Assessment of the American Flamingo distribution, trends, and important breeding areas

Ethics statement

Do not apply to this study.

We extracted American Flamingo records from eBird (https://ebird.org/home) on 14 November 2018, covering 1960–2018 (31 October). Across the literature, different terms are used to refer to eBird information; hence, for clarity sake, we define the terms used in this study. We used “record” to refer to a given observation (sighting) registered by a given birder at a given time in a given place; “individuals” to refer to the number of individuals reported per record; “cumulative number of individuals” to refer to the cumulative number of individuals registered in a given place (a country or a region) meaning that this value probably contains repeat counts of the same individuals. The complete checklist contained 16,948 records, but we eliminated all information outside the potential geographic range for the species (18 European records of birds in captivity), leaving a checklist of 16,930 records. We then applied two filters: 1) to keep only one record for a given location on a given day to minimize redundancy, leaving out 6,324 records, and 2) to account only for quantitative assessments (i.e., records providing the number of individuals per record), leaving out 1,323 records. Our final database consisted of 9,283 records. These records allowed us to build maps in ArcGIS 10.2.1, and an animation file using the Time Series Animation tools in ArcGIS 10.2.1. All shapefiles for the map backgrounds were downloaded from http://tapiquen-sig.jimdofree.com [19].

Because there is no definitive agreement among scientists on the population structure of the American Flamingo, we decided to describe trends in the cumulative number of individuals considering a geographical division which included (1) North America, (2) Central America, (3) South America, and (4) the Caribbean. In addition to this regional level, we examined all data at the country or sub-country level (S1 Table for details on countries included). We used two metrics to report the data: (1) total number of records per region (four regions) and per country (32 countries), and (2) cumulative number of individuals observed per region and per country. We compared the data in this study with the most recognized distribution of the American Flamingo to determine their occurrence in areas that were not previously included within this distribution.

To assess the current importance of each country as a potential breeding site, we split the data between a breeding season (mostly from April to August; [8]) and a non-breeding season (September to March). We established a parameter over the last decade (2010–2018), which corresponds to the cumulative number of individuals reported in a given country relative to the total number of individuals reported globally during the nine breeding seasons. We only considered countries with data collected during the breeding season to avoid biases. The value obtained for the Galapagos Islands was set as a reference to identify the most critical potential breeding countries for the American Flamingo. The population of American Flamingo in the Galapagos Islands is a small reproductive population of 500 individuals [11], but an important breeding site for the species due to its isolation [11] from the rest of the Pan-Caribbean population (as defined by [20]).

To get a grasp on the effect of birders’ “effort” on eBird data, we determined the number of record days in eBird for the countries identified as most important for breeding over 1960 to 2018. We then explored the changes in the cumulative number of individuals relative to the total number of records for each country to understand the evolution of eBird data over time, thus coarsely accounting for birders’ effort on the number of individuals recorded.

Distribution of the American Flamingo

Detailed information for all eBird records in the four regions is presented in Table 1, and historical data per decade is found in S1 Table. Below, we present information for each of the four regions and for countries.

Updated distribution of the American Flamingo

When we projected all the eBird records from 1960s to 2010s (N = 9,283) onto a map, we found that most records coincide with the most recognized distribution of the American Flamingo (black dots in Fig 1). However, there were records (red dots in Fig 1) that extended to the North (US Gulf of Mexico/Florida) and to the South (Colombia, Guyana, Surinam, and Brazil).

American Flamingo breeding sites

During the last decade, the percentage of the cumulative number of individuals observed in Ecuador during the breeding season relative to the cumulative number of individuals observed during that season was 0.6827 (Table 2). Based on this value, we identified nine countries as critical breeding areas for the American Flamingo during the breeding season (in alphabetical order): Bahamas, Bonaire, Colombia, Cuba, Curaçao, Ecuador, Turks and Caicos Islands, Mexico (Yucatan Peninsula), and Venezuela (Table 2 and S2 Table for details per year considered).

Finally, the changes in the cumulative number of individuals relative to the total number of records for each country (S4 Fig) highlighted contrasting situations. In the nine countries with critical breeding areas for the American Flamingo (S4A Fig), the relative number of individuals tended to increase slightly except for (1) the colony in Ecuador that remained broadly stable, (2) the colony in Cuba that stopped growing in the 2000s, and (3) the colony in the Yucatan Peninsula that appeared to be decreasing since the 1990s. Most other countries (S4B Fig) showed a slight increase in the relative number of individuals since the 2010s. The marked increase in the total number of records and cumulative number of individuals in the last decade (2010s) appears therefore to reflect mainly birders’ reporting effort, as shown on S5 Fig.

Distribution of the American Flamingo

To revise the current distribution of the species using eBird data, we developed an overall picture of the situation of the American Flamingo and citizen's efforts to report this iconic bird. Our results demonstrate a bias in the total number of records according to the countries and regions. Countries with high tourist interest and/or high research investment generally reported the greatest total numbers of records and cumulative individuals. The situation of the American Flamingo in these countries is probably well described by the data. For instance, the lack of records in Costa Rica, which is renowned for its ecotourism, particularly ornithological tourism (see [22]), suggests the absence of the American Flamingo in this country. On the other hand, the low total number of records for some countries such as Brazil (Amapa state), Guyana, or Suriname, probably reflect a lack of tourism, the difficult access to flamingo colonies, and/or limited ornithological research, rather than the rarity of the species.

A tendency in many countries was the sharp increase in the total number of records (S1 Animation) in the 2010s and birders’ “effort” indicated by the number of record days in eBird. This may suggest an increase in (1) the worldwide availability of eBird in the 2010s that led to an increase in reported data [17], (2) improved access to technology to share data, particularly the development of smartphone applications that considerably facilitated the submission of data from field [17], (3) the global interest in conservation, (4) research investment, and (5) tourist facilities and especially interest in birdwatching tourism in the whole Caribbean region [3]. Additionally, the increased in the cumulative number of individuals on eBird during the 2010s may also indicate an increase of the American Flamingo global population, as indicated by the Red List assessment which reports the Flamingo population as “increasing” [5]. Our results on changes in the cumulative number of individuals relative to the total number of records for each country suggested a slight increase for most countries, and a decline only in the Yucatan Peninsula since the 1990s. The latter may be caused by decreasing habitat quality, possibly due to natural disturbance [23], which could induce flamingos to move to other sites, as observed in the Galapagos [7]. It might also be explained by the movement of individuals to nearby colonies or feeding areas that showed an increase in the cumulative number of individuals, such as Texas or Florida [2]. Although all other countries (S4B Fig) showed a slight increase in the cumulative number of individuals since the 2010s, very little data is available, so these results need to be interpreted with caution.

In North America, two areas stand out for the American Flamingo: the Yucatan Peninsula in Mexico and the coast of the Gulf of Mexico and Florida. The significance of the Yucatán Peninsula comes as no surprise, as the ecology of the American Flamingo has been studied there for decades (e.g. [4, 8, 24, 25]. Regarding the coasts of the Gulf of Mexico and Florida, eBird data show records there since the 1970s; the last decade (2010s) saw an explosion in numbers, especially in Texas and Louisiana as previously reported (see [13]). These authors suggested that the recent increase of the Florida population probably results from emigration from other Caribbean populations. This resurgence of the Florida flamingo population, eliminated by ca. 1900 due to anthropogenic pressure, has generated increased interest in their conservation [13, 20].

In the Caribbean, the three countries that stand out are located in the North: the Bahamas, Cuba, and the Turks and Caicos Islands. In the Bahamas, despite being the national bird [26], the flamingo population was tiny with approximately 100 individuals in the 1950s; however, the increase in the cumulative number of individuals (S1 Table) may suggest an increasing population (S4 and S5 Figs). The group sizes reported in eBird are generally large, with 41% of records reporting flocks of 100 to 1,000 individuals and up to 5,000 individuals in 9% of records. Baltz [26] reported that Flamingos are once again observed in North Andros where they had disappeared by 1950 and that the only current colony is on Great Inagua; in eBird, most records originated from Inagua. In Cuba, which has the highest total number of records and cumulative individuals in eBird since the 1980s, 63% of records report 1–100 flamingos per record; however, some records (5%) reveal flocks of over 1,000 individuals and up to 15,000 individuals. Such large cumulative numbers of individuals in Cuba suggest that some birds are not resident. Indeed, Cuba is an important area for the movement of flamingos during seasonal migration [12]. The Turks and Caicos Islands’ records for date back from the 1970s, but 96% of records did not appear until the 2010s. The status of the species is unclear in these islands: despite evidence of breeding activity and large numbers of individuals [27, 28], Walsh-McGehe et al. [29] were unable to confirm whether flamingos are residents or migrate from areas such as the Bahamas.

Most first records in eBird are from the 2010s in the other Caribbean islands except for Hispaniola and Aruba where the first records date back from the 1980s. On Hispaniola more records are reported in the Dominican Republic than in Haiti, but the population has decreased on the whole island principally due to anthropogenic factors [30]. Flamingo numbers on Hispaniola fluctuate, probably because it receives migrants during the non-breeding season [references in 30]. For Aruba, the flocks reported in eBird have no more than seven flamingos. Luksenburg and Sangster [31] suggested that Aruba’s flamingos may originate from Bonaire or the Los Olivitos colony in Venezuela. In Puerto Rico and the Virgin Islands, although flamingos are reported as vagrants or accidental [32], there is no information published for these locations to the best of our knowledge.

In South America, four countries are noteworthy: Colombia, Venezuela, Bonaire, and Curaçao. For Colombia, Murillo-Pacheco et al. [9] suggested that the recent occurrence of individuals in new areas indicates a possible range extension or the presence of vagrant/introduced individuals. In Venezuela, Espinoza et al. [33] suggested that the American Flamingo population represents 38% of the global population, a very similar proportion to that found in this study. The population of resident American Flamingos appears to have doubled in 20 years, from an estimated 37,110 individuals in 1996 [33] to 75,622 individuals in 2017 [34]. The western coast of Venezuela, specifically the Los Olivitos refuge, is the most crucial area in the country (~69% of eBird flamingo records of Venezuela) [33]. American Flamingos in this area continuously migrate, principally to Bonaire, searching for food [35]. Bonaire is a significant area with reports in eBird spanning several decades, most in the 2010s. The current American Flamingo population of Bonaire is estimated between 1,500 and 7,000 breeding individuals [35]. In nearby Curaçao, the species is considered a regular non-breeding visitor, and breeding attempts of a few hundred individuals were unsuccessful [36]. Although the literature mentions the Jan Thiel Lagoon (a salt pan) as the main location for the American Flamingo [36, 37], eBird reports only 86 records for this area, whereas it reports 293 records in Saliña Sint Michiel (another salt pan).

The literature on the American Flamingo in Guyana, Suriname, and Brazil is scarce. For Guyana, the status is unclear due to the lack of data [14], and recent eBird observations could reflect the opening of this country to eco-tourism [38]. For Suriname, most records are from the Nickerie District (Northern border with Guyana), but Schulz et al. [39] reported the species in the Wia-Wia Nature Reserve, in Central Suriname, over 40 years ago. For Brazil, eBird had only one record with 150 individuals on the north coast of Amapá State in Cabo Orange. This single record is surprising considering that the American Flamingo is a resident species in Brazil [40]; however, it may reflect the isolation and difficult access to the area. This example highlights some of the weaknesses of eBird data, and emphasizes caution during its analysis.

In the Galapagos Islands, the population of the American Flamingo is considered stable, with approximately 500 individuals [41], 45% of which are capable of reproducing [11]. The groups are tiny, with usually no more than 20–30 birds (eBird database, [41]). However, one record in eBird reported a group of 100 individuals observed out of the breeding season (January 2010). Vargas et al. [7] reported that the most important feeding and breeding sites are Quinta Playa and Cementerio Lagoon, both on Isabela Island. eBird data showed only two records for Quinta Playa and none for Cementerio Lagoon, indicating that birds in these sites are probably under-reported. Most eBird records providing a location are for Isabela and Floreana Islands, which have permanent flocks [41], and for which 73% of records date back from the 2010s. There has been a huge increase in land-based tourism in these islands [42], which might explain increasing reports especially around Puerto Villamil on Isabela, while traditional sites could be more difficult to reach.

Updated distribution of the American Flamingo

The American Flamingo distribution, as reflected by eBird data in this study, may suggest an expansion of its range to (1) the North on the south coast of the United States and in Central Mexico, and (2) to the South in Brazil, Guyana, Suriname, and inland Colombia. The remainder of the records coincide with the species most recognized distribution [2, 4, 5, 8, 12].

The colonization of the North coast of the Gulf of Mexico and the re-colonization of Florida represent the most significant range expansion for the American Flamingo. Although, the number of individuals reported remains modest with a cumulative 3,773 individuals, their increase in this area is dramatic, as just seven individuals were reported in the 1970s. There is evidence that natural dispersion is occurring in the United States with individuals migrating from the Caribbean [13]. This dispersion is consistent with the findings of other studies that demonstrate a northward range expansion for different taxa, possibly related to global warming [43, 44]. These studies suggest that climate change may affect and modify the distribution of the American Flamingo, which warrants further modelling of the potential geographical range.

In Mexico, flamingos were reported in different states (e.g., Mexico City, Jalisco, Tamaulipas, Nayarit, Veracruz, and Guanajuato), but the very low total number of records and cumulative individuals suggested that these individuals were introduced or vagrant. In Mexico, flamingos are uncommon out of their distribution area (the Yucatan Peninsula). However, some locations such as Laguna Chumbeño in the Biosphere Reserve Marismas Nacionales in Nayarit have small population descendent from individuals that have escaped from captivity [10]. Therefore, we can consider that eBird reports in Central Mexico do not represent an expansion of the range of distribution of the American Flamingo.

In Colombia, Murillo-Pacheco et al. [9] suggested the possibility of range expansion for the Orinoco region. However, the eBird data had no information for this region. Historically, the American Flamingo distribution occurred on the Caribbean coast of Colombia [5, 9], as shown by the majority of eBird records coming from coastal Departments (e.g., La Guajira, Magdalena). As in Mexico, inland records of flamingos could be exotic or introduced [9] and these probably do not equate to a range extension.

The southernmost regions where the American Flamingo occurs are Suriname, Guyana, and northeast Brazil [5]. The scarcity of information on flamingos in both Suriname and Guyana is probably due to limited tourism and research in the area. In Brazil, the status of the American Flamingo is unclear; it is reported in the Amapá State where it reproduces [40]. The bird list of Brazil considers the American Flamingo a resident species, but this is unconfirmed [45]. In summary, it seems that information in these countries is more a reflection of the presence of eBird as an easy and readily available means of reporting rather than a valid range expansion.

American Flamingo breeding sites

Previous literature recognized six countries with important breeding sites for the American Flamingo: Mexico (Yucatan Peninsula), Bonaire, Ecuador, Venezuela, Bahamas, and Cuba [5, 11]. Whitfield et al. [20] omitted Ecuador because they considered only the “Pan-Caribbean population”. Using recent eBird data (2010–2018) for the breeding season, and setting the Galapagos as the reference point, we identified nine countries that are potentially important breeding areas for the species. Six of these countries coincide with those previously mentioned. At the same time, an additional three were included: the Turks and Caicos Islands, Colombia, and Curaçao. The inclusion of these countries was based on our analysis of eBird data and is therefore not confirmed. However, this result should inform efforts to assess the status of the colonies in those countries.

The inclusion of the Turks and Caicos Islands should come as no surprise. When estimating the historical nesting range of the American Flamingo in the Caribbean, Allen [46] already mentioned this British territory. Norton and Clarke [27] found two abandoned nesting colonies on North Caicos with 6,000 to 7,000 mounds, and the presence of young flamingos in the Turks and Caicos Islands strongly suggests breeding in this country [29]. Contrary to the situation in many areas of the Caribbean, where wetlands are disappearing under anthropogenic pressures, wetlands in the Turks and Caicos Islands remain effectively unaltered by human activities [28]. Furthermore, these authors noted that the ornithological interest in the Turks and Caicos Islands remained low in the 1990s, and eBird records for the breeding season did not commence until 2010, which may explain the lack of information.

The situation of Colombia as a potential breeding area is far less clear. There are records of the American Flamingo nesting in the department of La Guajira during the first half of the XX Century, and recent observations suggest the possibility of reproduction [15]. Nevertheless, according to these authors, individuals move to Venezuela (“Ciénaga de los Olivitos” protected area) to reproduce. Indeed, no nesting areas were reported in the 2000s despite efforts to find active nests or mounds [47]. One hypothesis is that flamingos show high fidelity to their birth site [48], and therefore would most probably return to bordering Venezuela to reproduce. However, although eBird has reported the American Flamingo in Colombia since the 1990s, the first information during the breeding season appeared in 2010 with a significant increase in the cumulative number of individuals in 2016 (> 4,000). This data may suggest that flamingos find suitable nesting areas in Colombia, stressing the need to increase efforts to confirm its reproduction status in that country.

For Curaçao, recent eBird data may suggest a small incipient breeding population. Though considered a non-breeding visitor, several breeding attempts were documented [36], and eBird reported two recent events related to breeding. The first, in 2012, corresponded to territorial defense behavior in a group of 3 individuals in Saliñas Jan Kok; the second, in 2018, corresponding to the C code (meaning courtship, display, or copulation) in a group of 120 individuals in Jan Thiel Lagoon. Almost half of records during the breeding season counted less than 20 individuals and did not exceed 200 individuals [49]. Groups of less than ten pairs are unlikely to nest, because flamingos need social stimulation to develop breeding behavior [20]; although, as Lazell [50] found in the Virgin Islands, this is not impossible. Consequently, Curaçao is a potential breeding area, but more research is necessary to confirm it.

Perspectives

The American Flamingo is a charismatic bird, but knowledge related to its populations and movements remains surprisingly patchy despite its importance. Citizen science projects such as eBird allowed us to obtain valuable information regarding where and when the species is located and how much certain sites are crucial. Nevertheless, a complete understanding of the movements of individuals will only be possible using techniques such as capture-mark-recapture, satellite telemetry, and stable isotopes, among others [13]. Though flamingos are reported in many countries, very little literature is available for most countries. This underlines the necessity to develop more scientific programs that include information sharing and cooperation between countries, to firmly establish the status of the American Flamingo and gain a better understanding of its population dynamics. In the interim, eBird data allows us to pinpoint areas that may require more attention for their conservation, in addition to areas where the species is expanding. Of course, a more detailed analysis of eBird data for each country combined with information on the population dynamics of the American Flamingo is warranted. However, it is beyond the scope of this study. Finally, we suggest a re-evaluation of its status, especially for countries hosting important breeding colonies.

Our results appear to support the current view that the status of the American Flamingo is improving generally [5], in line with the significant efforts that have been made to conserve its populations. In 1978, the Flamingo Specialist Group (FSG; http://www.flamingo-sg.org/), today part of the IUCN Species Survival Commission, was created. This group is composed of flamingo specialists who study, monitor, manage, and conserve the populations of six species of flamingos [15]. At the same time, many countries where the American Flamingo inhabits increased efforts to gather information about the species and to ensure the protection of its habitat. In Mexico, for instance, the scientific interest in flamingos began in the 1950s in the Yucatan Peninsula, and by the 1970s the Mexican government had launched a monitoring program. In 1979, two federal reserves were created in the region to protect flamingo habitat (Ría Lagartos and Ría Celestún), and in the 1980s research efforts turned to habitat use, feeding ecology and behavior, as well as the effects of ecotourism on the species (references in [4]). In Cuba, the population of flamingos was declining due to habitat degradation and hunting, prompting the Cuban government to create, in 1978, the “National Company for the Protection of the Flora and Fauna” to develop a protection plan for the American Flamingo that would allow its recovery [51]. The Bahamas developed specific programs for the American Flamingo, mainly to identify critical habitat and assess the breeding status of the species [26]); Bonaire decreed wetland reserves protecting important areas for the species [52]; Colombia started to legally protect flamingos from hunting in 1971 and created national and regional reserves to protect its habitat [15]; and Venezuela decreed most of Los Olivitos, an important area for the flamingo, a wildlife refuge in 1986 [53]. These great efforts in the Caribbean region have been rewarded so far; however, they must continue due to the vulnerability of flamingos to natural and human disturbances, especially climate change.