Critical phenomena in human-environment systems

在我们面临环境挑战的时代，越来越明显的是，人类人口和环境系统相互影响，以至于我们可以将它们概念化为一个单一的、耦合的人-环境系统。人-环境耦合系统的研究正在迅速发展，但许多研究系统的重要问题仍未得到解答。例如，人类-环境系统何时会超越临界点，进入截然不同的社会或环境状态？这些新的州什么时候是有害的，也是有益的？是否有可能通过监测人类或环境系统的数据并寻找某些指标来预测这些临界点？我的研究计划将采取措施解决其中一些问题。该方法将开发四个人类环境研究系统的数学和计算机模型：(1)灰狼种群动态和人类对狼的扑杀，(2)森林病虫害的爆发和通过限制营地柴火活动的控制，(3)麻疹的爆发和通过接种疫苗进行控制，以及(4)社会网络上气候变化辩论的动态。这些模型将根据来自自然和人类系统的数据进行测试。然后将仔细分析这些模型，以了解它们预测的人类与环境系统之间的相互作用，以及对麻疹疫苗接种或猎杀灰狼等干预措施的预测反应。将对这些模型进行检查，看看它们是否以及何时显示出临界点，以及这些临界点是否可以通过早期预警信号提前检测到。一个特别的重点将是这些系统的空间方面，包括分析在线社交媒体中的社交网络，通过这些社交网络传播关于环境系统的想法、观点和情感。我们还将使用机器学习算法等数据科学方法来调查在线社交媒体数据和互联网搜索数据，从而了解模型预测的早期预警信号是否存在于任何现实世界的系统中。这些努力将得到新方法的补充，这些方法使用计算机算法直接从数据中“发现”模型，这与通常开发模型的方式形成了鲜明对比。这项研究将提高我们对人类-环境系统如何对干预做出反应的理解，当它们接近临界点时，以及我们是否可以对这些临界点提前发出警告。这项研究的现实意义和影响将来自于改善加拿大及其他地区人类及其环境的健康。通过监测在线社交媒体能够发现人类-环境系统初期临界点的高风险，将使我们能够采取措施防止或减轻临界点对可持续发展和健康的负面影响，或加速导致积极环境状态的积极临界点。

In our era of environmental challenges, it is becoming increasingly clear that human populations and environmental systems mutually influence one another, such that we may conceptualize them as a single, coupled human-environment system. The study of coupled human-environment systems is growing rapidly, but important questions remain unanswered for many study systems. For instance, when are human-environment systems pushed beyond tipping points into dramatically different social or environmental states? When are those new states harmful or beneficial? Is it possible to anticipate these tipping points by monitoring data from human or environmental systems and looking for certain indicators? My research program will take steps to address some of these questions. The methodology will develop mathematical and computer models of four human-environment study systems: (1) grey wolf population dynamics and culling of wolves by humans, (2) forest pest outbreaks and control through camp firewood movement restrictions, (3) measles outbreaks and control through vaccination, and (4) dynamics of the climate change debate on social networks. These models will be tested against data from both natural and human systems. The models will then be carefully analyzed to understand the interactions between human and environmental systems that they predict, as well as the predicted response to interventions such as vaccination for measles, or hunting for grey wolves. The models will be examined to see if and when they exhibit tipping points, and whether those tipping points can be detected in advance through early warning signals. A particular focus will be the spatial aspects of these systems, including analyzing the social networks in online social media through which ideas, opinions and sentiments about environmental systems propagate. We will also use data science approaches like machine learning algorithms to investigate online social media data and Internet search data, allowing us to see whether the early warning signals predicted by the models are present for any of the real-world systems. These efforts will be supplemented by novel approaches that use computer algorithms to "discover" models directly from data, which contrasts with the usual way that models are developed. This research will improve our understanding of how human-environment systems respond to interventions, when they are near tipping points, and whether we can obtain advance warning of those tipping points. The real-world significance and impact of the research will come from improving the health of humans and their environments in Canada and beyond. Being able to detect heightened risk of incipient tipping points for human-environment systems by monitoring online social media would enable us to take steps to prevent or mitigate tipping points with negative consequences for sustainability and health, or accelerate positive tipping points that lead to positive environmental states.