Modeling is a crucial approach for understanding the past and exploring the future of coupled human-natural systems. However, uncertainty in various forms challenges inferences from modeling results. Model-based support for decision-making has increasingly adopted an emerging exploratory approach. This approach addresses uncertainty explicitly through systematically exploring the implications of modeling assumptions, aiming to enhance the robustness of inferences from models. Despite a variety of applications, the extent and the way(s) that exploratory modeling can deal with the challenges that arise from the uncertainty and complexity of decision-making with stakeholders has not yet been systematically framed. We address this gap in two ways. First, we present a taxonomy of the ways that exploratory modeling can be used to inform robust inferences in coupled human-natural systems by mapping the technical capabilities of this approach in relation to the diversity of past applications. This subsequently guides an investigation of the practical benefits and challenges of these capabilities in handling uncertainty and complexity. Second, we discuss different ways for integrating genuine stakeholder engagement into exploratory modeling through transdisciplinary research. Finally we outline some priorities for future expansion of this research area.

建模是了解过去和探索人与自然系统耦合未来的关键方法。然而，各种形式的不确定性挑战了建模结果的推论。基于模型的决策支持已越来越多地采用一种新兴的探索性方法。该方法通过系统地探索建模假设的含义来明确解决不确定性，旨在增强鲁棒性模型的推论。尽管有各种各样的应用，但探索性建模可以应对与利益相关者进行决策的不确定性和复杂性所带来的挑战的程度和方式尚未得到系统地构架。我们通过两种方式解决这一差距。首先，我们通过分类探讨这种方法的技术能力（与过去应用程序的多样性有关），来探索探索性建模可用于在人与自然系统耦合中提供可靠推断的方法的分类。随后，这将指导对这些功能在处理不确定性和复杂性方面的实际好处和挑战进行调查。其次，我们讨论了通过跨学科研究将真正的利益相关者参与整合到探索性模型中的不同方法。