From imperfection to inference: issues of scale and uncertainty in global change biology

Invited seminar to Department of Environmental Science, Policy, and Management colloqium at UC Berkeley. File includes animations that are only visible when viewed with Adobe Acrobat. <strong>Abstract</strong> Integrated research across biological scales is vital to understand and anticipate the impacts of global change on ecosystem function and biodiversity. There is now a plethora of ecological, natural history, climate, and remotely sensed data available to address this important issue. Integrating these disparate (and often ‘messy’) datasets into models across spatial, temporal, and biological scales is tantalizing but perilous. Using examples ranging from fire ecology in South Africa to three-toed woodpecker distributions in North America to high resolution maps of global cloud cover, I will discuss several important challenges that must be addressed to rigorously harness these data streams (and keep track of their uncertainties) for scientific progress and prudent decision-making.

应邀在加州大学伯克利分校（UC Berkeley）环境科学、政策与管理系的学术专题讨论会上发表特邀报告。本文件包含仅可通过Adobe Acrobat查看的动画内容。

**摘要**
跨生物尺度的整合研究，对于理解并预判全球变化对生态系统功能与生物多样性的影响至关重要。当前已有海量生态学、自然史、气候学及遥感数据，可用于解决这一关键科研议题。将这些异构（且往往较为"杂乱"）的数据集整合至跨空间、时间与生物尺度的模型中，虽极具吸引力却也暗藏风险。我将以南非火灾生态学、北美三趾啄木鸟分布格局，以及全球云量高分辨率地图为例，探讨为实现科学研究进展与审慎决策，严格利用这些数据流（并同步追踪其不确定性）所需攻克的若干核心挑战。