Reefs are among the most diverse and productive environments on the planet (Odum and Odum, 1995), providing important ecosystem services such as coastal protection, maintenance of high biodiversity, tourism and fishing (Moberg and Folke, 1999). Despite their importance, reefs are under severe risk and threatened by anthropogenic actions (Hodgson, 1999; Hughes et al., 2017; Magris et al., 2018). In addition to the gradual reduction of fish stocks, overfishing and species extinction (Jackson et al., 2001), eutrophication (Fabricius, 2011) and the emission of pollutants into the atmosphere causing global changes are impacting reef ecosystems drastically (Hughes et al., 2018). Such increasing anthropogenic impacts has gained worldwide attention, leading to a growing concern about the conservation of reef systems (Magris et al., 2018). However, there is still a mismatch between social commitment and conservation actions (Kollmuss and Agyeman, 2010).

Involving the society in the conservation process is critical for its success. One way to promote such engagement is through citizen-based monitoring programs, in which people collect data on a voluntary basis, contributing to scientific research and taking action in conservation practices (Trumbull et al., 2000; Cohn, 2008; Bonney et al., 2009; Goffredo et al., 2010; Earp and Liconti, 2019). Such practice of voluntary enrolment of the public to collect scientific data and monitoring habitats is acknowledged as citizen-science (Cohn, 2008; Bonney et al., 2009). Beyond disseminating scientific knowledge, citizen-science includes the society in the scientific process and recognizes its contribution (Cohn, 2008; Nov et al., 2011), promoting engagement in environmental causes, curiosity, interest and responsibility in the subject, which may change behaviors and translate into conservation actions (Trumbull et al., 2000, 2005; Evely et al., 2011). Integrating volunteers in data collection allows the formation of large networks that take part of multiregional efforts, including different habitats and organisms (Costello et al., 2010; Teleki, 2012), which often improve the spatial and temporal scope of scientific monitoring (Cohn, 2008; Cerrano et al., 2017).

Despite the broad potential and applicability, and the rise of citizen-science initiatives in several fields all over the world (Cohn, 2008), the credibility and quality of data generated by citizen scientists have been questioned due to the non-scientific nature of people engaged on such programs (Cohn, 2008; Burgess et al., 2017; Freitag et al., 2016; Lukyanenko et al., 2016). Some citizen-based programs can generate data that are vague, variable and difficult to properly be included and published in traditional research (Cohn, 2008; Burgess et al., 2017), as a result of issues related to lack of accuracy and precision, small sample size and poor standardization of protocols (Lewandowski and Specht, 2015). A number of practices can be incorporated to citizen-based programs to ensure that the data will comply with the standards required by the scientific community. First, it is necessary to create an easy-to-follow protocol for data collection that demands realistic tasks and accounts for the volunteer´s prior experiences (Cohn, 2008; Freitag et al., 2016; Lukyanenko et al., 2016). Such protocol should also be flexible so the citizen-scientist may contribute to information other than the ones requested, which may help in evaluating the quality of the data provided (Lukyanenko et al., 2016). Volunteers can also be trained on data collection, including training in how to manipulate scientific equipment, to process and send the data if necessary (Cohn, 2008; Freitag et al., 2016). Leaders of citizen-based science should also establish communication channels with volunteers throughout the entire process, facilitating in situ staff guidance and data curation (Cohn, 2008; Freitag et al., 2016). Another powerful tool to ensure the quality of citizen-based data is to compare it to data generated by trained scientists using the same collection protocol (Cohn, 2008; Freitag et al., 2016; Lukyanenko et al., 2016). The implementation and association of these methods increase the chances of citizen-science data to be trusted by academic scientists, and consequently to be integrated to their researches and published, bringing more visibility to citizen-science (Burgess et al., 2017).

Citizen-science in marine systems is increasing in number of initiatives, scale and diversity of habitats and target taxa (Thiel et al., 2014; Earp and Liconti, 2019). In reef ecosystems, the participation of voluntary divers has already promoted important advances in data collection, generating a broader monitoring of coral and sponge biodiversity (Bell, 2007), abundance of several species of fish (Hodgson, 1999; Goffredo et al., 2006), distribution range of some species (Heard et al., 2019), occurrence of invasive species (Delaney et al., 2007), and large biodiversity assessment at continental scales (Stuart-Smith et al., 2017). The “Reef Life Survey” initiative, for instance, have assembled an impressive global database on reef biodiversity collected by SCUBA divers engaged in the citizen-science program (Edgar & Smith, 2014), which have been rigorously analyzed and published by scientists in high-impact journals (e.g.Duffy et al., 2016; Strain et al., 2019). A drawback of some of these initiatives is that they often require some level of previous training and/or a complex dive logistics that allows the use of transects or similar method (Stuart-Smith et al., 2017), usually meaning that the dive company has to change their routine operation. Despite resulting in good quality data, these approaches also restrict the reach and applicability of monitoring protocols in regular diving operations, both because not all recreational divers are experienced or willing to engage in previous training and not all diving companies are able to change their operations. Simplifying these protocols and adapting them to the routine of diving companies would likely expand its application, which is particularly important for monitoring multiple areas through time when research budget may be a constraint. We developed and tested a citizen-based monitoring protocol to monitor reef fish and sea turtles´ abundance during regular recreational diving operations at a tropical reef in Northeast Brazil. The effectiveness of the protocol was assessed by comparing data collected by volunteer divers and by trained scientists in the same area and season. We also assessed the influence of the volunteer diving experience in data collection and their level of satisfaction with the proposed protocol.

Volunteer divers were able to record all the species/groups present in the slate, counting on average 26.1 (±3.2) individuals of 7.8 (±0.4) species/groups per dive. From the 1,139 organisms recorded, around 9% were out of the size range published in the literature and all cases corresponded to underestimates (Fig. 2). Diving experience was well distributed among classes, with 40% of the 37 recreational divers rated as experienced divers, 30% being rated as beginners and 30% intermediate. The composition of the assemblage recorded by a subset of 33 recreational divers (11 in each category) was consistent regardless of diving experience level (PERMDISP: P = 0.317/PERMANOVA: Pseudo-F2,30 = 1.61; P = 0.108; Fig. 3).

Fig. 2.

Data collected by recreational divers using the monitoring slate. Total and per size class abundance (mean ± SE) recorded for the species/groups present in the monitoring slate.

(0.62MB).

Fig. 3.

Effects of diving experience on data collection. nMDS plot comparing the assemblages recorded by recreational divers with different diving experience: beginner (white), intermediate (light blue) and experienced (dark blue).

(0.06MB).

Among the 37 recreational divers, 77% recorded anthropogenic debris on the reef, mostly represented by ship debris (54%) or fishing gear (35%), while only 4% and 7% were plastic and other types of waste, respectively. Most recreational divers rated their experience as citizen-scientists as excellent (63%), 32% ranked as good, 5% as neutral, and none of the participants rated their experience as bad or terrible.

The lack of scientific experience did not affect the data collected by recreational divers, with no difference observed in total abundance (Shapiro-Wilk: P = 0.132 / Levene: P = 0.257/T-test: t = 0.24, df = 12, p = 0.813; Fig. 4A), size distribution (Shapiro-Wilk: P = 0.182/Levene: P = 0.001/ANOVA: Diver Type – F1,48 = 0.08, p = 0.784; Size Distribution – F3,48 = 20.92, p < 0.001, DT × SD – F3,48 = 0.64, p = 0.596; Fig. 4B) and assemblage composition (PERMDISP: P = 0.123 / PERMANOVA: Pseudo-F1,12 = 1.91, p = 0.141; Fig. 4C) observed when compared to data collected by scientific divers.

Fig. 4.

Effects of previous scientific training on data collection. Total (A) and per size class (B) abundance (mean ± SE), and the nMDS plot for assemblage structure (C), comparing assemblage data recorded by scientific (orange) and recreational (blue) divers. For comparisons between types of divers, ‘ns’ stands for non-significant effects. For comparisons among size classes, groups sharing the same letter are not statistically different.

(0.25MB).

We found that recreational divers could successfully identify, estimate abundance and size class of fish and turtles to characterize the reef assemblage, despite their lack of prior scientific experience. In addition to effectively recording all the species presented in the monitoring slate, volunteer divers were able to record data of large species such as rays, sharks and turtles, that are important for both reef ecosystem functioning and conservation. More traditional scientific survey methods, such as transects or stationary censuses, commonly underestimate these large-sized species as discussed in a previous study in the area (Morais et al., 2017). These results indicate that the citizen-based approach proposed here can complement traditional scientific monitoring (Cohn, 2008), corroborating our hypothesis that recreational divers can be powerful citizen-scientists. The use of different methodologies provides a more comprehensive view of reef assemblages optimizes data collection and, combined, generate more refined monitoring data (Schmitt et al., 2002; Baker et al., 2017; Bosch et al., 2017). Additionally, recreational divers characterized fish assemblages successfully regardless of their diving experience and provided a positive feedback of their experience as scientists. Therefore, there is a great potential for implementing similar monitoring protocols in other reef areas with recreational diving activities, particularly in marine protected areas where this activity is allowed and regulated, contributing to the popularization of citizen-science initiatives as a good tool to biodiversity monitoring (Burgess et al., 2017).

Our results showed no difference between recreational and scientific divers conducting the proposed protocol, but some departures arise when we contrast them with previous traditional surveys conducted by scientists in the same area (Morais et al., 2017). We are aware that different methods are not always comparable as diversity and abundance are related to sampling area effort (Schmitt et al., 2002), however a previous study in the area using 40 m2 belt transect surveys (Morais et al., 2017) counted ∼9.5 times more individuals of all the species selected for this study (standardizing by area and effort). Such contrasting pattern may have resulted from volunteers underestimating parrotfishes due to their diverse colors and juveniles that are quite different from adults, making it difficult for volunteers to recognize and count them. These results reflect the long training process of scientific divers, which improve their ability to properly identify, count and estimate size of all organisms during underwater surveys (Thompson and Mapstone, 1997). In some cases, recreational divers in our study also underestimated the size of individuals (9% of the 1,139 individuals recorded; Humann, 2008). This occurred for more cryptic species such as the green moray, Gymnothorax funebris, which stays in burrows during the day and may exhibit nocturnal behavior (Carvalho-Filho, 1999) impairing the size estimates. This was also the case for large-sized species such as Caranx spp., Lutjanus jocu and Sphyraena barracuda, which generally avoid divers, possibly because they are targeted by fishing activities (Floeter et al., 2006; Lindenmayer and Likens, 2009; Lindfield et al., 2014). Because recreational divers are not trained to estimate sizes underwater, the distance from the fish can make it difficult to properly estimate size, commonly resulting in underestimation (Bozec et al., 2011). Despite the ruler on the monitoring slate and the previous instructions, recreational divers show calibration errors. The support of scientific divers to recreational divers during and after the monitoring activity, allied to the use of novel technological tools, photography websites, social media profiles and applications (Newman et al., 2012; Mazumdar et al., 2018) may decrease the distance between the expertise of scientists and volunteers, improving estimates and increasing the reliability of the method (Cohn, 2008; Jiguet, 2009; Bonter and Cooper, 2012; Freitag et al., 2016; Lukyanenko et al., 2016). Also, a longer training to make recreational divers more familiar with the method may also be a good strategy to provide more accurate counts and size estimates (Thiel et al., 2014), ensuring data quality and suitability to scientific standard requirements (Burgess et al., 2017).

Despite the difficulties in counting and estimating sizes, recreational divers were efficient in recording species, particularly larger-sized and high-mobility ones, which are commonly not recorded by traditional surveys in the area (Morais et al., 2017). This possibly reflects the spatial limitation from the underwater visual census methodology (Williams et al., 2006; Lindfield et al., 2014). Larger species that have greater mobility and those targeted by fishing tend to fall out of the visual field surveyed by the scientific diver and are not recorded (Schmitt et al., 2002; Bozec et al., 2011; Dickens et al., 2011; Lindfield et al., 2014). This limitation makes traditional methods more efficient in the detection of species with low mobility (Bosch et al., 2017) as opposed to those with higher mobility, in the case of this stydy: the nurse shark Ginglymostoma cirratum, jacks Caranx spp. and rays Hypanus americanus and Hypanus marianae. Recreational divers in our study recorded these highly mobile species that are critical to the reef system functioning, in contrast to other studies in the area that used the transect survey method (Morais et al., 2017). Sharks, rays and turtles are considered charismatic (Frazier, 2005; Gallagher and Hammerschlag, 2011) and of interest to recreational divers, that tend to actively search for these species (Gallagher and Hammerschlag, 2011). This highlights important benefits from combining scientific and citizen-based approaches for monitoring reef communities, and from including flagship species to stimulate citizen engagement (Schmitt et al., 2002; Walpole and Leader-Williams, 2002; Cohn, 2008; Smith and Sutton, 2008).

Occurrence data generated by recreational divers may be important to identify population changes because the frequency of records may be directly related to changes in the number of individuals of a given species (Bender et al., 2014), as already observed by citizen-scientists monitoring monarch butterflies (Cohn, 2008). This becomes even more important for endangered and flagship species, which were efficiently recorded by recreational divers in our study (e.g. sea turtles and sharks). The colors and shapes of reef fishes, as well as the large sizes of sharks and sea turtles facilitate its identification to the species level, particularly when compared to other taxa with less evident traits such as some insects (Bonney et al., 2009; van Strien et al., 2013). In the largest citizen-based database on reef biodiversity, for example, nearly all of the 1,400 fish recorded were identified to the species level, largely because unique species traits enabled citizen-scientists to identify species using field guides and digital photographs that could also be sent to specialists to confirm it (Edgar and Smith, 2014). Proper species identification of fishes is a key advantage for the development of citizen-based monitoring programs in reef environments. Also, because recreational diving activities are well-stablished in several reefs worldwide and occur on a daily basis, they can be important allies to supplement the scientific data collection allowing a broader spatial and temporal monitoring of reef environments (Mason et al., 2018), often constrained by funding, particularly in developing countries that harbor great marine diversity. Recreational divers enjoyed their experience as citizen-scientists, collecting useful data regardless of their diving experience, reinforcing the great potential to implement similar protocol in other diving areas.

We demonstrated that recreational divers are efficient and committed when included in the scientific process. They can not only contribute to increase sampling effort, but also to supplement the scientific monitoring of reef environments when conducting standardized monitoring protocols, as the one we developed and implemented. More than turning citizens into scientific partners for reef monitoring, citizen-science activities are able to transform simple recreational interest into responsibility, bringing society closer to reef environments and promoting interest to preserve these ecosystems and their associated species (Trumbull et al., 2005; Cohn, 2008; Evely et al., 2011; Roy et al., 2012). Despite being developed for reef monitoring, our protocol has the potential to be adapted and applied in other systems and for other taxonomic groups as it has achieved the two mandatory issues that ensures the credibility of a citizen-based monitoring: (I) volunteer interest and enrollment by using an attractive design and considering charismatic and common species; (II) data quality by volunteer training, in situ scientific support and validation by comparison with data generated by scientific experts. We believe that the protocol presented in this paper can inspire data collection during recreational activities in other habitats and with other organisms, contributing to increase the credibility and popularity of citizen-science initiatives and reinforcing how powerful this tool can be to increase the scale and scope of scientific research.

We acknowledge that this work is original; has not been published or accepted for publication, nor is being considered for publication elsewhere; all the authors have read and approved the submitted version of the manuscript; all listed authors are entitled to authorship and agreed to be listed; all funding agencies have been properly acknowledged, had no role in the study conception, design, or results, and do not have a direct or indirect financial interest in the subject matter discussed in this manuscript. We have no conflicts of interests to declare.

This work was supported by the Serrapilheira Institute (Grant numbers :Serra-1708-15364 and R-1805-24275 awarded to GOL). Also, this study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) - Finance Code 001, with a post-doctoral scholarship awarded to EAV.