Agent-based modeling of incentives to encourage pre-disaster relocation in anticipation of coastal flooding

飓风卡特里娜和桑迪等极端风暴清楚地表明，由于洪水而被迫搬迁是一个主要的、日益严重的问题。与短期疏散相比，人们对如何管理长期搬迁知之甚少。广泛的社会科学研究证明了自愿搬迁与强制搬迁的优势。鼓励在灾难性洪灾发生前自愿撤离有风险的沿海地区的战略可以减少受影响的人数、由此造成的破坏和财产损失以及紧急救济的费用。此外，海平面上升的相对可预测性使其有可能提前很好地了解哪些地理区域最有可能面临日益严重的洪灾危险。这种可预测性有助于根据联邦紧急事务管理署的建议，进行前瞻性规划和预期搬迁，作为减少洪水破坏预计增加的一种方式。然而，对海平面上升的许多反应可能是有问题的。预计整个社区的搬迁成本极高(约为每人100万美元)。此外，许多社区和家庭规模的防洪努力的有效性可能会受到限制。例如，洪灾后买断被洪水破坏的财产并不能防止破坏和个人财产的损失；同样，海堤等保护性屏障往往鼓励高风险地区的开发。综上所述，这些观察结果表明，值得探索鼓励在严重洪灾之前自愿搬迁的方法。我们研究了在洪灾前鼓励搬迁的可能方法，重点关注两个易受沿海洪灾影响的地区(纽约州布鲁克林和德克萨斯州加尔维斯顿)，这些地区的住房成本不同，具有社会经济/人口特征。许多考察脆弱沿海地区风险缓解的研究采用了单一决策者(典型居民或地方政府)的观点，而没有考虑不同的利益相关者的目标、价值观或偏好。在这项工作中，博弈论被用来模拟政府和普通公民之间的互动，以确定政府可以激励居民在其他情况下搬迁之前进行搬迁的解决方案。基于代理的建模用于增强游戏的真实性，解决决策者的异质性，并探索可能难以分析建模的行为(例如受其他居民行为影响的紧急网络效应)。到目前为止，很少有研究人员将搬迁问题作为一个可以使用博弈论等正式方法解决的问题来处理。拟议的项目明确考虑了涉及搬迁决策的利益相关者(政府和私人)的不同贴现率，从而可以将问题框定为多人博弈，而不仅仅是单个决策者的决策问题。博弈论的观点有助于确定“双赢”解决方案，从而使政府能够比其他方式更早地鼓励搬迁。敏感性分析将解决洪水风险、网络效应(如级联或倾倒)、搭便车效应(例如，利用政府补贴但即使没有补贴也会搬迁的居民)以及居民决策的不同行为模式的不确定性。该项目有助于提高社会对减少海平面上升对沿海洪灾的脆弱性的机制的理解。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Extreme storms such as Hurricanes Katrina and Sandy have made clear that forced relocation due to floods is a major and growing problem. By contrast to short-term evacuation, relatively little is known about how to manage long-term relocation. Extensive social-science research has documented the advantages of voluntary vs. forced relocation. Strategies for encouraging voluntary retreat from at-risk coastal areas prior to catastrophic flooding can reduce the number of people affected, the resulting disruption and property loss, and the costs for emergency relief. Moreover, the relatively predictable nature of sea-level rise makes it possible to know well in advance which geographic areas are most likely to face increasing flooding hazards. This predictability facilitates proactive planning and anticipatory relocation as one way of reducing projected increases in flood damage, as recommended by the Federal Emergency Management Agency. However, many responses to sea-level rise can be problematic. Anticipatory relocation of entire communities is extremely costly (roughly $1 million per person). Furthermore, the effectiveness of many community and household-scale flood-protection efforts can be limited. For example, buying out flood-damaged properties after a flood does not prevent disruption and loss of personal property; similarly, protective barriers, such as seawalls, frequently encourage development in at-risk areas. Taken together, these observations suggest that it is worth exploring methods to encourage voluntary relocation before severe flooding. We study possible ways to incentivize relocation prior to flooding, focusing on two areas vulnerable to coastal flooding (Brooklyn, NY, and Galveston, TX) with different housing costs and with socioeconomic/demographic characteristics. Many studies examining mitigation of risks in vulnerable coastal areas adopt the viewpoint of a single decision maker (either a prototypical resident, or a local government), without considering different stakeholder goals, values, or preferences. In this work, game theory is used to model interactions between government vs. private citizens, to identify solutions whereby government can incentivize residents to relocate before they otherwise might. Agent-based modeling is used to enhance the realism of the game, address decision-maker heterogeneity, and explore behavior that may be difficult to model analytically (such as emergent network effects influenced by the behavior of other residents). Few researchers to date have approached the issue of relocation as a problem that can be addressed using formal methods such as game theory. The proposed project explicitly considers the different discount rates of stakeholders involved in relocation decisions (government vs. private individuals), making it possible to frame the problem as a multiplayer game rather than merely a decision problem for a single decision maker. A game-theoretic perspective facilitates the identification of “win-win” solutions whereby government can encourage relocation sooner than would otherwise occur. Sensitivity analyses will address uncertainty about flooding risks, network effects (such as cascading or tipping), free-rider effects (e.g., residents who take advantage of government subsidies but would have relocated even without subsidies), and differing behavioral models of resident decision-making. The project contributes to improved societal understanding of mechanisms for reducing vulnerability to coastal flooding from sea-level rise.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.