Collaborative Research: An innovative network to improve sea ice prediction in a changing Arctic

最近，北极海冰的范围、厚度和性质发生了重大变化，引起了各方的关注，并对各种利益攸关方构成了重大挑战，这些利益攸关方包括海上安全和安保、资源管理和开发、政治家、沿海社区、天气和气候预报员、气候变化研究人员以及越来越多的公众。季节性到年际时间尺度的海冰预报，特别是夏季和秋季的海冰预报，是特别有趣的。尽管每个利益相关者的优先级不同，但都需要改进对海冰状况的监测、预测和沟通。迄今为止，海冰模拟工作主要集中在气候尺度（即对温室气体强迫的响应）或有针对性的天气预报上，以支持导航。在季节尺度上预报海冰是一项挑战，因为：(1)大气和海洋影响的高度可变性；(2)初始化和验证的观测覆盖范围有限和/或不确定性高；(3)当前模式能力的局限性；(4)海冰可预测性的固有局限性；以及(5)北极系统的变化方式没有最近的历史先例。SEARCH海冰展望项目于四年前以一种特别的方式实施，要求自愿捐款，根据春末（6月1日）的情况估计9月份的海冰范围。使用不同的方法做出了贡献，从复杂的（部分和完全耦合的一般循环模型和统计关系）到基本的（趋势外推，启发式，民意调查）。《展望》将被组织和扩展为一个更加结构化、协调和正式的工作，重点是解决海冰预测的主要障碍，包括严格评估预测、协调和组织初始化的相关观测、评估方法，以及最后建立一个组织网络结构来管理工作并以新的方式交流结果。这项工作以过去四年的经验为基础，扩展现有的结构，利用国际范围内的资源和专业知识，帮助解决一系列被美国和国际项目和组织视为优先事项的挑战。知识价值-该项目促进了NSF？通过建立一个由科学家和利益相关者组成的创新网络来生成、评估和交流北极季节性海冰预报，为北极提供改进的预测工具。网络吗?S的重点是开发评估预测的新方法，综合和比较预测方法的新指标，以及利用目标观测初始化预测方法的新方法。最后，该网络将研究如何将不同的预测方法结合起来，以超越狭义方法的预测技能。研究小组将探索如何最好地规划观测，以改进季节性预测，以及这些预测如何反过来促进对北极海冰覆盖不断变化状态的理解。更广泛的影响——为了社会利益，将开发和传播改进的北极海冰季节性预报。通过培育和协调一个国际研究人员网络，并利用广泛的活动，该项目为利益相关者提供量身定制的信息。的需求。除了对利益相关者信息需求进行严格评估外，该网络还将为关键海冰变量和预测因子制定一个共同参考框架，为科学界和利益相关者生成综合数据集和预测领域。还将开展活动，向日益壮大的公民科学家网络社区提供信息和参与，其中许多人已经对北极科学表现出浓厚的兴趣，有能力进行自己的原创和潜在的有价值的分析，并与更广泛的媒体网络和公众讨论联系起来。这些活动将包括研究生和博士后，他们将在基础研究和应用研究的交叉点进行高质量的研究。通过将更多的年轻研究人员纳入该网络，该项目还满足了对合格专家的重要需求，这些专家可以帮助解决有关北极海冰快速变化对资源利用和生态系统服务影响的紧迫问题。

Recent major changes in the extent, thickness and properties of arctic sea ice have captured attention and posed significant challenges to a diverse group of stakeholders, ranging from maritime safety and security, resource management and development, politicians, coastal communities, weather and climate forecasters, climate change researchers, and a growing segment of the general public. Sea ice forecasting on seasonal-to-interannual timescales, especially over the summer and into fall, is of particular interest. Though each stakeholder is driven by different priorities, all require improved monitoring, prediction, and communication of sea ice conditions. To date, sea ice modeling efforts have largely focused on climate scales (i.e., response to greenhouse gas forcing) or targeted synoptic forecasting in support of navigation. Sea ice forecasting on seasonal scales is a challenge because of: (1) high variability in atmospheric and oceanic influence, (2) observations for initialization and validation have limited coverage and/or high uncertainties, (3) limitations of current model capabilities, (4) inherent limitations in sea ice predictability, and (5) an Arctic system changing in ways without recent historical precedent.The SEARCH Sea Ice Outlook was implemented four years ago in an ad hoc fashion, requesting voluntary contributions to estimate September sea ice extent based on late spring (June 1) conditions. Contributions have been made using different methods that vary from complex (partially- and fully- coupled general circulation models and statistical relationships) to basic (trend extrapolation, heuristic, public poll). The Outlook will be organized and expanded into a more structured, coordinated and formal effort that focuses on tackling key barriers to sea ice forecasting, including rigorous evaluation of predictions, coordination and organization of relevant observations for initialization, evaluation of methods, and finally, an organizational network structure to manage the efforts and communicate results in new ways. This effort builds on the experience of the past four years and expands on structures already in place, leveraging resources and expertise at an international scale to help address a set of challenges recognized as priorities by a range of U.S. and international programs and organizations.Intellectual Merit - This project advances NSF?s goal of providing improved predictive tools for the Arctic by creating an innovative network of scientists and stakeholders to generate, assess and communicate arctic seasonal sea ice forecasts. The network?s focus is to develop new methods to evaluate forecasts, new metrics for synthesis and comparison across forecast methods and new approaches to initialize forecast methods with targeted observations. Finally, the network will investigate how different forecasting methods can be combined to exceed predictive skills of narrower approaches. The research team will explore how best to plan observations for improved seasonal predictions and how these predictions in turn can advance understanding of the evolving state of the arctic sea ice cover.Broader Impacts - Improved seasonal forecasts of arctic sea ice will be developed and disseminated for societal benefit. By fostering and coordinating an international network of researchers and leveraging a broad range of activities, the project provides information tailored to stakeholders? needs. In addition to rigorous evaluations of stakeholder information needs, the network will develop a common reference framework for key sea ice variables and predictors, generating integrated datasets and predicted fields for the scientific community and stakeholders. Activities will also be initiated to inform and engage the growing internet communities of citizen scientists, many of whom already show a strong interest in arctic science, have capabilities for their own original and potentially worthwhile analysis, and connect with wider networks of media and public discussion. The activities will include graduate students and post-docs who are mentored to conduct high-quality research at the intersection of fundamental and applied research. By entraining more young researchers into the network, the project also addresses an important need for qualified experts that can help address urgent questions concerning resource uses and ecosystem services impacted by rapid arctic sea ice change.