AAG 2020 Garrison Award

Type: Paper
Sponsor Groups:
Poster #:
Day: 4/7/2020
Start / End Time: 5:35 PM / 6:50 PM
Room: Agate C, Hyatt Regency, Third Floor
Organizers: The American Association of Geographers
Chairs: Candida Mannozzi


Integrating Complex Systems Theory, Geographic Information Science, and Network Science for Modelling Geospatial Phenomena

Taylor Anderson
Spatial Analysis and Modeling Lab, Simon Fraser University, Burnaby BC V5A1S6, CAN

Keywords: geographic information science, complex systems science, network science, geographic automata systems, forest insect infestation, epidemiology

Complexity theory seeks to understand complex systems by examining how sets of interactions between individuals scale-up to form emergent system-level behavior and patterns. This approach is operationalized by cellular automata (CA) and agent-based models (ABM) that also include contexts of geographic space and time, referred to as geographic automata systems (GAS). Both CA and ABM represent local interactions between neighboring cells or discrete agents, respectively, from which patterns at large spatial scales emerge. Similarly, network models represent sets of interactions as discreate links between pairs of nodes that form an emergent network structure. Separately, GAS and network-based approaches offer unique advantages for representing and analyzing spatio-temporal complex systems, however the two are rarely integrated. Therefore, the purpose of this study is to explore the variety of ways GAS and network science can be integrated to develop a suite of network-based GAS approaches including: a geographic network automata (GNA) modelling framework for representing complex spatial systems as evolving spatial networks; a network-based ABM (N-ABM) modelling framework that leverages networks to analyze and communicate ABMs, and a network-based validation approach for testing network-based GAS. Results demonstrate that integrating GAS and network science offers new means for the representation, analysis, communication, and testing of GAS models and the complex systems they represent. The developed methodologies and results in application to insect infestation and epidemiology provide useful tools for forecasting and scenario-testing to enhance decision-making processes. This study contributes innovative geocomputational frameworks to the fields of geographic information science, GAS, and network science.


Type Details Minutes Start Time
Discussant Taylor Anderson Simon Fraser University 45 5:35 PM

To access contact information login