A visual analytics framework for conservation planning optimization
A novel visual analytics system that combines multicriteria analysis, optimization algorithms, and decision-making support to efficiently construct, compare, and modify conservation portfolios under multiple constraints.
Study link: Environmental Modelling & Software, Elsevier (2025)
Approximately 17% of global land is currently in some state of protection. Recent conservation research suggests the need for a drastic increase of protected lands by 2050. In order to reach this target, at least an additional 13% of lands need to be conserved or restored in a cost-effective and time-efficient manner in order to support the resiliency of our planet and its climate. While many individuals and foundations continue raising much-needed funding for the environment, the development of conservation portfolios is a complex multi-dimensional task.
Agencies have limited resources for investing in new conservation areas and have differing priorities for conservation in terms of species, land cover, human activities, etc. A novel interactive conservation portfolio development system is proposed, combining visualization, multicriteria analysis, optimization, and decision making that enables conservation planners and scientists to efficiently construct, compare, and modify conservation portfolios under multiple constraints. The system incorporates a multi-layer map view, a parallel coordinates attribute view, a control area for optimization modeling, and a multiple portfolio visualization for solution comparison.
This system can also be used in conjunction with AI agents to support automatic portfolio optimization by implementing a median ranking algorithm to allow parcel filtering based on an aggregated indicator of all the attributes combined with an integer programming model to generate land purchase recommendations within user-defined constraints and objective function. Multiple land portfolios can be generated and saved for comparison. Our system complements the existing body of tools by providing new visual, analytical, and mathematical features, while also allowing loading of a (shape compatible) conservation plan obtained with any other tool for further visual analysis.
Lead researcher: Rui Zhang, School of Computing and Augmented Intelligence, Arizona State University (ASU)
Contributing authors: Yafeng Lu (ASU), Katherine Adams (U Wisconsin), Jorge Sefair (ASU), Haley Mellin (Conserve), Miguel Acevedo (U Florida), Ross Maciejewski (ASU)