A Temporal Variant-Invariant Validation Approach for Agent-based Models of Landscape Dynamics Christopher Bone,* Bart Johnson, † Max Nielsen-Pincus, ‡ Eric Sproles* and John Bolte § *Department of Geography, University of Oregon † Department of Landscape Architecture, University of Oregon ‡ Institute for a Sustainable Environment, University of Oregon § Department of Biological and Ecological Engineering, Oregon State University Abstract Agent-based modeling provides a means for addressing the way human and natural systems interact to change landscapes over time. Until recently, evaluation of simulation models has focused on map com- parison techniques that evaluate the degree to which predictions match real-world observations. However, methods that change the focus of evaluation from patterns to processes have begun to surface; that is, rather than asking if a model simulates a correct pattern, models are evaluated on their ability to simulate a process of interest. We build on an existing agent-based modeling validation method in order to present a temporal variant-invariant analysis (TVIA). The enhanced method, which focuses on analyzing the uncertainty in simulation results, examines the degree to which outcomes from multiple model runs match some reference to how land use parcels make the transition from one land use class to another over time. We apply TVIA to results from an agent-based model that simulates the relationships between land- owner decisions and wildfire risk in the wildland-urban interface of the southern Willamette Valley, Oregon, USA. The TVIA approach demonstrates a novel ability to examine uncertainty across time to provide an understanding of how the model emulates the system of interest. 1 Introduction Many of the world’s most pressing problems require insight into how the future may unfold given current and projected circumstances. This is especially true for coupled human and natural systems (CHANS) that are governed by land use decision-making and biophysical processes in concert with anthropogenic and natural disturbances (Matthews and Selman 2006; Liu et al. 2007; Bennett and McGinnis 2008). The interaction of these processes leads to positive and negative feedbacks that produce emerging landscape patterns over time. Due to their complexity, the need to project the future states of CHANS is often incompatible with traditional modeling approaches that ignore the influence of feedback mechanisms on dynamic system properties. Conversely, computer simulations driven by agent-based modeling (ABM) provide a suitable venue for simulating CHANS because the bottom-up nature in which human-decision making influences landscape dynamics can be explicitly represented. ABM is a Address for correspondence: Christopher Bone, Department of Geography, University of Oregon, 107D Condon Hall, Eugene, Oregon 97403, USA. E-mail: cbone@uoregon.edu Acknowledgements: This research was funded in part by the National Science Foundation Dynamics of Coupled Natural and Human Systems program, Grant 0816475. The Envision model used is the joint effort of a larger research team under this award. Critical compo- nents and contributions to this modeling effort were made by (in alphabetical order): Alan A. Ager, Dominique Bachelet, John P. Bolte, Allan Branscomb, Scott D. Bridgham, Stewart Brittain, David Conklin, Chris Enright, Cody Evers, Peter J. Gould, David W. Hulse, Con- stance A. Harrington, Jane A. Kertis, Ronald P. Neilson, Homero Penteado, Robert G. Ribe, Timothy Sheehan, JamesSulzman, Nathan D. Ulrich, and Gabriel I. Yospin. Research Article Transactions in GIS, 2013, ••(••): ••–•• © 2013 John Wiley & Sons Ltd doi: 10.1111/tgis.12016