SECURE- network for modelling environmental change

SECURE是一个由统计学家、建模师和环境科学家组成的网络，我们的目标是就如何描述和量化环境变化达成共同愿景，以帮助决策。了解和预测环境变化对于制定减轻未来事件影响的战略至关重要。围绕环境变化的沟通和决策有时会受到证据的重要性、不确定性的性质和大小以及如何描述两者的问题的困扰。环境变化的证据来自许多来源，但这一提议的关键是数据(来自观测、监管监测和地球观测平台，如卫星和移动传感器)和模型(过程和统计)的最佳使用。强大和可靠的证据基础是决策过程中的关键，它由强大的统计模型和最佳数据提供信息。该提案将提供支持决策的统计工具。许多与变化有关的环境挑战要求开发统计建模和推理工具，以了解驱动因素和系统反应，这些反应可能是直接或间接的，并通过反馈和滞后联系在一起。随着新技术(例如，提供高分辨率数据流的传感器网络)的发展和更广泛的可获得性，环境数据的特征正在发生变化。新兴的传感器技术能够以前所未有的规模和规模提供增强的环境系统动态细节。通过公民科学，公众对环境科学的参与度也在不断提高。越来越多地使用公民科学观测站将带来新的统计挑战，因为这种观测的抽样基础很可能是优先的，不直接的，质量不同，收集工作也不同。融合不同的数据流将是具有挑战性的，但在向社会和监管机构通报变化方面是必不可少的。在环境领域，不同数据源的联系以及处理大数据的挑战在于汇集各种高吞吐量的数据源，分析、汇总信号和模型，并最终使用数据模型系统处理复杂和不断变化的环境变化问题，以支持决策。成功的关键在于产生可消化的产出，这些产出可以在学术界、政策制定者和其他利益攸关方中传播和批评。在气候变化、粮食安全、生态系统复原力、可持续资源利用、灾害预警和灾害管理方面，出现了新的大量数据来源，包括群众来源的数据流，它们在数据管理、综合、通信和实时决策支持方面提出了问题和尚未开发的机会。我们的研究将涉及：改进与不确定性和可变性有关的建模和通信工具，这些工具普遍存在于许多环境数据源中；开发和扩展处理多尺度问题的建模能力，特别是在数据流的不同空间和时间尺度上进行整合，以及模型输出的导出时间尺度；探索应用于环境变化问题的最新统计创新的力量和局限性，最后反思可视化和交流的新技术。

SECURE is a network of statisticians, modellers and environmental scientists and our aim is to grow a shared vision of how to describe and quantify environmental change to assist in decision making. Understanding and forecasting environmental changes are crucial to the development of strategies to mitigate against the impacts of future events. Communications and decision making around environmental change are sometimes troubled by issues concerning the weight of evidence, the nature and size of uncertainties and how both are described. Evidence for environmental change comes from a number of sources, but key to this proposal is the optimal use of data (from observational, regulatory monitoring and earth observations platforms such as satellites and mobile sensors) and models (process and statistical). A robust and reliable evidence base is key in the decision making process, informed by powerful statistical models and the best data. This proposal will deliver the statistical tools to support decision making. Many environmental challenges related to change require statistical modelling and inferential tools to be developed to understand the drivers and system responses which may be direct or indirect and linked by feedback and lags. The character of environmental data is changing as new technologies (e.g. sensor networks offering high resolution data streams) are developed and become more widely accessible. Emerging sensor technology is able to deliver enhanced dynamic detail of environmental systems at unprecedented scale and . There is also an increasing public engagement with environmental science, through citizen science. Increasing use of citizen science observatories will present new statistical challenges, since the sampling basis of such observations will most likely be preferential and not directed, be of varying quality and collected with different effort. Fusion of the different streams of data will be challenging but essential in terms of informing society and regulators alike about change. Linkage of the different data sources, and the challenges of dealing with big data, in the environmental sphere lie in drawing together diverse, high-throughput data sources, analysing, aggregating and integrating the signals with models and then ultimately using the data-model system to address complex and shifting environmental change issues in support of decision making. Key to success lies in generating digestible outputs which can be disseminated and critiqued across academia, policy-makers and other stakeholders. In climate change, food security, ecosystem resilience, sustainable resource use, hazard warning and disaster management there are new high-volume data sources, including crowd sourced streams, which present problems and untapped opportunities around data management, synthesis, communication and real-time decision-support.Our research will involve: improving modelling and communication tools concerning uncertainty and variability, which are ubiquitous in many environmental data sources; developing and extending modelling capabilities to deal with multi-scale issues, specifically integrating over the different spatial and temporal scales of the data streams, and the derived timescales of model outputs; exploring the power and limitations of recent statistical innovations applied to environmental change issues and finally reflecting on new technologies for visualisation and communication.