Great Barrier Reef Aesthetics Indicator Study 2017 Ratings

Online survey data from Australian residents, distributed nationally, in October 2017. Rating scores of aesthetic beauty are provided for 181 coral reef images that featured a set of specific attributes that were tested as correlates of aesthetic beauty ratings.\nLineage: Our aim was to provide rigorous testing of potential indicators for use in GBR reporting, monitoring and management. We chose the five factors that were both prominent amongst responses and amenable to quantification as part of a monitoring program, and tested for their influence on aesthetic ratings through regression analysis. They included: (i) coral cover, (ii) coral pattern, (iii) coral topography, (iv) fish abundance, and (v) visibility. (Coral pattern is likened to ‘coherence’ within the field of landscape aesthetics where a harmonious arrangement within a landscape composition, such as unity in colour, shape, or texture exists; Stamps, 2004; Rosley, Rahman, and Lamit, 2014). \nWe then identified 181 underwater coral reef photographs from those that were publically available (www.gbrmpa.gov.au) or existed in the combined image libraries of the study authors. They represented typical underwater images from the GBR, with a common oblique perspective taken from approximately 5-10 m above a coral substrate. This perspective characterised the image that a person would see as soon as they placed their head beneath the water, and it was similar to the visual perspective used in monitoring surveys conducted by manta-towing at the Australian Institute of Marine Sciences. Some photos were duplicated and placed randomly, and some were modified using photo editing software to manipulate one feature independent of others, for the purposes of ‘checking’ the consistency and subtleties associated with making aesthetic judgements. Each photo was rated for each of the five factors (on a scale of low, medium, high) by members of the research team with experience in coral reefs. Given that there were insufficient photos representing abundant fish and poor visibility, a total of 20 photos were manipulated to enhance or de-emphasise certain factors. These photos ensured that we could attribute differences in aesthetic appeal of each photo to at least one of the five factors. The final set of photos represented realistic coral reef images across all five factors, with a greater representation of images containing moderately high coral cover to capture the nuances across the scale of potential ratings and also to aide engagement during online rating sessions. \nOnce 181 photos were identified and rated for each of the five factors, they were delivered to an online service provider with the ability to reach a large number of random, representative Australians (www.pollinate.com.au). A survey was constructed to collect simple demographic information about each participant, the self-rated level of interest in coral reefs, and aesthetic ratings for each photo on a scale of 1-10 (where 1=extremely unattractive, and 10=extremely attractive). Once an individual agreed to partake in the survey, they were sent a survey with 50 photographs randomly chosen from the pool of 181 photographs. We were concerned that the quality of responses would be affected if more than 50 photos were viewed (where 50 photos represented a ten-minute survey). The style of the survey was not dissimilar from very popular online games in which individuals are asked to rank aesthetic preferences of fashion or interior design items.\nA total of 1,417 individuals participated in the study, where each photo was rated at least 380 times on the ten-point scale. Twenty-nine percent of the sample came from Queensland, and 71% were distributed across Australia. Some 62.3% of people came from Metropolitan Australia, whilst 37.7 came from rural/regional Australia. Some 51.4% were female. Participants represented a range of experiences with the Great Barrier Reef, where 7.2% had never visited, and 7.9% did not find coral reefs that interesting. Most participants (99.6%) were not part of a GBR based club or community groups, such as a spear-fishing club. The average age for the sample population was 46.96 (standard error=0.471), and ranged from 16 to 89.

本数据集为2017年10月面向澳大利亚全国居民收集的在线调查数据。数据集包含181张珊瑚礁图像的美学美感评分，这些图像均具备一组经测试可作为美感评分关联因素的特定属性。

数据集溯源：本研究旨在为大堡礁（Great Barrier Reef, GBR）的报告、监测与管理工作提供严谨的潜在指标验证。我们选取了在既往响应中占比突出且可通过监测项目实现量化的五项因素，通过回归分析检验其对美感评分的影响。这五项因素分别为：(i) 珊瑚覆盖率，(ii) 珊瑚格局，(iii) 珊瑚地形，(iv) 鱼类丰度，(v) 水体能见度。（注：珊瑚格局在景观美学领域可类比为“连贯性”，即景观构图中存在色彩、形状或纹理统一的和谐布局；Stamps, 2004; Rosley, Rahman, and Lamit, 2014）

随后，我们从公开渠道（www.gbrmpa.gov.au）或研究作者联合图像库中筛选出181张水下珊瑚礁照片。这些照片均为大堡礁典型水下图像，采用统一的斜拍视角，拍摄位置距珊瑚底质约5-10米，该视角契合人类潜入水下后即刻所见的视野，与澳大利亚海洋科学研究所（Australian Institute of Marine Sciences）开展的蝠翼拖曳监测调查所用视觉视角一致。部分照片经复制后随机置入样本集，另有部分照片通过图像编辑软件独立调整单一项特征，以验证美学判断的一致性与细微差异。

研究团队中具备珊瑚礁研究经验的成员，按照低、中、高三档对每张照片的五项因素分别进行评分。鉴于具备高鱼类丰度与低能见度特征的照片样本不足，我们额外调整了20张照片以强化或弱化特定因素，确保可将每张照片的美学吸引力差异归因于至少一项上述五项因素。最终的照片集合覆盖了五项因素的全部组合，其中中等偏高珊瑚覆盖率的图像占比更高，以覆盖评分尺度内的细微差异，并提升在线评分环节的参与度。

在完成181张照片的五项因素评分后，我们将数据集交付至可触达大量随机、具有代表性的澳大利亚居民的在线服务平台（www.pollinate.com.au）。调查问卷收集参与者的基础人口统计学信息、自我报告的珊瑚礁兴趣程度，并要求参与者以1-10分对每张照片的美学吸引力进行评分（1=极不美观，10=极美观）。参与者同意参与后，将获得随机选取的50张照片组成的问卷。考虑到若查看超过50张照片可能影响作答质量，50张照片对应的作答时长约为10分钟。问卷形式与广受欢迎的在线游戏类似，此类游戏通常要求参与者对时尚或室内设计物品的美学偏好进行排序。

本研究共招募1417名参与者，每张照片的10分量表评分至少获得380次有效作答。样本中29%来自昆士兰州，其余71%分布于澳大利亚其他地区；62.3%的参与者来自澳大利亚大都会区域，37.7%来自农村/区域区域；51.4%为女性。参与者的大堡礁游览经历存在差异：7.2%从未到访过大堡礁，7.9%表示对珊瑚礁兴趣平平。绝大多数参与者（99.6%）未参与大堡礁相关的俱乐部或社区团体，例如渔猎俱乐部。样本的平均年龄为46.96岁（标准误=0.471），年龄跨度为16至89岁。