Alistair G. B. Poore

Photographs of organisms taken by citizen scientists often contain a wealth of secondary data such as habitat composition information captured in the background. Computer vision has the potential to unlock these secondary data and increase the value of this resource for biodiversity monitoring. Here, we trained a computer vision model using approximately 4000 underwater fish photographs to detect the presence of the common kelp species Ecklonia radiata (hereafter Ecklonia ) in the background. The best performing model returned an average area under the precision-recall curve (AUPRC) value of 0.798 on the validation set. The model had an average precision of 0.79 across unseen test sets, meaning that approximately 4 out of every 5 Ecklonia detections were correct. However, the model had a relatively low recall of 0.52 across test sets, meaning it was only detecting half of the fish images that contained Ecklonia. When the best performing model was deployed on an unseen data set of 55000 iNaturalist photographs of fish, approximately 3900 secondary Ecklonia observations were obtained (after accounting for the false positive rate). This represents a 15-fold increase in the number of underwater photographs of Ecklonia in our study region when compared to iNaturalist observations where Ecklonia was the primary subject of the photograph. This study demonstrates how extracting secondary data using computer vision can considerably increase the information obtained from citizen science photographs.

Context: Diseases have affected coral populations worldwide, leading to population declines and requiring active restoration efforts. Aims: Describe population and individual impacts of necrotising disease in the Endangered octocoral Dendronephthya australis. Methods: We quantified population loss and recruitment by using reference photos, survey and GPS mapping and described disease lesions by using histopathology. Key results: From December 2019 to January 2020, we observed polyp loss, necrotic lesions and loss of large colonies of D. australis at Botany Bay, New South Wales, Australia. By September 2020, only a few scattered recruits remained, and all large colonies were lost. Histopathology of colonies sampled in January 2020 confirmed that the disease had resulted in necrosis, gastrovascular canal collapse and internal colony integrity loss, leading to mortality. New recruits were recorded within 10 months of disease onset, and large colonies within 18 months. Conclusions: Although the necrotising disease had significant impacts on both the individual and population level, natural recruitment began quickly. As such, unlike in other populations, restoration is not currently required in the Bare Island D. australis population. Implications: The extent of disease impact at the individual and population levels suggests that monitoring for lesions should be undertaken before developing conservation and restoration strategies for this species.

Vast numbers of nature photographs are shared online on various social media and citizen science platforms, providing an important and growing source of ecological data. In addition to the focal subject, which provides species occurrence data, photographs often contain secondary information such as habitat captured incidentally in the background. Here, we test whether the background of underwater photographs taken by recreational divers can be used to monitor reef habitat composition by comparing them to photoquadrats from standardised scientific surveys at three locations in Sydney, Australia. Furthermore, we assessed the effect of different photographic styles (wide‐angle vs. close‐up) and photographic subjects (fish vs. gastropods) on the habitat captured. Over 90% of underwater photographs displayed habitat in the background demonstrating the considerable amount of data available within this unutilised resource. Overall, the habitat composition captured in the background of photographs was broadly similar to that recorded by standardised survey quadrats. Although significant differences between photographs and photoquadrats were detected for most habitats, the differences were small. Furthermore, we found the difference between the two methods could be reduced by using subsets of photographs; for example, fish photographs detected kelp habitat at a similar rate to photoquadrats. Synthesis and applications: The background of underwater photographs from recreational divers contains valuable data that can be used to monitor habitat taxa that are typically under‐represented. To reduce the potential for bias we recommend restricting photographs to a single photographic subject (e.g. fish), using wide‐angle photographs only and obtaining at least 50 photographs per site from different divers when using recreational photographs to assess changes in habitat‐forming species.

Habitat complexity plays a critical role in shaping biotic assemblages and ecosystem processes. While the impacts of large differences in habitat complexity are often well understood, we know less about how subtle differences in structure affect key ecosystem functions or properties such as biodiversity and biomass. The late-successional seagrass Posidonia australis creates vital habitat for diverse fauna in temperate Australia. Long-term human impacts have led to the decline of P. australis in some estuaries of eastern Australia, where it is now classified as an endangered ecological community. We examined the influence of P. australis structural complexity at small (seagrass density) and large (meadow fragmentation) spatial scales on fish and epifauna communities, predation and sediment erosion. Fine-scale spatially balanced sampling was evenly distributed across a suite of environmental covariates within six estuaries in eastern Australia using the Generalised Random Tessellation Structures approach. We found reduced erosion in areas with higher P. australis density, greater abundance of fish in more fragmented areas and higher fish richness in vegetated areas further from patch edges. The abundance of epifauna and fish, and fish species richness were higher in areas with lower seagrass density (seagrass density did not correlate with distance to patch edge). These findings can inform seagrass restoration efforts by identifying meadow characteristics that influence ecological functions and processes.

Population characteristics (e.g. density and body sizes) of foundation species can affect their own persistence and provisioning of ecosystem functions. Understanding the drivers of population characteristics of foundation species at multiple spatial scales is therefore critical for maximizing ecosystem functions of restored habitats. We analyzed variation in population characteristics (densities, 95th percentile, and median lengths of live oysters) of the Sydney rock oyster, Saccostrea glomerata, on remnant oyster reefs at regional scales (among three estuaries) along an approximately 250 km of coastline in New South Wales, Australia. We then analyzed how population characteristics were further related to spatial attributes at smaller spatial scales including within‐patches (rugosity, distance to patch‐edge, and elevation), whole‐patches (size and shape), and among‐patch (connectivity) within each estuary. The densities and body sizes of S. glomerata were related to spatial attributes occurring within‐patch (e.g. elevation), whole‐patch (e.g. shape), and landscape (i.e. connectivity) scales, but these relationships varied among estuaries. The greatest variation in oyster density and size occurred at regional scales, suggesting that processes acting at larger spatial scales (e.g. water quality and/or climate) set the context for smaller scale influences on oyster characteristics. Our results highlight the potential importance of incorporating site‐specific, spatial attributes in the design of restored oyster reefs to maximize ecosystem services and functions provided by restoration efforts.

Citizen science is on the rise, with growing numbers of initiatives, participants and increasing interest from the broader scientific community. iNaturalist is an example of a successful citizen science platform that enables users to opportunistically capture and share biodiversity observations. Understanding how data from such opportunistic citizen science platforms compare with and complement data from structured surveys will improve their use in future biodiversity research. We compared the opportunistic fish photographs from iNaturalist to those obtained from structured surveys at eight study reefs in Sydney, Australia over twelve years. iNaturalist recorded 1.2 to 5.5 times more fish species than structured surveys resulting in significantly greater annual species richness at half of the reefs, with the remainder showing no significant difference. iNaturalist likely recorded more species due to having simple methods, which allowed for broad participation with substantially more iNaturalist observation events (e.g., dives) than structured surveys over the same period. These results demonstrate the value of opportunistic citizen science platforms for documenting fish species richness, particularly where access and use of the marine environment is common and communities have the time and resources for expensive recreational activities (i.e., underwater photography). The datasets also recorded different species composition with iNaturalist recording many rare, less abundant, or cryptic species while the structured surveys captured many common and abundant species. These results suggest that integrating data from both opportunistic and structured data sources is likely to have the best outcome for future biodiversity monitoring and conservation activities.

Global patterns of plant biomass drive the distribution of much of the marine and terrestrial life on Earth. This is because their biomass and physical structure have important consequences for the communities they support by providing food and habitat. In terrestrial ecosystems, temperature is one of the major determinants of plant biomass and can influence plant and leaf morphology. In temperate marine systems, macroalgae are major habitat‐formers and commonly display highly variable morphology in response to local environmental conditions. Variation in their morphology, and thus habitat structure on temperate reefs, however, is poorly understood across large scales. In this study, we used a trait‐based approach to quantify morphological variability in subtidal rocky reefs dominated by the algal genus Sargassum along a latitudinal gradient, in southeastern Australia (~900 km). We tested whether large‐scale variation in sea surface temperature (SST), site exposure, and nutrient availability can predict algal biomass and individual morphology. We found Sargassum biomass declined with increasing maximum SST. We also found that individual morphology varied with abiotic ocean variables. Frond size and intraindividual variability in frond size decreased with increasing with distance from the equator, as SST decreased and nitrate concentration increased. The shape of fronds displayed no clear relationship with any of the abiotic variables measured. These results suggest climate change will cause significant changes to the structure of Sargassum habitats along the southeastern coast of Australia, resulting in an overall reduction in biomass and increase in the prevalence of thalli with large, highly variable fronds. Using a space‐for‐time approach means shifts in morphological trait values can be used as early warning signs of impending species declines and regime shifts. Consequently, by studying traits and how they change across large scales we can potentially predict and anticipate the impacts of environmental change on these communities.

Aim: As urban areas continue to expand, it is increasingly important to quantify species‐specific responses to urban environments, and how these change across the full annual cycle. Our objective was to quantify urban tolerance for North American birds across the time. We tested (a) whether intra‐annual variability of urban tolerance differed between migrants and residents and (b) whether intra‐annual variability of urban tolerance was phylogenetically conserved. We then assessed how the relationship between ecological and life history traits and urban tolerance differed both across the year and between migrants and residents.Location: North America.Taxon: Birds.Methods: We integrated a large citizen science dataset of observations for 237 bird species, remotely sensed VIIRS night‐time lights data, and trait data on each species. We estimate, for each species and each month of the year, a continuous measure of urban tolerance (i.e. the median of their distribution of observations across an urbanization gradient). We then use phylogenetic linear models to assess the relationship between this measure of urban tolerance and various life history and ecological traits.Results: There was a distinct drop in the overall urban tolerance scores corresponding with the breeding period; this pattern was more pronounced for migrants compared to residents. Migrants also had greater intra‐annual variability than resident species. We also found that the strength of the relationships between ecological and life history traits and urban tolerance was highly seasonal for most traits considered, and some divergent patterns were noted between migrants and residents.Main conclusions: The urban tolerance of birds greatly changed throughout the annual cycle, with different patterns for migrants and residents. Compared to residents, migrants showed more intra‐annual variability of urban tolerance with a drop in the average urban tolerance score during the breeding season. Together, our results suggest that urban tolerance is a function of both species and season, and they highlight the importance of considering the dynamic nature of birds' use of urban ecosystems throughout the full annual cycle.