Recent studies have investigated connections between natural systems and related human impacts, such as those between the rainfall-runoff process and agricultural activities. This is primarily performed by using two-way connected, natural-human computer models. However, these coupled models require many parameters, which results in larger prediction uncertainty, especially when the observed data is limited. This study quantifies the relationship between model output uncertainty and model complexity. We found that, depending on the type of model used, the uncertainty of model outputs will likely increase with model complexity; however, active two-way feedback between natural and human systems could offset the impact of the natural system on output uncertainty by increasing human variations.
Read a representative journal paper on this line of research