Computer Simulations of Complex Ecological and Socio-Economic Systems

 

Alexey Voinov

(Erica Gaddis and David Landis)

 

Abstract

 

Most environmental management is conducted at the regional (state and county) scale and requires integration of natural and human components of landscapes. Management of landscapes, delineated by watershed boundaries, depends on analysis of the often complex dynamics and spatial relationships of ecological and socio-economic systems. Modeling, when used to conduct such analyses, is recognized as an effective decision support tool in environmental management. However, sophisticated modeling tools are typically not available or accessible for state and county employees to conduct comprehensive environmental simulations, due to the computer programming skills, computer power, and data collection effort required to develop and calibrate watershed models for new regions. The Gund Institute of Ecological Economics has developed the Landscape Modeling Framework (LMF), which allows researchers with minimal programming experience to develop and implement spatio-temporal to address complex management questions. An exciting aspect of this work is the ability to conduct optimization within the context of the existing framework. For example, we are working to implement a method to identify optimal spatial and temporal allocations of technologies and management practices for remediation of diffuse sources of nutrient and sediment pollution (agricultural, residential, and urban) in the short, medium, and long-terms. Management practices and technologies will be evaluated based on function (attenuation, buffering, or reduction) as well as cost and environmental impacts. The framework will be used in watersheds in the Lake Champlain Basin to maximize reduction of nutrients and sediment at a minimal cost while maintaining certain watershed properties of importance to the community. Goal functions will reflect the need to determine priority actions in the watershed based on available funds, water quality goals, community concerns, and ecological impacts. CS challenges include developing new optimization algorithms and linking existing optimization software (simulated annealing, genetic algorithm) with LMF, implementing LMF and its core Spatial Modeling Environment (SME) on a GRID-based execution engine to handle the computationally intensive aspects of the system, further developing the capabilities of a user interface for the SME, making it more accessible to researchers in other fields and decision makers, linking SME and GRASS (an open source GIS system) together and developing a new cyberinfrastructure that will allow the entire system to be launched through an interactive user-friendly web-interface, with intelligent data search and retrieval mechanisms, and module composition and recalibration methods.