichus chamissonia (Eschscholtz, 1821) (courtesy of Germán Muriziasco); (E) Hesperocharis hirlanda planasia (Fruhstorfer, 1910) (courtesy of Rodrigo Conte); (F) Tithorea harmonia caissara (J. Zikán, 1941) (courtesy of Elsie Laura)

Full size image

From 2015 to April 2023, images (photographs) of adult and/or larvae of butterflies from the Brazilian Red List were searched (using their current names, previous names, synonyms, and for subspecies determinations) in the world wide web, on the Google Images online platform, and other specific websites (Flickr (https://www.flickr.com/), Biofaces (https://www.biofaces.com/), iNaturalist (https://www.inaturalist.org/), YouTube (https://www.youtube.com/), and also on the social media website Facebook (https://www.facebook.com/), both in general search and in specific groups for photos of butterflies and animal-related topics (“Borboletas e Mariposas Neotropicais”, “Borboletas e Mariposas”, “Passarinhandos e Borboletando”, “Entomologia Brasileira”, “Entomos”, “Insetos”, “Insetos do Brasil”, “Fotonaturalistas – Insetos”, “Fauna Brasileira”, and “Borboletas”). Duplicate records in different databases were considered from only one source (in this case, the database where the image was posted first) (e.g. the same record on Facebook and iNaturalist). In order to obtain the most accurate geographical coordinates of the localities where the photographic record was made, each of the citizens was contacted, most of whom provided accurate data. Records from personal communication with citizen scientists (eg. Via personal email and Facebook Messenger) were also compiled, along with the photograph of the taxon to prove their identity. All photographs in this study were verified by the authors (both are experts on Brazilian butterflies) and/or by external specialists. Additionally, the photographs were also compared with images or specimens of the threatened Brazilian butterflies by the authors themselves and/or by external specialists (see also Rosa and Freitas 2023).

A citizen science project (CSP) entitled “Como a ciência cidadã pode ajudar na conservação das borboletas ameaçadas de extinção do Brasil?” (“How citizen science can aid the conservation of threatened Brazilian butterflies?”) was also developed from January 2019 to January 2021 on Facebook, using the page “Labbor – Laboratório de Ecologia e Sistemática de Borboletas” (“Labbor – Laboratory of Butterfly Ecology and Systematics”) (https://www.facebook.com/labborboletas/) as host. Initially, a photographic guide containing the target species was produced and previously shared in Facebook groups (same as cited before) to facilitate the identification of the target taxa by citizens (Rosa and Freitas 2023). Subsequently, an infographic containing the butterfly photographs and basic information (eg. Conservation status, habitat, host plant) was posted weekly on “Labbor” page and shared in ten groups of butterfly and animal-related photographers (as mentioned above) (see the complete material in Supplementary Material 1). Records informed by comments on posts or sent by messages/email were also compiled (Supplementary material 2), however, only records that contained photos to prove the unequivocal identification of the taxon were used in the analysis of this study. The dates of when each photograph was posted were compiled to show the temporal pattern of data availability over time. All records from photographs were hereafter considered “citizen records”. All “citizen science records” were separated into two categories: (1) active search (compiled by intensive search in the aforementioned databases) and (2) citizen scientists’ feedback (compiled by the citizen science project (CSP), coming exclusively from citizen contact via project as well as separate personal communication (records informed by citizens, but not from the CSP).

In addition, “Non-citizen science records” (data not provided by citizen photographs) were used only for some comparisons with the “Citizen science records” and for contrasting geographical range estimates (see Rosa et al. 2023a for data). To avoid multiple “new” geographical records overlapping in the same locality, all data outside a grid of 4 km2 where any published records was present were considered as new. The grid size (4 km2) follows the IUCN recommendations for the area of occupancy (see further information in the next section). The distance of each “Citizen science record” from the nearest known locality of that given species was was measured using the ruler tool in the Google Earth program. All records within a diameter of 2 km2 were considered as “same locality” and were not used to calculate the mean distances and other parameters.

Estimates of geographical range (extent of occurrence and area of occupancy)

Two metrics related to the restricted geographic distribution (International Union for Conservation of Nature (IUCN) criterion B) were used to assess the conservation status of each taxon: (1) the extent of occurrence (hereafter EOO) and (2) the area of occupancy (hereafter AOO) (IUCN 2012; IUCN Standards and Petitions Committee 2022). The EOO is defined as the area (in km2) contained in the shortest continuous imaginary limit that can be drawn to cover all known, inferred, or projected sites of the current occurrence of a taxon, excluding cases of vagrancy. The AOO is a metric (in km2) on a scale representing the area of suitable habitat currently occupied by a given taxon, within the limits of the EOO (IUCN 2012; IUCN Standards and Petitions Committee 2022). The EOO can be estimated as the area of the minimum convex polygon (the smallest polygon in which no internal angle exceeds 180 degrees and that embraces all locations of occurrence), while the AOO is estimated by the sum of all grid cells (usually with the size of 4 km2) containing at least one record of the target taxon (IUCN 2012; IUCN Standards and Petitions Committee 2022). The extinction risk categories based on the thresholds of criterion B for EOO are: Critically Endangered (CR) = 100 km2, Endangered (EN) = 5,000 km2, and Vulnerable (VU) = 20,000 km2. For the AOO, the thresholds are: CR = 10 km2, EN = 500 km2, and VU = 2,000 km2 (IUCN 2012; IUCN Standards and Petitions Committee 2022).

For the present study, values for both EOO and AOO were estimated based on two scenarios, here named the “without citizen science records” estimate and “with citizen science records” estimate. The “without citizen science records” estimate was obtained without including the data gathered from citizen photographic records. The estimate “with citizen science data” was obtained by combining all records, including photographs from citizens. These data were analyzed using the online open-source program GeoCAT (Geospatial Conservation Assessment Tool, available at http://geocat.kew.org/) (Bachman et al. 2011). Grid cells of 4 km2 were used following the recommendation of the IUCN for AOO analyses (IUCN 2012; IUCN Standards and Petitions Committee 2022).

Estimates were made as global (use of records inside and/or outside Brazil). Geographic data used to estimate EOO and AOO are the same used in Rosa et al. 2023a; with the addition of some data (Supplementary material 2). Localities where the original habitat of a given taxon was modified, replaced, or destroyed (based on recent visits or direct observation of Google Earth Pro images) were not used for AOO-EOO estimates (these are possible cases of local extinction). Finally, since all records refer to threatened taxa that have potential commercial trade interest (for private collections), specific geographic coordinates and specific geographical localities were deliberately omitted.

Results

A total of 37 taxa (58.7% of all threatened butterflies from the Brazilian Red List) were identified from 335 photographs (records) from 187 citizen scientists from various surveyed sources (Supplementary material 2). From this total, 121 (36%) records represent 67 new geographical localites, distributed in 48 new municipalities and two new Brazilian states (Supplementary material 2). Virtually all citizen data were concentrated in Brazil (only one record in Argentina), and are distributed in 14 states, 83 municipalities, and 106 different localities (“citizen science records only” in Fig. 2). Records and number of taxa in the Brazilian states are shown in Table 1.

Fig. 2
figure 2

All known records of threatened Brazilian butterflies. Red squares: records without citizen science data; Green circles: only citizen science data. Abbreviations of Brazilian states: AL. Alagoas, BA. Bahia, DF. Distrito Federal, ES. Espírito Santo, GO. Goiás, MG. Minas Gerais, MT. Mato Grosso, MS. Mato Grosso do Sul, PB. Paraíba, PR. Paraná, PE. Pernambuco, RJ. Rio de Janeiro, SC. Santa Catarina, SP. São Paulo, TO. Tocantins

Full size image
Table 1 Number of records and taxa of threatened Brazilian butterflies in different states of Brazil
Full size table

The family with the most records was Papilionidae with 146 records (43.6%) from six taxa, followed by Nymphalidae (120 records, 35.8%) from 20 taxa, then Pieridae (45 records, 13.4%) from four taxa, Riodinidae (15 records, 4.5%) from four taxa, Hesperiidae (5 records, 1.5%) from three taxa and Lycaenidae (4 records, 1.2%) from one taxon. The most recorded taxon was Parides ascanius (Cramer, 1775) with 99 records (29.6%), followed by Parides bunichus chamissonia (Eschscholtz, 1821) with 26 records (7.8%), then Anaea suprema (Schaus, 1920) (24 records, 7.2%), and Hesperocharis hirlanda planasia (Fruhstorfer, 1910) and Tithorea harmonia caissara (J. Zikán, 1941), both with 20 records (6%). All other taxa were represented by less than 19 records (Fig. 3, Supplementary material 2). Considering the two categories of search, 257 records (76.7%) for 35 taxa came from active search and 78 (23.3%) for 20 taxa came from citizen science feedback (Table 2, Supplementary material 2). The temporal distribution of records (posted on websites or citizen personal communications of threatened butterflies photographs) showed a clear increase in numbers in the last decade (2014–2023), with a well-marked peak in 2019 (Fig. 4).

Fig. 3
figure 3

Abundance rank of threatened Brazilian butterflies from citizen data

Full size image
Table 2 Number of records of threatened Brazilian butterflies from different sources of citizen data. CSP = Citizen science project
Full size table
Fig. 4
figure 4

Number of records of threatened Brazilian butterflies from citizen data per year (based on the year the record was posted on the internet platform or the contact of the citizen scientist)

Full size image

Based on the CSP alone, approximately 27,300 people were aware of the project (mainly through the posts advertising the photographic guide of the Brazilian threatened butterflies). The average number of views for the infographics that were posted and shared on Facebook groups was approximately 2,450 views. However, the feedback and correct identification rate was relatively low; only 197 records (personal communications with 151 citizen scientists) were shared by citizen scientists, and only 26 records (13%) were proven to be of threatened butterflies (11 taxa). The other putative records received were only visual (no photographs available) or were of other similar species. Total records of CSP are distributed in 110 municipalities from 19 states (Supplementary Material 2).

In total, 254 “Citizen science records” were considered as “same locality”; concerning the new records, five classes of distance were defined: 1) from 0 to 10 km (n = 28); 2) > 10 to 50 km (n = 38); 3) > 50 to 100 km (n = 2); 4) > 100 to 200 (n = 3) and 5) > 200 km (n = 3). Considering only records of new localities (see methods), the distances from the nearest known locality ranged from 4.4 to 486 km, with a mean of 34.4 km (median = 11.5 km).

Geographical range (EOO-AOO)

Comparing the estimates of species’ geographical ranges “without citizen science records” to “with citizen science records”, changes were reported for 26 taxa (70.3% of the taxa with new data obtained in the present study), which represent 41.3% of all taxa in Brazilian Red List (Table 3). Comparing “without citizen science records” to “with citizen science records” estimates for EOO, changes were reported for 20 taxa (54.1%) with increases in EOO for all these taxa. Concerning the AOO, changes were reported for 26 taxa, with increases in AOO for all these taxa (Table 3). There was an increase of more than 50% in the EOO of eight taxa, with this increase in percentage in AOO observed only in one taxon (Table 4).

Table 3 Estimates of extent of occurrence (EOO) and area occupancy (AOO) and their percentages of change for the 37 threatened Brazilian butterfly taxa studied
Full size table
Table 4 Percentages of change in extent of occurence (EOO) and area of ocupancy (AOO) and number of studied taxa of threatened Brazilian butterflies
Full size table

Discussion

In the present study, citizen scientists have contributed data for more than half of the taxa of threatened Brazilian butterflies. Citizen science records increased knowledge with 67 new geographic localities of threatened species in two new states in Brazil and various municipalities, as well as the first record of Zonia zonia diabo O. Mielke & Casagrande, 1998 in Argentina. These data represent new opportunities for further studies (e.g. population or immature biology) on threatened species in these areas.

Parides ascanius (Cramer, 1775) was the species most recorded by citizen scientists, representing alone almost 30% of all records. This is probably because this species is easily found in urban parks in the municipality of Rio de Janeiro and surrounding areas, as well as because of its conspicuous coloration, large size, and easy detection when present (as several other species of Papilionidae). Finally, this is an iconic threatened species that is well known to the local people in Rio de Janeiro (Otero and Brown Jr 1986).

The active search of data resulted in more records than the citizen scientist’s feedback (both in numbers of distribution records and in the number of registered taxa). Moreover, the active search was more efficient in terms of results. However, the citizen feedback was also important, contributing data for 20 butterflies on the Brazilian Red List. Regardless of the type of search, we briefly introduced ourselves to each citizen and explained the type of study that was being carried out, also asking for additional information for each record (e.g., more precise geographic coordinates, dates of the record, among other details). This type of interaction with the citizen scientist is very rewarding for both actors, as most were very eager in providing the additional requested information and mentioned that were happy to contribute in some way to a scientific study. It worth noting that in the data obtained through active search by the authors, most of the time the photographers did not know that the butterfly taxa were threatened. This information is important because it minimizes potential bias such as photographers actively searching for threatened species. However, this scenario, along with the low number of new information obtained from CSP, indicates a necessity for improved engagement strategies or educational initiatives to enhance the accuracy and quantity of citizen-contributed data in future projects centeresd on this source of information.

The citizen science project (CSP) developed on Facebook received a moderate number of views on the social network (average views of 2,450 per post), however, as previously mentioned, the feedback was relatively low; from the available data, only 13% represented threatened butterflies. Despite this relatively low rate, when the CS data were summed with the full dataset, it resulted in effective changes in the estimates of geographic range. With the inclusion of CS data, there was an increase in geographic range (for both EOO and AOO) of 26 taxa (41%) of the 63 butterfly taxa present in the Brazilian Red List. For some taxa, there was more than a 100% increase in EOO, with the most notable being Petrocerus catiena (Hewitson, 1875), a species previously known only from two municipalities in the mountainous region of the state of Rio de Janeiro. With the new record for the municipality of Castelo, in the state of Espírito Santo, the EOO increased by 2271% (see Rosa et al. 2023b). Both the records and geographic range estimates from this study can be used in future conservation status assessments of these taxa, resulting in a more realistic status than those in the current Brazilian Red List. We expect the greatest increases in species in the families Riodinidae (as is the case), Hesperiidae, and Lycaenidae. Butterflies in these three families are usually not easy to observe, and their rarity could many times be an artifact of low detectability. Accordingly, in two recent examples, a picture of a butterfly in the field expanded considerably the geographical ranges of two supposedly threatened species of Hesperiidae (Rosa et al. 2017) and Riodinidae (Greve et al. 2013).

The temporal distribution of records clearly showed an increase in the last decade, possibly as a result of the advance in the technology of mobile devices and online data sources (Cohn 2008; Silvertown 2009; Bik and Goldstein 2013; Arts et al. 2015). However, the noticeable reduction in the number of records in 2020 could be a direct effect COVID-19 pandemic and the restricted mobility imposed by the quarantine in Brazil. This underscores a potential vulnerability associated with reliance on citizen science data, as it can be susceptible to external influences. Other studies have also discussed the impacts of COVID-19 on the development of CS projects, supporting our results (Basile et al. 2021; Crimmins et al. 2021; Lynch and Miller 2023). Accordingly, the number of records began to grow after the first year of the pandemic and is now approaching the rates observed before the pandemic (Fig. 4).

Based on the distance between the new geographic record and the closest known locality, most of the new records (77%) fell into the category “same locality”. However, several records here considered as “new localities” were considerably distant from the previous known records, representing potential areas for specific studies on the biology of these organisms and contributing to changes in the conservation assessments of various species (see Rosa et al. 2023a). One valuable advantage of CS data is that information is consistently recent and up to date compared to data available in the literature and/or scientific collections and can serve as a confirmation that a given taxon is still present in a historical locality. A good example is the occurrence of Parides ascanius in the Deodoro neighborhood in the municipality of Rio de Janeiro; a single specimen was collected there in 1938 (voucher number CEIOC 20,707), and recently, four individuals were photographically recorded at the same locality (one in 2021 and three in 2022).

The current results illustrate the significance of citizen science in advancing our understanding of endangered Brazilian butterflies. However, while this study specifically addressed threatened butterflies in Brazil, the findings may not directly apply to other regions or species. However, a similar approach using data from various social media platforms could be used to gather information on a wider range of organisms and geographic locations. The involvement of the general public in scientific research should be constantly encouraged, as the simple act of sharing photographs on the internet of a living organism in nature has the potential to contribute to researchers in their studies by providing accurate and quality data, in addition to relieving them of limiting factors in research, such as time in the field and financial resources.