Family forest owners (FFO’s) play an important role in the complex interactions between development decisions and ecosystems. Determining the conditions of FFO engagement with natural systems will facilitate projections of future decision-making and subsequent ecosystem responses. This research seeks to establish agent functional types (AFT’s) of landowners as a form of dimension reduction, effectively assigning individual FFO’s to a particular decision-making class, each with unique behavior rules. By creating distinct classes of landowners, we will be able to model the evolution of (and between) classes over time.
To characterize the AFT’s we use a survey which was administered to New England FFO’s. The survey includes a choice experiment in which respondents indicate their willingness to cut their trees under various insect infestation scenarios. We use the landowner responses to the choice experiment, as well as stated demographics and motivations, to construct AFT’s in the form of a mixture model. Parcel- and town-level demographic and geographic data are then used to develop an AFT classification model. Ultimately, we want to be able to use parcel- and town-level information (which is freely available) to predict AFT's, and also to be used in conjunction with the classification to predict cutting behavior. This modeling framework provides a representation of the geographical, sociological, economic, and ecological drivers of human-land interaction in New England.
|University of Massachusetts Amherst|