Improving use of citizen science data in ecological surveillance with hidden process models.

学生战略优先领域：数学生物学关键词：监测、统计学、生态系统健康、海洋哺乳动物、环境变化这个项目将探索在海洋动物生物学的背景下收集和分析监测数据的新方法。几个世纪以来，海洋动物一直滞留在英国海岸，但这些事件的原因尚不清楚。发现搁浅的动物往往会引发这样的问题：为什么会发生这样的事件，以及它们对我们海洋的健康意味着什么。长期积累搁浅数据可以调查搁浅人数、健康指数和死亡原因的趋势，并可提供必要的基线信息，以发现新出现的疾病、异常死亡事件和人为影响。然而，将搁浅记录外推到海上人口是具有挑战性的，因为报告的病例是生物、物理和社会过程的复杂功能。这项研究力求改进搁浅数据的统计建模技术，以便利将其用于监测，并减少在量化人类活动对海洋种群的影响方面的不确定性。它将询问可供英国和其他北海沿岸国家使用的搁浅数据库，以模拟合并多个数据源的搁浅时空模式，并研究这些新颖的见解如何有助于为管理层决策提供信息，并制定更强大的未来监测战略。这适用于更广泛的机会主义野生动物监测计划，并为监测和同意机构在该项目上的合作工作提供切实的好处。学生与提供数据可视化和集成培训。特别是，在使用一类区分观察模型和过程模型的一般潜在变量方法来分析不完美的数据方面，特别是在调查工作信息匮乏的不同公民科学计划方面。具有影响力的科学已经成为学术事业不可分割的一部分，这一学生将同时提供学术和非学术环境(SMASS)的培训，从而在这一领域比其他博士毕业生获得相当大的领先地位。公民科学数据集是环境监测信息的一个迅速扩大和严重未得到充分利用的来源，在未来的环境研究中肯定会变得更加重要，因为强大的统计和建模背景将打开职业发展机会。

Studentship strategic priority area: Mathematical BiologyKeywords: Surveillance, statistics, ecosystem health, marine mammals, environmental changeThis project will explore new ways of collecting and analyzing surveillance data within the context of marine animal biology. Marine animals have become stranded along UK coasts for centuries but the reasons for these events remain unclear. Discovery of a stranded animal often leads to questions regarding why such events happen, and what they indicate about the health of our oceans. Long-term accumulation of stranding data allows the investigation of trends in stranding numbers, indices of health and causes of mortality and can provide essential baseline information to detect emerging diseases, unusual mortality events, and anthropogenic impacts. However, extrapolating the strandings record to the at-sea population is challenging as reported cases are a complex function of biological, physical, and social processes. This study seeks to improve the statistical modelling techniques for strandings data to facilitate their use for monitoring and to reduce uncertainty in quantifying anthropogenic impacts on marine populations. It will interrogate the strandings databases available for the UK, and other countries bordering the North Sea, to model spatiotemporal patterns of strandings incorporating multiple data sources, and examine how these novel insights can help inform management decisions and develop more robust future monitoring strategies. This has application to wider opportunistic wildlife surveillance schemes and offers tangible benefit to the work of the monitoring and consenting bodies collaborating on this project.The studentship with provide training in data visualization and integration. In particular, in the use of a general class of latent variable approaches that distinguish between observation and process models to analyze imperfect data, and especially in relation to diverse citizen-science schemes where the information on survey effort is scant. Science with impact has become an inseparable part of academic careers, and this studentship will provide training simultaneously in both an academic and non-academic environment (SMASS), thereby acquiring a considerable head start in this area over other PhD graduates. Citizen science data sets are a rapidly expanding and seriously underexploited source of environmental monitoring information, certain to become considerably more important in future environmental research where a strong statistics and modeling background will open up career development opportunities.