Electronic Citizen Science Platforms as Methods to Collect Data from Recreational Fisheries; Strengths and Limitations

The research conducted in relation to this thesis was aimed at investigating strengths and current limitations of electronic citizen science platforms as a method to collect data from recreational fisheries. Traditionally, data collection from recreational fisheries is a logistical and financial challenge due to the diffuse and diverse nature of the sector. Therefore, data tend to be lacking for many water bodies. The general lack of data is a hindrance for sustainable fishing and healthy fish stocks through viable management plans. Combined, the challenges of data collection, the importance of data, and the widespread lack of it highlight the need to rethink data collection approaches in recreational fisheries. A potential solution is offered through citizen science, which makes it possible to conduct cost-effective research on large temporal and spatial scales via members of the public (e.g., as data collectors). The potential of citizen science platforms is widened by the increasing availability of internet and smartphones. Electronic citizen science platforms for anglers working as smartphone applications (apps) and/or webpages are relatively cost-efficient and have the potential to supplement or replace traditional recreational fisheries survey methods. Yet, these platforms for anglers require validation to be successfully integrated in recreational fisheries management and research, e.g., with respect to participation and data quality. The need for validation and research is furthered by a likely inclusion of electronic platforms, as a tool to collect data from recreational fisheries in multiple countries, in the near future (i.e., 5-10 years). Indication has been given that angler apps and electronic citizen science projects for anglers can complement traditional survey methods, however, little is known about electronic citizen science platforms for anglers as data collection tools. This thesis investigates aspects of participation (i.e., recruitment and retention), explores quality of recreational fisheries data, and looks into research opportunities using a novel data collection approach, i.e., data generated by anglers via the electronic citizen science platform Fangstjournalen (see appendix A for a detailed description of the platform).

The first paper (MS I) focused on characterizing citizen science participants and their retention patterns (i.e., how long they stay engaged) on the electronic citizen science platform Fangstjournalen. The first part of MS I was based on data from an onsite survey that involved interviewing anglers while they were fishing (i.e., roving creel survey), which was conducted to investigate the coastal recreational fisheries on the Danish island of Funen. Using a roving creel survey approach, it was possible to ask anglers whether they were using the Fangstjournalen platform and thereby identify citizen science participants and non-participants in a population of anglers. Subsequently, participant characteristics (i.e., demography and recreation specialization) and catch rates (i.e., fish per trip) of participants and non-participants of the citizen science platform were investigated. The results suggested that citizen science participants were younger, more specialized anglers, and had higher catch rates compared to non-participants. The second part of MS I used data from the Fangstjournalen platform to explore retention patterns and thereby the level of sustained engagement after recruitment to the platform. Additionally, various characteristics of the retained participants were investigated to determine whether a specific segment of anglers used the platform continually. Results indicated that ~ 21% of the participants that contributed with data (i.e., register fishing trips) stayed engaged on the citizen science platform for more than three months after recruiting. Additionally, ~10% of the participants that contributed with data stayed engaged for more than twelve months. Compared to participants who stopped using the platforms within three months of recruitment, the sustained participants were older and considered angling a more important hobby.

In conclusion, the findings from MS I suggest that citizen science participants are a specific segment of the angling population (i.e., younger, more skilled, and more specialized), which points towards electronic citizen science platforms not being representative of the general angling population. Consequently, extrapolation of catch and effort from citizen science platforms to population level might be an issue. The findings from MS I also indicate that citizen science platforms for anglers can benefit from specific measures to increase retention and suggest a specific segment to target (i.e., younger and less committed anglers).

Paper two (MS II) investigated the quality of catch and effort data generated by the electronic citizen science platform Fangstjournalen. Additionally, MS II explored whether the demography of citizen science participants potentially differ from participants in other type of recreational fisheries surveys. In this study, data collected from the citizen science platform was compared to three traditional survey methods, namely a roving creel survey, a recall survey, and an aerial survey. These data comparisons were performed for a specific fishery within a limited temporal and spatial frame, i.e., the sea trout fisheries on the Danish island of Funen in spring 2017. Results from MS II indicated that the Fangstjournalen platform provided catch and effort data that were similar to the data generated from two concurrently conducted surveys (i.e., roving creel survey and recall survey). There were similarities in estimates between survey methods, despite some demographic differences (e.g., seen by a disproportionate distribution of tourist anglers on the Fangstjournalen platform compared to the roving creel survey). However, when focused only on Danish participants, age and geographical distribution of residences were similar between the citizen science platform and the creel survey. Findings from MS II also indicated a relationship between anglers counted during the aerial survey and number of citizen science participants that were active during the aerial survey, i.e., during the specific time slot on a given survey day where aerial counts were conducted. This result illustrated a possibility of using the number of active citizen science participants as a proxy for total angling effort in a limited area and for specific fisheries.

Overall, the findings from MS II suggest that electronic citizen science platforms can supplement traditional survey methods or, in certain cases, act as a standalone method, e.g., for the Funen sea trout fishery within a limited temporal frame. However, the results also indicate that participant characteristics of citizen scientists may not reflect that of participants captured via traditional surveys.

MS III was in part an extension of MS II. In this paper, three additional years were included to be a part of the catch and effort comparison between citizen science and recall survey data from the Funen sea trout fishery. In addition, citizen science and recall data were temporally and spatially compared for other Danish fisheries, such as the coastal sea trout fishery, the coastal/offshore cod (Gadus morhua) fishery, and the freshwater pike (Esox lucius) fishery. The comparisons for the Danish coastal sea trout fishery and coastal/offshore cod fishery were based on catch data (i.e., catches and release proportions) collected within each of the eight distinct ICES management areas that cover the Danish coastal and offshore areas. In these fisheries, catch data from different survey methods were compared for ten specific half-year periods from 2016 to 2020. Similarly, catch and effort data from the freshwater water pike fishery were contrasted in ten half-year periods from 2016 to 2020. The findings from MS III indicated that estimates of catch and effort from the two survey methods were similar in the Funen sea trout fishery and that the similarities were consistent over time (i.e., temporal consistency). Equivalently, similarity and temporal consistency of catch and effort estimates were found for the Danish freshwater pike fishery. In contrast, large differences were found when catch data was compared temporally and spatially (i.e., ICES management areas) for the Danish coastal sea trout and coastal/offshore cod fishery. Here, the catches from the recall survey were generally larger (i.e., 200% – 300%) compared to the recall survey. Additionally, a larger proportion of fish were being released on the citizen science platform compared to the recall survey (i.e., 20% -30%). These differences between survey methods were fairly consistent over time, which suggested that it was possible to identify a conversion factor between estimates. It is worth noting that sample size was low for the Danish coastal/offshore cod fishery on the citizen science platform.

In conclusion, the findings from MS III suggest that citizen science platforms can generate catch and effort data that are similar to the data of a traditional survey method, i.e., a recall survey, and that these similarities seem consistent, at least over 4- 5 years. Yet, these findings only relate to certain fisheries (i.e., the Funen sea trout fishery and freshwater pike fishery). Low sample sizes for the coastal/offshore cod fishery clearly indicate that certain fisheries are not captured by the citizen science platform. The results from the Danish coastal sea trout fishery indicate systematic bias in either recall survey data (e.g., framing of questions, number of recall periods, and length of the recall periods), citizen science data (e.g., underreporting caused by not reporting certain parts of the catch, such as smaller fish), or possibly both.

Understanding human dimensions (e.g., angler heterogeneity and angler behavior) are central components in modern management of recreational fisheries. Recreational fisheries are coupled social-ecological systems that rely on information about the fish (e.g., catch, effort, biological information) and the people that fish for them (e.g., behavior) for successful management. Paper four (MS IV) and five (MS V) highlight how electronic citizen science platforms can advance the exploration of human dimensions, e.g., by offering a novel way to collect research data.

Paper four (MS IV) was made as a response to the Covid-19 pandemic. The covid-19 pandemic caused a lockdown in spring 2020, where citizens to a wide extent were restricted to stay at home. The pandemic caused an abrupt change in the life of Danish citizens, which, for example resulted in a 20% increase in the sale of fishing licenses during the lockdown compared to the same period in previous years (i.e., spring 2018 and 2019). The increased sale of fishing licenses was interpreted as a heightened interest for recreational fishing during the Covid-19 lockdown and thus an increase in participation. In response, participant characteristics (e.g., demography and importance of angling as a hobby) of anglers recruited to the citizen science platform (i.e., created an account) during the lockdown were compared to anglers recruited in the same period in previous years (i.e., spring 2017, 2018, and 2019). Additionally, participant characteristics of data providers (i.e., participants that register fishing trips) recruited during the lockdown period were compared to data providers recruited in the same period in previous years. The comparison of data providers also included catch and effort
patterns (i.e., catch rates, release rates, and daily and hourly effort patterns). The results indicated that participants recruited during the lockdown were younger, less experienced as anglers, more likely to live in urban areas, considered angling a less important hobby, and were less likely to be from outside of Denmark compared to participants recruited in previous years. Most of these factors disappeared, when participant characteristics of data providers were compared between the periods, with the exception of angling experience. Compared to previous years, data providers had relatively higher effort on weekdays and fished later in the day during the lockdown period. Additionally, they had lower catch rates and were more inclined to harvest fish compared to previous years.

The findings from MS IV provide some insights into potential effects of the Covid-19 lockdown for the Danish recreational fisheries and indicate the value of citizen science platforms to collect recreational fisheries data, e.g., as a supplement to onsite surveys in times of travel restrictions and closed public offices. However, it also clearly demonstrates some of the potential biases in the data that electronic citizen science platforms aimed for anglers produce. For example, a sudden change in participation can be reflected in the generated data. Additionally, the transition from recruit (i.e., account creation) to data provider (i.e., fishing trip registration) seems to be done by a homogenous group of anglers, which could be another indication that data providers from citizen science platforms are not representative of the general angling population as also shown in MS I.

Paper five (MS V) investigated the potential of an electronic citizen science platform as a tool to collect data related to angler satisfaction, an important human dimension aspect, in recreational fisheries. In this study, fishing trips satisfaction from the citizen science participants was investigated using a human dimensions (HD) survey build into the Fangstjournalen platform. The HD survey was programmed to appear after fishing trip completion in approximately one of seven fishing trips that anglers reported to the platform. During the survey, citizen science participants were first asked to state their motive for fishing, i.e., “Why did you fish today?” with six different responses. Second, participants were asked to state their satisfaction, i.e., “How satisfied were you with the trip?” on a ten-point Lickert scale, from one being very dissatisfied to ten being very satisfied. Using this survey, several drivers related to fishing trip satisfaction were determined. Some of the findings were consistent with findings from other studies. For example, catch proved to be a strong driver of satisfaction. However, findings from MS V also revealed several additional drivers for fishing trip satisfaction, e.g., the motivational setting of a fishing trip moderated the satisfaction derived from catch. For example, anglers who were motivated by catch-related aspects (e.g., to catch a trophy fish) received higher satisfaction from catching fish compared to anglers who were motived by non-catch related aspects (e.g., to experience nature). Another novel finding was that fishing trips at higher air temperature caused somewhat higher satisfaction levels, indicating that fishing trips happening in more pleasant environments provide higher satisfaction. Importantly, an effect of recall period (i.e., time between conducting and submitting a fishing trip) was also found. This effect indicated that as time between conducting a fishing trip and submitting it the platform increased, so did satisfaction.

The findings from MS V provide novel insights into fishing trip satisfaction for anglers, e.g., by relating recall period, air temperature, motives, and catch to satisfaction. Additionally, this highlights the potential of electronic citizen science platforms as diverse data collection tools (e.g., by capturing recall period and weather automatically) and showcases the usefulness of citizen science platforms in a human dimensions context.

Together, the results from the papers presented in this PhD-thesis provide new insights into the strengths and current limitations of electronic citizen science platforms as tools for data collection in recreational fisheries. Electronic citizen science platforms are likely to attract a segment of the angling population that are younger, more specialized, and have higher catch rates compared to the general angler, which may be a problem for extrapolation to population level (MS I). Retention rates of ~20% after three months and ~10% after twelve months could be another source of bias (MS I). In addition, the participants that provide data and the participants that provide data continuously (i.e., stay engaged for more than twelve months) are also likely to be specific segments of anglers (MS I; MS IV). The specific angler segments do not engage on electronic citizen science platforms and the segments that stop using them (e.g., by a failure to transition from recruit to data provider) offer potential targets in measures to attract, engage, and retain participants (MS I; MS II; MS IV). Despite the potential sources of bias related to participation, findings in this thesis indicate that electronic
citizen science platforms still can generate catch and effort estimates that are similar to those of traditional survey methods (MS II), and that the similarities can be consistent over a four-to-five-year period (MS III). Nonetheless, it is also indicated that citizen science platforms might be limited to certain fisheries (MS III). Electronic citizen science platforms also hold value for collecting data that can further the knowledge of other aspects of recreational fisheries, e.g., angler behavior (MS IV; MS V) and human dimensions research (MS V). For example, it was possible to detect participational changes and associated behavioral changes in the Danish recreational fisheries during the Covid-19 lockdown in spring 2020 (MS IV) and identify several known and novel drivers for angler satisfaction (MS V) using the Fangstjournalen platform.

Combined, the findings from this thesis highlight some of the strengths and current limitations of electronic citizen science platforms for anglers and provide a starting point for future research. From the work conducted in relation to this thesis, it is indicated that the strengths relate to data quality and research opportunities (i.e., the ability to generate recreational fisheries data similar to that of traditional survey methods in some fisheries and the capacity to further research via a novel data collection approach). Most of the associated limitations relate to participation, e.g., the inability to recruit participants that are representative of the general angling population, the inability to reflect the participants captured in traditional surveys, low retention rates, and a failure to retain and receive contributions from certain angler segments.

Future research should extend from the central topics from this thesis. Specifically, participation, data quality, and research opportunities. The research into participation from this thesis has given some insights and subsequently new questions relating to recruitment and retention of participants. For example, demographic differences among citizen science participants and participants from traditional surveys does not necessarily seem to affect data quality (i.e., they generate similar estimates of catch and effort), which question the role of demographics in validating electronic citizen science platform. Hence, future research could focus on determining the role of different segmentation methods (e.g., demography, recreation specialization, consumptive orientation, and catch orientation) in their ability to capture variation associated with recreational fisheries data (e.g., how catch rates are affected by different angler segments engaging in different surveys). Speaking of segmentation, future work should also focus on methods to ensure recruitment of tourist anglers, recruitment and retention of less committed anglers, and recruitment of anglers from fisheries that so far are poorly represented on citizen science platforms. Such research is likely to benefit from inter-disciplinary approach (e.g., social and design sciences) and could involve platform design and advertisement, e.g., by determining the number of data entries different segments are willing to register from a fishing trip.

Further research should also focus on further data quality evaluations. Findings from this thesis indicate that the Fangstjournalen platform can supplement or replace traditional surveys but only in some fisheries. From a data quality perspective, this indicates that fishery specific evaluation is a necessity. This could for example, be in relation to participation and platform design in order to determine why electronic citizen science platforms works in some fisheries but not in others. Another important research area is biological information generated by anglers using electronic citizen science platforms. On one hand, it is important to understand how well citizen scientist can identify species, genders, and stages of maturity as well as, determine accuracies of lengths and weights. One the other hand, it is possible that most of these features can be done automatically, e.g., via image recognition software. Novel use of technology could both lessen the burden for participants and serve as an educational tool for anglers.