Using Data Science to Enhance Great Barrier Reef Conservation
Coral reefs are home to some of the most diverse ecosystems on earth (Ortiz et al. 2018). The marine ecosystem of the Great Barrier Reef (GBR), in particular, is recognised as one of global significance, in regard to its biodiversity and beauty, as well as its environmental, cultural, social, and economic value (McCook et al. 2010). Pressures from global climate change (Vercelloni et al 2010), as well as pollution and overfishing, are causing coral reefs to die at an increasingly alarming rate (Anthony et al., 2015); coral bleaching events are becoming more frequent, and with less time in between for coral to recover (Baker et al. 2008). If coral reefs are lost, it would not only lead to a terrible loss of biodiversity, but also to an economic disaster, with millions of people depending on them for income, fisheries, tourism, and coastal protection (Anthony et al., 2015).
It is becoming apparent that the GBR is facing many complex challenges that are making traditional conservation methods unable to support the effective protection of the ecosystems on the reef (Vercelloni et al. 2010), and this paper will discuss three of these challenges. The first is how, due to its immense size, scientists are struggling to get a detailed picture of the health of the GBR as a whole. The second is how current methods to assess the health of coral reefs in the GBR can’t keep up with the pace at which the conditions are changing. The last challenge discusses how imperfect compliance of zone restrictions in Marine Protected Areas (MPAs) prevents adequate protection of the ecosystems within them. It is essential to think innovatively in order to overcome these challenges and adequately protect the GBR.
Using Drones to Get a Better Picture of Reef Health
With around 3000 individual reefs, the Great Barrier Reef Marine Park spans over 344,000 km2 (Ortiz et al. 2018). The immense size of the area, and the extent of the reef that is being degraded, is proving to be a big challenge for marine scientists and policymakers alike when it comes to prioritising specific areas of the reef that are in most need of help (Bellwood et al. 2019). Traditional methods for assessing the health of the reef involve a mixture of in-water survey images, which have great detail but can’t provide the large quantities of data points that are needed (Kwan, 2020), and satellite images, which can provide thousands of data points but are hard to interpret due to low resolution and cloud coverage issues (Boyd, 2019). Because of the issues in these methods, the information that they provide is not sufficient enough (Kwan, 2020) to allow for the types of concise decision-making and actions that are required for effective conservation of the GBR.
This challenge has driven researchers to come up with new ways to obtain the amounts of data that they require (Madin et al. 2019), and with enough detail to support effective conservation efforts. An innovation that can achieve this, is the use of drones fitted with hyperspectral cameras (Parsons et al. 2018). By collecting and processing information from the electromagnetic spectrum, hyperspectral cameras can collect thousands of data points about the particular area of a reef that’s in a single image (Caras & Karnieli 2015); a vast improvement on the 30–40 data points from in-water survey images (Boyd, 2019). And the drone technology eliminates the cloud coverage issue that satellites face (Simic Milas et al. 2018). The data that is captured by these drones is able to differentiate between coral, algae, and sand (Parsons et al. 2018). It can also determine the specific type of coral, and provide information on its condition, including whether it is healthy or dying (Boyd, 2019). However, the improvement in resolution and amount of data points of hyperspectral images comes at a cost. Hyperspectral datasets are large and multidimensional, and therefore require substantial data storage capacity, so analysing this data is a complex and costly process (Caras & Karnieli, 2015).
The combination of drones and hyperspectral cameras is allowing researchers to assess a greater area of the reef than was previously possible, and with much more accuracy (Caras & Karnieli, 2015). By having this more detailed data from more of the reef, scientists and policymakers are able to make better-informed decisions and effectively allocate resources to the areas of the GBR that need them most.
Automating Image Analysis with Artificial Intelligence (AI)
Traditionally, researchers have been manually analysing in-water survey images to assess the health of the coral reefs; but analysing and labelling these images manually is much too time-consuming to keep up with the rapidly changing conditions of the reef (Baker et al. 2008). This makes it difficult for scientists and policymakers to make informed decisions about managing the GBR, and as a result, it is hindering preservation efforts (Thompson & Dolman, 2010). Hence a new method of image analysis is needed that will allow them to monitor the reef’s conditions accurately and in real-time. This motivated the development of automated image analysis.
Marine experts joined forces with data scientists, and together they used a combination of marine science expertise and machine learning technology, to develop an automated image analysis program (Williams et al. 2019). This program is able to obtain relevant information about things, such as reef composition and coral cover (Hamylton et al. 2020), and provide researchers with real-time analytics and trends from the data (Williams et al. 2019). By automating the process, this new approach allows for more advanced image analysis that is quicker, in greater detail, and at a much larger scale than was previously possible. Having better and more readily available information will allow for smarter decisions about protecting the GBR being made.
One example of this new approach is a project from the Australian Institute of Marine Science (AIMS) called ReefCloud (Hardisty, 2020). ReefCloud seeks to machine-learn from the marine science experts at AIMS and combine this with the use of artificial intelligence (AI), to automate the analysis of in-water survey images (Hamylton et al. 2020). This use of machine learning and AI enables ReefCloud to extract relevant data from massive amounts of images and provide relevant information on the conditions of large areas of coral reefs (Hamylton et al. 2020). By speeding up the process to deliver real-time and larger-scale analysis, this innovation aims to help the management of coral reefs by increasing efficiency, reducing resource limitations, and producing robust and timely reports (Hardisty, 2020). Allowing for smarter, data-driven decisions to be made, leading to more efficient management of the GBR.
Even with this advanced capability to analyse the images, this innovation is still subject to issues. One of which, is related to the physical collection of the images, and how using human divers to manually capture the images can disrupt marine life. The quantity and variety of fish in coral reef habitats are an important indicator of the overall health of the reef (Emsile et al. 2017), and so disrupting the natural state of an area can cause incorrect conclusions to be drawn. An example of a way to overcome this issue is demonstrated in a project from the Philippines called CORaiL. Where they used strategically placed cameras equipped with a Video Analytics Services Platform, created by Accenture, to detect, count, classify and photograph fish; and the data from the cameras is sent to the researchers, providing real-time analytics and trends (Business World 2020).
Upgrading Surveillance and Enforcement of Marine Protected Areas (MPAs)
Marine Protected Areas on the GBR have achieved many positive outcomes, such as increased coral cover in areas, increased population sizes of target fish, as well as providing many other ecological benefits (Bergseth et al. 2015). MPAs are fairly effective at limiting activities that are detrimental to the area, with evidence showing that no-take zones generally have higher fish biomass than the fished zones (McCook et al. 2010). However, the total effectiveness of MPAs relies heavily on total compliance (Bergseth et al. 2015), which is not currently being achieved. This is apparent in the significant discrepancies between the no-take zones and the no-entry zones, indicating the probability of substantial illegal fishing activity in no-take zones (McCook et al. 2010). A reason for this is how difficult it is to enforce the no-take restrictions in such large areas with the traditional surveillance and enforcement methods (Wilhelm et al. 2014), which involve using satellite vessel monitoring systems (VMS) and vessel-based surveillance. VMS doesn’t fully support real-time enforcement (Molenaar et al. 2000), and the resources that are currently available for vessel-based surveillance cannot cover enough area to provide total compliance (Davis et al. 2004). With no-take zones already being one of the more difficult zones to enforce, the rezoning of the Great Barrier Reef marine park in 2004, which increased the area of no-take zones from 4.6 percent to 33.4 percent, made it even more challenging for marine park management to conduct effective surveillance and enforcement of MPAs (Mellin et al. 2016).
The challenge of enforcing total compliance in MPA zones created an opportunity to use a combination of innovative satellite technologies and unmanned vehicles. Different satellite systems each have their own benefits and limitations (Cervera et al. 2011). Satellite-based Automatic Identification Systems (AIS), collect data on a vessel’s position, identification, speed, and course (Cervera et al. 2011). It is a useful tool for detecting illegal fishing in no-take zones because most vessels are fitted with an AIS device, and they are actually mandatory for the majority of vessels (Fournier et al. 2018). But AIS alone isn’t able to be totally effective in tracking illegal fishing activity, as it is possible for vessels to disable them, or to manipulate their data to show false coordinates (Fournier et al. 2018). Synthetic Aperture Radar (SAR) sensors can help combat this issue, as they are able to detect the presence of vessels even when they aren’t transmitting AIS (Graziano et al. 2019). Traditional vessel-based surveillance doesn’t have the resources required to cover enough ground to catch all the illegal activity (David et al. 2004), but Unmanned aircraft systems (UAS) and unmanned surface vehicles (USVs) are two innovations that can help to alleviate this issue. Supplied with coordinates of suspected illegal activities, from the satellites mentioned above, USVs can be sent out to the exact location to investigate (Turnbull, 2017). And UAS’s, like the Flexrotor, designed by Aerovel, fitted with electro-optical sensors for daylight use, and infrared cameras for night-time use, can conduct surveillance virtually undetected because of their small size (Ball, 2015). There are some issues with the implementation of these technologies, however, including the amount of time and resources that are required to train staff in the proper operation of these new highly technological systems (Davis et al. 2004).
Using a combination of AIS, SAR and VMS allows the satellites to accurately gather real-time data on suspected illegal activities (Graziano et al. 2019). This, along with these highly sophisticated unmanned vehicles, allows for efficient detection and investigation of illegal fishing in no-take zones and will provide a higher success rate for enforcement (Turnbull, 2017). Meaning that MPAs will be able to operate as intended and contribute to the restoration of the GBR.
Conclusion
The impact of losing the Great Barrier Reef, and other coral reefs around the world, would have such a devastating impact on the world, that there is no option but to continue to strive for a greater understanding of these complex ecosystems, as well as the threats that they face. As shown in the few challenges that are discussed in this paper, it is highly likely that coral reef conservation will continue to require many more transdisciplinary solutions for the many more challenges it faces. Necessitating the continued use of the best available scientific information from a range of backgrounds, with input from marine managers, researchers, experts, and traditional owners. And while the innovations discussed here present exciting opportunities for new ways to support the management of the reef, there is still a long road ahead in the mission to save coral reefs.
References
Anthony, K., Marshall, P., Abdulla, A., Beeden, R., Bergh, C., Black, R., Eakin, C., Game, E., Gooch, M., Graham, N., Green, A., Heron, S., Hooidonk, R., Knowland, C., Mangubhai, S., Marshall, N., Maynard, J., McGinnity, P., McLeod, E., … Wear, S. (2015). Operationalizing resilience for adaptive coral reef management under global environmental change. Global Change Biology, 21(1), 48–61. https://doi.org/10.1111/gcb.12700
Baker, A., Glynn, P., & Riegl, B. (2008). Climate change and coral reef bleaching: An ecological assessment of long-term impacts, recovery trends and future outlook. Estuarine, Coastal and Shelf Science, 80(4), 435–471. https://doi.org/10.1016/j.ecss.2008.09.003
Ball, M. (2015, September 27). Monitoring a Marine Reserve. Earth Imaging Journal. https://eijournal.com/print/articles/monitoring-a-marine-reserve
Bellwood, D., Pratchett, M., Morrison, T., Gurney, G., Hughes, T., Álvarez-Romero, J., Day, J., Grantham, R., Grech, A., Hoey, A., Jones, G., Pandolfi, J., Tebbett, S., Techera, E., Weeks, R., & Cumming, G. (2019). Coral reef conservation in the Anthropocene: Confronting spatial mismatches and prioritizing functions. Biological Conservation, 236, 604–615. https://doi.org/10.1016/j.biocon.2019.05.056
Bergseth, B., Russ, G., & Cinner, J. (2015). Measuring and monitoring compliance in no‐take marine reserves. Fish and Fisheries (Oxford, England), 16(2), 240–258. https://doi.org/10.1111/faf.12051
Boyd, J. (2019). Drones survey the great barrier reef: Aided by AI, hyperspectral cameras can distinguish bleached from unbleached coral — [News]. IEEE Spectrum, 56(7), 7–9. https://doi.org/10.1109/MSPEC.2019.8747297
Business World (India). (2020). Using Artificial Intelligence to Save Coral Reefs.
Caras, T., & Karnieli, A. (2015). Ground-level classification of a coral reef using a hyperspectral camera. Remote Sensing (Basel, Switzerland), 7(6), 7521–7544. https://doi.org/10.3390/rs70607521
Cervera, M., Ginesi, A., & Eckstein, K. (2011). Satellite‐based vessel Automatic Identification System: A feasibility and performance analysis. International Journal of Satellite Communications and Networking, 29(2), 117–142. https://doi.org/10.1002/sat.957
DAVIS, K., RUSS, G., WILLIAMSON, D., & EVANS, R. (2004). Surveillance and Poaching on Inshore Reefs of the Great Barrier Reef Marine Park. Coastal Management, 32(4), 373–387. https://doi.org/10.1080/08920750490487223
Emslie, M., Cheal, A., & Logan, M. (2017). The distribution and abundance of reef-associated predatory fishes on the Great Barrier Reef. Coral Reefs, 36(3), 829–846. https://doi.org/10.1007/s00338-017-1573-x
Fournier, M., Casey Hilliard, R., Rezaee, S., & Pelot, R. (2018). Past, present, and future of the satellite-based automatic identification system: areas of applications (2004–2016). WMU Journal of Maritime Affairs, 17(3), 311–345. https://doi.org/10.1007/s13437-018-0151-6
Graziano, M., Renga, A., & Moccia, A. (2019). Integration of Automatic Identification System (AIS) Data and Single-Channel Synthetic Aperture Radar (SAR) Images by SAR-Based Ship Velocity Estimation for Maritime Situational Awareness. Remote Sensing (Basel, Switzerland), 11(19), 2196–. https://doi.org/10.3390/rs11192196
Hamylton, S.M., Zhou, Z. & Wang. L. (2020). What Can Artificial Intelligence Offer Coral Reef Managers? Frontiers in Marine Science, 7. https://doi.org/10.3389/fmars.2020.603829
Hardisty, P. (2020, February 10) Submission to the consultation process guiding a new Australian development policy and performance framework. Australian Institute of Marine Science. https://www.dfat.gov.au/sites/default/files/australian-institute-of-marine-science.pdf
Kwan, C. (2019, June 26). How AI and drones are trying to save the Great Barrier Reef. Techepublic. https://www.techrepublic.com/article/how-ai-and-drones-are-trying-to-save-the-great-barrier-reef/
Madin, E., Darling, E., & Hardt, M. (2019). Emerging Technologies and Coral Reef Conservation: Opportunities, Challenges, and Moving Forward. Frontiers in Marine Science, 6. https://doi.org/10.3389/fmars.2019.00727
McCook, L.J., Ayling, T., Cappo, M., Choat, J.H., Evans, R.D., De Freitas, D.M., Heupel, M., Hughes, T.P., Jones, G.P., Mapstone, B., Marsh, H., Mills, M., Molloy, F.J., Pitcher, C.R., Pressey, R.L., Russ, G.R., Sutton, S., Sweatman, H., Tobin, R., … Williamson, D.H. (2010) Adaptive management of the Great Barrier Reef: A globally significant demonstration of the benefits of networks of marine reserves. Proceedings of the National Academy of Sciences, 107 (43) 18278–18285; DOI: 10.1073/pnas.0909335107
Mellin, C., Aaron MacNeil, M., Cheal, A., Emslie, M., Julian Caley, M., & Bellwood, D. (2016). Marine protected areas increase resilience among coral reef communities. Ecology Letters, 19(6), 629–637. https://doi.org/10.1111/ele.12598
Molenaar, E., Molenaar, E., & Tsamenyi, M. (2000). Satellite-based vessel monitoring systems for fisheries management: International legal aspects. The International Journal of Marine and Coastal Law, 15(1), 65–109. https://doi.org/10.1023/A:1008706319665
Ortiz, J., Wolff, N., Anthony, K., Devlin, M., Lewis, S., & Mumby, P. (2018). Impaired recovery of the great barrier reef under cumulative stress. Science Advances, 4(7), eaar6127–eaar6127. https://doi.org/10.1126/sciadv.aar6127
Parsons, M., Bratanov, D., Gaston, K., & Gonzalez, F. (2018). UAVs, hyperspectral remote sensing, and machine learning revolutionizing reef monitoring. Sensors (Basel, Switzerland), 18(7), 2026–. https://doi.org/10.3390/s18072026
Researchers at Institute of Information Science & Technology Target Marine Science and Engineering (Remote Sensing for Maritime Prompt Monitoring). (2019). In Journal of Engineering (p. 1576–). NewsRX LLC.
Simic Milas, A., Cracknell, A., & Warner, T. (2018). Drones — the third generation source of remote sensing data. International Journal of Remote Sensing, 39(21), 7125–7137. https://doi.org/10.1080/01431161.2018.1523832
THOMPSON, A., & DOLMAN, A. (2010). Coral bleaching: one disturbance too many for near-shore reefs of the Great Barrier Reef : Coral reefs in a changing environment. Coral Reefs, 29(3), 637–648.
Turnbull, J. (2017, October 19). Innovations and technology to protect our Overseas Territories. Marine Developments. https://marinedevelopments.blog.gov.uk/2017/10/19/innovations-and-technology-to-protect-our-overseas-territories/
Vercelloni, J., Liquet, B., Kennedy, E., González‐Rivero, M., Caley, M., Peterson, E., Puotinen, M., Hoegh‐Guldberg, O., & Mengersen, K. (2020). Forecasting intensifying disturbance effects on coral reefs. Global Change Biology, 26(5), 2785–2797. https://doi.org/10.1111/gcb.15059
Wilhelm, T., Sheppard, C., Sheppard, A., Gaymer, C., Parks, J., Wagner, D., & Lewis, N. (2014). Large marine protected areas — advantages and challenges of going big. Aquatic Conservation, 24(S2), 24–30. https://doi.org/10.1002/aqc.2499
Williams, I., Couch, C., Beijbom, O., Oliver, T., Vargas-Angel, B., Schumacher, B., & Brainard, R. (2019). Leveraging automated image analysis tools to transform our capacity to assess status and trends on coral reefs. Frontiers in Marine Science, 6. https://doi.org/10.3389/fmars.2019.00222
