Characterising human agents in the Mekong region

Abstract

Simulating complex social-ecological processes in an empirical situation faces the challenge of specifying realistic assumptions on system variables. In particular, the characterisation of human behaviour and behavioural responses to perceived changes requires careful implementation. This paper presents methodology developed for an agent-based model for the wider Mekong region, that is, the five countries that Mekong river flows through: China's Yunnan province, Laos, Thailand, Cambodia and Vietnam. The model aims to explore potential side-effects of major development investments, such as mainstream dams, monoculture plantations and large-scale irrigation projects; as well as adaptation strategies to exogenous drivers such as climate change and foreign investment flows. Linking such exogenous and top-down decisions to potential outcomes on the ground, for example, changes in poverty levels, requires modelling of household behaviours that relate to relevant livelihood strategies. Considering the size of the system at stake, inhabited by 270 million people and covering an area of 2.3 million kilometres square, characterisation and parameterisation of human agents and their behaviour requires robust methodology. This paper outlines a mixed resolution approach for the model design and discusses methodology employed to inform model assumptions with real-world data.