This essay argues that much of the concern over issues of scale in the modeling of complex human-environment systems – of which integrated assessment models are a special case – tends to be preoccupied with bottom-up aggregation and top-down disaggregation. Deep analysis of the underlying explanation of scale is missing. One of the intriguing propositions of complex systems theory is the emergence of new structures at a high level of scale that are difficult if not impossible to predict from constituent parts. Emergent properties are not the mysterious creation of ‘‘new material’’ in the system, but rather the placement of the components of the system into their logical contexts (scales) so that the observer=modeler can see structures arise from them for the first time. The stochastic interaction among low-level elements that gives rise to emergent properties may be part of a larger process of self-organization in hierarchical systems. Self-organization and attendant emergent properties constrain low-level elements through a network of downwardly propagating positive feedbacks. Those feedbacks not only tend to hold the system in a temporary stable state, but they also render it vulnerable to radical reorganization by rapid external forcing. The vulnerability of the USA agricultural production system to climate change is given as an example of how a self-organizing, hierarchical system paradoxically may become susceptible to large external shocks as a result of the emergence of high-level structures that seek to protect its low-level components from shortterm variability. Simulations of changes in Honduran maize production in the aftermath of Hurricane Mitch using the CLUE land use model demonstrate the influence of multi-scale complexity on the resilience of land use after disturbance. Finally, it is argued that improved understanding of emergent properties of scale may give fundamental insight into the conditions of surprise.
